Abstract
In this paper, the effect of cellulose nanofibrils (CNFs) loading levels on the conventional and dynamic mechanical, morphological, thermal and rheological properties of the polyhydroxybutyrate (PHB) biopolymers were studied. According to the results, adding CNFs from 1% to 20% generally didn’t provide any improvement in the flexural, tensile and izod impact strength attributable to void formation and pulling out and agglomeration of nanofibrils in the matrix, which was observed during morphological characterization, however adding CNFs substantially increased both flexural and tensile modulus of elasticity. Thermal analysis showed that adding CNFs generally decreased degradation at high temperatures of the biopolymer nanocomposites (BNCs) The addition of CNFs at 1, 10 and 20% increased the E' and E'' of neat PHB but the other loadings decreased them, and tan delta increased with CNF loadings of 3, 5, 10 and 20%, and finally adding CNFs didn’t change the rheological behavior of the composites.
Keywords
Introduction
Plastic usage increases in the global industry because of advantages including easy processing, durability, and low-cost, etc. Many plastics are generally obtained from petroleum-based materials, which is not an environmental friendly and the accumulation in the nature of plastics such as bottles, bags and plastic wraps harm to the environment. 1 To resolve the environmental problems, Biopolymers chemically or naturally synthesized from different biological sources has been studied as an alternative material to petroleum based polymers. 2 Nowadays, biopolymers such as polyhydroxybutyrate (PHB), polylactic acid (PLA) commonly have been studied.3–5 PHB are a polyester polymer that is produced with bacteria from renewable raw materials.6–8 Nowadays, PHB attract much attention because of their advantages including being environmentally friendly, bio-compatible, and requiring lower processing temperatures. 9 But PHB has some disadvantages including brittleness and poor mechanical properties, and the PHB exhibits low thermal degradation at below 190 °C.10,11 The structural properties make them a challenging polymer to be processed and the status also limits to their industrial applications of PHB. 8 Therefore, various reinforcing materials such as cellulosic natural fibers, carbon fibers, synthetic fibers etc. have been used to improve the mechanical properties of neat PHB. The reinforcement of the biopolymer matrix by fibers can provide anisotropicity in the resulting composites and proper processing can contribute to the homogenous dispersion and uniform orientation of the fibers. 7
Cellulose nanofibrils (CNFs) is an important value-added fibers of the cellulosic natural materials and CNFs can increase the mechanical properties of the biopolymers attributable to aspect ratios. 12 They have various advantages such as environmental friendly nature, low cost and high mechanical properties. 13 The advantages make them a superior reinforcing filler to increase the mechanical strength of biopolymers. 14 Several papers have been published on PHB composites with natural fibers such as nanofibrillated cellulose, 13 cellulose nanowhiskers and cellulose nanocrystals,7,15–17 cellulose nanofibrils,4,5,18,19 lignin, 20 bamboo microfibrils, 21 cellulose,4,5,14,19–22 bacterial cellulose, 23 microcrystalline cellulose, 24 plant fibers.25–27 In the previous studies, effects of the nature fibers on characterization of the PHB films were studied at low loading levels, the adding nano fibers was found to form a fibril network in the matrix, and thus, especially modulus of the films can be improved with the adding nano scaled fibers, which act as nucleating agents in the biopolymer matrices, 19 and Barnes et al. 28 expressed that large surface area and hydroxyl groups of CNFs provided an improvement effect on the mechanical properties of the biopolymers. Zhang et al. 16 and Benini et al. 9 found that nanofibrils in cellulose improved the mechanical properties by wrapping the biopolymer matrix as a nanofibril network.
In our studies, the effects of cellulose nanofibrils at loading range from 1% to 20% on the conventional and dynamic mechanical, thermal, morphological and rheological properties of the PHB biopolymers were studied. As seen the previous studies, low loadings of the nano scaled fibers were used, but our study investigated the changes in the structure of the PHB matrix at medium and high loading levels of the nano fibers and this paper also revealed the viscoelastic and rheological properties of the PHB composites with high CNF loadings.
Materials and methods
Polyhydroxybutyrate (PHB) was provided from Good Fellow Inc. (Huntingdon, England). The dimension of CNFs were 5–200 nm (width) and 130–225 µm (length) (Figure 1), and obtained by the University of Maine’s Process Development Center. Motaung and Linganiso 29 found that the interfacial adhesion between fillers and biopolymer matrix can be improved with the adding coupling agent such as polypropylene maleic anhydride (MAPP) and therefore, MAPP as a coupling agent was also used at 3 wt. % loading level in an effort to improve the interactions occurring between the matrix and CNFs. Both biopolymer matrix and CNFs were dried in an oven at 50 °C for few days prior to processing.
SEM images of cellulose nanofibrils.
Preparation of the biopolymer nanocomposites
CNFs and dried PHB biopolymer were compounded according to Table 1.
Formulation of the biopolymer nanocomposites.
A Speed mixer was used to mechanically blend the CNFs with PHB for 5 min., and the samples were produced with a twin screw extruder (Brabender GmbH & Co) at 170 °C and at 65 rpm. After the extrusion, the blends were cooled, and then granulated with a lab type-cutter. They were dried at 50 °C during few days, and an injection molder were used to obtain mechanical test samples including flexure, tensile and Izod impact strength. Barrel and mold temperature, and pressure of the injection molder were set at 180 °C and 80 °C, and 17 MPa, respectively.
Scanning electron microscope
Scanning electron microscope (SEM) characterization of the samples were conducted with a Hitachi tabletop TM3000 with a voltage of 5–10 kV. The fractured section of the tensile specimens were used but the fractured section of the samples didn’t coated with any materials.
Mechanical properties
An Instron 8801 with 10 kN load cell was used in the tensile tests, and speed parameter during testing was selected as 5 mm/min as given in ASTM D 638-03 Type I. The elongation of the sample was determined with an extensometer and the data were used to calculate tensile modulus of elasticity. An Instron 8801 with 4.48 N load cell was used in the flexure tests according to ASTM D 790-03. The flexure parameters were set at support span of 50.8 mm and a test speed of 1.27 mm/min. The Izod impact tests were conducted with a Resil 50 B tester with 2.75 J hammer (Ceast Company). Ten test specimens were used in each nanocomposites for the mechanical tests, and the average values were used.
Thermogravimetric analysis
A thermal analyzer (TA) Q50 was used to determine the thermal stability of the samples (about 10 mg). Test parameters were conducted at a temperature range from 25 °C to 800 °C, a heating rate of 10 °C/min under nitrogen, and a flow rate of 20 ml/min. The temperatures in weight loss at 10% (T 10% ), 50% (T 50% ), and 85% (T 85% ), and maximum derivative thermal gravimetry (DTGmax), and total weight loss (WL) were determined by using the thermogravimetric analysis (TGA) curves.
Differential scanning calorimetry
A thermal analyzer (TA) Q2000 was used to differential scanning calorimetry (DSC) isotherms of the samples at heating rate of 10 °C/min under nitrogen, flow rate of 20 ml/min, heating range from 0 °C to 200 °C. Glass transition temperature (Tg), crystallization temperature (Tc) and melting temperature (Tm), crystallization (Hc), and melting enthalpy (Hm) were determined with DSC curves, and crystallinity (Xc) were also found according to following Equation.
ΔH
f
is the heat of fusion of the BNCs,
Dynamic mechanical analysis
The bending tests at 3 points with helping a TA Q800 dynamic mechanical analysis analyzer was conducted to determine the viscoelastic properties including storage (E') and loss modulus (E''), and tan δ. Test frequency was at 1 Hz under room conditions, and the strain amplitude was selected as 0.01% in the linear viscoelasticity. The sample dimensions were 5 cm x 1.24 cm x 0.2 cm (L, W, T). The heating range was from 25 °C to 80 °C at a scanning rate of 5 °C/min.
Rheological properties
A stress-controlled Bohlin Gemini rheometer (Malvern Instruments, UK) was used to determine the rheological behavior of the samples at 175 °C. To obtain the elastic (G′) and loss modulus (G″), and tan delta, small-amplitude oscillatory shear tests with parallel plate geometry were used and test parameters were set as the plate diameter of 25 mm and the plate gap setting of 1.0 mm. Frequency sweep test in the linear scale were also conducted at range from 10−2 to 102 Hz to find the complex viscosities (η*) of neat PHB and the BNCs. Steady shear measurements were conducted to determine the apparent viscosity at range from 10−2 to 101 (1/s).
Statistical analysis
Statistical analysis (one-way variance analysis (ANOVA) and Duncan test) was conducted to find the effects of CNF loading levels on the mechanical properties of the samples. In statistical analysis, significant level was selected as 95% (p < 0.05), and the groups which is statistically significant was showed as the letters such as A, B, C, etc.
Results and discussions
Morphological properties
The dispersion of CNFs in the matrix and morphological properties of the fractured tensile samples were investigated with SEM. As seen in Figure 2, the dispersion of CNFs was seen in the PHB matrix and they were generally observed to be homogenously embedded in the matrix, however some voids and pulling out of nanofibrils were visible in the matrix and some fibril aggregations were also observed in all the BNC samples in the SEM images, and the morphological imperfections also had a negative effect on the mechanical strength of the composites as seen in Table 2. Nevertheless, the adding CNFs improved the flexure and tensile modulus of elasticity for all the composites because of the reinforcing effects of CNFs having high modulus of elasticity values30,31 and the good fiber-matrix interactions. The improvement in stiffness moduli also increased with the high CNFs loadings, although the SEM images showed various agglomerated fibrils in the PHB matrix when the loading level of the CNFs was increased, especially at high loadings of 10 and 20 wt.%. The increases in stiffness were attributed to hydrogen bonding of the CNFs.32,33
SEM images for dispersion of CNFs in the BNCs.
The mechanical test results of the neat PHB and the BNCs.
In a related study, Miao and Hamad 34 found strong interfacial adhesion between the CNF and a biopolymer matrix and the strong interactions provided improvement in the composite stiffness, however, the interfacial adhesion between the fibers within the aggregates are expected to be much weaker because of an insufficient matrix binding effect. In this study, these aggregates between the CNFs are thus considered defects contributing to poor adhesion within the matrix. As such, the increased defects in the PHB matrix contributed to the lower tensile and flexural strength. In other studies, the increase of CNF content, the small and large voids, various holes because of fiber pull-out and agglomerated CNFs and single CNFs were also observed.12,16,18,35
Mechanical properties
The tensile, flexure and izod impact test results were analyzed with SPSS 16 software. ANOVA and Duncan tests were conducted on the mechanical properties of the samples. The obtained results are given in Table 2.
In this study, MAPP was used in an attempt to enhance the interactions between CNFs and PHB, and the addition of MAPP slightly decreased the mechanical properties of neat PHB. CNFs didn’t provide an improvement on the flexural and tensile strength expect for 20 wt. % CNFs attributable to embrittlement formed by some agglomeration of the CNFs in the PHB matrix. Figure 2 clearly shows that some CNFs agglomerated in the matrix. Whereas, various loadings of CNFs improved both flexure and tensile modulus of elasticity, and while the loading levels of CNFs increased at range from 1 wt. % to 20 wt. %, the improvement ratio in both flexural and tensile modulus of elasticity increased with the reinforcement effect of CNFs in the PHB matrix. The largest increases in the flexure and tensile modulus of elasticity was found as 17% and 28% for the biopolymer nanocomposites with 20 wt.% CNFs, respectively. Adding CNFs decreased the izod impact strength slightly, however, the changes in the izod impact strength was found to be statistically insignificant. As a result, the adding CNFs has an important effect on the elastic structure of the PHB, and all CNF loadings improved both flexural and tensile modulus of elasticity. According to the results, it can be said that the elastic properties of PHB can be improved with adding CNFs. According to the obtained results of studies conducted with fibers including bamboo microfibrils, 21 jute fibers9,36 and nanofibrillated cellulose, 12 the addition of the fibers provides that stress distributes homogenously between the matrix polymer and the microfibrils at low loadings but some fiber agglomerations occur in the matrix at higher loading as shown in Figure 2 and this causes weak interactions and poor adhesion in the matrix and the decrease in the strength of the composites occurs. Similar results on the mechanical properties of PHB composites with bamboo fibers 37 and natural fillers including eggshell flour, walnut shell flour, and tuff 25 were found.
Thermogravimetric analysis
The TG and DTG curves of neat PHB and the BNCs were given in Figures 3 and 4, respectively. The thermal curves of neat PHB and the BNCs showed a slight weight loss attributable to moisture content and later a main degradation between 200 °C and 300 °C (Figure 3). The addition of CNFs improved the degradation temperature (Td) of neat PHB. The onset Td of the BNCs decreased with the presence of CNFs.
TG curves of neat PHB and the BNCs.
DTG curves of neat PHB and the BNCs.
Figure 4 showed a second shoulder at 380 °C in the DTG curves. The shoulder raised with an increasing amount of CNF loading level. The first and the second degradation might be due to the thermal decomposition of biopolymer matrix and CNFs, respectively. 7 Table 3 shows the summary of TGA of the neat PHB and the BNCs. Similar results were found in another studies about thermal degradation of PHB nanocomposites with cellulose nanofibrils.4,12,16 As shown in Table 3, adding CNFs generally was determined to slightly increase the degradation temperature of T50% and adding CNFs provided important improvements in T90% of the PHB matrix, and it improved the thermal stability at T90% to range from 2% (for 1% CNFs) to 12% (for 20% CNFs). The adding CNFs generally provided an improvement on the DTGmax, but as seen in the weight loss, CNFs didn’t have any increase on it and the weight loss of the BNCs were the same according to Table 3. As a result, it can be said that adding CNFs improved the thermal stability at especially high temperatures (T90%) of neat PHB because of cellulose fibrils having high thermal stability as reported in a previous study. 38
The summary of TGA curves of the neat PHB and the BNCs.
Differential scanning calorimetry
Tg, Tc, Tm1 and Tm2 curves of the neat PHB and the BNCs were given in Figures 5 to 8, respectively. DSC was performed with two heating to obtain the Tm2. A seen in Figure 5, Tg of the samples was not noticeably altered by the adding various loading levels of CNFs and the Tg was similar for all the BNCs. However, the Tc decreased with addition of CNFs as shown in Figure 6. The samples generally exhibited similar Tm1 (164–167°C) in the first heating (Tmf) as given in Figure 7. PHB exhibited that two Tm peaks are generally attributable to melting of the crystals with various lamellar thicknesses 39 in the 2nd heating (Tms) as given in Figure 8. In this study, the samples showed two Tm peaks at 157–167°C and Tm2 of the BNCs were determined to decrease with presence of CNFs.
Tg of neat PHB and the BNCs.
Tc of the neat PHB and the BNCs.
Tm1 of neat PHB and the BNCs.
Tm2 of neat PHB and the BNCs.
According to Table 4, the Tg and Tc of the samples were similar to each other and Tg and Tc changed at a range between 2 and 3 °C, and 111 °C and 115 °C, respectively. Tmf of the neat PHB and the BNCs measured with the 1st heating were found to be similar and in 2nd heating, neat PHB showed two Tm peaks (Tm1 and Tm2), which have a characteristic property but Tm2 of all the BNCs slightly decreased with the presence of CNFs. The crystallinity of neat PHB and the BNCs were calculated by using the ΔHm. The crystallinity of the BNCs increased from 45% to 50% with the addition of CNFs and MAPP. The lowest and highest value for crystallinity in the composites was calculated at 50% for the PHB-MAPP-3CNFs and 45% for the neat PHB and the BNCs with 20% CNFs. The increase in crystallinity of the BNCs in the adding CNFs could also be responsible for the increases observed on the mechanical properties of the BNCs as reported in Table 2. Similar results were reported by Srithep et al.12,40
The summary results of DSC curves of the neat PHB and the BNCs.
Dynamic mechanical analysis
Figures 9 to 11 show the E', E'' and tan δ of neat PHB and the BNCs, respectively. As seen in Figure 9, the E' of neat PHB decreased with the adding MAPP. The E' didn’t change with filler loadings of 3% and 5% CNFs, however the E' of the BNCs with 1, 10 and 20% CNFs increased in comparison to the neat PHB attributable to a reinforcing effect of CNFs in the matrix. It can be inferred that the E' represents the capability of a material to store mechanical energy without dissipation.41–44 The highest E' in the BNCs was found at 20% CNF loading level, whereas the lowest E' was found in the BNCs with 3% CNF loading level. The adding CNFs increased the E'' as shown in Figure 10. While the temperature raised, E' and E'' of the neat PHB and the BNCs were found to decrease because of the mobility of the polymer chains of the BNCs with increasing temperature. 44 Similar results have been expressed by Srithep et al. 12 and Melo et al. 45 The tan δ (damping factor) is defined the fractional energy lost occurring after deformation; a high tan δ generally indicates changes in the elasticity of a polymer matrix. 44 In Figure 11, Tan delta increased with 3, 5, 10 and 20% loadings of CNFs, and this is the result of stress removal, and the energy stored in deforming the material recovers more slowly in comparison to the neat PHB. However, it decreased with loadings of 1% and this means that it behaves in an opposite manner to the other composites due to the amorphous areas of PHB permits long-range motion.7,40
E′ of neat PHB and the BNCs.
E″ of neat PHB and the BNCs.
Tan Delta of neat PHB and the BNCs.
Rheological properties
Rheological properties including η*, G', G'', tan delta and steady viscosity were determined at 175 °C, and the obtained results as a function of ω were given in Figures 12 to 16, respectively.
Figure 12 shows that the η* decreased with a raise of ω, due to the shear thinning behavior of the biopolymer. The behavior might be attributable to the orientation of the CNFs which disturbs the formation of PHB chain entanglements in the BNCs with the done shear force to the BNCs and higher interactions of PHB-fibers, which requires higher shear stress and longer relaxation duration in the BNCs to flow18,46–48 Neat PHB has a higher η* compared to the BNCs and it can be said that the adding CNFs didn’t have an improvement on the η* of the BNC. The G' of neat PHB generally was higher in comparison with the BNCs as given in Figure 13. The G′ of the samples gradually decreased from 10−2 Hz to 10−1 Hz (low frequency range) and later it increased with the increasing of ω. G′ of the composites with 20% CNFs was higher than other nanocomposites in the low frequencies range, however, at the high frequencies range (from 10−1 Hz to 102 Hz), it decreased dramatically with the increasing of ω and increasing the CNFs loading levels from 0 to 20 wt.% didn’t provide an improvement on the viscoelastic melt behavior of the BNCs. Similar results generally were found for the G'' of neat PHB and the BNCs, and G'' of neat PHB also was larger than the PHB composites in Figure 14. The addition of CNFs loading levels from 0 to 20 wt. % didn’t show an increase on the G'' of the BNCs. Figures 12 to 14 show some drops in the rheological curves of the neat PHB and the BNCs at the frequency ranges because of possibly undergoing thermal degradation.44,49,50 Similar η*, G' and G'' results were found by Sadat-Shojai et al. 44 and Mousavioun et al. 51 Another viscoelastic property is tan δ, and the tan delta increased from 10−2 Hz to 10−1 Hz and later decreased at from 10−1 Hz to 102 Hz as shown in Figure 15. Tan δ of the samples generally was higher than neat PHB since the composites behave more like a viscous fluid (less elastically) than neat PHB as reported by Mousavioun et al. 51 Figure 16 shows the steady viscosity of the samples, and the neat PHB generally had higher steady viscosities as compare to the BNCs. Although steady viscosity of the samples first increased slightly until critical shear rate, and then decreased at the shear rate from 0.01 to 10. In similar study, the steady viscosity was found to decrease with an increasing of the shear rate, and this was reported to be attributable to the entanglement and intermolecular forces between polymer molecules.16,29,52,53
η* of neat PHB and the BNCs.
G′ of neat PHB and the BNCs.
G″ of neat PHB and the BNCs.
Tan delta of neat PHB and the BNCs.
Steady viscosity of neat PHB and the BNCs.
Conclusions
In this paper, a range of CNF loading levels were used in an attempt to enhance the mechanical, morphological, thermal, viscoelastic, and rheological properties of the PHB, when the CNFs loading levels were increased from 1% to 20%, all the strength values including flexure, tensile and Izod impact strength decreased because of morphological defects caused by some agglomeration of nanofibrils expect for the composites with 20% CNFs. However, all CNFs loadings provided improvements in both flexure and tensile modulus of elasticity of the composites with the reinforcement effect of CNFs in the PHB matrix. It can be implied that CNFs can be used to enhance the stiffness of the polyhydroxybutyrate composites. SEM images showed that some voids, fiber pull-out of nanofibrils and fibril aggregations in the matrix also were detected, and these observations supported the results obtained from the mechanical tests of the composites. Thermal properties of the PHB were generally improved at high temperature by adding CNFs as a reflection of the T50% and T90% results. Tg, Tc and Tm1 generally didn’t change with presence of CNFs but Tm2 decreased with adding CNFs, however, the addition of CNFs increased the crystallinity and this might also be responsible for the increases observed in the modulus of elasticity of the composites. The addition of 10 and 20 wt.% CNFs generally increased the storage modulus and loss modulus, and, adding CNFs didn’t show any deleterious effect on the PHB rheological melt properties.
Footnotes
Acknowledgements
The authors would like to express appreciation for the 2219 post-doc scholarship provided by the Scientific and Technological Research Council (TUBITAK) of Turkey (Funding Number: 1059B191800730). The authors would also like to acknowledge the contributions of Justin Crouse, Richard Fredericks and Chris West who provide the training for analytical test devices.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Scientific and Technological Research Council (TUBITAK) of Turkey (Funding Number: 1059B191800730).
