Journal Information
Vol. 8. Num. 1.
Pages 1-1592 (January - March 2019)
Download PDF
More article options
Vol. 8. Num. 1.
Pages 1-1592 (January - March 2019)
Original Article
DOI: 10.1016/j.jmrt.2018.05.005
Open Access
Rheological and thermal behavior of PHB/piassava fiber residue-based green composites modified with warm water
Eduardo Braga Costa Santosa, Janetty Jany Pereira Barrosb, Danusa Araújo de Mourab, Camila Gomes Morenob, Fabiana de Carvalho Fimb, Lucineide Balbino da Silvaa,b,
Corresponding author

Corresponding author.
a Universidade Federal da Paraíba, Graduate Program in Materials Science and Engineering, CEP: 58051-900 João Pessoa, PB, Brazil
b Universidade Federal da Paraíba, Department of Materials Engineering, CEP: 58051-900 João Pessoa, PB, Brazil
This item has received

Under a Creative Commons license
Article information
Full Text
Download PDF
Figures (8)
Show moreShow less
Tables (4)
Table 1. Textural properties of the fiber residue washed and treated at 50°C.
Table 2. Untreated and treated fiber diameters.
Table 3. TGA data of the PHB, piassava fiber residue, and their composites.
Table 4. DSC data of the PHB and its green composites.
Show moreShow less

The search for increasingly biodegradable materials motivated us to investigate composites of poly(3-hydroxybutyrate) (PHB) and piassava fiber residue. The composite mixing process for 10 and 30% (w/w) mixtures was conducted in an internal mixer at 180°C and 60rpm for 10min. The fiber residue was washed with a detergent solution, treated in warm water at 50°C, and then ground to a particle size of smaller than the 270 mesh. The fiber residue had a higher surface area, higher crystallinity index, and smaller particle size than the residue that was only washed. The fiber had thermal stability up to 224°C. The thermal treatment neither altered the constitution nor crystalline structure of the fiber, which suggested that a major concentration of the thermal fiber led to more effective defibrillation of the fiber residue, which facilitated the mixing process of the 30% (w/w) composite and increased its degree of crystallinity by 8.8%. However, the maximum degradation temperature of the fiber was significantly reduced (by 58°C) for this composite with respect to PHB. For the other composites, the presence of the residue caused a smaller decrease in the thermal stability of the polymer.

Green composites
Piassava fiber residue
Thermal treatment
Torque rheometry
Mass loss
Full Text

Biodegradable polymers, including poly(3-hydroxybutyrate) (PHB) derived from polyhydroxyalkanoates (PHAs), are being extensively investigated. PHB has been widely studied because its thermal properties are similar to those of polypropylene (PP) [1]. The thermal behavior of PHB involves the presence of multiple melting peaks, much like that of other polymers. It is well established that these multiple peaks, composed of two (or more) endothermic peaks with an exothermic peak between them, means that a melting-recrystallization-remelting (MRR) process has occurred. The melting of the crystals formed thus far is attributed to the lowest temperature peak, whereas the higher temperature peak results from the melting of the recrystallized crystals. PHB also presents MRR behavior, which involves the partial melting of crystals, followed by crystallization, with those that did not melt acting as nucleators. This behavior arises because, when PHB crystallizes, more perfect crystals are formed first, and then, other crystals with a higher concentration of defects are formed that are smaller and have less perfect lamellae. These defect laden crystals are accommodated between the larger and more perfect crystals of the dominant lamellae [2–4]. The lowest melting peak of PHB was found to be approximately 15°C lower than the higher temperature peak, and the PHB recrystallization process was hindered when using high heating rates in differential scanning calorimetry (DSC) [5].

Natural fibers have attracted the attention of scientists and technologists since the 1990s due to legislative requirements with respect to the use and final destination of composites obtained from polymer resin and synthetic fibers [6]. Therefore, these environmental concerns have motivated researchers to investigate natural fibers, such as sisal, flax, bamboo, coconut, curaua, jute, ramie, cotton, bagasse, and pineapple, as a reinforcement to polymer matrices [7]. Some advantages of natural fibers are their low cost and density and high specific properties, such as their biodegradability and their derivation from renewable sources, and they are less abrasive compared to synthetic fibers and readily available [8,9]. These fiber characteristics are desirable, which results in their usage in applications of composites such as outdoor and indoor decking, railing, fencing and furniture, as well as automotive applications instead of synthetic reinforcement [9]. Another advantage of natural fibers is their mechanical properties. In a comparison between natural and glass fibers, for example, the tensile strength of the glass fibers was higher than that of natural fibers, while the modulus of both was in the same order of magnitude. On the other hand, it is noteworthy that when the specific modulus of the natural fibers is considered, their values are comparable to or better than those of glass fibers. The specific properties of natural fibers have motived researchers around the world to investigate their potential to improve thermal and mechanical properties as well as the recyclability of biocomposites [10]. The suitable use of natural fibers in composite materials also benefits communities that grow crops and harvest the fibers to supply local industries [11]. Their structures are composed internally of both cellulose and hemicellulose and superficially by lignin with a hydrophilic characteristic that is incompatible with hydrophobic polymeric matrices [12]. Therefore, various superficial modifications have been proposed to improve the interactions between the fiber and polymer matrices, mainly when short fiber residues are to be used in composites materials [13–15]. More recently, fiber residues have gained increased prominence due to their ecological appeal, as the process reduces biomass incineration [11,16]. In the present study, a piassava fiber is obtained from the Attalea funifera Martius palm tree cultivated in north and northeast Brazil, even though the largest area of production is in the southern part of the state of Bahia, where it is considered a native and endemic species. Piassava fiber is used in the manufacture of brooms, brushes, paint brushes and similar items. The process of making brooms produces considerable amounts of piassava fiber residue, which has no industrial use but may be useful for the manufacture of polymeric composites.

Composites of biopolymers and plant fibers, also called green composites, have been increasingly studied in the search for biodegradable eco-materials, mainly in the automotive industry. This great interest stems from the advantages they offer, such as recyclability and light mass, which appeal ecologically and structurally. A new field of research is the investigation of green composites fabricated from the industrial byproducts of plant fibers, whose use mitigates the environmental problems of waste generation [17]. Many of the PHB composites investigated have used flax and carnauba plant fibers [9,18,19]. However, very little research has been published with PHB/plant fiber residue [20]. The study of the thermal behavior of piassava fiber residues and their green composites with PHB will provide knowledge on the degradation temperature and its effect on the crystallinity of the matrix, parameters that are fundamental to the future applications of these eco-friendly materials. Moreover, the production of biopolymers is costly, making it difficult to apply them on a larger industrial scale. Thus, the production of green composites has reduced the cost of obtaining biopolymers and improved their properties.

To the best of our knowledge, the flow and thermal behaviors of PHB/piassava fiber residue composites treated with warm water (50°C) have not been reported in the literature. Therefore, the main goal of this study is to evaluate the flow behavior using torque rheometry and melt flow index and to perform thermal measurements via DSC and thermogravimetric analysis (TGA) of green composites with various compositions. Structural characterizations of the piassava residue are also evaluated using Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), surface area (BET), particle size and morphology to better understand the behaviors of composites.

2Materials and methods2.1Materials

The poly(3-hydroxybutyrate) (PHB) was kindly donated by PHB Industrial SA from São Paulo (Brazil). It has a melting point of approximately 170°C. The residue of the piassava fiber of the species A. funifera Martius from Bahia, Brazil, was kindly donated by the Bruxaxá Company (Pernambuco, Brazil) as trimmings from the production of brooms. The PHB green composites were prepared with the fiber residue, washed, and heat treated with water at 50°C, at compositions of 10–30% by mass.

2.2Surface treatment of piassava fiber residue and microscopic analysis

The fiber residue, as received from the company, was initially cut to sizes ranging from 1 to 5cm in length. It was then washed with a neutral detergent solution (2% v/v) for 24±1h under magnetic stirring at room temperature. Then, the fiber residue was filtered with filter paper and rinsed with distilled water. The washed fiber residue (WFR) is considered in this paper to be the unmodified sample. The WFR sample was dried at 70°C for 120±3min and ground in a knife mill and then in a ball mill for 2h. After grinding, the sample was sieved at a rate of 5Hz for 15min, with a set of sieves, to obtain fiber particle sizes below 270 mesh.

The WFR sample was then treated with warm water at 50°C (TFR50) for 24h. Finally, the TFR50 sample was dried outdoors for 168±2h. The WFR and TFR50 samples were hermetically stored in a desiccator for later use in the preparation of the green composites.

2.3Fiber residue characterization2.3.1Fourier transform infrared spectroscopy (FTIR)

The untreated and treated fiber analyses were performed using an IR-Prestige 21 spectrophotometer, Shimadzu Model, by accumulating 32 scans with 4cm−1 resolution in the 4000–400cm−1 region range. Samples were prepared as KBr pellets (1wt% fiber content).

2.3.2Surface area, particle size and morphology

Untreated and treated fiber surface area analyses were performed using a surface area and porosity analyzer, model ASAP 2420, from Micromeritc. Initially, the samples were treated for 4h at 120°C in a BELPREP-II from Bel Japan under nitrogen flow. The samples were then subjected to sorption in gaseous nitrogen, by applying a He standard in a Dewar containing liquid nitrogen.

Particle size analyses of the untreated and treated fiber were performed using a S3500 unit from Microtrac. All analyses were performed three times using a wet flow. The samples were immersed in an ethanol-acetone mixture.

The surfaces of the untreated and treated fiber were coated with gold and analyzed in a Zeiss Leo 1430 scanning electron microscope with a signal generated by secondary electrons.

2.3.3X-ray diffraction (XRD)

X-ray diffraction patterns of the fibers were recorded using a D8 Advance Davinci unit from Bruker under the following conditions: CuKα radiation, 40kV and 40mA anode current, Ni filter, and 0.6°/min scanning speed. The crystallinity index (ICr) of the fibers was evaluated from the XRD intensity data of the crystalline and amorphous fiber phase, and it was calculated using Segal's formula [21]:

where I002 is the intensity of the crystalline part of the cellulose fiber (i.e., for pure cellulose), and Iam is the intensity of the very broad peak of the amorphous contents (i.e., hemicellulose and lignin); according to [22], these parts generate a peak at 2θ=16°.

2.4Processing and measurement of the green composite melt flow index

Green composites of PHB and either WFR or TFR50 fiber residue, at 10 and 30% (w/w), were oven dried for 6h at 80°C before processing in a Haake PolyLab OS internal mixer at 180°C and 60rpm for 10min. The composites made with 10% and 30% (w/w) unmodified fiber residue were named WFR9010 and WFR7030, respectively, and those with the treated residue were called TFR50-9010 and TFR50-7030, respectively.

The melt flow index (MFI) was determined using a CEAST Melt Flow MODULAS LINE tester, according to the ASTM D 1238 standard. The parameters included a temperature of 174°C and a load of 2.16kg; 5 samples were collected every 15s for each composition. The samples were dried for 6h at 80°C before processing and performing the MFI measurements.

2.5Thermal characterization of the green composites

Thermal characterization was performed by DSC using a SHIMADZU DSC-60. The mass of the samples was 5.0±1.0mg, and the assay was performed under a nitrogen atmosphere at a flow rate of 50mL/min and a heating rate of 10°C/min. First, the samples were heated to 200°C, which was held for 3min to eliminate the thermal history of the material. Then, they were cooled to 25°C and heated again to 200°C. From the second heating, the melting temperature (Tm) and enthalpy of melting (ΔHm) were obtained. The degree of crystallinity (Xc) was calculated from Eq. [2]:

where Xc is the degree of crystallinity (%); ΔHm is the enthalpy of melting (Jg−1); ΔH100% is the theoretical enthalpy of a 100% crystalline PHB (Jg−1), equal to 146Jg−1[23]; and Wf is the fiber mass fraction.

TGA of the fiber residue (WFR and TFR50), the PHB matrix, and the composites was performed using a Shimadzu DTG-60H analyzer with a heating rate of 10°C/min under an argon atmosphere from 25 to 750°C; the mass of the samples was 5±0.05mg.

3Results and discussion

Fig. 1 shows the FTIR spectra of the WFR and TFR50 samples. The spectra show the existence of the main functional groups of the components that form all vegetable fibers: cellulose, hemicellulose and lignin [24].

Fig. 1.

FTIR of samples WFR and TFR50.


The typical functional groups of cellulose appear at 3400cm−1, which is related to the stretching vibrations of the OH group resulting from the bonding with hydrogen in the cellulose. The vibrations at 2931 and 2882cm−1 refers to the aliphatic CH stretches of the methyl and methylene groups, respectively. The bands at 1269, 1163 and 1046cm−1 are attributed to the stretching of the C-O-C group present on the cellulose glucose ring [25,26]. The vibration at 898cm−1 is due to the β-glycosidic bond [27], and it is also representative of the occurrence of the crystalline phase of cellulose type I [26]. The presence of the vibration at 1723cm−1, characteristic of the CO group of unconjugated carbonyl is a peak of hemicellulose [28], as well as at 1246cm−1 refers to the acetyl group [25]. For lignin, the vibrations characteristics are determined by its structural composition. These vibrations are attributed to wavelengths at 1605, 1455 and 823cm−1, referring to the aromatic structure of lignin [29]. In addition, other characteristic peaks of lignin are at 1423cm−1, which are assigned to the methoxyl-O-CH3 stretch and the bands at 1163 and 1112cm−1, which refer to the COC and CO alcohol groups, respectively.

The spectra shown in Fig. 1 shows that the modified surface of the fiber residue with water at 50°C was not efficient in the total or partial removal of hemicellulose and lignin from the fiber because only differences in intensity of the vibration bands of hemicellulose (1723cm−1) and of lignin (1605, 1246, 1046 and 823cm−1) were observed. The nitrogen adsorption analysis verified the influence of the surface modification on the texture properties of the piassava fiber residues. Table 1 presents the specific surface area data by the BET method, parameter C and the pore average diameter by the BJH method for the washed fiber residue (WFR) and for the fiber residue treated at 50°C (TFR50).

Table 1.

Textural properties of the fiber residue washed and treated at 50°C.

Sample  SBETa (m2/g)  Dp(BJH)b (Å)  Parameter C 
WFR  50.82±0.46  53.40  3.16 
TFR50  103.73±1.20  49.97  2.88 

SBET, surface area via BET.


Dp(BJH), mean pore diameter via BJH method.

The results of the specific area showed that the modification of the surface with water at 50°C was effective at increasing the surface area of the fiber and decreasing the average diameter of the pores. This result is in agreement with SEM analyses that collaborate the possibility of defibrillation of the TFR50 sample with a probable reduction in the diameter of the fiber (Fig. 4c) compared to the WFR sample (Fig. 4a).

The diameter distribution curves of the untreated and treated fibers are shown in Fig. 2 and the values of the particle diameters are presented in Table 2. The shape of the piassava fiber diameter was considered to be cylindrical in accordance with the Gaussian frequency distribution curves presented in Fig. 2. The untreated fiber had a narrower range of diameter distribution than the treated fiber, which suggests that the WFR sample had more homogeneously sized particles than the TFR 50 sample, which had a larger distribution of small and larger particles. The time and temperature affected the particle size of the TFR50 sample and the diameter decreased and specific surface area increased as both increased, as shown in Table 2.

Fig. 2.

Particle size of the fibers: WFR and TFR50.

Table 2.

Untreated and treated fiber diameters.

Sample  D¯V (μm)  D¯N (μm)  D¯A (μm)  SA (m2/mL)  Icr (%) 
WFR  97.88  1.090  32.49  1.85×10−1  25.14 
TFR50  66.60  0.800  16.30  3.68×10−1  29.27 

D¯V = average diameter, in mm, volume distribution.

D¯N = average diameter, in μm, number distribution.

D¯A = average diameter, in μm, area distribution.

SA, specific surface area, m2/mL.

Icr, crystallinity index.

The crystalline structure of the vegetable fiber is well defined in the current literature [30–32], being classified as cellulose type I or II, where the cellulose-I structure is identified by 101, 101¯ and 002 planes with 2θ angles of approximately 14°, 16° and 23°, respectively, while cellulose-II has one other plane (021, in 2θ=20°). Fig. 3 depicts the XRD diffractograms of untreated and treated piassava fiber. The crystalline structure of the cellulose in both the WFR and TFR50 samples is type I with 2θ angles of approximately 22° (002 plane), which corresponds to cellulose, and 16° (101 plane), 35° (040 plane), which correspond to hemicellulose and lignin respectively, according to Miranda et al. [33]. Moreover, in the FTIR analysis, the presence of the 823cm−1 band can be related to a cellulose structure of type I according to Zhang et al. [26]. However, the crystallinity index value of the TFR50 sample was higher than the WFR value, as shown in Table 2. This finding is in agreement with that of Mulinari et al. [17], who found that a thermal treatment at 100°C resulted in a higher crystallinity index value for a textile fiber residue.

Fig. 3.

XRD diffractograms of the piassava fibers: WFR and TFR50.


Fig. 4a shows the morphological aspect of washed piassava fiber denoting that its surface is covered by a superficial layer of typical amorphous phase composed of plant fibers. Based on the FTIR analyses, hemicellulose and lignin covered the samples of the surfaces. After treatment with warm water, the fiber surface became smoother and its bundles of microfibrils were more visible, as shown by arrows in Fig. 4b. This finding is in agreement with Mulinary et al. [17] who observed modification on the morphological aspect of fiber that acquired flattened forms after warm water treatment at 100°C, which is attributed to the removal of the extractives. Furthermore, partial dissociation among the bundles of TFR50 sample was verified, which suggests defibrillation. The microfibril diameter was approximately 55μm, obtained from the Fig. 4c, which was approximated to be the average diameter of the volume distribution obtained from the size particle analysis (Table 2). The defibrillation mechanism of a plant fiber occurs due to a process of breaking fiber bundles into smaller fibers, which increases the surface area of the fiber [34]. In fact, the TFR50 sample has a smaller diameter and larger superficial area than the WFR sample, as shown in Tables 1 and 2. Furthermore, the untreated fiber residue had more connected bundles of microfibrils, as shown in Fig. 4a. These morphological aspects suggest that the warm water treatment caused partial fiber defibrillation in the PHB matrix, which was intensified during the processing of the composites. Therefore, morphological modification of the piassava fiber residue treated superficially with warm water can alter the flow behavior and thermal behavior of composites composed of piassava fiber and PHB. For example, Yan-Hong et al. [35] observed that changes in both the fiber aspect ratio and the morphology influenced the rheological behavior of the green composites and, consequently, their processing.

Fig. 4.

SEM images of the piassava fiber residues: (a) WFR, (b) TFR50 and (c) bundles of microfibrils dissociated from the TRF50 sample.


Fig. 5a shows the torque curves (M) as a function of time for PHB and its green composites. The inset shows the first minute of the feed, revealing that the matrix and the composites had different torque values. The flow behavior of PHB was influenced both by the treatment of the fiber residue and by its morphology and concentration. The 10% (w/w) composites showed higher torques than the 30% (w/w) composites. TFR50-9010 stands out because it has the highest torque among all the samples. Yan-Hong [35] suggested that greater contact area between the load and the matrix results in increased torque. The partial dissociation among the bundles of microfibrils, which leads to TRF50 fiber defibrillation, can increase the contact between the fiber and matrix in the 10% composite and thereby increase its torque. In the steady state, represented by the final 2min of mixing, there were no significant differences in the torque curves of the samples. Therefore, there was no degradation of the samples during processing. On the other hand, both composites with 30% fiber had decreased torques with respect to those with the 10% concentration. Although these results seem contradictory, the decrease in torque with the addition of fiber is related to the high lignin content, approximately 48.4%, present in the piassava fiber [13]. For example, Ayora et al. [36] suggested that the lignin present in coconut fiber, which is approximately 32.8% [37], acted as a lubricant in the studied composites and reduced the torques in the transient flow regime. The lubricant effect, based on this assumption, plus the fiber defibrillation in the TRF50-7030 composite resulted in more sliding among the bundles of microfibrils, which favored both flow and a decrease in the torque. Accordingly, the torque of WFR7030 was higher than that measured for the TRF50-7030 composite because the washed fiber used to prepare the WFR7030 sample did not show defibrillation. Another interesting feature of the flow behavior of these composites was that the fiber residue and the matrix mixed easily since a steady flow regime was reached immediately after the feed.

Fig. 5.

PHB and its composites: (a) torque vs. time curve with a feed area detail inside the figure and (b) Melt Flow Index measurements.


The flow behavior of the composites was also evaluated by the melt flow index (MFI), whose values, along with a statistical analysis of the reliability of the measurements, are presented in Fig. 5b. Based on this analysis, it can be assumed that addition of the fiber residue increased the PHB matrix fluidity by varying the MFI values of the samples within the upper control (UCL) and lower control (LCL) limits of the analysis. The addition of 30% (w/w) residue considerably increased the fluidity index of the biopolymer matrix. The samples were initially mixed in the torque rheometer and then extruded through the capillary matrix of the fluidity index meter. Thus, the samples passed through two cycles of temperature and shear. In this sense, the composites with higher residue contents were more affected and showed the highest fluidity index values among the samples. The alignment of the fiber along the flow direction, as it passes through the capillary of the melt flow meter, may also have contributed to the increased fluidity of these composites. The trend of the flow through the torque rheometer for the WFR7030 and TFR50-7030 composites is also observed in the fluidity analysis. The defibrillation of the treated fiber residue and the lignin lubricant effect in the TFR50-7030 composite contributed to it having the greatest MFI value, which is three times higher than that of the PHB matrix. Moreover, the higher crystallinity index of the TRF50 fiber collaborates with the alignment of bundles of microfibrils, which increases the flux. Therefore, the facilitation of integration and mixing between the PHB and piassava fiber residue, as well as the improved fluidity of all the composites, make the composites candidates for machine-executed processes (extrusion and injection molding) that are used with other thermoplastic polymers.

Fig. 6 shows the mass loss and derivative mass loss (DTG) curves plotted as a function of temperature for the WFR and TFR50 samples. In general, the degradation of the plant fibers occurs in three stages [38] as follows. The first step, which occurs below 150°C, involves moisture evaporation, water desorption, and the emission of volatile organic compounds. The second stage, involving the degradation of non-cellulosic polysaccharides (hemicellulose and pectin), is characterized by the presence of a shoulder and a peak of maximum mass loss corresponding to cellulose degradation, that is, the breakdown of glycosidic bonds of the glucose chain. The third step corresponds to lignin degradation, which is almost imperceptible because it occurs in a wide range of temperatures beginning at the hemicellulose degradation temperature. Finally, the carbonization stage of the plant fiber occurs at temperatures above 500°C. The WFR and TFR50 samples were evaluated between 100 and 680°C. The degradation of the samples did not occur until 224°C, and two degradation stages were clearly observed. The shoulder peak corresponding to hemicellulose degradation occurred at 283.5°C for WFR and at 274.7°C for TFR50, and the cellulose degradation peaks occurred at 329.7 and 335.0°C, respectively. Our results agree with those of D’Almeida et al. [39], who obtained 276.4 and 347.8°C for the degradation of hemicellulose and cellulose, respectively, in unmodified piassava fiber residues. The degradation temperatures of the WFR and TFR50 residues are very close, but their mass losses differ during the degradation stages, especially during cellulose degradation: 21.9% for WFR and 36.2% for TFR50. It is noteworthy that treating the piassava fiber residue at 50°C made the hemicellulose shoulder and the cellulose peak sharper and more defined than those of the unmodified residue, in agreement with Várhegyi et al. [40], who observed the same effect when chestnut wood was treated at 60°C. Lignin degradation occurred at relatively low rates [41], and the degradation is represented by the tailing region in the mass loss thermogram [42]. Fig. 6c compares the mass loss curves of both samples. For the WFR sample, the loss of lignin started at a lower temperature (approximately 10°C), degrading more slowly than in the TFR50 sample. Both samples showed a tailing lignin degradation over a wide temperature range, similar to that reported by other authors [41,42].

Fig. 6.

TGA of the piassava fiber residues of the samples: (a) WFR, (b) TFR50 and (c) the mass loss curves of the samples.


These degradation variations are closely related to changes in the chemical composition of the TFR50 sample arising from its treatment, which removed surface material (Fig. 4b), providing results unlike those of the unmodified residue (Fig. 4a). Thus, the warm water treatment caused lignin degradation to occur more rapidly. However, the onset of this lignin degradation occurred at a higher temperature for the TFR50 residue than for the WFR residue, at 360 and 350°C, respectively. Additionally, the TFR50 residue had an ash content of 4.8%, almost half that of WFR, 10.5%. Our results agree with those obtained by Várhegyi et al. [40], who attributed changes in degradation temperature and ash content to the removal of sugars, phenolic compounds, inorganic matter and metallic ions of the biomass.

Fig. 7 shows the mass loss and derivative mass loss (DTG) curves of PHB and its green composites. The degradation of the samples was analyzed using the temperature at which a mass loss of 20% occurred (T20), the temperature with the maximum rate of decomposition (Tmax) and the remaining ash content of PHB, WFR, TFR50, and their green composites at 600°C; these values are shown in Table 3. The 20% mass loss of PHB occurred at approximately 281.8°C, and it presented the highest decomposition rate at 293.3°C, with only one stage of degradation. The composites presented T20 and Tmax values lower than those of PHB. The TFR50-7030 composite showed the worst resistance to thermal degradation, with a decrease of approximately 58°C relative to the biopolymer, probably related to the greater defibrillation caused by the fiber residue treatment. Consequently, the fibrils and, thus, their hydroxyl groups (which make up the structure of cellulose, hemicellulose and lignin) were more exposed, which may have facilitated interactions with the carbonyl groups of PHB. It can be assumed that more interactions will facilitate cleavage of the PHB polymer chains because the tendency is for thermal degradation to occur first at the end groups of the chains, initiating self-oxidation cycles [43]. On the other hand, the temperature at 20% mass loss of the TFR50-9010 composite was mildly altered compared to the pure polymer, which suggests that less defibrillation occurred in this composite and it was not enough to decrease the temperature. As shown in Fig. 7a, all the composites showed two stages of degradation, the first at a lower temperature due to PHB and another due to the degradation of the fiber residue, a behavior also observed for other plant fiber-based composites [44]. The composites with higher percentages of fiber residue presented a more pronounced second stage than the other composites, WFR9010 and TFR50-9010. The presence of the second stage may indicate that the overall thermal stability of the composites was higher than that of the PHB matrix; that is, degradation occurred over a wider temperature range, despite having lower T20 and Tmax than those of the biopolymer, in agreement with other reports [45]. The ash content of WFR7030 was the highest of the composites, while that of TFR50-7030 was close to that of the 10% (w/w) composites. This result is in agreement with that of the TFR50 sample, which had a lower ash content than the unmodified fiber residue, as shown in Table 3.

Fig. 7.

(a) Mass loss and (b) DTG of PHB and its composites.

Table 3.

TGA data of the PHB, piassava fiber residue, and their composites.

Sample  T20 at 20% mass loss (°C)  Tmax maximum degradation rate (°C)  Ash (%) at 600°C 
PHB  281.8  293.3  1.9 
WFR9010  278.9  287.6  2.3 
WFR7030  277.4  284.4  6.7 
TFR50-9010  280.4  288.2  2.3 
TFR50-7030  225.9  234.7  2.6 
WFR  279.6  352.1  10.5 
TFR50  274.6  347.4  4.8 

Fig. 8 and Table 4 show the curves and data obtained by melting the samples from the second heating run of the DSC and show the degree of crystallinity calculated using Eq. [3]. PHB and its composites have multiple melting peaks, with an endothermic peak at the lowest temperature (Tm1), an intermediate exothermic peak and another endothermic peak at a higher temperature (Tm2). This behavior characterizes the melting-recrystallization-remelting (MRR) of these samples. First is the melting of the smaller crystals, i.e., those with a smaller lamellar thickness and a lower degree of order. These can also act as nuclei for the recrystallization that reorganizes the molecules, leading to the formation of larger and more stable crystals, which require more energy to re-melt [2]. The Tm1 and Tm2 values of the PHB are approximately 160 and 171°C, respectively, which are similar to those of the green composites. However, the enthalpy of fusion values (ΔHm1 and ΔHm2) is different from those of PHB. All the composites, except for TFR50-7030, showed lower enthalpy values than PHB, and WFR7030 had the smallest ΔHm1 and ΔHm2.

Fig. 8.

DSC curves of the PHB and its green composites obtained during second heating.

Table 4.

DSC data of the PHB and its green composites.

Sample  Tm1 (°C)  ΔHm1 (J/g)  χc1 (%)  Tm2 (°C)  ΔHm2 (J/g)  χc2 (%)  χc (%) 
PHB  160.75±0.62  19.57±1.78  13.42±1.20  172.01±0.65  57.5±9.36  39.38±6.41  52.8 
WFR-9010  161.19±0.28  13.88±6.21  10.56±4.7  171.84±0.1  24.92±6.15  18.96±4.6  29.52 
WFR-7030  161.53±0.26  11.63±4.85  11.38±4.75  171.79±0.07  16.24±6.14  15.89±6.01  27.27 
TFR50-9010  160.79±0.35  17.49±2.14  13.30±1.63  171.41±0.02  29.09±7.15  22.14±5.44  42.39 
TFR50-7030  161.09±0.22  28.98±0.53  28.35±0.51  171.47±0.23  33.99±3.27  33.26±3.21  61.61 

The higher WFR content may have limited the organization of the molecules to form PHB crystals in both the primary crystallization and the recrystallization processes, resulting in a lower degree of crystallinity of approximately 27.27% in the WFR7030 composite. The surface treatment of the fiber residue with 50°C warm water resulted in a smaller decrease in the degree of crystallinity of the TFR50-9010 composite relative to that of WFR9010. However, the surface treatment of the fiber residue had a more positive effect on χc for the green composite TRF50-7030, whose degree of crystallinity was increased by 8.8% in relation to PHB. Defibrillation of the fiber residue in the TRF50 composites may have caused this increase in χc. This conclusion is based on previous results in which a decrease in the residue diameter (Table 2) and an increase in its surface area (Table 1) led to the formation of more stable crystals in these composites than in the WFR composites. Consequently, nucleation of the PHB crystals occurred in the TFR50 composites when crystallization during the cooling scan (ΔHm1) and recrystallization during heating (ΔHm2) occurred. The defibrillation was more intensive in the TRF50-7030 composite because the higher content of fiber contributed to increasing the shear rate among the fibers. As a result, the bundles of microfibrils had a nucleant effect on the PHB, which increased its degree of crystallinity. Additionally, as the crystallinity index of the TRF50 fiber increased, it became more rigid, which contributed to the crystallization of the polymer because it became a more rigid surface over which the PHB crystals could nucleate and grow. The TFR50-9010 composite had a lower χc than the TRF50-7030 composite, which agrees with the conclusion that the defibrillation was important to crystallization and that it was more prominent in the TRF50-7030 composite. On the other hand, the composites with WFR had lower values of crystallinity than the TFR50 composites as seen in the morphological aspect in which the WFR did not have defibrillation. Therefore, the composites with WFR had lower degrees of crystallinity than both pure polymer and TFR composites. Moreover, the benefit of adding WFR on the crystallinity of the PHB was slightly higher for the composite with 10% PHB than with the composite with the 30% WFR.


PHB/piassava fiber residue green composites had improved flow behavior due to increases in their melt flow indexes and easier wetting and interaction during processing, especially for samples with higher residue concentrations, which is beneficial because this biopolymer is difficult to process. The fiber residue presented thermal stability until 224°C. The heat treatment improved some of the interesting structural characteristics of the fiber residue such as its use as a load, its increased specific area and crystallinity index and its reduced particle size. The green composites with untreated and 10% treated fibers had a slight decrease in the maximum degradation temperature with respect to the PHB, which further decreased for the 30% treated fiber. However, a positive finding is that the second decomposition stage of the composite occurred at higher temperatures than the pure biopolymer degradation temperature, meaning that the composites had overall thermal stability over a wider temperature range. PHB and its green composites showed multiple melting peaks at temperatures of 160 and 171°C. Furthermore, the degree of crystallinity of the biopolymer was further influenced by the treated fiber residue, especially when the concentration was 30% (w/w). Therefore, it is advantageous to treat the piassava fiber residue surface with warm water; besides being economical and non-polluting, it is very promising in regards to the development of eco-friendly green composites with some improvements to the thermal and rheological properties of the composite.

Conflicts of interest

The authors declare no conflicts of interest.


The Authors gratefully acknowledge the financial/technical support from the following institutions: the scholarships provided by both the Coordination for the Improvement of Higher Level or Education Personnel (CAPES) and the National Council for Scientific and Technological Development (CNPq), the financial support to the project number APQ-1481-3.03/15 provided by Pernambuco Research Foundation (FACEPE), and the technical support provided by both the Northeast Center for Strategic Technologies (CETENE) and the Fast Solidification Laboratory at the Federal University of Paraiba (LSR-UFPB).

L. Wei, A.G. McDonald.
Peroxide induced cross-linking by reactive melt processing of two biopolyesters: poly(3-hydroxybutyrate) and poly(l-lactic acid) to improve their melting processability.
J Appl Polym Sci, 132 (2015), pp. 1-15
R. Pearce, R.H. Marchessault.
Multiple melting in blends of isotactic and poly(β-hydroxybutyrate).
Polymer, 35 (1994), pp. 3990-3997
L. Wei, A.G. McDonald.
Thermophysical properties of bacterial poly(3-hydroxybutyrate): characterized by TMA, DSC, and TMDSC.
J Appl Polym Sci, 132 (2015), pp. 1-11
L.M.W.K. Gunaratne, R.A. Shanks.
Multiple melting behaviour of poly(3-hydroxybutyrate-co-hydroxyvalerate) using step-scan DSC.
Eur Polym J, 41 (2005), pp. 2980-2988
M.C. Righetti, M.L. Di Lorenzo.
Melting temperature evolution of non-reorganized crystals, poly(3-hydroxybutyrate).
Thermochim Acta, 512 (2011), pp. 59-66
P. Wambua, J. Ivens, I. e Verpoest.
Natural fibres: can they replace glass in fibre reinforced plastics?.
Compos Sci Technol, 63 (2003), pp. 1259-1264
S. Taj, M. Ali, S. Khan.
Review: natural fibre reinforced polymer composites.
Proc Pak Acad Sci, 44 (2007), pp. 129-144
H.P.S. Abdul Khalil, I.U.H. Bhat, M. Jawaid, A. Zaidon, D. Hermawan, Y.S. Hadi.
Bamboo fibre reinforced biocomposites: a review.
Mater Des, 42 (2012), pp. 353-368
S. Singh, A.K. Mohanty, T. Sugie, Y. Takai, H. Hamada.
Renewable resource based biobased composites from natural fiber and polyhydroxybutyrate-co-valerate (PHBV) bioplastic.
Compos A: Appl Sci Manuf, 39 (2008), pp. 875-886
D.N. Saheb, J.P. Jog.
Natural fiber polymer composites: a review.
Adv Polym Technol, 18 (1999), pp. 351-363
M.A. Arbex, J.E.D. Cançado, L.A.A. Pereira, A.L.F. Braga, P.H.N. Saldiva.
Biomass burning and health effects.
Jornal Brasileiro de Pneumologia, 30 (2004), pp. 158-175
A.S. Singha, A.K. Rana.
Effect of aminopropyltriethoxysilane (APS) treatment on properties of mercerized lignocellulosic grewia optiva fibe.
J Polym Environ, 21 (2013), pp. 141-150
E.B.C. Santos, C.G. Moreno, J.J.P. Barros, D.A. Moura, A. Ries, R.M.R. Wellen, et al.
Effect of alkaline and hot water treatments on the structure and morphology of piassava fibers.
M.J. John, R.D. Anandjiwala.
Recent developments in chemical and characterization of natural fiber-reinforced modification.
Polym Compos, 29 (2008), pp. 187-207
A. Jahn, M.W. Schroder, M. Futing, K. Schenzel, W. Diepenbrock.
Characterization of alkali treated flax fibres by means of FT Raman spectroscopy and environmental scanning electron microscopy.
Spectrochim Acta A: Molec Biomolec Spectrosc, 58 (2002), pp. 2271-2279
A. Mahmoud, H. Abdel-Goad.
Dynamic mechanical properties of agricultural residues-filled polystyrene.
J Appl Polym Sci, 93 (2004), pp. 37-40
D.R. Mulinari, H.J.C. Voorwald, M.O.H. Cioffi, C.A.A. Lima, C.A.P.R. Batista, G.J.M. Rocha.
Composite materials obtained from textile fiber residue.
J Compos Mater, 45 (2011), pp. 543-547
S.G. Lee, S.S. Choi, W.H. Park, D. Cho.
Characterization of surface modified flax fibers and their biobased composites with PHB.
Macromolec Symp, 197 (2003), pp. 89-99
J.D.D. Melo, L.F.M. Carvalho, A.M. Medeiros, C.R.O. Souto, C.A. Paskocimas.
A biodegradable composite material based on polyhydroxybutyrate (PHB) and carnauba fibers.
Compos B: Eng, 43 (2012), pp. 2827-2835
E.V. Torres-Tello, J.R. Roblexdo-Ortíz, Y. González-García, A.A. Pérez-Fonseca, C.F. Jasso-Gastinel, E. Mendizábal.
Effect of agave fiber content in the thermal and mechanical properties of green composites based on polyhydroxybutyrate or poly(hydroxybutyrate-co-hydroxyvalerate).
Ind Crops Prod, 99 (2017), pp. 117-125
L. Segal, L. Creely, A.E. Martin, C.M. Conrad.
An empirical method for estimating the degree of crystallinity of native cellulose using X-ray diffractometer.
Textile Res J, 29 (1959), pp. 786-794
R. Kumar, G. Mago, V.E. Balan, C.E. Wyman.
Physical and chemical characterizations of corn stover and poplar solids resulting from leading pretreatment technologies.
Bioresour Technol, 100 (2009), pp. 3948-3962
P.J. Barham, A. Keller, E.L. Otun, P.A. Helmes.
Crystallization and morphology of bacterial thermoplastic: poly-3 hydroxy butyrate.
J Mater Sci, 19 (1984), pp. 2781-2794
J.I. Morán, V.A. Alvarez, P.V. Cyras.
Extraction of cellulose and preparation of nanocellulose from sisal fibers.
Cellulose, 15 (2008), pp. 149-159
E.N. Pires, C. Merlini, A.Q.A. Hazim, G.V. Salmoria, G.M.O. Barra.
The influence of alkaline treatment on jute fiber-reinforced epoxy resin composite.
Polímeros, 22 (2012), pp. 339-344
T. Zhang, M. Guo, L. Cheng, X. Li.
Investigations on the structure and properties of palm leaf sheath fiber.
Cellulose, 22 (2015), pp. 1039-1051
N. Abidi, L. Cabrales, C.H. Haigler.
Changes in the cell wall and cellulose content of developing cotton fibers investigated by FTIR spectroscopy.
Carbohydr Polym, 100 (2014), pp. 9-16
S.M. Mortazavi, M.K. Moghaddam.
An analysis of structure and properties of a natural cellulosic fiber (leafiran).
Fibers Polym, 11 (2010), pp. 877-882
H. Yang, R. Yan, H. Chen, D.H. Lee, C. Zheng.
Characteristics of hemicellulose, cellulose and lignin pyrolysis.
Fuel, (2007), pp. 1781-1788
J.I. Moran, V.A. Alvarez, V.P. Cyras, A. Vazquez.
Extraction of cellulose and preparation of nanocellulose from sisal fibers.
Cellulose, 15 (2008), pp. 149-159
M. Aydin, H. Tozlu, S. Kemaloglu, A. Aytac, G. Ozkoc.
Effects of alkali treatment on the properties of short flax.
J Polym Environ, 19 (2011), pp. 11-17
L. Xiao, Z. Sun, Z. Shi, X.F. Feng, R. Sun.
Impact of hot compressed water pretreatment on the structural changes of woody biomass for bioethanol production.
BioResources, 6 (2011), pp. 1576-1598
C.S. Miranda, R.P. Fiuza, R.F. Carvalho, N.M. José.
Effect of surface treatment on properties of bagasse piassava fiber Attalea funifera Martius.
Química Nova, 38 (2015), pp. 161-165
S. Mishra, M. Misra, S.S. Tripathy, S.K.K. Nayak, A.K. Moha.
The influence of chemical surface modification on the performance of sisal-polyester biocomposites.
Polym Compos, 23 (2002), pp. 164-170
F. Yan-Hong, L. Yi-Jie, X. Bai-Ping, Z. Da-Wei, Q. Jin-Ping, H. He-Zhi.
Effect of fiber morpholology on rheological properties of plant fiber reinforced poly(butylene succinate) composites.
Compos B: Eng, 44 (2013), pp. 193-199
M. Ayora, R. Rios, J. Quijano, A. Marquez.
Evaluation by torque-rheometer of suspensions of semi-rigid and flexible natural fibers in a matrix of poly(vinyl chloride).
Polym Compos, 18 (1997), pp. 549-560
H.P.S. Abdul Khalil, M. Siti Alwani, A.K. Mohd Omar.
Chemical composition, anatomy, lignin, distribution, and cell wall structure of Malaysian plant waste fibers.
BioResources, (2006), pp. 220-232
M. Benítez-Guerrero, J. Lopés-Beceiro, P.E. Sánchez-Jiménez, J. Pascual-Cosp.
Comparison of thermal behavior of natural and hot-washed sisal fibers based on their main components: cellulose, xylan and lignin. TG-FTIR analysis of volatile products.
Thermochim Acta, 581 (2014), pp. 70-86
J.R.M. D’Almeida, R.C.M.P. Aquino, S.N. Monteiro.
Tensile mechanical properties, morphological aspects and chemical characterization of piassava (Attalea funifera) fibers.
Compos A: Appl Sci Manuf, 37 (2006), pp. 1473-1479
G. Várhegyi, M.G. Gronli, C. Di Blasi.
Effects of sample origin, extraction, and hot-water washing on the devolatilization kinetics of chestnut wood.
Ind Eng Chem Res, 43 (2004), pp. 2356-2367
C. Di Blasi.
Modeling chemical and physical processes of wood and biomass pyrolysis.
Prog Energ Combust, 34 (2008), pp. 47-90
T.H. Nam, S. Ogihara, H. Nakati, S. Kobayashi, J. Song.
Mechanical and thermal properties and water absorption of jute fiber reinforced poly (butylene succinate) biodegradable composites.
Adv Compos Mater, 21 (2012), pp. 41-58
B. Singh, N. Sharma.
Mechanistic implications of plastic degradation.
Polym Degrad Stab, 93 (2008), pp. 561-584
M.L.C. Machado, N.C. Pereira, L.F. Miranda, M.C. Terence, J.G.C. Pradella.
Study of mechanical and thermal properties of the polymer poly-3-hydroxybutyrate (PHB) and PHB/wood flour composites.
Polímeros, 20 (2010), pp. 65-71
M. Payam, O.S. William, G.G. Doherty.
Thermal stability and miscibility of poly(hydroxybutyrate) and soda lignin blends.
Ind Crops Prod, 32 (2010), pp. 656-661
Copyright © 2018. Brazilian Metallurgical, Materials and Mining Association
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.