Enhancing the thermal and mechanical properties of epoxy resins by addition of a hyperbranched aromatic polyamide grown on microcrystalline cellulose fibers

Abstract

In this study, microcrystalline cellulose fibers (MCFs) derived from sisal were treated with a hyperbranched aromatic polyamide (HBAP). The modified sisal fibers were used to produce composites with epoxy resins. Firstly the MCFs were treated with a silane coupling agent, then a HBAP was grown on the modified surface. The HBAP-MCFs were used to reinforce epoxy resins. The HBAP-MCF/epoxy composites were studied by Fourier transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM), thermal gravimetric analysis (TGA), dynamic mechanical analysis (DMA), and mechanical properties analysis. The results show that the HBAP-MCFs enhanced the thermal and mechanical properties of the epoxy resin. For instance, the impact strength, tensile strength, Young's modulus and toughness of the HBAP-MCF/epoxy composites with 2.0 wt% HBAP-MCFs were 32.1 kJ m⁻², 59.4 MPa, 695 MPa, and 4.37 MJ m⁻³. These values represent improvements of 83.4%, 34.7%, 25%, and 178.3%, respectively, compared to a neat epoxy resin. Moreover, the addition of HBAP-MCFs produced composites with higher thermal degradation temperatures and glass transition temperatures. The HBAP-MCF swere effective in improving the thermal and mechanical properties due to a strong affinity between the fillers and the matrix.

---

Introduction

The dwindling of fossil resources, growing concerns about the environment and ever-increasing prices of petroleum-based materials are some of the driving forces towards the exploitation of renewable and eco-friendly materials.^1–12 In the last two decades, natural fiber polymeric materials have gained great interest and insights from laboratory work and this basic research has been transferred to industrial implementation. The main selling point of natural over synthetic fibers is that they are abundant in nature, derived from renewable raw materials and have a low cost. Compared with conventional inorganic fillers such as glass fibers and carbon fibers, natural fibers provide many advantages. Some of these advantages are their abundance in nature, low cost, biodegradability, flexibility during processing, low machine wear, health hazards and densities and desirable fiber aspect ratios, while also providing a relatively high tensile and flexural modulus. Incorporating the tough and lightweight natural fibers into polymer matrices produces composites with a high specific stiffness and strength.^13–21

Sisal is one of the most widely used natural fibers and is very easy to cultivate. It has short renewal times and grows wild in the hedges of fields and along railway tracks. Nearly 4.5 million tons of sisal fibers are produced every year throughout the world.^18,22 Sisal fibers are composed of cellulose (50–74 wt%), lignin (8–11 wt%), hemicellulose (10–14 wt%), pectin (1 wt%) and wax (2 wt%).^20,23 Because of its high cellulose content, cellulose extraction from these fibers could lead to large quantities of microcrystalline cellulose.²³

Lu et al.²⁴ reported that better and stronger adhesion between microfibrils and epoxy matrices was observed for treated fibers, which resulted in better mechanical properties of the composite materials. Xie et al.²⁵ investigated the effects of silane treatments on the mechanical and outdoor performance of the resulting composites. Abdelmouleh et al.²⁶ showed that silane functional groups, appended to fiber surfaces, could participate in the chain growth of appropriate monomers to give a covalent continuity between the fibers and the ensuing polymer matrix. Alamri et al.²⁷ found that the values of the maximum water uptake and diffusion coefficient rose with an increase in fiber content. Peng et al.²⁸ indicated that the surface properties of single fibers can be tailored to facilitate the interface of composites and laccase treatment could be a potential alternative to traditional chemical treatments or pretreatments to modify cellulose fibers. Kalia et al.²⁹ reported that modification of sisal fibers with both bacterial cellulose and methyl methacrylate resulted in better properties with regard to composite fabrication. Lu et al.³⁰ showed that hyperbranched polymer lubricant (HBPL) treatment could significantly improve the melt flow rate of lignin/polypropylene composites. They also observed that the dispersion of lignin treated with HBPL in the polymer matrix was improved compared to neat lignin. Chand et al.³¹ reported that sisal fibers treated with polysulphide and used to reinforce lignin epoxy composites led to electrical anisotropic behaviour. Spoljaric et al.³² investigated polypropylene microcrystalline cellulose (PMC) composites. They showed that the storage modulus, loss modulus and glass transition temperature increased with PMC concentration due to effective interactions between polypropylene and PMC. Maleic anhydride treated polypropylene was used as a compatibilizer for polypropylene/microcrystalline cellulose/wood flour composites by Ashori et al.³³ The results showed that PMC along with wood flour could be effectively used as a reinforcing agent in thermoplastic matrices.

Very few works, have been published so far on composites containing microcrystalline cellulose modified with hyperbranched polymers (HBPs). The objective of this research was to synthesize a new kind of reinforcement agent by growing hyperbranched aromatic polyamide on the surface of microcrystalline cellulose fibers (HBAP-MCFs). The aim was to use this agent to improve the physical and thermal properties of polymer composites. Epoxy resins were chosen as the polymer matrix because of their wide use in aerospace, automotive and electronic industries.³⁴ In this study, firstly we extracted microcrystalline cellulose fibers (MCFs) from sisal, then modified the surface of the MCFs by treatment with a silane coupling agent to introduce amine groups as nucleation sites to grow the hyperbranched aromatic polyamide. The strong covalent bond formation between the silane and the fibers and the silane and the polyamide was intended to produce a strong interaction between the fiber filler and the epoxy matrix.

Experimental

Materials

Sisal fibers were obtained from the Guangxi Sisal Company, China. The epoxy resin used in this study was the diglycidyl ether of bisphenol A (DGEBA, E-44, epoxy value = 0.44) supplied by Yueyang Chemical Plant, China. The growing monomer, 3,5-diaminobenzoic acid (DABA, purchased from Aladdin Chemistry Co., Ltd.) was purified by recrystallization from water and dried under vacuum at 80 °C for 12 h. 4,4-Diaminodiphenylsulfone (DDS) was purchased from Aladdin Chemistry Co., Ltd., with a molecular mass of 248.31 and purity of 96% according to the supplier. Lithium chloride (LiCl) was dried at 100 °C overnight before use. Other chemicals, such as triphenylphosphine (TPP), N-methyl-2-pyrrolidone (NMP), N,N-dimethylacetamide (DMAc), N,N-dimethylformamide (DMF), methanol, acetone and pyridine of analytical grade were provided by Sinopharm Chemical Reagent Co., Ltd., China, and were used without further purification. γ-Aminopropyl-triethoxysilane (γ-APTEOS) purchased from GE silicones was used as the coupling agent. The other reagents, sodium hydroxide (NaOH) of analytical grade and 30% hydrogen peroxide (H₂O₂), 99.5% acetic acid (CH₃COOH), 65–68% hydrogen nitrate (HNO₃), 36% hydrochloric acid (HCl), xylene (Xyle), and disodium tetraborate decahydrate were purchased from Guang Zhou Jin hua du Chemical Reagent Co., Ltd. in China and were used as received without further purification.

Preparation of microcrystalline cellulose fibers (MCFs) from sisal

The treatment of the sisal was adapted from the work done by Vazquez et al.²³ where as received sisal fibers (see Fig. S1†) were preconditioned before cellulose extraction took place. Firstly, the fibers were washed with distilled water several times, then dried in a vacuum oven at 80 °C for 24 h. Then the fibers were chopped down to a length of 5–10 mm. Finally a de-waxing step was carried out; the fibers were extracted with toluene–ethanol (2 : 1 v/v) in a 250 ml three-necked flask for 6 h. The de-waxed fibers were then filtered, washed several times with ethanol and then dried. Subsequently the MCFs (see Fig. S2†) were extracted as follows; firstly the fibers were extracted with 0.1 M NaOH at 45 °C for 3 h and then treated with H₂O₂ for 3 h at 45 °C at pH = 11.5 (buffer solution). This was followed by submerging the fibers in a solution of NaOH and Na₂B₄O₇·10H₂O for 15 h at 28 °C and then HNO₃ at 120 °C for 15 min. The residual MCFs were washed with ethanol and distilled water several times and dried in a vacuum oven at 60 °C until they were a constant weight.

Silanization of the MCFs

Prior to silane modification, the MCFs were dried in a vacuum oven at 80 °C for 24 h to remove moisture absorbed on the surface. In a 250 ml three-necked flask, equipped with a mechanical stirrer and a reflux condenser, 1 g of dried MCFs and an appropriate amount of the silane, (as reported in the literature³⁵) were added into the flask and stirred in high purity ethanol at 60–95 °C for at least 5 h. After filtration, the crude product of the γ-APTEOS treated MCFs (denoted γ-APTEOS-MCFs) was washed several times with ethanol and then dried under vacuum at 80 °C for 12 h.

Growing hyperbranched aromatic polyamide on the surface of the MCFs

The DABA monomer was polymerized onto the nucleation sites (the active amino groups of the γ-APTEOS-MCFs). A typical experiment was carried out as follows: in a three-necked flask, 1 g of γ-APTEOS-MCFs and 1 g of DABA were placed in 20 ml of NMP and stirred. 2.5 g of pyridine and 2.6 g of TPP were added to the flask after the DABA had fully dissolved. The solution was heated to and kept at 100 °C and was stirred under nitrogen for 6 h and then allowed to cool to room temperature and was poured into 50 ml of methanol to precipitate the polymer coated fibers. To collect these coated fibers a solution of DMF containing 0.1% LiCl was added to the methanol. The polymer coated fibers were filtered and dried under vacuum at 90 °C for 12 h. The HBAP grown on the γ-APTEOS-MCFs is denoted as HBAP-MCFs in the following text. The suspensions of the MCFs, γ-APTEOS-MCFs and HBAP-MCFs (0.5 mg ml⁻¹) in ethanol for 2 hours after ultrasonic treatment are shown in Fig. S3.†

Preparation of the HBAP-MCF/epoxy composites

The HBAP-MCF/epoxy composites were prepared as follows: firstly, 30% of DDS (30 g/100 g of epoxy resin) was used to prepare the HBAP-MCF/epoxy composites. Secondly, the HBAP-MCFs were dispersed in acetone, then treated ultrasonically for 30 min. Thirdly, the HBAP-MCFs were added to a stoichiometric amount of epoxy resin under vacuum and were degassed at 120 °C for about 1 h. Then, the DDS was added. Subsequently, the solution was degassed for 30 min. Finally, the mixture was poured onto preheated stainless steel molds with a silicone resin. All of the samples were cured at 120 °C for 2 h, 160 °C for 2 h and 180 °C for 2 h. The content of HBAP-MCFs was calculated based on the amount of epoxy resin. Composites containing different weight fractions (0.5, 1.0, 1.5, 2.0%) of HBAP-MCFs were prepared. The experimental details of the synthesis of the epoxy composites are shown in Scheme 1.

Scheme 1 Preparation process of the epoxy composites.

Characterization

Fourier transformed infrared spectroscopy (FTIR) was recorded between 4000 and 450 cm⁻¹ on a PerkinElmer 1710 spectrophotometer using KBr pellets at room temperature. The size distribution of the fibers was determined by a dynamic/static light scattering method using a Zetasizer Nano ZS90 instrument (Malvern Instruments Co., UK) at 25 °C. The samples were dispersed in water under ultrasound for 15 min prior to the measurements. Dynamic mechanical analysis (DMA) was performed on a DMA Q800 dynamic mechanical analyzer (TA Instruments, USA), operating in a single cantilever bending mode at an oscillation frequency of 1.0 Hz. The testing temperature was set from room temperature to 250 °C at a heating rate of 3 °C min⁻¹. Thermal gravimetric analysis (TGA) was performed with a NETZSCH STA 449C instrument at a heating rate of 20 °C min⁻¹ from 50 to 600 °C under a nitrogen atmosphere. The morphology of the composites was examined using field emission scanning electron microscopy (FE-SEM, JEOL JSM-6701F, Japan) at an accelerating voltage of 20 kV, and the fractured surfaces of the composites were sputter-coated with gold before observation. The impact strength was measured on a tester of type XJJ-5 without a notch in the specimen according to the National Standard of China (GB1043-79). The specimen had a thickness of 4 mm, a width of 10 mm and a length of 80 mm. The tensile strength was examined on an electron omnipotence tester of type RGT-5. The tensile experiments used the universal testing machine with a crosshead speed of 2 mm min⁻¹, according to the National Standard of China (GB1040-92). All of the presented results are an average of five specimens.

Results and discussion

Characteristics of the surface modified MCFs

FTIR spectra of the MCFs, γ-APTEOS-MCFs and HBAP-MCFs are shown in Fig. 1. For the as received MCFs, the strong absorption at 3411 cm⁻¹ was attributed to the hydroxyl group (–OH) stretching. The peak at 1639 cm⁻¹ was associated with the –OH bending and thought to be due to absorbed water in the cellulose, which was hygroscopic and able to form strong hydrogen bonds. The peak at 1732 cm⁻¹ was attributed to C=O in hemicellulose, which was either present as an impurity, was formed from cellulose due to opening of some of its pyranose ring moieties, or from C–OH group oxidation. The peak at 2900 cm⁻¹ confirmed the presence of –CH₂ (stretching peak). The peak at 1430 cm⁻¹ was the –CH₂ and O–C–H (bending peaks). The spectrum of the γ-APTEOS-MCFs was very similar to that of the MCFs, but there were some differences between them. The peak around 3500 cm⁻¹ could be attributed to –NH₂ existence, and the peak became broad due to the association of –NH₂ with –OH. At 1525 cm⁻¹ the N–H bending peak was observed. The peak at 1469 cm⁻¹ corresponded to the peak associated with C–N–H. For the spectrum of the HBAP-MCFs, the peaks at 1660 and 1545 cm⁻¹ indicated the existence of –NHCO– stretching, and the peaks at 1601 and 1443 cm⁻¹ were characteristic of an aromatic structure. The peak at 1228 cm⁻¹ was due to C–N–H. These changes in the characteristic peaks proved that silane and hyperbranched aromatic polyamide had been grown successfully on the surface of the MCF. This was confirmed by the XPS results (see Fig. S4†).

Fig. 1 FTIR spectra of the MCFs, γ-APTEOS-MCFs, and HBAP-MCFs.

Histograms of the distribution of the MCFs, γ-APTEOS-MCFs, and HBAP-MCFs in water are given in Fig. 2, which show the diameter of pristine MCF in the 100–4000 nm size range. Some agglomeration leading to a second peak around 600–2500 nm was also observed. The γ-APTEOS-MCFs showed a small peak in the 500–1000 nm size range. Meanwhile, the HBAP-MCFs showed a bimodal distribution with one peak in the 400–1000 nm range and the other in the 200–400 nm range. Compared to pristine MCFs, the γ-APTEOS-MCFs and HBAP-MCFs diameter distributions have improved significantly. In other words, surface modification can significantly improve the dispersion in water.

Fig. 2 Distribution histograms of the diameters of the (a) MCFs, (b) γ-APTEOS-MCFs, and (c) HBAP-MCFs measured by dynamic light scattering.

The TGA curves of the MCFs, γ-APTEOS-MCFs, and HBAP-MCFs are shown in Fig. 3. The temperature for 10 wt% weight loss (T_10%) of the MCF was 293 °C. While the T_10% values of the γ-APTEOS-MCFs, and HBAP-MCFs were 296 and 227 °C, respectively. It was obvious why the γ-APTEOS-MCFs had a higher decomposition temperature as silane reduced the moisture content due to hydrophobicity on the surface via the long chain hydrocarbon attachments. In addition, these coupling agents penetrate the cell wall through surface pores and accumulate in the fibrillar regions, restricting further ingress of moisture.¹⁹ The present low molecular weight impurities in the HBAP-MCF sample lead to the HBAP-MCFs being less stable than MCFs. Hence the HBAP-MCFs had a lower temperature for its 10% weight loss than a corresponding sample of MCFs. The TGA confirmed the growing of the silane coupling agent on the MCFs and the HBAP molecules being successfully attached to the surface of the MCFs.

Fig. 3 TGA curves of the MCFs, γ-APTEOS-MCFs, and HBAP-MCFs measured under a N₂ atmosphere.

Morphology of the composites

The SEM images of the impact fracture surfaces of neat epoxy and the HBAP-MCF/epoxy composites after tensile tests are shown in Fig. 4. From Fig. 4(a) it can be observed that river patterns appear on the fracture surface of the neat epoxy. In addition, the mirror-like fracture surface was very smooth and the structural deformation showed a brittle failure of a homogeneous material, which is consistent with poor tensile strength. Fig. 4(b)–(e) present considerable differences in the fracture morphology of the epoxy composites in comparison to the neat epoxy, the surface of the composite is rough. This indicates that the epoxy matrix and the HBAP-MCF additive experienced interfacial adhesion due to the MCF surface functionalization.

Fig. 4 SEM images of the fractured surface of: (a) neat epoxy and epoxy-based composites with a HBAP-MCF content of: (b) 0.5 wt%, (c) 1.0 wt%, (d) 1.5 wt% and (e) 2.0 wt%.

Thermal properties of the composites

Fig. 5 shows the TG and DTG curves for the neat epoxy and its composites. As shown in Fig. 5(a), the curves showed only one thermal decomposition platform, indicating a one step process. All of the samples displayed similar degradation profiles, suggesting that the existence of the HBAP-MCFs did not significantly alter the degradation mechanism of the epoxy matrix.³⁰ Furthermore, the addition of the HBAP-MCFs caused the TGA curves to shift towards higher temperatures. A higher loading of the HBAP-MCFs did not bring further improvement to the thermal stability. Firstly, we attribute the improvement in thermal stability. The improvement was attributed to the formation of a char which acted as a mass transport barrier and an isolator between the bulk polymer matrix and the surface, where combustion occurred. The improvement was due to the bond formation between the HBAP-MCFs and the epoxy resin between the epoxide group in the epoxy matrix and the amide groups of the HBAP-MCFs. Finally, because of the low HBAP-MCFs content, no significant changes were observed in the thermal stability of the composites in the temperature range studied. The DTG curves indicate that the maximum degradation temperature (T_max) also increased upon the addition of the HBAP-MCF, as shown in Fig. 5(b). The T_max of the HBAP-MCF/epoxy composites decreased as the HBAP-MCF content increased. However, with a change from 0.5 wt% to 2.0 wt%, the T_max only varied a little (about 3 °C) as shown in the DTG curves.

Fig. 5 (a) TG and (b) DTG curves of neat epoxy and its composites.

The storage modulus (E′) of the neat epoxy and epoxy composites as a function of temperature is shown in Fig. 6(a). In the literature for epoxy resins, a rubbery state has been described that is expressed by a plateau in the storage modulus graph. The temperature at which this plateau extends, increases with the degree of cross-linking density, which is primarily observed in the elastomeric modulus of DMA. In the DMA test, as the temperature increases, the modulus passes through the glass transition region to the elastomeric region, where more segments in the chains move in a cooperative way. Above T_g, the network chains have sufficient thermal energy to undergo fast conformational changes using cooperative segmental motions, but cross-linking structures can prevent any long range flow.³⁵ In Fig. 6(a), the storage modulus of the neat epoxy was at a lower glass transition temperature due to the neat epoxy chains being able to move faster than the chains in the composites. Large reactive groups between the HBAP-MCFs and the epoxy exist, such as epoxy and amine groups, which can easily participate in cross-linking networks and, hence, impede the movement of the chains.

Fig. 6 (a) Storage modulus and (b) tan δ as a function of temperature of neat epoxy and its composites.

Fig. 6(b) shows tan δ as a function of temperature of the neat epoxy and its composites. The loss factor (tan δ) is defined as the ratio of the loss modulus to the storage modulus, which is very sensitive to changes in the solid material’s structure. The values of the tan δ peaks on the tan δ curves also gave the glass transition temperature (T_g) of the composites.³⁶ The neat epoxy's T_g was 154.65 °C. The T_g of the composites shifted to higher temperatures but no clear trend was discernable. The modulus depended on the dispersion of the HBAP-MCFs in the epoxy resin matrix. The HBAP-MCFs dispersed non-uniformly in the resin. This led to poorer interface adhesion between the HBAP-MCFs and the epoxy and higher damping at the interfaces, which were due mainly to the shear stress concentrations at the HBAP-MCF ends, in association with the additional viscoelastic energy dissipation in the epoxy matrix.³⁷ The tan δ values of the composites were consistent with the storage modulus behavior.

Mechanical properties of the composites

Representative tensile stress–strain curves of the neat epoxy and epoxy composites for 0.5, 1.0, 1.5, and 2.0 wt% of the HBAP-MCF additives are plotted in Fig. 7. The tensile specimens are shown in the inset of Fig. 7. From the stress–strain curves we observed that the materials extended in an almost linear fashion right up to their points of fracture, without plastic deformation. Compared with neat epoxy, the composite materials possessed higher elongation at break. In other words, brittle epoxy could be toughened significantly by the addition of the HBAP-MCF’s elastomeric phase. The HBAP-MCFs could make epoxy not only stronger but also tougher. Again no trend was discernible, indicating further issues with regard to sample uniformity and distribution of the HBAP-MCFs within the epoxy resin.

Fig. 7 Representative tensile stress–strain curves of neat epoxy and the epoxy composites.

Fig. 8, 9 and Table 1 show the impact strength, tensile strength and flexural strength of neat epoxy and the epoxy composites. The results presented in Fig. 8(a) and Table 1 show that the impact strength reached its highest level when the HBAP-MCF content was 2.0 wt%. The addition of even small amounts of HBAP-MCFs impacted the performance of the composite materials which increased by about 83.66%. The HBAP-MCFs played an important role in the impact resistance of the composites as they interacted with the crack formation in the matrix by acting as a stress transferring medium.³⁸ From Fig. 8(a) it is also observed that the tensile strength enhanced with the increase in HBAP-MCF content. For a HBAP-MCF content of 2.0 wt%, the maximum tensile strength value was achieved. For neat epoxy, the impact strength and the tensile strength were 17.5 kJ m⁻² and 59.4 MPa, respectively. The impact strength and tensile strength were enhanced effectively with the addition of HBAP-MCFs. At 2.0 wt% HBAP-MCF loading, the impact strength and tensile strength of the epoxy composites were 32.1 kJ m⁻² and 80.0 MPa, corresponding to increases of 83.4% and 34.7%, respectively. The Young's modulus and toughness (area under the stress–strain curve) of the neat epoxy and epoxy composites are shown in Fig. 8(b). The change in the trend of the Young's modulus plot was similar to the tensile strength plot. The toughness and the Young's modulus of the neat epoxy were 1.57 MJ m⁻³ and 556 MPa, respectively. For the epoxy with 2.0 wt% HBAP-MCF loading, the toughness and Young's modulus were 4.37 MJ m⁻³ and 695 MPa, corresponding to increases of 178.3% and 25%, respectively.

Fig. 8 (a) Impact strength and tensile strength and (b) corresponding toughness and Young's modulus of neat epoxy and epoxy composites.

Fig. 9 The flexural strength and flexural modulus of neat epoxy and epoxy composites.

Table 1 Influence of the amount of HBAP-MCFs on the impact and flexural properties of the composites

Samples (wt%)	Impact strength (kJ m^−2)	Flexural strength (MPa)	Flexural modulus (MPa)
0.0	17.50 ± 0.90	79.50 ± 2.5	1029 ± 63
0.5	23.58 ± 0.80	122.54 ± 2.8	2670 ± 70
1.0	26.11 ± 1.00	126.00 ± 2.3	2880 ± 65
1.5	26.47 ± 0.85	123.12 ± 2.0	2730 ± 84
2.0	32.14 ± 0.78	118.56 ± 2.9	2820 ± 80

---

Furthermore, Fig. 9 shows the flexural strength and flexural modulus of neat epoxy and the epoxy composites. With the increasing content of HBAP-MCFs, the flexural strength and flexural modulus of neat epoxy and epoxy composites were enhanced and achieved a maximum value at 1.0 wt% HBAP-MCF loading. The flexural strength and flexural modulus of the neat epoxy were 79.5 and 1029 MPa, respectively. For the epoxy with 1.0 wt% HBAP-MCF loading, the flexural strength and flexural modulus were 126 and 2880 MPa, increasing by 116.5% and 179.8%, respectively. The mechanical properties of the epoxy composites were enhanced by improving the interfacial interaction between the HBAP-MCFs and the epoxy matrix.

Conclusions

In this work, the effect of functionalized MCFs on the morphology and the mechanical and thermal properties of epoxy composites was investigated. The results showed that the HBAP-MCFs enhanced the thermal and mechanical properties due to the strong covalent bonds formed between the silane and the fibers, and the silane and the polyamide. This entailed a strong interface interaction between the fiber filler and epoxy matrix. For instance, the impact strength, tensile strength, Young's modulus and toughness of the composites with 2.0 wt% HBAP-MCF loading were 32.1 kJ m⁻², 59.4 MPa, 695 MPa, and 4.37 MJ m⁻³, respectively. These values represent improvements of 83.4%, 34.7%, 25%, and 178.3%, respectively, compared to neat epoxy. Moreover, the addition of HBAP-MCFs produced composites with higher thermal degradation temperatures and glass transition temperatures.

Acknowledgements

The authors gratefully acknowledge the financial support of the National Natural Science Foundation of China (no. 51163004 and 51303034), the Opening Funding of Guangxi Key Laboratory for Advance Materials and New Preparation Technology (12KF-8), the Innovation Team of Guangxi Universities' Talent Highland and the Guangxi Funds for Specially-appointed Expert.

References

1. D. Liu, J. Song, D. Anderson, P. Chang and Y. Hua, Cellulose, 2012, 19, 1449–1480.
2. Y. Zhang, X. Li, H. Li, M. E. Gibril, K. Han and M. Yu, RSC Adv., 2013, 3, 11732–11737.
3. T. G. Vijayabaskaran and S. Vitta, RSC Adv., 2013, 3, 12765–12773.
4. H. Song, Z. Luo, H. Zhao, S. Luo, X. Wu, J. Gao and Z. Wang, RSC Adv., 2013, 3, 11665–11675.
5. H. Sharma, E. Carmichael, M. Muhamad, D. McCall, F. Andrews, G. Lyons, W. McRoberts and P. Hornsby, RSC Adv., 2012, 2, 6424–6437.
6. Y. Li and A. J. Ragauskas, RSC Adv., 2012, 2, 3347–3351.
7. K. M. Gupta, Z. Hu and J. Jiang, RSC Adv., 2013, 3, 4425–4433.
8. K. Gupta, Z. Hu and J. Jiang, RSC Adv., 2013, 3, 12794–12801.
9. J. A. F. Gamelas, M. G. Evtyugina, I. Portugal and D. V. Evtuguin, RSC Adv., 2012, 2, 831–839.
10. D. Dai and M. Fan, RSC Adv., 2013, 3, 4659–4665.
11. A. Casas, S. Omar, J. Palomar, M. Oliet, M. V. Alonso and F. Rodriguez, RSC Adv., 2013, 3, 3453–3460.
12. E. E. M. Ahmad, A. S. Luyt and V. Djokovic, Polym. Bull., 2013, 70, 1265–1276.
13. J. M. L. Reis, Constr. Build. Mater., 2012, 37, 177–180.
14. P. Noorunnisa Khanam, H. P. S. Abdul Khalil, G. Ramachandra Reddy and S. Venkata Naidu, J. Polym. Environ., 2011, 19, 115–119.
15. T. P. Mohan and K. Kanny, Composites, Part A, 2012, 43, 1989–1998.
16. T. Madera-antana, H. Soto Valdez and M. Richardson, Polym. Eng. Sci., 2013, 53, 59–68.
17. K. Joseph, S. Varghese, G. Kalaprasad, S. Thomas, L. Prasannakumari, P. Koshy and C. Pavithran, Eur. Polym. J., 1996, 32, 1243–1250.
18. M. Idicula, A. Boudenne, L. Umadevi, L. Ibos, Y. Candau and S. Thomas, Compos. Sci. Technol., 2006, 66, 2719–2725.
19. A. Campos, G. D. Tonoli, J. Marconcini, L. C. Mattoso, A. Klamczynski, K. Gregorski, D. Wood, T. Williams, B.-S. Chiou and S. Imam, J. Polym. Environ., 2013, 21, 1–7.
20. A. C. H. Barreto, D. S. Rosa, P. B. A. Fechine and S. E. Mazzetto, Composites, Part A, 2011, 42, 492–500.
21. C. Albano, J. Reyes, M. Ichazo, J. Gonzalez and M. I. Chipara, Polym. Degrad. Stab., 2001, 73, 225–236.
22. Y. Li, Y. W. Mai and L. Ye, Compos. Sci. Technol., 2000, 60, 2037–2055.
23. J. I. Moran, V. A. Alvarez, V. P. Cyras and A. Vazquez, Cellulose, 2008, 15, 149–159.
24. J. Lu, P. Askeland and L. T. Drzal, Polymer, 2008, 49, 1285–1296.
25. Y. Xie, C. A. S. Hill, Z. Xiao, H. Militz and C. Mai, Composites, Part A, 2010, 41, 806–819.
26. M. Abdelmouleh, S. Boufi, M. N. Belgacem, A. P. Duarte, A. Ben Salah and A. Gandini, Int. J. Adhes. Adhes., 2004, 24, 43–54.
27. H. Alamri and I. M. Low, Polym. Test., 2012, 31, 620–628.
28. X. Peng, L. Zhong, J. Ren and R. Sun, Composites, Part A, 2010, 41, 1848–1856.
29. S. Kalia and S. Vashistha, J. Polym. Environ., 2012, 20, 142–151.
30. S. R. Lu, S. R. Li, J. H. Yu, D. Guo, R. H. Ling and B. Huang, Wood Mater. Sci. Eng., 2013, 8, 1–7.
31. N. Chand and D. Jain, Composites, Part A, 2005, 36, 594–602.
32. S. Spoljaric, A. Genovese and R. A. Shanks, Composites, Part A, 2009, 40, 791–799.
33. A. Ashori and A. Nourbakhsh, Composites, Part B, 2010, 41, 578–581.
34. J. Yu, X. Huang, C. Wu, X. Wu, G. Wang and P. Jiang, Polymer, 2012, 53, 471–480.
35. L. J. Luo, Y. Meng, T. Qiu and X. Y. Li, J. Appl. Polym. Sci., 2013, 130, 1064–1073.
36. J. Yu, X. Huang, L. Wang, P. Peng, C. Wu, X. Wu and P. Jiang, Polym. Chem., 2011, 2, 1380–1388.
37. S. R. Lu, S. R. Li, J. H. Yu, Z. K. Yuan and B. Qi, RSC Adv., 2013, 3, 8915–8923.
38. L. A. Pothan, Z. Oommen and S. Thomas, Compos. Sci. Technol., 2003, 63, 283–293.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra45732j

---