- Email: [email protected]
Preparation, structure and mechanical properties of all-hemp cellulose biocomposites
Composites Science and Technology 69 (2009) 2119–2126
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Preparation, structure and mechanical properties of all-hemp cellulose biocomposites Sirisart Ouajai 1, Robert A. Shanks * Applied Sciences, RMIT University, GPO Box 2476V, Melbourne 3001, Australia
a r t i c l e
i n f o
Article history: Received 27 August 2008 Received in revised form 2 May 2009 Accepted 8 May 2009 Available online 18 May 2009 Keywords: A. Polymer matrix composites (PMCs) A. Short fibre composites B. Fibre–matrix bond C. Casting D. X-ray diffraction
a b s t r a c t All-hemp (Cannabis Sativa L.) cellulose composites were prepared by a mechanical blending technique followed by hot pressing and water–ethanol regeneration. The alkali treated fibres were ground and sieved to a size ranging from 45 lm to 500 lm. Introduction of fibres into 12% w/v cellulose Nmethyl-morpholine-N-oxide (NMMO) solution was performed with low solution viscosity at 100 °C. The solid mixtures were cut and heat pressed between heated glass and PTFE plates at 85 °C to obtain a flat smooth-surfaced composite sheet of approximately 0.2 mm thickness. The cellulose was regenerated in a 50:50 water–ethanol mixture that subsequently removed NMMO and stabilizer (Irganox 1010, Ciba) from the composite. FTIR and X-ray diffraction measurements were performed to investigate the structural change of cellulose from fibre into partially regenerated composite. Composition and thermal stability of composites were investigated using thermogravimetry. A broadening of the scattering of the main crystalline plane (0 0 2) and a depression of the maximum degradation temperature of fibre were observed. The observations revealed a structural change in the fibres. The mechanical properties of composites depended on size, surface area, crystallinity and the structural swelling of fibres. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction There is increased interest in sustainable biocomposites obtained from renewable resources driven by the potential for reduction of petrochemical feedstock and by the hazards of traditional fibreglass composite waste management [1]. Widely-used semibiocomposites, i.e. polypropylene (PP) or polyethylene (PE)-cellulose are not sufficiently eco-friendly because of their petroleumbased and non-biodegradable matrix. Poly(hydroxybutyrate-cohydroxyvalerate) (PHBV) and poly(hydroxyoctoate) (PHO) composites containing cellulose are examples of ‘green’ composites [2–8]. Nevertheless, modification of cellulose is still required to improve the interfacial adhesion in composites. Differing proportions of polymer and matrix caused variation of biodegradation times [2]. Usually good interfacial adhesion between fibre and matrix is the expected requirement for composites. This requires compatibility between both materials, so composites derived from fibre and matrix of similar chemical structure are of increasing interest. The main rationale is the anticipated better interfacial adhesion. Consequences of using the same component but with different * Corresponding author. Tel.: +61 3 9925 2122; fax: +61 3 9925 3747. E-mail address: [email protected] (R.A. Shanks). 1 Present address: King Mongkut Institute of Technology, North Bangkok, Thailand. 0266-3538/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2009.05.005
physical properties for composite production such as all-PP, allPE fibre–matrix composites has been reported [9–11], where interfacial adhesion was improved without surface modification requirement. This initiated the idea to prepare a composite (matrix and filler) comprised of materials from a sustainable resource without an interfacial problem and totally biodegradable. Cellulose is the most abundant biomass resource. It has potential to be an alternative feedstock for composite manufacturing. Cellulose is not meltable however it can be dissolved in several solvents [12]. Instead of a melt mixing process, a composite containing all cellulose could be prepared by an addition of cellulose fibres into a cellulose solution or vice versa. An example of a composite that has all components derived from cellulose was prepared by the impregnation of 3% pulp cellulose in DMAc (LiCl) solution into a rami fibre [13], and the partial dissolution of microcrystalline cellulose powder in a similar solution [14]. The mechanical properties of this composite relied upon the similarity between the matrix and the fibre. An alternative process to prepare a cellulose solution by using N-methyl-morpholine-N-oxide (NMMO) was claimed to be environmentally friendly and the most economic method for producing regenerated cellulose. The basic closed loop process with full recovery of solvent was an advantage of this process. Moreover the waste water produced from this process was less harmful
2120
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
and completely treated by the ozonisation and activated sludge method [15,16]. The main product of the NMMO process is Lyocell or Tencel fibre with outstanding properties [17]. Interestingly, the process can be adopted to manufacture regenerated cellulose in film form [18]. Generally the mechanical properties of cellulose or carbohydrate films could be improved by fibre reinforcement [19,20]. Hence it is possible to prepare a composite which has fibres embedded in cellulose matrix via the NMMO process. The raw materials used in this process are mainly from wood pulp. Cellulose fibres from crop plants such as hemp have advantages compared with wood pulp because of their short growing period. Within the same time-period, the crop-sourced fibre yielded tentimes the cellulose of wood over the land area. In addition, the strength of hemp fibres is suitable for reinforcement of cellulose films derived from the NMMO process. In summary, all-cellulose composites may obtain several advantages from an environmental point of view. Firstly, hemp fibre is a sustainable reproducible crop fibre that has a suitable strength for composite applications. Secondly, solvent used in the NMMO process is almost totally recovered and waste from the process can be treated efficiently. Thirdly, the formation of composites with the same chemical structure can improve interfacial adhesion. Moreover the waste management of composites of the same material causes less problems and useful recyclability. The aim of this research was to prepare composites with matrix and fibres derived from the same cellulose resource (hemp fibres) but different in crystalline structure (cellulose I and cellulose II). The hemp fibres were treated with 8% NaOH solution, then were ground and sieved to a uniform size to ensure that non-cellulosic components have been removed and fibres could be dispersed homogeneously in the composites. Composites containing 40% variation in fibre sizes were prepared. Subsequently, they were fabricated and regenerated to form films or composite sheets. Presence of the fibres in the film matrix was expected to enhance shape stability, modulus and strength. An investigation of the structure, composition and mechanical properties of the composite were performed by using wide-angle Xray diffraction (WAXD), FTIR and DMA techniques, respectively. Thermal decomposition of composites was studied using thermogravimetry (TGA). Morphology of fibres and films were investigated using optical microscopy (OM), scanning electron microscopy (SEM) and Brunauer–Emmett–Teller (BET) gas adsorption techniques. The results obtained for composites and pure cellulose films have been compared. Furthermore, composite properties depending on the initial fibre physical properties (surface area, crystallinity) were explored and discussed. A mixture of Tencel fibre and cellulose powder (Ajax chemicals) were also prepared and used as a reference system for quantitative evaluation of the mixtures between cellulose I and cellulose II in composites.
2. Experimental 2.1. Materials Hemp (C. Sativa L) was obtained from Australian Hemp Resource and Manufacture (AHRM) with an average fibre length of 10 cm. Nmethyl-morpholine-N-oxide (NMMO) from Aldrich Chemical Company was used as received. Irganox 1010 (Ciba Specialty Chemicals) was used as the stabiliser. Standard grade, ashless cellulose powder was purchased from Ajax Chemicals (Sydney, Australia). Tencel fibre (Acordis Group) was obtained from school of Fashion and Textiles, RMIT University.
2.2. Pre-treatment 2.2.1. Acetone extraction and alkalisation The fibres were subjected to Soxhlet extraction with acetone for 3 h to remove any waxes present and then air-dried. Dried fibres (2.5 g) were mixed with 8% w/v NaOH aqueous solution (100 mL) then placed in an oven at 30 °C for 1 h to remove lignin associated with the fibres. The alkaline treated fibres were subsequently washed with running tap water followed by distilled water until no alkali was present in the wash water. 2.2.2. Fibre grinding and sieving The fibres were cut in an IKA MF10 cutting mill and sieved to provide a size range between 45 and 600 lm. Firstly all cut fibres were placed in a sieve of mesh number 36. A gentle shaking of sieve was conducted until no more fibre falling into sieve number 72. The same procedure was applied for smaller sieve opening size. Since the fibre was laid parallel to the surface of sieve. Fibres with length (L) (or diameter (D)) shorter than the sieve opening size passed through. The average fibre diameter and length measured from SEM images were used for the L/D ratio calculation. Therefore the fibre retained on the sieve mesh number 36, 72, 140 and 325 would provide a minimum L/D ratio. 2.3. Dissolution of fibre using NMMO solution The cellulose-NMMO solution containing 12% w/v cellulose and 0.6% w/v stabiliser was prepared using the ground 200 lm hemp fibres. A three neck 500 mL round bottom flask was equipped with a mechanical stirrer, a nitrogen purge gas inlet and a water-cooled condenser. NMMO powder (15.5 g), hemp fibres (2.1 g) and water (0.4 g) were transferred into this reactor separately then mixed gently without heating. An oil bath was used to heat the mixture to 140 °C. with stirrer speed of 200 rpm and a nitrogen purge used. The mixture became viscose at 135 °C. The stirring speed was then increased to 300 rpm and a vacuum was applied to remove any water from the reactor. The reactor temperature was maintained at 135–140 °C until the fibres dispersed in the viscous solution was completely dissolved. This could be observed from the presence of a residue of clear brown paste. The total time for the dissolution of fibres was 12–15 min. The heating bath was removed and the reactor was allowed to cool for 5 min. 2.4. Composites preparation Fibres were then added to the viscous paste and mixed for 3 min to provide 40% w/w fibre in dissolved cellulose. In order to control the fibre content in the composite, an inclusion of calculated amount of fibres (0.8 g) into regenerated cellulose matrix was all conducted in a similar manner, i.e. 3 min mixing time and stirring speed of 300 rpm. The composites were prepared by spreading NMMO–cellulose–water solution on a heated glass sheet at 85–90 °C, since the solutions were solid at ambient temperature, then covered with a PTFE sheet [21]. The melted solution was manually spread outward from centre to form a round flat sheet by an applied compression force from roller above the PTFE covered sheet. The thickness of composites was gradually reduced and was restricted by the diameter of the embedded fibre resulted in approximate 0.2 mm thickness sheet. An ever pressuring caused a separation of melt solution and a discontinuous sheet was obtained. Generally the thickness of unfilled composite was lower than the fibre-filled one. The films were kept under compression to achieve a flat and smooth surface. After the composite films were cooled they were
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
washed with water–ethanol (50:50) mixed solvent. The dissolved cellulose was regenerated and the solvent (NMMO) was subsequently removed from the composite into the water–ethanol. The composites were finally washed with deionized water. The composite films were left to dry overnight at room temperature (25 °C) between two glass plates separated with a paper spacer. This minimised the shrinkage of composite during water removal. The films obtained were sealed in polyethylene prior to further characterisation. 2.5. Characterisation 2.5.1. Surface pore structure analysis 2.5.1.1. Gas adsorption. Nitrogen adsorption isotherms were measured for each size of ground hemp fibre using a Micromeritics ASAP 2000 apparatus. The specific surface area and pore size distributions of fibres were calculated from approximately 1 g of fibre after vacuum drying at 105 °C. BET analysis was performed to obtain the surface area.
2121
3. Results and discussion 3.1. Infrared spectroscopy Fig. 1 shows the FTIR spectrum of NaOH treated hemp fibre compared with the NMMO regenerated cellulose film and a typical Tencel fibre known to be of cellulose II morphology. The 8% NaOH treated fibre showed the characteristic of cellulose I although without a carbonyl peak at 1740 cm 1, suggesting non-cellulosic components, that is pectin, hemicelluloses and lignin were removed by alkali. The increase in the intensity of the band at 879 cm 1 in regenerated cellulose indicated structural transformation from cellulose I into cellulose II. Cellulose I has a parallel closed packed structure. The vibration of b-1,4-glycosidic linkage was limited and less intensity obtained. Transformation to cellulose II caused a structure with reduced packing thus the intensity of vibration band at 879 cm 1 increased. This was observed in a regenerated film from hemp and Tencel fibre. 3.2. Fibre characteristics for the all-cellulose composites
2.5.1.2. Microstructure and surface morphology. Optical microscopy (OM) and scanning electron microscopy (SEM) were used to observe the microstructure and the surface morphology of treated and untreated cellulose fibres. A Nikon Labophot 2 polarised optical microscope, using a 50 times magnification, was connected with a computer interface and digital camera. The SEM was a Phillips XL 30 Oxford 6650 with an acceleration voltage of 142 eV. The composite specimens were coated with gold to provide a 20 nm gold layer thickness using a vacuum sputter coater. 2.5.1.3. FTIR measurements. A mixture of 5.0 mg of dried fibres dispersed in 200 mg of KBr was pressed into a disk for FTIR measurement performed using a Perkin–Elmer 2000 spectrometer. 100 scans were taken for each sample with a resolution of 2 cm 1. 2.5.2. Wide angle X-ray diffraction The fibres (70 mg) were cut and pressed into a disk using a cylindrical steel mold (diameter = 1.3 cm) with an applied pressure of about 7000 kg/cm2 using a laboratory press. The composite sample with dimensions 10 25 0.2 mm was cut from the composite sheet. Ni-filtered CuKa radiation (k = 0.1542 nm) was generated at 40 kV and 35 mA using a Bruker AXS D8. The X-ray diffractograms were recorded from 5 to 60° of 2h (Bragg angle) by a goniometer equipped with scintillation counter at a scanning speed of 0.02°/s and sampling rate of 2 data/s.
Fig. 2a–d shows ground hemp fibre. The original length of sourced fibre was 10 cm which is suitable for a woven composite fabrication. A shorter length of fibre is more applicable for a nonwoven composite. The dimension of the fibre influenced the mechanical properties of the composite [2], therefore the fibre was cut and sieved. The mesh number presented in all figures was assigned for the fibre that remained on the sieve; the proportion of various fibre lengths in ground fibres is shown in Table 1. A length of 100 lm was the most frequently obtained fraction. The three different lengths of 500, 100 and 45 lm fibres were selected as filler for all-cellulose composite. The 200 lm fibre was used to prepare cellulose solution. The fibres had different surface area and pore structure (Table 1) as a consequence of their length and the cutting process. The surface area of 100 lm fibre was lower than 500 lm and 45 lm fibres, respectively. The 500 lm fibres showed partial disintegration of a fibre bundle (Fig. 2d). The 100 lm fibres (Fig. 2b) showed an even size distribution with a clean cut of fibre and less bundle disintegration.
2.5.3. Thermogravimetry The temperature programs for thermogravimetry were from 35 to 850 °C at heating rates of 2.5–30 °C/min using a Perkin–Elmer TGA7 instrument. The measurements were conducted under nitrogen (20 mL/min), which was switched to air at 700 °C. 2.5.4. Mechanical properties The mechanical properties of all-hemp cellulose composites were determined by static tensile testing. Gauge length of the composite sheet was set at 10 mm. The tests were performed using a Rheometric DMTA IV at a speed of 0.03 mm min 1. The cross-section area of specimen was determined using a Mitutoyo digital micrometer. An average value was taken from at least ten specimens of each specimen. Tensile creep-recovery was tested using a constant stress of 0.8 MPa for 900 s followed by 2700 s of recovery with 500 Pa applied stress. The strain was measured by the displacement of the upper and lower clamps. The gap between clamps is set at 10 mm. The specimen dimension is 2 10 0.2 mm (width:length:thickness).
Fig. 1. FTIR spectrum of: (a) regenerated hemp film, (b) Tencel fibre and (c) 8% NaOH treated hemp fibres.
2122
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
Fig. 2. SEM images of ground hemp fibres of different sizes: (a) 45 lm (mesh 325), (b) 100 lm (mesh 140), (c) 200 lm (mesh 72) and (d) 500 lm (mesh 36).
Table 1 Surface area and size distribution of ground fibres. Hemp fibres
BET Specific surface area (m2/g)
Ground mixed size (45–600 lm) 600 500 200 100 45
3.3. Wide angle X-ray diffraction measurement Sieve size distribution (%)
Pore size (Å)
Volume (cm3/g)
0.787
76.8
0.00151
100
N/A 0.939 0.601 0.662 1.018
N/A 35.8 138.8 95.0 98.3
N/A 0.00084 0.00208 0.00157 0.00250
4.7 3.4 30.1 50.7 11.1
Fig. 3a–d exhibits a surface considered to be smooth, for an all-cellulose composite to which calendering was not applied. Fibres were completely covered by a regenerated cellulose matrix. A light brown translucent film was a characteristic of the matrix. Therefore further observation using an optical microscope was conducted to obtain information regarding internal composite structure. Dispersion of fibres within the composite can be seen from optical microscope images (Fig. 4a–d), some fibres are indicated by arrows. Small internal voids, indicated by circles drawn on the images, were distributed throughout the composite body. Escape of solvent (NMMO) during the regeneration step may be a reason. The pores may have occurred since the solution was required to be at high temperature during regeneration [22]. The fibres exhibited random orientation throughout the composite especially the composite containing 45 lm fibres. Interestingly the unfilled cellulose contained a small amount of very short fibres (Fig. 4a. This indicated partial dissolution of the fibre into NMMO solution. The insoluble fraction of fibre was determined using the following X-ray technique.
Fig. 5 shows the X-ray diffractograms of composites. The diffractograms exhibited a structural combination of cellulose I and cellulose II. A slight difference in the intensity of the main crystalline plane (0 0 2) at 22.7° in cellulose I can be observed. This result was attributed to variation of fibre (cellulose I structure) composition in the composite. Cellulose II structure showed a main crystalline plane at a lower Bragg angle, 20.4°. The X-ray scattering measurement, due to the main diffraction planes present at different diffraction angles can be applied quantitatively. The intensity ratio of these crystalline planes was calculated to obtain the fraction of each composition in the all-cellulose composite by Gindl and Keckes assuming a linear relationship [14]. A mixture of Ajax cellulose (pure cellulose I) and Tencel fibre (regenerated cellulose II) was prepared at various compositions to use as a reference material. Fig. 6 shows the X-ray diffractogram of the mixture of Ajax cellulose and Tencel fibre. The ratio of intensity at 22.7° and 20.4° was calculated to estimate the content of each fibre. The relationship between the intensity ratio and cellulose fraction is shown in Fig. 7 with a slight negative deviation from a linear relation. The diffractogram of a reference mixture with Tencel composition of 60% or greater was similar to that of all-cellulose composites. The fractions of regenerated cellulose were calculated and shown in Table 2. The unfilled regenerated film contained 85% of regenerated cellulose, confirmed by the presence of undissolved fibre in this material as observed by optical microscopy. The composite containing 500 lm fibres maintained the original composition of fibre at 60%. A slight reduction of fibre fraction was found in the composites containing 45 lm and 100 lm. It is proposed that this can be explained if the cellulose fibres swelled and were partially dissolved during the mixing to prepare composites. The solvent penetrated between the molecular cellulose sheets [23], although the temperature at mixing was lower than
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
2123
Fig. 3. SEM images of all-cellulose composites: (a) regenerated, (b) 40% of 45 lm, (c) 40% of 100 lm and (d) 40% of 500 lm.
Fig. 4. Optical microscopy (magnification 50) of composites with different fibre lengths: (a) no added fibres, (b) 45 lm, (c) 100 lm and (d) 500 lm.
130 °C at which cellulose started to dissolve and the 12% w/v of cellulose in NMMO solution was not the saturated concentration. Mechanical shearing was applied. Hence some degree of swelling or dissolution may occur during the mixing, especially with a low crystallinity fibre. Molecules of coagulants (water–ethanol, (50:50) solution) diffused into the cast composite sheet. The
NMMO molecules that were restricted between solvated cellulose chains interacted with diffused NNMO molecules and dissolution of NMMO into coagulant occurred. Due to the high degree of solvation of cellulose chains in the irreversible swelling process, the original cellulose I crystal structure was almost completely transformed into the cellulose II modification [24].
2124
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
Fig. 5. X-ray diffractogram of all-cellulose composite containing: (a) no added fibres, (b) 45 lm, (c) 100 lm and (d) 500 lm fibre.
100 lm and 500 lm fibres presented a maximum weight loss at 401.4, 396.6 and 404.3 °C, respectively. Pretreatment with 8% NaOH solution improved the purity of fibre by removal of non-cellulosic components. Therefore the fibre bundle was disintegrated into fine fibres that were purer and of increased structural order. After the grinding and sieving processes, a large fraction of this finer fibre was found in mesh 36 (500 lm) than mesh 325 (45 lm) and mesh 140 (100 lm), respectively. Hence the difference in thermal stability of fibre was based on its properties with respect to fibre size. The lower degradation temperature of 100 lm fibres represented a slightly lower structural order of this fibre than the 45 lm and 500 lm fibres, respectively. The all-cellulose composite showed a broader degradation temperature (Fig. 8b). This signified the combination of two different crystalline structures in the composite. Moreover the degradation temperature of the fibre component was shifted to a lower temperature. A greater reduction of maximum degradation temperature was found in 100 lm fibre. A lower structural order indicated by a low differential thermogravimetry (DTG) peak of 100 lm fibre may be because it was easier to swell by NMMO solution than the 45 lm and 500 lm fibres. The result was consistent with the reduction in cellulose I fibre composition found by X-ray measurement (Table 2). There was no reduction of 500 lm fibre fraction measured by X-ray therefore the reduction of the maximum degradation temperature found in 500 lm fibre may be attributed to the structural swelling rather than dissolution or transformation into cellulose II structure. Dissolution and/or swelling of cellulose fibre took place during the shear-mixing step. A reduction of time and temperature of mixing was a suitable means to control fibre fraction at a desired composition, however the increase in viscosity of solution must be considered. 3.5. Mechanical properties
Fig. 6. X-ray diffractogram of a mixture of Ajax (cellulose I) and Tencel (cellulose II) at different compositions.
Fig. 7. The intensity ratio of major diffraction plane of Ajax fibre and Tencel fibre.
3.4. Thermogravimetry of fibres and composites The differential thermograms of regenerated cellulose and ground fibre are shown in Fig. 8a. The regenerated cellulose from hemp exhibited a maximum weight loss at 359.3 °C. The 45 lm,
3.5.1. Tensile mechanical properties Mechanical properties of the composites are shown in Table 2. The error in the results at the 95% confidence level is provided alongside the data for the mechanical properties in Table 2. The composition of fibres in composites was fixed at 40% w/w for all composites so mechanical properties depended mainly on the physical properties of the fibres. The relatively poor performance of the composites was due to the small-scale laboratory fabricating technique that cannot create compacted structures for the composites and left considerable void contents in the composite. The mechanical properties of each matrix were improved as seen from an increase in modulus and yield stress of each composite except for a slight reduction in modulus of the composite containing 500 lm fibre. The composite containing 45 lm fibre showed the highest mechanical properties improvement. This behaviour signified that the greater surface area and pore volume of 45 lm fibre played an important role in the mechanical properties of this composite. Moreover a shorter fibre permitted better dispersion. A change of fibre composition may cause some variation to the mechanical properties. Since the reduction of fibre contents by partial dissolving of cellulose might cause interface swelling of fibre. This would enhance the interface adhesion between cellulose fibre and matrix. Therefore the high degree of fibre dissolution that found in the composite containing 45 lm and 100 lm fibre showed an increase in mechanical properties. As the composites were prepared by hot compaction and direct regeneration into water–ethanol solution without any calendaring, the imperfection of composite bulk structure from a presence of internal voids may depress mechanical properties from their optimum value. The voids in composites with different fibre content were observed in the optical microscope images. A high void content and size was observed in the 500 lm composite and the voids declined in
2125
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126 Table 2 Tensile mechanical properties and X-ray intensity (I22.7/I20.4) ratio of all-cellulose composites with different fibre length. Composites
Modulus (GPa)
Yield stress (MPa)
Elongation at break (%)
I22.7/I20.4 (Cellulose II content)
Regenerated film +45 lm composite +100 lm composite +500 lm composite
1.44 ± 0.17 1.82 ± 0.27 1.58 ± 0.37 1.33
23.2 ± 3.13 28.9 ± 3.63 24.7 ± 8.32 25.3 ± 2.81
21.2 ± 3.76 20.8 ± 4.66 26.0 ± 7.47 20.6 ± 4.34
1.04 1.44 1.32 1.83
(84%) (70%) (73%) (60%)
Table 3 Model fitted creep parameters of all-cellulose composites with different fibre lengths. Composites
Maxwell modulus (GPa)
Maxwell viscosity (GPa s)
Voigt modulus (GPa)
Voigt viscosity (GPa s)
Relaxation time (s)
Regenerated film +45 lm composite +100 lm composite +500 lm composite
1.6 32.5 9.6 10.3
2.15 0.48 0.66 0.68
1.7 65.6 6.36 4.3
41.0 303.7 111.3 130.2
23.9 4.6 17.5 30.5
Fig. 9. Creep of all-cellulose composites with an applied stress of 0.8 MPa for 900 s.
Fig. 8. TGA results of (a) different sizes ground fibres and (b) all-cellulose composites.
100 lm and 45 lm composites, respectively. The presence of large void sizes found in the 500 lm composite drastically reduced the mechanical properties. 3.5.2. Tensile creep-recovery The creep-recovery experiment was conducted to investigate the shape stability improvement by addition of fibres into cellulose. The four-element model of Maxwell and Kelvin-Voigt [25] was used to describe both creep and recovery of this composite. The model parameters were computed and presented in Table 3. Inclusion of fibre caused a reduction of the simultaneous creep strain and/or recovery strain (Fig. 9). The simultaneous recovery strain was used to calculate the modulus (spring constant) of the first element using an applied stress of 0.8 MPa. The modulus of the spring element was greatly improved, indicating the ability of fibres to retain their original shape under a rapidly applied tensile stress. The highest modulus was found in the 45 lm fibre composite which demonstrated excellent interfacial adhesion. The partial dissolution and inter-planar swelling of fibre may enhance
the compatibility between fibre and matrix. When fibre was not included there was more creep (3 times) under constant load than was measured for the composites. Nevertheless the unrecovered strain of unfilled cellulose was less than composites resulting in a high viscosity of the dashpot element. The creep zone was used to calculate the viscosity of KelvinVoigt elements. The slope of this creep section indicated a flow of composite structure under tensile load. All fibre-filled composites showed a similar slope but lower than unfilled cellulose. This signified retardation of the matrix flow by fibre under load. The creep modulus (spring constant in Voigt element) was calculated by using the recovery strain and the applied stress of 0.8 MPa. The recovered strain calculation was started 1–2 s after the applied tensile stress was released, as indicated by a deviation from a linear vertical line of the recovery curve. Recovery was observed to be complete when the measured strain reached a plateau. The significant high modulus in the 45 lm fibre composite resulted from a small recovery strain compared with regenerated cellulose matrix (Fig. 10a). Since the retardation time is g/E, in order to obtain the viscosity of the dashpot element the retardation or relaxation time was required. The retardation or relaxation time is the time for the Voigt element to deform to 63.21% of its total deformation or recover to 36.79% of its original value. In this case the relaxation from recovered strain was preferred since the permanent (unrecovered) strain
2126
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
lose–NMMO–water solution. The dissolution of fibre transformed cellulose I morphology to cellulose II, as confirmed by X-ray scattering. An improvement of mechanical properties of regenerated cellulose films was obtained by addition of fibres, and the extent of improvement depended on the crystallinity and surface area of the fibres. Mixing caused partial dissolution and swelling of the fibre structure, especially lower crystallinity fibres. TGA and X-ray scattering techniques were particularly useful for indicating the structural change in composites and to calculate the composition of cellulose with different crystalline structure quantitatively. Acknowledgements The authors gratefully thank King Mongkut Institute of Technology, North Bangkok (KMITNB), Thailand for a PhD scholarship and Lu Guang Chen for an optical microscope measurement. References
Fig. 10. Recovery of (a) 45 lm fibre composite and (b) regenerated hemp cellulose after removing an applied stress of 0.8 MPa; (where s is experimental and — indicates fitted line).
could be eliminated. The selected recovery zone of the regenerated cellulose and composite was fitted to calculate the relaxation time of the composite (Fig. 10a and b). The modulus and viscosity of Voigt element of composites was increased (Table 3) signifying that inclusion of fibre improved the composite properties of especially that containing 45 lm fibres. 4. Conclusions All-cellulose composites were prepared by solution-slurry blending between ground hemp (C. Sativa L.) fibres and cellu-
[1] Mohanty AK, Misra M, Drzal LT. J Polym Environ 2002;10:19–26. [2] Shibata M, Takachiyo K-I, Ozawa K, Yosomiya R, Takeishi H. J Appl Polym Sci 2002;85:129–38. [3] Dubief D, Samain E, Dufresne A. Macromolecules 1999;32:5765–71. [4] Zhang Ming Qiu, Rong Min Zhi, Lu Xun. Compos Sci Technol 2005;65:2514–25. [5] Zhang MQ, Rong MZ, Lu X. All-green composites. In: Fakirvor S, Bhattacharyya D, editors. Handbook of engineering biopolymers, homopolymers, blends, and composites. Hanser Publishers; 2007. p. 747–72. [6] Lu Xun, Zhang Ming Qiu, Rong Min Zhi, Shi Guang, Yang Gui Cheng. Compos Sci Technol 2003;63:177–86. [7] Lu Xun, Zhang Ming Qiu, Rong Min Zhi, Yue Da Lei, Yang Gui Cheng. Compos Sci Technol 2004;64:1301–10. [8] Lu Xun, Zhang Ming Qiu, Rong Min Zhi, Yang Gui Cheng. Polym Adv Technol 2003;14:676–85. [9] Flores A. J Macromol Sci Part B: Phys 2001;40:749–61. [10] Houshyar S, Shanks RA, Hodzic A. Polym Test 2005;24:257–64. [11] Lacroix FV, Loos J, Schulte K. Polymer 1999;40:843–7. [12] Woodings C. Regenerated cellulose fibres. Cambridge: Wood Head Publishing Ltd.; 2001. [13] Nishino T, Matsuda I, Hirao K. Macromolecules 2004;37:7683–7. [14] Gindl W, Keckes J. Polymer 2005;46:10221–5. [15] Stockinger H, Kut OM, Heinzle E. Water Res 1996;30:1745–8. [16] Meister G, Wechsler M. Biodegradation 1998;9:91–102. [17] Kreze T, Strnad S, Stana-Kleinschek K, Ribitsch V. Mater Res Innov 2001;4:107–14. [18] Fink H-P, Weigel P, Ganster J, Rihm R, Puls J, Sixta H, et al. Cellulose 2004;11:85–98. [19] Dufresne A, Vignon MR. Macromolecules 1998;31:2693–6. [20] Curvelo AAS, de Carvalho AJF, Agnelli JAM. Carbohyd Polym 2001;45:183–8. [21] Chanzy H, Peguy A, Chaunis S, Monzie P. J Polym Sci Polym Phys Ed 1980;18:1137–44. [22] Kim DB, Lee WS, Jo SM, Lee YM, Kim BC. J Appl Polym Sci 2002;83:981–9. [23] Michael M, Ibbett RN, Howarth OW. Cellulose 2000;7:21–33. [24] Chanzy H, Noe P, Pailet M, Smith P. J Polym Sci Appl Polym Sym 1983;37:239–59. [25] Menard KP. Dynamic mechanical analysis a practical introduction. Florida: CRC Press LLC; 1999.
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Preparation, structure and mechanical properties of all-hemp cellulose biocomposites Sirisart Ouajai 1, Robert A. Shanks * Applied Sciences, RMIT University, GPO Box 2476V, Melbourne 3001, Australia
a r t i c l e
i n f o
Article history: Received 27 August 2008 Received in revised form 2 May 2009 Accepted 8 May 2009 Available online 18 May 2009 Keywords: A. Polymer matrix composites (PMCs) A. Short fibre composites B. Fibre–matrix bond C. Casting D. X-ray diffraction
a b s t r a c t All-hemp (Cannabis Sativa L.) cellulose composites were prepared by a mechanical blending technique followed by hot pressing and water–ethanol regeneration. The alkali treated fibres were ground and sieved to a size ranging from 45 lm to 500 lm. Introduction of fibres into 12% w/v cellulose Nmethyl-morpholine-N-oxide (NMMO) solution was performed with low solution viscosity at 100 °C. The solid mixtures were cut and heat pressed between heated glass and PTFE plates at 85 °C to obtain a flat smooth-surfaced composite sheet of approximately 0.2 mm thickness. The cellulose was regenerated in a 50:50 water–ethanol mixture that subsequently removed NMMO and stabilizer (Irganox 1010, Ciba) from the composite. FTIR and X-ray diffraction measurements were performed to investigate the structural change of cellulose from fibre into partially regenerated composite. Composition and thermal stability of composites were investigated using thermogravimetry. A broadening of the scattering of the main crystalline plane (0 0 2) and a depression of the maximum degradation temperature of fibre were observed. The observations revealed a structural change in the fibres. The mechanical properties of composites depended on size, surface area, crystallinity and the structural swelling of fibres. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction There is increased interest in sustainable biocomposites obtained from renewable resources driven by the potential for reduction of petrochemical feedstock and by the hazards of traditional fibreglass composite waste management [1]. Widely-used semibiocomposites, i.e. polypropylene (PP) or polyethylene (PE)-cellulose are not sufficiently eco-friendly because of their petroleumbased and non-biodegradable matrix. Poly(hydroxybutyrate-cohydroxyvalerate) (PHBV) and poly(hydroxyoctoate) (PHO) composites containing cellulose are examples of ‘green’ composites [2–8]. Nevertheless, modification of cellulose is still required to improve the interfacial adhesion in composites. Differing proportions of polymer and matrix caused variation of biodegradation times [2]. Usually good interfacial adhesion between fibre and matrix is the expected requirement for composites. This requires compatibility between both materials, so composites derived from fibre and matrix of similar chemical structure are of increasing interest. The main rationale is the anticipated better interfacial adhesion. Consequences of using the same component but with different * Corresponding author. Tel.: +61 3 9925 2122; fax: +61 3 9925 3747. E-mail address: [email protected] (R.A. Shanks). 1 Present address: King Mongkut Institute of Technology, North Bangkok, Thailand. 0266-3538/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2009.05.005
physical properties for composite production such as all-PP, allPE fibre–matrix composites has been reported [9–11], where interfacial adhesion was improved without surface modification requirement. This initiated the idea to prepare a composite (matrix and filler) comprised of materials from a sustainable resource without an interfacial problem and totally biodegradable. Cellulose is the most abundant biomass resource. It has potential to be an alternative feedstock for composite manufacturing. Cellulose is not meltable however it can be dissolved in several solvents [12]. Instead of a melt mixing process, a composite containing all cellulose could be prepared by an addition of cellulose fibres into a cellulose solution or vice versa. An example of a composite that has all components derived from cellulose was prepared by the impregnation of 3% pulp cellulose in DMAc (LiCl) solution into a rami fibre [13], and the partial dissolution of microcrystalline cellulose powder in a similar solution [14]. The mechanical properties of this composite relied upon the similarity between the matrix and the fibre. An alternative process to prepare a cellulose solution by using N-methyl-morpholine-N-oxide (NMMO) was claimed to be environmentally friendly and the most economic method for producing regenerated cellulose. The basic closed loop process with full recovery of solvent was an advantage of this process. Moreover the waste water produced from this process was less harmful
2120
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
and completely treated by the ozonisation and activated sludge method [15,16]. The main product of the NMMO process is Lyocell or Tencel fibre with outstanding properties [17]. Interestingly, the process can be adopted to manufacture regenerated cellulose in film form [18]. Generally the mechanical properties of cellulose or carbohydrate films could be improved by fibre reinforcement [19,20]. Hence it is possible to prepare a composite which has fibres embedded in cellulose matrix via the NMMO process. The raw materials used in this process are mainly from wood pulp. Cellulose fibres from crop plants such as hemp have advantages compared with wood pulp because of their short growing period. Within the same time-period, the crop-sourced fibre yielded tentimes the cellulose of wood over the land area. In addition, the strength of hemp fibres is suitable for reinforcement of cellulose films derived from the NMMO process. In summary, all-cellulose composites may obtain several advantages from an environmental point of view. Firstly, hemp fibre is a sustainable reproducible crop fibre that has a suitable strength for composite applications. Secondly, solvent used in the NMMO process is almost totally recovered and waste from the process can be treated efficiently. Thirdly, the formation of composites with the same chemical structure can improve interfacial adhesion. Moreover the waste management of composites of the same material causes less problems and useful recyclability. The aim of this research was to prepare composites with matrix and fibres derived from the same cellulose resource (hemp fibres) but different in crystalline structure (cellulose I and cellulose II). The hemp fibres were treated with 8% NaOH solution, then were ground and sieved to a uniform size to ensure that non-cellulosic components have been removed and fibres could be dispersed homogeneously in the composites. Composites containing 40% variation in fibre sizes were prepared. Subsequently, they were fabricated and regenerated to form films or composite sheets. Presence of the fibres in the film matrix was expected to enhance shape stability, modulus and strength. An investigation of the structure, composition and mechanical properties of the composite were performed by using wide-angle Xray diffraction (WAXD), FTIR and DMA techniques, respectively. Thermal decomposition of composites was studied using thermogravimetry (TGA). Morphology of fibres and films were investigated using optical microscopy (OM), scanning electron microscopy (SEM) and Brunauer–Emmett–Teller (BET) gas adsorption techniques. The results obtained for composites and pure cellulose films have been compared. Furthermore, composite properties depending on the initial fibre physical properties (surface area, crystallinity) were explored and discussed. A mixture of Tencel fibre and cellulose powder (Ajax chemicals) were also prepared and used as a reference system for quantitative evaluation of the mixtures between cellulose I and cellulose II in composites.
2. Experimental 2.1. Materials Hemp (C. Sativa L) was obtained from Australian Hemp Resource and Manufacture (AHRM) with an average fibre length of 10 cm. Nmethyl-morpholine-N-oxide (NMMO) from Aldrich Chemical Company was used as received. Irganox 1010 (Ciba Specialty Chemicals) was used as the stabiliser. Standard grade, ashless cellulose powder was purchased from Ajax Chemicals (Sydney, Australia). Tencel fibre (Acordis Group) was obtained from school of Fashion and Textiles, RMIT University.
2.2. Pre-treatment 2.2.1. Acetone extraction and alkalisation The fibres were subjected to Soxhlet extraction with acetone for 3 h to remove any waxes present and then air-dried. Dried fibres (2.5 g) were mixed with 8% w/v NaOH aqueous solution (100 mL) then placed in an oven at 30 °C for 1 h to remove lignin associated with the fibres. The alkaline treated fibres were subsequently washed with running tap water followed by distilled water until no alkali was present in the wash water. 2.2.2. Fibre grinding and sieving The fibres were cut in an IKA MF10 cutting mill and sieved to provide a size range between 45 and 600 lm. Firstly all cut fibres were placed in a sieve of mesh number 36. A gentle shaking of sieve was conducted until no more fibre falling into sieve number 72. The same procedure was applied for smaller sieve opening size. Since the fibre was laid parallel to the surface of sieve. Fibres with length (L) (or diameter (D)) shorter than the sieve opening size passed through. The average fibre diameter and length measured from SEM images were used for the L/D ratio calculation. Therefore the fibre retained on the sieve mesh number 36, 72, 140 and 325 would provide a minimum L/D ratio. 2.3. Dissolution of fibre using NMMO solution The cellulose-NMMO solution containing 12% w/v cellulose and 0.6% w/v stabiliser was prepared using the ground 200 lm hemp fibres. A three neck 500 mL round bottom flask was equipped with a mechanical stirrer, a nitrogen purge gas inlet and a water-cooled condenser. NMMO powder (15.5 g), hemp fibres (2.1 g) and water (0.4 g) were transferred into this reactor separately then mixed gently without heating. An oil bath was used to heat the mixture to 140 °C. with stirrer speed of 200 rpm and a nitrogen purge used. The mixture became viscose at 135 °C. The stirring speed was then increased to 300 rpm and a vacuum was applied to remove any water from the reactor. The reactor temperature was maintained at 135–140 °C until the fibres dispersed in the viscous solution was completely dissolved. This could be observed from the presence of a residue of clear brown paste. The total time for the dissolution of fibres was 12–15 min. The heating bath was removed and the reactor was allowed to cool for 5 min. 2.4. Composites preparation Fibres were then added to the viscous paste and mixed for 3 min to provide 40% w/w fibre in dissolved cellulose. In order to control the fibre content in the composite, an inclusion of calculated amount of fibres (0.8 g) into regenerated cellulose matrix was all conducted in a similar manner, i.e. 3 min mixing time and stirring speed of 300 rpm. The composites were prepared by spreading NMMO–cellulose–water solution on a heated glass sheet at 85–90 °C, since the solutions were solid at ambient temperature, then covered with a PTFE sheet [21]. The melted solution was manually spread outward from centre to form a round flat sheet by an applied compression force from roller above the PTFE covered sheet. The thickness of composites was gradually reduced and was restricted by the diameter of the embedded fibre resulted in approximate 0.2 mm thickness sheet. An ever pressuring caused a separation of melt solution and a discontinuous sheet was obtained. Generally the thickness of unfilled composite was lower than the fibre-filled one. The films were kept under compression to achieve a flat and smooth surface. After the composite films were cooled they were
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
washed with water–ethanol (50:50) mixed solvent. The dissolved cellulose was regenerated and the solvent (NMMO) was subsequently removed from the composite into the water–ethanol. The composites were finally washed with deionized water. The composite films were left to dry overnight at room temperature (25 °C) between two glass plates separated with a paper spacer. This minimised the shrinkage of composite during water removal. The films obtained were sealed in polyethylene prior to further characterisation. 2.5. Characterisation 2.5.1. Surface pore structure analysis 2.5.1.1. Gas adsorption. Nitrogen adsorption isotherms were measured for each size of ground hemp fibre using a Micromeritics ASAP 2000 apparatus. The specific surface area and pore size distributions of fibres were calculated from approximately 1 g of fibre after vacuum drying at 105 °C. BET analysis was performed to obtain the surface area.
2121
3. Results and discussion 3.1. Infrared spectroscopy Fig. 1 shows the FTIR spectrum of NaOH treated hemp fibre compared with the NMMO regenerated cellulose film and a typical Tencel fibre known to be of cellulose II morphology. The 8% NaOH treated fibre showed the characteristic of cellulose I although without a carbonyl peak at 1740 cm 1, suggesting non-cellulosic components, that is pectin, hemicelluloses and lignin were removed by alkali. The increase in the intensity of the band at 879 cm 1 in regenerated cellulose indicated structural transformation from cellulose I into cellulose II. Cellulose I has a parallel closed packed structure. The vibration of b-1,4-glycosidic linkage was limited and less intensity obtained. Transformation to cellulose II caused a structure with reduced packing thus the intensity of vibration band at 879 cm 1 increased. This was observed in a regenerated film from hemp and Tencel fibre. 3.2. Fibre characteristics for the all-cellulose composites
2.5.1.2. Microstructure and surface morphology. Optical microscopy (OM) and scanning electron microscopy (SEM) were used to observe the microstructure and the surface morphology of treated and untreated cellulose fibres. A Nikon Labophot 2 polarised optical microscope, using a 50 times magnification, was connected with a computer interface and digital camera. The SEM was a Phillips XL 30 Oxford 6650 with an acceleration voltage of 142 eV. The composite specimens were coated with gold to provide a 20 nm gold layer thickness using a vacuum sputter coater. 2.5.1.3. FTIR measurements. A mixture of 5.0 mg of dried fibres dispersed in 200 mg of KBr was pressed into a disk for FTIR measurement performed using a Perkin–Elmer 2000 spectrometer. 100 scans were taken for each sample with a resolution of 2 cm 1. 2.5.2. Wide angle X-ray diffraction The fibres (70 mg) were cut and pressed into a disk using a cylindrical steel mold (diameter = 1.3 cm) with an applied pressure of about 7000 kg/cm2 using a laboratory press. The composite sample with dimensions 10 25 0.2 mm was cut from the composite sheet. Ni-filtered CuKa radiation (k = 0.1542 nm) was generated at 40 kV and 35 mA using a Bruker AXS D8. The X-ray diffractograms were recorded from 5 to 60° of 2h (Bragg angle) by a goniometer equipped with scintillation counter at a scanning speed of 0.02°/s and sampling rate of 2 data/s.
Fig. 2a–d shows ground hemp fibre. The original length of sourced fibre was 10 cm which is suitable for a woven composite fabrication. A shorter length of fibre is more applicable for a nonwoven composite. The dimension of the fibre influenced the mechanical properties of the composite [2], therefore the fibre was cut and sieved. The mesh number presented in all figures was assigned for the fibre that remained on the sieve; the proportion of various fibre lengths in ground fibres is shown in Table 1. A length of 100 lm was the most frequently obtained fraction. The three different lengths of 500, 100 and 45 lm fibres were selected as filler for all-cellulose composite. The 200 lm fibre was used to prepare cellulose solution. The fibres had different surface area and pore structure (Table 1) as a consequence of their length and the cutting process. The surface area of 100 lm fibre was lower than 500 lm and 45 lm fibres, respectively. The 500 lm fibres showed partial disintegration of a fibre bundle (Fig. 2d). The 100 lm fibres (Fig. 2b) showed an even size distribution with a clean cut of fibre and less bundle disintegration.
2.5.3. Thermogravimetry The temperature programs for thermogravimetry were from 35 to 850 °C at heating rates of 2.5–30 °C/min using a Perkin–Elmer TGA7 instrument. The measurements were conducted under nitrogen (20 mL/min), which was switched to air at 700 °C. 2.5.4. Mechanical properties The mechanical properties of all-hemp cellulose composites were determined by static tensile testing. Gauge length of the composite sheet was set at 10 mm. The tests were performed using a Rheometric DMTA IV at a speed of 0.03 mm min 1. The cross-section area of specimen was determined using a Mitutoyo digital micrometer. An average value was taken from at least ten specimens of each specimen. Tensile creep-recovery was tested using a constant stress of 0.8 MPa for 900 s followed by 2700 s of recovery with 500 Pa applied stress. The strain was measured by the displacement of the upper and lower clamps. The gap between clamps is set at 10 mm. The specimen dimension is 2 10 0.2 mm (width:length:thickness).
Fig. 1. FTIR spectrum of: (a) regenerated hemp film, (b) Tencel fibre and (c) 8% NaOH treated hemp fibres.
2122
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
Fig. 2. SEM images of ground hemp fibres of different sizes: (a) 45 lm (mesh 325), (b) 100 lm (mesh 140), (c) 200 lm (mesh 72) and (d) 500 lm (mesh 36).
Table 1 Surface area and size distribution of ground fibres. Hemp fibres
BET Specific surface area (m2/g)
Ground mixed size (45–600 lm) 600 500 200 100 45
3.3. Wide angle X-ray diffraction measurement Sieve size distribution (%)
Pore size (Å)
Volume (cm3/g)
0.787
76.8
0.00151
100
N/A 0.939 0.601 0.662 1.018
N/A 35.8 138.8 95.0 98.3
N/A 0.00084 0.00208 0.00157 0.00250
4.7 3.4 30.1 50.7 11.1
Fig. 3a–d exhibits a surface considered to be smooth, for an all-cellulose composite to which calendering was not applied. Fibres were completely covered by a regenerated cellulose matrix. A light brown translucent film was a characteristic of the matrix. Therefore further observation using an optical microscope was conducted to obtain information regarding internal composite structure. Dispersion of fibres within the composite can be seen from optical microscope images (Fig. 4a–d), some fibres are indicated by arrows. Small internal voids, indicated by circles drawn on the images, were distributed throughout the composite body. Escape of solvent (NMMO) during the regeneration step may be a reason. The pores may have occurred since the solution was required to be at high temperature during regeneration [22]. The fibres exhibited random orientation throughout the composite especially the composite containing 45 lm fibres. Interestingly the unfilled cellulose contained a small amount of very short fibres (Fig. 4a. This indicated partial dissolution of the fibre into NMMO solution. The insoluble fraction of fibre was determined using the following X-ray technique.
Fig. 5 shows the X-ray diffractograms of composites. The diffractograms exhibited a structural combination of cellulose I and cellulose II. A slight difference in the intensity of the main crystalline plane (0 0 2) at 22.7° in cellulose I can be observed. This result was attributed to variation of fibre (cellulose I structure) composition in the composite. Cellulose II structure showed a main crystalline plane at a lower Bragg angle, 20.4°. The X-ray scattering measurement, due to the main diffraction planes present at different diffraction angles can be applied quantitatively. The intensity ratio of these crystalline planes was calculated to obtain the fraction of each composition in the all-cellulose composite by Gindl and Keckes assuming a linear relationship [14]. A mixture of Ajax cellulose (pure cellulose I) and Tencel fibre (regenerated cellulose II) was prepared at various compositions to use as a reference material. Fig. 6 shows the X-ray diffractogram of the mixture of Ajax cellulose and Tencel fibre. The ratio of intensity at 22.7° and 20.4° was calculated to estimate the content of each fibre. The relationship between the intensity ratio and cellulose fraction is shown in Fig. 7 with a slight negative deviation from a linear relation. The diffractogram of a reference mixture with Tencel composition of 60% or greater was similar to that of all-cellulose composites. The fractions of regenerated cellulose were calculated and shown in Table 2. The unfilled regenerated film contained 85% of regenerated cellulose, confirmed by the presence of undissolved fibre in this material as observed by optical microscopy. The composite containing 500 lm fibres maintained the original composition of fibre at 60%. A slight reduction of fibre fraction was found in the composites containing 45 lm and 100 lm. It is proposed that this can be explained if the cellulose fibres swelled and were partially dissolved during the mixing to prepare composites. The solvent penetrated between the molecular cellulose sheets [23], although the temperature at mixing was lower than
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
2123
Fig. 3. SEM images of all-cellulose composites: (a) regenerated, (b) 40% of 45 lm, (c) 40% of 100 lm and (d) 40% of 500 lm.
Fig. 4. Optical microscopy (magnification 50) of composites with different fibre lengths: (a) no added fibres, (b) 45 lm, (c) 100 lm and (d) 500 lm.
130 °C at which cellulose started to dissolve and the 12% w/v of cellulose in NMMO solution was not the saturated concentration. Mechanical shearing was applied. Hence some degree of swelling or dissolution may occur during the mixing, especially with a low crystallinity fibre. Molecules of coagulants (water–ethanol, (50:50) solution) diffused into the cast composite sheet. The
NMMO molecules that were restricted between solvated cellulose chains interacted with diffused NNMO molecules and dissolution of NMMO into coagulant occurred. Due to the high degree of solvation of cellulose chains in the irreversible swelling process, the original cellulose I crystal structure was almost completely transformed into the cellulose II modification [24].
2124
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
Fig. 5. X-ray diffractogram of all-cellulose composite containing: (a) no added fibres, (b) 45 lm, (c) 100 lm and (d) 500 lm fibre.
100 lm and 500 lm fibres presented a maximum weight loss at 401.4, 396.6 and 404.3 °C, respectively. Pretreatment with 8% NaOH solution improved the purity of fibre by removal of non-cellulosic components. Therefore the fibre bundle was disintegrated into fine fibres that were purer and of increased structural order. After the grinding and sieving processes, a large fraction of this finer fibre was found in mesh 36 (500 lm) than mesh 325 (45 lm) and mesh 140 (100 lm), respectively. Hence the difference in thermal stability of fibre was based on its properties with respect to fibre size. The lower degradation temperature of 100 lm fibres represented a slightly lower structural order of this fibre than the 45 lm and 500 lm fibres, respectively. The all-cellulose composite showed a broader degradation temperature (Fig. 8b). This signified the combination of two different crystalline structures in the composite. Moreover the degradation temperature of the fibre component was shifted to a lower temperature. A greater reduction of maximum degradation temperature was found in 100 lm fibre. A lower structural order indicated by a low differential thermogravimetry (DTG) peak of 100 lm fibre may be because it was easier to swell by NMMO solution than the 45 lm and 500 lm fibres. The result was consistent with the reduction in cellulose I fibre composition found by X-ray measurement (Table 2). There was no reduction of 500 lm fibre fraction measured by X-ray therefore the reduction of the maximum degradation temperature found in 500 lm fibre may be attributed to the structural swelling rather than dissolution or transformation into cellulose II structure. Dissolution and/or swelling of cellulose fibre took place during the shear-mixing step. A reduction of time and temperature of mixing was a suitable means to control fibre fraction at a desired composition, however the increase in viscosity of solution must be considered. 3.5. Mechanical properties
Fig. 6. X-ray diffractogram of a mixture of Ajax (cellulose I) and Tencel (cellulose II) at different compositions.
Fig. 7. The intensity ratio of major diffraction plane of Ajax fibre and Tencel fibre.
3.4. Thermogravimetry of fibres and composites The differential thermograms of regenerated cellulose and ground fibre are shown in Fig. 8a. The regenerated cellulose from hemp exhibited a maximum weight loss at 359.3 °C. The 45 lm,
3.5.1. Tensile mechanical properties Mechanical properties of the composites are shown in Table 2. The error in the results at the 95% confidence level is provided alongside the data for the mechanical properties in Table 2. The composition of fibres in composites was fixed at 40% w/w for all composites so mechanical properties depended mainly on the physical properties of the fibres. The relatively poor performance of the composites was due to the small-scale laboratory fabricating technique that cannot create compacted structures for the composites and left considerable void contents in the composite. The mechanical properties of each matrix were improved as seen from an increase in modulus and yield stress of each composite except for a slight reduction in modulus of the composite containing 500 lm fibre. The composite containing 45 lm fibre showed the highest mechanical properties improvement. This behaviour signified that the greater surface area and pore volume of 45 lm fibre played an important role in the mechanical properties of this composite. Moreover a shorter fibre permitted better dispersion. A change of fibre composition may cause some variation to the mechanical properties. Since the reduction of fibre contents by partial dissolving of cellulose might cause interface swelling of fibre. This would enhance the interface adhesion between cellulose fibre and matrix. Therefore the high degree of fibre dissolution that found in the composite containing 45 lm and 100 lm fibre showed an increase in mechanical properties. As the composites were prepared by hot compaction and direct regeneration into water–ethanol solution without any calendaring, the imperfection of composite bulk structure from a presence of internal voids may depress mechanical properties from their optimum value. The voids in composites with different fibre content were observed in the optical microscope images. A high void content and size was observed in the 500 lm composite and the voids declined in
2125
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126 Table 2 Tensile mechanical properties and X-ray intensity (I22.7/I20.4) ratio of all-cellulose composites with different fibre length. Composites
Modulus (GPa)
Yield stress (MPa)
Elongation at break (%)
I22.7/I20.4 (Cellulose II content)
Regenerated film +45 lm composite +100 lm composite +500 lm composite
1.44 ± 0.17 1.82 ± 0.27 1.58 ± 0.37 1.33
23.2 ± 3.13 28.9 ± 3.63 24.7 ± 8.32 25.3 ± 2.81
21.2 ± 3.76 20.8 ± 4.66 26.0 ± 7.47 20.6 ± 4.34
1.04 1.44 1.32 1.83
(84%) (70%) (73%) (60%)
Table 3 Model fitted creep parameters of all-cellulose composites with different fibre lengths. Composites
Maxwell modulus (GPa)
Maxwell viscosity (GPa s)
Voigt modulus (GPa)
Voigt viscosity (GPa s)
Relaxation time (s)
Regenerated film +45 lm composite +100 lm composite +500 lm composite
1.6 32.5 9.6 10.3
2.15 0.48 0.66 0.68
1.7 65.6 6.36 4.3
41.0 303.7 111.3 130.2
23.9 4.6 17.5 30.5
Fig. 9. Creep of all-cellulose composites with an applied stress of 0.8 MPa for 900 s.
Fig. 8. TGA results of (a) different sizes ground fibres and (b) all-cellulose composites.
100 lm and 45 lm composites, respectively. The presence of large void sizes found in the 500 lm composite drastically reduced the mechanical properties. 3.5.2. Tensile creep-recovery The creep-recovery experiment was conducted to investigate the shape stability improvement by addition of fibres into cellulose. The four-element model of Maxwell and Kelvin-Voigt [25] was used to describe both creep and recovery of this composite. The model parameters were computed and presented in Table 3. Inclusion of fibre caused a reduction of the simultaneous creep strain and/or recovery strain (Fig. 9). The simultaneous recovery strain was used to calculate the modulus (spring constant) of the first element using an applied stress of 0.8 MPa. The modulus of the spring element was greatly improved, indicating the ability of fibres to retain their original shape under a rapidly applied tensile stress. The highest modulus was found in the 45 lm fibre composite which demonstrated excellent interfacial adhesion. The partial dissolution and inter-planar swelling of fibre may enhance
the compatibility between fibre and matrix. When fibre was not included there was more creep (3 times) under constant load than was measured for the composites. Nevertheless the unrecovered strain of unfilled cellulose was less than composites resulting in a high viscosity of the dashpot element. The creep zone was used to calculate the viscosity of KelvinVoigt elements. The slope of this creep section indicated a flow of composite structure under tensile load. All fibre-filled composites showed a similar slope but lower than unfilled cellulose. This signified retardation of the matrix flow by fibre under load. The creep modulus (spring constant in Voigt element) was calculated by using the recovery strain and the applied stress of 0.8 MPa. The recovered strain calculation was started 1–2 s after the applied tensile stress was released, as indicated by a deviation from a linear vertical line of the recovery curve. Recovery was observed to be complete when the measured strain reached a plateau. The significant high modulus in the 45 lm fibre composite resulted from a small recovery strain compared with regenerated cellulose matrix (Fig. 10a). Since the retardation time is g/E, in order to obtain the viscosity of the dashpot element the retardation or relaxation time was required. The retardation or relaxation time is the time for the Voigt element to deform to 63.21% of its total deformation or recover to 36.79% of its original value. In this case the relaxation from recovered strain was preferred since the permanent (unrecovered) strain
2126
S. Ouajai, R.A. Shanks / Composites Science and Technology 69 (2009) 2119–2126
lose–NMMO–water solution. The dissolution of fibre transformed cellulose I morphology to cellulose II, as confirmed by X-ray scattering. An improvement of mechanical properties of regenerated cellulose films was obtained by addition of fibres, and the extent of improvement depended on the crystallinity and surface area of the fibres. Mixing caused partial dissolution and swelling of the fibre structure, especially lower crystallinity fibres. TGA and X-ray scattering techniques were particularly useful for indicating the structural change in composites and to calculate the composition of cellulose with different crystalline structure quantitatively. Acknowledgements The authors gratefully thank King Mongkut Institute of Technology, North Bangkok (KMITNB), Thailand for a PhD scholarship and Lu Guang Chen for an optical microscope measurement. References
Fig. 10. Recovery of (a) 45 lm fibre composite and (b) regenerated hemp cellulose after removing an applied stress of 0.8 MPa; (where s is experimental and — indicates fitted line).
could be eliminated. The selected recovery zone of the regenerated cellulose and composite was fitted to calculate the relaxation time of the composite (Fig. 10a and b). The modulus and viscosity of Voigt element of composites was increased (Table 3) signifying that inclusion of fibre improved the composite properties of especially that containing 45 lm fibres. 4. Conclusions All-cellulose composites were prepared by solution-slurry blending between ground hemp (C. Sativa L.) fibres and cellu-
[1] Mohanty AK, Misra M, Drzal LT. J Polym Environ 2002;10:19–26. [2] Shibata M, Takachiyo K-I, Ozawa K, Yosomiya R, Takeishi H. J Appl Polym Sci 2002;85:129–38. [3] Dubief D, Samain E, Dufresne A. Macromolecules 1999;32:5765–71. [4] Zhang Ming Qiu, Rong Min Zhi, Lu Xun. Compos Sci Technol 2005;65:2514–25. [5] Zhang MQ, Rong MZ, Lu X. All-green composites. In: Fakirvor S, Bhattacharyya D, editors. Handbook of engineering biopolymers, homopolymers, blends, and composites. Hanser Publishers; 2007. p. 747–72. [6] Lu Xun, Zhang Ming Qiu, Rong Min Zhi, Shi Guang, Yang Gui Cheng. Compos Sci Technol 2003;63:177–86. [7] Lu Xun, Zhang Ming Qiu, Rong Min Zhi, Yue Da Lei, Yang Gui Cheng. Compos Sci Technol 2004;64:1301–10. [8] Lu Xun, Zhang Ming Qiu, Rong Min Zhi, Yang Gui Cheng. Polym Adv Technol 2003;14:676–85. [9] Flores A. J Macromol Sci Part B: Phys 2001;40:749–61. [10] Houshyar S, Shanks RA, Hodzic A. Polym Test 2005;24:257–64. [11] Lacroix FV, Loos J, Schulte K. Polymer 1999;40:843–7. [12] Woodings C. Regenerated cellulose fibres. Cambridge: Wood Head Publishing Ltd.; 2001. [13] Nishino T, Matsuda I, Hirao K. Macromolecules 2004;37:7683–7. [14] Gindl W, Keckes J. Polymer 2005;46:10221–5. [15] Stockinger H, Kut OM, Heinzle E. Water Res 1996;30:1745–8. [16] Meister G, Wechsler M. Biodegradation 1998;9:91–102. [17] Kreze T, Strnad S, Stana-Kleinschek K, Ribitsch V. Mater Res Innov 2001;4:107–14. [18] Fink H-P, Weigel P, Ganster J, Rihm R, Puls J, Sixta H, et al. Cellulose 2004;11:85–98. [19] Dufresne A, Vignon MR. Macromolecules 1998;31:2693–6. [20] Curvelo AAS, de Carvalho AJF, Agnelli JAM. Carbohyd Polym 2001;45:183–8. [21] Chanzy H, Peguy A, Chaunis S, Monzie P. J Polym Sci Polym Phys Ed 1980;18:1137–44. [22] Kim DB, Lee WS, Jo SM, Lee YM, Kim BC. J Appl Polym Sci 2002;83:981–9. [23] Michael M, Ibbett RN, Howarth OW. Cellulose 2000;7:21–33. [24] Chanzy H, Noe P, Pailet M, Smith P. J Polym Sci Appl Polym Sym 1983;37:239–59. [25] Menard KP. Dynamic mechanical analysis a practical introduction. Florida: CRC Press LLC; 1999.