Flexural properties of sawdust-reinforced epoxy composites post-cured in microwaves

Introduction

‘It has been established in recent years that polymer reinforced with a small percentages of strong fillers can significantly improve the mechanical and thermal properties.’ ¹ This research project is to investigate the yield strength, tensile strength and Young’s modulus of epoxy composites reinforced with varying percentage by weight of sawdust, the filler, with a view to finding out the optimum percentage by weight of the sawdust that can be added to the composites. The sawdust particles were sieved into three different sizes, which were <425 µm, 425–600 µm and 600–1180 µm, respectively. The epoxy resin used in this study is Kinetix R246TX Thixotropic Laminating Resin, an opaque liquid, and the hardener used is kinetic H160 medium hardener which has a pot life of 120 min. Other hardeners like H126, H128, H161 and H162 can also be used. ² The resin has been supplied by ATL Composites, Australia. The sawdust was first mixed with epoxy resin, after this, the hardener, kinetic H160 medium was added. The by weight ratio of resin to hardener used was 4:1. ² The composite was then cast to moulds of tensile test pieces and left to cure under ambient conditions for 24 h. The tensile test specimens were taken out of the moulds and then post-cured in oven at 40°C for 16 h, and then at 50°C for 16 h and finally at 60°C for 8 h. This is to ensure the heat distortion temperature is above 63°C. To bring the ultimate heat distortion temperature to 68°C, another 15 h of post-curing will be required. ² The specimens were then subjected to tensile tests.

Experimental

Three-point bending test

The three-point bending flexural test provides values for the modulus of elasticity in bending, E_B, flexural stress, σ _f , flexural strain, ε _f , and the flexural stress–strain response of the material. The main advantage of a three-point flexural test is the ease of the specimen preparation and testing. However, this method has also some disadvantages: the results of the testing method are sensitive to specimen and loading geometry and strain rate. ¹²

The standard used is ISO 14125:1998(E) because the results can then be compared with the work of others. ¹³ The centre uses a universal machine MTS Alliance RT/10 at 10 kN couple with the software TESTWORK 4. The dimensions of the specimens of resins were 250 mm × 10 mm × 4 mm and tested at a crosshead speed of 4 mm/min.

The equation used for calculating the flexural stress is

σ f = 3 PL 2 bh 2

(1)

The equation used for calculating the flexural strain is

ɛ f = 6 Dh L 2

(2)

The equation used for calculating the Young’s modulus is

E B = L 3 m 4 bh 3 = Flexural strength Flexural strain

(3)

where: σ _f : stress in outer fibre at midpoint, MPa;

ε _f : strain in the outer surface, %;

E_B: modulus of elasticity in bending, MPa;

P: load at a given point on the load deflection curve, N;

L: support span, mm;

b: width of test beam, mm;

h: depth of test beam, mm;

D: maximum deflection of the centre of the beam, mm;

m: slope of the tangent to the initial straight line portion of the load deflection curve, N/mm.

The slope, modulus of elasticity, of a sample is illustrated in Figure 1. OPEN IN VIEWER

Figure 1.

Deflection curve for determining the modulus of elasticity of a sample.

Composite preparation

The reinforcement was sawdust and they were made 0–15% by weight in the cured epoxy composite, EP/sawdust (X%), where X is the percentage by weight of the filler. Above 15% by weight of filler, the slurry would be too sticky to be cast into moulds. As the raw materials of the composites are liquid and sawdust, the flexural test specimens were cast to shape. The resin is an opaque liquid and is first mixed with the catalyst. After that the sawdust was added to the mixture, they were then mixed to give the uncured composite. Table 1 shows the mass in grams of resin, catalyst and glass powder required, respectively, to make 1000 g of uncured composite of 10 wt % of sawdust.

OPEN IN VIEWER

Table 1.

Weight of materials required to make 1000 g of EP/sawdust (10%)

	Materials	Resin (R)	Hardener (H)	R + H	Sawdust	Composite
Parameters
Ratio by weight		4	1	–	–	–
Percentage by weight		–	–	90	10	–
Weight of materials in 1000 g of EP/GP (20%)		720 (g)	180 (g)	900 (g)	100 (g)	1000 (g)

The mixture of sawdust, resin and hardener was blended with mechanical blender to ensure a more homogenous mixture. The uncured composite was then cast into the moulds and cured in ambient conditions. After initial 24-h curing, when the test pieces were removed from the mould, they were post-cured. This was achieved by curing the pieces in an oven or in microwaves. Oven temperatures and times were:

16 h at 40°C

16 h at 50°C

8 h at 60°C

The test pieces were then tensile-tested.

The post-curing in microwaves was achieved by placing the samples in the cavity of a modified Sanyo microwave oven. The maximum power of the microwave oven was 800 W. There are 10 settings in the power level. The power can be increased in steps of 80 W. The microwave post-curing was achieved in three stages to ensure that the sample had achieved the specified temperature. After each step of the microwave post-curing, the temperature of the sample was measured with an infrared thermocouple. If the sample was not at the required temperature of the stage, the sample was placed back in the microwave oven until the correct temperature was achieved. Upon achieving the required stage temperature, samples were then allowed to cool to room temperature inside the oven cavity before the next stage began. The power used for the process was 160 W and details are shown in Table 2. It can be found that the desired temperature at each stage is the same to that in conventional post-curing.

OPEN IN VIEWER

Table 2.

Times and temperatures for curing in microwaves

Microwave power: 160 W
Time (minutes)	Temperature (°C)
6	40
8	50
10	60

Results and discussion

The flexural strengths of samples containing 15 wt % of sawdust with varying percentages of weight of palm oil post-cured in microwaves are shown in Figure 3. The neat epoxy samples exhibit the highest peak flexural strengths. The samples with <425 µm sawdust have higher peak flexural strengths than those of the 425–600 µm and 600–1180 µm sawdust. The samples with 425–600 µm sawdust have a marginally higher flexural strength than those with 600–1180 µm. It is fair to say that the flexural strengths decreased with increasing sizes of sawdust. Referring to Figure 3 and samples without palm oil, it can be found that the neat epoxy sample had a flexural strength of 99.00 MPa. The sample with <425 µm sawdust had a flexural strength of 57.07 MPa which was 42.4% lower than that of the neat epoxy sample. The sample with 425–600 µm sawdust had a flexural strength of 38.25 MPa, 49.2% lower than that of <425 µm sawdust sample. Finally, the sample with 600–1180 µm sawdust had a flexural strength of 39.27 MPa, 3% higher than that of 425–600 µm sawdust sample. The flexural strengths of samples with 5 wt % sawdust, 10 wt % sawdust and 20 wt % sawdust exhibited a similar pattern to those of 15 wt % sawdust. OPEN IN VIEWER

Mosiewicki et al. claimed that increasing the composite fibre weight fraction would increase the voids in the composites, which would affect the physical and mechanical properties of the composite. ¹⁶ Thus, the greater the amount and size of the sawdust in the samples, the larger would the voids in the composites and hence the lower the physical and mechanical properties of the composites. This was found to be correct in this study.

Gruenwald stated that smaller particle sizes are generally more beneficial in improving the mechanical properties of composites. The above results clearly exhibited this pattern; the specimens with smaller particle sizes had the highest peak flexural strengths, while specimens with the largest particle sizes had the lowest flexural strengths. ⁸

O’Donnell et al. discovered that natural fibres exhibited many advantageous properties; they were a low-density material yielding relatively lightweight composites with high specific properties. ¹⁷ In this study, the natural fibre was sawdust and by having 5 wt % of palm oil and sawdust, the flexural strength of the composite was lower than its neat resin counterpart.

Nakamura et al. also found that the flexural strengths of silica-filled epoxy composites decreased steadily with increasing filler particle size. The results were in-line with those of this study. ¹⁸

Figure 3 also illustrates the flexural strengths of varying wt % of palm oil and sawdust reinforced epoxy matrix post-cured in microwaves. The strengths of the neat epoxy samples decreased with increasing amount of palm oil. It can be found that the flexural strengths of the neat epoxy samples were higher than those of composites with any wt % of sawdust. The neat epoxy samples exhibited the plasticizing effect of the palm oil. The flexural strengths of the samples with sawdust increased marginally with 5 wt % of palm oil and then decreased with 10 wt % palm oil. It can be found that the flexural strengths of the samples with <425 µm of sawdust and 0 wt % of palm oil was 57.07 MPa; the flexural strength increased by 6.8% to 60.96 MPa with samples with 5 wt % of palm oil; the flexural strength decreased by 12.65% to 53.25 MPa with samples with 10 wt % palm oil. With more wt % of palm oil, the plasticizing effect was too much, making the samples softer. Table 3 shows the standard deviations of flexural strengths of varying wt % of palm oil and 15 wt % sawdust reinforced epoxy matrix post-cured in microwaves. It can be observed that except for one or two data, the standard deviations were less than 5% of the values of the flexural strength and it can be argued that the results are accurate and reliable.

OPEN IN VIEWER

Table 3.

Standard deviations of flexural strengths and maximum flexural strain of varying wt % of palm oil and 15 wt % sawdust reinforced epoxy matrix post-cured in microwaves

Wt % of palm oil composites	Flexural strength			Maximum flexural strain			Flexural modulus
	0	5	10	0	5	10	0	5	10
Neat epoxy resin	3.46 (99.00) a	1.86 (92.92)	0.77 (78.58)	1.76 (4.89)b	0.53 (5.04)	0.36 (5.00)	0.11 (2.58) c	0.03 (2.22)	0.06 (1.50)
15 wt % of <425 µm sawdust	3.05 (57.07) a	4.87 (60.96)	1.10 (53.25)	0.16 (2.42)b	0.10 (3.03)	0.15 (2.82)	0.04 (2.27) c	0.08 (2.12)	0.10 (1.64)
15 wt % of 425–600 µm sawdust	4.70 (35.25) a	3.18 (44.41)	4.00 (48.99)	0.26 (1.59)b	0.20 (2.55)	0.05 (2.34)	0.02 (2.45) c	0.16 (2.22)	0.06 (1.82)
15 wt % of 600–1180 µm sawdust	4.69 (39.37) a	5.94 (48.80)	1.40 (46.68)	0.18 1.61)b	0.37 (2.28)	0.04 (2.36)	0.10 (2.18) c	0.03 (2.00)	0.13 (2.00)

When the composites were not filled with any palm oil, the amount of sawdust in the sample did not considerably affect the flexural strengths of the samples; the flexural strengths appeared to stay relatively stable with increasing amounts of sawdust. When the composites were filled with palm oil, the flexural strengths appeared to decrease marginally with increasing amounts of sawdust.

Figure 4 illustrates the flexural strength of different mica flakes-filled epoxy resin (Epikote 828) composites of Inubushi et al. The epoxy resins exhibited 30–40% reduction in flexural strength when a small fraction of mica was added. It can be found that the trend of the curves were very similar to that of this study. ¹⁹ However, the particle sizes had even less effect to the flexural strengths of the composites than those in this study because of different fillers used. OPEN IN VIEWER

Marcovich et al. made composites from sawdust and polyester and found that ‘all the composites present a smaller flexural strength than that of the unloaded matrix. Flexural strength data for all samples are rather dispersed, and there is no significant difference between them. However, there is a slight increment with filler content for all the composites tested.’ ²⁰ The neat resin was strongest and is similar to the case in this study. In this study, there was also no significance difference between the flexural data for all samples, but samples with smaller particle sizes had higher flexural strengths than their counterparts.

The maximum flexural strains of samples containing 5 wt % sawdust post-cured in microwaves are depicted in Figure 5. The flexural strains of samples with 5 wt % sawdust exhibited a similar pattern to those of 10 wt % sawdust, and the flexural strains of samples with 15 wt % sawdust showed a similar pattern to that of 20 wt % sawdust. From Figure 5, it can be found that the neat epoxy samples had the greatest maximum flexural strains. The maximum flexural strains decreased with increasing sizes of sawdust. From Figure 5, it can be observed that the maximum flexural strains of the <425 µm, 425–600 µm and 600–1180 µm samples were clearly distinguishable. By analysing the results of samples with 5 wt % sawdust, it can be found that the maximum flexural strain of the neat epoxy sample with 0 wt % palm oil started at 4.89 GPa, and decreased 50.51% to 2.42 GPa with <425 µm sawdust, further decreased 34.30% to 1.59 GPa with 425–600 µm sawdust, and finally increased 1.24% to 1.61 GPa with 1180 µm sawdust. Table 3 also shows the standard deviations of the maximum flexural strains of samples containing 5 wt % sawdust post-cured in microwaves. As only a few of the standard deviations were over 5% of the values of the maximum flexural strains, it can be argued that the results are accurate and reliable. OPEN IN VIEWER

A project conducted by Ku et al. found that the only drawback for using finer particles as filler was their tendency to agglomerate. ²¹ Fine sawdust particles were difficult to disperse, and they agglomerated and behaved as large single particles. The research undertaken in this project confirmed the finding of the research undertaken by Ku et al., the <425 µm sawdust particles of samples with higher particulate ratio acted similar to that of the 425–600 µm and 600–1180 µm particles. Therefore, it can be argued that the <425 µm particles agglomerated and behaved as large single particles. This agglomeration of particles started to occur when the epoxy composites was reinforced with 15 wt % of sawdust. However, the effect of agglomeration in flexural strengths of samples with <425 µm sawdust particles occurred only at the 20 wt % filler. It can be argued that the effects of agglomeration in flexural strengths have occurred at 15 wt % of sawdust and that more effects occurred with increasing wt % of filler.

The amounts of palm oil in the samples affecting the maximum flexural strains of the samples are depicted in Figure 5. The maximum flexural strains in the samples with sawdust increased marginally with increasing amounts of palm oil; this was a clear example of the plasticizing affects of palm oil. When a plasticizer was added to an epoxy sample, the product was softened, which in turn increased in flexibility.

The flexural modulus of samples containing 20 wt % sawdust post-cured in microwaves is depicted in Figure 6. The flexural modulus of samples with 5 wt % of sawdust, 10 wt % of filler and 15 wt % of reinforcement exhibited a similar pattern to that of 20 wt % of sawdust. From Figure 6, it can be observed that all samples shared a similar flexural modulus. It can be argued that the size of the sawdust particles have minimal effects on the flexural modulus; the neat epoxy sample had a similar flexural modulus to that of other samples, so the size and wt % of sawdust had a minimal effect on the flexural modulus of the samples. The flexural modulus in the composites decreased linearly with increasing amounts of plasticizer. This is an example of the plasticizing affect of palm oil; the resistance of the sample to bend should decrease with increasing amounts of palm oil. The highest flexural modulus was found for neat epoxy resin sample with a flexural modulus of 2.58 GPa. The sample with the lowest flexural modulus was the one with a 1.50 GPa. Without palm oil, the plasticizer, the behaviour of the graphs would become opposite as found by Marcovich et al., Inubushi et al. and Mosiewicki et al.^16,19,20 OPEN IN VIEWER

Table 4 shows the average flexural strength, average maximum flexural strain and average flexural modulus of samples post-cured conventionally and in microwaves. It can be found from Table 4 that the average flexural strength was on average 8.91% higher when post-cured conventionally; the average maximum flexural strain was similar with no noticeable variance, and the average flexural modulus was on average 7.74% higher when post-cured conventionally. The conventional oven allowed the samples to stay at an elevated temperature for an extended period of time, while the samples post-cured in microwaves achieved the elevated temperature but could not maintain it for an extended period of time. The extra period of time at an elevated temperature allowed more cross-linking to occur, therefore, further strengthening the samples. Thus, conventional curing was more effective unless some changes were made in the post-curing of the samples by microwave irradiation like maintaining the targeted temperatures for a specified period of time. Alternatively, microwave curing at various conditions, including one-, two- and three-step heating cycle made by Tanrattanakul and Jaroendee has to be carried out, in which each cycle employed different power level and time of exposure. They found that the microwave-cured composites produced mechanical properties, including the flexural strength, as good as the thermally cured composites. ²² This is attributed to the slow increase in temperature during the beginning of the microwave curing process, whereby the very low power level was applied in the first cycle of the multistep heating process. This affected the slower rate of viscosity increment, resulting in better wettability of the glass fibre with enhanced interfacial adhesion between the fibres and the resins. The viscosity of resins affected the homogeneity of the cross-linked structure. ²²

OPEN IN VIEWER

Table 4.

Average flexural strength, average maximum flexural strain and average flexural modulus of samples post-cured conventionally and in microwaves

Post-curing method	Average flexuralstrength (MPa)	Average maximum flexural strain (%)	Average flexural modulus (GPa)
Conventional	52.2	2.4	2.14
Microwaves	47.5	2.4	2.32

Nightingale and Ray chose two epoxy systems and the selection was based on their dielectric loss factors. ‘System 1 contains resin and hardener with higher loss factor than system 2.’ The systems were cured by three methods: autoclave curing, autoclave plus microwave post-curing and microwave curing. They found that ‘autoclave cured composites had the best flexural properties followed by the microwave-cures samples. Composites that were partially autoclave cured and then post-cured in microwaves had the worst flexural properties.’ The mechanical properties of fully microwave-cured specimens were better than their rivals cured in autoclave and then post-cured in microwaves despite their larger void content. This implied that composites fully cured by microwaves would have comparable mechanical properties but shorter curing duration than their counterparts, if the specimens were cured under correct conditions. ²³ In this study, time saving was achieved and it can be argued that by adopting correct curing conditions, composites with comparable mechanical properties can be acquired.

The average T_gs of the composites post-cured conventionally and in microwaves plasticized with varying wt % palm oil is illustrated in Figure 7. From the data collected from the DMA tests, it can be claimed that the amount and size of sawdust particles and palm oil did not affect the T_gs. The T_gs value should decrease with increasing amounts of palm oil; however, no significant change was recorded. This means that the strength of the epoxy cross-linking is not weakened with increasing amounts of palm oil. It can be observed that the T_gs of samples post-cured conventionally were higher than those post-cured in microwaves. The graph also illustrated no significant change in results with increasing amounts of palm oil. Table 5 shows the average T_gs of samples. The relevant standard deviation is also incorporated in the table to show the reliability of the results. Since the standard deviations of the various samples were low, confirming that the T_g of the samples did not vary when various amounts and sizes of sawdust particles and palm oil were added to them. It can be found from Table 5 that the T_g was on average 13.62% higher when post-cured thermally. This data once again showed the effect of cross-linking between conventional and microwave post-cured samples. By studying Tables 4 and 5, it can be observed that the conventionally post-cured samples had higher average glass transition temperature, and hence stiffer; this was reflected by the fact that they also had higher average flexural strengths and modulus. ‘However, the differences in T_g may not always relate to mechanical properties’ as pointed by Tanrattanakul and Jaroendee in their study with carbon fibre and polyimide composites in which the conventionally cured samples had lower T_g but higher flexural strengths than their microwave-cured counterparts. They added by saying that the interfacial adhesion between the matrix and fibre also played a vital role on the flexural properties. ²² OPEN IN VIEWER

OPEN IN VIEWER

Table 5.

Average glass transition temperature of plasticized epoxy resins

Post-curing	Average glass transition temperatures (°C)	Standard deviation
Conventional	84.71	1.46
Microwaves	73.17	1.27

The microscopic analysis was performed on different samples to determine the reasons for porosity formation, as well as the amount and size of air bubbles formed. This key characteristic has a great impact on the flexural properties of the composites.

Figure 8 shows the scanning electron microscope image of neat epoxy resin post-cured for a total of 40 h at 40°C, 50°C and 60°C, respectively at a magnification of 200 times. Faint striations followed by a ‘turbulent flow’ can be found in the fractured surface of the neat resin. This shows that plastic deformation had taken place in the resin. Figure 9 shows a sample with <425 µm, 5 wt % of sawdust and 0 wt % of palm oil; the sample exhibited a minimal amount of air bubbles and also indicated dirt in the sample, which had darkened the sample. Figure 10 illustrates a sample with 425–600 µm, 20 wt % of sawdust and 0 wt % of palm oil; the sample exhibited a vast quantity of large air bubbles, the largest bubble found using the microscope had a circumference of 209 µm. Big holes were spotted and this explained why the flexural strength of neat epoxy resin was stronger than that of epoxy composite with 20 wt % of sawdust. The sample also exhibited a reduced contamination by dirt. Figure 11 shows a sample with 600–1180 µm, 20 wt % of sawdust and 0 wt % of palm oil; the sample exhibited a vast quantity of large air bubbles, the largest bubble found using the microscope had a circumference of 402 µm. This suggests that the size and quantity of air bubbles increased with the size of the sawdust; the larger the sawdust, the larger the air bubbles and the amount of air bubbles would be. It can be found from Figures 9–11 that there were no voids around the sawdust particles; it can be argued that the adhesion between sawdust and epoxy resin was very good. OPEN IN VIEWER

OPEN IN VIEWER

Figure 11.

Optical microscope image of a sample with 600–1180 µm, 20 wt % of sawdust and 0 wt % of palm oil, ×60.

OPEN IN VIEWER

Figure 8.

Scanning electron microscope (SEM) image of fractured neat epoxy resin, ×200.

OPEN IN VIEWER

Figure 9.

Optical microscope image of a sample with 425 µm, 5 wt % of sawdust and 0 wt % of palm oil, ×60.

Conclusions

This study has evaluated the flexural strength, maximum flexural strain and flexural modulus of varying percentage by weight of sawdust, of different range of particle sizes and of varying percentage by weight of palm oil reinforced epoxy resin; in all cases, the fluidity of the slurry composite was high and could be cast into moulds. Some air bubbles were found due to imperfect manufacturing of the samples. The neat epoxy sample exhibited the highest flexural strength. The flexural strengths of samples post-cured in microwaves exhibited similar relationships to those post-cured conventionally. The flexural strengths:

Decreased with increasing size of sawdust;

When the composites were not reinforced with any palm oil, the amount of sawdust in the composites did not affect the flexural strengths of the sample considerably; the strengths stayed relatively stable with increasing amounts of sawdust. When the composites were plasticized with 5 wt % and 10 wt % of palm oil, the strengths decreased slightly with increasing amounts of sawdust. The strength of the neat epoxy samples decreased with increasing amounts of palm oil. The palm oil acted as a plasticizing agent and increased flexibility of the samples, which in turn reduced the flexural strengths. The strengths in the samples with sawdust increased marginally with 5 wt % palm oil and then decreased again with 10 wt % of plasticizer.

The flexural strains of samples post-cured in microwaves exhibited similar relationships to those post-cured conventionally. The neat epoxy sample exhibited the highest flexural strain. The flexural strains:

Decreased with increasing size of sawdust;

The flexural modulus of samples post-cured in microwaves exhibited similar relationships to those post-cured conventionally. The neat epoxy sample had a similar flexural modulus to other samples containing different weights and sizes of sawdust, thus, it can be argued that the size and weight of sawdust had minimal effect on the flexural modulus of the samples. The flexural modulus in the samples with sawdust decreased linearly with increasing amounts of palm oil. This is an example of the plasticizing affect of palm oil; the resistance of the sample to flexural bending decreased with increasing amounts of palm oil. The flexural modulus:

Decreased with increasing amount of palm oil and

It can also be argued that the interfacial adhesion between epoxy resin (matrix) and sawdust (reinforcement) would be improved by treating the sawdust, and the properties of the composites would then be improved. The trend of the results of this study was also supported by the results of other similar study using sawdust as filler.