Introduction

Starch is employed in numerous industrial applications on making food and nonfood products, owing to its low cost, availability and ability to impart a broad range of functional properties. Starch has become a most promising candidate for future materials and alternatives to the fossil resources that have recycling limitations and cause environmental concerns, because of its attractive cost and performance [1, 2]. Starch is composed of both amylose molecules and amylopectin molecules with α-(1–4)-linked d-glucose backbones and α-(1–6)-linked branches. Native starch from plants is granular and partly crystalline and is not suitable for thermoplastic materials. When a plasticizer (typically water) is added, gelatinization will occur under heat and high shear, and the native granular starch is disrupted and converted to thermoplastic starch (TPS) that can be processed similarly to other more conventional thermoplastic polymers [3,4,5,6,7]. Various sources of starch have been extensively investigated to obtain biodegradable thermoplastic starch, such as potato starch [3, 8], corn starch [8], rice starch [9] and wheat starch [10]. Cassava or tapioca is among the most important starch sources in the tropics and is extensively used in tropical Africa and South America as well as parts of Asia, especially Indonesia and Thailand [11]. Thailand was one of the top five producers globally in 2012 [12]. In addition, cassava flour is substantially rich in starch with only small amounts of other components and is a good source of inexpensive good purity starch [13]. Thermoplastic cassava starch has been prepared for bioactive, biodegradable, flexible and edible films or packages [8, 14,15,16,17,18,19].

However, TPS has as its disadvantages: brittleness, poor mechanical properties, high water sensitivity and poor barrier properties [20]. To improve the properties of thermoplastic starch, it is blended with other polymers, such as natural rubber [21], polystyrene [22], polyvinyl alcohol [23,24,25], polyethylene [26], polypropylene [27], polylactic acid [28,29,30], polycaprolactone [31], chitosan [32, 33], poly(butylene adipate-co-terephthalate) [34], or polyhydroxyalkanoates [35], and reinforced with organic or mineral fillers, such as cellulose [3, 36,37,38,39,40,41], talc [42], or clay [43, 44], ZnO [45]. Titanium dioxide (TiO2) is a multipurpose filler used with polymer matrices, and is nontoxic, inert and inexpensive with inherent photocatalytic activity against microorganisms, staining, allergens and odors [46]. Several studies have assessed the influences of TiO2 addition on the properties of various biopolymers. For example, Zhu et al. [47] prepared poly(lactic acid)/TiO2 composite films. It was found that the addition of TiO2 nanoparticles increased tensile strength and crystallinity, as well as improved the barrier properties. Amin and Panhuis [48] found that the mechanical properties and water resistance were improved by addition of TiO2 in chitosan matrix. Khan et al. [49] prepared regenerated bacterial cellulose/TiO2 nanocomposites to enhance bactericidal activity and tissue regeneration properties. Oleyaei et al. [46] studied the effects of TiO2 concentration on functional, UV transmittance and mechanical properties of potato starch/TiO2 nanocomposite films, with a view to food packaging applications. Ostafińska et al. (2017) prepared thermoplastic starch/TiO2 composites with highly homogenous good dispersion of the filler by a two-step method, including solution casting followed by melt mixing. The modulus increased with TiO2 loading [50]. Aluminum oxide (Al2O3) or alumina is commonly used as filler in commercial applications due to its low cost, electrical insulation ability, and high temperature tolerance [51]. The incorporation of Al2O3 has been reported for polymer matrices including natural rubber [52], silicon rubber [53], poly(lactic acid) [54], epoxy [55], and chitosan [56, 57].

Individually TiO2 or Al2O3 fillers improve the properties of polymers. Combinations of TiO2 and Al2O3 have been used to enhance the properties of thermoplastic starch composites. The aim of this work was to study the effects of Al2O3 filler particle size on dielectric, mechanical, morphology and thermal properties of thermoplastic starch–TiO2–Al2O3 composites.

Experimental

Materials

Food grade native cassava starch was provided by Kriangkrai Co., Ltd., Thailand. It was first heated for 24 h in a vacuum oven at 60–70 °C to remove moisture and was immediately placed in a desiccator to prevent moisture re-absorption. The glycerol used as a plasticizer had 1.26 g/cm3 density and was manufactured by Unilever Co., Ltd., USA. Titanium dioxide (TiO2) nanopowder (< 100 nm, mixture of rutile and anatase TiO2) and alumina, Al2O3 (0.05 μm, 1 μm and 5 μm) were purchased from Sigma-Aldrich Co.

Preparation of TPS composites

Native cassava starch was weighed and introduced in a kitchen blender. Glycerol was then added and continuously pre-mixed at room temperature and 1000 rpm for 10 min to obtain a homogeneous dispersion. The weight ratio of glycerol and cassava starch was 30:70. The mixture was placed in a desiccator for 24 h, after which it was further dry-mixed with TiO2 and Al2O3 particles. The weight ratio TPS:TiO2:Al2O3 was held fixed at 97:2:1. In order to obtain thermoplastic starch (TPS), the mixture was continuously melt blended in an internal mixer at 140 °C with 60 rpm rotor speed for 6 min. During mixing, the torque was monitored, in order to assess plasticization and processability of the starch. The starch lost its original granular structure and was transformed into a molten homogeneous TPS. The TPS composite was sheeted out on a two-roll mill, and 1-mm-thick sheets were prepared by 1500 psi compression at 165 °C for 15 min.

Testing mechanical properties

Tensile testing was performed to characterize the mechanical properties of the thermoplastic starch composites, using a Hounsfield Tensometer, model H 10 KS. The sample sheet was cut by a cutting machine into the dumbbell shape specified in ASTM D680-14 standard [58]. Specimens were tested for each formula at 25 °C with a cross-head speed of 500 mm/min. The stress–strain curves are reported. The tensile strength (σb) and elongation at break (εb) were determined from the curves. The Young’s modulus (E) was determined as the initial slope of the linear portion of each stress–strain curve. The Shore A hardness was measured according to ASTM D2240-15 with an indentation durometer [59].

Sample characterization

Thermogravimetric analysis was performed on a TA Instruments® TGA Q 500 with a heating rate of 10 °C/min from room temperature to 600 °C in nitrogen atmosphere. Dynamic mechanical thermal analysis used a Rheometric Scientific® DMTA V. The experiment was carried out in the dual-cantilever bending mode at a frequency of 1 Hz with 0.01% strain amplitude, and the heating rate was 3 °C/min. The temperature range was from − 120 to 80 °C. Dielectric constant (ε′) and dissipation factor or dielectric loss tangent (tan δ) were obtained by dielectric measurements. The sample with ~ 1 mm thickness was coated with high-purity silver paint before measurement with Precision LCR meter (Agilent, model 4285A) at room temperature in the frequency range from 75 kHz to 30 MHz. The evaluation of the dispersion of fillers in TPS was performed by using a scanning electron microscope (SEM) (FEI Quanta, 400, USA). Fresh cross sections were obtained by fracturing after immersion in liquid nitrogen. All specimens were sputter coated with gold prior to SEM examination.

Results and discussion

Dielectric constant and dissipation factor

The dielectric constant and dissipation factor over the frequency range from 75 kHz to 30 MHz are shown in Fig. 1, for the thermoplastic starch–TiO2–Al2O3 composites with Al2O3 particle sizes 0.05 μm, 1 μm, and 5 μm. The dielectric constant rapidly decreased at low frequency and then gradually decreased at frequencies above 2 MHz. This behavior was almost similar in all cases tested. At low frequencies, maximal polarization was possible [60], while the dissipation factor rapidly decreased with frequency in the low range (0–10 MHz), because both ionic relaxation and dipole relaxation are comparatively slow. The further increase in dissipation factor at the higher frequencies (> 10 MHz) might be due to relaxation of polar groups and ionic polymerization. The polar molecules cannot orient rapidly enough as the electromagnetic frequency is increased [61]. The dielectric constant decreased with Al2O3 particle size. The larger particles had lower specific surface and therefore less interfacial polarization. Thus, the smaller Al2O3 particles had comparatively improved dielectric properties in the thermoplastic starch–TiO2–Al2O3 composites, with also stronger mechanical stress effects and extrinsic grain boundary effects [62]. In addition, the polarization contribution by dipole–dipole interactions increased with closely packed particles, and the smaller particles had better particle–particle contacts and increased packing density [63].

Fig. 1
figure 1

The frequency dependence of a dielectric constant and b dissipation factor for the thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes

Full size image

Thermal analysis

Figure 2 exhibits the TGA and DTA of thermoplastic starch–TiO2–Al2O3 composites, respectively. All the samples showed two main stages in the loss of mass. Within the temperature range from 30 to 280 °C, the loss corresponded to the volatilization of both water and glycerol from the TPS. The second stage corresponded to starch decomposition (280–350 °C) [64]. Finally, the char yield at 600 °C was similar for all the samples because they had the same mixture proportions. As can be observed in Fig. 2a, TGA curves showed that increased Al2O3 particle size improved thermal stability of the composites. The DTA curves (Fig. 2b) show the thermal degradation characteristics more clearly. The maximum decomposition peaks in the DTA curves slightly shifted toward higher temperature as the particle size of Al2O3 increased, which agrees with the results from TGA. The maximum decomposition (DTAmax) peak with larger Al2O3 particles was lower than with smaller particles, indicating that the larger Al2O3 particles increased the rate of degradation of starch.

Fig. 2
figure 2

a TGA and b DTA of thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes

Full size image

Dynamic mechanical thermal analysis

Storage modulus (E′) and loss modulus (E″) are related to stiffness, representing elastic and viscous components in deformation response, respectively, while the loss factor (tan δ) relates to molecular motions [65]; these are shown in Fig. 3. Figure 3a shows the storage modulus (E′) as a function of temperature for the thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes. Thermoplastic starch filled with 5 μm Al2O3 provided higher storage modulus than the other cases. Glass transition temperature (Tg) was determined from the maximum peak of tan δ and is presented in Table 1 and Fig. 3c. Two Tg values were detected. The first one is attributed to relaxation of glycerol rich phases and did not significantly change with Al2O3 particle size. The second relaxation is attributed to starch chain mobility and exhibited clear changes, indicating that Al2O3 can interact with starch rich phases more strongly than with glycerol rich phases [66]. It was found that the second relaxation of starch rich phases shifted to lower temperatures with increasing Al2O3 particle size, indicating that small-sized Al2O3 filler hindered chain mobility of starch. This suggests that good dispersion of the finer filler was achieved. Tg tended to decrease with Al2O3 particle size, from 39.6 °C with 0.05 μm Al2O3 to 15.1 °C with 5 μm Al2O3 as seen in Table 1. The increased Tg with smaller Al2O3 particles was attributed to increased interfacial interactions with the polymer matrix, because the finer particles had higher specific surface and restricted molecular chain motions by the interactions [67]. However, the Tg of these thermoplastic starch composites was not detectable by DSC. Zhang et al. [68] reported that DSC is not sensitive enough to measure the heat-flow changes of starch samples.

Fig. 3
figure 3

Evolution of a storage modulus, b loss modulus and c tan δ with temperature for thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes

Full size image
Table 1 Tensile properties and glass transition temperature of the thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes
Full size table

Contact angle

The hydrophilicity of material surfaces can be characterized by water droplet contact angles. A low contact angle indicates high hydrophilicity. The contact angles for the thermoplastic starches with different sized Al2O3 fillers are shown in Fig. 4. It was found that the contact angle increased with Al2O3 particle size, indicating that the composite surface became more hydrophobic.

Fig. 4
figure 4

Water contact angles on the thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes

Full size image

Scanning electron microscopy

Figure 5 exhibits SEM images of native cassava starch and fracture surfaces for the thermoplastic starch matrix and the thermoplastic starch–TiO2–Al2O3 composites, at 2500× magnification. Spherical and truncated shapes and some oval shapes were observed for the native cassava starch granules shown in Fig. 5a. In addition, various size granules and agglomerates were found, as results reported also earlier [69, 70]. After adding glycerol and under heat and high shear, the original granular structure was transformed into a molten homogeneous TPS, as seen in Fig. 5b. The fracture surface of thermoplastic starch had holes and remnants of starch granules. However, no cracks were observed in any of the samples. Furthermore, good adhesion between the fillers and TPS phase is likely, since no holes or voids are seen from loss of filler particles. The smaller Al2O3 particles were more homogeneously dispersed in the TPS phase than the larger particles that tended to agglomerate, as seen in Fig. 5e.

Fig. 5
figure 5

SEM images of a native cassava starch, b pure TPS, and of the TPS–TiO2–Al2O3 composites with Al2O3 particle sizes, c 0.05 μm, d 1 μm and e 5 μm

Full size image

Mechanical properties

Nonlinear stress–strain curves are seen in Fig. 6. The stress continuously increased with strain until breakage without necking, indicating typical plastic behavior [71]. Young’s modulus, tensile strength and elongation at break are summarized in Table 1.

Fig. 6
figure 6

Stress-strain curves of the thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes

Full size image

The modulus refers to the stiffness of the material. It is obtained from the ratio of stress and strain in elastic region of the stress–strain curve. It was found that the particle size of Al2O3 in the range studied did not affect the Young’s modulus. Similar lack of particle size effects on the elastic modulus has been reported for epoxy resin–alumina trihydrate composite (1, 2, 5, 8 and 12 μm) [72] and for polyester-aluminum composite (100 nm, 3.5 and 20 μm) [73]. When the particle size is larger than a critical limit, it will not affect modulus of the composite. In addition, the tensile strength tended to decrease with filler particle size because the larger particles had lower specific surface and less efficient stress transfer [74]. Furthermore, the larger Al2O3 particles improved the matrix surface resistance to indentation, as the surface hardness of a composite mainly is attributed to hardness of the solid filler [75]. Thus, hardness increased with Al2O3 particle size because of increased contact of the filler, as seen in Fig. 7.

Fig. 7
figure 7

Hardness of the thermoplastic starch–TiO2–Al2O3 composites with different Al2O3 particle sizes

Full size image

Conclusions

Biocomposites based on thermoplastic cassava starch containing TiO2 and Al2O3 combination filler, with varied Al2O3 particle size (0.05 μm, 1 μm or 5 μm), were successfully prepared by melt mixing in an internal mixer followed by thermo-compression. The results showed that water contact angle, hardness and thermal stability slightly increased, whereas dielectric constant, dissipation factor and glass transition temperature decreased with Al2O3 particle size. However, the tensile properties were not significantly affected by Al2O3 particle size. Aggregates of the larger sized Al2O3 particles were observed in SEM images.