Thermal and mechanical properties of poly(lactic acid)/poly(butylene adipate-co-terephthalate)/calcium carbonate composite with single continuous morphology

2.1 Materials

PLA pellets (FY801, density: 1.24 g‧cm⁻³; melt mass-flow rate = 4 g·10 min⁻¹ at 190°C and 2.16 kg) were purchased from Anhui Fengyuan Co. Ltd. Polybutylene terephthalate/adipate (PBAT) (blow-molding grade) was purchased from Lanshantunhe Co. Ltd. Maleic anhydride-grafted polylactic acid (LGM, G1603) used as the compatibilizer was purchased from Foshan Zuogao plastic materials Co., Ltd. Micro-scaled calcium carbonate (m-CaCO₃) powders and nano-scaled calcium carbonate (n-CaCO₃) were supplied by Nanzhao Dingcheng calcium industry Co. Ltd. The characteristic parameters are listed in Table 1.

2.2 Preparation of samples

The formulation in the present study is shown in Table 2. The components were dried in a vacuum oven at 60°C for 12 h before use. Each component was taken according to the formulation and then premixed in a mechanical mixer at room temperature at 40 rpm for 5 min.

A twin-screw extruder was used for melt blending and granulation. There were six heating zones at the temperatures between 170°C and 190°C in the extruder. The screw speed was set at 60 rpm, and the braces were naturally cooled. The PLA/PBAT/CaCO₃ tablets were pressed and formed in a flat vulcanizer with the temperature set at 183°C and the pressure set at 10 MPa. The tablets were cut into test strips. The thickness of tensile test strip was 2 mm.

The tensile property was tested using an electronic universal testing machine according to GB/T 1040.1-2006. Tensile rate was set at 20 mm·min⁻¹. The test was repeated five times for each sample and took the average value.

2.3 DSC measurements

A differential scanning calorimeter (DSC) (DZ-DSC 300, Nanjing Dazhan Testing Instrument Co., Ltd.) was used to investigate the melting and crystallization behaviors of PLA/PBAT/CaCO₃ composites. The samples used for DSC characterization were from the above-mentioned strips. Each sample of 9 ± 1 mg was sealed into an aluminum crucible, and the test was carried out in a nitrogen atmosphere. The heating and cooling rates were set both at 20°C·min⁻¹.

The sample was first heated from 40°C to 210°C and held in the molten state for 5 min. The first heating DSC scan was recorded. The glass transition temperature (T _g), cold crystallization temperature (T _cc) and cold crystallization enthalpy (∆H _cc), melting temperature (T _m), and melting enthalpy (∆H _m) were obtained from the first heating DSC scans, which reflected the thermal behavior of the nascent sample. The crystallinity (X _m) was determined from ∆H _cc and ∆H _m using Eq. 1:

where Δ H m 0 = 93.6 J·g⁻¹ for 100% crystalline PLA (24) and w PLA is mass% of PLA in the sample.

The sample in the molten state was considered to have eliminated the thermal history. The sample without the thermal history was cooled at 20°C·min⁻¹ to 40°C, and the cooling DSC scans were recorded. The melt crystallization temperature (T _c) and the corresponding crystallization enthalpy (∆H _c) were obtained from the cooling curves. The crystallization half-time (t _1/2) was defined as the time spent from the onset of the crystallization to the point where the crystallization was 50% complete. The onset of the crystallization was labeled as T _co. After 3 min at 40°C, the cooled sample was reheated to 210°C, and the second heating DSC scan was also recorded.

2.4 Non-isothermal crystallization kinetics

To observe the non-isothermal crystallization characteristics, the aforementioned DSC technology was used. The sample was heated to 210°C at a heating rate of 20°C·min⁻¹ and held for 5 min in a nitrogen atmosphere. Subsequently, the sample was cooled down to room temperature at cooling rates of −4, −6, −8, and −10°C·min⁻¹. In the non-isothermal crystallization process, the crystallization temperature and crystallization time were transformed by Eq. 2:

where t was the crystallization time, T _o was the initiate crystallization temperature, T was the temperature at time t, and φ was the cooling rate (°C·min⁻¹).

2.5 SEM observation

The fractured surfaces of the impact bars were coated with gold and observed in a scanning electron microscope (Regulus 8100, Hitachi, Japan).

3.1 Melting and crystallization behaviors

The morphology of the PLA/PBAT blend was reported in literatures. Deng et al. (25) presented a schematic diagram showing the phase morphology of the PBAT/PLA blend over a full range of composition, as deduced from the melt viscosity, optical micro graph, tensile properties, and SEM fracture surfaces. PBAT and PLA formed bicontinuous phase in the composition range between 19.0 and 40 wt% in the schematic diagram. Li et al. (26) observed PBAT spherical droplets dispersed in a PLA matrix when the PBAT content was lower than 20 wt%. In the present study, the effect of the CaCO₃ filler on the thermal behavior of the composite with the PLA and PBAT content ratio of 90–10 was investigated, and the corresponding DSC scanning curves were recorded and are shown in Figure 1a–c when heating and cooling were carried out at 20°C·min⁻¹.

Figure 1

DSC scans of PLA/PBAT blend and PLA/PBAT/CaCO₃ composites for both heating and cooling at 20°C·min⁻¹: (a) the first heating, (b) cooling, and (c) the second heating.

The nascent samples of ⁹⁰L¹⁰B and its CaCO₃ composites went through the glass transition, cold crystallization and melting in the first heating from 40°C to 200°C as shown in Figure 1a. The diverse curve shapes showed that either the particle size of the CaCO₃ filler or the content ratio of n-CaCO₃ and m-CaCO₃ particles had an effect on the melting and crystallization of the composites. PBAT is a random copolyester and therefore does not have a sufficiently symmetrical structure to give high levels of crystallinity. Generally, PBAT has a very broad and shallow endotherm around 100–120°C, indicating some crystallization of PBAT (25). However, the low content of PBAT in the present study made the PBAT endotherm more easily masked by the crystallization and melting of PLA. Therefore, it was reasonable to attribute the endotherm and exotherm in the DSC scanning to the PLA phase in the study. The exothermic peak around 80–120°C was attributed to the cold crystallization of PLA. The endothermic peak around 150–175°C was due to the PLA melting. The characteristic parameters coming from the DSC scanning are given in Table 3.

It was found by Yang et al. that the glass transition temperatures of PLA and PBAT did not change significantly after blending, indicating that the compatibility between the PLA and PBAT was poor (27). Compared with ⁹⁰L¹⁰B without the filler, the addition of the CaCO₃ particles significantly increased T _g of PLA as shown in Table 3, regardless of whether the particle size was nano-sized or micro-sized.The significant increase in glass transition temperature in the present study should be attributed to the introduction of CaCO₃ particles into the composites. As the CaCO₃ particles were activated by stearic acid in the present study, the ester was formed on the surface of the particles. The formed ester underwent transesterification with PLA and PBAT during melt mixing, thus establishing a connection between CaCO₃ particles and the matrix. Both n-CaCO₃ and m-CaCO₃ particles in the composites were strongly connected to PLA chains by transesterification during melt mixing, which reduced the motility of the PLA chain segments. The highest T _g at 69.4°C in ⁹⁰L¹⁰B⁵ⁿCa implied a more efficiently blocking effect of n-CaCO₃ particles on the motility of the PLA chain segments in the composite. Accordingly, the strong connection between five parts n-CaCO₃ and PLA also made the nucleation of n-CaCO₃ more effective. The strong connection and effective nucleation improved the melting temperature and crystallinity of PLA, and maximized them in ⁹⁰L¹⁰B⁵ⁿC as shown in Tables 3 and 4.

The cooling DSC scans at 20°C·min⁻¹ were recorded after the first heating as shown in Figure 1b. The samples without thermal history were divided into three types based on the PLA crystallization shown in Figure 1b. The first type involved ⁹⁰L¹⁰B and ⁹⁰L¹⁰B^5mC with double peaks of crystallization, the second one ⁹⁰L¹⁰B^11mC and ⁹⁰L¹⁰B^11n5mC with sole crystallization peak, and the third one ⁹⁰L¹⁰B⁵ⁿC, ⁹⁰L¹⁰B¹¹ⁿC and ⁹⁰L¹⁰B^5n11mC without crystallization peak. It is well understood that PLA is a polymorphous polymer. PLA exhibits two different crystalline phases termed α and α′. And α′ was described as a disordered form of the stable α phase. Differences between the crystalline structures were associated to chain conformation and packing mode of chains between the disordered and ordered forms. Crystallization at temperatures below 100°C lead predominately to the α′ crystalline form whereas at temperatures above 120°C, the α phase was the main form (28–30). Within the 100–120°C range, a mixture of these crystalline phases was present. Different thermal properties for α′ and α phases stem from the structural difference and interchain interactions (31,32). The double crystallization peaks at high and low temperature in ⁹⁰L¹⁰B and ⁹⁰L¹⁰B^5mC were attributed to α and α′ phase, respectively.

But the crystallization temperatures attributed to α′ and α phases in ⁹⁰L¹⁰B without the filler were different from those in ⁹⁰L¹⁰B^5mC with five parts m-CaCO₃ particles as shown in Figure 1b, respectively. The increase in α′ crystallization temperature from 89.2°C to 93.1°C was owing to the heterogeneous nucleation of m-CaCO₃ particles in ⁹⁰L¹⁰B^5mC, and the decrease in α crystallization temperature from 126°C to 123.7°C was attributed to the blocking effect of the connection between CaCO₃ particles and the matrix on the mobility of PLA chains. The sole exotherm around 94.2°C and 94.5°C in ⁹⁰L¹⁰B^11mC and ⁹⁰L¹⁰B^11n5mC, respectively, indicated that the connection completely blocked the ordered arrangement of PLA chains, but made the chains low orderly arranged on cooling at 20°C·min⁻¹. Further, the low order of arrangement was completely blocked in ⁹⁰L¹⁰B⁵ⁿC, ⁹⁰L¹⁰B¹¹ⁿC, and ⁹⁰L¹⁰B^5n11mC as shown in Figure 1b. The strong blocking effect was attributed to the huge specific surface area of n-CaCO₃ particles, correspondingly resulting in the tight connection to the matrix. Higher n-CaCO₃ amounts were less effective as they tended to agglomerate in ⁹⁰L¹⁰B^11n5mC, and the aggregation made the blocking effect of the connection was discounted in the composite compared with that in ⁹⁰L¹⁰B^5n11mC. The blocked PLA chains orderly rearranged in the second heating, so that the cold crystallization occurred, and the corresponding melting occurred at a higher temperature as shown in Figure 1c. In case of ⁹⁰L¹⁰B^5mC, ⁹⁰L¹⁰B^11mC, and ⁹⁰L¹⁰B^11n5mC, there were discernible crystallization exotherm as shown in Figure 1b. On subsequent heating, no crystallization peak was observed as shown in Figure 1c since the PLA was already crystallized to its maximum in the cooling cycle.

3.2 Non-isothermal crystallization of ⁹⁰L¹⁰B and its composites with CaCO₃ particles

In order to further describe the α′ crystallization behaviour of the composites, the non-isothermal method was used in the differential scanning calorimeter. The cooling rate was set at −4, −6, −8, and −10°C·min⁻¹ at 20°C·min⁻¹ after melting. The DSC scans are presented in Figure 2. X _t versus time is shown in Figure 3. Here, X _t was defined as the X _c at time t. The characteristic parameters shown in Figures 2 and 3 are listed in Table 5.

Figure 2

DSC thermograms of PLA/PBAT blend and PLA/PBAT/CaCO₃ composites for cooling at the scan rate of (a) −4°C·min⁻¹, (b) −6°C·min⁻¹, (c) −8°C·min⁻¹, and (d) −10°C·min⁻¹.

Figure 3

Plots of X _t versus t at cooling rates of −4, −6, −8, and −10°C·min⁻¹ for (a) AB, (b) ABCa-m₅, (c) ABCa-m₁₁, and (d) ABCa-n₁₁m₅.

T ^α _c of α crystalline of ⁹⁰L¹⁰B and ⁹⁰L¹⁰B^5mC remained stable between 122°C and 123°C with increase of the cooling rate from −2 to −10°C·min⁻¹, which was lower than that at a more rapid cooling rate of −20°C·min⁻¹ as shown in Table 4. The half time of crystallization, t _1/2, defined as the time required to reach 50% of the α′ final crystallinity. t _1/2 of each of the samples decreased with increasing of the cooling rate from −2 to −10°C·min⁻¹ as shown in Table 6. However, there was only a small reduction in t _1/2 of ⁹⁰L¹⁰B, which was consistent with the coinciding curves in Figure 3a. This indicated that the PLA chains in ⁹⁰L¹⁰B were more mobile than those in the CaCO₃ composites and could keep up with the cooling rate increasing.

3.3 Jeziorny method

The α′ crystallization behavior was further investigated using Jeziorny method. The isothermal crystallization kinetic was widely described by the Avrami model (19) in which X _t was expressed in the following form:

where n is the Avrami crystallization exponent and dependent on the nucleation mechanism and growth dimensions; t is the crystallization time, and Z _t is the crystallization rate constant, which depended on the nucleation and growth of crystal. The above equation can be rewritten into a linear equation:

However, the Avrami model was not suitable for non-isothermal crystallization because the temperature constants varied and both nucleation and growth of crystal were temperature dependent. Jeziorny proposed that Z _t was affected by the cooling rate ( φ ) and was verified by having a cooling rate term as shown in Eq. 5 (33):

where Z c is the corrected Jeziorny crystallization constant. According to Eq. 4, n and Z _t were obtained from the slope and intercept of the curve between ln[−ln (1−X _t )] and lnt shown in Figure 4. Then, the value of Z _c was calculated using Eq. 4. The values of n and Z _c were listed in Table 7.

Figure 4

Avrami plots for (a) AB, (b) ABCa-m₅, (c) ABCa-m₁₁, and (d) ABCa-n₁₁m₅.

Compared with that of ⁹⁰L¹⁰B, the n values of ⁹⁰L¹⁰B^5mC, ⁹⁰L¹⁰B^11mC, and ⁹⁰L¹⁰B^11n5mC increased slightly, and the n value did not change regularly with increase in the cooling rate for each of the composites, indicating that their nucleation patterns were close to each other. As for the crystallization rate, Z _c of each of the composites almost always increased significantly as the cooling rate increased, which was consistent with the parameters in Table 5.

3.4 Mo method

Assuming that the non-isothermal crystallization process was a small step of isothermal crystallization by developing Avrami’s theoretical model, Ozawa’s theoretical model was used to describe the non-isothermal crystallization. The fractional crystallinity ( X t ) of time t was shown in Eq. 6:

The linear form of this equation was more commonly used as Eq. 7:

where m and K _o were obtained from the slope and interception on the y-axis of the graph from the relationship between ln [ − ln ( 1 − X t ) ] and ln ( φ ) , respectively.

Mo’s method combined Avrami equation with Ozawa equation to describe non-isothermal crystallization process. Mo equation was as follows:

where F ( T ) is the Mo-modified crystallization rate parameter, γ is the ratio of n in Avrami equation to m in Ozawa equation, which related to crystallization dimension. Obviously, plotting ln φ versus ln t at a given value of relative crystallinity yielded a series of straight lines, as shown in Figure 5. Then, F ( T ) and γ were obtained from the intercept and slope of the straight lines. Table 8 lists the parameters obtained from the plots of ln φ versus ln t for all samples listed in Table 2. It showed that all plots exhibited good linear relationship which indicated that Mo’s method was fit to describe the non-isothermal crystallization kinetics in this study.

Figure 5

Plots of lg φ versus lgt for the cold crystallization of (a) AB, (b) ABCa-m₅, (c) ABCa-m₁₁, and (d) ABCa-n₁₁m₅.

As shown in Table 8, the obtained F(T) increased with an increase in X _t , but the γ value hardly changed in the samples without n-CaCO₃ particles involving ⁹⁰L¹⁰B, ⁹⁰L¹⁰B^5mC, and ⁹⁰L¹⁰B^11mC except for ⁹⁰L¹⁰B^11n5mC. This indicated that the faster the cooling rate, the higher the α′ crystallinity achieved in unit crystallization time, and the crystallization mode did not change at different cooling rates. In other words, the achievement of higher X _t required faster cooling rate in each of the samples. On the other hand, the γ values of ⁹⁰L¹⁰B^5mC and ⁹⁰L¹⁰B^11mC were closer and both were lower than that of ⁹⁰L¹⁰B, indicating that α′ crystals grew in the same way in ⁹⁰L¹⁰B^5mC and ⁹⁰L¹⁰B^11mC, but in a different way from that in ⁹⁰L¹⁰B. The above-mentioned analysis showed that the Mo method was suitable for describing the α′ crystallization in ⁹⁰L¹⁰B, ⁹⁰L¹⁰B^5mC, and ⁹⁰L¹⁰B^11mC.

Different from the earlier samples, F(T) and the γ value of ⁹⁰L¹⁰B^11n5mC decreased with the increase of X _t . This indicated that the achievement of higher X _t required slower cooling rate in the sample. This difference between ⁹⁰L¹⁰B^11n5mC and the others was presumably due to the presence of n-CaCO₃ particles in ⁹⁰L¹⁰B^11n5mC.

3.5 Tensile property analysis

Tensile testing was used to determine the mechanical properties of ⁹⁰L¹⁰B and its CaCO₃ composites. The results are shown in Figure 6a–d. The structure in the composites is usually considered to involve three components: dispersed filler, continuous phase, and interface. A strong connection among the components ensured that external forces were transferred to the dispersed filler, so that the mechanical properties were significantly improved. Therefore, the strong connection at the interface was necessary to obtain high mechanical properties. The relatively high tensile modulus and strength of ⁹⁰L¹⁰B⁵ⁿC should be attributed to this strong connection of PLA and n-CaCO₃ particles at the interface.

Figure 6

(continued)

As shown, ⁹⁰L¹⁰B⁵ⁿC exhibited the highest tensile properties involving the tensile modulus, strength, and yield strength in the present study, owing to the strong connection between the five parts n-CaCO₃ particles and the PLA matrix. In contrast, ⁹⁰L¹⁰B¹¹ⁿC had lower tensile properties, which indicated that the increased amount of n-CaCO₃ was not beneficial to the improvement of mechanical properties. The decrease in tensile properties of ⁹⁰L¹⁰B¹¹ⁿC was owing to the tendency of the nano-sized particles to aggregate at a high concentration, which reduced the overall connection strength between the filler and the PLA matrix. The decrease in other composites compared to ⁹⁰L¹⁰B⁵ⁿC was also attributed to the decrease in the overall connection strength. As for ⁹⁰L¹⁰B^5mC and ⁹⁰L¹⁰B^11mC without n-CaCO₃, the lower tensile strength stemmed from the low specific surface area of m-CaCO₃ particles compare with the n-CaCO₃ particles, which decreased the overall connection strength between the micro-sized filler and the PLA matrix. But the connection strength decreasing was obviously beneficial to the improvement of impact strength of the composites as shown in Figure 6e. ⁹⁰L¹⁰B^11mC and ⁹⁰L¹⁰B^5n11mC with the connection decreasing had the highest impact strength.

3.6 SEM observation

Figure 7 shows SEM micrographs of the fractured surface of impact test strips of PLA/PBAT/CaCO₃ composites. In Figure 7a, the fractured surface displayed “sea-island” structure with PBAT beads dispersing in the PLA matrix, and some beads were detached from the matrix after the impact test. However, the interface between PLA and PBAT phases was ambiguous, which was attributed to the compatibilization of LGM in ⁹⁰L¹⁰B. The comparison of micrographs in Figure 7b and c showed that excessive n-CaCO₃ particles in ⁹⁰L¹⁰B¹⁰ⁿCa caused accumulation, which reduced the mechanical properties compared with ⁹⁰L¹⁰B⁵ⁿCa. In Figure 7b, the coarse fractured surfaces indicated a strong adhesion between CaCO₃ particles and PLA matrix. Like that of ⁹⁰L¹⁰B⁵ⁿCa, the fractured surface of ⁹⁰L¹⁰B^5mCa was also coarse as shown in Figure 7d, while excessive m-CaCO₃ particles tended to gather together in Figure 7e, just as excessive n-CaCO₃ particles tended to gather together as in Figure 7c. Although the fractured surfaces of both ⁹⁰L¹⁰B^5n11mCa and ⁹⁰L¹⁰B^11n5mCa were rough as shown in Figure 7f and g, the morphology was different from each other, which probably resulted the difference in the thermal and mechanical properties.

Figure 7

SEM images of the impact-fractured surfaces of PLA/PBAT/CaCO₃ composites: (a) ⁹⁰L¹⁰B, (b) ⁹⁰L¹⁰B⁵ⁿCa, (c) ⁹⁰L¹⁰B¹⁰ⁿCa, (d) ⁹⁰L¹⁰B^5mCa, (e) ⁹⁰L¹⁰B^10mCa, (f) ⁹⁰L¹⁰B^5n10mCa, and (g) ⁹⁰L¹⁰B^11n5mCa.