The effect of hydroxyapatite nanoparticles on crystallization and thermomechanical properties of PLLA matrix

Materials

The PLLA used in this study was purchased from Nature Works Co. (USA). More specifically, the grade Ingeo PLLA 4032D is a semicrystalline material in pellet form with L-lactide: D-lactide ratio 98:2. The glass transition temperature and melting point are 55–60°C and 165–170°C, respectively, as reported by the manufacturer. The nano filler Hydroxyapatite (HA) powder with particle size smaller than 200 nm, was purchased from Sigma-Aldrich (MO, USA).

Preparation of PLLA based nanocomposites

HA/PLLA composites with compositions 100/0, 5/95, 10/90, 20/80 and 0/100 (w/w) were prepared by extrusion melt blending in a Haake co-rotating twin-screw extruder, with L/D=25 and 16 mm diameter (Haake Polylab System, Haake Rheomex PTW16, Thermo Electron Corporation). The extruder was heated at five zones along the cylinder and the die. The temperature profiles of the barrel from the hopper to the die were 200–205–210–210–210–215°C and the screw speed 40 rpm. Prior to the melt extrusion, raw PLLA pellets were pulverized cryogenically into uniform powder with a typical particle size distribution from 0.2 to 0.5 mm (FRITSCH, pulverisette 14). PLLA and HA were dried in a vacuum oven at 80°C for 4 h, in order to eliminate hydrolytic degradation reactions. Before processing, the appropriate amounts of PLLA and HA were dry-mixed. After melt extrusion, the obtained material, in the form of continuous strands, was granulated into regular, cylindrical pellets using a Brabender knife pelletizer.

X-ray diffraction

X-ray Diffraction analyses of HA, PLLA and HA/PLLA nanocomposites were made using a BRUKER D8-ADVANCE (twin/twin) diffractometer (40 kV, 40 mA) with a Cu X-ray tube (λ=1.5418 Å). The recording speed was 0.05°/s in the range of angles of 2θ=(5°–55°). Samples for X-ray analysis were obtained from compression-molded films.

Differential scanning calorimetry (DSC)

DSC measurements were run in a Mettler Toledo model DSC 1 differential scanning calorimeter with pure indium as a calibration standard.

1. For the non-isothermal experiments, a total of 8–10 mg of each sample was heated from room temperature to 190°C and equilibrated at 190°C for 3 min to completely eliminate previous thermal history. The samples were then cooled to 20°C at a cooling rate of 10°C/min and kept at 20°C for 3 min to expose all samples to the same thermal treatment. Finally, they were reheated to 190°C at a rate of 10°C/min.

2. For the isothermal crystallization experiments, a total of 8–10 mg of each sample was heated from room temperature to 190°C and equilibrated at 190°C for 3 min to completely eliminate previous thermal history. The samples were then cooled to one of the examined isothermal crystallization temperatures (100°C, 110°C, 115°C, 120°C) at a cooling rate of 20°C/min to avoid the crystallization of the polymer matrix and equilibrated at the predetermined isothermal crystallization temperature for 45 min until crystallization was completely developed. Finally, they were reheated to 190°C at a rate of 10°C/min.

All runs were conducted under nitrogen to avoid thermo-oxidative degradation.

The degree of crystallization (X_c, %) of PLLA and HA/PLLA nanocomposites was evaluated using Eq. (1) [2]:

where ΔH_m indicates the melting enthalpy (J/g) that was calculated from the fusion peak in DSC second heating curves, ΔHm0 is the melting enthalpy of the 100% crystalline polymer matrix and f is the weight fraction of PLLA in the composite. Various theoretical values of 100% crystalline PLLA melting enthalpy (ΔHm0) have been reported ranging from 91 to 148 J/g [26], [27], [28], [29]. The value of 93.1 J/g was used to calculate the degree of crystallization (X_c) [29], [30].

X-ray diffraction (XRD)

In order to elucidate the effect of HA content on the crystallinity of PLLA due to the incorporation of inorganic nanoparticles, the pure polymer as well as its nanocomposites containing HA were characterized by X-ray diffraction analysis. The X-ray diffraction patterns of the pure PLLA and the examined compositions of HA/PLLA (w/w) are presented in Fig. 1.

Fig. 1:

Diffraction patterns of (a) pure PLLA, (b) 5/95, (c) 10/90, (d) 20/80 HA/PLLA (w/w) nanocomposites and (e) pure HA.

HA nanoparticles have crystalline structures with sharp diffraction peaks as can be seen in Fig. 1e. In fact, the two peaks observed at 2θ=25.7° and 31.7° are clear and sharp for pure HA, showing the crystalline nature of this reinforcement. In Fig. 1b–d it is observed that all the characteristic peaks for HA remain in the patterns of HA/PLLA nanocomposites. They appear at the same angle and with slightly lower intensity, as compared to those of pure HA. The intensities of all peaks for HA do not significantly change because of the relatively simple blending process followed for the incorporation of HA into the PLLA matrix. These results are very similar to those of Sui and Gang, who have prepared HA/PLLA membranes using the solvent casting technique [6]. Differences observed in these spectra originate from interactions with the PLLA matrix. The same behavior has been also observed by other researchers [2], [6]. The observed dominating diffraction peak in the spectra of pure PLLA and HA/PLLA composites (Fig. 1a–d) centered at 2θ=16.7° and smaller peaks at 2θ=19.1° and 35.6° are consistent with reports in literature and can be corresponded to the (110/200), (010) and (203) reflections of the α crystal form with significant disorder [18], [31], [32]. It was also found that the intensity of this main sharp peak was increased with increasing HA content. Its position remained unaffected by the incorporation of HA. In Fig. 1a, the broadness of the area in the range 2θ=10–25°, under the main sharp diffraction peak at 2θ=16.7° is indicative of pure PLLA’s less organized structure [33], [34]. It was observed in the spectra of HA/PLLA composites that the broadness of this area was decreased with increasing HA content in the PLLA matrix. This behavior is consistent with the melting peaks obtained by the DSC experiments (Fig. 2b) and can be attributed to the transformation of the less ordered crystalline PLLA phase into a better organized crystalline structure, which resulted in a shift of T_m,PLLA at the higher melting rate to a higher temperature range.

Fig. 2:

Thermal analysis of pure PLLA and HA/PLLA (w/w) nanocomposites: (a) cooling and (b) second heating cycle curve.

Thermal analysis

DSC is a useful tool for the study of phase transitions of PLLA, such as glass transition, melting and cold crystallization. In this work, DSC thermograms (Fig. 2) were used to identify the thermal transitions of pure PLLA and HA/PLLA nanocomposites.

The glass transition temperature (T_g) of the PLLA matrix remained almost unaffected by the incorporation of HA. Moreover, no crystallization peak was observed during the cooling cycle (Fig. 2a) at the examined cooling rate of 10°C/min. An exothermic peak, associated with the cold crystallization process of PLLA, was detected in all the examined samples. This peak is due to the reorganization of macromolecular chains in the amorphous domains quenched before, which again obtain the enhanced flexibility and mobility during the DSC heating process, into crystalline regions [35].

From Table 1 it can be observed that for pure PLLA, the cold crystallization peak was detected at 120.2°C. With increasing HA content in the nanocomposites, the cold crystallization process is initiated at lower temperatures, which means that the inorganic nanoparticles facilitate packing of PLLA chains into crystalline structures. In particular, a sharper cold-crystallization peak was observed, shifted to lower cold crystallization temperatures by about 10°C, since the dispersed HA nanoparticles played the role of nucleation centers in the crystallization process of PLLA’s chains. In comparison to pure PLLA’s crystallization enthalpy (ΔH_cc≈36 J/g), the crystallization enthalpy of the PLLA phase in the nanocomposites was slightly decreased (ΔH_cc≈32 J/g). This decrease remains almost constant as the HA content increases in the composite. This behavior is attributed to two competitive mechanisms induced by the HA nanoparticles: the enhancement of nucleation ability and the inhibition of PLLA chains mobility, because of their presence.

The melting enthalpy of PLLA phase follows the same trend with the exothermic enthalpy of cold-crystallization, suggesting that the crystallinity of PLLA phase mainly originates from the cold crystallization process. From Fig. 2b it is obvious that by incorporation of HA nanoparticles in the polymer matrix, the symmetric melting peak of pristine PLLA was analyzed in a main peak which shifts to higher temperatures, followed by a shoulder which appears in the low-temperature flank of the curve, coming from the less ordered crystalline material.

The melting temperature depends on the size and perfection of the obtained crystalline structure of PLLA affected by the presence of HA particles in the bulk of the polymer matrix [9]. The shape of the melting peak could be a result of the polymorphic crystalline transition, since PLLA can display three different kinds of crystal structures, i.e. α, α′, β and γ. As it has been verified by the XRD diffraction patterns, the PLLA crystals obtained under the examined conditions belong mainly to the α form which is more stable and it is characterized by relatively high melting temperature [18]. The increase in the intensity of the diffraction peak attributed to the α phase of the PLLA matrix (Fig. 1) with the incorporation of HA, is accompanied by an increase in melting temperature of the PLLA phase in the nanocomposites (~170°C), in comparison with that of pure PLLA (~165°C) (Table 1). This increase remains unaffected by the increase of HA concentration in the nanocomposites.

Isothermal crystallization

Isothermal crystallization behavior of pure PLLA

In order to examine the crystallization behavior of pure PLLA, isothermal crystallization experiments were performed. All the examined samples of PLLA were imposed to the same thermal treatment until their melting and then they were rapidly cooled down to the different examined isothermal crystallization temperatures, from 90 to 130°C, until full crystallization was reached.

From the DSC isothermal crystallization curves shown in Fig. 3 it is observed that there is an increase in the crystallization rate from 100 to 110°C reaching a maximum at 115°C. At the higher examined T_ic temperatures (120, 130°C) crystallization rate decreases.

Fig. 3:

Isothermal DSC curves of pure PLLA, at various temperatures.

Effect of HA content

Figure 4a–d show the crystallization exotherms vs crystallization time of PLLA and its nanocomposites with different compositions, isothermally crystallized at 100°C (a), 110°C (b), 115°C (c) and 120°C (d).

Fig. 4:

Isothermal crystallization curves of pure PLLA and HA/PLLA (w/w) nanocomposites at 100°C (a), 110°C (b), 115°C (c) and 120°C (d).

With the addition of HA at 100°C and 110°C, the crystallization rate was increased in comparison with pristine PLLA and the shape of the exothermic crystallization peaks became narrower (Fig. 4a, b). This effect was more significant with increasing HA content. The inverse effect was observed at higher T_ic (i.e. 110°C and 120°C, being closer to PLLA’s melting temperature, where it was observed that crystallization process took more time and the crystallization exothermic peaks became wider in shape (Fig. 4c, d). Therefore it is concluded that crystallization mechanism of PLLA is affected not only by the incorporation and the content of inorganic nanoparticles, but also by the crystallization temperature of the test.

Based on the above graphs, the relative crystalline conversion degree (X_t) can be obtained from the ratio of the area of the exotherms up to time t (ΔH_t) divided by the total exotherms (ΔH_∞) (Eq. (2)) leading to the Avrami equation [22]:

(2)Xt=ΔHtΔH∞=∫0t(dHdt)dt/∫0∞(dHdt)dt=1−exp(ktn)

where dH/dt is the DSC heat flow rate.

By evaluating the degree of crystalline conversion (X_t) vs crystallization time at a constant temperature, for 0/100 (a), 5/95 (b) 10/90 (c) and 20/80 (d) HA/PLLA (w/w) as shown in Fig. 5a–d it can be observed that all the isotherms curves exhibited a sigmoid dependence with time.

Fig. 5:

Time dependence of relative crystallinity for isothermal crystallization of (a) 0/100 (b) 5/95 (c) 10/90 (d) 20/80 (w/w) HA/PLLA nanocomposites at different T_ic.

From Fig. 5a it was verified that PLLA shows the higher crystallization rate at 115°C, as already presented in Fig. 3. Regarding the examined HA/PLLA nanocomposites, from Fig. 5b–d it was noticed that with the increase of crystallization temperature, the characteristic isotherms were shifted to higher values of time, indicating a progressive decrease in the crystallization rate which leads to an increase in the time required for completion of crystallization. In fact, in the case of nanocomposites the maximum spherulitic growth rate occurred at the lower examined T_ic (100°C) and at 120°C it took the longest time.

Avrami analysis

The crystallization kinetics of isothermal crystallization of PLLA and its nanocomposites at the examined temperatures, were analyzed using the Avrami equation (Eq. (2)), where k is the crystallization rate constant related to the crystallization kinetics, for nucleation and growth rate, and n is the Avrami exponent, related to the mechanism of crystal nucleation as well as on the form of crystal growth. Based on the data of Fig. 5, the double logarithmic form of the Avrami equation (Eq. (3)) [22] was plotted in Fig. 6:

Fig. 6:

Avrami double-logarithmic plots for isothermal crystallization of pure PLLA and HA/PLLA (w/w) nanocomposites at (a) 100°C, (b) 110°C, (c) 115°C and (d) 120°C.

(3)log[−ln(1−Xt)] = logk + nlog(t)

applied to pure PLLA and HA/PLLA nanocomposites, at the examined crystallization temperatures.

The values of n, k were calculated from the part of the curves representing 10–90% of crystalline conversion, where a linear relation between log[–ln(1–X_t)] and log(t) appeared during the crystallization process, indicating that the isothermal crystallization behavior of the samples can be described by the Avrami equation. The values of n were calculated from the slope of the linear part and the kinetic constant k was determined by the intercept for pure PLLA and HA/PLLA nanocomposites, under the examined isothermal temperatures. The results are listed in Table 2. The Avrami exponent n represents the dimensionality of the growth and is affected by many factors, such as nucleation density and restriction of crystalline formation due to the presence of inorganic filler.

Table 2:

Avrami equation parameters for isothermal crystallization of pure PLLA and PLLA in HA nanocomposites.

HA/PLLA (w/w)	0/100		5/95		10/90		20/80
Tic (°C)	n	k (min−n)	n	k (min−n)	n	k (min−n)	n	k (min−n)
100	2.73	8.80×10−4	2.62	2.16×10−3	2.52	3.71×10−3	2.75	4.84×10−3
110	2.76	7.30×10−4	2.44	2.83×10−3	2.48	3.54×10−3	2.76	2.75×10−3
115	2.72	2.05×10−3	2.88	1.04×10−3	2.82	8.60×10−4	2.76	9.00×10−4
120	2.68	1.14×10−3	2.49	2.94×10−3	2.67	4.50×10−4	2.58	7.70×10−4

For spherulite three-dimensional growth, n is 4 for homogeneous nucleation system and 3 for heterogeneous nucleation system. The values of n are decimal due to the presence of crystalline branching and/or two stage crystal growth during the crystallization process and/or mixed growth and nucleation mechanism [36], [37]. The obtained n values for pure PLLA (2.7–2.8) in the examined T_ic range, are in good agreement with those reported in literature [11], [38], [39]. Therefore, a spherulitic growth is suggested from nuclei initiated at time zero or a disc like growth from nuclei initiated over time or a combination of the above mechanisms [38]. Regarding HA/PLLA nanocomposites, n values closer to 2 were observed, indicating a 2-D growth on a lamellar structure [2]. When the n values were closer to 3, three-dimensional spherical growth took place. A tendency of reduction in the n values obtained by the nanocomposites indicates an enhanced level of heterogeneous nucleation verifying the contribution of the nanoparticles in the formation of nuclei [38], [40].

An increase of k values (Table 2) for PLLA with increasing temperature of isothermal crystallization T_ic was observed, reaching a maximum at 115°C, and then (at 120°C) a decrease took place again. No significant effect of T_ic in k values of PLLA in nanocomposites with low HA content (5 wt%) was observed. In nanocomposites with higher HA content (10 & 20 wt%) the calculated k values at 100°C and 110°C revealed faster crystal growth rate, in comparison with pure PLLA matrix, and this rate decreased at higher crystallization temperatures.

The half-time of crystallization t_1/2 is defined as the time to reach 50% crystallization. It can be obtained either directly from the experimental data in Fig. 5 or from the n and k values (as seen in Table 2), using Eq. (4) [41]:

(4)t1/2 = (ln2/k)1/n

From Table 3 it was confirmed that pure PLLA presented the lowest value of t_1/2 at 115°C. The incorporation of HA nanoparticles significantly accelerated the crystallization process of PLLA matrix at the lower examined T_ic (100 & 110°C). In this area of temperatures, as HA content increases in the nanocomposites, shorter t_1/2 is needed to attain 50% crystallization of the PLLA matrix. In the case of nanocomposites, it can also be observed that with the increase of examined T_ict_1/2 values also increases and, therefore, longer time is needed for the completion of crystallization process.

Table 3:

t_1/2 values during the isothermal crystallization experiments of pure PLLA and PLLA in HA nanocomposites.

Tic (°C)	t1/2 (min)
HA/PLLA (w/w)	0/100	5/95	10/90	20/80
100	11.55	8.36	8.04	6.09
110	11.98	9.50	8.39	7.41
115	8.54	12.20	10.71	11.07
120	10.93	14.43	15.62	13.99

t_1/2 can also be used to directly characterize the crystallization rate, since the corresponding half-time of crystallization (1/t_1/2) can be considered approximately proportional to the crystal growth rate (G) (Eq. (5)) [41].

(5)G = 1/t1/2

From the results of Table 4 it can be seen that the values of G for pure PLLA followed a bell-shaped temperature dependence showing a maximum at 115°C, which is consistent with related literature data [38], [41], [42]. With the addition of HA in the PLLA matrix, the crystallization rate was faster at 100°C and it decreased with increasing temperature becoming slower at higher testing temperatures (115, 120°C). Relatively higher T_ic can be considered disadvantageous for the development of crystallinity in the PLLA matrix as it tends towards its melting temperature. This behavior is in agreement with the decreased crystallization rate of the PLLA matrix after the turning point at 115°C. The aforementioned observations verify that as the mobility of polymeric molecules is low, crystal nucleation is more favorable at low temperatures. However, crystal growth is more favorable at higher temperatures where the viscosity of a polymer is low [43].

Table 4:

Radial growth rate (G) of pure PLLA and PLLA in HA nanocomposites after isothermal crystallization experiments at different examined T_ic.

Tic (°C)	G (min−1)
HA/PLLA (w/w)	0/100	5/95	10/90	20/80
100	0.087	0.110	0.124	0.164
110	0.083	0.105	0.119	0.135
115	0.117	0.082	0.093	0.090
120	0.091	0.069	0.064	0.071

Melting behavior after isothermal crystallization

Figure 7 presents the melting endotherms of PLLA and HA/PLLA nanocomposites. It can be observed that the shape of the melting peak for PLLA matrix is dependent on the crystallization history of the samples. By heating the examined samples after isothermal crystallization at 100°C, two endothermic peaks were observed, namely T_m1 and T_m2, respectively (Fig. 7a, Table 5). A broad melting peak was observed after isothermal crystallization at 110°C (Fig. 7b). However, in the case of isothermal crystallization at higher temperatures (115°C, 120°C), a single melting peak appears (Fig. 7c, d). Multiple melting behavior has been observed for many semi-crystalline polymers corresponding to different crystal structures. According to He & Fan, when the degree of supercooling (Tm0−Tic) is relatively high, the initially formed lamellae can hardly evolve because of the high transportation energy resulting in a small endothermic peak of T_m1 followed by melt-recrystallization [44]. On the contrary, with lower degree of supercooling, almost perfect crystals are formed during melt-crystallization, thus suppressing melt-recrystallization. As a result, a single melting endotherm was observed [44]. According to Di Lorenzo [1] when crystallization is conducted at low temperatures, small and/or defective crystals develop. The occurrence of two recrystallization exotherms or a broad exotherm as PLLA’s melting behavior can be linked to crystal reorganization during heating. Part of the less ordered crystalline material (α′ phase) can be transformed to α phase at elevated crystallization temperatures >120°C [18]. This behavior has been verified by the shape of the melting peak of the PLLA matrix after the isothermal crystallization of all the examined samples at 120°C (Fig. 7) where a less broad melting peak is observed.

Fig. 7:

Melting peaks of pure PLLA and HA/PLLA nanocomposites after isothermal crystallization at (a) 100°C, (b) 110°C, (c) 115°C and (d) 120°C.

After completion of crystallization at lower T_ic, a double melting mechanism is observed in the melting curve of the examined samples indicating the gradual melting of α′ and α PLLA. This behavior appears to be independent of the HA content in the nanocomposites. After the isothermal crystallization of the nanocomposites at higher temperatures only a single melting peak is observed that becomes sharper with increasing T_ic. It is known that the less ordered α′ crystalline phase formed at lower crystallization temperatures (100, 110°C) can be transformed to the α phase at elevated temperatures >120°C [17], [18].

As far as the degree of crystallization (X_c %) is concerned (Table 6), the presence of HA may promote nucleation thermodynamically but, at the same time, it can impend crystal growth dynamically, resulting in a variation of the percent crystallinity [38], [45].

Comparing the values of t_1/2 (Table 3) with the corresponding crystallinity of PLLA nanocomposites, determined by heating after isothermal crystallization, for the lower T_ic (100, 110°C) it was observed enhanced crystallinity accompanied by lower values of t_1/2. Therefore, it is obvious that, for these systems, not only the crystallization ability is improved but the phenomenon proceeds with high crystallization rate. Isothermal crystallization study at T_ic 120°C showed both, increased crystallinity with higher values of t_1/2.

Equilibrium melting point (Tm0)

According to the theoretical consideration by Hoffman and Weeks [46], the equilibrium melting point temperature Tm0 can be obtained by linear extrapolation of the T_m versus T_c plot to the line T_m=T_c (Fig. 8a) and the dependence of the T_m on the T_c is given by the following equation (Eq. (6)):

Fig. 8:

Equilibrium melting point (a) and Lauritzen-Hoffman (L-H) plot (b), for pure PLLA and PLLA in HA nanocomposites.

(6)Tm=Tm0(1−1γ)−(Τicγ)

where Tm0 is the equilibrium melting point and γ is the thickening ratio [46] which describes the growth of lamellar thickness during crystallization. Under equilibrium conditions γ equals to 2 [35], [41]. According to Souza et al. [39], who also studied the isothermal crystallization kinetics of Poly(lactic acid) nanocomposites, in order to determine the Tm0 of PLA, values corresponding to the lower-temperature endothermic peak (T_m1) were used and taken from Table 5.

The equilibrium melting points obtained for samples with different HA content are presented in Table 7 and they are within the range of values (166–225°C) reported in the literature for the different compositions of PLLA [14], [39], [41], [47], [48]. For the examined type of PLLA the value of Tm0=173.6 °C is within the above range, representing the melting temperature of the crystals with the most perfect lamellae or, according to the theory derived by Hoffman and Weeks [46], the melting temperature of infinitely extended crystals of this polymer. Liao et al. [41] have attributed the variation of the values of Tm0 of PLLA matrix in the nanocomposites, to the addition of the HA nanoparticles that might have attenuated the formed crystals and, correspondingly, decreased the thickness of the crystal lamellae.

Table 7:

Equilibrium melting point Tm0 of pure PLLA and PLLA in HA nanocomposites.

HA/PLLA (w/w)	Equation	R2	Tm0 (°C)
0/100	y=0.134x+150.33	0.9813	173.6
5/95	y=0.187x+144.69	0.9438	177.9
10/90	y=0.145x+149.87	0.983	175.4
20/80	y=0.194x+143.39	0.9991	177.9