PLA filament is not equal to PLA filament: Experimental studies of the influence of the type of pigment on the mechanical and thermal properties of poly(lactic acid) products

PLA filament was purchased from 3D Printers Company (Poland). Eleven filaments with different pigments and one filament without pigment were tested, which was a reference to the other samples. PLA samples were printed in the following colours: natural (without pigment), white, silver, iron gray, black, green leaf, yellow, orange, red, sky blue, dark blue and magenta. Detailed information on the chemical composition of the pigments, their concentration and the type of polylactide is not available. Figure 1 shows all printed sample types (in all colours) used in testing.

Fig. 1.

Test samples were printed using the HBOT 3D F300 printer. The accuracy of its printing (provided by the manufacturer) is +/− 0.1 mm. The working area of the printer is 300x300x300 [mm]. The working platform is made of borosilicate glass. The diameter of the PLA filament was 2.85 mm. All samples were printed with the same printing parameters as shown in Table 1. After the printing process, the samples were not subjected to any finishing treatment. They were left for 72 hours in order to season.

Figure 2 shows the bending curves of natural samples. The maximum flexural stress was reached at the moment of sample fracture. Figures 3 and 4 show the relationship between bending strength and bending modulus as a function of the type of pigment. The manufacturer’s specification does not include parameters measured in the bend test. The highest flexural strength is found for natural samples (103 MPa), and the lowest for silver (84 MPa) and white samples (85 MPa). Generally, it can be seen that the addition of any pigment more or less significantly reduces the strength of the material (pure PLA), the change in strength is in the range of 5–18%.

Fig. 2.

Fig. 3.

Fig. 4.

Analysing the flexural modulus, the highest values have the white (3.17 GPa) and natural (3.15 GPa) samples, and the lowest have magenta and dark blue samples (slightly over 2.90 GPa). The differences in properties are less visible here. The addition of pigment does not affect the module value so much, which in a few cases remains at a similar level within the margin of error (white, black, yellow), and the maximum decrease in the module value is about 7%. Apart from the modulus value for white and black samples, the differences in its values for individual materials are qualitatively similar to the differences in flexural strength.

The obtained differences in the results of flexural strength and flexural modulus for individual materials are similar to the values of tensile strength and Young’s modulus. The results of tensile strength tests were not presented in detail due to their very similar nature and analogous differences for individual materials. The modulus values for tensile and bending are very similar (for tensile they are higher by approximately 10%), while the tensile strength values are approximately 40% lower than the flexural strength for individual materials.

Figure 5 shows the obtained compression curves for natural samples, while Figures 6 and 7 show the results of compressive strength and compression modulus measurements. There are no properties of the compression test in the manufacturer’s specification. Analysing the measured values, it can be seen that the white (90 MPa) samples are by far the most resistant to compression, in contrast to the iron gray (73 MPa) samples. We can see that the addition of any pigment not only does not reduce the value of compressive strength, but in some cases even improves it, unlike in the case of bending. The improvement is significant for the white material and minimal for the red material. This is due to the different characteristics of the test – in bending the sample is subjected to simultaneous stretching and compression, unlike the compression process itself. Another reason may be probably the highest pigment content in the white material, which is confirmed by the results of density tests of individual materials. Although these results were not presented graphically, it is worth adding that the measured density values were approx. 1.22±0.01 g/cm³ for all types of materials except for white, whose density was determined to be 1.30 g/cm³, and thus significantly differs from the values for PLA in other colours. Compressive strength for individual materials varies (except for white) in the range of 73–82 MPa. However, taking into account the white material, the difference between the material with the highest and the lowest strength is about 19%.

Fig. 5.

Fig. 6.

Fig. 7.

Analysing the compression modulus of tested materials, the white material again exhibits the highest stiffness in compression, probably due to the previously described higher density of this material compared to the other filaments, which is important in the compression test. The biggest difference is between white and black material and amounts to almost 18%. “Coloured” materials (from green leaf to dark blue) have similar modules (within the error limits), amounting to less than 800 MPa.

Figure 8 shows the dependence of the unnotched Charpy impact strength as a function of the type of pigment. The impact strength value for pure PLA is approx. 18 kJ/m². Different effects of particular pigments can be observed in the case of impact strength, although the addition of pigment reduces impact strength in general. The silver material (21.6 kJ/m²) has the highest impact strength – the improvement in impact strength compared to pure PLA is 18%. Silver is the only material whose impact strength is higher than that of pure PLA (natural). The least resistant is green leaf (11.8 kJ/m²), and for this material, the deterioration of impact strength is approximately 35% compared to natural material. When analysing the impact strength of all materials, the difference between silver and green leaf is in the order of 45%.

Fig. 8.

Figure 9 shows the dependence of the softening temperature as a function of the type of pigment. The highest value was obtained by natural material (59.7°C) and iron gray (59.8°C) materials. The addition of any other pigment lowers the PLA’s softening point. The effect of the pigment on the deterioration of heat resistance is very small, what means the order of 1°C. The exception is the white material, which has the lowest softening point of 56.5°C. Here, the reduction of the Vicat’s temperature in relation to the natural material is 3.2°C (5.4%).

Fig. 9.

The melting temperatures (T_m) of the samples ranged from 151.8 to 162.8°C (during the first heating) and from 149.2 to 161.3°C (during the second heating) in DSC tests carried out in this work. The glass transition temperatures (T_g) (read from the first heating) ranged from 61.8 to 66.2°C. In the case of pure PLA (natural), T_m during the first and second heating and T_g were, respectively: 162.5°C, 158.6°C and 63.4°C. Other PLA samples contain pigments as additives. Analysing the literature data, it can be assumed that the tested polymer samples may have a semi-crystalline structure, in which the share of D enantiomers does not exceed 12%. Melting points, not exceeding 162.8°C, exclude the possibility of the tested samples in the form of blends. They may, therefore, be PLA copolymers. However, referring to the data from Table 1, in the case of PLLA/PDLA copolymers, the glass transition temperature does not exceed 60°C, and the melting point ranges from 125 to 164° C, depending on the contents of D enantiomers. In the case of PLLA homopolymer, the glass transition temperature is 63°C, but the melting point is much higher than the T_m of the tested samples, and it reaches 178°C [31], while the range of glass transition temperatures for semi-crystalline PLLA is from 60 to 70°C, according to [32]. On the other hand, for semi-crystalline PLA, the range of glass transition temperatures is from 55 to 60° C (with a melting point range from 160 to 180°C) [33]. Taking into account the fact that most of commercial PLA are copolymers of the L and D enantiomers of lactic acid (meso-lactide) with the L-lactide content above 95% [34] and taking into account the values of the melting points of the samples tested in this work, it can be concluded that that the analysed PLA samples are semi-crystalline PDLLA copolymers with the content of D enantiomers not exceeding 10%. The shift of the glass transition temperature above 60° C can be explained by the presence of modifying additives and pigments. In the case of the α and α′ polymorphs, the melting points are approximately 180 and 150°C, respectively [32], which suggests that the α′ polymorphic structure prevails in the tested PLA samples.

Tables 2 and 3 summarise the basic thermal parameters obtained from the first and second heating of the samples, respectively, and the degree of crystallinity (X) calculated according to the formula (1). When one temperature peak was observed during melting, it was marked as Tm1. When two temperature peaks were observed during melting, they were marked as Tm1 and Tm2.

All samples had similar curves (Figure 10) during the first heating from 20° C to 200° C, in which the glass transition temperature (T_g) was distinguished, then the exothermic process of cold crystallisation and the endothermic melting process were observed (Figure 11). Two temperature peaks were observed during melting for the natural, white, black and red samples (Table 2).

Fig. 10.

Fig. 11.

The first heating process reveals thermal effects related to the thermal history of the sample, such as sample preparation (thermal and mechanical impacts related to the production of the sample – the production of filaments and the 3D printing process in this case), aging or annealing the sample. In the case of samples that are 3D-printed products, their properties may be affected by the temperature of the nozzle (235°C in this case) as well as the storage conditions of the samples. It should also be noted that during the printing process, the temperature of the platform was 60°C, and the printing process itself is time-consuming. Therefore, 3D printed PLA products can be treated as partially aged/annealed, because during their production they are subjected to ambient temperature, close to the range of glass transition temperatures of the tested samples. During this time, further spatial ordering of polymer chains may occur (system with lower energy, more thermodynamically stable), and thus an increase in the proportion of crystal structures. The presence of double temperature peaks during PLA melting was explained in the work [35], where the following theories were presented:

the lamellar thickness model – the double thermal peak in the endothermic transformation is associated with the presence of two types of lamellae of different thickness. Thinner lamellae have a lower melting point, and thicker ones – a higher one;

The difference between the total enthalpy of endothermic melting and exothermic cold crystallisation processes is the confirmation of the last model assuming the existence of three crystal structures. The presence of the β form (with a melting point lower by 10°C compared to the α form, that is, approximately 170° C) can be expected, in addition to the α and α′ forms. α′ crystal structures nucleate at temperatures below 100°C. Both α and α′ polymorphic forms nucleate in the range between 100 and 120°C. Only the α form crystallises above 120° C. Polymorphic transition from the α′ to α structure takes place at a temperature of approximately 115°C [36].

In order to analyse the obtained results in terms of possible correlations between the results obtained from DSC measurements and the mechanical and thermal properties of the tested materials, the linear correlation coefficients R were determined. The coefficients of determination R² are presented in Tables 4 and 5. Coefficients (I) and (II) indicate the heating number.

Observation of the results presented in Table 4 does not generally indicate any correlation between the degree of crystallinity X and other thermal parameters, with the exception of two negative correlations related to crystallisation parameters: the correlation R between X and Tc (II) is -0.60, and between X and Hc(I) is -0.75.

The analysis of the correlation of mechanical test results with the degree of crystallinity of the material also did not bring any effects, except for a small positive correlation between X and the bending modulus (R = 0.45). It seems that none of the determined parameters (strength, impact strength and softening point) is in any way positively correlated with the degree of crystallinity of the tested PLA systems. This result confirms the observations contained in the work [4], which also did not indicate any significant correlations in these studies. However, other correlations can be observed. There is a negative correlation between Tg and compressive strength (R= -0.62). It is difficult to clearly indicate the reasons for this correlation at this moment.

It is worth noting that only Pearson’s linear correlation coefficients were measured. Therefore, it is not possible to exclude the occurrence of nonlinear correlations between the examined parameters, although linear correlation was the most expected here.

Thermal history is an important factor influencing the mechanical properties of polymers [24]. It includes sample aging [32], heating and cooling rates and sample processing [37]. An amorphous layer (interphase) can be distinguished in the semicrystalline structure of the polymer, directly adjacent to the crystalline structures. This layer is called rigid amorphous (RA), in contrast to the layer not directly adjacent to the crystal structures (mobile amorphous, [MA]). The RA layer partially crosses the crystalline-amorphous boundary, affecting the modulus of elasticity, which is similar to the modulus of the crystalline part. The RA layer is also strongly dependent on the crystallisation temperature (Tc). As the value of Tc increases, the amount of RA fraction decreases. The thermal effects observed near the Tg concern the MA fraction. Thermal effects are observed well above the Tg [30] in the case of the RA fraction, which is significantly limited in terms of mobility. Formulas (2–4) were used to determine the share of the RA fraction [32]. However, it should be taken into account that they do not include the enthalpy values during the sample cooling process (when the formation of the RA fraction is induced at temperatures above Tg), and therefore they do not take into account the entire possible thermal history of the sample. 2 wc=ΔHm−ΔHcΔHm0 \[\begin{align}{{w}_{c}}=\frac{\Delta {{H}_{m}}-\Delta {{H}_{c}}}{\Delta H_{m}^{0}} \end{align}\] 3 wma=ΔcpΔcp,a \[\begin{align}{{w}_{ma}}=\frac{\Delta {{c}_{p}}}{\Delta {{c}_{p,a}}} \end{align}\] 4 wra=1−wc−wma \[\begin{align}{{w}_{ra}}=1-{{w}_{c}}-{{w}_{ma}} \end{align}\] where:

ΔH_m – melting enthalpy of the sample, [J/g]

The exothermic crystallisation peak was observed during the cooling stage from 200^°C to 50°C only for the black and yellow samples (Figure 12). The other materials did not show changes in this area in the form of characteristic exothermic effects (such as the natural sample in Figure 12). The enthalpy values of the temperature peak during cooling for the black and yellow materials were 20.12 J/g (99.4°C) and 10.66 J/g (96.8°C), respectively. These temperatures significantly exceed the glass transition temperatures (64.6°C and 65.5°C for black and yellow samples, respectively), but are much lower than melting temperatures observed during the first heating (for black sample: T_m1=155°C and T_m2=148°C, for yellow sample: T_m1=151.8°C). These effects may be related to the previously discussed RA fraction. These samples show the highest proportion of the crystalline phase among all the tested samples (black: 16.8%, yellow: 18.5%). Therefore, a large share of RA interphase bordering between crystalline and amorphous structures (MA fraction) can be expected. Furthermore, the lowest temperatures for the cold crystallisation process were observed for both samples during the second heating process (from 50°C to 200°C) amounting to 102.7°C (black) and 105.9°C (yellow). The temperatures of the cold crystallisation process observed during the second heating oscillated from 113.1°C (sky blue) to 131.4°C (natural) for the remaining samples without a clear relationship between the observed temperature and the degree of crystallinity. Comparing the values of cold crystallisation temperature from the first and second heating process it can be observed that the highest values of these temperatures are characteristic for the natural material (respectively, 106.4°C for the first heating and 131.4°C for the second heating). Reducing the cold crystallisation temperatures in the case of samples with pigments may mean the nucleation of crystal structures at lower temperatures, and thus the acceleration of the processes related to the further ordering of polymer chains and the relaxation of sample stresses.

Fig. 12.