Influence of Thermal Annealing Temperatures on Powder Mould Effectiveness to Avoid Deformations in ABS and PLA 3D-Printed Parts

Fused deposition modelling (FDM)-printed parts can be treated with various post-processes to improve their mechanical properties, dimensional accuracy and surface finish. Samples of polylactic acid (PLA) and acrylonitrile butadiene styrene (ABS) parts are treated with annealing to study a ceramic powder mould’s effectiveness in order to avoid dimensional part deformation. The variables chosen are annealing temperatures and the usage of a ceramic powder mould to avoid part deformations. A flexural strength test was carried out to evaluate the mould’s influence on the mechanical properties of the part. The effectiveness of the mould has been evaluated mainly attending to the length of the part, because this is the dimension most affected by deformation. A polynomial approximation to a deformation’s length and the effectiveness of the mould allows for their prediction. Results obtained show that effectiveness increases with the annealing temperature. Nevertheless, mould effectiveness decreases when parts are fabricated with PLA, because it is a semi-crystalline thermoplastic, and it suffers a lower shrinkage during thermal post-process than amorphous polymers such as ABS. Attending to the flexural strength test, mould has no significant influence on the mechanical properties of the treated parts in both materials studied.


Introduction
Additive manufacturing (AM) is becoming a technological revolution into industrial processes. These technologies have great potential, and their number of industrial applications is widely increasing [1]. One of the most widely used and affordable AM techniques is fused deposition modelling (FDM) [2]. This technology has several advantages, such as the capability to produce complex shapes without manufacturing restraints, design flexibility, low equipment and material costs and a wide range of plastic materials, as well as the ability to manufacture large-scale components [3]. As a result, FDM has emerged as a flexible and powerful technique in the advanced manufacturing industry. This technology is becoming widely used as a manufacturing process in both aerospace and automotive industries [4]. However, 3D printing processes present numerous challenges in its applications and clarify the advantages and disadvantages, and comparison with conventional manufacturing methods is needed [5].
FDM technology allows for the use of a wide range of thermoplastic types. Acrylonitrilebutadiene-styrene (ABS) and polylactic acid (PLA) are among the most widely used deformations during the annealing post-process and mould effectiveness at a wide range of temperatures. To reach this objective, a variation of part dimensions (length, width and height) will be measured according to annealing temperature. Mechanical property flexural strength of ABS and PLA are studied, as well as how they can be influenced by the use of the mould during the annealing process.

Test Specimens Design and Manufacturing
Samples are designed according to standard ISO 178:2019 [35] in order to determine the flexural properties. This geometry allows for an easy analysis of the dimensional variations occurring during the annealing post-process.
ABS and PLA samples are manufactured using a FDM 3D printer Ultimaker 3 Extended (Ultimaker B.V., Utrecht, The Netherlands) equipped with a 0.4 mm-diameter nozzle. All building lines of each specimen are printed without differences between infill lines, wall lines and bottom and top layer lines. Unidirectional building lines printed in 0 degrees orientation in the XY plane (code XY + 0 according to [36]) with a 100% density infill line pattern are chosen for all samples. Specimen dimensions are 80 mm × 10 mm × 4 mm. (Figure 1). Three-dimensional-printed specimen geometry is modelled using Inventor software (Autodesk, Inc., San Rafael, CA, USA). Cura software (Ultimaker B.V., Utrecht, The Netherlands) was used to generate G-code files and to command and control all the FDM 3D printing parameters.
In order to reduce the interaction between lines of the same layer, a layer height of 0.2 mm and a line width of 0.5 mm values are selected. In the 3D manufacturing process, printing parameter values are shown in Table 1.

Thermal Post-Process Treatment
The thermal post-process described by Lluch et al. [34] has been followed. The samples to be treated are introduced inside the mould over a 1 cm layer of dry alumina powder with an average grain size of 150 µm, (Protechno, Spain) ( Figure 2) and covered by another powder layer of the same thickness. A pressure of 12 g/cm 2 to avoid creep deformations is exerted on the alumina powder. No binders are used, as the pressure is sufficient for the ceramic powder to behave as a solid.

Thermal Post-Process Treatment
The thermal post-process described by Lluch et al. [34] has been followed. The samples to be treated are introduced inside the mould over a 1 cm layer of dry alumina powder with an average grain size of 150 μm, (Protechno, Spain) ( Figure 2) and covered by another powder layer of the same thickness. A pressure of 12 g/cm 2 to avoid creep deformations is exerted on the alumina powder. No binders are used, as the pressure is sufficient for the ceramic powder to behave as a solid. Samples are heat-treated in a convection furnace at different temperatures above the ABS Vicat softening temperature and the PLA glass transition temperature, respectively, for 120 min. A 10 °C/min ramp is used to ensure that the temperature inside the mould is as uniform as possible. Before unpacking samples, the mould was kept in the furnace until it reached room temperature. Samples are heat-treated in a convection furnace at different temperatures above the ABS Vicat softening temperature and the PLA glass transition temperature, respectively, for 120 min. A 10 • C/min ramp is used to ensure that the temperature inside the mould is as uniform as possible. Before unpacking samples, the mould was kept in the furnace until it reached room temperature.

Design of Experiments
The variables used to study the dimensional deformations and flexural strength after the thermal treatment are the material (ABS or PLA), the annealing temperature and the usage of a ceramic powder mould. The material and the usage of ceramic mould are studied at two levels, and the annealing temperature is studied at seven levels ( Table 2). Annealing temperatures are different for the studied materials. The temperature range starts at 3 • C above Vicat for the ABS material and at the glass transition point for PLA, and it ends near the temperature fusion point for both materials. Table 2 shows the annealing temperatures for ABS (without brackets) and PLA (in brackets).
A full factorial design is used to determine the combination of variable levels to use for each experimental case (Table 3). Five replicas of each experiment are performed. A total of 140 specimens (70 ABS and 70 PLA) were printed at 0 degrees in the XY plane orientation, according to Figure 1.

Dimensional Analysis
Samples' deformation due to thermal post-processing is determined in all three dimensions of the material. The length (L), width (W) and height (H) values of each specimen before and after heat treatment are measured to evaluate dimensional changes. The height (H) and width (W) values are obtained as an average of the values in three sections of each specimen ( Figure 3). Measurements are carried out according to standards using an electronic digital caliper instrument (resolution = 0.01 mm, accuracy = ±0.03 mm). L, W and H variation are calculated according to ISO 294-4:2018 [41] using Equations (1)-(3). The subscripts "f " and "o" indicate "after" and "before" thermal post-processing, respectively.  The subscripts "f" and "o" indicate "after" and "before" thermal post-processing, respectively.

Flexural Test Analysis
Three-point bending tests ( Figure 4) are performed on the heat-treated specimens (with and without mould) to determine the influence of the mould on the mechanical properties.

Flexural Test Analysis
Three-point bending tests (Figure 4) are performed on the heat-treated specimens (with and without mould) to determine the influence of the mould on the mechanical properties. The subscripts "f" and "o" indicate "after" and "before" thermal post-processing, respectively.

Flexural Test Analysis
Three-point bending tests (Figure 4) are performed on the heat-treated specimens (with and without mould) to determine the influence of the mould on the mechanical properties.  Tests are performed under ISO 178:2019 standards [35] using an Instrom 5967 (Illinois Tool Works Inc., Glenview, IL, USA) 30 kN load cell with a loading rate of 0.5 mm/s. All specimens are evaluated in the same orientation.
Experimental data are processed to obtain the flexural strain curve graphs and to calculate the maximum flexural stress (flexural strength). The mean and standard deviation of the maximum flexural stress values of the five specimens for each annealing temperature are taken as results (mean ± standard deviation). To obtain the flexural stress in megapascals, Equation (4) is used, where F is the load in Newtons, L is the span of the support in mm, and W and H are, respectively, the width and height of each tested specimen in mm. Tables 4 and 5 show average measurements observed in variations of length (∆L), width (∆W) and height (∆H) for all specimens of each sample. Positive values represent an expansion, whereas negative values represent a shrinkage of every evaluated dimension.

Results and Discussion
Values obtained for average dimensional variations are represented in a bar chart (Figures 5-7). Dimensional variations are directly proportional to the annealing temperature in all tests. ABS shows much higher deformations than PLA in all dimensions studied. It can be observed that the use of a ceramic powder mould has a significant effect on reducing dimensional variations in samples after the thermal treatment in both materials. Nevertheless, using a powder mould has a more significant effect on ABS than on PLA, because ABS undergoes greater deformations. In ABS and PLA, deformations during heat treatment are more important in the main length of the part (Figure 5), because the direction of the construction line coincides with the length dimension [34]. Similar behaviour can be observed at all annealing temperatures studied in this paper.

Mould Effectiveness to Avoid Deformations at Different Annealing Temperatures
Using a ceramic mould during the thermal post-process allows for a decrease in deformations in both materials ( Figures 5-7), although length is the most affected dimension for deformation. For this reason, length will be used to evaluate mould effectiveness in both materials.
Mould effectiveness is measured as it is shown in Equation (5), with E being the mould effectiveness to avoid deformations in the length L during the thermal post process.
A mould effectiveness value (E) of 100% means that samples treated with the mould avoid all deformations during the annealing post-process. A value of 0% means that the samples suffer identical deformations with or without the mould. The mould's effectiveness for each annealing temperature is shown in Tables 6 and 7.  The values obtained for the mould's effectiveness are represented in a bar chart (Figure 8). In both cases, the effectiveness of the mould increases with the annealing temperature. Nevertheless, it can also be observed that the mould's effectiveness is higher in amorphous polymers, or ABS, than in semi-crystalline polymers, or PLA. The annealing temperatures chosen are between the glass transition temperature, Tg, and the melting temperature of the polymers. It is well known that exceeding Tg increases molecular mobility, causing significant changes in thermal properties. Above the glass transition temperature, the polymer tends to expand isotropically, and hysteresis is observed in the expansion or dimensional changes upon cooling [42].   Figure 9 shows specimens treated at annealing temperatures studied with and without the mould. ABS presents high deformations and signs of degradation at 240 • C thermal treatment temperature. PLA presents less deformation in all temperatures, and no degradation is observed. The mould's effectiveness is adequate to reduce deformations in both materials, but the mould is less effective at low annealing temperatures in PLA parts. On the other hand, the mould is highly recommended for ABS parts or parts with high annealing temperatures.  The polynomial approximation is less precise when a mould is used, because deformations are minor in samples with the mould, as measurement error is greater. Nonetheless, the adjustment to a polynomial of third degree is still correct.

Prediction of Deformations and Mould Effectiviness
The third degree polynomial fitting of each graph can be expressed in equation form to obtain variations of each dimension as a function of the annealing temperature (Tables 8 and 9).  Figures 10-16 approximate the results to a polynomial function to predict dimensional deformations and mould effectiveness of ABS and PLA parts during the annealing post-process. The best results are obtained by fitting a third-degree polynomial function.          Table 9. PLA deformation as a function of annealing temperature. As shown in Figure 16 the mould's effectiveness also fits well to the third polynomial in both materials studied. Therefore, it can be expressed in equation form (Table 10). Table 10. Mould effectiveness as a function of annealing temperature.

ABS PLA
The behaviour of the mould at different temperatures is similar in both materials, despite the difference between amorphous (ABS) and semi-crystalline (PLA) materials (Table 10). They differ in the scale on the effectiveness axis, which is smaller in semicrystalline materials such as PLA. It can be concluded that dimensional variation and mould effectiveness are predictable at different annealing temperatures for ABS and PLA.

Mould Influence on the Mechanical Properties at Different Annealing Temperatures
The annealing post-process is performed to improve the mechanical properties of FDM parts. The use of the mould during annealing is highly recommended to avoid deformations, but it could lead to variations in the mechanical properties of the material. For this purpose, the flexural strength of both materials studied at the different annealing temperatures will be analyzed.
ABS presents a slight improvement in the flexural mechanical properties of the heattreated specimens compared to the untreated ones ( Figure 17). In the treated specimens, improvement is lost from 170 • C and worsens drastically, with premature failure of the specimens occurring from 205 • C onwards. If a mould is not used above 135 • C, the specimen deformations are so high that the flexural test cannot be carried out. Figure 18 shows PLA flexural stress curves. A significant improvement in flexural mechanical properties is observed at 86 • C and above. Results are similar if mould is not used. Tables 11 and 12 show the average maximum flexural stress (flexural strength) of specimens from each sample. In tests 4-7 (Table 11), specimen deformations are so high that the flexural test cannot be carried out.   Values obtained for flexural strength are plotted in a graph ( Figure 19). Low annealing temperatures using a ceramic powder mould slightly reduce the flexural strength in both materials. In contrast, there is no influence on this mechanical property at high annealing temperatures.  Values obtained for flexural strength are plotted in a graph ( Figure 19). Low annealing temperatures using a ceramic powder mould slightly reduce the flexural strength in both materials. In contrast, there is no influence on this mechanical property at high annealing temperatures.  In ABS, at low annealing temperatures, the slight decrease in flexural strength is compensated for by the clear advantage of using a mould to reduce deformations significantly. In PLA, annealing temperatures above 86 • C improve the flexural properties significantly. There is no significant difference in flexural strength if a mould is used or not at these temperatures. On the other hand, using a mould is necessary to reduce part deformations at these temperatures.
An analysis of variance (ANOVA) is performed to quantify the influence of a mould and temperature during post-processing thermal treatment on the flexural strength of 3D-printed samples. An F test with a level significance of 0.05 has been used. Due to strong ABS deformations, if a mould is not used, analysis can only be applied at a temperature range between 100 and 135 • C (Table 13). PLA analysis has been performed for a low temperature range (Table 14) to compare with ABS and to the whole temperature range studied (Table 15). As "p values" are higher than 0.05 in all ANOVAs, there is no significant effect of the factor on the response. For this reason, in all annealing temperature ranges, a mould has no significant influence on the flexural strength of the treated parts, and it can be recommended in both materials.

Conclusions
In this research, a ceramic powder mould's influence on the annealing post-process of the two most widely used thermoplastic materials in this type of additive technique-ABS and PLA-has been studied. The annealing post-process has a significant influence on the mechanical properties of FDM-printed parts. However, in this post-process some deformations can appear that can cause the part to be rejected. The present study showed that a ceramic power mould could be a solution to preventing part deformation during the annealing post-process.
Length, width and height specimen variation during the annealing process at different temperatures and mould usage effectiveness have been evaluated. Deformations are directly proportional to the annealing temperature in all tests. ABS shows much higher deformations than PLA in all dimensions studied. However, length is the dimension most affected by deformation; therefore, the effectiveness of the mould has been evaluated mainly according to the length of the part. Usage of the ceramic powder mould has a great effect in reducing deformations in both materials. The mould's effectiveness increases with annealing temperature. Its behaviour at different temperatures is similar in both materials, although the mould's effectiveness is higher in amorphous polymers (ABS) than in semi-crystalline polymers (PLA). Dimensional variations and mould effectiveness have been fit to third order polynomial equations in order to predict them as a function of the annealing temperature and material used.
The flexural strength of ABS and PLA was tested. A wide range of annealing treatment temperatures and ceramic powder mould usage were considered. An analysis of variance (ANOVA) was performed to study the mould's influence on flexural strength during the annealing treatment. It was found that using this kind of mould during thermal postprocessing had no significant impact on flexural strength. It can be concluded that in order to avoid deformations in the annealing of ABS and PLA thermoplastics parts, the usage of a powder mould during annealing is highly recommended.