Effect of PA12 powder recycling on properties of SLS 3D printed parts including their hygroskopicity

Effect of PA12 powder recycling on properties of SLS 3D printed parts including their hygroscopicity Jana Machotov´ a a , Marek Pag´ aˇ c b , Roman Svoboda c , Jan Jansa b , ˇ Stˇ ep´ an Podzimek a , Eva ˇ Cernoˇ skov´ a d , Jiˇ rí Palarˇ cík e , Zuzana Koutov´ a f , Petr Kut´ alek g , Lucie Z´ arybnick´ a h,i,* a Institute of Chemistry and Technology of Macromolecular Materials, Faculty of Chemical Technology, University of Pardubice, Studentsk´ a 573, 532 10 Pardubice, Czech Republic b Department of Machining, Assembly and Engineering Technology, Faculty of Mechanical Engineering, VSB-TU Ostrava, 17. Listopadu 2172/15, 708 00 OstravaPoruba, Czech Republic c Department of Physical Chemistry, Faculty of Chemical Technology, University of Pardubice, Studentsk´ a 573, 532 10 Pardubice, Czech Republic d Department of General and Inorganic Chemistry, Faculty of Chemical Technology, University of Pardubice, Studentsk´ a 573, 532 10 Pardubice, Czech Republic e Institute of Environmental and Chemical Engineering, Faculty of Chemical Technology, University of Pardubice, Studentsk´ a 573, 532 10 Pardubice, Czech Republic f MemBrain s.r.o., Pod Vinicí 87, 471 27 Str´ aˇ z pod Ralskem, Czech Republic g Center of Materials and Nanotechnologies, Faculty of Chemical Technology, Cs. Legii 565, 530 02 Pardubice, Czech Republic h Department of Technical Studies, College of Polytechnics Jihlava, Tolst´ eho 16, 586 01 Jihlava, Czech Republic i Institute of Theoretical and Applied Mechanics of the Czech Academy of Sciences, Centre Telˇ c, Proseck´ a 809/76, 190 00 Praha 9, Czech Republic ARTICLE INFO Keywords: Selective laser sintering Print orientation Polyamide 12 Recycling Hygroscopicity ABSTRACT The ageing and recycling of polyamide 12 (PA12) powder are key issues in selective laser sintering (SLS) processing, while hygroscopicity is an important defect of PA12 products during use. This paper aims to correlate PA12 powder recycling with changes in the hygroscopicity of printed parts, which is significant for a comprehensive understanding of the impact of ageing on the performance of PA12 parts. PA12 powder was utilized in different grades: virgin, post-industrial recycled, and virgin-post-industrial recycled mixture (virgin-to-post-industrial recycled powder weight ratio of 25/75). The mutual effects of the powder grade, print orientation used in the building process, and absorbed moisture on the properties of printed parts were evaluated. Molar mass increase and secondary crystallization have been demonstrated in the post-industrial recycled powder, being responsible for the worse sintering quality in the SLS process. Therefore, the printed parts made from the postindustrial recycled powder or the virgin-post-industrial recycled powder mixture exhibited worse mechanical strength, a higher porosity, and hence a higher level of hygroscopicity in terms of moisture absorption and wettability. On the contrary, the printed parts made from the virgin powder were found to be more moisturesensitive in terms of the decline in mechanical strength in the wet state because of the higher content of the amorphous phase holding the plasticizing water. 1. Introduction Powder bed fusion (PBF), selective laser sintering (SLS) is a widely used additive manufacturing technique that involves the use of a highpower laser to selectively fuse small powder (typically polymeric) particles into a mass that has a desired three-dimensional (3D) shape [1]. The laser selectively sinters powdered material by scanning crosssections generated from a 3D digital description of the part (for example, from a Computer-Aided Design (CAD) file) on the surface of a powder bed. After each cross-section is scanned, the powder bed is lowered by one layer, a new layer of material is applied on top, and the process is repeated until the part is completed. In contrast with other widely used additive manufacturing techniques, namely stereolithography (SLA) or fused filament fabrication (FFF) [2–4], which most often require special support structures to fabricate overhanging designs, SLS does not need a separate feeder for support material because the part being constructed is always surrounded by unsintered powder. This allows the construction of previously impossible geometries with high accuracy and resolution [5,6]. Therefore, SLS technology is particularly popular in many industries due to its ability to produce * Corresponding author. E-mail address: [email protected] (L. Z´ arybnick´ a). Contents lists available at ScienceDirect European Polymer Journal journal homepage: www.elsevier.com/locate/europolj https://doi.org/10.1016/j.eurpolymj.2024.113432 Received 27 June 2024; Received in revised form 1 September 2024; Accepted 2 September 2024 European Polymer Journal 220 (2024) 113432 Available online 6 September 2024 0014-3057/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). complex geometries with little added manufacturing effort. Its most common applications are in automotive, electronics, aerospace, medicine, and pharmacy, where it has been widely used to rapidly produce prototypes, low-volume parts, and biomedical devices [7–10]. It should be noted that compared to SLA and FFF techniques, SLS is characterized by the following disadvantages: expensive cleaning of the printer, utilization of fine powder, the high price of 3D printers, and non-uniform material shrinkage [11]. Commercially available materials used in SLS come in powder form and include polymers such as polyamides (PA), polystyrene, thermoplastic elastomers, and polyaryl ether ketone [12]. Polyamide 12 (PA12) is one of the most used materials for SLS due to its ideal sintering behaviour as a semi-crystalline thermoplastic polymer, resulting in printed parts with desirable mechanical and thermal properties [13–18]. In terms of applications, SLS printing using PA12 has been used in various fields, such as dentistry, water applications, and composite material manufacturing [17–20]. The properties of laser-sintered PA12 structures in terms of their mechanical characteristics, porosity, and surface roughness [21], depend on several factors, including process parameters (such as powder bed temperature, laser power, atmospheric conditions, layer thickness, scan spacing, heating and cooling rates) [22–24], print orientation [25], presence of additives, and finally polymer powder properties such as size distribution, morphology, and shape [26–31]. In addition, material properties (such as melt viscosity, melting and crystallization behavior [32]) also affect the achievable density, microstructure, and mechanical properties of printed parts [33]. Surface properties can be positively influenced by surface treatments such as electrodeposition [34]. The polymer powder utilization efficiency during SLS 3D printing is low, meaning that 85 −95 % of the volume ratio of powder remains unused and experiences physical and chemical changes due to the preheating, sintering, and cooling involved in the SLS process [35]. One challenge in the SLS technology is the recycling of polymer powder, which typically consists of mixing the unused polymer powder with the virgin powder. However, printed parts made from virgin-post-industrial recycled powder mixtures usually exhibit properties that differ from those of virgin powder [36,37]. Several research works have investigated the effect of PA12 powder recycling in SLS on the properties of printed parts, but the conclusions are not consistent. Alo et al. [38] observed a lowering of the size of postindustrial recycled PA12 powder compared to virgin powder because of particle fragmenting during the SLS process, which resulted in dimensional variation of laser-sintered objects. On the contrary, Martynkov´ a et al. [39] proved the sintering of PA12 post-industrial recycled powder, resulting in an increase in particle size and enhanced porosity of printed parts. Dadbakhsh et al. [32] found that PA12 powder ageing during SLS increased the melt viscosity and molar mass while the crystallinity decreased drastically, changing the crystal structure from α - to γ-form. Drummer et al. [40] also reported the increase in molar mass of postindustrial recycled PA12 due to post-condensation and crosslinking reactions proceeding at elevated temperatures during SLS, which resulted in undesirable surface finish effects, namely orange peel, after long recycling times [41,42]. In terms of mechanical properties, the effect of powder recycling has been less established, or the reported values are rather contradictory. For example, some reports [43–45] show no significant change in the tensile properties of printed parts made from postindustrial recycled powder, but others [46–48] present a lowered tensile strength and/or elongation. Despite these previous works, there is still poor knowledge about the effect of PA12 powder recycling in SLS on the hygroscopic nature of printed parts, which can affect their final properties, sometimes irreversibly. This study may be the first one to correlate PA12 powder recycling with changes in the hygroscopicity of printed parts. Three different grades of PA12 powder were tested, namely virgin powder, post-industrial recycled powder (i.e., a powder that remained unused during the SLS process), and the virgin-post-industrial recycled powder mixture (virgin-to-post-industrial recycled powder weight ratio of 25/ 75). The presented research involved the following investigations: (i) the PA12 powder properties, such as size and shape distribution; (ii) the PA12 material properties, particularly the melting behaviour and chemical structure; and (iii) the properties of printed parts. The effect of print orientation and absorbed moisture were also included in the study. The emphasis was placed on evaluating the hygroscopicity, microstructure, and mechanical properties of printed parts. The obtained information can guide designers in creating an effective PA12 powder recycling protocol in SLS to ensure consistently good-quality printed parts. 2. Materials and methods 2.1. Materials A commercial PA12 powder (PA 2200, EOS GmbH, Germany) was investigated in three different grades: (i) virgin, (ii) post-industrial recycled, and (iii) virgin-post-industrial recycled mixture. The postindustrial recycled powders were those industrially exposed to a variety of different SLS cycles (typically 2 −10 cycles), where the powder bed surface was heated to ~ 170 ◦C and the baseplate to ~ 130 ◦C during SLS. The virgin-post-industrial recycled powder mixture comprised 25 wt% of the virgin powder and 75 wt% of the post-industrial recycled powder, which is a similar ratio reported in a recent reference [49], showing that adding 30 wt% of the virgin to the aged PA12 powder offered a good compromise between maintaining part performance. As the present study aims to follow a positive economic and environmental impact, the economization of the input material was prioritized. To obtain commercially viable cost savings, 25 wt% of virgin powder was used in a virgin-post-industrial recycled powder mixture (about 17 % of cost savings can be achieved compared to utilizing 50 wt% of the virgin powder, which manufacturers standardly recommend). The detailed calculation based on recent PA12 powder prices is presented in Supplementary Material. 2.2. Selective laser sintering (SLS) processing The SLS printing was performed on an EOS P 396 PBF printer (EOS GmbH, Germany) working under a nitrogen atmosphere and utilizing a CO 2 laser (laser maximal power 70 W). The process parameters (Table 1) were constant for all the building cycles. Models have been created and edited using SOLIDWORKS 2020 (version 2020, Dassault Syst` emes, France) and exported in STL format. Digital models were sliced using EOS RP-Tools 6.210 (version 2016, EOS GmbH, Germany) and were converted for 3D printing using PSW 3.8 EOS (version 2016, EOS GmbH, Germany). In each building cycle, five tensile test specimens, five bending test specimens (8 ×1 ×0.4 cm 3 ), five water absorption test specimens (2 ×2 ×0.2 cm 3 ), five specimens for water contact angle measurement (7.6 ×2.6 ×0.2 cm 3 ), and two specimens for microscopic analysis (0.2 ×0.2 ×0.2 cm 3 ) were prepared. The tensile test specimens were printed as dog bones with dimensions specified in ˇ CSN EN ISO 527–2 [50]. The dimensions of the bending test specimens were in accordance with ˇ CSN EN ISO 178 [51]. The print orientation and position in the build chamber are shown in the example of tensile test specimens in Fig. 1. The printed parts differing in the powder grade and print orientation are marked as X_Y, where X is the grade of PA12 powder (virgin, mixture, and post-industrial recycled), and Y is the print orientation (A1, B1, and G1). 2.3. Methods The PA12 powder properties, such as size and shape distributions, were analyzed using a Morphologi 4 (Malvern Panalytical, UK) equipped with a fully automated static image analysis system. The molar mass distribution of PA12 powders was determined by size exclusion J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 2 chromatography (SEC) with a multi-angle light scattering (MALS) detector. The experimental set-up consisted of an Agilent 1200 Series isocratic pump, an autosampler coupled with a MALS detector DAWN NEON, and a refractive index detector Optilab NEON. ASTRA 8 software was used for the data collection and processing. The software and detectors were from Wyatt Technology (Santa Barbara, CA, USA). The system included two Shodex 8 ×300 mm HFIP-806 M columns using hexafluoro isopropanol with an addition of 0.02 M sodium trifluoroacetate as the mobile phase at a flow rate of 0.8 mL⋅min −1 . The samples were prepared in the mobile phase at the concentration of ≈3 mg⋅mL −1 , filtered with 0.45 μ m filters, and injected in the volume of 100 μ L. The PA12 powder grades were tested by differential scanning calorimetry (DSC) using a heat flow DSC Q2000 calorimeter (TA Instruments, USA) equipped with an autosampler, RCS90 refrigerating unit, and T-Zero technology. The DSC measurements were realized as a two-step temperature program, where the sample was first heated at 5 ◦C⋅min −1 from 25 to 225 ◦C, and then it was cooled at the same rate back to 25 ◦C. The samples were placed in hermetically sealed aluminum pans; the sample masses varied between 3.7–3.9 mg (accurately weighted to 0.01 mg). The sintering window was determined as a difference between the extrapolated onsets of the melting peak from the 1st (heating) step and the crystallization peak from the 2nd (cooling) step. The degree of crystallinity ( χ c ) of PA12 was calculated according to the literature [52] using the value of melting enthalpy (ΔH m ) of the 100 % crystalline PA12 polymer of 209 J⋅g −1 [43]. The thermogravimetric analysis (TGA) using an STA 504 thermal analyzer (TA Instruments, Germany) was utilized to test the thermal decomposition of PA12 powders. Measurements were performed in a nitrogen atmosphere at a heating rate of 20 ◦C⋅min −1 in the 30–600 ◦C temperature range. The sample weights ranged from 10 to 15 mg. The melt mass-flow rate (MFR) and melt volume-flow rate (MVR) of PA12 powders were determined according to ˇ CSN EN ISO 1133–2 [53] by an extrusion capillary plastometer LMI 5000 (Dynisco, Franklin MA, USA) under 230 ◦C/1.2 kg load. The chemical structure of printed parts was analyzed using Fourier-transform infrared (FT-IR) spectroscopy with a Nicolet iN10 instrument (Thermo Fisher Scientific, Waltham, MA, USA). Infrared spectra were obtained using the attenuated total reflectance (ATR) with a diamond crystal in the 3500–600 cm −1 range. The microstructures of PA12 powders were observed using a scanning electron microscope (SEM) Quanta 450 FEG (FEI, CZ) in the secondary electron mode at an acceleration voltage of 20 kV on a sample that was sputtered with a 7 nm thick layer of gold. The surface of printed parts in a dry and wet state was compared (without sputter coating) using a scanning electron microscope Quanta 250 FEG (FEI, USA). The wet state of samples was provided by conditioning the specimens in distilled water (conductivity 7 µS⋅cm −1 ) for 16 days at room temperature (23 ±2 ◦C). At the end of the immersion period, the wet specimen was immediately transferred to a holder (watered with distilled water) and scanned in ESEM mode (humidity 98 %, 1.5 ◦C) with an accelerating voltage of 10 kV. The surface roughness Table 1 SLS parameters set in the 3D printing process. Laser power / W Scanning speed / mm⋅s −1 Scan spacing / μ m Layer thickness / mm Temperature / ◦C Beam offset / mm Material dependent scaling / % Build chamber Feeder chamber X Y Z (0) Z (600) 5 600 150 0.12 171 130 0.35 3.15 3.23 2.55 1.40 Fig. 1. Positions and print orientations (A1, B1, and G1) of tensile test specimens in the build chamber. J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 3 of printed parts in a dry and wet state was measured using a mechanical profilometer SSC-01 (R.M.I., CZ) with a steel tip in a contact mode. The wet state was achieved by conditioning the specimens in distilled water for 16 days at room temperature. The device works on the principle of measuring the torsion of a carrying tip (stylus) and its conversion to height in the range of 15 nm −15 mm. The height resolution is 50 nm. The topography measurement conditions were as follows: 5 mm distance with a step of 5 µm (resolution). The data were always fitted with a polynomial to remove long-distance tilting and curvature of the sample, and then the average roughness parameter (Ra) was used to assess the roughness of the material, which is defined as the arithmetic mean of the absolute deviations of the roughness profile (roughness) from the mean value (line) over the entire range of the measured length (DIN EN ISO 4287) [54]. The moisture content absorbed by printed parts during their conditioning in distilled water was determined using immersion experiments. The specimens were immersed in distilled water for 16 days at room temperature. At the end of the experiment, the soaked specimen was removed, carefully wiped using wood pulp, and weighed. The content of absorbed moisture (A) was calculated from five measurements using Eq. (1) [61,62]: A=100(wT−w0)/w0(1) where w 0 is the initial weight of the dry specimen, and w t is the weight of the wet specimen. The porosity of printed parts in a dry state was measured with a mercury intrusion porosimeter (MIP) Autopore IV 9500 (Micromeritics, UK). The water contact angle (WCA) of printed parts in a dry state was analyzed using an instrument, “See System E” (Advex Instruments, Brno, Czech Republic). The specimen was placed on a movable stage of a CCD camera while a drop of distilled water (with a volume of 10 μ L) was put onto the surface. Images were recorded after 10 s. At least five measurements at different positions were made. The analysis was performed at room temperature. A universal mechanical device, Instron 3345 (Instron, USA), with a maximum capacity load of 5 kN, was used to test the tensile and bending properties of the printed parts. Both tests were undertaken at a 1 mm⋅min −1 speed with a load of 5 kN. The tensile test was performed according to ˇ CSN EN ISO 527–2 [50], and the bending test was performed according to ˇ CSN EN ISO 178 [51]. The printed parts were tested in a dry and wet state, the latter was achieved by conditioning in distilled water for 16 days at room temperature. The principal components analysis (PCA) between the final properties of 3D printed samples differing in the PA12 powder grade and the effect of print orientation was carried out using the software OriginPRO 2018b (Academic version, b9.5.5.409). 3. Results and discussion 3.1. Powder properties The size and shape distributions of three grades of PA12 powder, namely the virgin powder, the virgin-post-industrial recycled powder mixture (virgin-to-post-industrial recycled powder weight ratio of 25/ 75), and the post-industrial recycled powder, were analyzed to evaluate the effect of thermal load in the SLS process on the post-industrial recycled PA12 powder characteristics. In comparison with the virgin powder, the post-industrial recycled powder was found to contain a slightly lower number of smaller particles with a size range of 5–10 µm, while the representation of the fraction with a particle size of 50–100 µm gently increased (Fig. 2) probably because of particle sintering, which was also observed in the reference [39]. Regarding the shape, the postindustrial recycled powder showed a decrease in the circularity of the particles (Fig. 3), which can also be attributed to the formation of sintered particles with a heterogeneous morphology. The particle size and morphology of the three grades of PA12 powder were also investigated by SEM and are shown in Fig. 4. 3.2. Material properties Table 2 lists the number-average molar mass (M n ), the weightFig. 2. Particle size distribution for different PA12 powder grades. J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 4 average molar mass (M w ), the z-average molar mass (M z ), and the dispersity (M w /M n ) for three grades of PA12 powder. The cumulative molar mass distribution curves are compared in Fig. 5. The results revealed a significant increase in the molar mass of the post-industrial recycled PA12 powder, which is a well-known phenomenon attributed to endchain groups that undergo post-condensation and branching/crosslinking reactions at higher temperatures [32,40]. The thermal behaviour of the PA12 powders was analyzed using DSC, applying the two-step temperature program (see the Experimental part). During the first heating step, a significant change in the shape and position of the melting peak can be observed (Fig. 6). The post-industrial recycled powder exhibited a later initiation of the melting process compared to the virgin powder (extrapolated onset temperature shifted by crystal. 1 ◦C to higher temperature); the maximum of the melting peak increased with the SLS degradation by crystal. 2 ◦C. Similar temperature shifts were observed also in [28,35]. Note that the characteristic melting temperatures of the mixture of the two powders indeed correspond to the expected merger result, i.e., onset temperature similar to the virgin powder and the melting maximum similar to the mass-dominant component – the post-industrial recycled powder. The shift of the melting peak to the slightly higher temperature in the case of the post-industrial recycled powder can be associated with the combination of the following two factors: 1) increased thermal gradients within the post-industrial recycled sample, the composition of which includes a larger amount of the high-sized particle fraction (see Fig. 2), similar behaviour was observed, e.g. in [35]; 2) increased molar mass do to post-condensation and crosslinking reactions proceeding at the elevated temperatures during SLS. In all cases, the melting indicates the presence of the stable γ-form crystalline structure. The melting enthalpy associated with the melting peak measured during the first heating step very slightly increased (from 108.9 to 112 ◦C), resulting in an increase of crystallinity χ c from 52.1 to 53.6 % (a complete list of the characteristic quantities is given in Table 3). This increase may be associated with the secondary crystallization process [35,49], where an additional crystalline phase (lamellar thickening) is formed due to the more efficient packing of the polymer chains during the prolonged annealing at elevated temperatures and contributes to the changes in the melting behaviour. In the second step of the two-step temperature program, the crystallization from the molten state was observed during the cooling of the PA12 samples (Fig. 7). Here, again in agreement with the literature [28,35], the post-industrial recycled PA12 material exhibited a delayed crystal growth, with the bulk crystallization process occurring at a lower temperature. This behaviour is again consistent with the idea of the Fig. 3. Circularity distribution for different PA12 powder grades with illustrative particle morphologies. Fig. 4. SEM images of different PA12 powder grades: A) the virgin powder, B) the virgin-post-industrial recycled powder mixture, C) the post-industrial recycled powder, D) the virgin powder at a lower magnification, E) the virgin-post-industrial recycled powder mixture at a lower magnification, and F) the post-industrial recycled powder at a lower magnification. Table 2 Molar mass averages and the dispersity values of different PA12 powder grades. Powder grade M n / g . mol −1 M w / g . mol −1 M z / g . mol −1 M w / M n Virgin 19,000 30,000 43,000 1.58 Mixture 34,000 79,000 314,000 2.32 Post-industrial recycled 40,000 120,000 305,000 2.55 J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 5 Fig. 5. Cumulative molar mass distribution curves for different PA12 powder grades. Fig. 6. Comparison of 1st (heating) curves of DSC measurements for different PA12 powder grades. Table 3 Characteristic thermo-analytical quantities obtained for different PA12 powder grades using DSC: extrapolated onset temperature of the melting peak T m ons , maximum peak temperature of the melting peak T m , half-height width of the melting peak FWHM m , melting enthalpy ΔH m , extrapolated onset temperature of the crystallization peak T c ons , maximum peak temperature of the crystallization peak T c , half-height width of the crystallization peak FWHM c , crystallization enthalpy ΔH c . Powder grade*T m ons / ◦C T m / ◦C FWHM m / ◦C ΔH m / J⋅g −1 T c ons / ◦C T c / ◦C FWHM c / ◦C ΔH c / J⋅g −1 Virgin 178.2 183.1 3.8 108.9 155.8 152.6 3.9 57.5 Mixture 178.2 184.8 6.0 111.3 155.2 151.9 4.3 57.4 Post-industrial recycled 179.4 185.3 5.3 112.0 154.7 151.2 4.4 57.2 * The errors associated with the determination of the characteristic temperatures and enthalpies are ±0.05 ◦C and ±0.3 J⋅g −1 , respectively. J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 6 PA12 molar mass being increased during SLS [36], which results in a lower tendency toward crystallization and, thus, the need for larger undercooling (the increased driving force of the melt-originating crystal growth). Also, the amount of the crystalline phase formed during the cooling of the molten PA12 very slightly decreased with the SLS-based degradation (based on the values of the crystallization enthalpy ΔH c ), further confirming the overall decrease of the crystallization tendency. DSC was also used to simulate changes in the virgin powder induced by the industrial 3D printing process (see blue dashed line in Fig. 7); an almost identical shift of the crystallization signal was obtained for the virgin PA12 powder annealed 10 times for 10 h (overall 100 h) at 130 ◦C. The characteristic quantities associated with the evaluation of the DSC crystallization peaks are listed in Table 3. PA12 is a polymer with a relatively large temperature interval between the onsets of the melting (during heating) and the crystallization (during cooling) processes. This temperature interval, the so-called sintering window (ΔT), is important information for the SLS parameters setting. The evaluation of ΔT from the DSC curves is shown in Fig. 8; the evaluation is based on the difference between the extrapolated onsets of the crystallization and melting processes. It was found that the post-industrial recycled PA12 powder exhibits a slightly broader sintering window (by ~ 2.3 ◦C) than the virgin PA12 powder, which is beneficial as the polymer remains in a liquid state for a longer time during the cooling process, thus preventing the build-up of residual stresses and printed part distortion [55]. In addition to the standard ΔT evaluation procedure based on the extrapolated onsets, several other methodologies can be applied for the ΔT evaluation. The second most common approach is based on the determination of the first processcaused signal deviation from the DSC baseline (often expressed as a percent of the given DSC peak height). In this case, the increase of the sintering window associated with the SLS degradation would be even larger because the post-industrial recycled PA12 powder did not exhibit the slow, prolonged melting onset observed for the virgin PA12 powder. As such, the post-industrial recycled powder also provided a significantly higher upper limit (given in this case by the initial T m ) for the sintering process, allowing for better adjustment of the sintering temperature. As the thermal stability of PA12 powder may also change due to thermal load during the SLS process, the PA12 powders were analyzed by TGA in the nitrogen atmosphere (Fig. 9). (The nitrogen atmosphere, present during the SLS process, is significant for avoiding thermooxidative degradation and keeping printed materials stable for longterm applications.) The onset decomposition temperature of the postindustrial recycled PA12 powder was found to be about 5 ◦C lower than that of the virgin powder (specifically 418.8 and 423.2 ◦C for the post-industrial recycled and the virgin powder, respectively). This phenomenon was also observed in the relevant literature [56], indicating macrostructure chain changes (e.g., post-condensation induced cross-linking or chain scission) caused by thermal ageing during SLS, accelerating the decomposition of the long-chain PA12 backbone of polyamide 12 during TGA. All grades of PA12 powder were subjected to MFI measurements to provide information on the rheological properties in the melt state. This information is crucial not only from a technological but also from a practical point of view, as high melt viscosity, represented by low MFR and MVR values, reduces the sintering ability of polymer particles, thus leading to porosity and, in turn, a decrease in mechanical strength of printed parts [57]. It was shown that the virgin powder exhibited significantly higher MFR and MVR compared to the other powder grades. In contrast, the lowest MFR and MVR values were determined for the post-industrial recycled powder (Table 4). These results correlate with the experiments described in the literature [42,47], which illustrated a decrease in the MFI of PA12 because of thermal load in the SLS process. This phenomenon is attributed to the thermally induced molar mass increase of polymer chains, which increases the melt viscosity of PA12 polymers [58]. Furthermore, the chemical structure of the printed materials made from different PA12 powder grades was characterized using FT-IR spectroscopy (Fig. 10). By interpreting the FT-IR spectra, one can see an absorption band at around 3294 cm −1, which is characteristic of the N−H stretching vibration, together with a peak at 3083 cm −1 which belongs to the Fermi resonance of ν (N−H) stretching. Two intense peaks Fig. 7. Comparison of 2nd (cooling) curves of DSC measurements for different PA12 powder grades including the experimentally obtained recycled powder (i.e., the virgin powder subjected to DCS simulated annealing for overall duration of 100 h at 130 ◦C). J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 7 at 2913 and 2844 cm −1 correspond to asymmetric and symmetric stretching vibrations of CH 2 . Amide characteristic bands are detected at 1635, 1538, 1268, 713, and 621 cm −1 . Amide bonds are sensitive to thermal ageing and can be used as a benchmark for determining chemical changes during SLS 3D printing. The intensities and positions of these bands are almost the same, so we can conclude that the SLS process did not affect the destruction of amide bonds. The polyamides are prone to post-condensation during thermal ageing [32,40]. Thus, we can focus on the changes of the bands at 1159 and 1061 cm −1 that are characteristic of the skeletal motion involving the CO−NH groups and the peak at 946 cm −1 corresponding to the CO−NH deformation [59–61]. From the spectra recordings, an increase in these band intensities can be seen in the post-industrial recycled PA12 powder and the virgin-post-industrial recycled powder mixture, which indicates a significant occurrence of the post-condensation reaction, being responsible for the observed molar mass and melt viscosity increase (in line with the literature [62]). 3.3. Properties of printed parts Products made from polyamides are repeatedly exposed to environments with variable humidity or moisture during their lifetime. Due to their hygroscopic nature, these polymers can absorb moisture and change their properties, sometimes irreversibly. Many studies have already shown a strong influence of absorbed water on the chemical structure, morphology, and physical properties of various polymers [63]. Therefore, the effect of moisture on the properties of printed parts was investigated in more detail. By observing the surface morphology of dry printed parts built in the A1 print orientation, no pronounced effect of PA12 powder grade was detected among the samples (Fig. 11). On the contrary, significant differences among the printed parts were found after their exposure in water. In this case, the wet sample built from the virgin powder showed a more compact surface than the samples prepared from the post-industrial recycled powder or the virgin-postindustrial recycled powder mixture (Fig. 12). The surface morphology of the latter samples exhibited a rough topography, where initial polymer particles became visible, probably because of a higher level of moisture absorption, larger particle size fluctuations, indicating the poorer quality of sintering. The level of moisture absorption by a polymer may depend not only on its molecular or supramolecular character but also, to a large extent, on its surface roughness and material porosity. Therefore, these characteristics for the dry printed parts were determined and compared from the point of view of PA12 powder grade and print orientation (Table 5). Regarding the powder grade, the printed parts made from the postindustrial recycled powder and the virgin-post-industrial recycled powder mixture exhibited slightly increased surface roughness and porosity in contrast to the samples made from the virgin powder. This phenomenon indicates a deterioration in the sintering quality, probably due to thermally induced changes in the post-industrial recycled PA12 powder, which has already been discussed in more detail in the text above. Regarding the print orientation, the porosity and surface roughness values of samples built in A1 and B1 print orientations did not vary significantly, whereas the G1 print orientation provided samples with increased surface roughness and porosity, suggesting the worst sintering quality. When comparing the surface roughness of corresponding printed parts in a dry and wet state, the results showed enhanced surface roughness for the wet samples, which is in agreement with SEM and ESEM results (Figs. 10 and 11, respectively). The material hygroscopicity was evaluated according to the content of absorbed moisture and WCA (Table 5). It was confirmed that a lower level of hygroscopicity, expressed by a decreased content of absorbed moisture and an increased WCA, was generally determined for the printed parts, which were simultaneously characterized by a lower surface roughness and, in particular, a lower porosity. Specifically, the samples from virgin powder in the A1 and B1 print orientations appeared the least hygroscopic. On the contrary, the samples made from the post-industrial recycled and the mixture of powder grades in the G1 print orientation exhibited a more pronounced hygroscopic nature, probably due to a porous structure enabling easier water transport. The mutual effects of PA12 powder grade, print orientation, and absorbed moisture on the mechanical properties of the printed parts both in a dry and wet state (before and after 16 days-long water conditioning, respectively) were tested using tensile and bending tests (Tables 6 and 7). Regarding the powder grade, the use of post-industrial recycled powder, solely or in the virgin-post-industrial recycled powder mixture, decreased the mechanical strength of the printed parts. This phenomenon is probably a consequence of the increased molar mass of post-industrial recycled PA12 powder due to thermal load, as has been Fig. 8. DSC curves used to determine the sintering window (ΔT): A) the virgin powder, B) the virgin-post-industrial recycled powder mixture, and C) the postindustrial recycled powder. J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 8 demonstrated above. As a result, a higher melt viscosity and poorer sintering occurred, which provided more residual voids acting as local stress concentrators in the printed parts [64,65]. Regarding the print orientation, the mechanical properties of samples built in the A1 and B1 print orientations were found not to differ significantly, whereas the G1 print orientation provided samples with a lowered tensile strength, while the bending strength was not influenced distinctly. On the contrary, the printed parts built in the G1 print orientation exhibited higher Young and bending moduli in comparison with the samples built in the A1 and B1 print orientations, which is the opposite result to that obtained by Sanders et al. [49], where samples vertically oriented had lower Young modulus than the horizontally oriented samples. Nevertheless, the observed behaviour is not unusual for printed parts built layer-by-layer using the SLS technology, where anisotropic microstructure and anisotropic mechanical properties are a natural consequence of the building process [66]. Furthermore, the influence of the absorbed moisture on the mechanical strength of the printed parts was investigated, as water molecules absorbed into a polymer material are known to act as an effective plasticizer [67–69] An illustrative comparison of samples from the point of view of the mechanical strength decrease caused by absorbed moisture is presented in Fig. 13. It was found that the samples made from the virgin powder were the most moisture sensitive, exhibiting the highest decrease in tensile strength and especially in bending strength when comparing corresponding wet and dry printed parts, whereas the mechanical strength of the samples made from the post-industrial recycled powder, or the virgin-post-industrial recycled powder mixture achieved a similar rate of decline. This behaviour can be attributed to the degree of crystallinity of the PA12 powder, as the plasticizing water can only be present in the amorphous phase. Therefore, the printed parts made from virgin powder with the lowest crystallinity were mostly plasticized with water. At this point, it is worth mentioning that the content of plasticizing (bound) water may not follow the total amount of absorbed water. This phenomenon is especially true for porous materials with a high proportion of non-plasticizing (bulk) water trapped in the pores [70]. Statistical evaluation of the mechanical properties and the print orientation is presented in Figs. S1 −S3 in Supplementary Material. 4. Conclusion This study deals with PA12 powder recycling during SLS 3D printing and aims to correlate PA12 powder ageing with changes in the Fig. 9. TGA curves for different PA12 powder grades. Table 4 Mass flow rate (MFR) and melt volume flow rate (MVR) for different PA12 powder grades measured using a capillary plastometer. Powder grade MFR / g/10 min MVR / cm 3 /10 min Virgin 55.7 ±2.6 63.3 ±2.2 Mixture 18.9 ±0.9 20.9 ±0.4 Post-industrial recycled 12.1 ±1.8 15.0 ±0.2 Fig. 10. FT-IR spectra of the printed materials made from different PA12 powder grades. J. Machotov´ a et al. European Polymer Journal 220 (2024) 113432 9