Influence of manufacturing process in pores, f ibredistribution and mechanical properties of ASA-CFfor 3Dlargeformatadditive manufacturing pieces

Influence of manufacturing process in pores, fibre distribution and mechanical properties of ASA-CF for 3D large format additive manufacturing pieces Daniel Moreno Sánchez Departamento de Ingeniería Mecánica y Diseño Industrial, Escuela Superior de Ingeniería, Universidad de Cádiz, Campus Río San Pedro, C adiz, Spain Jesús Hern andez Saz Departamento de Ingeniería y Ciencia de los Materiales y del Transporte, Universidad de Sevilla, Sevilla, Spain, and Jos e Javier Relinque Madroñal, Pedro Burgos Pintos, Sergio I. Molina Rubio and Miriam Herrera Collado Departamento de Ciencia de los Materiales e Ingeniería Metalúrgica y Química Orgánica, IMEYMAT, Universidad de Cádiz, Campus Río San Pedro, C adiz, Spain Abstract Purpose –The purpose of this study was to evaluate parts manufactured by fused granular fabrication (FGF) using two different acrylonitrile styrene acrylate (ASA) compounds reinforced with carbon fiber (CF) and compared them with those produced by injection molding (IM). Design/methodology/approach –The difference in both ASA_CF composites is the mixing process, achieved using twin-screw extruders with different mixing shear rates. The porosity proportion and distribution, and the CF length and alignment, have been carefully quantified and compared using X-ray computerised tomography (CT). Also, the correlation with mechanical properties has been evaluated for FGF and for IM parts in both materials. Findings –Reinforcement of polymeric matrices with different fibres is a common strategy to improve the mechanical properties of 3D printing pieces, especially relevant in FGF. The manufacturing process of these composites needs to be carefully designed to optimise the structural and mechanical characteristics of the material. CT analyses show that the use of higher shear mixing rates increases the amount of pores within the polymeric pellets and produces a smaller fibre size distribution in the composites, which negatively affects the mechanical performance of the final part. Originality/value –To the best of the authors’knowledge, this is the first time that a global study of the complete fabrication process has been carried out, providing detailed information of every step of the process. Keywords Design, Mechanical characterisation Paper type Technical paper 1. Introduction Additive manufacturing (AM) comprises a set of technologies that allow objects to be materialised layer by layer directly from a digital file (Gardan, 2016;Ngo et al.,2018;Tofail et al.,2018). In recent years, these technologies are gaining particular attention because of the advantages they offer compared to traditional manufacturing processes (i.e. subtractive and conformative), such as freedom of design, part customisation or inventory reduction, to name a few (Attaran, 2017). Amongst AM technologies defined by ASTM (ASTMF42.91 Subcommittee, 2015), those classified in the material extrusion (ME) category are the most widespread because of their low cost, scalability, availability of materials and ease of use. In ME, the material is forced to pass through a nozzle, and the layers and parts are created following .gcode paths. For polymers, the main ME technologies lie between fused filament The current issue and full text archive of this journal is available on Emerald Insight at: https://www.emerald.com/insight/1355-2546.htm Rapid Prototyping Journal 31/11 (2025) 218–230 Emerald Publishing Limited [ISSN 1355-2546] [DOI 10.1108/RPJ-03-2024-0103] © Daniel Moreno Sánchez, Jesús Hern andez Saz, Jos e Javier Relinque Madroñal, Pedro Burgos Pintos, Sergio I. Molina Rubio and Miriam Herrera Collado. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) license. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this license may be seen at http://creativecommons.org/licences/by/4.0/legalcode Received 31 March 2024 Revised 9 July 2024 15 September 2024 Accepted 25 September 2024 218 fabrication (FFF), commonly known as fused deposition modelling (FDM), and fused granular fabrication (FGF). The most prominent differences between these two systems are the format of the raw material and the size of the printers. FFF systems are fed with a polymer filament that is melted at the tip of the nozzle and extruded through it, whereas FGF is fed with pellets that are melted inside a single-screw extruder. Also, FFF printers are usually intended for small-format fabrication (less than 1 m 3 ), while FGF printers are generally conceived to produce parts larger than 20 m 3 . In the industrial sector, FGF is very promising, as it allows creating large parts in a short time with a continuous flow of material. For example, these technologies have been successfully used to create a boat toilet (Moreno Nieto et al., 2018), cars (Curran et al.,2016) and moulds for boats (Post et al.,2019) or wind turbine blades (Post et al.,2017). Because of their capabilities in large format, these technologies are often known as large format additive manufacturing (LFAM), big area additive manufacturing or large-scale additive manufacturing. As in other manufacturing processes, the materials used in FGF are required to fulfilspecific requirements. Among others, they need to have a low coefficient of thermal expansion (CTE) and a proper rheology (a Melting Flow Index, between 10 and 20 g/10 min) (Moreno Nieto et al.,2018). FGF printers usually work with no heated chamber; thus to obtain suitable parts, the incorporation of fibres is a common practice to reduce CTE, thus preventing distortions during printing and after cooling. Carbon fibre (CF) and glass fibre appear in the literature to be the more used fibres for the described purpose (Billah et al., 2020; Economidou and Karalekas, 2016;Love et al., 2014;S anchez et al., 2020a). Moreover, the addition of fibres increases the strength of the material and stiffness; however, it also turns the material from ductile to fragile and induces the formation of porosity (Vaxman et al.,1989). These last effects may result in a premature breakage of the parts. Additionally, it has been reported that the presence of fibres hinders the adhesion between layers in printed parts; therefore, a compromise between performance and the amount of fibres needs to be achieved (S anchez et al., 2020a). The quality of the raw material in terms of inhomogeneous integration of fillers or inner defects (i.e. porosity or debonding between filler and matrix) also might affect the final properties of printed pieces. Regarding materials, acrylonitrile butadiene styrene (ABS) (Billah et al., 2020;Hassen et al., 2016; Hill et al.,2018;Rodríguez et al.,2003), polyphenylene sulphide (Liu et al.,2018), poly ether ketone (Tseng et al., 2018)orpolyetherimide(Ajinjeru et al.,2017) reinforced with CF have been most used in FGF. Acrylonitrile styrene acrylate (ASA) is an amorphous terpolymer, which combines excellent resistance to ultraviolet (UV) rays, good mechanical behaviour, good resistance to chemicals and low CTE (John Scheirs, 2003;Oepen and Gottschalk, 2011;Xiang et al.,2017). The chemical structure of ASA consists of alkyl acrylate particles (elastic component), made compatible by a styrene, co-acrylonitrile coating, homogeneously dispersed in a polystyrene co-acrylonitrile matrix (rigid component). It is a polymer that does not present double carbon bonds in its molecular structure, which is the reason why it presents such stable behaviour against degradation under UV rays (John Scheirs, 2003;Moghbeli and Tolue, 2011;Tolue et al.,2009;Yu et al.,1974). It is a polymer little studied in literature, probably for its similitude with ABS, that was recently introduced in FGF, principally reinforced with glass or carbon fibres to create exterior parts (S anchez et al., 2020a). The improvement in mechanical properties and reduction in CTE by fibres is strongly affected by the adhesion fibre-matrix, length of fibres, porosity, and fabrication process, among other aspects (Li et al.,2022;Spoerk et al., 2018;Tekinalp et al.,2014), which determines a threshold in the effective wt.% of fibre added to polymer matrices. Upon previous investigations by the authors using ASA_CF it was concluded that 20–30 Wt.% of CF addition was compatible with mechanical properties enhancement (S anchez et al., 2020b). In that work, ASA_20CF composites were studied because higher CF rates did not significantly improve the mechanical performance while increasing the probability of nozzle clogging in FGF printing because of the agglomeration of fibre. Other studies in the literature confirmed that the aforementioned range of CF wt.% is the most suited to improve mechanical properties (Duty et al., 2017;Love et al.,2015;Solutions, 2015;Yasa and Ersoy, 2018). On the other hand, X-ray computed tomography (CT) is gaining attention in the analysis of polymer composites as it provides detailed information on the phase distribution of the material without altering the structure. This nondestructive technique can reconstruct in 3D the inner structure of materials, procuring the necessary correlation between the microstructure of the printed parts and their properties. In this sense, X-ray CT has been used to analyse the distribution of fibresandporosityinAMinprevious works; for example, Savandaiah et al. (2021) studied the distribution of pores and fibres in polypropylene (PP) in ME; Yang et al. (2021) studied the distribution of the fibres through the nozzle; or Keleset al. (2018) evaluated the distribution of fibreandthepresenceporosityinABS, among other relevant references. In this work, the structural characteristics of ASA and ASA_CF composites were quantitatively analysed by X-ray CT and correlated to the mechanical properties of 3D-printed final pieces. In particular, the distribution and length of fibres and the porosity during different stages of the manufacturing process are studied: from the production of pellets obtained by two different extrusion lines to the subsequent fabrication of parts using both FGFandIM.IMisoneofthemostwidespreadprocessesforthe industrial fabrication of thermoplastics (Kazmer et al.,2023). Because AM techniques are claimed to fulfil the current environmental awareness of the conventional manufacturing processes, the output assessment of alternative manufacturing techniques is usually driven through the comparison with a conventional technique such as IM, as confirmed by the literature (Bezzina and Refalo, 2023;Curmi and Rochman, 2024;de Toro et al.,2020). This comparison highlights pros and cons among conventional and alternative manufacturing methods, finally concluding the feasibility of the use of the latter. To the best of the authors’knowledge,thisisthefirsttimethata global study of the complete fabrication process has been carried out, providing detailed information of every step of the process. The results derived from this work highlight the significant effect of 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 219 the manufacturing process on the characteristics and performance of 3D final parts. 2. Materials and methods Two different ASA_CF composites (23Wt.% CF), as well as neat ASA used as a reference, were tested in this research. All materials were purchased at Matersia S.L. (Spain). ASA composites differ in the compounding conditions: the first one will be referred to as low shear mixing (LS), with a 1.2m long, 40°C water bath cooling system; and the second one as high shear mixing (HS) with a 4 m long water bath at 15°C. Prior to any use (i.e. IM or FGF), raw materials and composites were dried at 80°C for at least 4 h in a PIOVAN DPA30 compressed air dehumidifier. Figure 1 summarises the experiments carried out in this communication, and Table 1 gathers the variables studied in this research: materials (ASA and ASA_CF); compounding conditions (LS and HS); and manufacturing processes of the samples (IM and FGF). LFAM pieces were fabricated in a single-screw FGF printer, Super Discovery 3D Printer Compact. A temperature profile of 250°C–260°C (from top to nozzle) and a heating bed of 100°C were used. Pieces intended for structural characterisation by X-ray CT scanning and for mechanical characterisation by means of tensile testing were prepared. The fabrication of parts followed a two-stage procedure. In the first stage, two-layered rectangles with an A4 dimension were deposited onto the printer platform. The infill was 100%, and a 2-mm nozzle was used with a thickness layer of 1 mm. In the second stage, tensile test and prims specimens were cut from those printed sheets by using a desktop CNC machine. Samples for structural characterisation were rectangular prisms with dimensions 5 mm 5mm2 mm, whereas tensile specimens were fabricated according to the UNE-EN-ISO 527 standard (type 1BA specimens). The tensile specimens were cut in X orientation which means that the deposition of the beads was in parallel to the direction of load application in the tensile testing. Additionally, tensile test specimens produced by IM were prepared using a BabyPlast 10/12 injector at 250°C–260°C (plasticiser, chamber and nozzle) with 60 and 80 bar in injection and compaction pressure, respectively. X-ray CT samples were also cut from injected parts for a proper comparison. X-ray CT scanning was carried out using a Zeiss Xradia 610 Versa X-ray microscope. At least two scans were recorded for most of the samples. Initially, a large field of view scan was performed to obtain information about the general features of the material, with an optical resolution of 0.39. Then, a higher resolution scan was performed to analyse in detail the microstructural features of each material (optical resolution 3.97). The working accelerating voltage was 60kV. A total of 1,000–2,000 2D projections were recorded as the sample was rotated by 360° and computationally reconstructed via a filtered back projection algorithm (Zeiss XM Reconstructor). To visualise the internal features inside the sample in 3D space, software DragonFly was used. Tomograms were imported into Figure 1 Summary of the experiments performed in this research 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 220 the Avizo software suite to analyse the CF orientation and length and pore distribution of the pieces. For this, features in the composites were segmented using a greyscale threshold applied to the absorption contrast data. Small isolated regions in the segmentation arising from noise in the scan data were removed. The tensile properties of the different composites were determined using a Shimadzu AGS-X universal testing machine, attached with a 10 KN load cell and screw action grips, shown in Figure 1(h). The tests were performed at a constant speed of 1 mm/min, with data acquisition every 0.1s, using Trapezium X software. The grips were properly tightened manually to prevent the specimens from slipping during the test. The results were expressed as an average of 5 experiments according to UNE EN ISO 527. 3. Results and discussion 3.1 Pore analysis Initially, pellets from both compounding lines and neat ASA were analysed before further fabrication using IM or FGF. Figure 2(a) shows a rendered volume of ASA_CF_LS_Pellet, and Figure 2 (d) includes a magnified part of a 2D slice from the 3D greyscale data. As it can be observed, a homogeneous distribution of CF can be observed in the material (white contrasts), some of them marked with red arrows in Figure 2(d). Interestingly, CF appear mainly as circular contrasts, indicating that the fibres are observed in cross section. Thus, CF were mostly aligned parallel to the extrusion direction, as should be expected (Fallon et al., 2019;S anchez et al., 2020a;Tekinalp et al., 2014). Attention will be paid to the CF orientation and length further on. Additionally, it is worth noting that some pores are found in the material, located preferentially at the central part of the pellet. The pore volume percentage calculated by taking the ratio of the total segmented pore volume to the total investigated sample volume has been quantified as 4.9%. In the case of ASA_CF_HS_Pellet (high shear mixing conditions), Figure 2(b) shows a reconstruction of the CT data of the pellet, and Figure 2(e) includes a 2D slice corresponding to this set of data. The pore distribution is different in this composite. In spite of some degree of porosity still being perceptible at the central part of the pellet, it is not as conspicuous as in the low shear processed pellet; additionally, the size of the pores is remarkably smaller (see a comparative histogram of the size distribution in supplementary information Figures S1 and S2). In addition, pores are found up to the edge of the pellet, whereas in the low shear processed composite, almost no pores were found in this region. The amount of pores in that material has been quantified as 10.1% in volume, which is approximately double that in the precedent case. To understand whether this porosity is related to the introduction of the CF in the polymer, a reference sample of neat ASA was also analysed by CT, and a 3D reconstruction of the data is shown in Figure 2(c) and (f) . As it can be observed, no pores are present in the ASA_Pellet, evidencing that the Table 1 Classification of the materials studied, identifying their compounding conditions and manufacturing process Denomination Compounding condition Manufacturing process ASA_Pellet –– ASA_IM –Injection moulding ASA_FGF –Fused granular fabrication ASA_CF_LS_Pellet Low shear mixing – ASA_CF_LS_IM Low shear mixing Injection moulding ASA_CF_LS_FGF Low shear mixing Fused granular fabrication ASA_CF_HS_Pellet High shear mixing – ASA_CF_HS_IM High shear mixing Injection moulding ASA_CF_HS_FGF High shear mixing Fused granular fabrication Source(s): Table by the authors Figure 2 Rendered volumes of ASA_CF_LS_pellet (a), ASA_CF_HS_pellet (b) and neat ASA_pellet (c); 2D slices of the ASA_CF_LS_pellet (d), ASA_CF_HS_pellet (e) and neat ASA_pellet (f) 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 221 porosity stems from the introduction of the CF in the polymer. The presence of porosity in polymers reinforced with fibres has been ascribed to a double effect: air entrapment during mixing and shrink-void by an uneven contraction during cooldown (Tronvoll et al.,2018;Vaxman et al.,1989). In the present case, pores are not located around the CF, as it is usual in fibrecontaining composites because of different thermal expansion of the phases involved (Kumar et al.,2021;Vaxman et al., 1989). Instead, at the edge of the pellet, CF without pores can be clearly observed in ASA_CF_LS_Pellet. The absence of pores around the fibres is associated with a proper integration of the fibres inside the polymers (S anchez et al., 2020a;Song et al.,2016). The amount of pores might not only be related to entrapped air inside the material during mixing. In ASA_CF_LS_Pellet, the presence of pores in the middle of the pellets could also be attributed to a slow cooling down of the material, producing the last contraction of the polymer in the centre of the pellet (Rauwendaal, 2010). In ASA_CF_HS_Pellet, the distribution of small pores in the material might be because of the faster cooling down of the material (4m long in 15°C water in HS, instead of 1.2m long in a 40°C water bath for LS configuration). A higher cooling rate sets the geometry of the extruded material quickly. After die swell effect (expansion of material by the pressure difference between the interior of the extruder and atmospheric pressure), the polymer shrinks almost homogeneously in a nonpreferential area and the microporosity generated appears in this random distribution. Also, other aspects, such as a longer residence time inside the extruder line and the presence of fibres, might affect the nucleation of this microporosity in this second configuration. Given that such porosity appears not to be around the fibres and the fact of observing smaller pores homogeneously distributed in ASA_CF_HS_Pellet than in ASA_CF_LS_Pellet, these observations necessarily might be linked to the severity of the ASA/CF mixing process. As a matter of fact, the extrusion line used to produce ASA_CF_HS_Pellet includes more mixing elements and procures a higher residence time than the extrusion line used to produce ASA_CF_LS_Pellet; therefore, a higher shear rate can be expected to be provided by the first extrusion line. This shear rate should be responsible for the porosity size distribution and their location. Because the shear rate of the HS line is higher than in the LS line, pores are not allowed to coalesce in the inner part of the pellet, as observed in the ASA_CF_LS composite. Conversely, a higher shear rate allows a better distribution of the pores through all the volume, thus reducing its average size. Considering the round shape of the observed pores, they should be ascribed either to air bubbles or remnant moisture as well as to polymer matrix contraction because of the heat dispersion enhanced by the CF. Moisture is minimised upon dry pretreatment of ASA and CF; thus, the contractions should be the dominant phenomenon to explain pore formation. This hypothesis is in good agreement with the absence of pores in the pure ASA, where the lower heat dissipation, not favoured by the presence of CF, would not provoke the emergence of pores. In any case, for a higher total volume of pores, microfractures will be more likely to grow and nucleate by porosity, and, coherently, lower mechanical strength will be expectable. The comparison of 3D parts fabricated by FGF using both ASA_CF compounds and neat ASA are exhibited in Figure 3. Concretely, in Figure 3(a) and (d),3DX-rayCTdata reconstructions of ASA_CF_LS_FGF and ASA_CF_HS_FGF are shown, respectively (obtained from the region marked with a red square in the 2D images in Supporting Information, Figure S3), and Figure 3(g) shows the data corresponding to ASA_FGF. A uniform distribution of the roads, characteristic of the FGF process, can be observed in the three objects, with an approximate widthof2mmandheightof1mm,inagreementwiththe parameters used in the printing process (nozzle size and layer thickness). Figure 3(c) and (f), shows details of 2D slices of the 3D greyscale CT data of ASA_CF_LS_FGF and ASA_CF_HS_FGF, respectively. Besides the CF that can be observed as bright contrasts (that will be analysed in more detail in section Fibre Analysis), some porosity can be found in both materials, which is absent in the pure ASA piece, Figure 3(h). Observations in the middle of the beads show that the very large pores found in the central parts of the ASA_CF_LS_Pellet [Figure 2(a)] are not present in the printed object. However, it is noticeable there is a larger density of pores at the middle of the bead than at the edge, as expected. Subregions in the centre of the roads [marked with red squares in Figure 3(a) and (d)] have been evaluated by a higher resolution X-ray CT analyses. The segmentation of the pores included in Figure 3(b) and (e), clearly shows that the density of pores is larger in the ASA_CF_HS_FGF piece. The quantitative analysis of the observed porosity showed that the pore volume percent in the ASA_CF_LS_FGF specimen was 3.4%, while in the ASA_CF_HS_FGF specimen, it reached 8.7%, evidencing that the amount of pores in the middle of the roads is more than twice in the high shear compounded material. Regarding the size of the pores, they were slightly larger in the ASA_CF_LS_FGF object than in ASA_CF_HS_FGF, as can be seen in the histogram of pore size in Figure S2. In FGF printed objects, the tendency found in the amount of pores is similar to that in the pellet study, indicating that in spite of the printing process melting the material again, the printer cannot eliminate the residual porosity inside the original feedstock. This evaluation highlights that even considering the additional macroporosity intrinsically generated between beads during printing, porosity from feedstock remains after printing, which might negatively affect further properties of the printed pieces. In any case, for a higher total volume of pores, microfractures will be more likely to grow and nucleate, and, coherently, lower mechanical strength will be expectable because, as reported in the literature, porosity acts as a stress concentrator (Du Plessis et al.,2018;Wang et al., 2019). Regarding the pore shape, in the literature, elongated pores are commonly observed in composites containing fibres (Keles et al.,2018;Kumar et al.,2021). In this case, elongated porosity is partially related to the deformation of the molten material during deposition (Keleset al.,2018) or likely because of the affinity of the entrapped gas for the fibres that also indicates a poorer integration of the fibres in the matrix (Kumar et al., 2021). In relation to this, the segmentation of the pores found in the ASA_CF_LS_FGF and ASA_CF_HS_FGF pieces included in Figure 3(c) and (f), shows that the pores do not present an evident elongation. To analyse the pores shape, the sphericity of the pores (which is a measure of how closely the shape of an object resembles that of a perfect sphere) has been calculated, and a histogram with the obtained results is shown in Figure 3(h). The sphericity, being 1 that of a sphere 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 222 by definition, for any particle which is not a sphere will be less than 1. Sphericity ( c ) can be calculated with the following formula (1), where V means volume of particle and A means surface area of the particle. W¼ p 1=36V ðÞ 2=3 A(1) The observed sphericity for both composites presents distributions centred about 0.9, i.e. the majority of the pores can be approached as spheres. The distribution corresponding to ASA_CF_LS_FGF seems to exhibit an unexpected peak around sphericity 0.5. This could be related to the number of pores: because the number of pores is lower in the latter composite than in the corresponding ASA_CF_HS_FGF, pores exhibiting a lower sphericity should be statistically more significant. In any case, the frequency is quite lower than those of sphericity 0.9; hence, both distributions could be reasonably assumed to be mostly unimodal. This means that they are not likely because of the affinity of the entrapped gas for the fibre or decohesion between fibres and polymer. Discarding so, the spherical pore distribution found can be related mostly to the shrinkage effect of the polymer in the composite material (Rauwendaal, 2010). For comparative purposes, parts were also fabricated by IM and analysed using X-ray CT. Figure 4 shows rendered volumes of parts obtained by injection moulding using the two compounding systems considered, ASA_CF_LS_IM in Figure 4(a) and ASA_CF_HS_IM in Figure 4(c). Data corresponding to a neat ASA-injected specimen is also included [Figure 4(e)]. In contrast to previous observations related to pellets and FGF specimens, no pores have been found in any of the injected solid pieces. Figure 4(b) and (d), shows 2D slices of the CT data corresponding to the ASA_CF_LS_IM and ASA_CF_HS_IM pieces, where the absence of pores is evident. These results clearly determine that injection of pellets eliminates its inherent porosity, likely because of the higher pressure involved in the process. 3.2 Fibre analysis Orientation and length of CFs have a strong influence on the mechanical properties of composites. Because of this, a detailed analysis of these parameters has been carried out in the pellets, FGF and IM pieces, using high-resolution X-ray CT analysis. Concerning the orientation analysis, CF segments are presented in a spherical coordinate system, which allows for quantification of the offset angle ( u ) from the longitudinal direction (Billah et al., 2020). Figure 5 shows the results of the analysis of the CF orientation distribution for the ASA_CF_LS_FGF (a) and ASA_CF_HS_FGF (b) pieces in the region of the red square in Figure 3(a) and (d), where CF are represented colour-coded regarding their orientation. As can be observed, in both pieces, CF is mostly oriented parallel to the extrusion direction. This is corroborated by the CF orientation histogram shown in Figure 5 (c), which shows a reasonably narrow orientation distribution for both pieces. This preference orientation is expected in materials fabricated by extrusion processes, as noted (Kumar et al.,2021; Spoerk et al., 2018;Tekinalp et al.,2014). In this case, that prealignment implies a positive effect in terms of an increase in stiffness and strength in the direction of the deposited material (Love et al.,2015;S anchez et al., 2020a). However, it also Figure 3 Rendered volumes of the ASA_CF_LS_FGF (a), ASA_CF_HS_FGF (d) and ASA_FGF (g) parts; pieces and pore segmentation images of the ASA_CF_LS_FGF (b) and ASA_CF_HS_FGF (e) from the region marked with a red square in (a) and (d); magnified regions of 2D slices of the ASA_CF_LS_FGF (b), ASA_CF_HS_FGF (e) and ASA_FGF (h), respectively, from the region marked with a black square; (i) histogram of the pore sphericity of the ASA_CF_LS_FGF and ASA_CF_HS_FGF pieces 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 223 increases the anisotropy of the object not only in mechanical terms (Hmeidat et al.,2020;Spoerk et al., 2018;Torrado et al., 2015) but also in CTE behaviour (Hoskins et al., 2019). A detailed analysis of the CF length has been carried out using the high-resolution CT data obtained from the FGF and IM pieces. Figure 5(d) and (e), shows the results of the analysis of CF length in ASA_CF_LS_FGF and ASA_CF_HS_FGF pieces, respectively, represented with a colour scale. As it can be observed, a narrow distribution of CF length values is observed, with some larger fibres observed in cyan colour in the ASA_CF_LS_FGF piece. To obtain quantitative information on the CF length distribution in both materials, Figure 5(f) shows a comparative histogram of the measured CF length. Results show that CFs have a length inferior to 250 m minboth materials. These values obtained are quite similar to those found in the pellets histogram of CF length in the pellets, Figure 5(g), showing that compounding implies fracture and a remarkable reduction in length of CFs (originally the length of the CF was 6 mm, according to the distributor). Breakage of fibres during composite fabrication occurs because of the shear force and screw configuration during mixing. This is of major importance, as it is well known that longer fibres offer higher reinforcement than shorter ones in the compounds (Adeniran et al., 2022;Yeole et al.,2020). In our case, the fibre length obtained after compounding seems to be above the critical value, as reinforcement is demonstrated in the mechanical properties measured, as shown in the next section. On the other hand, the fact that the fibre length is almost the same in printed parts as it is in pellets highlights that the configuration of the single screw of the printer does not produce a significant reduction in fibre length, evidencing that the printing process used does not degrade the properties of the feedstock material used. A closer look at Figure 5(f) shows that, on average, CF has a larger length in the ASA_CF_LS_FGF piece than in the ASA_CF_HS_FGF one. This indicates that the LS extrusion line is less aggressive than the HS extrusion line, as expected. The reduction in mixing elements, together with a lower residence time inside the extruder, produces a breaking level of the fibres which is lower in line LS than in HS. As a homogeneous distribution of fibres inside the polymer has been observed in both cases, the level of reinforcement might be better in a configuration with low shear mixing than in high shear mixing, as the longest fibres in the material ASA_CF_LS are the ones with a major effect in the reinforcement mechanism. This will be further discussed below in relation to the mechanical properties measured in both materials. In connection with this, it should be mentioned that the amount of porosity observed in the composites might also be related to the length of fibres, as the larger the number of fibre breakage events in the material, the larger the number of porosity nucleation sites. The larger number of pores and the shorter average CF length found in ASA_CF_HS_FGF regarding ASA_CF_LS_FGF would be in line with this possibility. Also, a Figure 4 Rendered volumes of the ASA_CF_LS_IM (a), ASA_CF_HS_IM (c) and ASA_IM (e) pieces; magnified regions of the 2D slices of the ASA_CF_LS_IM (b), ASA_CF_HS_IM (d) and ASA_IM (f) 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 224 lower ‘fibre-free’volume in the polymer reduces the possibility of pore coalescence, avoiding the increase in size of the porosity. Regarding the IM specimens, an analogous analysis of CF orientation and length has been carried out in high-resolution CT data (obtained from the region in the 2D images in Supporting Information, Figure S4) to understand the microstructure of the materials to be correlated to their mechanical properties. Figure 6 depicts a representation of the CF orientation in the ASA_CF_LS_IM (a) and ASA_CF_HS_IM (b) objects, and Figure 6(c) shows a histogram of the orientation data in both pieces. As illustrated by the results, in both materials, most of the CF are oriented parallel to the injection direction, similarly to what is observed in FGF printed specimens. Nevertheless, in this case the injected specimens exhibit a central region where some disorientation occurs, as can be observed in the green colour in Figure 6(a) and (b), . It is well known that during the injection process, the material in contact with the mould walls experiences a sharp reduction in material velocity because of an increase in shear rate. It is this material in contact with the mould that solidifies first (referred to as the frozen layer), and where the reduction in velocity promotes a preferential alignment of the fibres in the direction of flow by being in a laminar regime. The molten material continues to flow through the centre of the cavity until it fills the mould completely. This melt has a higher Figure 5 Colour-coded representation of the CF orientation in the ASA_CF_LS_FGF (a) and ASA_CF_HS_FGF (b) pieces; (c) statistical distribution of the CF orientation angles regarding the longitudinal direction, measured from the data in (a) and (b); colour-coded representation of the CF length in the ASA_CF_LS_FGF (d) and ASA_CF_HS_FGF (e) pieces; (f) statistical distribution of the CF length measured in (d) and (e); (g) statistical distribution of the CF length measured in the LS and HS ASA_CF_pellets 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 225 velocity and a lower shear rate, and because of this, the fibres are more randomly arranged (Malloy, 2025;Rauwendaal, 2010), explaining the variation in fibre distribution observed in the CT analysis. Concerning the size of the CF, Figure 6(d) and (e), shows in colour code the quantification of the CF length in the ASA_CF_LS_IM and ASA_CF_HS_IM pieces, respectively, and Figure 6(f) shows the comparative length data histogram. As can be observed in Figure 6(f) and S5, these pieces also show length values smaller than 250 m m, analogously to the pellets used for the injection process, indicating that the CF is not remarkably further fragmented in the injection process applied after the material is compounded. Narrow length distributions are observed in both pieces, assuring the homogeneity of the material during the IM manufacturing process. Similarly to the FGF pieces, Figure 6(f) shows that, on average, the CF length in the ASA_CF_LS_IM piece is larger than in the ASA_CF_HS_IM one, which is expected to have some effect on their mechanical properties. This difference in length might also influence the orientation distribution previously discussed, as the larger the fibres, the more tendency to align in the flow direction, while the shorter fibres might be affected by the turbulent flow of molten polymer. 3.3 Tensile strength characterisation Mechanical properties determined from the composites fabricated by IM and by FGF are depicted in Figure 7. The corresponding engineering stress-strain curves can be found in Figure S5 as supporting material, while a representative curve of each material is depicted in Figure 7(d). As evidenced by the results, the incorporation of CF clearly enhances the elastic properties of neat ASA. Thus, the stiffness, evaluated in terms of tensile modulus (E), increases more than 150% for both injected and printed pieces, as shown in Figure 7(a). Also, the Figure 6 Colour-coded representation of the CF orientation in the ASA_CF_LS_IM (a) and ASA_CF_HS_IM (b) pieces; (c) statistical distribution of the CF orientation angles regarding the longitudinal direction, measured from the data in (a) and (b); colour-coded representation of the CF length in the ASA_CF_LS_IM (d) and ASA_CF_HS_IM (e) pieces; (f) statistical distribution of the CF length measured in (d) and (e) 3D large format additive manufacturing pieces Daniel Moreno Sánchez et al. Rapid Prototyping Journal Volume 31 · Number 11 · 2025 · 218–230 226