A hybrid material extrusion device with local debinding and sintering

.


Introduction
The fabrication of high-quality parts with intricate shapes has been a longstanding goal for engineers since the early days of the manufacturing industry. Currently, the sector is being transformed by the widespread adoption of additive manufacturing (AM), which is providing unprecedented improvements in the performance of engineering products. The allowable part complexity enabled by AM [1,2] can decrease the amount of total parts within a system [3,4], resulting in a decrease of joints and interfaces and increasing product reliability. In this way, the disadvantages of traditional assembly with bolts and rivets, which introduces stress concentrations and adds structural mass, can be eliminated. Highly-integrated complex systems are being manufactured on a single build, decreasing the overall mass, reducing single points of failure, and lowering development and tool costs, as shown in the aerospace industry with AM of spacecraft and rocket systems [5][6][7], on the automotive sector with combustion engines [8,9], or in the biomedical field with patient implants [10,11].
Compared to other AM methods, an important benefit of this technique is the immobilization of the powder particles which are embedded within a polymeric matrix acting as binder material (Fig. 1b, Step 1), with stearic acid typically used to facilitate uniform dispersion of the particles within the filament [25]. This polymeric matrix reduces the hazards of handling micron-sized particles or volatile resins. The main polymer (i.e., primary binder), comprising 50%− 90% of the polymer matrix, is typically removed from the "green part" by a soluble agent [39][40][41][42], catalysts [43,44], or through a thermal debinding [45,46] process (Fig. 1b, Step 2). The output of the debinding stage is comprised of metal or ceramic particles and a secondary binder that keeps these together, denoted as the "brown part". The subsequent sintering process, at temperatures below the melting temperature, thermally degrades the residual secondary binder and causes the metal or ceramic particles to coalesce, resulting in part densification (Fig. 1b, Step 3).
A major disadvantage of this method lies in the time-consuming processing and complicated heating procedures, often involving the use of toxic debinding agents (e.g., acetone [40], hexane [47], oxalic acid [42], or nitric acid [43]). Additionally, traditional debinding and sintering are done separately after shaping the part using sophisticate equipment, and little advancement has been made to develop the technology to process parts locally, within the same volume. It has been shown [48] that local debinding can be achieved efficiently through an application of local energy by selectively irradiating the built layers with a laser source, attaining fast debinding rates through the ablation of the main binder matrix through the top surface open pores. The polymeric mass can be removed by focusing sufficient energy on the material, which layer by layer vaporizes the primary binder along bulk of the shaped geometry. This approach allows to debind without consumables, detached equipment, extended debinding cycles, nor toxic agents that pose significant threat to both environmental and public health. Moreover, the capability of sintering MEAM metal parts with an efficient concentration of energy has been shown using indirect induction heating [49], which can render dense bodies in only 360 s of soak time, as opposed to hours via the traditional approach [50][51][52]. Compared to other heating methods, induction heating offers several advantages, including the capability to efficiently heat a crucible to high temperatures while minimizing energy waste and heating times. In this approach, and utilizing the Joule heating effect, electromagnetic energy is transformed to thermal energy to heat the crucible. Temperature profiles can be predefined through control techniques, and advanced capabilities such as local heating can be achieved, providing high selectivity in depth and across the surface of the workpiece.
In this study, the combination of FFF metal parts with both local laser debinding and induction sintering is investigated. The integration of MEX with local debinding and sintering is an innovative hybrid AM method and apparatus with the patent application registered by the Technische Universität Berlin [53]. The method described in this manuscript involves printing with an extruder head, and thermally debinding the main polymeric binder matrix by employing a low intensity diode laser, with a subsequent indirect induction sintering treatment (Fig. 2). The results are characterized through optical and electron microscopy, as well as through 3-point bending tests. Being able to debind and sinter in-situ decreases significantly the processing times and costs of FFF parts.

Materials and methods
A custom-built machine was developed within the European Space Agency's (ESA) funding framework combining a debinding heater and a sintering heater, with the schematic illustration and the assembled device shown in Fig. 3a and b-d, respectively. Cubic samples of 10 × 10 x 2.5 mm 3 were produced on this device using a filament of 1.75 mm diameter that was acquired from PT+A GmbH (Dresden, Germany). The extruder was fitted with a nozzle of vanadium alloy and high-speed steel from Slice Engineering (Gainesville, FL, USA) with an aperture size of 0.4 mm. The binder matrix is based on a polyamide and was embedded with micron-sized austenitic stainless steel 316 L powder, with the composition of the 316 L given in Table 1. Scanning electron micrographs of the as-purchased filament are provided in Fig. 4a and b at mag. 50x and 300x, respectively. The granule size distribution of the 316 L powder in the debinded state is mapped in Fig. 4c, exhibiting an average diameter of D 50 = 3-4 µm. The powder size distribution was determined from the SEM measurements in the debinded state, whereby the diameter of more than 150 individual particles has been measured. The parameters to shape the parts, i.e., temperature of the extruder, temperature of the heated bed, thickness of the layer, and print speed, are provided in Table 2. Pre-test measurements were made to identify an adequate combination of building parameters for the shaping phase by following the supplier guidelines. For comparison, control as-built (AB) specimens which underwent traditional solvent debinding and sintering were fabricated (Fig. 4d).

Debinding
An infrared diode laser was coupled to the extruder of the machine as illustrated in Fig. 3c-d, with the laser operated in continuous mode during the debinding treatment. The top surface of each layer was irradiated using parallel laser scans after fully shaping the layer. Representative laser scans for three different layers are illustrated in Fig. 5a-c. To compare the effect of laser debinding, two consecutive laser    scans with different processing parameters were conducted per layer; one with fast scan, and another with slow scan. The laser debinding parameters for each condition, i.e., wavelength, spot size of the beam, power of the laser, scan speed and hatch distance, are provided in Table 3. The control samples were debinded via traditional solvent debinding using a 99.5% pure acetone acquired from Höfer Chemie (Germany). Samples were placed on the acetone bath at room temperature for 24 h to ensure full removal of the binder mass, and were subsequently removed and left to dry at room conditions for an additional 24 h.

Sintering
After debinding, the parts were exposed to induction sintering testing. Sintering was done with a 5 kW induction heater working at 65 kHz resonant frequency, whereby after shaping the part and debinding with the laser, the specimens were lowered along the z-axis to undergo a sintering treatment. The indirect approach involved using a graphite crucible that was provided by the LLF Smelting Lab. The graphite crucible is heated rapidly and indirectly heats the debinded samples, with the test set-up shown in Fig. 3c. The sintering temperature was monitored and controlled utilizing a CSmicro 2MH pyrometer from Optris GmbH (Berlin, Germany), with the graphite crucible reaching a maximum temperature of 1365 • C. Samples were sintered at this temperature for 360 s of hold time. To avoid oxidation of the 316 L powder, indirect induction sintering was done within a vacuum level of 1.8•10 − 2 millibar. The sintering parameters were selected based on previous trials made on indirect induction sintering of stainless steel 316 L shaped via FFF [49].

Microstructure and atomic composition measurements
Specimens for microstructural analysis were prepared and micrographs were taken with an Axioscope 7 optical microscope (Zeiss, Germany), at 50x and 500x magnification. Surface topography features were measured at 200x magnification with a digital optical microscope VHX-7000 (Keyence GmbH, Germany), with the results characterized utilizing the VHX 3D measurement software. The local surface roughness was quantified from the surface topography using the surface assessment package from Fiji Image-J. A Hitachi S-2700 scanning electron microscope (SEM, Japan) was used to create electron backscatter diffraction (EBSD) maps at 3000x magnification. The EBSD maps were assessed using the Oxford Instrument Channel 5 software. The average density was calculated utilizing the Fiji Image-J analysis software. A JEOL JXA-8530 F microprobe operated at 12,000x magnification and 15 kV was used to generate wavelength-dispersive X-ray spectroscopy (WDX) and energy dispersive X-ray spectroscopy (EDX) measurements, with the results analyzed utilizing the JEOL WDX and EDX system.

Three-point flexural measurements
Three-point flexural tests were performed on a servohydraulic Zwick Roell HB100 (Germany) test machine to investigate the static behavior of indirect induction sintered parts. Cylindrical specimens were manufactured horizontally, with six samples produced along the XY directions, with Z being the build direction. Due to limitations in build volume for the custom device, the horizontally built beams had an asbuilt (AB) longitudinal dimension of 30 mm with a 3 mm diameter, shown in Fig. 6a and b in the XY and YZ planes, respectively. The indirectly induction sintered beams of the solvent debinded and laser debinded treatments are shown in Fig. 6c and d in the XY and YZ planes, respectively. Control commercial 316 L cylindrical samples of 30 mm × 3 mm were tested for comparison. A total of 6 samples have been employed for the tests, having 2 specimens per condition. The flexural surface strain of the specimens was calculated using Eq. (1).

Local debinding
Micrographs of the AB, solvent debinded, one laser scan and two   laser scans are provided in Fig. 7a-d, with 3D surface topography measurements given in Fig. 7e-h and color-coded surfaces in Fig. 7i-l. No shrinkage of the samples was detected after debinding in any of the conditions. For the green AB specimens (Fig. 7a, e, i), individual deposited tracks exhibiting inter-raster voids are observed as indicated by the yellow arrows on the figures, with a quantified height of the deposited layer of 0.1 mm. Raster voids are a well-known detrimental effect of depositing melted layers on FFF [54]. Their presence is attributed to the lack in control of the filament pressure after the extruded mass exits the print head [55], combined with rapid cooling of the melt with a subsequent decrease in flowability of the viscous raster [56]. This stretches the rasters into elliptical cylinders from the nozzle movement resulting in voids in between the deposited tracks [57][58][59].
On the solvent debinded specimen (Fig. 7b, f, j), deposition tracks are also observed, without any major quantified deviation compared to the AB state. For the laser debinded specimen after one pass (Fig. 7c, g, k), clumps of powder forming hills and valleys are observed. This is attributed to the laser fluence, which upon impacting the surface tends to re-organize the layer agglomerating the powder, congruent with previous trials [48]. Superficial inter-raster voids start to be removed within the pore network in this condition as the polymeric mass ablates and redistributes the tracks of the viscous binder due to capillary forces [44]. Upon a second laser debinding treatment (Fig. 7d, h, l), the surface topography improves compared to one single pass. The residual primary binder mass is ablated and the clumps and valleys of powder are eliminated due to the laser fluence. No inter-raster voids can be detected on the samples after two laser treatments. The total debinding time was 247.5 min for the full laser debinding cycle with two passes and over 24 h for the solvent debinded condition. This represents a decrease in debinding time of 482% when concentrating local energy on a layer-by-layer approach compared to traditional solvent debinding. Optical micrographs of the laser debinded condition after two scans are provided in the XY (top view) and ZX (side view) planes in Fig. 8a-d. A noticeable contrast is detected between the non-debinded AB region and the local laser debinded region in Fig. 8a, with the magnification shown in Fig. 8b highlighting full debinding of the specimen. Some residual binder is found within the bottom left corner of Fig. 8b, together with several micropores (Fig. 8b-c). It is important to fine-tune the laser parameters, as delivering a high threshold of energy may lead to detrimental pores, while a low energy intensity will not result in full debinding of the layer. Therefore, careful calibration of the laser energy is necessary to achieve optimal debinding while mitigating the generation of debinding defects. However, despite having ablated the polymer on a debinded layer, it is observed that consecutive extruded layers adhere to the previously debinded layer using the binder mass of the current track being deposited (Fig. 8c-d). The binder thus adheres optimally to the previously debinded layer. Choosing the ideal laser parameters for local debinding is ultimately dependent on the specific binder system employed, the layer thickness, and the composition of the powder. As such, future work should concentrate on modelling the kinetics of laser debinding to ensure high quality debinded layers.
To further analyze the effects of the laser debinding process, SEM measurements on the bulk of the AB and laser debinded conditions were taken in the ZX plane, as shown in Fig. 9a-d. In the AB state, the binder  mass encapsulates the metallic particles fully embedding them across the bulk of the printed part. No discernible defects or imperfections can be detected in the bulk of the AB state. On the laser debinded specimens, an increase in number of fully visible particles can be observed within the bulk, showing a clear debinding effect compared to the AB state. In agreement with [48,60], the primary binder is ablated by focusing photothermal energy on the top surface of the deposited layer. The energy vaporizes the polymeric matrix when the laser threshold fluence provides sufficient energy to the binder which absorbs it and increases its temperature to sufficient levels allowing the decomposition of the polymer. On the other hand, in contrast to the AB state, the laser debinded specimens exhibit some pores due to the laser fluence. To mitigate the creation of these macropores, it is crucial to carefully control the heating rates during the laser debinding treatment. Such control through tailored debinding profiles can prevent the formation of detrimental defects and improve the overall quality of laser debinded parts. Laser debinding is a promising alternative to avoid the increase in vapour pressure within the bulk of debinded parts, a typical occurrence in thermally debinded FFF geometries, as the polymeric mass is ablated through the top surface open pores, mitigating cracking, blistering and bloating defects [46,61].

Sintering: Density & microstructure
After debinding, samples were sintered for 360 s via the indirect induction approach, whereby a conductive graphite crucible was employed as susceptor to indirectly heat the debinded geometries. Surface 3D topography measurements and color-coded topographic surfaces in the XY plane and Z-build direction are provided in Fig. 10a-h for the solvent debinded and sintered, as well as laser debinded and sintered specimens. The as-sintered small-scale arithmetic average surface roughness is 15.44 µm and 13.67 µm for the solvent debinded in the XY and Z planes, respectively, and 16.17 µm and 13.92 µm for the laser debinded and sintered in the XY and Z planes, respectively. The sintered state of the solvent debinded specimen reveals inter-raster voids ( Fig. 10a-b). As the individual rasters undergo shrinkage and contraction upon sintering, there is a corresponding decrease in the diameter of their tracks, quantified by a percentage contraction of 6.3% compared to the AB state. Shrinkage of FFF parts is a well-known phenomenon observed upon solid-state sintering [62][63][64][65], being a result of the packing density of the powder within the filament feedstock. To minimize part shrinkage and achieve high densities, it is advised to embed the filament with a high powder volume, thereby increasing the packing density within the feedstock material. A typical recommended volume loading of the powder is between 55% and 70%, ensuring adequate filament flowability and high densities of the sintered bodies [66,67]. Additionally, increasing the raster overlap and employing fine micron-sized powders improves the packing density between particles, which reduces interstitial voids and promotes higher density while reducing shrinkage after sintering. On the other hand, inter-raster voids have been removed after two laser debinding treatments in the laser debinded and sintered condition (Fig. 10e-f), suggesting that these large surface pores can be removed around the near surface zone using the laser debinding technique. There is a detrimental effect of pores on the fatigue properties of stainless steel 316 L produced via additive manufacturing, as the concentration of stress around these voids promotes the origin of nucleation sites for crack initiation and propagation, particularly close to the surface [68]. Removing these surface pores thus Fig. 11. EBSD maps of the laser debinded and sintered specimens with a) inverse pole figures along z (IPFz), b) phase maps c) local misorientation maps, d) binarized map, e) Euler pole figures, (f) local misorientation distributions of the laser debinded and induction sintered (red curve) and solvent debinded and induction sintered (green curve), with the data for the solvent debinded extracted from [49], and g) grain size distribution of the sintered sample, with an average of ~ 9.5 µm.
promotes higher fatigue life and strength [69].
Electron back-scatter diffraction (EBSD) maps for the laser debinded combined with induction sintered samples are given in Fig. 11a-e. Porosities within the maps are displayed in black due to non-indexed points. The inverse pole figure map along z (IPFz) shown in Fig. 11a displays an equiaxed grain morphology, with the grain size distribution mapped in Fig. 11g presenting an average grain size of ~9.5 µm. Due to the uniform external heating during the sintering process, which provides an even delivery of heat throughout the sintered body, a homogeneous microstructure emerges across the part. Sinter-based AM processes do not involve external stresses before or during sintering, with the final microstructure mainly influenced by the initial powder morphology, print defects, chemical composition and sintering conditions (i.e., heating and cooling rates, as well as atmosphere), which promotes equiaxed grain morphology due to the proportional incremental grain growth in all spatial axes [70]. Several annealing twins are detected in the sintered samples after 360 s. This has been observed previously during the grain growth of sintered stainless steel 316 L powdered geometries shaped via FFF [50,71]. The emergence of annealing twins is linked to the grain growth in annealed FCC metals. Annealing twins are promoted at locations wherever there is a change in stacking sequence between crytallographic planes, with annealed metals exhibiting considerable twin band formation after extensive grain growth [72]. The colour-coded Euler pole figures in the planes {001}, {101}, and {111} are provided in Fig. 11e. These images reveal the texture which characterizes the orientation of the grains in the sintered parts. The texture is weak with 2.35 maximum intensity, presenting a random isotropic crystallographic texture. This indicates a random distribution of the grains, in agreement with previous work conducted on sintering stainless steel 316 L produced via FFF [73].
Phase maps providing information on the crystallographic structure are shown in Fig. 11b. The colour blue indicates the presence of a face centered cubic (FCC) structure, while red corresponds to a body centered cubic (BCC) structure. The prevalent crystallographic phase observed is the γ-austenitic FCC phase, which accounts for 96.28% of the material composition. In addition, a secondary phase of δ-ferrite BCC is present, comprising 3.71% of the material. The distribution of austenitic and ferrite phases is in agreement with previous studies of sintering stainless steel 316 L shaped via the FFF approach [49,71]. The observed retention of δ-ferrite is considered to arise from the rapid solidification of molten stainless steel 316 L powder feedstock during the gas atomisation process [73,74]. Specifically, the segregation of δ-ferrite stabilizers during this process, in particular Cr, Mo and Si, creates localized site-specific metastable conditions within the austenitic matrix.
The Kernel average misorientation (KAM) map depicting local crystallographic misorientations is shown in Fig. 11c. Using a scale between 0 • and 5 • , these maps quantify the average misorientation in between adjacent pixels within regions extending up to its seventh nearest neighboring pixels, without including misorientations exceeding 5 • . The results are drawn in Fig. 11f, with the red curve depicting the laser debinded and induction sintered specimens and the green curve representing the solvent debinded and induction speciments, with the later results extracted from [49]. The laser debinded and induction sintered samples display higher misorientation levels compared to the solvent debinded and sintered specimens after 360 s. This is attributed to the efficiency in the sintering treatment, as the solvent debinded and sintered achieved a higher density in previous trials [49], exhibiting a lower crystallographic misorientation due to an increased annealing of the initially stored dislocations [75] within the stainless steel 316 L powder. This is a known effect of heat treatments conducted on 316 L fabricated using additive manufacturing [76].
Following a duration of 360 s of indirect induction sintering, a number of confined pores are observed within the equiaxed grains as well as pores distributed on the grain boundaries as a result of particle densification [71]. The presence of these pores is an occurrence commonly encountered during the last phases of sintering [77]. The bulk density of the laser debinded and sintered sample is 94.77%, with the binarized maps given in Fig. 11d. This is inferior than prior tests made through traditional sintering which achieved above 95% density on stainless steel 316 L [28,52], and above 99% using the indirect induction sintering approach [49]. Higher densities can be achieved by tailoring the heating profiles to a specific material and geometry, as well as increasing the hold time of the sintering treatment. With over 50 porosities measured on the solvent debinded and induction sintered, and laser debinded and induction sintered conditions, the overall average porosity size is 1.75 ± 0.87 µm and 3.57 ± 1.50 µm, respectively, with the size distribution mapped in Fig. 12a-d. In the solvent debinded and sintered case, the mean porosity size is lower than the laser debinded and sintered, exhibiting percentage decrease in porosity size of 51.04%. Fig. 12b displays the ellipticity of the porosities, wherein Fig. 12c-d depict magnified representative porosities for the solvent debinded and sintered and laser debinded and sintered, respectively. These pores are fitted with an ellipse that is defined by a semi-major axis b and a semi-minor axis a. A ratio of a/b reaching 1 indicates spherical micropores, while values that tend towards 0 describe pores with high ellipticity. The observations suggest that there is a tendency for the micropores to develop more spherical given an increase in the effect of the sintering treatment.

Composition analysis
Wavelength-dispersive X-ray (WDX) chemical mapping of integral elements to stainless steel 316 L alloy, as well as potential contaminants, are shown in Fig. 13. These measurements include Fe, Cr, Ni, C, Mo, Mn, Si, S, and O. Only the elements reported to have a significant detrimental effect on the mechanical response of FFF 316 L are considered. Point energy-dispersive X-ray (EDX) results, together with the elementcomposition counts, are provided in Fig. 14 at three different locations, with the quantified values in wt% given in Table 4. Compared to the chemical composition of the original stainless steel 316 L powder, the laser debinded and induction sintered specimens exhibit a notable carbonization. This is believed to result from residual carbon originating from the organic binder material [78], as well as from the evaporated residues from the graphite crucible during the sintering treatment, which is consistent with previous findings [49]. In order to reduce carbonization upon sintering, further studies on the impact of crucibles made of other materials on the chemical composition of stainless steel 316 L sintered via indirect induction should be conducted. In addition, measurement 3 indicates that there is a greater concentration of carbon (C) along the boundaries of the grains. This observation, together with the apparent enrichment of chromium (Cr) and molybdenum (Mo) in the boundaries, suggest there are chromium-rich carbide precipitates at the boundaries being likely Cr 23 C 6 or Cr 7 C 3 [79,80]. This has been reported previously by similar work made on sintered stainless steel 316 L [28,49]. This is associated with sensitisation of the steel [81], whereby as carbide precipitation occurs, carbon atoms move quickly to the grain boundaries through interstitial diffusion. On the other hand, chromium atoms move at a much slower rate, resulting in areas depleted of chromium. This can have a detrimental effect on the mechanical properties, resulting in an increase in intergranular corrosion [78] and reduced elongation and embrittlement [28]. A further observation reveals that the closed pores exhibit an increased concentration of oxygen, which occurs due to the presence of residual oxygen during the vacuum sintering process. It is necessary to experiment with higher levels of vacuum to prevent any oxidation of the steel. Future work should focus on preventing sensitisation of the 316 L by adjusting the sintering loads, and reducing the oxygen content within the build chamber. However, it is essential to highlight that the choice of the EDX/EDS methodology is not the most adequate to quantify the carbon content, which can lead to inaccurate measurements due to matrix effects. These effects occur when the composition, structure, or contaminants in the sample impact the emission and detection efficiency of X-rays in EDS, leading to potential inaccuracies in the carbon measurements.

Mechanical properties
The mechanical properties were quantified by flexural 3-point bending tests to characterize the bending resistance of the sintered specimens. The results of the traditional solvent debinded and sintered (green curve), the local laser debinded and sintered (red curve), and commercial stainless steel 316 L (blue curve) samples are mapped in Fig. 15. The commercial 316 L bars did not break but were just bent exhibiting a ductile behavior (Fig. 15b), whereas both induction sintered bars completely fractured, with the fracture surfaces shown in Fig. 15cd. In the commercial 316 L condition, the engineering stress-strain curve displays an initial linear elastic region with a subsequent extended plastic non-linear region. For the solvent debinded and sintered, the elastic region is narrow with a more extended plastic behavior of the specimens. The laser debinded and sintered also presents an initial elastic region but does not present an inelastic plastic zone, showing a sharp brittle fracture. There is a clear embrittlement of the sintered specimens after reaching a maximum flexural stress value of 1124 MPa and 934 MPa for the solvent debinded and laser debinded, respectively, and an engineering strain of 0.0201 and 0.0098, respectively. The calculated Young's modulus in the elastic regions is 110.53 GPa, 102.64 GPa, and 145.04 GPa for the solvent debinded and sintered, laser debinded and sintered, and commercial 316 L, respectively.
Compared to the commercial 316 L, the decrease in strength and   ductility of the sintered specimens is attributed to interlayer delamination ( Fig. 15c-d), the presence of pores (Fig. 11d), as well as the sensitisation of the samples. Moreover, the surface roughness measurements and the presence of inter-raster voids provides regions promoting crack nucleation and propagation [68], thereby decreasing the mechanical response compared to as-purchased 316 L. This explains the decrease in Young's modulus in the sintered specimens. In addition, the brittle fracture in the laser debinded condition, exhibiting no plastic deformation before fracture, is attributed to the presence of larger pores with higher aspect ratio ( Fig. 12a-b), and decreased density, compared to the solvent debinded. Subsequent efforts should concentrate on enhancing the density of the sintered specimens and refining the surface quality of the components. The reduction of surface roughness can be readily achieved before the sintering process by mechanically eliminating the near-surface zone through machining, as the samples can be very easily worked prior to densification. However, implementing this approach necessitates the inclusion of additional steps between the debinding and sintering stages.

Conclusion
A novel hybrid device enabling local debinding and sintering of metallic parts shaped via FFF has been investigated in this work, revealing that geometries can be processed without part transportation, toxic consumables or complicated post-processing. The method, comprised of a low intensity debinding laser and an indirect induction heater for sintering, exhibits a promising alternative to efficiently process FFF AM metallic or ceramic parts. The following conclusions are made: 1. Layer-by-layer debinding with lasers can effectively remove the primary binder matrix, showing an increase in debinding efficiency of 482% compared to traditional solvent debinding. However, a careful choice of parameters need to be further investigated to ensure  full debinding on every layer while minimizing the generation of defects. A slow laser scan can result in clumps of powder particles, whereas too fast can result in having non-debinded regions. 2. Inter-raster voids are removed upon laser debinding with two passes.
The clumps of powder forming valleys are relaxed upon a second laser scan treatment. 3. Sintering of laser debinded parts results in densities 94.77% in only 360 s of soak time using an indirect induction sintering heater. Higher power levels and soak times are necessary to increase the density of stainless steel 316 L and sinter larger parts shaped via FFF. 4. An isotropic fine-grained FCC γ-austenitic microstructure is achieved within the bulk of the laser debinded and induction sintered parts. The final grain size depends on the soak time, with induction sintering being a technique offering precise control of the heating rates. Dedicated sintering profiles need to be developed in future work to expand the materials database whilst ensuring high density parts. 5. There is an apparent carbonization of the 316 L steel processed through this method which is connected to the use of a graphite crucible. The absorption of carbon residues can be avoided by efficiently extracting the fumes generated during the heating of the crucible. It is critical to point out that the chosen EDX/EDS methodology to quantify the carbon content can lead to inaccurate measurements. Future efforts should prioritize investigating the use of alternative crucible materials, and analyzing the carbon content through other techniques such as carbon combustion analyzers. 6. Flexural 3-point tests revealed that the stress-strain curves are reduced compared to commercial stainless steel 316 L. This is attributed to interlayer delamination during testing, sensitisation of the specimens, as well as the presence of pores and their subsequent decrease in density. Due to the high number of variables influencing the process (i.e., shaping, debinding, and sintering parameters), future work shall focus on conducting further mechanical tests according to international standards and thus gather a comprehensive set of statistical data. 7. The total processing time for the samples debinded and sintered through this method is 253.5 min, offering an increase in efficiency compared to debinding and sintering through traditional routes.
Despite the reduced mechanical properties, in situ debinding and sintering is an innovative method that offers promising routes to process MEX parts, reducing the complexity, costs, and risks of other conventional approaches.

Funding
This work was financially supported by the OSIP framework of the European Space Agency under Contract No. 4000133882/21/NL/MH/ic which is gratefully appreciated.