Additive Manufacturing in Space: Process Physics, Qualification, and Future Directions

Abstract

Additive manufacturing has emerged as a key enabling technology for in-space manufacturing, offering the potential to reduce logistics mass, enhance mission autonomy, and support long-duration exploration. The suppression of gravity-driven phenomena fundamentally alters melt pool dynamics, heat transfer, surface-tension-dominated flow, and defect formation, limiting the direct transferability of terrestrial AM process knowledge to space applications. This paper reviews the current understanding of metallic additive manufacturing process physics under reduced gravity, with emphasis on melt pool behavior, dimensional stability, and in situ monitoring constraints. Approaches for qualification and certification are critically examined, including the applicability of existing AM standards, the role of digital twins and model-based verification, and emerging strategies for space-based validation. Enabling technologies such as autonomous and AI-assisted fabrication, compact hardware architectures, and alternative energy sources are discussed in the context of reliable in-space operation. By synthesizing current developments and identifying key limitations and open challenges, the review provides a roadmap for advancing additive manufacturing toward operational readiness, supporting sustainable exploration, in-space infrastructure development, and long-duration human presence beyond low Earth orbit.

1. Introduction

Additive manufacturing (AM) has emerged as a transformative technology with the potential to revolutionize how components and systems are designed, fabricated, and sustained in terrestrial settings. As humanity extends its reach beyond Earth, the transition of AM into space environments—characterized by microgravity, vacuum conditions, and extreme thermal gradients—poses both unprecedented opportunities and formidable challenges. Space additive manufacturing (Space AM) promises significant benefits, including in situ production of mission-critical hardware, reduced dependency on Earth-based logistics, and enhanced mission flexibility. However, the unique physics governing material behavior in microgravity and vacuum, coupled with the stringent requirements of space operations, necessitate a fundamental rethinking of process control, materials science, and qualification methodologies.

Even with the significant advances made in space-based additive manufacturing over the last ten years, a number of core challenges still control its transition from experimental systems to a mature, fully functional manufacturing capability. The first and most persistent limitation concerns material diversity. To this point, most in-space fabrication has relied on thermoplastic polymers such as acrylonitrile butadiene styrene (ABS), high-density polyethylene (HDPE), and ULTEM 9085, which offer low processing temperatures and manageable handling in microgravity. However, these polymers restrict the achievable mechanical performance and thermal stability of printed parts, highlighting the need to extend fabrication to metallic, ceramic, and composite materials compatible with the thermal and environmental conditions of orbital and planetary applications.

A second challenge is the lack of standardized certification pathways for components produced in orbit. Terrestrial aerospace manufacturing benefits from well-defined qualification frameworks (e.g., NASA-STD [1], ECSS [2], ISO ASTM 529XX [3]), yet these cannot be directly applied to microgravity production, where process repeatability, traceability, and defect detection remain only partially understood. Accordingly, the implementation of robust in situ monitoring and verification standards is a key requirement for the qualification and acceptance of additively manufactured components for mission-critical applications in space.

Finally, the physics of the printing process in microgravity remains an active research gap. Reduced convection, altered melt flow, and unpredictable layer bonding complicate thermal management and interlayer adhesion, leading to mechanical anisotropy and uncertain quality metrics. Experimental data from the Additive Manufacturing Facility (AMF) and related ISS payloads have provided valuable insight, yet comprehensive models of material behavior under microgravity are still emerging. Overcoming these bottlenecks will require coordinated advances in materials science, process modeling, and autonomous control systems, elements discussed in the following sections.

This paper serves as a companion to the previously published “Additive Manufacturing in Space: Technologies, Flight Heritage, and Materials” paper [4], which highlighted flight heritage and demonstrated in-space manufacturing systems, including polymer printers deployed on the ISS (AMF, Zero-G Printer, POP3D), the Ceramic Manufacturing Module, and in-orbit recycling technologies enabling closed-loop material utilization. That work also provided a comprehensive review of materials for in-space manufacturing, covering polymers and composites, metals and alloys, ceramics and photopolymers, regolith-based feedstocks for in situ resource utilization (ISRU), and recycling strategies supporting circular manufacturing.

In contrast to existing reviews that primarily emphasize materials, flight heritage, or isolated technology demonstrations, this review adopts a process-physics- and systems-oriented perspective on space additive manufacturing. The essential distinction lies in treating Space AM as a coupled cyber–physical manufacturing system, in which microgravity-altered process physics, in situ monitoring, autonomous control, and qualification requirements are inherently interdependent. Accordingly, the paper first examines the fundamental melt pool dynamics, heat transfer mechanisms, powder handling challenges, and defect formation processes under microgravity (Section 2), establishing the physical constraints that govern process stability and repeatability. It then addresses the qualification, standardization, and verification frameworks required to certify additively manufactured components produced in orbit (Section 3), highlighting the gap between terrestrial standards and space-based manufacturing realities. Building on these foundations, the review discusses digital twin frameworks as a unifying link between physics-based modeling, real-time sensing, and predictive control (Section 4). Finally, emerging trends are examined, with particular emphasis on autonomous and AI-assisted fabrication, closed-loop manufacturing, and in-space resource utilization (Section 5), culminating in a critical assessment of remaining challenges and research gaps. This structured approach distinguishes the present work from prior reviews by providing a coherent roadmap from microgravity process physics to autonomous, certifiable, and scalable in-space manufacturing systems.

2. Process Physics Under Microgravity and Vacuum

The transition from terrestrial additive manufacturing (AM) to orbital or lunar environments introduces a fundamentally different process-physics regime. Microgravity, vacuum, radiation, and thermally extreme surroundings modify how melts behave [5], how powders [6] and resins flow [7], how heat is redistributed [8], and how defects nucleate or evolve [9]. Understanding these deviations is essential for designing space-qualified AM hardware and for predicting part quality, reliability, and mechanical performance. This section summarises the dominant physical mechanisms governing AM processes under microgravity and vacuum, drawing from ISS demonstrations, parabolic flight experiments, drop tower campaigns, and numerical studies.

2.1. Melt Pool Behavior, Surface Tension and Buoyancy Effects

2.1.1. Gravity-Dependent Melt Pool Flow Regimes

Melt pool dynamics represents one of the most profoundly altered physical phenomena when additive manufacturing transitions from terrestrial gravity to microgravity environments. Under Earth gravity, the shape, stability, and flow within a melt pool are governed by a balance between buoyancy-driven convection, Marangoni forces [10] (forces on melt pool are presented in Figure 1), surface tension gradients, and gravity-induced drainage. For typical terrestrial laser-based AM processes, reported melt pool depths range from approximately 100 to 500 μm [11,12], with buoyancy-induced flow velocities that are comparable to, or in some cases exceed, thermocapillary-driven velocities for commonly processed metallic alloys.

In microgravity, buoyancy effects are effectively suppressed, shifting the melt pool transport regime toward surface-tension-driven and thermocapillary flows. CFD simulations of laser metal deposition for potential space additive manufacturing show that, as gravity is reduced toward zero, surface tension becomes the principal driving force in the melt pool, fundamentally altering melt flow dynamics relative to terrestrial conditions. Moreover, the literature reviews have identified that deposition irregularity and melt pool behavior become increasingly dominated by capillary effects as gravitational forces diminish, consistent with the suppression of buoyancy-driven convection [11,12]. As a consequence, melt pool behavior in microgravity must be described within a capillary-dominated flow framework rather than the mixed buoyancy–Marangoni regime characteristic of Earth-based AM.

Experiments conducted using parabolic flights, drop towers, and orbital platforms such as the International Space Station (ISS) have provided key insights into melt pool dynamics and solidification behavior under reduced-gravity conditions. Wire-fed laser melting and metal droplet solidification tests consistently demonstrate more spherical melt pools, reduced internal convection, slower collapse kinetics, and an increased susceptibility to pore entrapment compared with terrestrial conditions [14]. Foundational investigations of microgravity melt behavior include early droplet solidification experiments [6], studies of thermal convection under microgravity [5], and more recent analyses of resin and metal melt stability under partial-g environments [13]. Collectively, these studies establish that melt transport and heat transfer in microgravity are dominated by surface-tension-driven mechanisms rather than buoyancy, underscoring the need for microgravity-specific models to accurately predict bead formation and defect evolution in space-based additive manufacturing processes [15].

Ground-based reduced-gravity platforms play a complementary role by enabling controlled and repeatable experimentation. Parabolic flight campaigns are widely used for scientific research and training, offering short-duration microgravity conditions with flexible experimental access, rapid reconfiguration, and the capability to conduct human-in-the-loop experiments, making them particularly valuable for exploratory studies [16]. However, residual accelerations and gravity fluctuations (g-jitter) during parabolic flights and sounding rocket experiments can degrade microgravity quality and have been shown to artificially enhance heat transfer by deforming bubbles and inducing additional flow within the melt [17,18].

Drop towers are an important class of ground-based microgravity facilities, providing high-quality reduced-gravity conditions (≈10⁻⁶–10⁻² g) in a cost-effective and accessible manner. Modern systems such as the Einstein-Elevator at the Hanover Institute of Technology (HITec) support large experimental payloads (up to 1000 kg) within vacuum-compatible chambers and allow high experimental throughput (up to 300 runs per day) [19,20]. A structure of the Einstein-Elevator is illustrated in Figure 2.

Figure 3 illustrates a powder-based metal additive manufacturing process investigated under microgravity at HITec using the Einstein-Elevator facility [22]. The study focused on developing reliable powder handling strategies, adapting a laser metal deposition (LMD) system for reduced-gravity operation, and maintaining a stable melting process under near-zero-g conditions. The results show that the suppression of buoyancy-driven flow in microgravity leads to localized heat accumulation, reduced dilution, and altered penetration profiles, which in turn increase the likelihood of pore retention and lack-of-fusion defects.

Under microgravity, buoyancy is effectively suppressed, eliminating a major driver of melt pool convection. Without gravity-induced fluid motion, molten metal remains more static, heat accumulates locally, and surface-tension-driven flows become dominant. This shift alters melt pool geometry, reduces internal mixing, changes dilution and penetration profiles, and increases the probability of pore retention. Understanding the reduction in buoyancy forces, and the resulting rise in capillary-dominated flow regimes, is therefore essential for predicting material behavior in orbital or lunar AM. Several studies help quantify how buoyancy contributes to melt pool dynamics and how its suppression transforms fluid flow in microgravity. Xiao and Zhang [23] conducted foundational modeling of direct metal laser sintering, demonstrating that melt pool flow on Earth is strongly influenced by the combined action of buoyancy-driven convection and Marangoni forces. Their simulations showed that buoyancy enhances circulation and influences melt pool depth and shape, making it a critical parameter in terrestrial AM.

Arunachalam et al. [15] conducted an experimental investigation of droplet transfer and bead formation in arc-based directed energy deposition under reduced gravity conditions. Their results indicate that suppressed buoyancy leads to more spherical droplet formation, reduced detachment forces, and smoother melt pool dynamics. These observations are consistent with theoretical expectations that, under normal gravity, buoyancy promotes droplet elongation and stretching during the deposition process.

Table 1. Comparison of melt pool metrics under terrestrial and reduced gravity conditions.

Parameter	Terrestrial Gravity (1 g)	Microgravity
Melt pool area [24]	0.538 mm2	0.551 mm2 (μg)
Maximum internal flow velocity (melt pool) [24]	~1.022 mm/s (normal g, Y-direction)	~0.419 mm/s (μg, Y-direction) − ~59%
Gravity effect on melt pool flow regimes [12]	Buoyancy present; stronger convective contribution	Buoyancy suppressed; dominated by surface tension
Melt pool morphology/dimensions (LPBF) [25]	Width ≈ 84.4 ± 12.6 μm, Depth ≈ 68.0 ± 10.0 μm	N/A
Cooling rate (terrestrial) [26,27]	104–106 K/s typical	Lower in microgravity levitation experiments (≈40–110 °C/s lower)
Melt pool convection vs. surface tension effects [12]	Buoyancy & Marangoni both contribute	Surface tension dominates in μg (irregular tracks predicted)

presents quantitative comparisons of melt pool characteristics under terrestrial and reduced-gravity conditions, drawing on available experimental and numerical studies. This includes melt pool area, internal flow velocity, and cooling trends, providing direct numerical support for the influence of buoyancy suppression and surface-tension-dominated transport in microgravity.

2.1.2. Heat Transfer and Thermal Transport

On Earth, convective heat transfer within the melt pool plays a major role in thermal homogenization, with cooling rates typically ranging from about 10⁴ to 10⁶ K/s in laser powder bed fusion processes, depending on process parameters and alloy system. These rapid rates arise from the localized high thermal gradients and the combined effects of conduction and fluid flow within the melt pool, and they are key drivers of microstructural evolution and defect formation in the solidified material [28,29].

Experimental microgravity studies of condensation and two-phase heat transfer report that heat transfer coefficients can be reduced by roughly 20–50% relative to normal gravity, consistent with an expected reduction in convective heat transfer under suppressed buoyancy. Numerical simulations of melt pool behavior under reduced gravity further indicate that thermal fields and isotherm distributions differ substantially from terrestrial conditions, reflecting diminished convective heat transport and altered thermal gradients. These effects are expected to directly impact solidification morphology and dimensional accuracy [30,31]. This reduction directly influences melt pool geometry, dimensional accuracy, and solidification morphology. Localized heat accumulation becomes more pronounced, leading to altered penetration profiles and increased sensitivity to process parameter variations.

Microgravity flow boiling experiments were conducted aboard the ISS using the Flow Boiling and Condensation Experiment (FBCE) with subcooled n-perfluorohexane flowing in a single-side-heated rectangular channel. The study investigated wide ranges of mass velocity, inlet subcooling, and pressure, with heat transfer behavior analyzed through flow boiling curves, wall temperature profiles, heat transfer coefficients, and high-speed video imaging. Results show that mass velocity and inlet subcooling strongly affect heat transfer and flow regimes, while inlet pressure has a minor influence. Parametric trends in microgravity are similar to those observed in vertical up flow under Earth gravity, although flow instabilities occurred at low mass velocities, high heat fluxes, and high subcooling. The experiments produced a comprehensive database under long-duration, well-controlled microgravity conditions [32].

In microgravity, boiling heat transfer suffers from a reduced critical heat flux (CHF) due to vapor accumulation near the surface. Experiments performed during parabolic flights show that micro-finned surfaces alone do not enhance CHF, despite improving heat transfer coefficients. Applying an electric field promotes vapor removal and significantly increases CHF. The combined use of micro-fins and electric fields produces a strong synergistic effect, raising CHF by up to 2.2 times and restoring surface enhancement in microgravity [33]. Pool boiling experiments with FC-72 were performed under normal gravity and microgravity using plain and microstructured surfaces, with and without an applied electric field. Microstructured surfaces enhance the critical heat flux (CHF) in microgravity, while electric fields improve CHF on plain surfaces. The best performance is achieved by combining both methods, showing a strong synergistic effect. This approach increases the microgravity CHF up to 257 kW/m², exceeding the terrestrial CHF on a plain surface (168 kW/m²), highlighting its potential for efficient two-phase heat transfer in microgravity [34].

Figure 4 shows high-speed video images of the boiling process under all operating conditions. In microgravity without the EF (EF off, Figure 4e–h, green frame), bubbles nucleate and grow on the heated surface, merging into a large bubble that hovers and detaches periodically. This contrasts with normal gravity (Figure 4a–d, blue frame), where bubbles remain small and detach promptly. Applying the EF in microgravity disrupts the large vapor bubble, producing smaller bubbles and a boiling pattern qualitatively similar to normal gravity. On Earth, the EF has little effect, as the electric body force is weak compared to buoyancy and does not alter the overall boiling dynamics [35,36].

2.1.3. Implications for Solidification, Microstructure, and Defect Formation

Beyond melt pool flow, microgravity directly affects solidification behavior by suppressing buoyancy-driven convection and reducing thermal homogenization. Heat transfer becomes predominantly conduction-controlled, leading to steeper thermal gradients, longer liquid lifetimes, and slower cooling rates. These conditions modify grain growth and dendritic morphology, often resulting in coarser, more isotropic microstructures and increased susceptibility to pore entrapment. Consequently, solidification in microgravity is governed by transport mechanisms fundamentally different from those under terrestrial gravity, underscoring the need for microgravity-specific modeling to accurately predict microstructural evolution and resulting material properties.

Figure 5 compares the surface morphology, internal defects, and grain structure of Al–Bi–Sn samples solidified under space, anti-gravitational, and normal gravitational conditions. The space-solidified sample exhibits a smooth surface and no visible gas cavities, whereas terrestrial samples show rough surfaces with grooves and significant porosity, including large gas cavities under anti-gravitational conditions and pinholes under normal gravity. Grain structure analysis reveals predominantly equiaxed α-Al grains in both space and anti-gravitational samples, while the gravitational sample contains a substantial fraction (~32%) of columnar grains. Moreover, the space-solidified sample shows the smallest average grain size (≈414 μm), which is about 7.5% smaller than the anti-gravitational sample and approximately 44% smaller than the gravitational sample [37]. These results indicate that microgravity conditions effectively suppress convective flow and Stokes motion of minority-phase droplets and gas bubbles, thereby reducing macrosegregation and porosity while promoting equiaxed grain formation.

Additional insights come from studies on Al–Cu alloys, which demonstrate that gravity and composition jointly control both solidification and mechanical behavior at the nanoscale. Sarker et al. [38] report that Earth and Martian gravity accelerate solidification, producing heterogeneous microstructures with higher dislocation densities, whereas microgravity delays nucleation and yields more uniform, defect-poor FCC crystals. This leads to gravity-enhanced hardening in low-Cu alloys through dislocation multiplication, while in Cu-rich alloys microgravity suppresses solute segregation and increases hardness via solid solution strengthening.

Figure 6 presents the dendritic morphologies in the longitudinal sections of Al–Cu and Al–Si samples. Zhang et al. [39] show that gravity significantly affects alloys that form columnar dendrites, such as Al–3.5 Si, causing dendrites to tilt (~8.5°) from the growth axis and promoting more randomly oriented dendrites near the crucible wall. Under microgravity, columnar dendrites grow nearly parallel to the axial direction, and fewer random dendrites form. In contrast, alloys that solidify predominantly as equiaxed dendrites, like Al–10 Cu, show little difference between 1 g and μg conditions, indicating that the influence of gravity is strongly dependent on dendrite morphology and crystallization mode.

Williams et al. [40] investigated bottom-up solidification of Al–Cu alloys containing 4, 10, and 18 wt% Cu under terrestrial and microgravity conditions. Under microgravity, all samples exhibited fully equiaxed grain structures, whereas terrestrial samples showed a columnar-to-equiaxed transition (CET), with columnar grains near cooled surfaces attributed to convective melt flow. Grain sedimentation and flotation were observed in terrestrial samples depending on alloy composition, while inverse segregation driven by shrinkage flow occurred under both gravity conditions. However, terrestrial samples exhibited more pronounced solute non-uniformity due to convection. The reported grain structure, eutectic fraction, temperature, and macro-segregation data provide valuable benchmark cases for the validation of solidification models under reduced-gravity conditions.

Table 2. Microgravity effects on solidification microstructure in metallic systems.

Alloy System		Solidification Characteristics	Grain Morphology	Microstructural Metrics
Al–3.6 wt% Bi–1 wt% Sn [40]	ISS microgravity vs. terrestrial	Suppressed macro-segregation and melt convection	Equiaxed dominant in μg; up to ~32% columnar in gravitational sample	Space sample shows equiaxed α-Al grains with no visible gas cavities; larger porosity and phase-segregated microstructure on Earth
Al–4, 10, 20 wt% Cu (grain-refined) [41]	ISS microgravity	Purely diffusive solidification, minimal convection	Entirely equiaxed dendritic grains in μg; columnar only under some Earth 1g conditions	No significant macrosegregation; equiaxed grain structure with no large porosity
Al–7 wt% Si [42]	ISS microgravity vs. 1 g	Diffusion-controlled growth, suppressed convection	Equiaxed and PCET observed in grain-refined samples	Microgravity eliminates buoyant melt flow, allowing analysis of CET and dendritic structure
Al–3.5 wt% Si [39]	Drop tube microgravity vs. 1 g	Elimination of thermosolutal convection	Columnar vs. equiaxed grain variations	Under μg, dendrites are more parallel and less randomly oriented; DAS and eutectic content differ
Al–10 wt% Cu (drop tube) [39]	Drop tube microgravity vs. 1 g	Suppressed convection aided diffusive solidification	Predominantly equiaxed dendrites	μg samples show differences in dendrite arm spacing (DAS) and solidification morphology
Model transparent alloys [43]	ISS microgravity	Pure diffusion-controlled growth	Uniform growth patterns without convective distortion	Transparent alloy studies indicate absence of convective flow significantly alters pattern formation

summarizes the reported effects of microgravity on the solidification microstructure of representative metallic systems. Microgravity generally suppresses convection, promotes equiaxed and more isotropic grain structures, and reduces macro-segregation and porosity compared to terrestrial conditions. Quantitative metrics such as grain size, dendrite spacing, and columnar-to-equiaxed transitions are included where available, providing clear benchmarks for modeling and understanding additive manufacturing under reduced gravity.

2.2. Powder Handling and Feedstock Management

As previously mentioned, powder-based additive manufacturing particularly powder bed fusion (PBF) and blown-powder directed energy deposition (DED-powder), faces some of the most severe constraints when transferred from Earth to microgravity. In terrestrial environments, powders remain stable due to gravitational settling, frictional contacts, and predictable flow properties governed by bulk density and cohesion. Under microgravity, however, powders behave more like fine-grained, weakly bound aerosols, characterized by uncontrolled lofting, electrostatic adhesion, and fluid-like movement when disturbed. These behaviors create significant operational and safety challenges inside spacecraft, which is why current in-orbit additive manufacturing platforms rely mainly on wire or filament feedstocks instead of loose powders.

In powder bed systems, the absence of gravity prevents uniform spreading, causes particles to cluster via van der Waals and electrostatic forces [44], and eliminates the stabilizing downward force needed for consistent recoating. Attempts to recreate the powder bed in reduced gravity, such as parabolic flight experiments and drop tower studies [45], show that powders tend to clump, float, or migrate toward charged surfaces, making layer definition and surface flatness nearly impossible without mechanical confinement. Pressurized chambers with forced airflow can mitigate lofting, but they introduce new risks: contamination of life-support systems, inhalation hazards, and uncontrolled particle trajectories during laser or electron interactions [46]. As a result, fully open powder bed processes are currently incompatible with ISS operations and future crewed platforms unless powders are fully enclosed and mechanically constrained.

Feedstock delivery in powder-blown DED systems faces similar barriers. Powder streams rely on gas flow dynamics whose behavior changes drastically in low pressure and microgravity. Reduced drag forces and modified turbulence cause poor powder focus, lower deposition efficiency, and erratic particle trajectories [47]. Vacuum conditions further complicate matters by altering gas expansion rates and reducing the carrier gas’s ability to collimate the powder plume, leading to overspray and contamination of optics or nearby equipment. These limitations highlight the need for tightly sealed nozzles, magnetic or electrostatic focusing systems, or mechanical auger-based delivery concepts suitable for lunar or Martian partial-gravity environments.

Given these constraints, most near-term and demonstrated in-space manufacturing platforms favour feedstocks inherently stable in microgravity, including: metal wire (used in ESA/Airbus Metal 3D Printer [48]), polymer filament (AMF [49], Zero-G Printer [50]), UV-curable resin cartridges (CMM [51], DCUBED extrusion [51,52,53]) and pellet or granule feeders embedded in sealed chambers. A suitability analysis of different AM feedstock types is provided in Table 3.

These feedstocks avoid the risks associated with particulate dispersion [54] and allow precise mass flow control under altered gravitational conditions. Nevertheless, sealed powder modules remain an active research direction for ISRU-relevant manufacturing on the Moon or Mars, where gravity, though reduced, is sufficient to enable powder behavior closer to Earth norms. Future developments may include hybrid mechanical–electrostatic spreaders, microgravity-tolerant recoaters, and regolith powder handling systems optimized for partial gravity.

Table 3. Suitability of different AM feedstock types under microgravity and vacuum conditions.

Feedstock Type	Examples/Systems	Microgravity Stability	Key Advantages	Major Limitations	Overall Suitability
Metal powders	LPBF, powder-DED [55]	Very low	High material versatility; fine feature resolution	Powder lofting, electrostatics, safety hazards, recoating impossible, contamination risk	Poor
Regolith powders	ISRU sintering, lunar PBF concepts [56]	Low (micro-g), Moderate (partial-g)	In situ resource utilization, high-temperature stability	Needs vacuum containment; poor flow in micro-g; cohesion; dust hazards	Low–Moderate (depends on gravity)
Metal wire	ESA Metal Printer (Airbus/AddUp [57], WAAM	Very high	No loose particles, controlled feed, good for microgravity, sealed melt zone	Limited alloy selection; lower resolution	Excellent
Polymer filament	Zero-G Printer, AMF [49,58]	Very high	Fully stable in micro-g; low-power; safe operation	Limited thermal/mechanical properties	Excellent
Polymer pellets or granules	FDM pellet extruders (ground), potential ISS designs [59]	High when enclosed	Lower-cost feedstock; scalable material supply	Must be fully enclosed; pellet bridging possible	Good
UV-curable resins	CMM [51,60], DCUBED UV extrusion systems	High (sealed cartridges)	No powder handling; fine resolution; adaptable chemistry	Bubble retention; cure kinetics modified in micro-g	Excellent
Pre-ceramic polymers	CMM ceramic green-body printing [60]	High	Ceramic component capability; stable flow	Thermal decomposition gases may not escape in micro-g	Very good
Thermoset pastes/viscous slurries	Regolith geopolymer extrusion [61]	Moderate–High when confined	ISRU potential; no powder lofting	High viscosity; gas entrapment; water management	Good

Table 3. Suitability of different AM feedstock types under microgravity and vacuum conditions.

Feedstock Type	Examples/Systems	Microgravity Stability	Key Advantages	Major Limitations	Overall Suitability
Metal powders	LPBF, powder-DED [55]	Very low	High material versatility; fine feature resolution	Powder lofting, electrostatics, safety hazards, recoating impossible, contamination risk	Poor
Regolith powders	ISRU sintering, lunar PBF concepts [56]	Low (micro-g), Moderate (partial-g)	In situ resource utilization, high-temperature stability	Needs vacuum containment; poor flow in micro-g; cohesion; dust hazards	Low–Moderate (depends on gravity)
Metal wire	ESA Metal Printer (Airbus/AddUp [57], WAAM	Very high	No loose particles, controlled feed, good for microgravity, sealed melt zone	Limited alloy selection; lower resolution	Excellent
Polymer filament	Zero-G Printer, AMF [49,58]	Very high	Fully stable in micro-g; low-power; safe operation	Limited thermal/mechanical properties	Excellent
Polymer pellets or granules	FDM pellet extruders (ground), potential ISS designs [59]	High when enclosed	Lower-cost feedstock; scalable material supply	Must be fully enclosed; pellet bridging possible	Good
UV-curable resins	CMM [51,60], DCUBED UV extrusion systems	High (sealed cartridges)	No powder handling; fine resolution; adaptable chemistry	Bubble retention; cure kinetics modified in micro-g	Excellent
Pre-ceramic polymers	CMM ceramic green-body printing [60]	High	Ceramic component capability; stable flow	Thermal decomposition gases may not escape in micro-g	Very good
Thermoset pastes/viscous slurries	Regolith geopolymer extrusion [61]	Moderate–High when confined	ISRU potential; no powder lofting	High viscosity; gas entrapment; water management	Good

2.3. Defect Formation, Dimensional Stability, and Monitoring Under Microgravity

Deviations from terrestrial physics complicate process predictability and create strong motivation for advanced simulation and real-time monitoring approaches in the case of defect formation in microgravity. Without gravity-assisted heat and mass transport, melt pools become more surface-tension-dominated, demanding high-fidelity multi-physics models capable of reproducing convection-free thermal fields, altered wetting behavior, and modified defect migration patterns. Simultaneously, monitoring techniques are increasingly important for detecting gas entrapment, insufficient fusion, thermal runaway, or layer delamination in situ. Together, modeling and monitoring provide the base for autonomous defect detection and closed-loop control. These capabilities are crucial for ensuring reliability in future sustained in-space manufacturing ecosystems. A short example of how CFD can be used to simulate possible defects is depicted further in Figure 3.

Noori et al. [31] expanded this understanding using a CFD framework for laser-directed energy deposition under lunar, microgravity, and vacuum conditions. Their results showed that as gravitational acceleration decreases, buoyancy becomes negligible, and surface tension gradients and recoil pressure become the primary drivers of liquid metal motion. This work clearly illustrates the transition from mixed convection to surface-tension-dominated regimes. Figure 7 presents the results of the study, where (Figure 7a) represents the calculation settings and (Figure 7b) the experiment and predicted results.

3. Qualification, Standards, and Verification

3.1. AM Standards

Additive manufacturing is increasingly used in the space industry for components ranging from brackets and fuel-system elements to heat exchangers and structural parts. However, the extreme conditions of space, vacuum, radiation, thermal cycling, vibration, and long-duration reliability requirements create challenges not addressed by conventional, terrestrial AM standards. Existing ISO/ASTM methodologies such as the ISO/ASTM 52900 series [62] provide general terminology and broad process definitions but are not tailored to space environments. Agencies such as NASA and ESA complement these efforts with their own internal guidelines, including NASA STD 6016 [63] for materials testing, NASA process specifications for aerospace alloys, and ESA’s ECSS engineering standards. However, these frameworks are not internationally harmonized and differ in their testing requirements, which points to a clear need for unified, space-specific additive manufacturing standards.

Table 4. Contrast between space-specific AM needs and the more general standards.

Aspect	Space-Specific AM Needs	ASTM F42/ISO/ASTM 52900
Environment	Must account for vacuum, microgravity, extreme thermal cycles, radiation	Terrestrial conditions; no explicit guidance for space environments
Material Performance	Long-term stability under space conditions, outgassing, radiation resistance	General material behavior; mechanical properties under standard Earth conditions
Part Qualification	Critical for flight hardware; requires formal verification for launch and operation	Provides general testing and characterization methods, not space certification
Design Considerations	Launch loads, thermal stresses, fatigue under orbital conditions	Focuses on design for manufacturability, but not space-specific loads
Process Control/Reproducibility	Must guarantee reliability across multiple builds for mission-critical parts	Emphasizes repeatability and process parameters but under terrestrial manufacturing tolerances
Inspection/Testing	Non-destructive testing for space-qualified components; verification in vacuum/thermal conditions	Recommends general testing methods (CT, mechanical tests) but not tailored to space flight

illustrates that space-based AM has far stricter and more specialized requirements than those covered by general ASTM/ISO standards. While existing standards address terrestrial materials, testing, and process control, they do not account for the vacuum, microgravity, radiation, and extreme thermal loads relevant to space. As a result, space AM demands additional qualification, environmental testing, and reliability measures beyond what current standards provide.

Beyond environmental challenges, AM itself introduces additional complexity because it couples geometric fabrication with material property formation in a single process. Small variations in laser wavelength, powder size distribution, scan speed, or alloy chemistry can significantly alter material outcomes, sometimes detectable only through advanced inspection such as ultrasonic imaging [16]. While ASTM Committee F42 has created process-specific standards (e.g., F3187 [64], F3303 [65]), these were developed for homogeneous feedstocks, stable terrestrial conditions, and commercial AM systems, limiting their applicability to mission-critical spaceflight hardware.

To mitigate these risks, NASA has developed space-specific AM guidelines, including STD-3716, SPEC-3717, and NASA-STD-6030 [1], which extend beyond ASTM’s scope. These documents define foundational process-control requirements such as AM process qualification, machine operation and calibration, personnel training, material characterization, and the establishment of material property limits, design values, and statistical process control (SPC) monitoring [65,66,67]. They also address powder and support structure removal, risk analysis, and the evaluation of creep, fatigue, and other long-term behaviors relevant to space environments. Space resource utilization, especially AM using lunar or Martian regolith, introduces further challenges due to heterogeneous particle size and composition, requiring rigorous documentation of feedstock chemistry, process parameters, environmental conditions, and inspection methods [68,69]. While powder bed fusion technologies hold promises for off-Earth manufacturing, achieving consistent material properties remains difficult, reinforcing the need for robust, space-specific AM standards.

Building on this guideline framework, NASA and ESA implement structured qualification pathways that ensure flight-ready reliability. These pathways define material performance limits, including statistically derived properties such as yield strength, ultimate tensile strength, elongation, fatigue limits, and creep behavior. These values are determined from extensive mechanical testing of multiple specimens printed under controlled conditions, capturing process variability and material anisotropy [12,65,66]. Mechanical tests typically include tensile, compression, fatigue, fracture toughness, and creep evaluations in multiple orientations. In parallel, non-destructive evaluation (NDE) techniques such as X-ray computed tomography (CT), ultrasonic inspection, dye penetrant, or radiography are used to detect internal defects such as porosity, cracks, or inclusions. Qualification relies on a combined mechanical-testing and NDE approach to ensure that components meet mission-specific requirements for loads, environments, and operational lifetime. For ISRU-based AM on the Moon or Mars, these pathways must adapt to accommodate variable feedstocks and novel processes.

3.2. Certification Strategies and Space-Based Verification

Certification of AM hardware follows a hierarchical logic that progresses from coupons to subcomponents and finally to the functional part. At the sample level, simple test specimens define the fundamental material and process behavior, providing the data needed to calibrate and validate material and process models. Subcomponent tests then introduce realistic geometrical features to evaluate scaling effects, defect evolution, and load-bearing behavior under more representative conditions. Only after these steps are successful is the full functional part assessed, demonstrating that the entire manufacturing path, from feedstock to final geometry, meets performance and safety requirements. This staged approach reduces uncertainty, builds model confidence, and forms the backbone of model-based and DT-enabled certification strategies.

Certification and verification of AM hardware follow a structured, multi stage approach that moves from test coupons to subcomponents and ultimately to the final functional part. Verification is intended to demonstrate that a product or system meets its design requirements [70] and includes both qualification and acceptance activities. Qualification is performed once to confirm that the design and manufacturing process are controlled, repeatable, and reliable, while acceptance ensures that each individual unit is free of defects, meets defined performance margins, and is ready for use [71]. At the coupon level, simple test specimens are used to characterize fundamental material behavior and process performance. These data support the calibration and validation of material and process models and enable statistical control of the manufacturing process. Subcomponent testing then introduces more realistic geometries, allowing assessment of scaling effects, defect development, and load-bearing behavior under representative conditions. Only after these stages are successfully completed is the full functional part evaluated, demonstrating that the entire manufacturing workflow, from feedstock to final geometry, satisfies performance, quality, and safety requirements [72,73]. Conventional qualification approaches for additive manufacturing can require thousands of tests and may span 10 to 15 years, reflecting the complexity of material properties that vary with build orientation and location [74]. To support ongoing process monitoring, auxiliary test features such as tensile or fatigue coupons, density cubes, or powder storage pyramids are often manufactured alongside flight hardware [75,76,77]. This structured and model-informed approach, particularly when integrated with digital-twin-based frameworks, reduces uncertainty, strengthens confidence in predictive models, and forms the foundation for more efficient and reliable certification strategies for additive manufacturing hardware.

Ground-to-orbit correlation campaigns, such as ESA’s LAMP and NASA’s ISM roadmap, provide essential data linking terrestrial qualification and in-space performance. By comparing sample, subcomponent, and functional-part behavior on the ground with measurements in orbit, these campaigns validate predictive models and DTs, refine uncertainty quantification, and support reliable certification of AM hardware for spaceflight.

The European Space Agency’s LAMP project [49] brings into practice laser-based metal additive manufacturing in space, producing high-precision components like brackets, tools, and antenna parts for orbiting platforms and future habitats. LAMP is advantageous for space applications because it minimizes waste, supports topology-optimized lightweight designs, and eliminates cutting tools. Challenges include powder handling in microgravity and thermal management without atmospheric convection, requiring enclosed chambers, controlled gas flows, electrostatic manipulation, and precise heat dissipation [78]. In situ resource utilization (ISRU) is being explored to use lunar or Martian regolith as printable material, with hybrid sintering and laser-melting systems and ruggedized platforms developed for extreme environments.

NASA’s In-Space Manufacturing (ISM) program, managed by the Marshall Space Flight Center, develops technologies for on-demand production of mechanical and electrical components during exploration missions, both in transit and on planetary surfaces [79]. The initiative addresses the limitations of current ISS logistics, which rely on Orbital Replacement Units (ORUs), by enabling Earth-independent manufacturing for deep-space operations where resupply is delayed or impossible. ISM focuses on reducing logistics requirements, mitigating operational risks, and supporting the fabrication of large or fragile structures that are difficult to launch fully assembled. Its portfolio incorporates materials, processes, and infrastructure for non-terrestrial manufacturing by aligning them with strategic priorities identified by NASA’s STMD as well as other government agencies, including DARPA, NIST, NSF, AFRL, the Space Force, and NRL [80].

Figure 8 illustrates the four primary pursuit areas of the ISM program, showing how these strategic activities address both NASA shortfalls and broader national challenges.

The integration of digital twins and model-based verification represents a transformative advance in the qualification and certification of AM hardware, particularly for space applications. By combining hierarchical certification, from coupons to subcomponents and functional parts, with predictive, multi-scale digital twin models, manufacturers can reduce uncertainty, optimize process parameters, and accelerate qualification timelines. Ground-to-orbit correlation campaigns, such as ESA’s LAMP and NASA’s ISM roadmap, provide critical empirical data that validate predictive models, refine uncertainty quantification, and bridge terrestrial and in-space performance. Initiatives like NASA’s IMQCAM and ESA’s GSTP further illustrate how DT-enabled frameworks can support reliable, first-time-right additive manufacturing processes, while ISM and LAMP demonstrate the implementability of producing components on-demand in orbit or on extraterrestrial surfaces. Together, these efforts point toward a future in which AM hardware can be certified efficiently, with high confidence in performance, safety, and resource efficiency, enabling sustainable human exploration and manufacturing beyond Earth.

4. Digital Twins

An emerging component of modern qualification frameworks is the integration of digital twins (DTs) and model-based verification. Originally developed by NASA to monitor satellites and simulate configuration changes, DTs are virtual counterparts of physical systems [67]. In additive manufacturing (AM), DTs provide high-fidelity representations of geometry, materials, process history, and environmental conditions, enabling predictive evaluation of performance under thermal cycling, launch loads, radiation, and microgravity. They also allow optimization of process parameters, correlation of defects and residual stresses with final part performance, and reduced reliance on extensive physical testing. When embedded in AM operations, DTs can form a closed-loop feedback system, synchronizing real-time monitoring, simulation, and adaptive control, thereby mitigating the uncertainties in process physics and material behavior that have historically hindered qualification and certification [68].

Pioneering work by DebRoy, Yang, and colleagues [69,70,71] established a foundational framework combining mechanistic modeling, sensing, statistical analysis, and machine learning for early-generation DTs in AM. They proposed a comprehensive framework that integrates mechanistic modeling, sensing and control, statistical modeling, and data-driven machine learning for a first-generation digital twin of 3D printing. Despite their promise, current research shows notable gaps. Many studies focus on individual components, such as geometry prediction or structural health monitoring, without achieving the interdisciplinary integration needed for full DT-enabled AM [81]. Significant interactions among material behavior, thermal management, and structural integrity remain insufficiently explored [82]. Similarly, although AI and machine learning are increasingly recognized as critical for process optimization and defect mitigation, their integration within DT-driven AM systems is still limited.

Useful insights can be drawn from established DT applications in aerospace and conventional manufacturing, where DTs usually represent the operational behavior of an existing and largely static asset. Additive manufacturing, however, introduces fundamentally different challenges [83]. In additive manufacturing, the physical twin evolves continuously, with its geometry, temperature distribution, and material state changing layer by layer. This demands DT models capable of updating boundary conditions and computational meshes in real time, far exceeding the complexity of simulating fixed systems. While DTs in industrial applications typically focus on performance metrics such as stress or fatigue, DTs for additive manufacturing must also capture the underlying physics of material transformation, including melting, fusion, and solidification [84]. The objective extends beyond predicting part failure to actively controlling microstructure formation in real time, making this a uniquely difficult scientific challenge.

Looking ahead, Tudorache et al. [85] identify three transformative directions for DTs in AM: (1) “first-time-right” DTs, prescribing optimal, machine-specific process parameters before production; (2) “certified part” DTs, which integrate multiscale models and in situ data to ensure reliable microstructure and performance predictions for safety-critical components; and (3) universal DTs, scalable across machines via standardized models, data formats, and communication protocols. These capabilities directly address qualification challenges by reducing uncertainty, improving predictability, and enabling model-based certification.

NASA and ESA are increasingly adopting digital-twin-based verification to accelerate qualification of additively manufactured metal parts for space applications. A key initiative is NASA’s IMQCAM [86], a Space Technology Research Institute co-led by Carnegie Mellon University and Johns Hopkins University, with multiple academic and industrial partners [87]. IMQCAM’s framework integrates multi-physics, multi-scale models linking feedstock, AM process parameters, microstructure, and structural response to predict location-specific fatigue life and support model-based certification [88]. SwRI translates these models into industrial tools, enabling application to flight-critical hardware. Similarly, ESA’s GSTP [89] is developing predictive digital twin frameworks for advanced manufacturing, including AM, to support real-time process monitoring, adaptive control, and future certification in European space-manufacturing workflows.

In addition to digital twins (DTs), emerging artificial intelligence techniques—particularly large language models (LLMs)—are poised to transform additive manufacturing (AM) process optimization and qualification. While DTs rely on physics-based and data-driven models to predict process outcomes, LLMs provide complementary capabilities by extracting high-level patterns from heterogeneous data sources, including multi-sensor measurements, simulation outputs, standards, and unstructured literature. Through semantic reasoning, LLMs can identify hidden correlations between process parameters, defects, and performance metrics, and support near-real-time decision-making by proposing adaptive process adjustments, reducing reliance on empirical trial-and-error approaches [90,91].

More broadly, LLMs are emerging as enabling technologies for Industry 4.0 and the emerging Industry 5.0 paradigm, facilitating the transition toward smart, resilient, and human-centered manufacturing systems. Recent reviews indicate that LLMs enhance process optimization, data structuring, innovation, and human–machine interaction, while also introducing ethical and reliability considerations [91]. These models have demonstrated the potential to improve operational efficiency, redistribute human roles toward higher-level supervision, and accelerate innovation across digital manufacturing workflows.

In metal AM, the complexity and interdependence of materials, processes, and post-processing steps create fragmented knowledge that is difficult to access. Integrating LLMs with structured knowledge graphs allows natural-language querying to assess material–process compatibility, enforce design constraints, and provide design-for-AM guidance, delivering intuitive, transparent, and explainable decision support for AM design and planning—capabilities that are particularly critical for autonomous or remotely operated manufacturing in microgravity [92].

Recent experiments demonstrate the feasibility and utility of operating LLMs in space. Generative AI LLMs have been successfully deployed aboard the International Space Station (ISS) using the HPE Spaceborne Computer-2, performing onboard data processing, summarization, and mission support tasks without Earth-based connectivity, illustrating the viability of LLM processing in resource-constrained orbital environments [93]. Advanced multimodal LLM stacks, such as “Space Llama,” have been tested aboard the ISS to interpret text and visual inputs, enabling rapid access to technical documentation and problem solving in real time [94]. Recent studies on fine-tuning large language models (LLMs) for autonomous spacecraft control indicate that these models can process real-time mission telemetry and generate actionable control commands in simulated spacecraft environments, demonstrating their potential for high-level decision-making in space operations [95], while agentic LLM architectures show promise for high-level planning in space missions [96]. Additionally, LLM-based code generation for aerospace systems underscores the relevance of these technologies for future engineering workflows and autonomous in-orbit operations [97].

These developments suggest that LLMs could play a direct role in in-space additive manufacturing, supporting autonomous process control, dynamic adjustment of build parameters, and real-time defect mitigation in microgravity environments, complementing DT-based predictive models. Integrating LLMs with DTs could further reduce reliance on trial-and-error experimentation, improve uncertainty quantification, and accelerate the qualification and certification of AM parts. This integration is expected to be transformative for both terrestrial and space-based manufacturing, enabling safer, more efficient, and adaptive deployment of additive manufacturing in extraterrestrial environments.

5. Emerging Trends and Future Directions for Addressing Key Challenges

5.1. Autonomous and AI-Assisted Fabrication

Autonomous and AI-assisted fabrication has rapidly become a central pillar of in-space manufacturing, driven by the constraints of microgravity, communication latency, and the need for continuous operation with minimal crew intervention. In microgravity, the absence of stable gravitational settling, the amplification of small perturbations, and the sensitivity of thermal and mechanical dynamics require fabrication processes to operate within closed-loop, self-correcting control architectures rather than relying on human supervision. These constraints have accelerated the transition from operator-driven fabrication systems toward autonomous robotic platforms, AI-guided monitoring, and predictive digital twin frameworks that ensure repeatability, safety, and fault tolerance across the full manufacturing life cycle.

Recent AI-assisted additive manufacturing (AM) systems increasingly adopt layered cyber–physical architectures that explicitly integrate sensing, learning, decision-making, and actuation. A representative architecture consists of: (i) a perception layer combining optical cameras, infrared thermography, acoustic emission, and force/torque sensing; (ii) a learning-based inference layer employing deep neural networks for state estimation and anomaly detection; (iii) a decision layer that fuses learned models with physics-based constraints using model predictive control (MPC) or reinforcement learning (RL); and (iv) an actuation layer implementing adaptive toolpath, feed rate, or thermal control in real time. Such architectures have been demonstrated in recent closed-loop AM platforms, where convolutional neural networks (CNNs) process layer-wise images to detect porosity or geometric deviation, while recurrent neural networks (LSTM/GRU) track temporal melt pool dynamics and predict defect accumulation across layers [98,99,100,101].

Recent hardware demonstrations illustrate the readiness of autonomous robotic manufacturing systems for orbital deployment. The C3 Space-Wire Bender payload establishes a complete end-to-end autonomy framework for robotic forming operations in low Earth orbit, integrating multi-modal sensing, automated feed-rotate-bend-cut sequences, and fault detection with autonomous recovery logic [102]. Operating without real-time human oversight, the system detects anomalous loads, mechanical misalignment, or feed irregularities and initiates safe-mode routines, demonstrating autonomous decision-making imperative for fabricating structural components under microgravity. The payload also incorporates vacuum-rated sensing, thermal modeling, and outgassing-aware control, offering a model for future on-orbit manufacturing units capable of executing complex multi-step tasks independently of ground control.

Several recent case studies move beyond conceptual discussions and demonstrate AI-enabled closed-loop AM in operational settings. In directed energy deposition systems, deep-learning-driven monitoring pipelines have reduced defect rates by dynamically adjusting process parameters based on real-time sensor fusion, achieving up to 30–50% improvement in dimensional accuracy compared to open-loop operation [103,104,105]. For material extrusion, vision-based neural networks combined with MPC have enabled automatic compensation for filament flow instability and thermal drift, maintaining part fidelity under changing environmental conditions [106,107]. Although these demonstrations are largely terrestrial, the underlying AI control strategies—sensor fusion, real-time inference, and autonomous decision-making—map directly onto in-space manufacturing constraints, where similar disturbances arise from microgravity-induced flow instability and spacecraft motion.

Parallel advances in AI-driven monitoring and control have expanded the precision and robustness of AM processes in space-relevant environments. Machine-learning-enhanced defect detection, as demonstrated in recent AM process monitoring research, uses real-time computer vision, sensor fusion, and predictive analytics to identify layer-wise anomalies, melt pool instabilities, and geometric deviations before they propagate [108]. These systems allow closed-loop parameter adaptation, including feed rate, temperature, extrusion rate, and toolpath modulation—compensating for microgravity-induced perturbations such as oscillatory filament flow, altered melt pool convection, or thermal drift. Predictive ML models can forecast error accumulation trends based on historical and live data, allowing in-flight calibration and automated correction steps that maintain part fidelity despite dynamic spacecraft conditions.

Concrete algorithmic implementations show how AI enhances AM robustness beyond rule-based control. Convolutional neural networks (CNNs) trained on in situ imagery have been used to classify defects and monitor quality in additive manufacturing processes, enabling classification of layer-wise anomalies and supporting parameter adjustment during builds rather than relying solely on post-process inspection [109]. In extrusion-based AM, real-time vision-based CNN classifiers have been implemented to detect good-quality, under-extrusion, or over-extrusion conditions and automatically modify printing parameters such as flow rate in response to successive defect identifications [110]. These studies demonstrate the feasibility of integrating deep-learning-enabled monitoring with automated corrective strategies in AM [99]. Gaussian process regression and Bayesian neural networks have been used to provide uncertainty-aware predictions of melt pool geometry and thermal gradients, allowing controllers to balance productivity and quality under variable boundary conditions [111]. More recently, reinforcement learning frameworks have been coupled with reduced-order thermal models to autonomously tune laser power or extrusion temperature, outperforming static control laws in non-stationary environments [112,113]. These algorithmic approaches are directly relevant to microgravity AM, where stochastic disturbances, altered convection, and limited human intervention demand adaptive, learning-based control.

Autonomous fabrication is further shaped by spacecraft–fabrication interactions, where manufacturing operations directly influence spacecraft attitude, momentum, and structural stability. Analytical and numerical studies of ISM dynamics demonstrate that attitude control system (ACS) torque limits, angular momentum storage, and flexible-body responses can dominate fabrication time and quality, especially for large extruded or assembled structures [114]. AI-assisted control algorithms are therefore required to predict and regulate variable-mass rigid-body dynamics during material deposition, adjust toolpath timing to match ACS authority, and prevent control–structure interactions. For multi-spacecraft manufacturing missions, machine learning models can coordinate distributed fabrication schedules, optimize workload distribution, and minimize coupled dynamics that could destabilize formation-flying architectures, making coordinated robotic manufacturing feasible at scales beyond single-vehicle capability.

Beyond monitoring and control, digital twin frameworks have emerged as a key enabler of predictive, AI-assisted fabrication systems in microgravity. Originally developed for astronaut health applications, recent microgravity-oriented digital twin architectures demonstrate capabilities directly applicable to advanced additive manufacturing, including real-time data assimilation, multi-scale physics-based simulation, and uncertainty-aware predictive analysis [115]. These fabrication-oriented twins integrate physics-based thermal and mechanical models with data-driven surrogates, such as neural networks or Gaussian processes, enabling rapid prediction of process outcomes under dynamic and uncertain conditions [116,117]. Through Bayesian updating and learned residual models, the twin continuously refines its predictions using live sensor data, correcting for inaccuracies in the underlying physics models. This hybrid AI–physics approach allows anticipation of defects, thermal distortion, extrusion irregularities, and toolpath deviations, providing proactive, algorithm-driven control rather than reactive correction [118]. By continuously integrating sensor data from thermal fields, mechanical loads, process signatures, spacecraft motion, and environmental conditions, the digital twin maintains a synchronized virtual model that predicts deviations, recommends parameter adjustments, and ensures process continuity despite incomplete information or shifting mission constraints. These capabilities are particularly critical for in-space manufacturing, where rework is costly and continuous human supervision is limited.

Autonomy in in-space manufacturing also depends on human–machine interaction systems capable of supporting operations under communication delays ranging from seconds (Lunar Gateway) to minutes (Mars transit) or complete isolation in deep space. The SpaceCHI autonomy-level framework provides a scalable model for ISM operations, demonstrating how AI-supported decision systems, adaptive interfaces, and explainable recommendations must increase as missions move from Earth-guided environments to fully autonomous deep-space scenarios [119]. For ISM, Level 1 autonomy (Moon vicinity) corresponds to conditional autonomous fabrication with intermittent teleoperation. By Level 2 (Mars transit), fabrication systems must execute complex repairs, component production, and corrective routines without real-time support, using robust onboard AI guidance. Level 3 autonomy—required for deep-space exploration—demands self-learning, fault-tolerant manufacturing ecosystems, capable of diagnosing internal failures, adapting toolpaths to degraded hardware, and autonomously rerouting fabrication tasks under constrained resources.

Finally, emerging research on autonomous recycling and robotic upcycling underscores the growing convergence between fabrication, robotics, and AI-driven decision systems. Vision-based identification, manipulation under tumbling dynamics, and automated disassembly planning, key components of future circular in-space manufacturing architectures, depend on robust autonomy and AI reasoning [120]. As in-orbit manufacturing shifts toward Earth-independent operations, these multi-layered autonomous systems will permit spacecraft and orbital facilities to fabricate, repair, and recycle materials without continuous human supervision.

Over the last years, the AM literature has shifted decisively toward data-driven and AI-enabled fabrication, with experimental studies validating real-time monitoring, closed-loop correction, and autonomous optimization across multiple AM modalities. This trend directly supports the feasibility of AI-assisted AM for in-space manufacturing, as the same algorithms—CNN-based defect detection, Bayesian inference, reinforcement learning, and hybrid digital twins—address the dominant challenges of microgravity fabrication, including process instability, limited sensing redundancy, and delayed human oversight. Consequently, AI-assisted AM is no longer a conceptual extension but an experimentally grounded pathway toward reliable, autonomous manufacturing beyond Earth.

Taken together, these advances show that autonomous and AI-assisted fabrication is a fundamental requirement for scalable, sustainable, and resilient production beyond Earth. By combining robotic autonomy, AI-based process control, predictive digital twins, and human–machine interfaces designed for communication delays, next-generation ISM platforms can carry out complex manufacturing tasks under microgravity and deep-space constraints. These capabilities form the basis of a self-sufficient, adaptive manufacturing ecosystem essential for long-duration exploration missions and large-scale infrastructure development beyond Earth.

5.2. New Energy Sources and Compact Hardware

New energy sources and compact hardware architectures are reshaping the landscape of in-space manufacturing by replacing bulky, energy-intensive terrestrial equipment with lightweight, high-efficiency systems capable of operating autonomously in microgravity and on extraterrestrial surfaces. Because spaceborne fabrication must function under severe mass, volume, and power constraints, next-generation AM systems increasingly rely on semiconductor-driven microwaves, highly efficient laser-diode arrays, and solar-concentrator systems that provide high thermal flux without the logistical burden of consumables. These technologies not only reduce launch mass but also enable ISRU-driven material processing for the Moon and Mars, forming the energetic backbone of long-duration sustainable manufacturing infrastructure.

Laser-based systems remain the core of metal AM in space due to their directional energy delivery, compact packaging, and compatibility with microgravity-adapted processes. As outlined earlier in this review, experiments conducted using the Einstein-Elevator microgravity platform have shown that laser beam melting and laser metal deposition can be adapted to operate without conventional substrates. In these approaches, focused laser radiation is used to consolidate wire-fed or powder-fed material directly in free space [121]. The microgravity-compatible LMD process benefits from the directional nature of laser energy, enabling stable melt pool formation even in the absence of buoyancy-driven convection. The experiments also underline the relevance of miniaturized laser optics and vibration-insensitive beam delivery systems for AM in microgravity, where dynamic perturbations can adversely affect melt pool geometry. Compact high-power diode lasers and fiber-coupled emitters therefore represent an essential energy source for future on-orbit metal deposition platforms, where precise and localized heating must be achieved using minimal hardware footprint.

Microwave-based manufacturing technologies are gaining traction due to their extraordinary efficiency in processing ceramic, oxide-based, and regolith-derived materials. Recent work of Tsubaki et al. demonstrated that single-mode, semiconductor-driven microwave systems can achieve ultrarapid sintering of lunar regolith simulants without the use of microwave susceptors or thermal insulation, owing to the formation of intense, frequency-auto tracked electric fields within resonant cavities [122]. Operating at 2.45 GHz and 915 MHz, the system produced sintered regolith gravel at temperatures up to 300 °C lower than conventional furnace processing, with heating rates exceeding 700 °C/min. The semiconductor oscillators automatically adjusted frequency to match cavity resonance, compensating for regolith dielectric property changes during heating. This ability is a critical advantage for in situ lunar construction where thermal drift and vacuum conditions alter energy coupling. Unlike multimode microwave ovens, which require heavy SiC susceptors and thermal shields, the single-mode resonator enables direct sintering with no additional mass. This architecture forms the basis for compact in situ fabrication units for landing pads, radiation shielding, and structural components on the Moon, avoiding the logistical burden of transporting construction materials from Earth.

Solar-concentrator-driven thermal systems constitute a third major energy source for compact in-space fabrication hardware. Recent assessments of solar concentration technologies for space applications demonstrate that both Fresnel-lens and parabolic-dish concentrators can deliver high-flux thermal power for melting, sintering, and chemical processing, with system efficiencies strongly dependent on environmental exposure, material degradation, and orbital thermal cycling [123]. The review highlights that solar concentrators, although simple and mass efficient, require careful optical stability and thermal management to minimize misalignment, coating degradation, and dust accumulation, which are particularly relevant considerations for long term deployment on the Moon or Mars. Nevertheless, concentrators offer substantial advantages: complete independence from electrical power systems, scalability from handheld to meter-scale units, and suitability for regolith melting, glass production, and thermal post-processing. Such systems provide a promising pathway for future hybrid AM approaches that combine solar-driven bulk sintering with laser or microwave precision finishing.

Developments in microgravity research infrastructure further support the design of compact AM hardware. The actively driven Einstein-Elevator drop tower enables microgravity experiments at repetition rates exceeding 100 tests per day, providing a cost-effective platform for validating compact AM systems, laser–material interactions, and substrate-free processing under 10⁻⁶ g conditions [124]. This facility permits rapid iteration of prototype hardware, including small-footprint laser systems, powder-handling mechanisms, and hybrid energy sources, supporting accelerated development of deployable manufacturing units for space habitats and robotic surface missions.

Additional insights from recent AM hardware studies reinforce the trend toward highly integrated, energy-efficient processing units. Advances in laser–material interaction modeling, low-power beam control, and compact scanning mechanisms enable miniaturized metal AM systems that maintain build quality while reducing mass and consumption [125]. Meanwhile, research on atmospheric and vacuum thermal-processing units demonstrates the feasibility of compact, modular heating chambers suitable for on-orbit repair or component fabrication in small spacecraft [126]. In parallel, hybrid microwave–thermal processing strategies offer directions to scalable manufacturing architectures where local sintering, pre-heating, or energy-assisted shaping can be achieved with minimal electrical input [127].

Taken together, these developments show that next-generation in-space manufacturing will rely heavily on compact, electrically efficient, and environmentally adaptive energy systems. Semiconductor-driven microwaves, high-power diode lasers, and solar-concentrator platforms collectively form a versatile energy toolbox capable of granting AM processes ranging from metal deposition to regolith sintering. As missions expand into cislunar space and planetary surfaces, these lightweight, high-efficiency hardware solutions will play a central role in underpinning sustained off-Earth manufacturing and minimizing dependence on terrestrial resupply.

5.3. ISRU and Habitat Construction

In situ resource utilization (ISRU) has become a cornerstone of current lunar infrastructure strategies, as transporting construction materials from Earth is steeply expensive and logistically restrictive. The lunar surface is dominated by regolith rich in silicates, oxides, and metallic minerals, providing a viable feedstock for additive manufacturing and rapid construction processes. Experimental studies on binder-assisted manufacturing of lunar regolith simulant demonstrate that extrudable regolith–polymer composites can achieve sufficient mechanical strength for early-phase structural components, even under vacuum conditions and extreme temperature cycles, indicating that such materials can support initial shelter creation and protective barriers on the Moon [128].

Research on lunar regolith utilization via direct ink writing (DIW) has primarily focused on several material systems, including biopolymer binding, alkali activation, matrix transformation, and polymer-based curing or sintering approaches. Each system exhibits distinct advantages and limitations with respect to printability, mechanical performance, and environmental adaptability. In regolith simulant-based geopolymer composites, bioinspired 3D printing strategies have integrated carbon fiber and quartz sand reinforcements into geometric configurations such as honeycomb, triangular, wave, and rectangular architectures (Figure 9).

Despite these encouraging results, significant scientific and technological gaps remain before ISRU-enabled additive manufacturing can be considered robust and certifiable. A primary challenge lies in the inherent variability of extraterrestrial regolith, whose composition, particle morphology, and impurity content differ across landing sites, complicating process standardization and material qualification. In addition, limited understanding of long-term mechanical degradation under combined lunar environmental stressors—including radiation exposure, thermal cycling, and micrometeorite impacts—currently constrains confidence in structural reliability. The absence of established certification and verification protocols for ISRU-derived materials further represents a critical barrier to their adoption in safety-critical habitat and infrastructure applications.

Research on large-scale automated construction has identified contour crafting (CC) as one of the most promising fabrication methods for lunar environments due to its ability to extrude thick, load-bearing layers rapidly and with high geometric precision. CC combines robotic extrusion and trowelling to create smooth, continuous walls and other architectural elements using locally sourced materials. The process supports multi-material deposition, embedded conduits for utilities, and reinforcement strategies suited for compressive structures, which are advantageous in reduced-gravity environments where tensile capacity is limited [129].

Building on this foundation, lunar-oriented CC research funded by NASA has explored robotic deployment of regolith-based construction systems for producing large structural elements directly on the surface. The studies propose using robotic platforms to fabricate protective enclosures, landing pad surfaces, dust-mitigation structures, and load-bearing vaults using regolith that is either extruded, sintered, or melted. These investigations detail how robots can construct essential infrastructure such as landing pads, roads, blast walls, and support structures using locally processed regolith, with an emphasis on minimizing crew exposure, improving operational efficiency, and reducing mission risk [130].

While large-scale additive construction techniques such as contour crafting demonstrate clear potential for lunar habitat fabrication, their extension from experimental demonstrations to certifiable, long-lived structures remains unresolved. Key challenges include ensuring structural integrity and defect tolerance at architectural scales, managing anisotropic mechanical behavior inherent to layer-wise fabrication, and implementing reliable in situ inspection and non-destructive evaluation methods under reduced-gravity and vacuum conditions. Moreover, current studies largely lack validated pathways for structural certification, load verification, and long-term performance assessment, which are essential for human-rated surface habitats.

Thermal processing methods offer additional pathways for transforming regolith into durable construction materials. Investigations of vacuum thermal consolidation show that regolith can be progressively densified at high temperatures, producing ceramic-like materials suitable for structural applications. These processes take advantage of the absence of an atmosphere, which reduces convective losses and allows for more efficient heat transfer in vacuum conditions [131]. Complementary studies of regolith composites reveal their mechanical performance across a range of temperatures and processing conditions, demonstrating that high-strength components can be formed from regolith-derived ceramics and composites without reliance on Earth-imported binders [132].

Broader analyses of material selection, structural behavior, and environmental constraints further reinforce the viability of ISRU-based construction. These studies address the effects of thermal gradients, radiation exposure, micrometeorite impacts, low-gravity buckling behavior, and dust abrasion on structural performance. They also evaluate the performance of different construction materials, including molten regolith, sulfur-based concretes, and sintered ceramics, under lunar thermal cycling and mechanical loading, as well as their integration into large-scale, long-lived surface installations [133].

Taken together, these developments show that ISRU-enabled manufacturing approaches, including regolith composite extrusion, robotic contour crafting, thermal sintering, and molten regolith processing, offer a practical and increasingly mature route for constructing habitats and critical infrastructure directly on the lunar surface. By combining robotic automation with locally sourced materials, future lunar settlements can significantly reduce their reliance on Earth, enabling safer, faster, and more sustainable expansion of human presence beyond low Earth orbit

5.4. Closed-Loop In-Space Manufacturing and Recycling

The long-term viability of human presence in space requires engineering a manufacturing ecosystem capable of operating independently from Earth. Such an ecosystem must integrate additive manufacturing, recycling, material recovery, and on-orbit testing into a unified and resilient production chain. NASA’s On-Orbit Servicing, Assembly, and Manufacturing (OSAM) initiatives provide the conceptual and technological foundation for this transition. The OSAM-1 mission focuses on robotic servicing and refuelling, demonstrating that spacecraft can autonomously manipulate, repair, and extend the life of existing assets, setting the groundwork for orbital infrastructure capable of supporting in situ production and maintenance [134]. OSAM-1 introduces a robotic architecture designed for precision manipulation, autonomous targeting, and multi-step servicing operations, forming the foundation of a manufacturing ecosystem capable of monitoring, adapting, and correcting processes in microgravity without continuous human oversight.

Beyond servicing, the OSAM-2 mission (Archinaut One) expands these capabilities to structural manufacturing in space. OSAM-2 integrates additive manufacturing with robotic assembly, catalyzing the production of large structural elements directly in orbit, including deployable booms and support structures far exceeding the size constraints of launch vehicles [135]. By demonstrating that AM and robotic assembly can function as a single continuous process in microgravity, OSAM-2 establishes the principle that future spacecraft, solar arrays, and optical systems can be built, expanded, or repaired on-orbit using only compact feedstock transported from Earth. This hybrid manufacturing–assembly paradigm is central to achieving closed-loop production, where material usage, structure fabrication, and system deployment occur entirely off-Earth.

Within NASA’s technology roadmap, the development of integrated AM and recycling units is highlighted as a critical step toward reducing logistical resupply and minimizing waste accumulation. TechPort documentation identifies the need for continuous material loops, post-processing in microgravity, and autonomous inspection systems to validate on-orbit production, emphasizing how manufacturing, recycling, and non-destructive evaluation must converge to support sustainable operations [136]. This aligns with the increasing prioritization of circularity and material closure across in-space manufacturing efforts.

Recycling technologies form the second major pillar of a closed-loop ecosystem. Research on polymer recyclers for microgravity environments details how waste plastics can be processed into reusable feedstock for AM, facilitating the transformation of packaging materials, habitat components, and mission consumables into AM-ready filament without returning waste to Earth [137]. These studies demonstrate the efficacy of grinding, melting, and reforming polymers under space-relevant conditions, supporting the concept of a fully regenerative material loop. Additionally, work on in-orbit waste recovery proposes the collection, processing, and reintegration of materials from discarded payloads, tools, and structural fragments, outlining how reuse and remanufacturing can be integrated into station-based or free-flying manufacturing platforms [138].

Biological recycling pathways offer a promising parallel to mechanical recycling. Experiments demonstrating fungal biodegradation of polyurethane in microgravity confirm that microorganisms can be employed to transform certain waste polymers into usable byproducts, supporting long-duration missions that require low-energy, self-sustaining waste-conversion methods [139]. Such bioreactors could complement mechanical recyclers by targeting soft polymers and mixed-material waste streams, allowing a broader spectrum of mission waste to be reintroduced into the manufacturing pipeline.

The broader concept of a closed manufacturing ecosystem is reinforced by studies proposing collaborative networks of distributed fabrication units, demonstrating how integrated AM clusters, recycling nodes, and inspection systems can function collectively as a dynamic production network in space [116,140]. Within this framework, small, modular fabrication units can exchange materials, coordinate production based on resource availability, and distribute tasks across multiple platforms. This model mirrors terrestrial digital manufacturing ecosystems but operates autonomously and independently of Earth’s industrial base.

Taken together, these developments demonstrate the potential of in-space AM and closed-loop manufacturing for supporting human and robotic operations beyond Earth. However, critical gaps remain.

• Process robustness and repeatability: Current systems have been validated at demonstration scales, but the reliability of AM processes under continuous microgravity operation is unproven.
• Structural validation: Large-scale components fabricated in orbit require rigorous testing to ensure structural integrity, load-bearing capacity, and defect tolerance.
• Inspection and certification: In situ non-destructive evaluation methods are still limited, and formal certification pathways for human-rated structures do not yet exist.
• Integration of recycling pathways: Mechanical and biological recycling methods must be seamlessly integrated into AM workflows while maintaining material quality and safety.
• Autonomous operation and coordination: Distributed networks of fabrication units need robust autonomy, error detection, and adaptive control mechanisms to operate without continuous human oversight.

Addressing these gaps is essential to transition from feasibility demonstrations to operational, certifiable systems, enabling reliable infrastructure for sustained human presence in orbit.

Despite advances in space-based additive manufacturing, critical gaps remain. ISRU-based manufacturing is limited by feedstock variability, process repeatability, and lack of qualification frameworks, while orbital and large-scale habitat construction face challenges in structural validation, defect tolerance, inspection, and long-term durability. Addressing these gaps across both lunar and orbital domains is essential to translate experimental demonstrations into reliable infrastructure supporting sustained human presence from Earth orbit to the Moon.

5.5. Challenges and Research Gaps

Despite substantial progress, additive manufacturing continues to face key challenges that restrict its broader adoption, especially in safety-critical and extraterrestrial applications. Additively manufactured materials often exhibit pronounced anisotropy and strong dependence on build orientation due to their layer-by-layer fabrication. This results in heterogeneous microstructures, direction-dependent mechanical properties, and variable defect distributions, which complicate process control, limit the accuracy of mechanical property predictions, and hinder consistent part performance [141,142]. In microgravity environments, these challenges can be amplified: altered melt flow, powder handling, and solidification dynamics may produce microstructural differences that reduce ductility or increase brittleness compared to Earth-manufactured counterparts, raising concerns about the reliability and mechanical integrity of space-fabricated components [38]. Furthermore, thermophysical property measurements for high-temperature alloys, such as superalloys and metallic glasses, are difficult to obtain on Earth due to gravity-driven effects like convection, sedimentation, and contamination [143]. Long-duration microgravity experiments, such as those conducted on the ISS, are essential to produce benchmark-quality data that enable accurate material models and predictive simulations [144].

Space-specific process and automation challenges also remain significant. Extreme environments—including altered gravity, vacuum, temperature fluctuations, and radiation—interact with AM machines, processes, and materials in complex ways [11,19]. These conditions influence processes, machines, and materials in complex ways, requiring integrated frameworks that consider all parameters in tandem. Feedstock availability is another central challenge: transporting sufficient materials from Earth is impractical, and in situ resource utilization (ISRU) remains underdeveloped. Lunar and Martian regolith simulants vary widely in composition, and it is unclear whether they can be used directly or require chemical processing to produce printable feedstocks [9].

Process monitoring, automation, and digital integration in space-based AM remain limited because real-time feedback systems, adaptive control, and predictive digital twins are still at an early stage of maturity. Similar gaps exist in terrestrial additive manufacturing, where AI-driven optimization and digital-twin-based control are largely under development rather than widely deployed [145]. The certification and qualification of additive manufacturing components, particularly for safety-critical space applications, are both time-intensive and costly, often requiring extensive testing at the coupon and full-part level to ensure consistent performance and compliance with rigorous standards [146]. NASA’s Technical Standard 6030 provides some guidelines for evaluating part risk, establishing appropriate levels of process control, qualification, and inspection, and facilitating trust in AM-produced components [147]. However, further work is needed to integrate such standards with operational AM systems in orbit.

Environmental constraints introduce additional uncertainties. Meteoroid impacts threaten both unfinished structures and printing equipment, particularly during unsupervised early construction stages. Temperature extremes and near-vacuum conditions on Mars significantly affect material curing and binder behavior, necessitating a fully integrated design of machines, processes, materials, and habitat locations. Specifically, the combination of low reaction rates at low temperatures (≈2 °C) and rapid water evaporation under low pressures (0.01–0.1 bar) has been shown to produce unsatisfactory curing of printed specimens under Martian conditions [148]. The unique advantages of microgravity, such as enhanced surface tension and reduced convection, improve bonding, reduce void formation, and enable high-quality crystal growth, yet these benefits can only be realized with precise understanding of material behavior and process adaptation [149].

Logistical and operational gaps further hinder sustainable in-space AM. Feedstock storage, handling, and recycling systems are underdeveloped, and standardization across materials, processes, and data formats is insufficient, limiting interoperability and scalability. Energy consumption and thermal management pose additional constraints: many AM processes, including stereolithography and cement-based contour crafting, require substantial energy inputs that are limited in spacecraft or lunar habitats [150]. Limited access to space-based experimental platforms, which provide only short or infrequent microgravity windows, slows iterative testing and the development of reliable in-space AM processes [19]. Finally, economic and logistic viability analyses remain sparse, making it difficult to assess the practicality of large-scale AM for long-duration missions [151,152].

Addressing these gaps is essential for translating experimental demonstrations into operational, certifiable systems capable of supporting extraterrestrial habitats, spacecraft component production, and extreme environment operations on Earth. Importantly, each of these gaps is directly linked to the emerging trends and solutions outlined in Section 5.1, Section 5.2, Section 5.3 and Section 5.4:

• Autonomous and AI-assisted fabrication can mitigate variability in material behavior and improve process repeatability.
• ISRU and habitat construction address raw material supply limitations and reduce dependence on Earth-launched feedstock.
• Closed-loop in-space recycling tackles resource efficiency, waste management, and feedstock regeneration.
• New energy sources and compact hardware enable reliable, scalable AM under environmental and operational constraints.

Table 5. Major challenges for additive manufacturing (AM) in space.

Category	Challenge	Description	Potential Solutions
Material behavior	Microgravity effects	Materials can exhibit brittleness, anisotropy, and unpredictable thermophysical properties due to lack of gravity-driven convection, sedimentation, and void formation	Conduct long-duration microgravity experiments (ISS, orbital platforms) to generate benchmark data; develop predictive material models and digital twins
	Feedstock availability	Transporting sufficient material from Earth is impractical; in situ resource utilization (ISRU) is underdeveloped	Develop ISRU methods for lunar and Martian regolith; investigate binder synthesis from local materials; standardize feedstock properties
Process & Equipment	Temperature and pressure extremes	Low temperatures and near-vacuum conditions affect curing, binder behavior, and process stability	Design integrated machine–process–material systems; develop binders compatible with low pressure/temperature; optimize habitat location
	Process monitoring & automation	Limited real-time feedback, adaptive control, and predictive modeling; underutilization of AI/ML	Implement sensor networks, automated control, and AI/ML-driven process optimization; develop digital twin frameworks
	Energy and thermal management	AM processes are energy-intensive; spacecraft or lunar base power constraints limit scalability	Develop energy-efficient processes; optimize thermal control; explore solar or alternative energy sources
Environmental hazards	Meteoroid impacts	Threat to equipment, unfinished structures, and printed parts, especially during early construction	Design protective shielding for printers and habitats; implement autonomous repair or fail-safe mechanisms
Experimental & Research	Limited microgravity platforms	Parabolic flights, sounding rockets, and orbital missions provide short or infrequent testing opportunities	Increase access to long-duration microgravity experiments; develop Earth-based simulators that closely mimic space conditions
Certification & Safety	Qualification and certification	Extensive coupon- and part-level testing is required; time-consuming and costly	Develop physics-based predictive models to reduce testing; adopt NASA-STD-6030 framework for consistent risk assessment
Standardization & Data	Lack of standardization	Insufficient standards for feedstock, process parameters, and data formats hinder interoperability, digital twin integration, and reproducibility	Standardize data formats; integrate digital twin protocols
Economic & Logistical	Cost and logistics	High transport costs, limited storage, and immature recycling systems challenge sustainable in-space AM	Optimize payload efficiency; develop recycling and reuse systems; evaluate economic feasibility and mission logistics

outlines the major challenges of in-space additive manufacturing, from unpredictable material behavior in microgravity and limited feedstock availability to harsh environmental conditions and energy constraints. It also highlights lacunae in process monitoring, automation, and access to long-duration microgravity testing. The “Potential Solutions” listed in the table correspond to the emerging trends and future directions presented in Section 5: Autonomous and AI-assisted fabrication (Section 5.1), New energy sources and compact hardware (Section 5.2), ISRU and habitat construction (Section 5.3), and closed-loop in-space manufacturing and recycling (Section 5.4). This explicitly maps core challenges to technological approaches, showing how ongoing and future research can address critical gaps and enable robust, certifiable deployment of in-space additive manufacturing systems.

5.6. Outlook

The challenges discussed above show that progress in in-space additive manufacturing is limited less by individual technologies than by how these technologies are developed and integrated. Moving beyond proof-of-concept demonstrations will therefore require a shift toward coordinated research strategies that explicitly link process physics, system autonomy, qualification, and resource sustainability.

A first priority is a deeper understanding of process physics under sustained microgravity conditions. Although short-duration experiments have clarified several gravity-dependent mechanisms, they cannot capture the cumulative effects of altered melt flow, heat transfer, and solidification over realistic build times. Long-duration experiments, particularly those combined with detailed numerical modeling, remain essential for establishing reliable relationships between process parameters, microstructure, and material performance in space-fabricated metallic components. Without such data, predictive modeling and robust process windows will remain highly uncertain.

Equally important is the transition from supervised operation toward genuinely autonomous manufacturing. In-space AM systems must be capable of responding to process disturbances, material variability, and environmental fluctuations with minimal human intervention. This requires tighter integration of in situ sensing, data-driven control, and physics-based models than is currently available. While AI-assisted approaches show promise, their effectiveness will ultimately depend on the quality of underlying physical models and the availability of validated training data from relevant microgravity environments.

Qualification and certification are expected to remain among the main limiting factors for the operational use of additive manufacturing in space. Terrestrial standards provide a useful baseline but do not account for the combined influence of microgravity, vacuum, radiation, and long service lifetimes. Model-based verification and digital twin approaches offer a practical way to reduce reliance on extensive physical testing, provided that they are supported by experimental validation and clear uncertainty bounds. Future efforts should therefore focus on aligning emerging digital qualification methods with agency-specific certification requirements.

The long-term viability of in-space additive manufacturing will also depend on reducing dependence on Earth-supplied materials. ISRU-based feedstocks and closed-loop recycling systems are essential for scalable manufacturing beyond low Earth orbit, yet their interaction with AM processes remains poorly understood. Research in this area must address not only material extraction and processing, but also feedstock consistency, recyclability, and compatibility with autonomous manufacturing platforms operating under constrained energy budgets.

Taken together, these directions highlight the need for a system-level approach to in-space additive manufacturing. Progress in process physics, autonomy, qualification, and resource utilization is strongly interconnected, and advances in one area are unlikely to be effective without parallel development in the others. Coordinated experimental campaigns, shared data frameworks, and long-duration validation platforms will be critical for reducing uncertainty and accelerating the transition from experimental demonstrations to reliable, certifiable manufacturing capabilities for sustained human and robotic exploration.

6. Conclusions

Additive manufacturing in space represents a paradigm shift in how hardware is produced, qualified, and sustained beyond Earth, requiring a departure from terrestrial, gravity-dominated assumptions toward a fundamentally physics-driven understanding of processes under microgravity, vacuum, and extreme thermal environments. This review has shown that reduced gravity profoundly alters melt pool dynamics, heat transfer, solidification behavior, and defect evolution, invalidating direct transfer of terrestrial process windows and qualification practices. Suppression of buoyancy-driven convection, dominance of surface-tension-controlled flow, modified cooling rates, and altered microstructural development collectively redefine process stability and material performance in space-based additive manufacturing.

Across the reviewed experimental, numerical, and in-orbit studies, it is evident that reliable in-space manufacturing cannot be achieved through hardware deployment alone. Instead, it demands an integrated framework combining microgravity-specific process physics, robust in situ monitoring, and adaptive control strategies. Advances in diagnostic techniques, including thermal imaging, optical monitoring, and sensor fusion, together with high-fidelity multi-physics modeling, provide the foundation for understanding and mitigating defect formation in environments where post-process inspection and repair are severely constrained. These capabilities are particularly critical for powder-based processes, ceramic systems, and regolith-derived feedstocks, where material handling, gas entrapment, and thermal management remain significant challenges.

Qualification and certification emerge as persistent bottlenecks for the operational deployment of additive manufacturing in space. Existing ASTM/ISO standards, developed for terrestrial manufacturing, do not adequately address the coupled effects of microgravity, vacuum, radiation, and long-duration service conditions. Space-agency-specific frameworks partially fill this gap yet remain fragmented and resource-intensive. The review highlights model-based verification and digital-twin-enabled certification as promising pathways to reduce dependence on exhaustive physical testing. When grounded in validated physics-based models and correlated through ground-to-orbit experimental campaigns, digital twins offer a scalable approach to predict process outcomes, quantify uncertainty, and support risk-informed certification of mission-critical hardware.

Looking forward, the convergence of autonomous fabrication, AI-assisted process control, compact energy-efficient hardware, and closed-loop material cycles defines the trajectory of in-space additive manufacturing. Autonomous systems are no longer optional but essential for operations under communication latency and crew-time constraints. In parallel, ISRU-driven manufacturing and recycling architectures offer the only viable route toward sustainable long-duration missions, enabling habitat construction, infrastructure development, and logistics independence from Earth. However, these advances must be accompanied by harmonized standards, long-duration material durability data, and validated inspection and verification methodologies to ensure safety and reliability.

In summary, additive manufacturing in space is transitioning from isolated demonstrations toward an integrated manufacturing ecosystem. Progress will depend on continued investment in microgravity process physics, coordinated international qualification strategies, and the tight coupling of digital twins, autonomy, and experimental validation. Addressing these challenges will enable additive manufacturing to evolve into a reliable, certifiable, and indispensable capability for sustainable human and robotic presence beyond low Earth orbit.