4.3.1 Residual stresses

Residual stresses are caused by the unique thermal cycle of metal AM, which consists of rapid heating, cooling rates and melt-back involving simultaneous melting of the top material layer and re-melting of the underlying previously solidified layer [Reference Li, Liu, Fang and Guo15]. It is also worth noting that when residual stresses are concerned, one should bear in mind the distribution, direction and magnitude as well as the constrain conditions of the body’s surroundings whilst keeping in mind that residual stresses are developed and altered in nearly every stage of the manufacturing process [Reference Xie, Lv, Yang, Shen, Tian, Shuai, Chen and Zhao16, Reference Bartlett and Li17].

The primary mechanism behind residual stress formation in AM is driven by the non-uniform material expansion and contraction. Therefore, the thermal and mechanical properties of the material being processed would naturally influence the magnitude of the residual stress [Reference Xie, Lv, Yang, Shen, Tian, Shuai, Chen and Zhao16]. For example, materials with higher-yield strength tend to exhibit lower-residual magnitudes under the same processing conditions [Reference Bartlett and Li17]. A study comparing Invar 36, 316 stainless steel and Ti6Al4V found that Invar 36, which has a low coefficient of thermal expansion, had the lowest thermal stresses and thus residual stresses [Reference Yakout, Elbestawi, Veldhuis and Nangle-Smith18].

4.3.2 Porosity

Porosity is a common and critical defect in metal AM that directly affects mechanical strength, fatigue life and gas tightness of rocket engine components. It refers to voids or un-melting gaps trapped within the printed part typically arising from incomplete melting, gas entrapment or lack of fusion (LoF) between layers. The pore morphology, distribution and size depend significantly on the process parameters, such as laser power, scanning speed, hatch spacing and powder quality [Reference Oko, Mbakaan and Barki19].

LPBF, for example, can exhibit high levels of porosity, which can be characteristically sorted as a lack of fusion pores, gas pores and keyhole pores [Reference Wang, Ning, Zhu, Yang, Yan, Dun, Xue, Xu, Bose and Bandyopadhyay20]. Gas porosity is small spherical pores with sizes ranging from submicron to several tens of microns predominantly caused by the moisture in the powders, the gas trapped in the particles after gas atomisation and the dynamics of the PBF process [Reference Martucci, Aversa and Lombardi21, Reference Tan, Liu, Fan, Zhang, Yin and Zhang22].

In contrast to gas porosities, LoF porosities are rather big in size, typically characterised by irregular shapes and usually located at the melt pool boundary. They are mainly caused by insufficient energy involved during the PBF process, which results in an incomplete adhesion of the ongoing melt to the surrounding part [Reference Martucci, Aversa and Lombardi21, Reference Tan, Liu, Fan, Zhang, Yin and Zhang22].

Last but not least, a dangerous type of porosity is that of keyhole pores. If an excessive laser power is involved in the PBF process, the temperature of the melt pool might exceed the material boiling temperature, resulting in the material’s vaporisation. The recoil pressure caused by vaporisation pushes the surface of the liquid downwards, forming a vacuum zone also known as a keyhole [Reference Martucci, Aversa and Lombardi21]. Its entrapment in the solidification front results in the formation of undesirable pores in the final component [Reference Kouraytem, Li, Cunningham, Zhao, Parab, Sun, Rollett, Spear and Tan23].

Due to the formation of these pores in the material, the mechanical properties such as thermal conductivity and fatigue performance will change significantly.

For aerospace applications, the accepted porosity limitations can be categorised according to the criticality of the part. For high-criticality parts, an ideal maximum porosity of 0.1%; for medium-criticality parts, porosity up to 1% is considered acceptable; and for low-criticality parts, porosity levels of up to 2% are tolerated [Reference Ferreira, Brandão, Borille, Gonçalves, Leite and Ribeiro24].

4.3.3 Cracking

Cracks typically form due to the thermal stresses from rapid solidification and cooling in fusion-based AM processes. The two major types of cracking are solidification cracking and liquidation cracking. Solidification cracking occurs in the final stages of solidification, when the material is partially solid and partially liquid. The formation of dendritic structures during this stage leads to thin films of liquid metal along grain boundaries, which fail to accommodate contraction stresses and ultimately form cracks [Reference Herzog, Seyda, Wycisk and Emmelmann25].

On the contrary, liquidation cracking occurs in previously solidified material that partially melts during the previous layer’s deposition. This type of cracking often appears in the heat-affected zones [Reference Liu, Yu, Wang, Zhang, Li, Chen and Liu26]. Liquidation cracking is most prevalent in superalloys with high (Al +Ti) content, and unlike solidification cracking, the source of cracking is low-melting-point eutectics in the HAZ [Reference Liu, Yu, Wang, Zhang, Li, Chen and Liu26]. These solidified eutectics will be re-melted to form a liquid film when the reheated temperature is above $T_{E}$ , which will lead to intergranular cracking under the action of contraction force [Reference Liu, Yu, Wang, Zhang, Li, Chen and Liu26].

4.3.4 Distortion

Distortion is a significant challenge in metal AM and is particularly critical in aerospace applications, where dimensional accuracy and tight tolerances are vital for parts such as injectors, cooling channels and combustion chambers. Distortion is defined as the deviation of the AM part from its designed shape or dimensions [Reference Wang, Qiu, Jiao, Gu, Xie, Wang, Tang, Wei and Shen27], and it can occur either during the additive process or after the part is separated from the substrate [Reference Kumar, Kumaraguru, Kumar, Pandey and Wimpenny28]. In PBF, for example, severe distortion can interfere with the movement of the recoater blade, causing build failure.

The key factors influencing distortion can be broadly categorised into material properties, process parameters and part geometry. As discussed earlier, Mukherjee et al. [Reference Mukherjee, Zhang and DebRoy29] showed that alloys with lower heat capacity and higher thermal diffusivity tend to exhibit higher strain. Dunbar et al. [Reference Dunbar, Denlinger, Gouge, Simpson and Michaleris30] observed that Inconel 718 parts distorted 80.9% more than Ti6Al4V when built under identical conditions, while Denlinger et al. [Reference Denlinger, Heigel, Michaleris and Palmer31] reported that Inconel 625 exhibited at least twice the distortion of Ti6Al4V at the same heat input. These results highlight the strong dependence of distortion on material properties and phase transformations during solidification.

Process conditions also play a major role. Wu et al. [Reference Wu, Brown, Kumar, Gallegos and King32] found that distortion in prism structures decreased with increasing energy density, whereas Yakout et al. [Reference Yakout, Elbestawi, Veldhuis and Nangle-Smith18] reported that cantilever distortion increased with higher laser energy density. This suggests that the relationship between input energy and distortion is non-linear and material dependent. More generally, studies have shown that optimising laser parameters in SLM or nozzle temperature in fused deposition modelling (FDM) can reduce distortion to manageable levels [Reference Xie, Lv, Yang, Shen, Tian, Shuai, Chen and Zhao16].

Although distortion remains a critical barrier to achieving dimensional accuracy in AM, ongoing research into tailored process parameters, improved support design and preheating strategies shows promise in minimising its effects.

4.3.5 Anisotropy

Anisotropy describes the various orientation characteristics of a material [Reference Huangfu, Liu, Liu, Wu and Bai33]. It occurs due to the directional energy input and the encountered thermal gradients. Unfortunately, these characteristics can limit the practical applications of LPBF-produced components [Reference Shrivastava, Subramaniyan, Rao and Nagesha34]. According to Condruz et al. [Reference Condruz, Matache, Paraschiv, Frigioescu and Badea35] a material’s anisotropic behaviour is initiated due to seven factors, from which only the following three properties will be discussed due to their relevance to this review:

• • Crystallographic texture: Crystallographic texture exists when specific crystal orientations occur in a higher frequency when compared to a uniform or random distribution [Reference Weaver and Rosenthal36]. This leads to plastic anisotropy, which means the material deforms differently depending on the direction of the applied force.

• • Grain morphology: Grain morphology refers to the shape, size and orientation of grains within a metallic microstructure. It is one of the most critical microstructural characteristics that governs the mechanical performance of metals and alloys [Reference Wei, Wang, Zhang and Ramamurty37].

• • Lack of fusion: Lack of fusion is characterised by irregular shapes or sharp edges that occur when insufficient melting occurs [Reference Weaver and Rosenthal36]. The irregular defects act like stress concentrators, and combined with their alignment along the layer make it easy for cracks to propagate parallel to the layer structure.

Similar to distortion, these factors can be mitigated through targeted post-processing techniques such as heat treatment, hot isostatic pressing (HIP) and thermomechanical processing.

5.1.1 SLM

In SLM, three-dimensional components are fabricated directly from CAD data by selectively melting thin layers of metallic powder using a focused laser beam [Reference Hussein, Hao, Yan and Everson43–Reference Zhang, Li and Bai45]. Typical layer thickness ranges from 20 to 100 $\mu$ m, while laser spot diameters are commonly 20 $\mu$ m [Reference Nandhakumar and Venkatesan46]. The build chamber is filled with an inert gas, usually argon, to maintain oxygen levels below approximately 0.2%, depending on material requirements [Reference Na, Kim, Yang, Kwon, Park, Kim, Jung, Lee, Park and Kim47]. After each layer is scanned, the build platform lowers by one layer increment and a new powder layer is deposited, enabling fabrication of components with relative densities often exceeding 99.5% under optimised processing conditions [Reference Gong, Ye, Chi, Zhao, Wang, Xia, Du, Tian, Yu and Chen48].

The process is characterised by steep thermal gradients, typically in the range of $10^{4}$ to $10^{5}$ K/s (cooling rate) but is also material dependent [Reference Pauly, Wang, Kühn and Kosiba49]. These solidification conditions promote fine cellular or columnar microstructures with characteristic feature sizes in the micrometre range, frequently enhancing mechanical strength compared to wrought equivalents. For example, SLM-processed Ti-6Al-4V commonly exhibits yield strengths above 900 MPa in the as-built condition. However, the pronounced thermal gradients and cyclic reheating generate residual stresses that may approach a significant fraction of the material yield strength, necessitating stress relief heat treatments, typically conducted at 600 to 800 $^{\circ }$ C for titanium alloys.

Overall, the combination of high-dimensional resolution, near-full density and geometric freedom makes SLM particularly well suited to compact, high-performance rocket propulsion components.

5.1.2 EBM

EBM shares the same layer-by-layer fusion principle as SLM but uses an electron beam as the energy source [Reference Cho, Kawabata, Hayashi, Yasuda, Nakashima, Takeyama and Nakano50]. Unlike SLM’s inert atmosphere, EBM operates under high vacuum conditions [Reference Qin, Damodaram, Maity, Zhang, Yang, Wang and Prashanth51, Reference Alquaity and Yilbas52]. Although the vacuum environment limits oxygen contamination, a more significant advantage is that it enables substantially higher preheat temperatures of the powder bed prior to and during melting. Elevated preheating reduces thermal gradients between the melt pool and surrounding material, thereby decreasing cooling rates and mitigating residual stress accumulation [Reference Nagahari, Nagoya, Kakehi, Sato and Nakano53]. This reduction in thermal stress lowers the risk of distortion and cracking, particularly in alloys such as $\gamma$ -Ti-6Al, which are prone to stress-induced cracking under the rapid solidification conditions typical of SLM [Reference Zhai, Wang, Ng, Zhou, Nai and Wei54]. Consequently, EBM is well suited for components with large cross-sectional areas, such as turbine blades or nozzle supports, where residual stress management is critical. For very-fine, high-tolerance components requiring superior surface finish and dimensional resolution, SLM remains the preferred method.

5.2.1 WAAM

WAAM uses an electric arc to melt feedstock wire for layer-by-layer deposition [Reference Ralph58]. In propulsion, WAAM is used for large, structural parts such as fuel tank domes, nozzle skirts and jacket sections. WAAM can be implemented via gas metal arc welding (GMAW), gas tungsten arc welding (GTAW) or plasma arc welding (PAW) [Reference Ding, Colegrove, Mehnen, Ganguly, Almeida, Wang and Williams59]. Each offers different deposition rates, penetration depths and precision, as summarised in Table 2. Additionally, the built surface roughness (Ra) for WAAM is approximately 200 $\mu m$ but is ultimately dependent on the material and method used to produce the part [Reference Ding, Pan, Cuiuri and Li60].

Table 2.

Comparison of GTAW, GMAW and PAW in WAAM applications for rocket manufacturing

Table 2. Long description

A table comparing three welding methods: GTAW, GMAW, and PAW. The table has four rows and seven columns. Column headers are Feature, GTAW (TIG), GMAW (MIG), and PAW. Row labels are Electrode, Filler addition, Deposition rate, Weld precision, Penetration depth, Typical shielding, and Best for. Row 1: Electrode, Non-consumable tungsten, Consumable wire, Non-consumable tungsten. Row 2: Filler addition, Separate rod or wire, Continuous wire feed, Separate rod or wire. Row 3: Deposition rate, Material dependent, Material dependent, Material dependent. Row 4: Weld precision, Very high, Medium, High. Row 5: Penetration depth, Shallow to moderate, Moderate, Deep (keyhole possible). Row 6: Typical shielding, Argon/Helium, Argon + CO2 mix, Argon/Hydrogen. Row 7: Best for, Thin, precise aerospace parts, Large structural welds, Deep welds on thick aerospace components.

$^{{\rm a}}$ Deposition rates are material- and parameter-dependent. For example, aluminium and steel WAAM deposition rates range from 1 to 4 kg/h [Reference Williams, Martina, Addison, Ding, Pardal and Colegrove61].

5.2.2 DMD

DMD uses a high-energy laser to melt metal powder as it is deposited [Reference Caiazzo, Alfieri, Argenio and Sergi62]. It produces parts with better surface quality, minimal heat-affected zones, and less distortion compared to WAAM, making it suitable for intricate components like injector heads or cooling manifolds. However, DMD’s low deposition rate (about 1 kg/h) is a drawback compared to approximately 1-8 kg/h for WAAM [Reference Williams, Martina, Addison, Ding, Pardal and Colegrove61, Reference Wu, Pan, Ding, Cuiuri, Li, Xu and Norrish63–Reference Vimal, Srinivas and Rajak65], so WAAM remains preferred for large structures.

7.1.1 Combustion chambers

Rocket engine combustion chambers are subjected to extreme conditions, including exceptionally high temperatures and immense pressures. Traditional manufacturing of these components involves complex multi-part assemblies, often requiring welding and brazing when regenerative cooling channels are incorporated [Reference Kerstens, Cervone and Gradl85, Reference Gradl86]. This increases the number of potential failure points and stress concentrations, potentially reducing reliability.

AM overcomes these limitations by enabling the direct fabrication of combustion chambers with integrated cooling channels in a single build or significantly fewer parts. This improves thermal performance by using optimised cooling geometries that are difficult or impossible to achieve with conventional manufacturing methods [87].

SLM is the preferred AM process for combustion chamber applications due to its superior dimensional accuracy and surface finish capabilities, which are critical for pressure vessel requirements. SLM enables the fabrication of intricate regenerative cooling channels with precise geometries necessary for effective heat management [Reference Gradl, Protz, Cooper, Ellis, Evans and Garcia8]. Copper-based alloys, particularly GRCop-84 and GRCop-42, are commonly employed due to their excellent thermal conductivity properties, essential for heat dissipation in high-temperature combustion environments [Reference Ellis9].

However, AM for combustion chambers faces key limitations. SLM processes have build volume constraints, typically limiting build chamber dimensions to approximately 400 mm diameter [Reference Khorasani, Gibson, Veetil and Ghasemi40]. For larger combustion chambers, DED may be employed for outer jacket structures, though this often requires hybrid manufacturing approaches that combine AM with traditional machining for critical surfaces.

7.1.2 Nozzles

Nozzles are amongst the most challenging to manufacture using AM techniques due to complex thermal gradients, high pressures and significant mechanical load. In rockets, regenerative cooling is essential to maintain structural integrity, which requires incorporating complex internal channels for cryogenic propellants. Conventional manufacturing of these parts is very difficult and relies on machining, brazing and welding of multiple parts that can decrease the structural integrity of the part drastically due to the increased number of failure points.

AM unlocks unprecedented flexibility in rocket nozzle design. With AM, integrated regenerative cooling channels can be built directly into the nozzle structure, reducing assembly requirements and minimising leakage risk. This geometric freedom enables cooling channels that follow the nozzle’s thermal profile more effectively, improving thermal performance while reducing mass.

In related research on conformal cooling for injection moulds, which serves as an analogue for cooling channel efficiency, tapered cooling channels designed via AM reduced cooling time by more than 20% compared to traditional 8 mm straight-drilled channels [Reference Tan, Wang, Ma, Chen, Chen, Yang and Zhou88]. Furthermore, incorporating lattice structures into these channels improved heat transfer rates by up to 62.9% and reduced overall system mass due to optimised structures [Reference Kanbur, Zhou, Shen, Wong, Chen, Shocket and Duan89]. While these figures are from moulding applications, similar performance improvements are achievable in regeneratively cooled rocket nozzles because of the shared physics of internal cooling.

Since nozzles use cryogenic propellants for regenerative cooling, it is essential to consider the materials that are being used to produce them due to the corrosive nature of these propellants. Metals like the GRCop-42 and GRCop-84 developed by NASA are perfect for such parts because they were specifically designed for such harsh environments [Reference Gradl, Protz, Cooper, Ellis, Evans and Garcia8].

The second most important factor for the manufacturing of an AM part is the selection of the AM process. As for nozzles, the method used can depend on the size and complexity of the part. For small- to medium-sized nozzles with intricate designs, PBF methods like SLM and EBM are used. Whereas for larger nozzles, DED is preferred due to its significantly larger build size [Reference Khorasani, Gibson, Veetil and Ghasemi40, Reference IvánTabernero, Alvarez and Suarez64]. Wire-based AM methods are also being researched because it enables rapid and cost-effective production of large-scale channel wall structures, which is a key part of rocket engine nozzle manufacturing [Reference Gradl and Protz90].

In addition to supporting innovative design strategies, AM simplifies the production cycle by reducing part count. Nozzles that would traditionally be built from several joined segments can instead be produced as single components, which lowers the risk of joint-related failures and eases inspection. The reduced reliance on specialised tooling, combined with shorter lead times, makes AM particularly well-suited for iterative design and testing during propulsion system development. Nevertheless, certain challenges persist. The as-built surface finish often requires post-processing to meet aerodynamic requirements in the throat and expansion sections. Build orientation can significantly influence the quality of internal features, with cooling channels aligned parallel to the build direction being prone to rougher surfaces. Furthermore, the layer-wise nature of AM introduces anisotropic properties that may reduce durability under severe thermal cycling. These limitations will be explored in greater depth in the following section.

7.3.1 Turbopump components

Turbopumps are some of the most complex subsystems in LREs. Their primary role is to deliver cryogenic propellants from storage tanks to the combustion chamber at extremely high pressures and flow rates, often exceeding tens of thousands of revolutions per minute. These operating conditions expose the turbopump to severe challenges, including high centrifugal forces, vibration, cavitation and steep thermal gradients caused by cryogenic propellants on one side and hot gases or mechanical heating on the other [Reference Gradl, Protz, Cooper, Ellis, Evans and Garcia8]. The combination of rotating machinery, precision tolerances and harsh environments makes turbopumps among the most difficult components to design and manufacture.

When selecting materials for turbopump components, factors such as fluid compatibility, ignition risk and mechanical strength must be carefully considered. For example, aluminium and titanium cannot be used in oxidiser pumps due to ignition risks in a pure oxygen environment. Instead, alloys such as Inconel are favoured for their ignition resistance [Reference Zink, Bourdon, Junior, Sias, Kitsche and Wagner96]. For fuel pump impellers, which typically operate at higher speeds than liquid oxygen impellers, titanium is widely used because of its high strength-to-weight ratio. Stainless steel may also be used where tip-speed requirements are less demanding [Reference Zink, Bourdon, Junior, Sias, Kitsche and Wagner96].

AM offers options for both new manufacture and repair of turbopump hardware. For precision-critical components such as impellers and blades, SLM is preferred due to its ability to produce tight tolerances and complex curved blade geometries. In contrast, DED is used for repair applications, as it enables localised material addition to damaged regions without requiring full component replacement [Reference Caiazzo, Alfieri, Argenio and Sergi62]. DED is also suitable for large structural components like housings and casings, where deposition rate and repairability are more important than fine resolution [Reference Valente, Gundlach, Christiansen and Somers69]. This balance between SLM for new builds and DED for repairs or large structures resolves the practical contradiction in method selection for turbopump applications.

Beyond replacing conventional methods, AM enables several innovations in turbopump design. Complex impeller geometries, such as blades with variable thickness or integrated flow passages, can be manufactured directly without multi-part assembly or extensive machining [Reference Adiaconitei, Vintila, Mihalache, Paraschiv, Frigioescu, Vladut and Pambaguian97]. Topology-optimised housings allow significant weight reduction while maintaining structural integrity [Reference Blakey-Milner, Gradl, Snedden, Brooks, Pitot, Lopez, Leary, Berto and du Plessis2]. Internal cooling channels can be embedded into bearing housings or shaft supports, improving thermal management and reliability [Reference Kaur and Singh98, Reference Han, Zhang and Lee99]. Furthermore, AM reduces the number of fasteners and potential leak paths while simplifying assembly and inspection.

Compared to traditional manufacturing, AM provides turbopump designers with greater flexibility and speed. Impellers with optimised curvature can be rapidly prototyped and tested, mounting features can be directly integrated into casings and multi-component assemblies can be consolidated into single near-net-shape parts. These advantages, combined with the dual capabilities of SLM for high-precision new manufacture and DED for repair and large structures, make AM a transformative approach to turbopump development.

8.1.1 Support structures

Support structures are essential for AM processes, specifically for SLM or EBM methods, to stabilise overhangs, complex internal channels and thin walls. Removing these supports without damaging the part is challenging, especially for components like regenerative cooling channels in combustion chambers or intricate injector assemblies. Mechanical techniques such as wire EDM, cutting and abrasive blasting are common, while chemical dissolution is employed when metal supports are used in materials like aluminium or titanium alloys. Inadequate removal can leave residual stresses or micro-cracks, which may propagate under high-pressure or thermal loads, reducing component reliability.

Figure 4.

Support structures in AM parts [103].

8.1.2 Heat treatment

Rapid thermal cycles during AM introduce residual stresses and localised variations in microstructure, which can significantly affect mechanical performance [Reference Li, Liu, Fang and Guo15]. To address these effects, heat treatment processes including annealing, stress relief and ageing are employed to relieve stresses and homogenise the microstructure [Reference Laleh, Sadeghi, Revilla, Chao, Haghdadi, Hughes, Xu, De Graeve, Qian, Gibson and Tan104]. For nickel-based superalloys, solution treatment dissolves existing precipitates and homogenises alloying elements within the $\gamma$ matrix, the nickel-rich solid solution that forms the structural `background’ of the alloy. Subsequent ageing induces $\gamma$ ^′ (Ni $_3$ (Al,Ti)) and $\gamma$ ^′′ (Ni $_3$ Nb) precipitates within this matrix, which strengthen the alloy by preventing dislocation motion and enhancing yield strength, creep resistance, and high-temperature fatigue performance [Reference Gao, Zhang, Li, Gong, Yuan and Song105, Reference Zhang, Zhu, Chen, Qin and Wang106].

Beyond stress relief and precipitation strengthening, heat treatment also governs phase transformations, grain morphology and crystallographic texture, which directly influence thermal expansion behaviour, oxidation resistance and crack propagation under cyclic thermal and mechanical loading. These factors are critical in rocket propulsion components such as regeneratively cooled combustion chambers, turbine blades and injector assemblies [Reference Ruan, Rezaei, Yang, Gross and Xu107–Reference Wang, Mokadem, Rappaz and Kurz109]. Advanced heat treatment strategies, including tailored multi-step solution and ageing cycles, can further refine precipitate size, distribution and coherency with the matrix, optimising mechanical and thermal performance for specific load conditions.

Properly designed post-processing heat treatments serve as a crucial bridge between the as-built microstructure of AM components and the high-performance requirements of aerospace propulsion systems, ensuring that parts meet dimensional and mechanical specifications and maintain reliability under extreme operating conditions. By controlling both macro- and microstructural features, heat treatment enables AM components to achieve service lifetimes comparable to or exceeding conventionally manufactured counterparts.

8.1.3 Surface finishing

Surface roughness is a recurring challenge in AM due to the layer-wise nature of fabrication, partially fused powder particles and stair-stepping effects along inclined surfaces [Reference Maleki, Bagherifard, Bandini and Guagliano110, Reference Barari, Kishawy, Kaji and Elbestawi111]. These irregularities not only reduce dimensional accuracy but also act as stress concentrators, lowering fatigue resistance and potentially serving as initiation sites for cracks under cyclic thermal and mechanical loading [Reference de Jesus, Ferreira, Borrego, Costa and Capela112]. In propulsion applications, rough surfaces can disrupt fluid flow, increasing turbulence and local heat transfer rates, which may cause hot spots in regenerative cooling channels or affect injector spray uniformity [Reference Sikora, Bohdal, Tandecka, Kacalak and Mathia113].

To address these limitations, a range of finishing techniques is applied depending on geometry and material. Mechanical methods such as grinding, machining and abrasive blasting are effective for reducing roughness on accessible surfaces, but they are less suitable for complex internal features. Advanced approaches such as abrasive flow machining (AFM) and magnetic abrasive finishing (MAF) use abrasive particles suspended in a carrier medium or guided by magnetic fields to polish internal channels [Reference Davoodi, Taghian, Carbone, Saboori and Iuliano114]. While these techniques can achieve notable improvements in surface quality, they are sometimes associated with geometric distortion due to uneven or excessive material removal.

Another promising category is ultrasonic-assisted finishing, often combined with abrasives in electropolishing or magnetic-assisted processes. In this method, an ultrasonically vibrated tool is used in conjunction with a liquid slurry for material removal [Reference Wang, Zhu and Liew115]. Tyagi et al. demonstrated that both chemical polishing and electropolishing can significantly reduce the surface roughness of AM 316L, achieving submicron finishes and improving the quality of intricate or internal geometries [Reference Tyagi, Goulet, Riso, Stephenson, Chuenprateep, Schlitzer, Benton and Garcia-Moreno116, Reference Han and Fang117].

The choice of surface finishing strategy depends strongly on the component geometry, material and functional requirements. Electropolishing of internal cooling passages can significantly improve heat transfer efficiency and reduce pressure drop by smoothing surface irregularities and removing micro-scale defects, leading to enhanced thermal performance in components such as turbine blades and heat exchangers. AFM is particularly effective at eliminating surface-connected defects, thereby improving fatigue life. Studies report that AFM can achieve average surface roughness (Ra) values of approximately 4 $\,\mu m$ , reducing stress concentrators and enhancing structural durability [Reference Cherian and Issac118]. For high-value or complex components, hybrid approaches that combine mechanical and chemical methods are increasingly adopted to balance accessibility, material removal rate and dimensional precision. Optimised surface finishing not only improves fatigue resistance and surface integrity but also contributes directly to more reliable thermal management and combustion stability in rocket engines.

8.1.4 HIP

HIP is a post-processing technique in which components are subjected to high temperature and isostatic gas pressure, typically using inert gases such as argon, to eliminate internal porosity and reduce microstructural defects [Reference Sripada, Tian, Chadha, Saha, Jahazi, Spray and Aranas119]. During additive manufacturing, trapped gas, incomplete fusion or lack-of-fusion defects can generate pores and microscopic voids within the material. These defects act as stress concentrators, significantly reducing fatigue life, creep resistance and overall mechanical performance. HIP densifies the material by closing voids and promoting diffusion bonding between particles, increasing component density to near 100% and restoring mechanical properties comparable to those of wrought or conventionally cast alloys [Reference Ci, Hu, Cheng, Liu, Xie, Cai, Dong and Wang120, Reference Martelli, Bassini and Ugues121].

In addition to densification, HIP can homogenise the microstructure, reduce residual stresses and influence grain morphology depending on the applied temperature, pressure and hold time [Reference Sun, Li, Sun, Liu and Sun122]. For nickel-based superalloys, HIP is typically performed prior to ageing treatments to ensure that $\gamma $ ^′ and $\gamma$ ^′′ precipitates form in a defect-free $\gamma$ matrix, maximising strength, high-temperature durability and creep resistance [Reference Zhao, He and Li123]. The process is particularly beneficial for components with complex internal geometries, such as regeneratively cooled combustion chambers and turbine blades.

Although HIP is highly effective, it requires specialised equipment and precise process control, and it increases production time and cost. Nevertheless, the improvements in mechanical reliability, fatigue life and thermal performance often justify its use for critical propulsion components. When combined with other post-processing techniques such as heat treatment and surface finishing, HIP allows AM components to achieve the density, microstructural uniformity and high-temperature performance necessary for aerospace propulsion systems.

8.2.1 Non-destructive testing (NDT)

NDT assesses the integrity of AM components without altering or destroying them. Additive processes can introduce voids, lack-of-fusion defects, microcracks and inclusions that may not be visible externally but can critically impair performance. X-ray CT can resolve internal pores in three dimensions, quantify their volume fraction and assess their distribution within complex geometries [Reference Khosravani and Reinicke124, Reference Sagbas, Durakbasa, Durakbasa and Gençyılmaz125]. CT also allows correlation of defect populations with processing parameters [Reference Zhou, Dai, Chu, Wang, Li, Zhou and Cheng126].

Ultrasonic inspection detects planar defects and delamination in thicker sections. Modern phased-array ultrasonics allow mapping of flaw orientations, which is vital in anisotropic AM microstructures [Reference Liu, Wang, Oliveira, He and Guan127]. Dye penetrant and fluorescent penetrant inspections remain widely used for surface-connected flaws [Reference Chen, Peng, Kong, Dong, Remani and Leach128]. Combining multiple NDE methods provides both volumetric and surface-level insight, ensuring defects are caught early before costly post-processing or integration steps.

8.2.2 Mechanical and thermal testing

Mechanical testing characterises the structural capabilities of AM alloys and is often more extensive than in conventional manufacturing. Tensile testing across multiple-build orientations captures anisotropy effects, as vertical builds may differ significantly in yield strength and elongation compared to horizontal builds [Reference Huangfu, Liu, Liu, Wu and Bai33, Reference Shrivastava, Subramaniyan, Rao and Nagesha34, Reference Weaver and Rosenthal36]. Hardness testing provides a quick indicator of microstructural uniformity, while compression testing is relevant for components subjected to intense hydrostatic loads.

Fatigue testing simulates cyclic loading during engine startup, shutdown and steady operation. High-cycle fatigue tests simulate long-term operation, while low-cycle fatigue tests replicate extreme transients. Creep testing is critical for nickel-based superalloys operating near their melting temperature [Reference Bridges, Shingledecker, Clark and Crudden129]. Fracture toughness and crack-growth testing provide insight into how residual porosity or microstructural inhomogeneity affects crack propagation. Mechanical testing establishes design limitations, identifies potential failure modes and creates the performance baseline against which post-processing and optimisation strategies are measured.

8.2.3 Hot-fire and environmental testing

Hot-fire testing (Figure 5) is the most rigorous qualification phase for AM propulsion components, exposing them to the actual thermo-mechanical environment of rocket engine operation. Combustion chambers, injectors and nozzles must endure extreme pressures and thermal loads reaching thousands of degrees Kelvin [Reference Gradl, Protz, Cooper, Ellis, Evans and Garcia8, Reference Gradl and Protz90]. Hot-fire campaigns confirm mechanical stability and reveal subtle failure mechanisms, such as microcrack propagation, liner delamination or localised hot spots that may arise from imperfect surface finishing [Reference Maleki, Bagherifard, Bandini and Guagliano110]. Infrared thermography and high-speed imaging can pinpoint localised overheating and erosion [Reference Ciampa, Mahmoodi, Pinto and Meo131].

Comprehensive testing ensures that AM rocket propulsion components meet aerospace requirements. NDE identifies internal and surface defects, mechanical testing quantifies strength, fatigue and creep performance, while microstructural characterisation links process parameters and post-processing to material behaviour. Hot-fire and environmental testing validate components under realistic operational conditions. When combined together, these approaches help in guiding design improvements, post-processing strategies and certification efforts, bridging the gap between as-built parts and fully flight-ready propulsion systems.

10.1.1 NASA RS-25 engine modernization

NASA’s RS-25 modernisation programme illustrates a systematic integration of AM into a proven propulsion system. L3Harris restarted RS-25 design and production for the Space Launch System in 2015, leveraging modern manufacturing techniques such as 3D printing to reduce costs while maintaining reliability. A prominent example of AM application is the Pogo Z-Baffle assembly, which reduced welds from 127 to just 4, demonstrating the potential of part consolidation. Additional AM components included injector elements with complex internal flow passages, turbine blades incorporating internal cooling channels and brackets and housings optimised for weight reduction [Reference Gradl132].

The RS-25 programme included rigorous testing and validation, encompassing incremental power-level testing, full-duration hot-fire tests at 111% power, and repeated start-stop cycles to evaluate long-term durability. Components were manufactured using L-PBF with Inconel 625 and 718 alloys, followed by hot isostatic pressing and precision machining. Quality control employed X-ray computed tomography and ultrasonic inspection to ensure defect-free parts. The programme successfully demonstrated performance equivalence to conventionally manufactured components, reduced overall engine cost by approximately 30%, shortened manufacturing lead times and enabled integration of AM components into flight-certified engines for Artemis missions [Reference Gradl132].

Beyond this success, NASA’s broader AM development programme has also documented failure modes that inform process qualification practice. In one case, an L-PBF copper-alloy combustion chamber failed during hot-fire testing as a direct result of degraded material quality. This failure occurred during a risk-tolerant development test intended to demonstrate high duty-cycle performance, where the consequences of failure were deliberately minimised by the test architecture. NASA has highlighted this incident as evidence of the necessity for robust material characterisation and rigorous in-process quality control when qualifying AM components for flight use [Reference Gradl132]. This example illustrates that, even within a programme with demonstrated successes such as the RS-25 Pogo Z-Baffle, AM qualification requires systematic failure analysis and process control alongside reported performance gains.

10.1.2 Relativity Space Terran 1

Relativity Space’s Terran 1 rocket showcases the near-complete adoption of AM for rocket production. Approximately 85% of the rocket’s dry mass, including critical components such as the combustion chamber, injector and turbopumps, was manufactured using AM techniques. This high degree of integration enabled complex geometries and reduced part count by over 70%, simplifying assembly and maintenance [133]. To further accelerate production and reduce risk, Relativity Space employed a hybrid approach combining traditional manufacturing methods with additive processes (Terran R) [134]. The extensive use of AM allowed critical engine parts to be produced within 24 hours, achieving material efficiency gains of 30–40% compared to subtractive manufacturing methods [133].

10.1.3 Ursa Major Hadley engine

Ursa Major’s Hadley engine showcases AM’s ability to streamline production and improve performance in commercial rocket engines. The company’s advanced manufacturing lab enabled in-house fabrication of complex components, reducing dependency on external suppliers and accelerating development timelines. Flight testing validated AM components under operational conditions. AM reduced production timelines from six months to one month and shortened prototyping lead times by approximately 80%. Multi-part assemblies were consolidated into fewer components, demonstrating AM’s capacity for complex geometries [135, 136].