The three-regime porosity problem in SLM Inconel 718

Residual porosity in selective laser melted Inconel 718 originates from three distinct mechanisms, each responding to different process levers — and each demanding a different engineering response. Understanding which regime is active is the prerequisite for eliminating pores without touching the laser power dial.

The three mechanisms are: lack-of-fusion (LOF) porosity, caused by insufficient energy to wet the preceding layer, producing irregular inter-layer voids that are highly detrimental to fatigue and tensile performance; gas and spherical porosity, from entrapped shielding gas or atomised powder content, generating near-spherical pores typically 10–100 µm in diameter; and keyhole porosity, deep irregular pores formed when excessive energy density causes vapour depression collapse. Critically, raising laser power pushes the process directly into this third regime.

The porosity window for IN718 is quantitatively documented: at a volume energy density (VED) of 60 J/mm³, porosities as low as 0.15% are achievable. Deviation in either direction — too little energy or too much — produces 1.26–3.46% porosity. This LOF → optimal → keyhole transition defines the central engineering challenge: reaching the low-porosity window without raising power. The challenge is compounded by the fact that 60 J/mm³ at fixed power is a multi-variable optimisation problem across scan speed, hatch spacing, and layer thickness, each of which carries its own machine-specific sensitivity.

According to standards bodies including ASTM and the ISO technical committees on additive manufacturing, porosity characterisation and process qualification are now central to AM component certification — making porosity reduction not merely a materials science challenge but a regulatory one. For aerospace components, the consequences are direct: porosity is identified in the research record as a fatigue crack initiation site that is life-limiting in rotating hardware such as turbine discs and impellers.

Scanning strategy and parameter redistribution: achieving optimal VED without touching power

Redistributing energy via scan speed, hatch spacing, and layer thickness — while holding laser power constant or even reducing it — is the most direct route to the 60 J/mm³ optimum. Research on SLM IN718 demonstrates layer-by-layer cycling of scan speed between 500 and 1900 mm/s at fixed power (100–200 W) to control melt pool size and transition between porous and dense microstructure within a single specimen. High energy density achieved via low speed — rather than high power — produced sound, columnar-grained regions.

A particularly significant finding for high-productivity industrial deployments concerns inter-layer cooling time (ILCT). In multi-laser, high-productivity SLM configurations, very short ILCT creates thermally unstable melt pools that increase defect density in IN718 components. This is an architectural consequence of running multiple lasers simultaneously: the reduced dwell time between layers leaves residual heat from the previous layer unresolved, altering melt pool viscosity and surface tension at the moment of subsequent laser irradiation.

The implication is significant: ILCT optimisation is a design-stage intervention that requires no hardware modification. Engineers can reduce porosity on multi-laser machines by controlling scanning geometry — island size, stripe width, scan sequence — to extend the effective dwell between successive passes over any given area. According to NIST research on additive manufacturing process monitoring, melt pool thermal history is a dominant predictor of defect formation, reinforcing ILCT as a first-order process variable.

A pulse parameter framework — encompassing beam radius, pulse energy, frequency, and duration — also provides a systematic basis for single-track porosity management, with direct application to commercial SLM machine optimisation and numerical model validation. This framework, documented as early as 2017 in the literature, enables engineers to map the porosity response surface without requiring access to laser power controls.

“The LOF → optimal → keyhole transition is the central engineering constraint: the 60 J/mm³ optimum that delivers 0.15% porosity is achievable via speed, hatch, and thickness — laser power is the one variable that creates more problems than it solves.”

In-situ layer remelting: 5× porosity reduction without a power increase

Layer remelting — passing the laser over an already-solidified layer before the next powder layer is deposited — delivers a 5× greater porosity reduction rate than the base deposition rate, according to published research. The mechanism is surface tension-driven pore collapse in the molten pool: pores collapse under surface tension forces when the melt pressure exceeds a critical threshold, with collapse dynamics governed by the relationship between melt pressure and pore size.

An analytical model of this process defines a critical melting rate threshold below which pore collapse remains feasible — this informs the upper bound on remelting scan speed. Operating above this threshold preserves pores rather than closing them, so remelting parameter selection requires deliberate calibration rather than arbitrary speed reduction.

Validation in the Ti6Al4V SLM system — which shares melt pool physics with Inconel systems — confirms that lower scanning speeds during remelting most effectively reduce defect numbers. A Rolls-Royce plc patent (US, 2016) specifically claims a laser surface remelting strategy for intermetallic components to eliminate surface porosity, using controlled shallow melting (≤300 µm depth) in an inert atmosphere. The inert atmosphere control is identified as a key enabling feature preventing oxidation during the remelting pass.

More recent work on femtosecond laser remelting in thermal barrier-coated superalloys demonstrates that HF-fs laser remelting heals surface pores and microcracks while producing refined, homogeneous grains — at energy levels far below those required for conventional laser remelting. This suggests a near-surface porosity remediation pathway applicable to IN718 components where conventional remelting depths are contraindicated by geometry or thermal budget constraints.

The primary limitation of layer remelting is machine access: the technique requires firmware or control access to the laser scan sequence that is not universally available on commercial SLM platforms. It is therefore more applicable to OEM-controlled systems or research-grade machines. For engineers with such access, the analytical model from the 2017 literature on pore collapse dynamics provides a rigorous design basis for remelting parameter selection.

Hot isostatic pressing: the most industrially mature post-process porosity elimination route for SLM IN718

Hot isostatic pressing (HIP) applies isostatic pressure — typically 100–200 MPa — at elevated temperature (1120–1200 °C for IN718) to collapse enclosed pores via creep and diffusion mechanisms. This approach is entirely independent of laser power and operates on the finished or near-finished component, making it applicable regardless of the SLM machine platform or scan strategy used during build.

HIP cycles at 1120–1200 °C applied to laser powder bed fusion IN718 produce equiaxed grain structures with 30–49 µm average grain diameter, annealing twins, and carbide precipitation at grain boundaries. Crucially, HIP closes both LOF open interconnected pores and spherical gas pores that survive standard heat treatment — the blowhole category being most responsive. The mechanistic evidence comes from 316L stainless steel HIP studies, where the analogous defect morphology directly transfers to IN718.

The most complete porosity elimination with concurrent precipitation strengthening is achieved through a sequence of stress-relief anneal → HIP → solution treatment → double aging. A Chinese patent by South China University of Technology (2020) explicitly claims that HIP prior to solution heat treatment eliminates both Laves phase and microscopic pores in SLM IN718 while promoting strengthening phase precipitation. This sequence positions HIP not merely as a densification step but as an integral element of the full metallurgical processing route.

AVIC Commercial Aircraft Engine — a subsidiary of China’s state aerospace conglomerate — has filed two patents (CN 2018, CN 2020) claiming HIP followed by abrasive flow machining to address both internal crack and porosity and inner-cavity surface integrity in aerospace turbine components. This combination targets geometrically complex components, such as turbine discs with internal cooling channels, where post-HIP surface finishing of internal cavities is required for dimensional compliance. As documented by international standards bodies including ASTM, HIP for AM components is now the subject of dedicated specification development.

From an IP perspective, HIP as a generic post-process step for AM components is not protectable — but specific HIP + downstream heat treatment combinations (as claimed by South China University of Technology and AVIC) should be reviewed before adopting identical protocols. IP freedom-to-operate in HIP process sequencing is largely available to teams that develop differentiated temperature-time profiles.

Alloy modification and feedstock-level interventions: expanding the printable process window

A smaller but strategically important cluster of research addresses porosity at the feedstock level, modifying alloy composition to improve melt pool behaviour, reduce gas entrapment susceptibility, or alter solidification dynamics — without any change to laser power or post-processing.

Yttrium addition at 0.1–2 wt% during SLM of IN718 has been evaluated through both single and double laser processing per layer. At 0.1 wt% Y, elongation increased, consistent with porosity reduction effects. However, at higher yttrium concentrations, undissolved Y particles create new defect sites — indicating a narrow optimisation window for this approach. The double-pass per layer approach used in this research functionally overlaps with the layer remelting strategy, suggesting that alloy modification and remelting can be combined in a single protocol.

At the thermodynamic modelling level, the solidification temperature range, melt pool geometry, and defect formation (keyhole, balling, LOF) can be linked through computational frameworks. This approach — applied to 316L stainless steel but directly generalisable — enables alloy composition selection to minimise defect susceptibility without power changes, by optimising melt pool geometry through thermophysical property tuning. Research published in peer-reviewed literature and indexed by organisations including Nature increasingly validates this computational alloy design approach for AM-specific applications.

The strategic value of alloy-level intervention lies in its potential to expand the printable process window — reducing sensitivity to parameter variation. This is particularly valuable for series production qualification, where parameter drift is a certification risk. If the alloy composition shifts the LOF boundary to lower VED values, the operator has more margin before entering the lack-of-fusion regime without requiring tighter laser power control. At current Technology Readiness Levels, however, composition modification remains an early-stage approach relative to HIP and scanning strategy optimisation.

Patent landscape and emerging directions: what the 2022–2024 filing record reveals

The patent record for SLM IN718 porosity reduction is geographically concentrated: six of seven identifiable patents in this dataset were filed in China, signalling that Chinese aerospace and manufacturing organisations are most actively seeking IP protection for production-ready porosity mitigation methods. The 2024 Shengu Group patent for a processing method for SLM-formed IN718 impellers represents active industrial deployment — moving beyond generic material research toward production-certification-grade protocols for compressor impeller manufacture.

Key assignees in the Chinese patent cluster include South China University of Technology (HIP + heat treatment, 2020), AVIC Commercial Aircraft Engine (HIP + abrasive flow machining, 2018 and 2020), FairyRobot Rapid Manufacturing Technology (SLM parameter + heat treatment protocol, 2019), Huazhong University of Science and Technology (induction preheating + laser irradiation, 2019 and 2020), and Shengu Group (impeller processing method, 2024). The US patent record in this dataset is limited to the Rolls-Royce plc laser surface remelting patent (2016, now inactive) and a Japan Atomic Energy Research Institute femtosecond laser treatment patent (2010, inactive).

Four emerging directions are identifiable in the 2022–2024 record. First, multi-laser machine ILCT management as a production-scale porosity lever — the 2023 ILCT study directly addresses high-productivity industrial machines where laser power cannot be arbitrarily increased due to thermal management constraints, and ILCT optimisation represents a largely unoccupied IP space as of this dataset. Second, VED optimisation at fixed power for stress rupture and creep targets: the 2023 study identifies 60 J/mm³ as the optimal VED for maximum rupture life and elongation, achieved by tuning speed, hatch, and layer thickness without power excursion. Third, impeller-level processing protocols combining stress relief, solution, and aging — the 2024 Shengu Group patent represents the industrialisation of porosity and residual stress elimination into component-specific, qualified manufacturing routes. Fourth, in-situ elemental modification for melt pool stabilisation remains at early TRL but offers the prospect of expanding the printable process window and reducing sensitivity to parameter variation — strategically valuable for series production certification.

For engineering teams seeking to understand the competitive and IP landscape in this field, the PatSnap R&D intelligence platform provides structured access to the full patent record across all four strategy clusters, including machine-readable claim analysis and assignee mapping.