Metal Additive Manufacturing —
Processes, Materials
and New Developments

Metal additive manufacturing has moved decisively out of the prototype shop and into production. Aerospace structures, medical implants, oil and gas downhole tools, nuclear reactor components, and defence hardware are being built from powder bed and wire-fed processes that did not exist as production technologies fifteen years ago. The rate of development has been high enough that the technology landscape of 2025 is materially different from 2020 — new process variants, new materials, new monitoring approaches, and the beginning of mature qualification frameworks are all making the technology accessible to industrial sectors that previously treated it as a laboratory curiosity.

This article covers the principal metal additive manufacturing processes, their material capabilities, the properties achievable relative to wrought equivalents, design considerations, the qualification challenge, and the specific developments in the last two to three years that are changing what the technology can deliver commercially.

The Principal Processes

Laser Powder Bed Fusion (LPBF)

The dominant process for precision metal components. A laser melts successive layers of metal powder (typically 20–60 μm layer thickness) on a build platform within an inert atmosphere chamber. Parts are built layer by layer, with the unmelted surrounding powder supporting overhangs and acting as a heat sink. After building, support structures are removed and the part undergoes post-processing — heat treatment, HIP, and machining of critical surfaces.

LPBF produces the finest feature resolution of any metal AM process (minimum feature ~0.1mm for well-optimised parameters), the best surface finish in the as-built state (Ra 5–15 μm on vertical surfaces, better on upward-facing surfaces), and a wide material range. It is the dominant process for aerospace, medical, and precision industrial applications. Limitations: build volume constrained by powder bed size (typical machines 250–500mm cube), slow build rates compared to DED and WAAM, and the powder handling and inert atmosphere requirements represent a significant facility investment.

Directed Energy Deposition (DED)

A nozzle deposits powder or wire into a melt pool created by a laser, electron beam, or plasma arc, building up material directly onto a substrate or existing part. The process is less precise than LPBF (typical layer heights 0.25–1mm) but substantially faster, operates in open air for laser/plasma variants (unlike the sealed chamber of LPBF), and critically can add material to existing parts — enabling repair of high-value components (turbine blades, tooling, dies) and the addition of features to near-net castings.

Laser DED with powder feed (also marketed as LENS, DED-LB/P) is the most common variant, capable of building onto complex substrate geometries and transitioning between multiple powder feeds to create functionally graded materials — for example, depositing a wear-resistant surface layer in a different alloy over a tough structural substrate in a single operation.

Wire Arc Additive Manufacturing (WAAM)

Uses a welding arc (MIG, TIG, or plasma) as the heat source to melt wire feedstock, building large metal components layer by layer. WAAM operates in open air (for most ferrous and many non-ferrous materials), uses standard welding wire as feedstock (significantly cheaper than AM powder), and has build rates orders of magnitude higher than LPBF — depositing several kilograms of metal per hour rather than grams.

The trade-off is geometric resolution and surface finish. WAAM parts require significant machining to achieve final dimensions and surface finish, making it a near-net-shape process rather than a precision process. But for large titanium, high-strength steel, or nickel alloy components where raw material is expensive and buy-to-fly ratios are high — landing gear components, pressure vessel heads, submarine components — WAAM offers dramatic material savings over traditional machining from solid. A titanium component machined from a large billet with a 10:1 buy-to-fly ratio can potentially be WAAM-built with a 1.5:1 ratio and machined to final tolerances.

Electron Beam Melting (EBM / SEBM)

Similar to LPBF but uses an electron beam rather than a laser as the energy source, operating in a high vacuum rather than inert gas atmosphere. The vacuum environment is mandatory (electron beams scatter in gas) but provides complete oxidation protection — beneficial for reactive metals such as titanium. The elevated build chamber temperature (typically 600–700°C for Ti-6Al-4V) reduces thermal gradients and residual stress compared to LPBF, often producing parts that require less post-build stress relief. EBM parts have a characteristic rough surface finish (Ra 25–35 μm in the as-built state) due to the sintered powder layer surrounding the melt, requiring more extensive machining than LPBF parts. Primary application: titanium medical implants and aerospace structural components.

Binder Jetting

A printhead deposits a liquid binder onto a powder bed, layer by layer, in a cold process — no melting occurs during printing. The "green" part (held together only by the binder) is then sintered in a furnace, burning off the binder and partially or fully densifying the metal. No support structures are required during printing (the surrounding dry powder supports the green part), enabling very complex internal geometries, and the process is fast — large numbers of parts can be produced in a single powder bed.

The sintering step introduces approximately 20% linear shrinkage that must be compensated in the build file. Material properties after sintering are slightly below wrought equivalents due to residual porosity (typically 97–99.5% density). Binder jetting is emerging as the leading AM process for volume production of medium-complexity precision parts — tool steel components, stainless valve bodies, wear parts — where the combination of high throughput, no support removal, and no residual stress from a melt-solidification cycle are compelling advantages over LPBF.

Materials

The available material range for metal AM is substantially wider than the common perception of "titanium and stainless" would suggest. Commercially qualified materials include:

Properties vs Wrought — The Critical Question

Whether AM-produced metal meets the properties of wrought equivalents is the most important question for engineering application — and the answer is more nuanced than either the optimistic claims of AM advocates or the sceptical dismissal of traditional metallurgists.

For LPBF 316L stainless steel, the as-built (before heat treatment) microstructure is cellular-dendritic with a strong crystallographic texture — the rapid solidification rates produce fine microstructure with grain sizes significantly smaller than typical wrought plate. Ultimate tensile strength and yield strength in as-built LPBF 316L are typically higher than wrought annealed equivalents; elongation to fracture is lower. After solution annealing, properties more closely approach conventional wrought product. Fatigue strength is more sensitive to defects (pores, lack-of-fusion voids, surface roughness) than static strength — AM parts in fatigue applications require either HIP (hot isostatic pressing) to close internal voids, electropolishing or machining of critical surfaces, or both.

For Ti-6Al-4V, LPBF as-built produces a martensitic α' microstructure with very high strength but limited ductility. Stress relief and subsequent annealing or STA (solution treat and age) treatment converts this to an α+β structure with ductility comparable to wrought equivalents. HIPped and heat-treated LPBF Ti-6Al-4V can meet AMS 4928 (wrought bar) equivalency for static properties; fatigue equivalency requires HIP plus surface treatment.

The general principle: with appropriate post-processing (stress relief, HIP, heat treatment, critical surface finishing), metal AM parts can meet wrought equivalency for static properties in most commercially important alloys. Fatigue equivalency requires more care and is application-specific. Dynamic loading and fracture mechanics applications still require careful qualification.

Design for Additive Manufacturing (DfAM)

The full value of metal AM is only realised when components are designed for the process — not when existing designs are reproduced in AM as a substitute for conventional manufacturing. DfAM considers features that are uniquely enabled by AM:

Topology optimisation

Mathematical optimisation of material distribution within a design envelope to maximise structural performance for a given load case while minimising mass. The output is typically an organic, skeletal structure impossible to machine but buildable in LPBF or DED. Applications: aerospace brackets, structural frames, heat exchanger headers. The resulting geometry typically requires careful interpretation — topology optimisation produces a conceptual structure that must be engineered for manufacturability and post-processing access.

Lattice structures

Regular or stochastic internal lattice geometries replace solid volumes, reducing mass while maintaining stiffness or providing controlled mechanical properties (e.g. graded stiffness in medical implants to match bone). LPBF can produce lattice strut diameters below 0.3mm. Lattice structures in medical implants promote osseointegration — bone grows into the lattice. In heat exchangers, lattice core structures provide dramatically increased surface area per unit volume relative to conventional baffle designs.

Internal channels and conformal cooling

Complex internal channel networks for conformal cooling of injection moulds and die casting tools — cooling channels following the mould surface contour rather than straight drilled holes — significantly reduce cycle times and improve part quality. This is one of the most commercially mature and cost-justified AM applications in tooling.

Part consolidation

Assemblies of multiple conventionally manufactured and joined parts can be consolidated into a single AM build. Eliminating joints removes potential leak paths, reduces assembly cost, and often reduces mass. The economic case requires careful analysis — the saving on assembly must justify the AM build cost.

New Developments — What Has Changed Recently

Multi-laser LPBF systems

The productivity constraint of single-laser LPBF has been directly addressed by multi-laser systems. Machines with 4, 8, and now 12 laser sources scanning simultaneously across a large build plate are commercially available from EOS, Nikon SLM, Trumpf, and Velo3D. A 12-laser machine (such as the Nikon SLM BLT-S1200) can achieve build rates approaching 1,000 cm³/hour compared to 20–60 cm³/hour for a single-laser machine — a step change that shifts the economic case for LPBF from prototype and low-volume to genuine production volumes for aerospace and energy components. The challenge with multi-laser systems is maintaining metallurgical consistency at the stitch zones where laser zones overlap — melt pool interaction between adjacent lasers must be carefully managed to avoid porosity or inconsistent microstructure at the stitch.

In-situ monitoring and closed-loop process control

Melt pool monitoring using high-speed cameras, pyrometry, and optical coherence tomography (OCT) during the build has moved from research to standard machine capability on current-generation hardware. EOS's EOSTATE MeltPool monitoring and Additive Industries' ProcessControl systems capture thermal and geometric data on every layer. The data volumes are enormous — a single build can generate terabytes of image data — but the payoff is the ability to detect voids, lack-of-fusion defects, and delamination events during the build rather than after, allowing builds to be halted before further material is deposited on a fundamentally defective region. The more ambitious development is closing the loop — using in-situ data to automatically adjust laser power, scan speed, or hatch pattern in real time to correct defects as they form. This is being actively developed and partially deployed commercially.

AI and machine learning in process parameter development

Developing print parameters (laser power, scan speed, hatch spacing, layer thickness) for a new material or geometry has historically required extensive and expensive Design of Experiments testing. Machine learning approaches — in particular Gaussian process models, neural networks trained on existing parameter sets, and reinforcement learning — are compressing this development time significantly. Materialise, Autodesk, and several AM machine manufacturers now offer AI-assisted parameter development tools. The practical impact: new alloys that previously required 12–18 months of parameter qualification can potentially reach production-ready parameters in 3–6 months.

Copper by LPBF — green laser breakthrough

Pure copper and copper alloys were previously very difficult to process by LPBF because their high optical reflectivity at the infrared wavelengths (1064nm) of standard ytterbium fibre lasers meant most of the laser energy was reflected rather than absorbed, producing inconsistent melt pools and high porosity. Green laser LPBF systems (515nm wavelength, from manufacturers including Trumpf and Elementum3D) have resolved this — copper absorptivity at 515nm is approximately 10× higher than at 1064nm, enabling consistent, high-density copper builds. The application driving this development is rocket propulsion — copper combustion chamber liners for liquid-fuelled rocket engines benefit enormously from AM's ability to produce conformal cooling channels in copper. But the technology also opens copper AM to heat exchanger cores, inductors, and electrical busbars.

WAAM for large structural components — industry adoption

Wire arc additive manufacturing for large titanium and high-strength steel structural components has moved from university research to active aerospace supply chain use. Norsk Titanium (now Amaero) has produced WAAM titanium structural components for Boeing 787 and other commercial aircraft programmes. Cranfield University's WAAM3D and similar systems are being adopted by defence primes for submarine and shipbuilding components. The economics are compelling for large, high-buy-to-fly-ratio parts: titanium billets cost £30–60/kg and a traditional machined part might use 90% of the billet as chips — WAAM near-net-shapes the component and leaves 10–20% to be machined, with deposited metal costs of £80–120/kg but on dramatically less material.

High-entropy alloy development for AM

High-entropy alloys (HEAs) — alloys containing five or more principal elements in near-equimolar proportions — show remarkable combinations of strength, toughness, and radiation resistance that make them attractive for extreme-environment applications in defence and nuclear. The challenge has been that HEAs are difficult to produce in wrought form at scale. Metal AM, where the alloy composition can be varied across a build using multiple powder feeders, is an ideal production route for HEAs and is accelerating their development from laboratory curiosity to engineering material. Compositions including CrMnFeCoNi (the Cantor alloy), AlCoCrFeNi, and refractory variants for temperatures above 1200°C are in active development for specific applications.

Cold spray for repair and coating

Cold spray — supersonic impact of metal powder particles that bond through plastic deformation without melting — is not AM in the traditional sense but fills a complementary role. Where DED and WAAM require a heat-affected zone and can distort thin substrates, cold spray deposits material at room temperature with no thermal input to the substrate. It is used for corrosion-resistant coatings, repair of aluminium aircraft structures (replacing panels damaged by corrosion or impact), and restoration of worn bearing surfaces and shafts on large rotating equipment. The process has been accepted by the US Navy for submarine and aircraft carrier maintenance, which represents a significant qualification milestone for general industrial adoption.

Qualification and Certification — The Persistent Challenge

The technical capability of metal AM has outpaced the qualification and certification frameworks, and this remains the primary constraint on adoption in regulated industries. Three parallel qualification challenges exist:

• Process qualification — demonstrating that a specific machine, powder batch, and parameter set consistently produces material meeting the required specification. ASTM F3001 (Ti-6Al-4V), F3055 (IN718), and F3056 (IN625) define material requirements for LPBF parts; AS7003 (aerospace process specification for LPBF) defines process control requirements. Qualification is machine-specific and parameter-specific — a qualified process on one machine does not automatically transfer to another machine of the same model.
• Part qualification — demonstrating that a specific part produced by AM meets its functional requirements, including fatigue, fracture mechanics, and non-destructive examination. The challenge is that AM porosity and microstructure are geometry-dependent (thin walls have different thermal histories than thick sections), so coupon testing does not directly qualify the part. Computed tomography (CT) scanning is increasingly used as the qualification NDE method for critical AM parts.
• Feedstock qualification — powder properties (particle size distribution, morphology, chemistry, flowability) must be controlled and documented. Powder reuse — the unmelted powder recycled after each build — changes properties over multiple build cycles, and maximum reuse cycles must be validated for each application.

API is developing standards for AM in oil and gas equipment (API TR 5LE5 covers AM for line pipe and tubular applications). ASME is developing AM sections within existing pressure equipment codes (ASME VIII Division 1 Addendum for AM parts is in progress). These frameworks will be enabling for process industry adoption, where the absence of code qualification currently forces the expensive bespoke-qualification route for each application.

When to Use Metal AM — and When Not To

Metal AM is not universally cheaper or better than conventional manufacturing. The economic and technical case varies significantly by application:

AM is well-suited when: geometry is complex and internal (lattices, conformal cooling, integrated channels), buy-to-fly ratio in conventional manufacture is high (large titanium or nickel alloy parts machined from solid), part quantities are low (1–100 parts), lead time is critical (tooling-free production), or functional performance justification outweighs unit cost (aerospace, medical, defence).

AM is poorly suited when: parts are simple solids that machine efficiently from bar or plate, material quantities are large relative to what AM can deposit, dimensional tolerances are tight throughout (AM still requires machining of mating and sealing surfaces), or the regulatory qualification cost cannot be amortised across sufficient production volume.

Summary

Metal additive manufacturing is no longer a technology to monitor — it is a production technology with a defined process landscape, a growing material library, improving qualification frameworks, and a rate of technical development that continues to extend its capability. The multi-laser productivity step change, in-situ monitoring moving toward closed-loop control, green laser enabling copper, WAAM entering aerospace supply chains, and AI-compressed parameter development are all expanding the viable application envelope. The qualification challenge remains real — particularly for pressure-containing and safety-critical components in regulated industries — but the ASME and API framework development underway will progressively remove that barrier. The engineer who understands the process capabilities and limitations, the design freedoms AM genuinely offers, and when the economics justify its use is better placed than one who waits for the technology to become universally proven before engaging with it.