Common Defects in Metal AM

Metal additive manufacturing (AM)—especially powder bed fusion (PBF) processes such as DMLS / SLM—can produce highly optimized geometries, consolidated assemblies, and rapid iteration for defense, aerospace, and advanced industrial programs. But the same thermal gradients, powder handling risks, and layer-by-layer build physics that enable performance also create recurring failure modes. For engineers, the impact is straightforward: defects can reduce fatigue life, degrade leak-tightness, shift dimensions out of tolerance, or create inspection and certification headaches. For procurement and program teams, defects show up as late deliveries, nonconformances, and unplanned rework costs.

This article focuses on the most common metal 3D printing defects encountered in production PBF workflows and—more importantly—how qualified suppliers prevent, detect, and disposition them under regulated quality systems (e.g., AS9100, customer flowdowns, ITAR/DFARS requirements). The goal is not academic theory; it is a practical view of what drives yield and reliability for flight- and mission-critical hardware.

Porosity types

Porosity is one of the most frequent and consequential classes of defects in PBF. Not all pores are created equal; the morphology often indicates root cause and informs whether downstream densification (e.g., Hot Isostatic Pressing (HIP)) will be effective.

Gas porosity (spherical pores) typically presents as small, round voids. Common contributors include entrapped gas in powder particles (especially if powder has been repeatedly reused beyond validated limits), inadequate shielding gas flow, or process parameters that trap gas during solidification. Gas porosity can be partially mitigated by parameter optimization and powder controls; it may also be reduced by HIP depending on size, location, and connectivity.

Keyhole porosity forms when energy density is too high, causing deep vapor depressions that collapse and trap voids. In CT scanning, it often appears as larger, irregular or elongated pores aligned with scan tracks. Keyhole porosity is a parameter-window problem: excessive laser power, low scan speed, insufficient hatch spacing control, or poor thermal management can push the melt pool into keyhole mode.

Lack-of-fusion (LoF) porosity is typically irregular, crack-like, and often planar, reflecting un-melted boundaries between tracks or layers. LoF is more damaging than spherical porosity because it behaves like a pre-existing crack, severely reducing fatigue strength. HIP may not fully close LoF defects if they are connected to surfaces or contain oxide films, and they can remain as unbonded interfaces even after densification.

Interconnected porosity / leakage paths matter for fluid systems, housings, and pressure components. Even relatively low porosity by volume can produce leakage if pores are connected. For procurement, this is where specifying the right acceptance tests (e.g., helium leak, pressure decay, or dye penetrant where applicable) becomes as important as setting a density target.

How suppliers prevent porosity in practice usually involves a combination of (1) validated parameter sets for each alloy and thickness regime, (2) powder quality management (particle size distribution, morphology, chemistry, oxygen/nitrogen limits, reuse tracking), (3) oxygen monitoring and inerting controls in the build chamber, and (4) build layout strategies that maintain stable thermal conditions across the plate. High-performing suppliers treat porosity control like a closed-loop manufacturing process, not a one-time “dial-in.”

Warping and residual stress

PBF processes impose steep thermal gradients: each layer is rapidly melted and resolidified while constrained by cooler surrounding material. That creates residual stress, which can lead to warping, delamination, or dimensional drift during the build or during stress relief and support removal.

Common symptoms include upward curling at edges, bowed walls, holes that go out of round, or parts that “spring” after being cut off the build plate. In extreme cases, a part can collide with the recoater, causing a build crash and scrapping the plate.

Primary drivers include high aspect ratio geometries, uneven cross-sections, large flat surfaces, insufficient support structure stiffness, and scan strategies that concentrate heat in one region. Material also matters: some alloys (and certain heat-treatment conditions) are more prone to stress-driven distortion.

Supplier controls for warping/residual stress typically combine design-for-AM (DfAM) guidance and disciplined process planning. For example: (1) part orientation chosen to minimize overhangs and distribute heat; (2) support strategy designed for both thermal conduction and mechanical restraint; (3) scan strategy (stripe/island rotations, contour-first vs. hatch-first decisions) tuned to reduce stress accumulation; and (4) in-process temperature control such as platform preheat where available. Successful suppliers will also schedule and document a stress relief heat treatment prior to support removal or aggressive post-processing to reduce the chance of distortion later.

Where machining meets AM reality: defense and aerospace components commonly require tight geometric tolerances and true position callouts that are not achievable “as-printed.” The practical approach is to treat the printed component as a near-net preform and then apply CNC machining—often 5-axis machining—after stabilization heat treatment. A robust supplier will define machining datums that survive stress relief, include machining stock where it is needed, and verify stability through first-article builds before committing to rate production.

Lack of fusion

Lack of fusion (LoF) deserves special attention because it is both common and performance-critical. LoF occurs when adjacent melt pools or successive layers do not adequately overlap, leaving unbonded regions or trapped unmelted powder at boundaries.

Why it happens is usually a combination of insufficient energy density (low power, high scan speed), hatch spacing that is too wide, poor layer thickness control, or local heat-sinking that cools the melt pool too quickly. But LoF can also be triggered by less obvious causes: spatter redepositing on the powder bed, recoater streaks, inconsistent powder spreading, or subtle mis-calibration of focus and beam parameters.

Why it matters: LoF defects behave like planar cracks and dramatically reduce fatigue strength. In dynamic flight structures or rotating hardware, LoF can dominate life even when density appears acceptable by simple measurements. That is why high-end programs often require volumetric inspection methods (e.g., CT scanning) on representative builds, coupons, or critical zones.

Practical prevention starts with parameter qualification and continues with operational discipline. A mature supplier will: (1) lock a qualified parameter set by machine model and alloy; (2) control layer thickness and recoater health; (3) monitor oxygen and gas flow to minimize spatter; (4) use witness coupons on the plate for density and mechanical testing; and (5) apply in-process monitoring where available (melt pool monitoring, layer imaging) to detect anomalies early. Importantly, the supplier should have a defined response plan when monitoring flags a risk: pause, adjust, abort, or quarantine for enhanced inspection rather than hoping post-processing “fixes it.”

HIP is not a cure-all for LoF. HIP can close voids when surfaces are clean and compressible, but it cannot reliably eliminate oxide films or fully bond un-melted interfaces. If your application is fatigue-driven, treat LoF prevention and detection as non-negotiable upstream requirements, not a downstream gamble.

Surface defects

Surface condition affects fatigue initiation, corrosion performance, sealing, and the ability to achieve final tolerances after machining. PBF parts frequently exhibit surface-related defects that are not captured by density metrics.

Typical surface issues include partially fused particles (“sintered” powder attached to walls), balling, stair-step effects on shallow angles, micro-notches at contour transitions, and support interface scarring. Internal surfaces—such as lattice cores, conformal channels, and manifolds—can trap powder or exhibit roughness that cannot be reached by conventional finishing tools.

Root causes vary: suboptimal contour parameters, inadequate shielding gas flow leading to spatter adhesion, overhang angles beyond process capability, and build orientations that place critical surfaces in down-skin regions. Surface defects can also be introduced in post-processing, such as aggressive grit blasting that embeds media, or inconsistent chemical processing.

Real-world post-processing stack often includes multiple steps, chosen based on function and accessibility. Common combinations include: (1) support removal; (2) stress relief; (3) optional HIP for density and fatigue improvement; (4) surface conditioning (blasting, tumbling, machining, or controlled chemical processes); and (5) final CNC machining on functional interfaces, sealing lands, and datum features. The key is to specify the final functional requirement (e.g., Ra/Rz, sealing, fatigue) and let the supplier propose a validated finishing route that meets it without damaging subsurface integrity.

Inspection for surface-driven risks typically includes visual inspection to defined criteria, surface roughness measurement where required, and NDE methods appropriate to the geometry (e.g., dye penetrant on accessible surfaces if permitted by material and requirements, or CT for internal channels). For aerospace-grade workflows, these inspection steps should be documented and traceable, not informal “shop checks.”

Prevention controls

Defect prevention in metal AM is less about a single magic parameter and more about a controlled workflow that ties engineering intent to shop-floor execution and inspection evidence. Below is how successful defense/aerospace suppliers typically prevent metal 3D printing defects from RFQ through delivery.

1) RFQ and requirement capture. A qualified supplier will ask questions that directly relate to defect risk: critical-to-quality (CTQ) features, fatigue vs. static loading, leak-tightness, allowable internal defects, surface finish needs, heat-treatment requirements, and inspection/acceptance criteria. Procurement should provide applicable drawings, specs, revision levels, and required certs/flowdowns (e.g., ITAR handling, DFARS material compliance, serialization, special process requirements). If acceptance is ambiguous, the supplier should propose a plan (e.g., CT on first articles, coupon testing per build) before price is locked.

2) Material and powder control. Suppliers prevent defects by treating powder like a controlled raw material, not a consumable. Expect: material identification, lot traceability, incoming inspection (chemistry, particle size distribution, morphology), oxygen/moisture handling controls, validated sieve/blend practices, and documented reuse limits. The powder lot used on your job should be traceable to your build records and ultimately to the certificates of conformance (CoC) in the ship package.

3) Machine qualification and calibration. In regulated production, the machine is not “good because it printed last week.” A disciplined supplier maintains calibration and preventative maintenance for laser power, optics condition, recoater performance, inert gas flow, oxygen sensors, and build plate flatness. Parameter sets are version-controlled and locked. If a machine is repaired or optics are changed, the supplier should have a defined re-qualification trigger to prevent silent drift that leads to porosity or LoF.

4) Build planning and simulation. Build orientation, support design, and scan strategy directly influence warping, LoF risk, and surface condition. Mature suppliers use build-prep checklists and, when warranted, simulation to predict distortion. They also plan for downstream machining: adding stock, defining datum schemes, and ensuring that supports do not interfere with tool access.

5) In-process monitoring and build records. Many PBF systems capture layer images, melt pool signals, oxygen trends, and machine alarms. The value is in how the supplier uses the data: establishing normal ranges, correlating anomalies to defects, and defining decision points (continue, pause, abort, quarantine). For defense and aerospace programs, build records should be retained and tied to part serial numbers for traceability.

6) Post-processing as a controlled process. Stress relief, HIP, solution/age heat treatment, and any special processes must be controlled and documented. If HIP is used, the supplier should specify the cycle, verify equipment capability, and ensure that cycle selection matches the alloy and property targets. In many aerospace supply chains, special processes are performed under accreditation such as NADCAP (depending on customer requirements and scope). Even when NADCAP is not mandated, adopting equivalent control rigor reduces risk.

7) Machining and metrology. Most mission hardware requires post-print machining. Best practice is to establish stable datums (often after stress relief/HIP), then apply 5-axis machining to achieve tolerance and surface finish. Dimensional verification is typically performed using CMM for prismatic features and GD&T, with additional methods for complex geometries as needed.

8) Non-destructive evaluation (NDE) strategy. Suppliers should propose NDE based on defect mechanisms: CT scanning for internal porosity/LoF, dye penetrant for surface-breaking cracks (where applicable), and other methods per program requirements. For high-consequence parts, CT is often used during process qualification and then reduced to sampling plans once the process is stable—provided the acceptance plan is agreed to up front.

9) Quality system integration. Defect prevention is reinforced when the supplier’s AS9100 quality management system ties together contract review, document control, training, calibration, nonconformance management, and corrective action. For buyers, it is reasonable to request evidence of process controls, not only a certification badge.

Acceptance and rework

Even with strong controls, some level of nonconformance is inevitable in metal AM. The difference between a production-ready supplier and a prototype shop is how acceptance criteria are defined, how deviations are contained, and how rework is executed without compromising integrity or compliance.

Define acceptance criteria early. On defense and aerospace programs, acceptance is typically a combination of: (1) dimensional conformance (CMM and/or other metrology), (2) material condition (heat treat records, chemistry/lot traceability), (3) internal integrity (CT/NDE results or coupon test results), and (4) surface condition (visual/roughness). If you need a density target, specify it in a way that can be verified (e.g., CT-based porosity limits in critical volumes, or coupon density with correlating CT during qualification). Avoid vague language that forces late-stage disputes.

Common dispositions include use-as-is (if within tolerance/criteria), rework/repair, or scrap. For regulated workflows, rework requires an approved plan, documented execution, and re-inspection. A serious supplier will quarantine suspect parts, perform root-cause analysis, and document corrective action rather than quietly “touching up” and shipping.

Typical rework paths for AM parts include: (1) additional machining if extra stock exists and dimensional issues are correctable; (2) additional surface finishing if roughness is high but geometry allows; and (3) heat treatment reprocessing if allowable by the material specification. Rework becomes risky when it changes microstructure unpredictably, removes too much material, or masks underlying defects.

Weld repair is not automatically acceptable. Some programs allow weld repair with strict procedures and subsequent inspection; others prohibit it outright for certain alloys or critical parts. If weld repair is on the table, it should be treated as a special process with qualified procedures, documented welder qualifications, and NDE after repair. Buyers should ensure repair allowances and approval paths are defined before production starts.

Certification pack expectations. Procurement teams can reduce delays by clearly stating what must be included at shipment. A typical pack for metal AM parts may include: CoC, material traceability records (powder lot, heat lot where applicable), build ID and machine ID, heat-treatment/HIP records, NDE reports (CT summary or acceptance statement), CMM or dimensional inspection report, and any customer-required forms. For ITAR programs, ensure data handling and marking requirements are met. For DFARS, ensure material compliance documentation is captured as required by your flowdowns.

How to make acceptance practical: For many organizations, the most effective approach is a two-phase plan. Phase 1 (qualification) uses heavier inspection—CT on representative builds, full coupon testing, and capability studies. Phase 2 (production) uses the established process window plus sampling plans and lot-based controls, with clear triggers that return you to Phase 1-level scrutiny if drift is detected. This aligns cost with risk while maintaining defensible quality evidence.

When metal AM is treated as an engineered, controlled manufacturing process—integrated with HIP/PM-HIP where appropriate, robust machining, and documented inspection—defects become manageable variables rather than program surprises. The suppliers who consistently deliver for defense and aerospace are the ones who can explain, in detail, how they prevent porosity, residual stress distortion, lack of fusion, and surface-driven failures—and can back those claims with traceable records.