Choosing the Right NDT Method for AM Parts

Non-destructive testing is not a formality at the end of AM production — it is the primary means by which a buyer or certifying authority can have confidence in a part that cannot be disassembled, sectioned, or tested to failure. Choosing the wrong method means spending money without obtaining information; in the worst case it means passing defective parts into service.

AM creates NDT challenges that do not exist in the same form for cast or wrought parts. This article sets out a practical decision framework for quality engineers and certifying authorities.

Why AM Microstructure Complicates NDT

Anisotropy. LPBF and EBM produce strongly textured columnar grain structures aligned with the build direction. Columnar grains scatter ultrasound differently depending on whether the beam propagates along or across the build direction — inspection results are orientation-dependent in a way that wrought material is not. Ultrasonic grain noise is typically higher in AM material, reducing the signal-to-noise ratio for flaw detection.

Rough as-built surfaces. As-built LPBF surfaces have Ra 8–20 µm; EBM surfaces are rougher at 25–50 µm. Surface roughness scatters both ultrasound and eddy current signals, generating false indications and masking real flaws at the surface and near-surface. The practical consequence: most contact UT inspection of as-built AM surfaces requires machining first, which changes the inspection workflow and adds cost.

Internal channels and complex geometry. Conformal cooling inserts, lattice structures, and internal manifolds are common AM design features that are impossible to inspect by line-of-sight methods. Internal surfaces cannot be accessed by penetrant or magnetic particle inspection. Channels smaller than approximately 0.5 mm cannot be reliably resolved by benchtop CT systems with standard focal spots.

Keyhole and LOF porosity distributions. Unlike casting shrinkage porosity (which concentrates at hot spots) or weld porosity (which follows the fusion line), LPBF porosity can be distributed throughout the bulk in regular patterns matching the scan strategy. This makes sampling-based inspection unreliable — a coupon from one area of a large part may not be representative of the whole.

Residual stress. High residual tensile stress at the surface (common in as-built LPBF parts without stress relief) can cause stress corrosion cracking during liquid penetrant inspection if the penetrant is aqueous and the part is left wet. This is not unique to AM but is more severe due to the higher residual stress levels.

The Six Main Methods

Computed Tomography (CT)

CT scanning is the only method that provides full volumetric defect mapping of internal features in a non-destructive manner. An X-ray source rotates around the part; a 2D detector captures projections; reconstruction algorithms produce a 3D volume. Resolution is limited by the focal spot size of the X-ray source and the geometric magnification achievable given the part size.

AM-specific notes: For detecting LPBF porosity, a voxel size of ≤ 10 µm is needed to reliably resolve pores > 30 µm. Most lab-scale CT systems can achieve this on parts < 50 mm diameter. For production-scale LPBF parts (100–250 mm), voxel sizes of 50–100 µm are more common, which means pores < 100–200 µm will not be resolved. CT also provides dimensional verification — wall thickness, channel diameter, and geometric deviations from CAD can all be extracted from the same scan. ASTM E1570 covers CT for aerospace; ASTM E2767 covers digital industrial CT.

Limitations: Cost (£300–2000 per part depending on size and resolution requirements), long scan times (30 minutes to several hours), and beam hardening artefacts in dense alloys (tungsten, iron, nickel) that can mask or simulate flaws. Parts containing trapped powder (a common LPBF quality escape) require CT before that powder can be safely removed — CT is the only way to verify depowdering completeness for internal features.

Conventional X-ray Radiography

2D projection radiography (DR or film) produces a through-thickness image. Defects are visible only if they produce sufficient contrast against the background — typically requiring a void of ≥ 1–2% of the through-thickness dimension.

AM-specific notes: Radiography is significantly less sensitive than CT for AM porosity. A 100 µm keyhole pore in a 10 mm wall section has a projected thickness fraction of 1% — at the limit of detection. LOF porosity, being planar, is essentially invisible in radiography unless the X-ray beam is tangent to the pore plane. Radiography is primarily useful as a quick screening tool for gross defects and missing internal features, not for AM process qualification. ASTM E1742 and EN 17630 govern radiographic testing of AM parts.

Ultrasonic Testing (UT)

High-frequency sound (typically 5–25 MHz for AM metals) is transmitted into the part and reflected from discontinuities. Time-of-flight and amplitude of reflections locate and size flaws.

AM-specific notes: Conventional contact UT suffers on as-built AM surfaces (see surface roughness issues above). Immersion UT, which couples through water, gives significantly better coupling and surface noise performance. Phased array UT (PAUT) combined with full matrix capture (FMC) and total focusing method (TFM) reconstruction can improve sensitivity and allow inspection of complex geometries without the geometric shadowing of conventional probes.

For LPBF-qualified aerospace parts (AMS 2630 / ASTM E2375 class scope), minimum detectable flaw size is typically 0.4–0.8 mm diameter using conventional UT on machined surfaces. Smaller flaws require immersion UT or phased array with TFM at shorter working distances. The anisotropic grain structure of AM parts means UT sensitivity must be demonstrated on AM-representative calibration blocks, not wrought blocks — this is an explicit requirement of ASTM E3166 (Guide for NDE of Metal AM Parts).

Fluorescent Penetrant Inspection (FPI)

FPI (or LPI — liquid penetrant inspection) reveals surface-breaking discontinuities by wicking a dye-penetrant into surface cracks or pores, removing the excess, and applying a developer that draws the penetrant back to the surface. Under UV illumination, fluorescent indications identify surface flaws.

AM-specific notes: FPI is highly sensitive to surface cracks in any material — it will detect cracks open to the surface as narrow as 0.5–1 µm. The limitation for AM is that the rough as-built surface holds penetrant in every surface valley, producing a high background that can mask relevant indications. In practice, FPI on as-built AM surfaces produces an uninterpretable result; the surface must be machined, electropolished, or at minimum bead-blasted to Ra < 6 µm. FPI is the standard method for inspecting post-machined AM aerostructure and engine brackets. ASTM E1417 / EN 3452 govern penetrant testing; NADCAP AC7114 is the audit checklist for aviation applications.

FPI detects surface discontinuities only. It will not detect sub-surface porosity, LOF, or cracks that do not break the surface. This is not a limitation unique to AM — it applies to all materials — but for AM parts with known sub-surface porosity risk, FPI alone is insufficient.

Magnetic Particle Inspection (MPI)

MPI applies a magnetic field to a ferromagnetic component. Surface and near-surface discontinuities distort the field and attract magnetic particles (usually fluorescent dry powder or wet suspension), making them visible.

AM-specific notes: Applicable only to ferromagnetic materials: 17-4PH stainless (martensitic condition), tool steels (H13, M2), low-alloy steels, and maraging steels. Not applicable to austenitic stainless (316L), titanium, aluminium, nickel superalloys, or cobalt-chrome alloys — which covers the majority of AM production volume. Where it is applicable, MPI is more sensitive than FPI for near-surface flaws (up to ~3 mm depth) and is less affected by surface roughness than FPI. ASTM E1444 / EN ISO 9934 govern MPI procedures.

Visual Inspection

Systematic visual examination under controlled lighting, with calibrated optical aids where applicable. Relevant standards: ASTM E2949 (visual examination of AM parts), ISO 17637 (visual testing of fusion-welded joints, applied by analogy to DED processes).

AM-specific notes: Visual inspection is the first-pass screen for gross build failures: delamination, missing features, gross dimensional deviation, obvious surface cracks, trapped powder spillage. It cannot detect internal defects, surface cracks below ~50 µm width (without optical aid), or process-induced microstructure anomalies. It is nonetheless mandated at every stage of production in aerospace quality systems — AS9100D requires documented visual inspection at defined hold points.

Decision Framework

Select NDT method based on defect type first, then geometry constraints, then cost:

Surface vs. volumetric defects — why a single method rarely covers both. FPI and MPI are surface-only. CT and UT are volumetric but less sensitive at the surface due to near-surface dead zones (UT) and resolution limits (CT). For fracture-critical AM parts, most aerospace specifications require both a volumetric method (CT or UT) and a surface method (FPI or MPI after machining). This is the approach taken in AMS 7010 for LPBF nickel and titanium parts.

CT Scanning for AM: Capabilities, Limitations, and Cost Reality

CT is frequently specified for AM parts without full understanding of what it can and cannot deliver.

What CT can do:

• Detect and size internal pores down to ~1.5–2× the voxel size
• Provide full 3D dimensional verification (GD&T from scan volume)
• Verify internal channel geometry and depowdering completeness
• Characterise porosity type (LOF vs keyhole) by pore morphology
• Detect delamination, cracking, and inclusions

What CT cannot do:

• Reliably detect cracks with apertures below ~10 µm (surface cracks need FPI)
• Achieve voxel sizes < 5 µm on parts larger than ~20 mm (geometric magnification limit)
• Inspect parts that contain highly absorbing features (dense inserts, trapped metallic powder in fine channels)
• Provide real-time or in-process inspection

Cost (2026 indicative): £300–800 for small parts (< 50 mm) at 5–10 µm voxel; £600–2000 for parts up to 150 mm at 20–50 µm voxel; series scanning with automated analysis can reduce unit cost to £200–400 at volume. CT cost must be weighed against defect risk and part value — it is routinely justified for aerospace components but rarely economical for commodity brackets.

AS9100 / NADCAP Context for Aerospace Parts

AS9100D (Quality Management Systems — Requirements for Aviation, Space, and Defense Organisations) does not specify NDT methods but requires that inspection is planned, documented, performed by qualified personnel, and traceable to the part record. NDT personnel qualification per EN 4179 / NAS 410 (aerospace) or ISO 9712 is required for all methods.

NADCAP (National Aerospace and Defense Contractors Accreditation Program) accreditation is required by most prime contractors for NDT on aerospace AM parts. Relevant checklists: AC7114 (Penetrant Testing), AC7114/5 (Computed Tomography, introduced 2019), and AC7114/7 (Radiographic Testing). NADCAP CT accreditation is now increasingly required for the combined dimensional verification and defect detection capability that CT provides; the audit scope covers equipment calibration, reference standard traceability, operator qualification, and procedure validation.

Use the NDT Method Selector to generate a method recommendation based on material, geometry, and defect type. For fracture-critical AM parts, cross-reference with the Fatigue-Critical Part wizard which steps through the full inspection requirement chain for aerospace certification.

Further reading

• NDT Method Selector — method recommendation based on material, geometry, and defect class
• Fatigue-Critical Part wizard — full inspection requirement chain for fatigue-critical AM parts
• ASTM E3166 — Standard Guide for Nondestructive Examination of Metal Additively Manufactured Aerospace Parts
• AMS 7010 — Laser Powder Bed Fusion Process for Nickel Alloy Parts