New 175-Cycle Study Separates Ti-6Al-4V Powder Conformance From Stability

A peer-reviewed study published on September 3, 2026 followed Ti-6Al-4V ELI powder through 175 reuse cycles in a certified medical-device manufacturing facility. The measured means remained within the powder and produced-material limits used by the researchers, but several properties did not remain stationary. Powder oxygen rose, material oxygen rose faster, and the produced material became harder (The International Journal of Advanced Manufacturing Technology).

That distinction changes the useful purchasing question. A reuse process should not be approved by asking only whether cycle 175 passed a specification. It should be controlled as a material-lifecycle system whose top-up ratio, handling environment, sampling location and product response travel with every result.

This was one controlled route, not a universal reuse limit

The campaign used one Ti-6Al-4V Grade 23 powder lot, one EOS M290 machine and a climate-controlled facility certified to ISO 13485 requirements. After each build, recovered powder was processed and topped up with virgin powder from the same production lot. The authors classify the approach as SAE AMS7031 Scheme 3.

The machine, process parameter set, powder lot, sieve mesh and processing equipment remained constant. Production nevertheless included a mix of spinal cages, maxillofacial implants, qualification specimens and other medical devices over about 2 years. Powder and witness material were characterized after 23, 104 and 175 reuses.

Those boundaries matter. The amount of virgin powder added after each build was not recorded, only three reuse points were sampled, and the product mix created different thermal histories. The study can show a long-run trend for this facility and route. It cannot establish that 175 cycles is a safe limit for another machine, powder lot, sieve, top-up rule or part family.

Flow improved while chemistry and hardness drifted

Between reuse 23 and reuse 175, the Hall flow time fell from 29.81 to 28.20 s/50 g, a 5.4% improvement. The angle of repose fell 3.0%, apparent density rose 0.8%, and the D10, D50 and D90 particle-size measures showed net changes no greater than 2.1%. Particle elongation decreased by about 2.9%, while the form-factor distribution narrowed around an essentially unchanged mean.

The chemical and material responses moved differently. Mean powder oxygen rose from 842.2 to 910.8 ppm, an 8.2% increase. Mean oxygen in the produced material rose from 851.5 to 1,065.0 ppm, a 25.1% increase. Mean microhardness increased from 325.6 to 337.6 HV 0.5, or 3.7%, across the measured depths.

The important mechanism is not simply degradation. Handling and sieving can improve the measured flowability of the retained powder population, even while repeated exposure and fusion increase interstitial content. A flow result can therefore improve at the same time that the built material changes. “Better powder flow” is not evidence that the material state is unchanged.

The specification result also needs precision. The means used for conformance remained within the researchers’ selected limits. Individual hydrogen measurements did not all behave that way: one powder result at reuse 104 exceeded its upper limit, and a material sample taken from the bottom of the build at reuse 175 measured 0.0135 wt.% hydrogen, above the 0.0125 wt.% limit used in the study, while the corresponding top sample remained within it. Averages can hide location-dependent excursions.

Reuse count is a weak material identity

“Reuse 50” sounds precise but omits the exposures that change powder. Two facilities can report the same cycle count while using different virgin-to-reused ratios, build heights, laser-on time, sieves, atmospheres, storage practices and product mixes. The paper itself warns that its results are specific to the reuse approach and processing environment.

A stronger record binds seven fields:

This record makes the lot a moving material state rather than a static name. It also prevents an acceptable average from erasing a local result that needs investigation.

For suppliers working with special titanium alloys, the same logic applies beyond powder: a chemistry certificate identifies the starting material, not every downstream state. For buyers combining near-net-shape production with titanium CNC machining, final dimensions cannot replace evidence about the feedstock and build state inherited by the part.

Passing current limits is not a lifetime claim

The researchers tested powder flow, apparent density, particle size and morphology, interstitial chemistry, and witness-specimen microhardness after stress relief and duplex annealing. They did not report tensile, ductility, fatigue, corrosion, biological or device-performance endpoints for this campaign. The measured hardness increase is attributed to oxygen pickup, but it is not a complete qualification of a medical component.

The limited sampling also means the observed trend should not be extrapolated beyond 175 cycles. At each selected reuse point, the produced-material chemistry came from one witness specimen; the two measurements represented top and bottom locations, not independent build replicates. Replicates from composite powder samples quantified measurement repeatability; they did not map variability across the powder batch. The unrecorded top-up mass prevents a precise material-balance reconstruction.

The defensible conclusion is therefore narrower than “powder can be reused 175 times.” Under one controlled Scheme 3 route, extensive reuse coexisted with specification-level conformance and measurable drift. A buyer or quality team should release the defined process envelope, trend both powder and built material, specify sampling locations, and treat a changed top-up or handling route as a qualification change.

Source boundary: This analysis uses the complete publisher PDF and publication page for DOI 10.1007/s00170-026-19064-8. It preserves the three-point sampling limit, the unrecorded top-up quantities, the individual hydrogen excursions and the absence of tensile, fatigue, corrosion or device-performance testing.