In the previous article I argued that reusing conforming powder beats sending it off for recycling. There is a catch, and for titanium it has a name: oxygen.
Titanium loves oxygen. Each time powder sits in a hot build chamber, gets spattered, is sieved and is handled, it can pick up a little more. Oxygen in solid solution strengthens titanium alloys, which sounds like good news, but it also reduces ductility. A part can pass its strength test and still fail where it matters.
So I find it useful to think of each powder batch as starting with an oxygen budget. The budget is the gap between the oxygen in the virgin powder and the limit your parts must meet. Every build spends some of it.
What the limits actually are
Two references come up most often for Ti-6Al-4V:
- ELI, Grade 23: typically 0.13 wt% oxygen (1,300 mass ppm). This is the grade usually relevant to implants, and it is sometimes specified for aerospace too.
- Standard, Grade 5: typically 0.20 wt% oxygen (2,000 mass ppm).
It is tempting to call these the “medical” and “aerospace” limits. That is a shortcut I would avoid. The real limit is whatever the drawing, the approved material specification and the customer require. The powder acceptance limit and the finished-part limit are also not the same thing: processing changes the chemistry, and you need a margin for measurement uncertainty.
One more distinction that causes confusion: the oxygen reading on the machine’s chamber sensor, in volume ppm, is not the oxygen content of the powder, in mass ppm. Both matter, but they answer different questions.
What the literature shows
Published reuse studies on Ti-6Al-4V give a consistent warning. In one study, ELI powder went past 0.13 wt% oxygen after eight recycles. In another, across nine reuse stages, oxygen rose from about 0.099 to 0.126 wt%. Tensile strength went up, but reduction of area fell from about 10% to about 3%. A third, ten-cycle study reported oxygen rising from 0.107 to 0.129 wt%, close to the ELI limit.
The lesson for me is not that eight or ten cycles is the answer. It is that a favourable short trial does not establish a 100-build process, and that strength alone is the wrong pass criterion.
A model you can do on paper
Before collecting plant data, I like a model simple enough to check by hand. Let:
- f be the fresh fraction added to each 200 kg charge,
- Cv the oxygen content of virgin powder,
- Cn the oxygen content of the mixed powder before build n,
- d the increase in oxygen of recovered powder during one build.
Then the mixed powder before the next build is:
Cn+1 = (1 − f)(Cn + d) + f Cv
This settles to a steady value of Cv + (1 − f)d/f before each build, and Cv + d/f in the recovered powder. In words: the less fresh powder you add, and the more oxygen each build picks up, the higher the pool settles.
I ran it with illustrative values: virgin powder at 0.090 wt%, a target for the mixed powder of 0.120 wt% (a 0.130 part limit with a 0.010 allowance), and a ceiling of 0.130 wt% for recovered powder.
- At a pickup of 0.002 wt% per build, the pool settles around 0.104 wt%. Plenty of budget left.
- At 0.004 wt% per build, it settles around 0.118 wt% with normal losses. That is very close to the target.
- At 0.006 wt% per build, it settles above the target. Oxygen, not loss, now decides how much fresh powder you must buy.
The surprising part: lower losses can raise oxygen
Here is the result that made me write this article. In the previous piece I showed that cutting handling losses from 5 kg to 3 kg per build saves about 200 kg of powder a year. That is true, but only if oxygen is not the limit.
Less loss means less fresh top-up, which means less dilution. At a pickup of 0.004 wt% per build, cutting losses to 3 kg pushes the pool over the illustrative target after about 32 builds. To stay under it, you need to replace about half a kilogram of extra powder per build on purpose. The real saving falls from 200 kg to about 147 kg a year.
At 0.006 wt% per build, cutting losses saves nothing at all in the long run, because oxygen control already decides how much fresh powder goes in.
This is why I think powder management deserves an engineer, not just a procedure. Two good improvements can partly cancel each other.
Low-oxygen powder can pay for itself, sometimes
The same model lets you put a value on a premium powder. In one illustration, lowering the oxygen of virgin powder from 0.090 to 0.070 wt% is enough to make the normal top-up sufficient again, and avoids about 53 kg of extra replacement a year. At CAD 200/kg, that justifies a premium of only about CAD 4.60/kg before qualification costs.
That is a useful answer. It does not say low-oxygen powder is always worth it. It says exactly how much more you could pay for it in this situation, and gives the supplier and the buyer a number to discuss.
How I would manage the budget
- Measure oxygen, nitrogen and hydrogen in virgin, recovered and blended powder, and in parts or coupons from the same builds. Learn your own pickup per build; do not borrow it from a paper.
- Attack the sources of pickup: leaks, moisture, how long powder sits warm in the machine, open transfers and storage. Sealed handling protects the budget more than any blending rule.
- Track history, not only averages. Know which lots are in the pool and the oldest exposure in it.
- Never blend back into compliance. If recovered powder is out of specification, it leaves the qualified loop. Deoxidation or reprocessing is a separate, separately qualified process.
- Test what the part needs. For titanium parts in service, ductility and fatigue tell you more about the cost of oxygen than strength does.
Oxygen is a budget you spend slowly and cannot easily earn back. Knowing your spending rate is what lets you keep powder working as long as it should, and not one build longer.
The model and figures are illustrations with stated assumptions, not a qualified release specification or measurements from a particular facility.
Further reading
- Harkin et al. (2020). Reuse of Grade 23 Ti6Al4V powder during the laser-based powder bed fusion process. Metals 10, 1700.
- Harkin et al. (2022). Powder reuse in laser-based powder bed fusion of Ti6Al4V. Materials 15, 2238.
- Koushik et al. (2023). Effective Ti-6Al-4V powder recycling in LPBF additive manufacturing considering powder history. Sustainability 15, 15582.
- ASTM F3001-14(2021). Standard specification for additive manufacturing Ti-6Al-4V ELI with powder bed fusion.
- NASA-STD-6030 (2021). Additive manufacturing requirements for spaceflight systems.
