Amorphous Metal 3D Printing Process: A Working Guide

• 1
  Key takeaways
• 2
  What the amorphous metal 3D printing process actually does
• 3
  Geometry that survives the cooling rate
• 4
  Powder, particle size, and atmosphere control
• 5
  What goes wrong and how it shows up
• 6
  When this process is the wrong answer
• 7
  Running the process, step by step
• 8
  Amorphous metal AM vs. CNC machining from bar stock

Metallic glasses have no grain structure. Atoms sit in a disordered arrangement instead of a lattice, which is why the alloy is strong, elastic, and hard to corrode. The catch is that this structure is metastable. Heat it slowly or cool it slowly and crystals appear, and the properties you wanted disappear with them.

Laser powder bed fusion gives you the fast cooling you need. A laser melts a thin layer of powder, and the melt pool touches the solid metal below and the gas above. Heat leaves through the substrate, so the pool solidifies in microseconds. Cooling rates land between 10^3 and 10^6 K/s, which is fast enough to freeze the disordered structure in place.

That is the whole trick. The process is not exotic equipment. It is a normal metal printer running a narrow window where the thermal history stays inside the glass-forming range. Step outside the window and you get a crystalline part that looks fine and tests badly.

We run this class of work alongside conventional metal AM in Dongguan and Singapore. The same machines build crystalline alloys every day. What changes for amorphous feedstock is the parameter set, the atmosphere, and how much post-machining the part will need.

• 1
  Zr-based alloysThe most common glass formers. Good strength, moderate cost, sensitive to oxygen.
• 2
  Fe-based alloysCheap and hard, magnetic. Lower glass-forming ability, so thinner layers and tighter control.
• 3
  Cu- and Ti-basedUseful for wear and medical work. Narrower process windows than Zr.

Cooling rate scales with surface area over volume. A thin rib dumps heat quickly. A 20 mm cube cannot, so the core cools slowly and crystallizes. This is the single hardest constraint in the process, and it decides most part designs before you ever pick a laser parameter.

Keep walls between 0.5 mm and 3 mm where the drawing allows. Above roughly 4 mm in the build direction, expect a crystalline core and a glassy skin. That gradient is sometimes useful, but you have to design for it instead of discovering it in a test report.

Sharp internal corners are a second problem. They concentrate heat and trap unmelted powder. Add a 0.5 mm radius where the function allows it. For overhangs below 45 degrees, add supports; unsupported melt pools curl and drag powder into the surface.

Drain holes matter more than people expect. Trapped powder inside a hollow amorphous part is hard to remove and adds weight. Put 2–3 mm holes at the lowest point of each cavity and orient them downward in the build.

• 1
  Wall thickness0.5–3 mm. Below 0.4 mm the layer may not form cleanly.
• 2
  Solid sectionsAvoid above 4 mm. Use ribs or lattice instead.
• 3
  Internal radii0.5 mm minimum to reduce heat concentration.
• 4
  Drain holes2–3 mm, at the lowest cavity point.

Gas-atomized powder is the standard feedstock. The atomization step already quenches each droplet fast, so the particles arrive amorphous. If the powder is partly crystalline, no laser setting will fix the part. Ask for an XRD trace on the lot, not just a chemistry certificate.

Particle size usually sits between 15 μm and 45 μm for a 20–30 μm layer. Coarser cuts flow better but need more energy to melt fully. Fine cuts give better surface finish and are more sensitive to oxygen pickup.

Build chamber oxygen should stay below 100 ppm, and lower is better for Zr-based alloys. Argon is the usual choice. Every time the chamber is opened, the powder picks up moisture, so keep the feedstock in sealed containers and dry it before reuse.

Reuse is possible but not unlimited. After several cycles, oxygen content rises and the fines fraction shifts. Track the cycle count and blend used powder with virgin stock rather than running 100 percent reclaimed material.

• 1
  Particle size15–45 μm for a 20–30 μm layer.
• 2
  Chamber oxygenBelow 100 ppm; lower for Zr-based alloys.
• 3
  Powder checkXRD plus chemistry on each lot.

Crystallization is the failure mode that matters most, and it is quiet. The part builds without sparks or alarms. It looks correct. Then a bend test shows low ductility, or an XRD scan shows sharp peaks where there should be a broad halo.

Porosity comes from two directions. Gas trapped in powder particles leaves spherical voids. Lack of fusion between layers leaves irregular voids, usually near the edges of the melt pool or at the end of a scan vector. Cross-section a witness coupon from each build and look at both.

Residual stress builds because the cooling is so fast and uneven. Thin parts are fine. Thick or tall parts can crack at the base or lift off the plate. Preheating the substrate to 200–300 °C helps, but it also slows cooling in the part, so there is a trade.

Surface roughness is a given. Downskin surfaces sit near Ra 15–20 μm, upskin near Ra 8–12 μm. If the drawing calls for Ra 0.8–1.6 μm, plan a machining allowance of 0.2–0.5 mm per finished face.

• 1
  Crystal peaksXRD shows sharp peaks. Recheck laser power and scan speed.
• 2
  Lack of fusionIrregular voids at layer boundaries. Raise energy density.
• 3
  Edge crackingPreheat the plate or reduce part height per build.

If the part is a simple prismatic block with tight tolerances and no internal channels, print it and machine it and you have paid twice. Cut it from bar stock or a casting instead. Amorphous bar is available in common diameters, and CNC gets you to ±0.005 mm without a powder lot qualification.

If the part needs a fully glassy 30 mm section, no parameter set will deliver it. The physics does not allow it. Redesign with ribs or a lattice, split the part into thinner pieces, or pick a crystalline alloy that tolerates slower cooling.

If you need thousands of identical small parts, powder bed AM is the wrong cost structure. Die casting or metal injection molding wins once tooling is amortized, provided a suitable alloy exists.

Where the process earns its cost: thin-walled, complex, low-volume parts that need wear resistance or elastic behavior, and internal features that cannot be machined. Lattice heat sinks, flexure elements, and conformal-cooled inserts fit that description.

• 1
  Good fitThin walls, internal channels, low volume, wear or elasticity needs.
• 2
  Poor fitSimple blocks, thick sections, high-volume small parts.