Steam smoothing for 3D printed parts

• 1
  How steam smoothing 3d printed parts melts the surface
• 2
  Which polymers respond to steam smoothing 3d printed parts
• 3
  Cycle parameters that decide the outcome
• 4
  Where steam smoothing 3d printed parts fails
• 5
  When to smooth and when to machine instead
• 6
  How to verify a smoothed part
• 7
  Finish routes compared

In powder-bed printing, a laser or fusing agent bonds polymer particles. Unmelted powder grains stay attached to the outer skin, which is why an as-built SLS or MJF part feels like fine sandpaper. Steam smoothing, also called vapor smoothing, attacks that skin rather than removing it mechanically.

The part sits in a sealed chamber. A solvent such as acetone for ABS or a controlled solvent blend for other polymers is heated until it fills the chamber as vapor. The vapor condenses on the cooler part surface. That condensed layer swells and softens the polymer chains in the top 5–30 μm.

Surface tension does the rest. Softened polymer flows toward the valleys between powder particles because that is the lower-energy state. Peaks round off. The result is a glossy, continuous skin that follows the original geometry rather than the powder grain.

Nothing is added and nothing is cut. The part loses a small amount of wall thickness as polymer redistributes, so a 1.2 mm rib can become 1.1 mm. That redistribution is the whole trade-off: you gain finish, you give up a little dimensional margin.

Because the reaction is limited to the surface, bulk properties stay close to the printed state. Tensile and flexural strength are typically unchanged or slightly improved, since the softened layer re-fuses and closes surface porosity. Impact resistance depends on how deep the solvent penetrates.

The process is not a coating and not a sanding step. It is a controlled partial dissolution. That distinction explains both its strengths, such as reaching internal channels, and its weaknesses, such as leaving no material behind if you overrun the cycle.

Amorphous thermoplastics smooth well because they have no sharp melting point. ABS is the classic case and acetone is the classic solvent. ASA behaves similarly with better UV resistance. Polycarbonate and PC-ABS blends respond to controlled solvent blends at lower temperatures.

Semi-crystalline polymers are harder. PA 12, the workhorse of SLS, has a crystalline phase that resists solvent attack. Dedicated vapor systems for PA 12 exist, but the window is narrower and the finish is satin rather than mirror. Go past the window and the part turns chalky.

PLA can be smoothed with solvents, but its low heat deflection temperature means the chamber must stay cool. PETG and TPU are possible with the right chemistry, though TPU tends to go tacky because the solvent stays mobile in the soft segments.

FDM parts can be vapor smoothed too. Layer ridges are larger than powder grains, so the cycle runs longer and the risk of sag on overhangs goes up. If you need a smooth FDM part with tight tolerances, machining is usually the safer route.

Material choice should follow the application, not the finish. If the part sees load, heat, or chemicals, pick the polymer first and ask whether it tolerates vapor at all. Some filled grades, glass-filled PA in particular, barely respond.

Three variables control the result: vapor concentration, chamber temperature, and dwell time. Concentration sets how fast the surface softens. Temperature sets how deep the solvent travels. Time sets how much polymer flows before the part is removed.

A typical ABS cycle runs at 40–60 °C with a dwell of 5–20 minutes, depending on part mass. Small brackets may need only a few minutes. A thick housing holds heat longer and can keep softening after the vapor is purged, so the dwell is shortened to compensate.

Solvent penetration scales roughly with the square root of time. Doubling dwell does not double depth, but it does widen the softened zone by about 40%. That is why a part that looks perfect at 10 minutes can lose edge sharpness at 20.

Parts should be supported during the cycle. Softened polymer will sag under its own weight if a thin section is unsupported, especially at 60 °C or above. Racks and fixtures are part of the process, not an afterthought.

After the vapor step, the part needs a purge and a drying period. Trapped solvent escapes slowly. A part that smells of acetone two days later was not dried long enough, and residual solvent can affect later bonding or painting.

Thin walls are the first casualty. Below roughly 1 mm, the softened zone spans the full wall thickness and the section can warp or collapse. If a design has 0.8 mm ribs, vapor smoothing is the wrong finish.

Sharp edges round over. A 90° corner becomes a radius of roughly 0.2–0.5 mm after a full cycle. That matters on mating faces, seal grooves, and any edge that locates against another part.

Deep blind holes and long internal channels smooth less than the outer skin. Vapor reaches them, but solvent exchange is slower and condensate drains poorly. Expect a gradient rather than a uniform finish inside a 10:1 channel.

Text and laser markings lose contrast. Fine lettering below about 0.5 mm stroke width can bridge or blur. Mark after smoothing, not before, if legibility matters.

Tolerance-critical features should be masked or machined after the cycle. A vapor-smoothed bore will not hold ±0.005 mm. The realistic path is to smooth the cosmetic surfaces and then re-cut the functional ones.

Vapor smoothing earns its place on parts with complex internal geometry, low load, and appearance requirements. Ducting, covers, manifolds, and anatomical models are good candidates because the finish has to reach surfaces that no tool can touch.

It is the wrong choice when the part is a functional interface. Bearing seats, threaded bosses, and dowel holes need dimensions that stay put. Smoothing softens the skin those features depend on.

At GreatLight we run both routes. Printed parts get vapor smoothed when the customer wants a sealed, paintable skin. When the same geometry has to hold ±0.005 mm, we machine it from 6061, 316L, or PEEK on 5-axis centers instead. The material and the tolerance decide, not the finish.

A hybrid works for many programs. Print and smooth the cosmetic shell, then machine the mating features in a second op. That gives a sealed surface and a true locating face without compromising either.

Cost is not the deciding factor at prototype volume. The deciding factor is whether the softened surface can do its job after the solvent is gone. If the answer is unclear, test one part before committing a batch.

Measure before and after. Record wall thickness and critical dimensions on the as-built part, then repeat after the cycle. A shift greater than 0.05 mm on a functional face means the dwell was too long.

Check gloss with a gloss meter at 60°. A consistent reading across the part confirms even vapor distribution. Patches that read dull usually sat in a low-flow corner of the chamber.

Weigh the part. Vapor smoothing removes nothing, but drying removes solvent. A part that keeps losing mass over 48 hours is still outgassing and should not be assembled yet.

Cut one part open if the program is large. A cross-section shows the softened depth under a microscope and confirms whether internal channels were reached.

Finally, run the same fit and function checks you would run on any finished part. A glossy surface that fails a leak test is still a failed part.