Guides / Materials

Cleanliness, Surface Finish and Material Rules for Vacuum Parts

September 26, 2026

A vacuum part can pass every dimensional check on the report and still be rejected. The reason is almost always something the drawing did not say: how clean is clean, which face needed the finish, what the part was allowed to touch on its way out of the building.

Cleanliness behaves like a dimension in this niche. It has a target, it has a measurement, it has a cost, and it can be out of spec. The difference is that most drawings dimension the geometry carefully and leave the cleanliness to be guessed.

Why cleanliness is a dimension

At atmosphere, a fingerprint is cosmetic. At 10-6 Torr it is a source of hydrocarbons that keeps evaporating for as long as the chamber is under vacuum. The pump has to remove it, the pumpdown takes longer, and if the part sits above the wafer the contamination lands somewhere it matters. This is outgassing, and it turns a cleaning shortcut into a schedule problem weeks later in a fab, not a defect the shop can see on the bench.

Geometry contributes to the same problem. A blind tapped hole with a screw in it holds a pocket of gas that leaks slowly into the chamber and never fully clears. That is why vented screws and through holes appear on UHV hardware, and why a design that is fine at 10-3 Torr can be unusable at 10-9 Torr without a single dimension changing.

Then there is particle generation, which is a machine shop question rather than a design question. Particles come from a short and predictable list:

• Blast media left in a corner from the previous job
• Coolant residue dried into a pocket
• Shop cloths and gloves, especially anything with silicone on it
• The bag, if it is the wrong bag or if it was sealed in the wrong room

Silicone deserves its own line. It migrates, it is almost impossible to remove once it is on a part, and it shows up as a defect far downstream. Shops that run semiconductor work keep it out of the building entirely. Shops that do not have never had a reason to think about it.

Surface finish, face by face

Three Ra values cover most vacuum drawings, and they are not three steps on one scale. They are three different manufacturing decisions.

Callout	How it is reached	What it is usually for
Ra 0.8 um	A normal finishing pass	General surfaces, non-critical faces
Ra 0.4 um	Slower pass, sharper tool, inspection step	Sealing faces, O-ring grooves
Ra 0.2 um and below	A secondary process, usually outside	Wetted interiors, wafer contact, UHV

The number is only half the callout. The other half is the area it applies to, and that is where the money is. One O-ring groove face at Ra 0.4 um costs very little. The same callout sitting in the title block prices the whole part at Ra 0.4 um, including faces nobody will ever look at.

Three kinds of face need a finish for three different reasons, so they deserve three separate notes:

Sealing faces. The finish exists so the elastomer seals and keeps sealing through thermal cycles. Scratches across the groove matter more than the average Ra.

Wafer contact faces. The finish exists so the part does not generate particles or mark the wafer. Flatness usually travels with it.

Interior wetted surfaces. The finish exists so there is less surface area holding adsorbed gas and less texture trapping residue. This is the one that usually means electropolish.

Finish also has to be read together with the coating. Type III hard anodize grows the part and changes the surface, so the drawing has to state whether the Ra and the dimension apply before or after coating. Parts that measure correctly and do not fit are almost always this mistake.

Materials that are accepted

The material list in this niche is short, and the reasons are specific.

6061-T6 is the default for chamber bodies, lids and brackets. It machines predictably and anodizes well. Where corrosion resistance matters more than strength, 5052 or 5083 is often the better answer, and for large plate parts the temper and the flatness of the incoming stock matter as much as the grade, because a plate that is not stress relieved will move during machining.

316L is the default for wetted gas paths, and the L is not optional. Low carbon keeps the grain boundaries clean through welding, which is what makes the joint survive in a corrosive line. 304L is acceptable for structure and for vacuum hardware that never sees process gas, and using it on a wetted path to save money is a decision that gets found later.

Titanium appears in specific corrosive areas and in parts that need the strength to weight ratio, usually wafer handling arms.

PEEK, Vespel and alumina are where the job leaves a metal shop. Insulators, seal carriers, wafer contact pads and rings. A plastics or ceramics house is the right supplier here, and a metal shop that says yes to an alumina ring is usually about to subcontract it without telling you.

Restricted material lists sit on top of all of this. A fab may exclude particular alloying elements, particular mills or particular coating chemistries. Each restriction narrows the supplier pool, and a shop that buys from whoever has stock this week cannot serve the requirement at all. That is worth knowing before the first order rather than after the first rejection.

Specifying cleaning

Cleaning is a scope of work, so it should read like one.

Degrease is what any shop does. Solvent, wipe, done. Suitable for frames, jigs and anything that stays at atmosphere.

Precision clean is a multi stage process with controlled water, controlled drying and a controlled room at the end. It costs real money and takes real time, and the shop has to have somewhere to do it.

A named fab procedure is the strictest case. The customer supplies the document, the shop follows it, and both sides know what was agreed.

Verification is usually a wipe test, and it is worth knowing what one actually checks. A controlled wipe over a defined area, then an inspection of the wipe, tells you about visible residue and particles. It does not tell you about molecular contamination, so a clean wipe test and a silicone problem can coexist.

Packing is the last chance to undo all of it. Double bag as standard, cleanroom bag where the fab asks for it, label on the outside so nobody opens the inner bag to find out what is inside, and nitrogen purge for the parts that need it. Fab cleanliness classes come from ISO 14644, and if your receiving area works to one of them the packing has to match or the part is dirty by the time it arrives.

Writing it on the drawing

A vacuum part becomes quotable when these notes are present. Most drawings that arrive here are missing at least one.

1. Material grade, temper and whether a mill cert or heat lot traceability is required.
2. Surface finish called against specific faces, with a note saying whether it applies before or after coating.
3. The finishing chain by name: anodize type, electropolish, passivation, and which faces are masked.
4. The cleaning level by name, and the procedure document if there is one.
5. The packing standard, including the bag and whether a purge is required.
6. Handling restrictions, with silicone called out explicitly if the part goes near a fab.

The three that go missing most often are the cleaning level, the packing standard and the before or after coating note. Each one is a single line. Each one, left out, is a quote that comes back too low and a part that comes back wrong.

If you would rather not audit a shop against this list yourself, that is the job this site does: you send the drawing and it goes to shops that already run this class of part.

Frequently asked questions

Keep reading

• Molybdenum and Titanium Parts in Vacuum Tools: Where Each Metal Earns Its Place
• Choosing Between 6061, 5083 and MIC-6 Aluminum Plate for a Vacuum Chamber