UHV Chambers & Components — Ultra-High Vacuum Fabrication

Below about 1 × 10⁻⁹ mbar the residual gas in a chamber no longer comes from leaks. It comes off the internal surfaces. Reaching UHV is therefore a fabrication and treatment problem before it is a pumping problem, and that is the part we control.

What ultra-high vacuum actually requires

Four things together: an all-metal-sealed chamber, an internal surface with the deformed layer removed, a bake-out cycle to drive off adsorbed water and hydrogen, and a pumping train that can reach the required base pressure. Miss any one and the chamber stalls at high vacuum no matter how good the others are.

The surface is the whole problem

Water is the dominant residual gas in a freshly assembled chamber. It desorbs from the oxide layer and from any deformed metal left by machining. Electropolishing removes 20–40 µm and leaves a chromium-enriched, lower-area surface; a bake at 250–400 °C then drives off what remains. A mechanically polished chamber and an electropolished one can differ by an order of magnitude in pump-down time.

Sealing: why elastomers are not an option

Viton permeates and outgasses above about 150 °C, which is below any useful bake-out temperature. Every joint on the high-vacuum side of a UHV system therefore has to be ConFlat with a copper gasket, so the seal is metal and survives the bake. KF is fine on the roughing side and nowhere else.

Materials

316L is the default for its corrosion resistance and its behaviour after electropolishing. 316LN is specified where magnetic permeability must stay low, such as near an electron column or a magnetic lens. Aluminium is used for very large chambers where weight drives the design, but it needs its own welding and cleaning route and its outgassing is harder to control at the low end.

Components alongside the chamber

UHV systems need more than a vessel: CF flanges, viewports, feedthroughs, bellows, manipulators and the pumping and gauge ports, all with metal seals and all bakeable to the same temperature as the chamber. Building the chamber and its ports in one machine setup avoids the misalignment that appears when a port is fitted after the shell is finished.

What we control in-house

304 / 316L stainless steel and 6061 / 5083 aluminium fabrication
Clean-environment TIG welding with full penetration and no micro-porosity
Helium mass-spectrometer leak testing to 1 × 10⁻⁹ mbar·l/s
In-house electropolishing and passivation — not subcontracted
Chamber bake-out to 150–400 °C for UHV outgassing control
Precision-machined ISO-KF, ISO-K and CF (ConFlat) flanges
Zeiss CMM verification to ±0.02 mm across large work envelopes
12 m gantry CNC machining — 12 m × 4 m × 2 m envelope
SEMI-referenced documentation and inspection protocols

Frequently asked questions

What pressure counts as ultra-high vacuum?

Below about 1 × 10⁻⁹ mbar (roughly 7.5 × 10⁻¹⁰ Torr). High vacuum ends near 1 × 10⁻⁷ mbar; between the two there is a region where elastomer seals are still usable but the chamber will not go lower without a full metal-sealed, baked construction.

Why does a UHV chamber need to be baked?

Because water and hydrogen desorb from the internal surfaces, and that desorption is what holds the pressure up. Heating to 250–400 °C with continuous pumping drives them off. Without a bake, the surfaces re-adsorb water each time the chamber is opened to air.

What is 316LN and when is it needed?

316LN is a low-carbon, nitrogen-bearing stainless steel with controlled magnetic permeability. It is specified where a magnetic field would disturb the process or the measurement, for example around an electron beam column.

Can a UHV chamber be welded rather than bolted?

Yes, and welded construction is normal for the shell because a weld is a permanent metal seal with no gasket to bake out. It also means the chamber cannot be taken apart, so the design has to be right before welding — which is why we machine the sealing faces after welding rather than before.

How long does it take to pump a chamber to UHV?

A properly electropolished and baked chamber of a few tens of litres typically reaches 10⁻⁹ mbar within a day of pump-down. A chamber with a smeared surface finish and no bake can take weeks and may never get there at all.