Computer Services Calendar Safety     Search PSFC


Plasma Science and Fusion Center

Massachusetts Institute of Technology


Alcator C-Mod

Alcator C-Mod Criteria for Design of Vacuum Components


This document is meant to be a guideline for design and construction of components that interface with the CMOD vacuum system. It does not intend to cover all situations but instead is designed to start one thinking about the problems encountered in constructing a successful device that will operate in the Alcator vacuum environment without causing any unwanted effect on the quality of that vacuum.

Material Selection

The Alcator vacuum vessel is made from 304L SS, as are most of the support devices and diagnostic assemblies in the vacuum. Other than Molybdenum on the limiters and in the divertor, this is the predominate material used in the vessel. Other acceptable materials are listed below along with the known unacceptable materials. The lists below are considered comprehensive but by no means complete. Any material not listed here is considered unacceptable until tested for out gassing and approved by the vacuum lab. This test would be done in our out gas testing system, which will test the material to 200C in UHV with RGA analysis. The material must show an out gassing rate < 1E-8 Torr-L/s to be accepted. 

Acceptable Materials

304,304L,304LN,316,316L,321 SS

Virgin Teflon (no colors added)


OFHC copper

Beryllium copper

Mild steel




Non porous ceramics such as:

Alumina and Beryllia

Macor machinable glass

Synthetic Mica

Braided fiber glass tubing-pure with no binders or colors


Unacceptable Materials  




303 and 306 series SS


Torlon , Vespel, PEK, etc.

G-10, Micata, Bakalite, Plexiglass, Lexan

Kapton, PVC, 5-minute epoxy, RTV Silicone rubber

Plated and anodized connectors


The use of epoxies in general has to be very limited to the extent that they can be used to make a seal on a material that is difficult to seal by any other means. An example would be very thin Beryllium X-ray windows or a fiber optic bundle vacuum interface. These epoxies must be specifically designed for use in vacuum with operating pressures in the low 10E-9 Torr range. The list at this time is very short and any use must be on a case by case basis.

The use of Teflon coated wire has to be approved since the inventory of Teflon is already rather large in Alcator and we have selected a very specific wire with a very thin coating to help reduce the degas time after pump down. Since the source has changed a few times please check with our staff when you are ready to construct your system.


Vacuum Seals

All the vacuum seals on Alcator are metal seals with minor exceptions for certain small windows with Viton seals and gate valve seals that are also Viton. The majority of the seals are copper gasketed Conflat seals. Non circular flange seals are special Helicoflex spring supported "C" seals. The use of ISO flanges like KF and NW is permitted but when connected directly to the Alcator vacuum they must be fitted with metal gaskets. The most popular metal seal for this purpose is the Helicoflex "C" seal but care must be taken to have the correct size centering ring. Since the metal seals adhere to the actual ISO dimensions and some of the O-ring/centering ring combinations sold by US vendors use US size O-rings this can mean a major difference in the centering ring fit to the metal gasket. To prevent this problem you should purchase the centering ring with the first gasket to insure a proper fit.

When connecting small tubes and gas lines to Alcator the use of Swagelok fittings is discouraged with Cajon VCR fittings being the preferred method. This is because of the ability to remake VCR fittings easily and reliably by changing the gasket. VCR fittings are also preferred over miniconflat flanges also, whenever possible.

When connecting a diagnostic beamline to Alcator with a ceramic electrical break the use of a flexible bellows coupling is necessary to reduce the mechanical load on the break. Care must therefore be taken to support the atmospheric load on the bellows as well.  

Material Preparation

All materials used inside the vacuum system must be thoroughly cleaned using generally accepted Ultra High Vacuum cleaning materials and practices. The use of cleaning solvents such as chlorothane and acetone is acceptable as long as they have a final rinse and wipe with Texwipe lint free cloths using 200 proof reagent grade ethyl alcohol. Materials fresh from a machining operation can be washed in hot soapy water and rinsed in multiple rinses of hot distilled water to remove the heavy machine oils. Again this should be followed by the ethyl alcohol final wipe. Porous materials like ceramics should not be cleaned in water but rather in solvents followed by ethyl alcohol and an air bake to drive off the solvents. Ultrasonic cleaning is also necessary for parts that have areas not accessible to mechanical cleaning.

As a general practice SS should be electropolished before welding or assembly. If all surfaces have been machined so that the mill finish of the metal has been removed then this step can be eliminated but it is recommended you do it anyway to help reduce surface area. As for other metals some sort of etching or mechanical method should be used to remove the mill finish if they are not machined on all surfaces. Above all do not weld or assemble things before electropolishing. This step has to be done first and the parts thoroughly cleaned before assembly. One material that is sometimes overlooked for electropolishing is SS tubing. All SS tubing must be electropolished.

When parts are cleaned for welding they should be kept clean in plastic bags. Do not wrap them in aluminum foil before welding since the foil can rub off and cause problems in getting a proper weld. After leak testing is done then the parts can be kept clean using the foil.

Parts that get electroplated such as silver, copper or nickel must be cleaned and vacuum baked to at least 200C for a few hours to drive off any residual moisture and make sure there won't be any blistering before assembly.

Parts that are going to be exposed to the direct contact with the plasma or in the near shadow must be able to withstand a bulk temperature of 300C and a surface temperature at the melting point of most metals. With this in mind devices that are positioned near the plasma need to have some Molybdenum limiter blocks mounted on the plasma-facing surface.

Optical devices should be concerned with the potential for coatings such as Molybdenum and Boron from the discharge cleaning and Boronization. If this is a problem then the design of a shutter or retraction system might be wise. Keep in mind these systems must be reliable so simplicity is key.


Design and Assembly

All parts that are welded to flanges or make a seal must be internally welded. Internal means on the vacuum side only. There are a few exceptions where it is impossible because of geometry that prevents internal welding. However, when this calls for an external weld it should be a full penetration weld with no follow up welding on the other side. Short tack welds can be used on the other side to add strength when necessary but in no circumstance should both sides of the joint be completely welded. Generally speaking most welding is TIG fusion welding without the use of fillers. The object of this procedure is to insure there are no trapped volumes to cause a slow bleed into the vacuum and make He leak testing impossible. During the welding process there should be a dedicated clean SS brush available to brush off the oxides while the part is still hot.

On the large horizontal racetrack ports the blank flange should start out with a thickness between 1.375" to 1.5" and get ground to between 1.375" to no less than 1.250". An initial grinding to insure parallel surfaces and fully flat-machined surface on the vacuum side is important if there is any doubt about the flatness to begin with. After this is done keep in mind that you want to be within the final thickness tolerance before you start cutting holes but leave at least .050" for a final grinding to the required vacuum finish that will be necessary because of distortion that will be present from machining and welding. This also means that any welding should be recessed from the vacuum surface so it won't get ground off. This final grinding should bring the vacuum surface of the flange to a smoothness of ~16 RMS. At installation, the final condition of the sealing area will be inspected, and all scratches will be removed by hand. Adequate soft protection, such as a ring of Plexiglas, should be in place on the sealing area during assembly and storage, to reduce the chance of scratches. When designing a "reentrant port" flange, one that is mounted on the internal horizontal port without the use of a port extension, keep in mind the need for some sort of cantilevered installation arm to facilitate the use of the overhead crane for installation of the flange. Close review by the vacuum lab personnel will be necessary for this design to be useful.

During assembly if there are any welds that cannot be rewelded at a later time and are vacuum seals they must be leak tested before assembly continues. Depending on the application this may mean temperature cycling and testing as well. Keep this in mind and try to avoid a design where this occurs. When this can't be avoided keep in mind how the part is to be leak tested when making the design.

Good vacuum design means there are no trapped volumes such as bottom tapped holes without vented hardware. When designing parts with tapped holes, it is important to use a coarse thread, whenever possible, to lessen the effect of galling on hardware and parts of similar metal. Parts that are a slip fit should have some way of venting as well as long tubes. Devices that consume power must have proper thermal shunts or be able to take the heat.

Final Assembly and Vacuum Testing

When the device is completed it will be vacuum tested on one of our bake out/test chambers. This test will include an initial He mass spec leak test with a minimum detectable leak rate of <1E-9 Torr-L/s. A bake out of the whole device to 200C will be performed with RGA analysis. Another leak test at temp followed by a final test back at room temp will conclude the testing. Following the successful testing the device must be stored such that it is kept clean to UHV standards and installed in like manner.



77 Massachusetts Avenue, NW16, Cambridge, MA 02139,


massachusetts institute of technology