Ion pumps use a combination of electric and magnetic fields to ionize free gas molecules, and then adhere those ions to cathode plates. There are two basic types of ion pumps. Holding ion pumps (HIPs) are equipped with all of the magnets needed for them to work as pumps. Distributed ion pumps (DIPs) are designed to use the magnetic fields of magnets used for other purposes (generally large bend magnets). Therefore DIPs do not work if the magnet that they are designed to use is turned off.
Ion pumps are generally constructed from an array of ion pump cells. Each cell consists of a cylindrical anode (generally made of stainless steel) and a titanium cathode plate at each end of the cylinder. A magnet is then situated so that the magnetic filed lines run axially through the cylindrical anode (see diagram).
Electrons are emitted by the cathode due to the 5000 to 7000 V potential between the anode and cathode. The magnetic field gives the electrons a long helical trajectory as they move toward the anode, which improves the odds that the electrons will collide with free gas molecules. When the electrons hit the gas molecules, the molecules are ionized. The positive ions then drift to the cathodes. The ions are accelerated by the large electric field that is present, and hit the cathodes with sufficient energy to both imbed themselves in the cathode plate, and sputter off titanium molecules. The titanium molecules sputter coat the inner surfaces of the pump, leaving a fresh reactive layer of titanium, which will adhere any non-ionized gas molecules that happen to hit the surfaces.
Ion pumps cannot be regenerated. They have a finite capacity to absorb gas. When that capacity to absorb gas has been exhausted, the pump must be replaced. The pumps can also be worn out as titanium molecules are sputtered off the cathode surfaces. The cathode plates gradually become thinner, and can be worn completely through. At that point, the ion pump will stop functioning properly. Ion pumps can be used at pressures as high as 10^-5 Torr, but use at higher pressures will decrease the lifetime of the pump considerably, and is therefore to be avoided.
Additionally, ion pumps can be used to measure vacuum. The current measured between the anode and cathode can (with proper calibration) be used to determine the vacuum pressure within the pump.
- Holding Ion Pumps (HIPs)
Holding ion pumps use an internal magnet to provide the magnetic field axial to the cell cylinder. Holding ion pumps are generally purchased as fully assembled units from vacuum equipment suppliers and installed onto the beam pipe.
- Distributed Ion Pumps (DIPs)
Distributed ion pumps contain anode cylinders and cathode plates much like those found in HIPs, but use the magnetic fields of surrounding magnetic devices, rather than providing separate magnets exclusively for the pump. Often DIPs are designed and custom built to fit a particular piece of the beam line, rather than being purchased from a supplier already assembled. Distributed ion pumping is often used in storage rings, where the magnetic field of the ring bend magnets can be used by distributed ion pumps to provide pumping in all regions where the beam is being bent. Conveniently, the bend regions are also the regions with high synchrotron light losses. The synchrotron light hits the vacuum chamber, and causes outgassing, leading to a need for higher pumping capacity in the bend regions than the straight sections. Distributed pumps are used for this purpose in the PEPII HER ring.
The HER also has a custom DIP near the interaction point, called the Radial Ion Pump, which was intended to use the magnetic fringe fields of the BaBar solenoid magnet to provide the magnetic field axial to the pump cell cylinders. This custom DIP is meant to provide ion pumping in a region were high vacuum is needed, but there is not enough space to install a conventional HIP. However, upon commissioning PEP II, it was determined that the Radial Ion Pump did not work as intended, and so it is not used.
The Refrigerated Baffle is also known as the Differential Pumping Station (DPS). It is a refrigerated section of LINAC disc-loaded waveguide that is used to maintain a difference in vacuum pressure between the LINAC and BSY, while allowing the beam to pass through it. The BSY does not require as high a vacuum as the LINAC, because there are no RF sections in the BSY, and therefore no need to prevent arcing in RF sections. The BSY vacuum system is maintained in the 10-4 Torr range, while the LINAC vacuum system operates in the 10-7 Torr range. The BSY uses oil diffusion pumping stations, which do not create as high a vacuum as the LINAC ion pumps. There is also concern that the oil from the diffusion pumps could contaminate the LINAC vacuum. If oil contamination were to occur, it would be very difficult to achieve the ultrahigh vacuum conditions that are currently maintained in the LINAC. In order to keep the higher pressure gas molecules from the BSY from leaking into the better vacuum in the LINAC, a section of vacuum pipe with low gas conductance is used to connect the two areas. The low gas conductivity of the refrigerated baffle is achieved by cooling the vacuum pipe. The cooling of the refrigerated baffle makes stray gas particles that hit the beam pipe wall give up energy and remain on the beam pipe surface, thereby preventing them from leaking from the high pressure area to the low pressure area.
For more information on the chiller system used to cool the refrigerated baffle, see the Utilities chapter of the AS02 manual.
Mechanical roughing pumps are used initially when pumping down a vacuum chamber after it has been vented. They work from atmospheric pressure to about 10-3 Torr. Generally, when roughing out a section of vacuum pipe, the vacuum tech will bring a cart with several mechanical vacuum pumps to the area that needs to be pumped down. Several different varieties of mechanical pump are used in stages to pump down the chamber before the high vacuum pumps can take over and the roughing valve can be closed.
Mechanical pumps come in various types, each of which work best in a different pressure range. At very high pressures, pumps with a rotor that physically displaces air volumes from the pump input port to its output port are used. Dry pumps use a design in which air volumes are mechanically moved while no oil is introduced into the vacuum system. At slightly lower pressures, trubopumps are used. Turbopumps contain a series of rotating blades that move gas molecules from the input to the output port by transferring momentum to individual gas molecules that contact the blades such that the molecules will migrate to the output side of the pump.
Cryogenic vacuum pumps work using a combination of condensation and cryosorption to remove gas molecules from the vacuum chamber. Water vapor, nitrogen, and argon are captured when they condense on cryopump surfaces. Noncondensable gasses such as hydrogen, neon, and helium ar captured by cryosorption. The portion of the cryopump used for cryosorption uses a high surface area material, such as charcoal, at a very low temperature. The grooves in the porous material help to confine the very low energy gas molecules, and keep them stuck on the surface even though their temperature is not low enough for them to be captured by condensing on the surface.
Cryogenic pumps do work by accumulating gas molecules on their internal surfaces, and therefore can become saturated with these materials. In order to regenerate cryopumps, they are isolated from the rest of the vacuum system, and then heated up to room temperature. This releases the condensed and adsorbed gasses, which can then be pumped out of the cryopump using a mechanical pump. The cryopump can then be cooled back down and returned to normal service.
SLAC currently doesn’t use any cryogenic vacuum pumps on accelerator systems aside from the specialized cryogenic system used to cool the refrigerated baffle.
Non-evaporable getter pumps are passive pumping devices. They pump gas by chemically interacting with molecules that hit the NEG surface and binding them to the NEG surface. The NEG’s surface is covered with a chemically active metallic compound which adsorbs gas can be made from many different transition metals, including zirconium, aluminum, titanium, vanadium, and others. The active compound is adhered to some substrate material in such a way that a large surface area of active compound is exposed. One common method of making NEG material is rolling powdered active TSP material into the surface of a stainless steel ribbon. The ribbon can be mounted in the vacuum chamber, or it can be cut up and modified to create a custom TSP with greater surface area, which can where needed into a complicated vacuum assembly. The PEPII HER uses custom NEG pumps, which have been fabricated from this material.
NEGs can be regenerated by heating. When the NEG material is heated, the adsorbed hydrogen is desorbed into the vacuum chamber. The other adsorbed gasses diffuse further into the NEG material, freeing up surface molecules to adsorb more gas. The released hydrogen can be pumped away by other vacuum pumps in the system (such as ion pumps) while the NEG is kept hot. When the hydrogen has been pumped away by the other pumps, the NEGs can be cooled and returned to normal service. NEG pumps are generally designed to include a heating element which is used to regenerate the pump.
NEG pumps are particularly effective in pumping hydrogen gas. They also pump oxygen, nitrogen, and carbon dioxide at a slower pumping rate. They do not pump methane or inert gases.
Titanium sublimation pumps work by sputtering a thin film of fresh (unoxidized) layer of titanium onto the inside of the vacuum chamber. The titanium layer is chemically active, and adsorbs gas molecules which come in contact with it. A titanium layer can become saturated with gas molecules. When the titanium layer becomes saturated, a titanium filament in the pump is heated in order to sputter a fresh titanium film over the old one. In vacuum systems running at higher pressure, it can be necessary to continuously sputter fresh titanium in order to keep the pumping speed up. In higher vacuum systems, continuous sputtering is not necessary. The pump can be turned on periodically to deposit a fresh layer of titanium when the pumping speed seems to begin to degrade. This greatly extends the life of the pump, since the titanium source is depleted much more slowly.
The PEPII LER uses TSPs. They are regenerated (or "flashed") periodically. These particular pumps seem to have unshielded wires which cause crosstalk to nearby magnets. When some of the TSP pumps are on, they have a steering effect on the beam. Collisions can be maintained, but average luminosity is lower, and having to constantly correct for TSP steering effects can interfere with other tuning. We therefore often attempt to flash the LER TSPs when the LER beam is off for other maintenance.
- Cold Cathode Gauges
Cold Cathode Ionization gauges (also known as Penning Gauges) consist of an anode ring placed between two cathode plates. A permanent magnetic field is applied perpendicular to the cathode plates. Electrons come off of the cathode plates, and spiral toward the anode. As they do this, the electrons ionize the gas molecules that they collide with, and the ionized gas changes the measured current flow. The change in current can be calibrated to read out in pressure units.
- Hot Filament Gauges
A hot filament vacuum gauge uses gas ionization to measure pressure. A hot filament gauge consists of a heated cathode filament which is near an anode. The anode is a coil of wire surrounding a collector wire. The collector is wired through an ammeter to ground. The hot cathode emits electrons, and those electrons are accelerated toward the anode. If those electrons collide with a gas molecule, the gas molecule is ionized. After ionization, the positively charged ions move toward the collector, and the electrons move toward the anode. The current measured to ground from the anode is proportional to the amount of ionization occurring within the gauge, which is in turn proportional to the vacuum pressure within the gauge.
- Pirani Gauges
Pirani Gauges indirectly measure pressure by measuring the thermal conductivity of the surrounding gas. A Pirani gauge consists of a Wheatstone bridge with one resistor component of the bridge within the vacuum system. Resistive heating heats the bridge resistor. Heat loss to the surrounding gas changes the temperature, and therefore resistance, of the resistor. The resistance change unbalances the bridge, and changes the current flow measured by a microammeter. Readout of the microammeter is then calibrated to read out in pressure units.
- Convection Gauges
Convection Gauges are similar to Pirani gauges in that they use a bridge circuit to measure heat lost from a resistor to the gas molecules in the vacuum system. However, convection gauges improve upon the pirani gauge setup by introducing a second resistor into the vacuum system, and using it to compensate for convecetive heat loss.
Two vacuum manifolds are use in the LINAC. There is an 8 inch diameter manifold located along the ceiling of the klystron gallery. This large manifold is known as the “Main Manifold”. The ion pumps used to maintain the ultra-high vacuum found in the LINAC beam pipe are mounted on the main manifold and pump on it directly. Lines run from the main manifold through the LINAC penetrations, and into the tunnel, where they connect the vacuum chamber of the main manifold to the vacuum of the smaller 5 inch manifold in the tunnel. The 5 inch manifold runs the length of the LINAC just next to the beam pipe. There are connections from it to the LINAC beam pipe, where the high vacuum actually needs to be maintained.
Control system interlocks exist both to protect the vacuum system from widespread accidental vents, and to protect the beam line vacuum valves from being damaged by beam. Vacuum sensors are used to detect possible leaks and actuate fast and slow valve closure if a pressure rise is detected. If beam line vacuum valves are in, or are in the process of falling in, they take away the beam permissive, and thereby prevent beam production.
- Isolation Valves (IV’s)
Beam line isolation valves (IVs) are used to either manually or automatically isolate portions of the vacuum system in order to maintain good vacuum in most regions in the event of a leak in one region. They can be actuated manually, or by vacuum system interlocks. The IVs consist of a copper sealing plate, which travels on rails to form a seal against a stainless steel plate. Valve movement is actuated by an air cylinder.
- Slow Valves
Slow valves are paired with fast valves to isolate vacuum regions in the event of a leak. They are both actuated simultaneously, but the fast valve moves into place more rapidly than the slow valve. Fast valves operate quickly, but do not form a good vacuum seal. Slow valves move in more slowly, but form a leak tight vacuum seal that prevents further gas leakage between adjacent vacuum regions.
- Fast Valves
Fast valves are designed to rapidly detect and interrupt shock waves and flying debris resulting from sudden major leaks opening. Fast valves need not form a vacuum tight seal in order to interrupt shock waves traveling within the vacuum pipe. In the event of a leak, interlocked slow valves will form a leak tight seal soon after the fast valves have closed.
Fast valves are spring loaded and held open against the spring force by an electromagnet. When electromagnet current is interrupted, the valve is forced closed by the spring. LINAC fast valves are actuated by the main manifold gauge controller, and are designed to close in 9 msec. They have no rolling or rubbing metal parts, and use a deforming indium ring to make a clean seal. After the valve has been cycled several times, the indium seal can be remelted in place to take out the dents that have formed when it was closed.
BSY fast valves are actuated by McClure switches and close in 20-30 msec. They consist of an aluminum plate which slides into place on rails.
- Manual Valves
Manual valves can be opened and closed locally by vacuum system technicians. They are used to isolate portions of the vacuum system, or isolate particular pumps or pump ports. Most manual valves do not have control system readback.