Achieving Ultrahigh Vacuum

Conductance can be thought of as the maximum quantity of a gas that can pass through a tube or opening in a given time period.  Conductance always reduces the 'effective' or net pumping speed of a system.  In UHV and XHV, particles are moving in molecular flow (i.e., the gas particles behave independently, travel in straight lines at a speed related solely to the temperature and have many more interactions with the walls of the vessel than with each other) so conductance is based on statistical probabilities of particles entering a vacuum pump.  HV and UHV pumps must be connected to the vacuum vessel with fittings and tubing that are as short and wide as possible to avoid significant pumping speed loss.

Creating ultrahigh vacuum (UHV) and extreme high vacuum (XHV) typically require a combination of vacuum pumps operating over specific ranges.  Rough vacuum pumps (eg: dry scroll pumps and rotary vane pumps) are required to evacuate a chamber or beamline from atmospheric pressure to 0.1 Torr or lower, where high-vacuum pumps can be engaged.  High vacuum pumps (e.g. turbo pumps, diffusion pumps, and cryopumps) are most effective between about 0.1 Torr and 10^-7 Torr.  Turbo pumps are very good at extracting the gas particles released during a vacuum system bake-out (getting to UHV and XHV pressures requires system bake-out).  Pumps with the highest pumping speed in UHV and XHV are sputter ion pumps, titanium sublimation pumps, and nonevaporative getter pumps.

Virtual leaks are gas load sources emanating from inside the vacuum chamber.  They are insidious, as they cannot be detected using traditional UHV/XHV tools like helium leak detection.  Virtual leaks encompass outgassing (desorption from surfaces inside the vacuum chamber or beamline and diffusion of gas from bulk materials inside the chamber or the chamber itself) and gas sources trapped by mechanical structures inside a chamber (e.g., mechanical polishing, or areas of poor conductance).   Virtual leaks are typically detected using a leak up or rate-of-rise test on an HV or UHV system.

Ion pump controllers provide the precision high voltage source that allows ultrahigh vacuum sputter ion pumps to operate.  Choosing an ion pump controller requires some knowledge of the ion pump design (connector type, polarity, pump size etc.).   Since ion pump controllers are often used to measure UHV and XHV pressure, choosing a controller with the highest current resolution and/or a step voltage capability will improve the accuracy of the pressure reading.

Since every UHV/XHV vacuum system is ultimately also a high vacuum and a rough vacuum system (since no UHV/XHV pumps can operate near atmospheric pressure), monitoring rough vacuum components with inexpensive thermal gauges can act as a low-cost insurance in protecting your UHV/XHV system.  Costly shutdowns and outages in beamlines and UHV/XHV systems can often be avoided if a failure or defect in a rough vacuum section is identified early and safeguards are put in place.

Hot filament and cold cathode gauges will both work through high vacuum, but for UHV and XHV, choices are reduced to the hot filament gauge or the inverted magnetron style of cold cathode.  The hot filament gauge will provide an accurate reading to lower pressure but will act as an emitter of heat and photons.  The advantages of the inverted magnetron gauge are its ruggedness and tolerance for less pristine environments (it is rebuildable).  On a single gauge system, this is less of an issue, but when installing gauge cabling for an entire beamline or particle accelerator, the 'extra' signal lead that the hot filament gauge requires can add significant cost.

Commercially available "UHV" gauges have a lower limit of about 10^-10 Torr, with only minimal variations based on whether a hot filament or inverted magnetron gauge is used.  Using ion pump current to measure vacuum is effective to about the same limit if a voltage optimization circuit (e.g. the Agilent step voltage function) is used.  Commercially available residual gas analyzers (RGAs) can measure partial pressures to about 10^-12 Torr, but their use as total pressure gauges declines decades earlier based on nonlinearity of the response curve. Measuring XHV pressure is typically done using variations on the hot filament ionization gauge that attempt to minimize the impact of soft x-rays on the gauge reading by physically displacing the ion collector. The lowest calculated pressure in scientific literature is probably the 10^-17 Torr, value derived from matter/antimatter annihilations observed in an experiment at CERN (ATRAP).

Combination vacuum gauges incorporate multiple technologies (typically two) inside a single gauge housing to measure vacuum pressure over a wider range than any single technology can.  Most combinations include a rough vacuum gauge (high accuracy capacitance manometer or less expensive thermal gauge) and a medium- or high vacuum gauge (low range capacitance manometer, or hot filament ionization gauge for example).  Combination gauges are best suited to smaller, self-contained high vacuum systems: True UHV and XHV systems typically separate the functions of monitoring the rough and high vacuum sections of the system from the UHV/XHV sections.

Leakage current is a low-level discharge observed in sputter ion pumps that distorts (artificially increases) the calculated pressure in a vacuum system. Sputter ion pumps function with very large (KV level) potential differences between the anode and the cathode.  In theory, leakage currents can be caused by the feedthroughs and cabling associated with ion pumps; however, in practice, it is most often the result of 'whiskers' or 'dendrites' formed during operation of the pump that provide a minor 'short' between the anode and cathode.  The 'whiskers' are most often removed through a process known as high potting the pump.

It's a relatively simple matter to convert ion pump current into a pressure reading, but unfortunately the current we measure also includes a component arising from leakage currents. Leakage current has a couple of sources, and each of these contribute to the distortion of the pressure reading. Leakage current can be dealt with by identifying and possibly eliminating the causes, either replacing worn out components or "high-potting" the pump. Since leakage current is proportional to operating voltage, lowering the voltage after start-up to the lowest possible level minimizes distortion of the pressure reading. The step function on the Agilent 4UHV and  Agilent IPCMini controllers allows them to read pressure down to 1e^-11 mbar with high resolution.

The ConFlat all-metal seal design was developed by William R Wheeler and the engineers at Varian Vacuum Products Division in the early 1960s.  Its main advantage over contemporary designs, and an advantage it maintains to this day, is the ability to remain leak-tight after repeated bake-out cycles on UHV systems.  All-metal seals are required in UHV systems because of the bake-out requirement, and because of o-ring seals' permeability, particularly to helium. ConFlat flanges use an oxygen-free high thermal conductivity copper gasket and knife-edge flange to achieve an ultrahigh vacuum seal.  Each face of the two mating flanges (typically stainless steel) has a knife edge, which cuts into the softer metal gasket, providing an extremely leak-tight, metal-to-metal seal.  ConFlat flanges operate down to 10⁻¹³ Torr (10⁻¹¹ Pa) pressure. Bake-out temperature is limited by the choice of gasket material.