Pumps for High and Ultra-High Vacuum

Most of us are familiar with processing in the vacuum range up to around 1.33 x 10^-3 Pa (1 x 10^-5 torr) or slightly lower (Fig. 1). There are also lessons to be learned from understanding the demands of ultra-high vacuum applications (Fig. 2). Let’s explore what’s involved.

What is an Ultra-High Vacuum? ^[1]

Practical high vacuum levels (Table 1) range down to approximately 1.33 x 10^-4 Pa (1 x 10^-6 torr) while ultra-high vacuum (UHV) levels are in the vacuum range characterized by pressures of about 10^-7 Pa (7.5 x 10^-10 torr) and greater.

Notes:

[a] The SI unit of pressure is the Pascal (1 Pa = 1 N m^-2)
[b] Normal atmospheric pressure of 1 atmosphere is 101,325 Pa or 1013 mbar (1 bar =10⁵ Pa)
[c] Normal atmospheric pressure of 1 atmosphere is 760 Torr (1 Torr = 133.3 Pa)
[d] Ultrahigh vacuum is defined as the pressure range between 10^-6 Pa (Europe) and/or 10^-7 Pa (USA) to 10^-10 Pa.

These vacuum levels demand the use of special materials of construction and processing techniques such as preheating (i.e. bake-out) of the entire system for several hours prior to processing to remove water and other trace gases, which adsorb on the surfaces of the chamber. At these low pressures the mean free path of a gas molecule (Table 2) is approximately 40 km (24.8 miles), so gas molecules will collide with the chamber walls more frequently than they collide with each other. Thus, almost all gas interactions, therefore, take place on various surfaces in the chamber.

Notes:

[a] Maintenance of a clean surface is defined as being able to achieve a pressure of less than 10^-9 torr.
[b] Collision free conditions are defined as being able to achieve a pressure of less than 10^-4 torr.
[c] ML designates a single monolayer corresponding to the maximum attainable surface concentration of adsorbed species bound to the substrate.

Achieving Ultra-High Vacuum

In order to achieve ultra high vacuum conditions, special materials and pumping procedures are needed. Seals and gaskets used between components in a UHV system must prevent even trace leakage. Therefore, nearly all such seals are all metal, with knife-edges on both sides cutting into a soft gasket, typically copper. These all-metal seals can maintain integrity to UHV ranges.

With respect to pump-down, initially, the vacuum chamber will be pumped down to 1 Pa (7.5 x 10^-3 torr) using a mechanical pump. Then the chamber will be pumped down to approximately 1 x 10^-4 Pa (7.5 x 10^-7 torr) using one or more of the following types of pumps: turbomolecular, ion, titanium sublimation, non-evaporative getter and/or cryopumps. UHV pressures are measured via ion gauges, either a hot filament or an inverted magnetron type.

At this point, the vacuum chamber is enclosed in heat-resistant material (boards commonly known as ovens), and baked to a temperature of about 180°C (355°F). After 24+ hours of baking, the ovens are removed, and the chamber allowed to cool down. Once at room temperature, the chamber should have a pressure in the UHV region. The process of baking removes gas atoms from the chamber wall surfaces (if the chamber was not baked it would literally take months before the chamber achieved UHV conditions).

Tips for achieving ultra-high pressure include:

1. Use a small chamber size (to minimize surface area);
2. High pumping speeds using multiple vacuum pumps in series and/or in parallel;
3. High conductance tubing to the pumps — large diameter, short runs with minimal obstructions (valves, etc.);
4. Use of low outgassing materials (stainless steel, aluminum, titanium, etc.);
5. Avoid creating pits of trapped gas behind bolts, welding voids, etc.;
6. All internal metal parts should be electropolished after machining or welding;
7. Use low vapor pressure materials (ceramics, glass, metals, Teflon if unbaked);
8. Bake the system to remove water or hydrocarbons adsorbed to the walls;
9. Chill chamber walls to cryogenic temperatures during use;
10. Use gloves to avoid all traces of hydrocarbons, including skin oils in fingerprints.

Ultra-High Vacuum Challenges

Vacuum chambers manufactured from stainless steel are widely used in high vacuum, ultra-high vacuum and even extreme high vacuum applications. However, there are other construction materials that may also have advantages [4] including aluminum, titanium, copper and specialized metals. For example, aluminum with less entrapped less hydrogen and carbon releases less hydrogen, water vapor and hydrocarbon vapors to the vacuum environment reportedly allowing faster cycles to high vacuum and ultra-high vacuum levels faster and quicker bake-out with less pumping.  Copper/copper alloy vacuum chambers are used for nonmagnetic, radio frequency (RF), and high thermal conductivity applications. Titanium vacuum chambers, with their very low hydrogen (and secondary gas) permeation rates, offer gettering properties in addition to other attributes.

Outgassing (from either internal surfaces or materials of construction) is a significant problem for UHV systems. Outgassing from internal materials is minimized by careful selection of those with low vapor pressures (e.g. glass, stainless steel, ceramics). Outgassing can come from many plastics as well. Vessels lined with a highly gas-permeable material (e.g. palladium) with a high capacity for hydrogen absorption also create special outgassing problems.

Outgassing from surfaces is a subtler problem. At extremely low pressures, more gas molecules are adsorbed onto the walls that are floating in the chamber, so the total surface area inside a chamber becomes more important than its volume. Water vapor is a significant source of outgassing, especially whenever the chamber is opened to air since any water vapor present absorbs other contaminates and evaporates from surfaces too slowly to be fully removed while pumping at room temperature, but just fast enough to present a continuous level of background contamination. Removal of water and similar gases generally requires baking at 180°C (355°F) to 400°C (750°F) while vacuum pumps are running. During use, the walls of the chamber may be chilled using liquid nitrogen to further reduce outgassing.

Hydrogen and carbon monoxide are the most common background gases present after baking from sources such as stainless steels. Helium diffusion is not considered significant.

Typical Ultra-High Vacuum Uses

Many surface analytic techniques would not be possible except for reduced surface contamination achieved by UHV. For example, at 1.33 x 10^-4 Pa (1 x 10^-6 torr) it takes only about one (1) second to cover a surface with a contaminant. Analysis techniques such as XPS (X-ray photoelectron spectroscopy), AES (Auger electron spectroscopy), SIMS (Secondary Ion Mass Spectrometry), TPD (Thermal Desorption Spectroscopy), ARPES (Angle Resolved Photoelectron Spectroscopy) and thin film techniques require high purity over a sustained analysis time and thus are prime candidates for UHV.

UHV is also used for particle accelerators, gravitational wave detectors, atomic physics and microscopy (e.g. atomic force and scanning tunneling). In addition, UHV is required for most surface science testing for two principal reasons:

1. To enable atomically clean surfaces to be prepared for study, and to maintain those surfaces in a contamination-free state for the duration of the experiment.
2. To permit the use of low energy electron and ion-based experimental techniques without undue interference from gas phase scattering.

For most surface scientific testing there are a number of factors necessitating a high vacuum environment:

1. For surface spectroscopy, the mean free path of a probe and detected particles (ions, atoms, electrons) in the vacuum environment must be significantly greater than the dimensions of the apparatus in order that these particles may travel to the surface and from the surface to detector without undergoing any interaction with residual gas phase molecules. This requires pressures better than 1.33 x 10^-2 Pa (10^-4 torr).
2. Most spectroscopic techniques are also capable of detecting molecules in the gas phase; in these cases, it is preferable that the number of species present on the surface substantially exceeds those present in the gas phase immediately above the surface – to achieve a surface/gas phase discrimination of better than 10:1 when analyzing.
3. In order to begin experiments with a reproducibly clean surface, and to ensure that significant contamination by background gases does not occur during an experiment, the background pressure must be such that the time required for contaminant build-up is substantially greater than that required to conduct the experiment (i.e. of the order of hours).

Variation with Pressure ^[3]

The following are some of the familiar interactions that take place with a change in pressure. Our goal of presenting this information is to ultimately estimate the period of time it will take for a clean surface to become covered completely with a monolayer contaminant (Equation 6) as expressed in Table 2.

Gas Density

The gas density will change and can be estimated from the ideal gas law (Equation 1) as:

               N          P

(1)  n = —– = —–

               V          kT

where:

n = number of molecules per cubic meter
P = pressure [N m^-2] k = Boltzmann constant (= 1.38 x 10^-23 J K^-1)
T = Temperature, °K

Mean Free Path (Particles in the Gas Stream)

The average distance that a particle (i.e. atom, electron, molecule) travels in the gas phase between collisions can be determined (Equation 2) as:

                 kT

(2)  λ =   ——

                 πPσ

where:

λ = mean free path, in meters (m)
P = pressure (N m^-2)
k = Boltzmann’s constant (=1.38 x 10^-23 J K^-1)
σ = Collision cross section in square meters (m²)
T = Temperature, °K

Surface Contamination (Incident Flux) Incident Molecular Flux on Surfaces

One of the critical factors in determining how long a surface can be maintained clean (or, alternatively, how long it takes to build-up a certain surface concentration of adsorbed species) is the number of gas molecules impacting on the surface from the gas phase.

The incident flux (Hertz-Knudsen formula) is the number of incident molecules per unit time per unit area of surface without regard for the angle of incidence. It is important to note that the molecular flux is directly proportional to the pressure. For a given set of conditions (i.e. P, T) the flux is readily calculated using a combination of the ideas of statistical physics, the ideal gas equation and the Maxwell-Boltzmann gas velocity distribution and is expressed in Equation 3.

                   P

(3)  F =   ———

               √2πmkT

where:

F = incident flux in molecules per square meter – second [molecules m-2 s-1] k = Boltzmann’s constant (=1.38 x 10^-23 J K^-1)
m = molecular mass, in kg
T = Temperature, °K

Gas Exposure

The gas exposure is a measure of the amount of gas that a surface has been subjected to. Although the exposure may be given in the SI units of Pascal-seconds (Pa-s), the normal and far more convenient unit for exposure is the Langmuir (L), where 1 L = 10^-6 torr -second and is expressed in Equation 4.

(4)   E = 10⁶ • t

where:

E = Exposure, in L

Sticking Coefficient & Surface Coverage

The sticking coefficient, S, is a measure of the percentage of incident molecules which adsorb upon a surface (i.e. a probability that lies in the range  0 – 1 , where the limits correspond to no adsorption (0) and complete adsorption (1) of all molecules). In general, S depends upon many variables including such items as surface coverage, temperature, crystal structure, etc.

The surface coverage, Θ, of an adsorbed species may be specified in several ways, as the number of adsorbed species per unit area of surface (e.g. in molecules cm^-2) in Equation 5a or relative to the atom density in the topmost atomic layer of the substrate, Equation 5b.

                    Actual Surface Coverage

(5a)  Θ = ————————————-

                   Saturation Surface Coverage

Θ is expressed in molecules per square centimeter and lies in the range of 0 to 1.

                           Number of Absorbed Species per Unit of Surface

(5b)  Θ_max = —————————————————————

                        Number of Surface Substrate Atoms per Unit Area

Θ_max is expressed in molecules per square centimeter and is usually less than 1 (but can for certain species such as hydrogen, exceed 1).

You might be wondering, after all, this, how long it will take for a clean surface to become covered by a monolayer of contaminant? The answer is dependent upon the flux of gas phase molecules incident upon the surface, the actual coverage corresponding to the monolayer and the coverage-dependent sticking probability. However, it is possible to get a minimum estimate of the time required by assuming a unit sticking probability (i.e. S = 1) and noting that monolayer coverage are generally of the order of 10¹⁵ per cm² (or 10¹⁹ per m²) yielding Equation 6.

          t            10¹⁹

(6)  ——  ≈  ——-

        ML             F

where:

ML = monolayer thickness
t = time (s)
F = Incident flux (Equation 3)

Final Thoughts

For heat treaters, consideration of high and ultra-high vacuum systems may not seem something worthy of their time, but the principles that are needed to achieve these vacuum ranges are worth noting and the practices necessary for sustaining these levels are valuable “lessons learned” in helping any vacuum system perform to its ultimate ability.

References