Components And Subsystems
Non Evaporable Getters: A coating that behaves like a pump
Getters are reactive materials used for removing traces of gas from vacuum systems. For systems which need to be opened to air for maintenance, non evaporable getters which work at high temperature are used instead. When getters are used within more general vacuum systems, such as in semiconductor manufacturing, they are introduced as separate pieces of equipment in the vacuum chamber, and turned on when needed. Dr Sunil Patel of the Vacuum Science Group, ASTeC at The Cockcroft Institute discusses how getters can be used as pumps with the vacuum environment.
Non Evaporable Getter (NEG) coatings are applied to the internal surface of a vacuum vessel and are most commonly used in accelerators and storage rings (such as DIAMOND, ESRF, SOLIEL, DESY and many others). Essentially these coatings are thin films of a getter material, which once activated has the ability to pump a vacuum vessel. This reduces the ultimate pressure attainable in the vessel thereby maintaining a “cleaner” environment, furthermore NEG coatings also help to minimise outgassing and secondary electron yield from the internal surface of the vessel, this eliminates beam scattering effects, which also helps to improve beam lifetime.
NEG coatings are particularly effective for pumping narrow gap vacuum vessels such as those used in modern insertion devices (eg. undulators and wigglers), which have limited conductance that makes pumping using conventional means (e.g. ion or turbo pump) extremely difficult.
One particularly novel application of these coatings is to use them in a titanium vacuum vessel for space exploration, where the coating will maintain a pressure in the high vacuum region for the lifetime of the mission (a period of approximately 20 months).
Some other advantages of NEG coatings are as follows, they are vibration free which makes them suitable for very delicate experiments (e.g. infrared), they do not require a power supply once activated (hence they can be used in remote environments such as space), and they are not influenced by strong magnetic or electric fields, which makes them particularly suitable for use on accelerators.
A good NEG coating must have good adhesion to the surface of the vacuum vessel, it should also be non-toxic and have high mechanical resistance. The cost of applying the coating also needs to be taken into consideration, especially during construction of new facilities or upgrades to existing facilities.
Typically NEG coatings are made using binary or tertiary alloys of elements in group 4 and 5 of the periodic table. The vast majority of coatings are based on titanium, zirconium, hafnium and vanadium, which are sputter coated (most commonly using magnetron sputtering) onto the internal surface of the vacuum vessel. These elements are used because of their high oxygen solubility and diffusion properties. NEG coatings are particularly effective at pumping gases such as hydrogen, carbon monoxide and carbon dioxide which often remain in a vacuum vessel after bakeout and that often limit the base pressure. They are less effective for pumping noble gases such as methane and argon.
The pumping effect of these coatings is initiated by first “activating” the coating. This involves heating the coating to a specific temperature at which hydrogen molecules on the surface are desorbed and oxygen molecules diffuse into the bulk of the coating. The precise activation temperature depends on the chemical composition of the surface.
Detailed surface analysis of these coatings has been used to investigate the actual pumping mechanism. Using x-ray photoemission spectroscopy it has been shown that during activation the surface changes from an oxide rich surface to one that is more metallic in nature. It is believed that these metallic sites act as the active sites for adsorption of residual gas molecules and hence result in the pumping effect.
Once activated the coating will continue to pump until saturation of the surface occurs, at which point the coating can be reactivated by re-heating the surface to the activation temperature (or slightly higher). Although activation typically occurs over a given temperature range, ideally the temperature selected should be as low as possible, as this has been shown to maximise the number of times the coating can be reactivated following venting2.
The activation temperature is also influenced by the nature of the substrate onto which the coating is applied. With the popularity of aluminium increasing as a material from which to manufacture vacuum vessels from, lower activation temperatures are required for coatings applied to this substrate relative to the more conventional stainless steel.
Currently our research aims to determine both the optimum activation temperature for different coatings and determination of the initial sticking co-efficient, which gives a good indication of pumping speed. These measurements are carried out using a dedicated vacuum system developed over the past 5 years by the group, which can accommodate both NEG coated vacuum tubes and flanges. The initial sticking co-efficient is measured using the static gas expansion method, in which high purity gases are exposed to the activated NEG coating. Total and partial pressure measurements are then used to calculate the initial sticking co-efficient along with Test Particle Monte Carlo simulation. This is a complicated calculation that takes into account the volume of the vacuum vessel and associated vacuum components such as elbows used on the RGA and tabulations on valves, the area of the NEG coated surface, the gas flux and changes in the flux during the course of the experiment and also gauge sensitivity for different gases.
One of the experimental challenges with using NEG coatings is how to prevent contamination and poisoning of the coating during bakeout of the rest of the vacuum system. Here at ASTeC, we have developed a procedure for preventing this happening that involves ensuring that the NEG coating is maintained at a steady temperature of 800C whilst the rest of the system is baked at 2500C. After 48 hours the system is allowed to cool to 1500C, at which point filaments in the residual gas analyser (RGA) and pressure gauge are carefully degassed.
The temperature of the NEG coated surface is then gradually raised initially from 800C to 1500C for a period of approximately 2 hours and thereafter to the desired activation temperature.
Our current method for determining the activation temperature for these coatings involves initially heating the NEG coated vacuum vessel for 24 hours at 1800C. If no evidence of a pumping effect is observed then the activation temperature is increased in steps of 200C up to a maximum of 3000C. We have also developed procedures for preventing the possibility of contaminating or poisoning of the coating during the activation process itself, by minimising heat transfer to uncoated parts of the vacuum system.
In future we aim to investigate the effects of electron stimulated desorption on these coatings and to learn more about how the sticking co-efficient changes after multiple ventactivation cycles.
(1) Physics World - Vacuum Special Edition July 2007 page 14
(2) C. Benvenuti et al Vacuum 60 (2001) 57-6