There are numerous forms of cryopumps. Liquid nitrogen traps, used in series with diffusion pumps, are a form of cryopump.¹ In 1960 Meissner reported on the use of liquid nitrogen cooled coils for pumping water vapor within a vacuum coating system.² Sorption roughing pumps, filled with some form of artificial zeolite (i.e., a sieve material) and chilled with liquid nitrogen, are another form of cryopump.^3,4 There are at least seven forms of liquid helium cryopumps. The use of traps or coils cooled with Freon-type liquid refrigerants have been with us for almost a half-century.⁵ This technology has been greatly refined in recent years by Missimer and his colleagues at Polycold, and used in a wide variety of coating and etch applications.⁶

This column will discuss closed-loop gaseous helium cryopumps (cryopumps hereafter). A comprehensive recommended practice on cryopumps was recently published by the American Vacuum Society.⁷ The practice defines how one quantitatively specifies the performance and characteristics of a cryopump, technical terms, and safety considerations of cryopumps. It is the work product of cryopump suppliers including APD, CTI, CVC, Genesis, Leybold and other colleagues active in the industry over the last 25 years. If you are in the market for a cryopump, you would be well advised to first study reference 7. Also, a peak at reference 8 wouldn't hurt.

Cryopump Configuration

A cryopump using a closed-loop, gaseous helium refrigerator (e.g., see Fig. 1), consists of the refrigerator, cryopanels or arrays, and a pump body. The refrigerator comprises a two?stage expander (1) connected to a helium compressor (2) with flexible metal hoses (3). The system is called a closed-loop refrigerator because on expansion of the high pressure (~300 psig), room temperature helium gas in the expander, the low pressure gas returns to the compressor. Work is done on the helium in the compressor, and refrigeration is produced when the helium is expanded in the expander. When the smoke clears, the heart of the cryopump is the refrigerator. The shiny nickel-plated arrays, fancy control systems, valves, heaters, gauges, etc., etc., mean nothing, absent a reliable refrigerator system. In fact, the major breakthrough in cryopump technology was the development of a compressor which could produce clean, oil-free, high pressure helium.⁸ Any substance in the helium stream other than helium, and perhaps a little hydrogen, is a contaminant, and will degrade the performance of the refrigerator. This will shortly become more obvious.

Thermal Sources

There are three sources of heat (i.e., power) input to the cooling stations: i) thermal radiation; ii) gaseous convection; and, iii) heat of sorption losses. Because it is more difficult to produce refrigeration capacity at 10K-20K, we shield the second stage array from room temperature radiation from the work chamber and pump body (7) by completely surrounding it with the first stage array, maintained at 50K-80K. This reduces the radiation load on the second stage array by greater than a factor of 200. Gaseous convection losses are negligible at pressures =10^-5 torr. Keep this in mind when speculating that the increasing temperature of the second stage of your cryopump is due to increasing pressure (i.e., is it a cause or an effect?).

Sputtering Applications

Most sputtering applications occur at pressures of 1-20 millitorr. Such pressures are required to sustain a sputtering glow discharge, and obtain sufficiently high sputtering rates. Assume that a cryopump is used to pump the chamber during a sputtering process, argon is used as the sputtering gas, and the pump speed for argon is ~1500 L/s. If the pressure in the sputtering chamber is, say, 10^-2 torr, the throughput of argon into the pump is just S × P = 1500 L/s × 10-2 torr = 15 torr-L/s. As a rough rule of thumb, the heat of sorption when pumping gases is ~0.7 watts per torr-L/s. This implies that the thermal load on the second stage of the cryopump, due to heat of sorption, would be ~ 10.5 watts. As light bulbs go, this isn't very much power. But, at 2^nd stage temperatures of 10K-20K it suggest a fairly large refrigerator.

One technique to get around this problem is to remove the 1^st stage inlet array (4), and install a variable aperture plate in its place. The variable aperture plate, secured to the first stage array with indium foil, can be opened fully during system pumpdown, and then throttled back to reduce argon pump speed during the sputtering operation. Because the variable aperture is maintained 50K - 80K, a high water vapor pumping speed is maintained during the sputtering process. This is desirable. I was awarded a patent on a cryopump variable aperture plate,⁹ but confess, in retrospect, that it was more marketing than technology. Let's think for a minute about possible alternatives. Let's assume that we removed the inlet array and attached a fixed aperture plate (i.e., an aluminum plate with holes) in its place. Let's size the holes in the plate to restrict the pumping speed for argon to 400 L/s. In order for argon to be pumped, it must pass through the holes in the aperture plate and be cryocondenced on the 2^nd stage array. Yet, the aperture plate, maintained at 50K - 80K, has a pumping speed for water vapor equivalent to a "black hole" of the same diameter as the pump inlet. If the speed for argon, due to aperture plate, is now 400 L/s, the throughput during sputtering has been reduced to 4.0 torr-L/s, and the thermal load on the second stage, due to the heat of sorption of argon, is now a modest 2.8 watts.

Some tool manufacturers solve the argon throughput problem by using a room temperature throttle valve between the pump and chamber. When doing this, they throttle the water vapor pumping as well as argon speed during the sputtering process. Not a good tactic unless there is another means of pumping the water vapor (e.g., Meissner coils) within the chamber.

 With the fixed aperture plate we've solved the argon throughput problem while maintaining a high water vapor pumping speed. What about system pumpdown? Assume that each chamber of the tool has a volume of 10³ L - rather large for a typical cluster-tool - and that the chamber is vented to 10^-1 torr on insertion of the work through a load lock. If the gas is predominantly water vapor, it will be instantaneously pumped away by the cold, fixed aperture plate. If the gases comprise predominantly air, pressure in time, P(t), will be as follows:

P(t) = 10^-1× e^-(S/V)t torr                           (1)

where S is 470 L/s, V is 1000 L and t is the pumpdown time in seconds. Let's see, after 1.0 second of chamber pumpdown the pressure in the chamber will be 0.062 torr; in 10 seconds it will be ~9 × 10^-4 torr; in 30 seconds it will be ~7.2 × 10^-8 torr. You get the picture. The key is restrict the argon throughput, but maintain a high speed for water vapor. But, we have created a new control problem. When using a throttle valve between the cryopump and chamber, we are able to control the sputtering pressure in the chamber with some form of feed-back loop between a vacuum gauge and the valve. With a fixed aperture plate, we must have feed-back control circuitry between the same vacuum gauge and a gas flow meter.

A Little About the Refrigerators

 An illustration of the cycle of a single-stage expander is given in Fig. 2. The cooling cycle progresses from left to right, and down the page from phase #1 to phase #6. The diagram seems a little busy, but let's walk through it one step at a time. First, let's understand the components. The expander comprises a motor-driven or pneumatic-driven displacer piston which reciprocates within a thin-walled stainless steel cylinder. There is a room-temperature gas seal located at the lower extremity of the displacer. In a two-stage expander the seal of the 2^nd stage displacer is located proximate to the 1^st stage cooling station (i.e., operate at a temperature of 50K to 80K). The displacer comprises a hollowed-out phenolic-like cylinder within which resides a regenerator bed material. This might comprise lead shot, bronze screens, or both. Because of the seal, gas when entering or leaving the expansion chamber at the top of the regenerator bed piston can do so only by coursing through the somewhat transparent (i.e., to gas) regenerator bed. The regenerator bed material is selected for its a high specific heat at cryogenic temperatures. High pressure (HP) and low pressure (LP) helium valves introduce or exhaust helium from the apparatus to the compressor, depending on the phase of the cycle.

At #1 (i.e., phase #1 of the cycle), the HP He valve starts to open, and HP He courses through the regenerator bed, eventually filling the void above the displacer piston with high pressure gas. At #3, the helium pressurization cycle is complete, and at #4 the LP He valve starts to open. The expanding He produces refrigeration at the cooling station. Also - and this is the secrete of how they work - as the cooled He passes back out through the regenerator bed, it cools down the regenerator bed material (i.e., picks up calories from the material; #4 & #5). In the next #1 cycle, as the room temperature He passes through the regenerator bed to again fill the expansion volume, it is precooled by the regenerator bed. During the next expansion cycle, the chilled He becomes even colder, until after several minutes a steady-state temperature is reached at the cooling station. Two-stage expanders work exactly the same, except He reaching the 2^nd stage expansion volume is prechilled by both the 1^st and 2^nd stage regenerator beds. Therefore, the regenerator bed serves as a thermal ballast in the steady-state cycle. High pressure helium entering the expansion volume is prechilled by the regenerator bed (i.e., the bed picks up calories), whereas expanded helium on exiting through the regenerator bed, picks up calories from the bed.

If there are contaminants in the He stream (e.g., oil or water vapor in a single-stage expander; and any of the air gases in a two-stage expander) they may coat or condense out on the regenerator bed material. When this happens, the contaminant forms a thermal barrier between the bed material and the He stream, and the refrigeration performance is degraded. This can lead to very strange and misleading effects. Some of these effects are discussed in reference 8.

A Few Safety Issues

Cryopumped gases are retained within the pump as long as the pumping arrays are maintained at cryogenic temperatures. On warming of the cryopump, these gases are released. A pressure relief valve is located on the pump body (e.g., see (8) of Fig. 1). This valve serves as a safety device to prevent over-pressurization of the system on warm?up of the cryopump. If you are pumping toxic or combustible gases, the relief valve should be plumbed into some form of sealed off-gas disposal system. Talk with the pump manufacturer prior to tinkering in any way with the pressure relief valve.

 A main valve should be used to isolate the cryopump from the system. This is essential when regenerating the pump, and has other safety considerations. Never use an ionization gauge or other possible source of ignition on the pump side of the main isolation valve.
Understand the chemistry of your process. It is possible that you may be creating ozone or other mixtures of gases which may spontaneously ignite when warming the pump. Consult with the pump manufacturer in this regard. Review reference 7 and reference 8 (e.g., work the problems at the end of Chapter 5) for other safety considerations.

Cryopumps, when properly used serve as an economical, safe and reliable pumping means for most applications. However, their limitation, as with all capture pumps, is that they have a finite capacity for gases, and must be periodically regenerated. The concepts of gas capacity and pump regeneration are discussed in detail in references 7 and 8.

References

1)  M. H. Hablanian, High-Vacuum Technology, A Practical Guide, 2nd Ed. (Marcel Decker, Inc., New York, 1997), pp 250-262.
2)  C. R. Meissner, "A Versatile Ultra-high Vacuum System for This Film Research", 1960 Vacuum Symposium Transactions (Pergamon Press, Oxford, 1961), p.196.

3)  F. Turner, "Cryosorption Pumping", Varian Tech Pub., VR-76, 1972, Varian Associates, Inc., Palo Alto, CA.
4)  K. M. Welch, Capture Pumping Technology, 2^nd Edition (Elsevier Science, Amsterdam,2001), pp. 284-295.
5)  R.C. Leever, "A Low Temperature Mechanically Refrigerated Cold Trap", (Vacuum Symposium Transactions, American Vacuum Society, 1954), p. 19.
⁶)  Polycold Systems is a subsidiary of Intermagnetics General.
7)  K.M. Welch, et al, "Recommended Practices for Measuring the Performance and Characteristics of Closed-loop Gaseous Helium Cryopumps@, J. Vac. Sci. Technol. A 17(5), Sept/Oct 1999, pp. 3081-3095.
⁸)  K. M. Welch, Capture Pumping Technology, 2^nd Edition (Elsevier Science, Amsterdam, 2001), pp. 301-330.
⁹)  U.S. Patent 4,285,710.