Rotary Mechanical Pump
Figure 2-12. Schematic of vacuum deposition system.
high (10" 10 torr) vacuum deposition and surface analytical equipment, but are being supplanted by turbomolecular and cryopumps.
2.3.2. Systems
The broad variety of applications requiring a low-pressure environment is reflected in a corresponding diversity of vacuum system design. One such system shown in Fig. 2-12 is employed for vacuum evaporation of metals. The basic pumping system consists of a nominal 15-cm diameter, multistage oil-diffusion pump backed by a 17-cfm (8.0 L/sec) rotary mechanical pump. In order to sequentially coat batch lots, the upper chamber must be vented to air in order to load substrates. To minimize the pumping cycle time, however, we desire to operate the diffusion pump continuously, thus avoiding the wait involved in heating or cooling pump oil. This means that the pump must always view a vacuum of better than ~ 10"' torr above and be backed by a similar pressure at the exhaust. A dual vacuum-pumping circuit consisting of three valves, in addition to vent valves, is required to accomplish these ends.
When starting cold, the high-vacuum and roughing valves are closed and the backing valve is open. Soon after the oil heats up, a high vacuum is reached above the diffusion pump. The backing valve is then closed, thus isolating the diffusion pump, and the roughing valve is opened, enabling the rotary pump to evacuate the chamber to a tolerable vacuum of about 10"1 torr. Finally, the roughing valve is shut, and both the backing and high-vacuum valves are opened, allowing the diffusion pump to bear the main pumping burden. By reversal of the valving, the system can be alternately vented or pumped rapidly and efficiently. This same operational procedure is followed in other diffusion-pumped systems, such as electron microscopes, where ease of specimen exchange is a requirement. To eliminate human error, pump-down cycles are now automated or computerized through the use of pressure sensors and electrically actuated valves. In other oil-less vacuum systems a similar valving arrangement exists between the involved fore- and main pumps.
Components worthy of note in the aforementioned evaporator are the high-vacuum valve and the optically dense baffle, both of which are designed to have a high conductance. Cryogenic cooling of the baffle helps prevent oil from backstreaming or creeping into the vacuum chamber. To ensure proper pressure levels for the functioning of the diffusion pump, thermocouple gauges are located in both the roughing and backing forelines. They operate from 10"3 torr to 1 atm. Ionization gauges, on the other hand, are sensitive to vacuum levels spanning the range 10"3 to lower than 10"10 torr and are, therefore, located to measure chamber pressure. Virtually all quoted vacuum pressures in thin-film deposition, processing, and characterization activities are derived from ionization gauges.
An actual vacuum deposition system is shown in Fig. 2-13.
- Figure 2-13. Vacuum deposition system. (Courtesy of Cooke Vacuum Products)
2.3.3. System Pumping Considerations
During the pump-down of a system, gas is removed from the chamber by (1) volume pumping and (2) pumping of species outgassed from internal surfaces. For volume pumping it is a relatively simple matter to calculate the time required to reach a given pressure. As an example, let us estimate the time required to evacuate a cylindrical bell jar, 46 cm in diameter and 76 cm high, from atmospheric pressure to a forepressure of 10"' torr. If an 8-L/sec mechanical pump is used, then substitution of Sp = 8 L/sec, V = (x/4)(46)2(76)/1000 = 126.3 L, P(t) = 10~' torr, P¡ = 760 torr, and P0 = 10~4 torr in Eq. 2-20 yields a pump-down time of 2.35 min. This value is comparable to typical forepumping times in clean, tight systems.
It is considerably more difficult to calculate pumping times in the high-vacuum regime where the system pressure depends on outgassing rates. There are two sources of this gas: (1) permeation and diffusion through the system walls and (2) desorption from the chamber surfaces and vacuum hardware. Specific vacuum materials, surface condition (smooth, porous, degree of cleanliness, etc.), and bakeout procedures critically affect the gas evolution rate. If the latter is known, however, it is possible to determine the necessary pumping speed at the required operating pressure through the use of Eq. 2-16. For example, suppose the vacuum system described has a total surface area of 1.5 m2, including all accessories. If the gas evolution rate (throughput) q0 is assumed to be 1.5 X 10 4 (torr-L/sec)/m2, then maintenance of a pressure P = 7.5 x 10"7 torr requires an effective pumping speed of S = 1.5q0/P = 300 L/sec. This value of S is necessary only to pump the quantity of gas arising through gas evolution from the walls, and is clearly a lower bound for the effective pumping speed of the system.
Lastly, it is appropriate to comment on vacuum system leaks. There is scarcely a thin-film technologist who has not struggled with them. No vacuum apparatus is absolutely vacuum-tight and, in principle, does not have to be. What is important, however, is that the leak rate be small and not influence the required working pressure, gas content, or ultimate system pressure. Leak rates are given in throughput units, e.g. torr-L/sec, and measured by noting the pressure rise in a system after isolating the pumps. The leak tightness of high-vacuum systems can be generally characterized by the following leakage rates (Ref. 7):
Adequately leak tight— ~ 10"5 torr-L/sec
One way to distinguish between gas leakage and outgassing from the vessel walls and hardware is to note the pressure rise with time. Gas leakage causes a linear rise in pressure, whereas outgassing results in a pressure rise that becomes gradually smaller and tends to a limiting value. The effect of leakage throughput on pumping time can be accounted for by inclusion in Eq. 2-20.
1. Consider a mole of gas in a chamber that is not being pumped. What is the probability of a self-pumping action such that all of the gas molecules will congregate in one-half of the chamber and leave a perfect vacuum in the other half?
2. A 1-m3 cubical-shaped vacuum chamber contains 02 molecules at a pressure of 10~4 atm at 300 K.
a. How many molecules are there in the chamber?
b. What is the ratio of maximum potential energy to average kinetic energy of these molecules?
c. What fraction of gas molecules has a kinetic energy in the x direction exceeding RT1 What fraction exceeds 2RT1
3. In many vacuum systems there is a gate valve consisting of a gasketed metal plate that acts to isolate the chamber above from the pumps below.
a. A sample is introduced into the chamber at 760 torr while the isolated pumps are maintained at 10~6 torr. For a 15-cm-diameter opening, what force acts on the valve plate to seal it?
b. The chamber is forepumped to a pressure of 10"2 torr. What force now acts on the valve plate?
4. Supersonic molecular beams have a velocity distribution given by where u0, the stream velocity, is related to the Mach number.
b. What is the average gas speed in terms of u0, M, and 7"? Note:
- Assume i>0 = 0.
5. The trap in Fig. 2-5 is filled with liquid N2 so that the entire trap surface is maintained at 80 K. What effect does this have on conductance?
6. Two identical lengths of piping are to be joined by a curved 90° elbow section or a sharp right-angle elbow section. Which overall assembly is expected to have a higher conductance? Why?
7. Show that the conductance of a pipe joining two large volumes through apertures of area A and A0 is given by C = 11.7 AA0 /(A - A0). [Hint: Calculate the conductance of the assembly in both directions.]
8. A chamber is evacuated by two sorption pumps of identical pumping speed. In one configuration the pumps are attached in parallel so that both pump simultaneously. In the second configuration they pump in serial or sequential order (one on and one off). Comment on the system pumping characteristics (pressure vs. time) for both configurations.
9. It is common to anneal thin films under vacuum in a closed-end quartz tube surrounded by a furnace. Consider pumping on such a cylindrical tube of length L, diameter D, and conductance C that outgasses uniformly at a rate q0 (torr-L/cm2-sec). Derive an expression for the steady-state pressure distribution along the tube axis. [Hint: Equate the gas load within length dx to throughput through the same length.]
10. After evacuation of a chamber whose volume is 30 L to a pressure of 1 X 10"6 torr, the pumps are isolated. The pressure rises to 1 X 10"5 torr in 3 min.
a. What is the leakage rate?
b. If a diffusion pump with an effective speed of 40 L/sec is attached to the chamber, what ultimate pressure can be expected?
11. Select any instrument or piece of equipment requiring high vacuum during operation (e.g., electron microscope, evaporator, Auger spectrometer, etc.). Sketch the layout of the vacuum-pumping components within the system. Explain how the gauges that measure the system pressure work.
12. A system of volume equal to 1 m3 is evacuated to an ultimate pressure of 10"7 torr employing a 200 L/sec pump. For a reactive evaporation process, 100 cm3 of gas (STP) must be continuously delivered through the system per minute.
a. What is the ultimate system pressure under these conditions?
b. What conditions are necessary to maintain this process at 10"2 torr?
13. In a tubular low-pressure chemical vapor deposition (LPCVD) reactor, gas is introduced at one end at a rate of 75 torr-L/min. At the other end is a vacuum pump of speed Sp. If the reactor must operate at 1 torr, what value of Sp is required?
References
1.* R. Glang, in Handbook of Thin Film Technology, eds. L. I. Maissel and R. Glang, McGraw-Hill, New York (1970).
2.* A. Roth, Vacuum Technology, North-Holland, Amsterdam (1976).
3.* S. Dushman, Scientific Foundations of Vacuum Techniques, Wiley, New York (1962).
4.* R. Glang, R. A. Holmwood, and J. A. Kurtz, in Handbook of Thin Film Technology, eds. L. I. Maissel and R. Glang, McGraw-Hill, New York (1970).
5. J. M. Lafferty, Techniques of High Vacuum, GE Report No. 64-RL-3791G (1964).
6. R. W. Roberts, An Outline of Vacuum Technology, GE Report No. 64-RL 3394C (1964).
7.* "Vacuum Technology: Its Foundations, Formulae and Tables," in Product and Vacuum Technology Reference Book, Leybold-Heraeus, San Jose, CA (1986).
*Recommended texts or reviews.
Crrr ee}- Chapter 3
Post a comment