As an introduction to vacuum technology, consider the air at ordinary atmospheric pressure inside a gallon jug on an average summer day. The jug encloses about a quarter of an ounce of gas, the molecules of which, driven by thermal energy, bounce endlessly off one another and the walls of the container like a load of ping-pong balls in a truck jouncing over a corduroy road. On the average the molecules move at about twice the velocity of a bullet from a high-powered rifle, but travel only a few millionths of an inch before experiencing a collision.

To clear the myriad molecules from a vessel in preparation for an experiment, one has the choice of three basic techniques, which may be used singly or in combination. First, the gas can be displaced by another substance, which is then withdrawn to leave a vacuum. Second, gas can be mechanically pushed out of a space connected to the vessel, so that molecules in the vessel can escape by their own random motion; the gas pressure is steadily lowered as the pushing is continued. Third, a solid material, usually a thin film of metal, that removes the gas by combining with it can be introduced into the vessel.

The first of these techniques was devised by Evangelista Torricelli to demonstrate that there is such a thing as a vacuum. Torricelli's experiment is described in a letter he addressed on June 11, 1644, to Michelangelo Ricci in Rome: "I have made an instrument of glass tubing," he wrote, "sealed at one end and about two cubits [35 inches] long. It was filled with quicksilver, its mouth closed with a finger and then it was inverted in a vessel containing quicksilver. When the finger was removed the tube was seen to empty partly, without anything entering while it emptied. The lower part of the tube remained always full, however, to the height of a cubit, a quarter and a finger extra. To show that the tube was empty above, we then filled the remaining part of the bowl with colored water up to the brim and slowly raised the tube until its mouth was seen to reach the water. The quicksilver suddenly dropped down from the tube and, with a horrible impact, the water shot up to the very top! The result makes it plain that the space above the quicksilver was empty." Torricelli then surmised that air pressing down on the surface of the mercury in the bowl exerted the force that supported the mercury in the tube. In modified form Torricelli's apparatus evacuated the first electric-light bulbs, and it still renders service as the mercury barometer. Incidentally, those who repeat the experiment are cautioned to keep a finger partly over the mouth of the tube as the water displaces the mercury; otherwise the "horrible impact" may shatter the glass. It is also well to keep in mind that mercury evaporates slowly even at room temperature, and that its fumes are toxic.

It is possible to produce useful vacuums by displacing air with mercury. But pumps based on Torricelli's principle are tedious to operate. Most modern experimenters therefore employ various combinations of the second and third basic techniques. Unwanted gas is cleared from the vessel by compressors designed to operate at low pressure. Residual gas is then trapped, if desired, by a "getter" (an evaporated film of a metal such as barium) or the gas can be adsorbed on charcoal.

Compressors take a variety of forms: piston pumps, rotary pumps in which a solid cylinder fitted with sliding vanes rotates inside a sealed hollow cylinder mounted eccentrically with respect to the rotor, high-speed drums or disks that turn inside a close-fitting housing and impart a preferred direction to gas molecules impinging on them from random directions, and jet pumps of various types. In no sense do these pumps "suck" gas from the vessel being evacuated. Pumps merely gather those molecules that chance directs their way and move them to an exhaust port, from which they escape into the atmosphere. No matter how long the compressors operate, the probability is low that all the molecules will escape from the vessel.

Although a perfect vacuum cannot be achieved, one can assemble a vacuum system capable of reducing atmospheric pressure inside a gallon vessel by a factor of 100 million or more at a cost of less than $30. The uses of such a system in the home laboratory are almost endless. A partial vacuum at the end of a pipe, for example, will produce the same flow of liquid as an equivalent pressure at the other end and will greatly reduce the time required for a fluid to pass through a filter. Or, as another example, the experimenter may wish to avoid the grosser physical or chemical effects of air: its movement, heat conduction, radiation absorption or oxidation. Atmospheric pressure influences the temperature at which many chemical reactions occur. The reduction of iron oxide by carbon, for instance, is normally accomplished in a furnace. But under an adequate vacuum the reaction proceeds at low temperature to yield free iron and carbon monoxide. Another area of interest has to do with vapor pressure. The boiling point of substances is directly related to their vapor pressure and thus to the pressure of the surrounding atmosphere. In consequence the amateur equipped with a vacuum system and a few ounces of dry ice can reduce the temperature of a test chamber almost to the boiling point of liquid air. Finally, to the experimenter interested in electronic and nuclear phenomena the vacuum system is as indispensable as the lathe is to the machinist.

A simple but effective system can be had by modifying the compressor unit from a used refrigerator. In most communities operable units can be purchased from appliance dealers at prices ranging from $5 to $10. F. B. Lee, a member of the faculty at the Erie County Technical Institute in Buffalo, N.Y., has investigated a number of makes and reports that three lend themselves to vacuum work: Frigidaire, Norge and Coldspot.

"Of these," he writes, "the Norge Rollator belt-driven unit rates best for vacuum produced. Frigidaire rates best for availability and second best for two-stage service below .020 millimeter of mercury. (Atmospheric pressure at sea level supports a column of mercury 760 millimeters high in a tube closed at the top.) The Coldspot, though unsuitable for pressures below .5 millimeter, is superior to the Frigidaire as a single unit.

"The modifications are not difficult. Those required to convert the Frigidaire 'Meter-Miser,' the unit that has been standard on this company's domestic refrigerators since 1936, are illustrative. The smaller refrigerators contain split-phase motors rated at less than 1/7 horsepower. The Imperial or Cold Wall series and all refrigerators larger than 135 cubic feet contain capacitor motors. The purchaser is advised to procure the capacitor as well as the compressor. The motor is not self-starting; it is therefore advisable to procure the starting relay as well. If this is not available, one may improvise a starter from a push-button switch. The motor is started by applying power to terminals 1 and 3 in the accompanying drawing [Figure 2] and short-circuiting terminals 2 and 3 (by means of the relay or push button) for a period of about four seconds.

"The pump is modified in three respects. The bypass line which runs between the housing and the check valve must be cut off and the ends sealed. The pump will then produce a vacuum of one millimeter if the check valve is open and the strainer is not wet with oil. The check valve opens automatically when the pressure in the system is above three millimeters. The pressure will not drop below 10 millimeters if the strainer is wet with oil. Oil can be removed from the strainer and the check valve opened by permitting air to flow through the pump for a few minutes prior to connecting the unit to the vacuum system. The screen may be removed, but great care must be exercised thereafter to prevent dirt or foreign material from entering the pump. To make this modification, cut the intake line about an inch away from the housing. Use a tube cutter, not a hacksaw, or particles from the saw will almost certainly find their way into the pump and cause it to stall. Bend the cut tube out of the way, then dig the strainer from the opening by means of a small hook made from a nail or steel wire. Inspect the opening carefully and remove all stray wires of the screen by means of tweezers. Cap the opening with a short length of rubber tubing and close the end with a pinch clamp.

"The pump operates best when tilted at an angle of 10 or 15 degrees, as shown in the drawing of the system [Figure 3]. The line to the oil trap should be pitched upward away from the pump to prevent the formation of oil pockets that would impede the free flow of air. The trap can be made from a quart milk-bottle.

"In addition to the compressor and oil trap, the system includes a dirt trap made from a half-gallon glass jug, and a pair of vacuum reservoirs, each a gallon glass jug. As a safety measure all the jugs are housed in wooden boxes to catch fragments in the event that atmospheric pressure shatters the glass. The various units of the system are interconnected by 3/8-inch copper tubing, perforated rubber stoppers and couplings of rubber hose. Five of the hose couplings are equipped with pinch clamps and act as valves as shown.

"To operate the system, first connect the vessel to be evacuated and close the clamp between the exhaust port [knife cut in rubber tube in illustration in Figure 2] and the rest of the system. Then open all the other clamps, and start the pump. This will reduce the pressure of the entire system, including that in the vacuum reservoirs, to about one millimeter. Now the clamp between the two reservoirs is closed, and operation is continued for about five minutes with the clamp between the pump exhaust and the reservoirs open. This has the effect of connecting the input of a second compressor to the exhaust port of the first, one vacuum reservoir serving as the added compressor. The clamp between this reservoir and the exhaust port of the pump, and the clamp between the reservoirs and the line leading to the oil trap, are now closed. The clamp between the reservoirs and the exhaust port of the pump is opened. This permits the system to exhaust into the second reservoir, now the one of lower pressure. With continued operation the pressure will then fall to the limit of the system's capacity. The compressor can operate for a whole day without increasing the pressure in the reservoir more than one or two millimeters."

An interesting project requiring a vacuum is the construction of the radiometer depicted in Figure 3. The bracket for supporting the vane assembly can be made of a short metal strip as shown. The point of the needle must rest either on hard, smooth metal or on glass. It is not difficult to make the entire bracket out of two glass stirring-rods of the kind used for mixing drinks. One rod is heated by a small gas torch successively at two points, until it is soft and bent in the form of a square-cornered "C," the opening of the "C" being made just slightly wider than the length of the sewing needle previously selected to serve as the shaft. Facing indentations are next made just inside the tips of the "C," to serve as pivot bearings. Heat the tips one at a time and press the point of a nail part way into the soft glass so that the impressions face each other. (Support the hot glass on Masonite or a sheet of asbestos.) The second stirring rod is then heated in the middle and bent to a right angle. The angular member is placed on the Masonite, and the "C" is supported vertically inside the apex of the angle so that one arm of the "C" rests on the Masonite and bisects the angle formed by the second rod. Both rods are heated until they become soft and join at the point of contact. The angular member now serves as a base support for the assembly. Next the point of the needle is set in the lower bearing and supported while the upper bend of the "C" is heated until the upper arm closes down over the eye end of the needle. Two flat sheets of aluminum foil, measuring about one by two inches, are centered on and cemented to the needle. When the cement has dried, the sheets are bent to form the four vanes illustrated. One side of each vane (the same side in each case) is now heavily smoked. (A candle produces a rich flow of smoke if a few drops of machine oil are added to the pool of melted wax surrounding the wick.)

The vane assembly should be reasonably well balanced and turn freely. It is placed in a wide-mouthed jar fitted with a rubber gasket and a screw cap into which a vacuum-tight nipple has been soldered. The radiometer is connected to the vacuum system and evacuated. When exposed to an infrared lamp of 200 watts (or an equivalent source of heat) at a distance of six inches, the vanes will spin vigorously after the pressure has been reduced to .05 millimeter.

An apparatus that opens many experimental opportunities is the dry-ice refrigerator. The ice container is made of a cardboard mailing-tube closed at the bottom by a disk of cardboard and supported on all sides by lightly crushed aluminum foil, as shown in Figure 5. A glass test-tube supported by a pierced rubber stopper serves as the cold chamber. The lid of the jar is fitted with an exhaust nipple and is pumped through a trap, also made of a wide-mouthed jar, charged with household lye (sodium hydroxide). The carbon dioxide combines chemically with the lye, thus reducing the load on the vacuum system. At a pressure of one atmosphere, dry ice (solid carbon dioxide) sublimes at a temperature of-78 degrees centigrade. At one millimeter the temperature drops to-135 degrees and at .001 millimeter to-166 degrees (-266 degrees Fahrenheit). Liquid oxygen boils at-183 degrees C.

To achieve pressures much below one millimeter within a reasonable interval (less than 30 minutes) one must insert a jet pump between the vessel being evacuated and the mechanical pumping system. A system of this type, constructed by Walter Semerau of Kenmore, N.Y., for evaporating films of reflecting aluminum onto telescope mirrors, is depicted in Figure 6. The mechanical, or "fore," pump appears at lower left; the jet, or "diffusion," pump, at lower right. The vacuum vessel, consisting of a bell jar and a base-plate assembly, appears at the top together with its accessories.

The diffusion pump consists essentially of a boiler fitted with a jet for discharging vapor (usually of mercury or oil) at high velocity into a tube of larger diameter that is maintained at a slightly lower relative temperature. Molecules of gas from the vessel being exhausted enter the cold tube at a point behind the jet and, by random motion, diffuse into the jet stream. The probability is high that collisions with the molecules in the jet will accelerate them in the direction of the exhaust port, where they will diffuse into the fore pump. The vapor condenses on the cold walls of the pump and is returned by gravity to the boiler for another cycle. Most diffusion pumps do not operate effectively above .1 millimeter, but below this pressure their speed is impressively higher than that of mechanical pumps. The diffusion pump used by Semerau employs two jets in series and pumps at the rate of 501iters per second; his fore pump is limited to 30 liters per minute-a ratio of 100 to 1 in favor of the diffusion pump.

An aluminum film can be evaporated onto a mirror at a pressure on the order of .001 millimeter, easily attained with a modified refrigerator compressor. But such a mirror is inferior, according to Semerau, to one coated in a vacuum system that includes a diffusion pump.

"The relatively poor quality of these mirrors," he writes, "is explained by at least two causes. First, at pressures above one millimeter, a relatively large amount of gas, including water vapor, is adsorbed on the surface of the glass to be coated. Unless much of this is removed during pump-down, the mirror is apt to peel shortly after it is put into use. Second, mechanical pumps are inefficient at low pressures, and in the case of small vacuum-systems a pumping time of several hours would be required to reduce the pressure substantially below one millimeter. During this interval trace amounts of organic material on the glass would oxidize, or undergo other chemical change, and discolor the film. I have never succeeded in eliminating trace contamination no matter how zealously I cleaned the glass.

"To deposit thin films, either for mirrors, nonreflecting coatings on lenses, interference filters or for evacuating electronic devices, one requires a vacuum system that includes at the minimum a diffusion pump, a bell jar large enough to accommodate the apparatus, a tungsten heating-element, a source of high-current, low-voltage power to energize the heater, and of course a stock of the material to be evaporated.

"The bell jar may be either of glass or metal. Whatever the material, the walls of the vessel must be capable of withstanding the total pressure exerted by the atmosphere when the system is evacuated or the vessel will implode. If glass is used, the jar must, as a safety measure, be enclosed by a substantial screen. In the case of my bell jar, which has walls 3/8 inch thick and measures 12 inches wide and 18 inches high, atmospheric pressure exerts a force on the glass of some five tons. Jars can be made by cutting the bottom from a glass jug and grinding the cut edges smooth to make a tight fit with the base plate. The opening of the jug is plugged by a rubber stopper. Do not make the bell jar by cutting off the rounded top of the jug; the flat bottom will flex under pressure and shatter the glass.

"The metal base-plate on which the bell jar rests can be made of any scrap material-iron, aluminum or brass-if the upper surface is smooth and the plate is thick enough to withstand atmospheric pressure without flexing perceptibly. My plate is a slab of polished aluminum .75 inch thick. The joint between the bell jar and the base plate may be sealed with vacuum wax (a mixture of beeswax and rosin in equal parts by weight applied while hot) or, as a more convenient alternative, with a rubber ring.

"The base plate is drilled in the center with a hole to match the exhaust port of the diffusion pump. The pump should be coupled as closely as possible to the plate by a gas-tight seal. Incidentally, the speed at which gas flows under low pressure varies directly with the diameter, and inversely with the length, of the conducting channel. Avoid long runs of tubing and bends. A right-angle elbow introduces as much resistance to the flow of gas at low pressure as three feet of straight pipe of matching diameter. The interval required for exhausting the bell jar can be reduced by equipping the base plate with a valve to cut off the diffusion pump. This permits the diffusion pump to be exhausted and placed in operation while the bell jar is being loaded. The valve must of course be designed for remote operation, because it is covered by the bell jar.

"The heating element is made of .03-inch tungsten wire and is formed by winding five turns spaced 1/8 inch apart on a mandrel 1/4 inch in diameter. Because tungsten is relatively brittle at room temperature, the wire is customarily heated to a dull red by means of a torch during the forming operation. (Overheating it is likely to result in sharp bends at which the heater may break during subsequent use.)

"Molecules of metal evaporated in a high vacuum move away from the heating element along radial paths. Hence when a coating is deposited on a relatively flat disk of glass centered above the heating element, the thickness of the metal film increases from the edge to the center of the disk. In an amateur telescope-mirror the effect is not serious if the mirror is spaced at least two diameters above the heating element. Otherwise multiple heaters must be used. The heaters are supported on studs set in gas-tight insulators that extend through the base plate. Power for the heater is taken from a step-down transformer that operates either from a rheostat or a variable transformer to control the output through the range of two to 10 volts at a maximum current of 50 amperes.

"A device for indicating pressure inside the bell jar, such as a glow tube, is useful. Aluminum films may be evaporated successfully when the tube darkens and the glass fluoresces.

"To aluminize a telescope mirror, the assembled apparatus is first pumped down until the glow tube fluoresces. The heater is then brought up to full temperature ( 10 volts ) . This reduces the film of oxide on the tungsten. The heater is cooled, air is admitted and the bell jar is removed. A hairpin loop of aluminum wire .03 inch in diameter and 3/4 inch long is suspended from each turn of the heater. Then the system is again pumped down. When the glow tube fluoresces, power is applied to the heater and gradually increased until the aluminum wire just melts and wets the tungsten. The heater is then cooled but left under vacuum.

"The glass to be coated is now thoroughly cleaned. Mere soap and water will not do; the surface must be chemically clean. I use trichloroethylene to dissolve grease, followed by acetone and a final rinse with alcohol. The glass is then dried with a well-washed cotton cloth. Permit nothing else to touch the dried surface or it will almost certainly be smudged. Next the vacuum is broken and the glass is transferred immediately to its supporting fixture. Exhaust the bell jar. When the glow tube indicates minimum pressure, apply full voltage to the heater. The aluminum will be fully vaporized within 45 seconds. Allow the system to cool at least 10 minutes before breaking the vacuum."

Amateurs who set up equipment for evaporating films will find the essential materials and apparatus (with the possible exception of the diffusion pump and tungsten wire) available in most communities. The diffusion pump used by F. B. Lee was made by a local glassblower at about twice the cost of the used refrigerator unit. Duplicates of this pump may be procured through Lee, who, in the interest of encouraging amateurs to take up high-vacuum experiments as a hobby, has volunteered to assist in procuring tungsten wire and hard-to-get supplies. His address is: F. B. Lee, 230 Hampton Parkway, Buffalo 3, N.Y.

Bibliography

HIGH VACUUM TECHNIQUE. J. Yarwood. John Wiley & Sons, Inc., 1955.

Suppliers and Organizations

The Society for Amateur Scientists
5600 Post Road, #114-341
East Greenwich, RI 02818
Phone: 1-401-823-7800

Internet: http://www.sas.org/