a Toepler can be seen in Fig. 7.27. In operation, the gas to be pumped enters the inlet at the top of the piston chamber. The pumping action begins when air is allowed into the lower chamber by the stopcock at C. The air forces the mercury up into the piston chamber and into the inlet and exhaust tubes. Finally, the mercury pushes up the float valves, preventing the mercury from going out the inlet. Once the mercury makes contact with the tungsten wire at D, the circuit with the tungsten wire at F is complete. Then, the stopcock at C is rotated to evacuate the lower portion of the pump. The mercury finally is drawn back into the lower portion of the pump. The float valve at exhaust tube (at B) prevents gas being pumped from re-entering the Toepler pump. As the mercury fills the lower chamber, it fills the small chamber (with tungsten wire at E) which completes the circuit with the tungsten wire at F. This completion activates the stopcock at C to re-admit air into the pump, and the process repeats until it is manually turned off.

The major problem with this type of setup is when the mercury makes (and unmakes) contact with the tungsten wire at D. A spark may form that reacts with the gas being pumped through the exhaust. Mechanisms to by-pass this problem have included either placing the air/vacuum cycling on a timer, thereby bypassing the need of electrical contacts, or placing a photosensitive relay on the side of the exhaust tube, which is activated when the opaque mercury fills the transparent glass tube.

As with most items on a vacuum line that contain mercury, place the Toepler pump in a secondary container that is firmly attached to another surface to contain any mercury that may spill from an accident as well as protect the pump from accidental bumps (see Fig. 7.28). Plastic containers (such as plastic milk cartons) are particularly good because the mercury will not affect plastic. Conversely, mercury may amalgamate with the metals in a metal can, which could destroy any containment capabilities. In addition, it is easier to get mercury out of a plastic container (with smooth walls) than out of a metal one (with a narrow rim). The plastic tub can be glued onto the table with some epoxy. The epoxy will stick better if you roughen up the bottom surface of the plastic container with sandpaper. Because not all epoxies stick to plastic, test the epoxy before assuming that it will hold.

7.4Traps

7.4.1The Purpose and Functions of Traps

Traps (and baffles and filters) are used on vacuum systems because of the need to remove, catch, or bind condensable vapors, paniculate matter, and aerosols. To obtain a vacuum, you must remove everything from a given area. However, like the environmental adage "you cannot throw anything away," you may not want everything that is removed from a vacuum system passed into the pumping sys-

Cold trap

Fig. 7.29 A cold trap is worth the trouble on even a simple vacuum system.

tem, or into the air you breath. Solvents (and their vapors) can impair or severely damage mechanical pumps by dissolving seals or by breaking down mechanical pump oils. Paniculate matter passing into a mechanical pump can severely damage veins and seals. Condensed vapors in pump oil can impair a pump's performance. Volatile compounds that pass through a pump and leave as aerosols can expose technicians to an unacceptable health risk. Thus, the use of traps in vacuum systems can protect equipment and the people who are using the equipment.*

Traps (specifically foreline traps) can also improve the vacuum in a vacuum system either by becoming a separate vacuum station (such as the cryogenic capabilities of a cold trap) or by limiting the backflow of oils from a pump (either a cold trap or a molecular sieve type trap).

Traps can also do more than protect vacuum systems. By placing a series of traps in a line, and by using successively cooler coolants (see Sec. 6.2 for different coolant materials)one can separate compounds by their condensation or freezing points (see Sec. 7.4.5).

The simplest way to protect a pump, pump oil, and pump seals, reduce maintenance time, and protect the worker from the effects of condensable gases in a vacuum system is to place a trap between the vacuum line and pump (see Fig. 7.29). For example, although water vapor is not likely to damage a mechanical pump or worker, it can still impair pump performance by hydrating the pump oil. This problem can be resolved (after the fact) by keeping the pump running for an extended period of time (overnight) against a small dry air (or dry nitrogen) leak. However, this problem can easily (and should) be prevented by trap use at the outset.

As mentioned, a trap prevents condensable vapors from damaging pump components as they are removed from the vacuum system. Additionally, a trap can improve the quality of the vacuum by preventing condensable vapors and pump oil vapors from drifting back into the system. Remember that once a pump has achieved its potential vacuum, gas flow drops to zero leaving materials to drift

Traps are supposed to prevent exhausts from being released into the working environment. However, they should not be relied on to the extent of the possible negligence of the user(s) or other possible accidents causing failure. Therefore,all exhaustsfrom mechanical pumps should be vented to fume hoods (see Fig. 7.14 on exhausting mechanical pumps to fume hoods).

the various traps and their effectiveness against various vacuum system contamination.

5.Mist traps limit the amount of the aerosols of mechanical pump oils from leaving the pump and drifting into the room containing the pump. These traps are different from the other traps in that they go on the exhaust of the mechanical pump and do not protect the pump or the system, only the operators.

Cold traps use either water, dry ice slush baths, or liquid nitrogen as a coolant. The low temperatures of a cold trap cause condensable vapors to change their state of matter into a liquid or a solid depending on the temperature and the vapor. As temperature decreases, there is an increase in price and an increase in performance.

Water-based cold traps have limited efficiency, although they are the most costeffective. They can liquefy water vapor and many hydrocarbon solvents, but they can't entrap these materials. Thus, they are capable (and likely) to revaporize back into the system. The cost of running tap water through the cooler and down the sink can be eliminated by running a recirculating pump in a large plastic bucket. The effectiveness can be enhanced by periodically dumping crushed ice into the bucket, but this does take constant maintenance of placing more ice as necessary and removing excess water as necessary (a drain placed on the top can decrease that concern).

Slush baths (see Sec. 6.2.7) are effective for a greater range of vapors and can entrap many vapors by freezing. These types of cold traps require a lot of maintenance.

The most efficient type of cold trap is one that uses liquid nitrogen. This is because it is so cold that once captured, water, hydrocarbon solvents, and most other vapors will not be released into the system. Liquid nitrogen cold traps also enhance pumping within the system due to their cryogenic capabilities. Unfortunately, they are not without problems or complications. Probably the biggest handicap is that they require constant attention to maintain filling levels. Every time the coolant level goes below the level of frozen vapors, the pressure in the system will increase (see Fig. 7.32). Additionally, liquid nitrogen and its equipment are expensive.

Molecular sieve traps (and Micromaze traps from the Kurt J. Lesker Co.) have low maintenance and low long-term costs. Their main advantage is that, once bound, vapors are trapped till they are released by heating their active materials. Thus they do not have the periodic pressure fluctuations that cold traps have. Their main disadvantage is that they require regeneration of the sieve by baking at 150300°C about every 300 hours (or less) of use, and the regeneration requires many hours. The vacuum section must be closed off during the bakeout; otherwise, already-trapped hydrocarbons are dumped right back into the system. Because it is recommended that any valve between the trap and the system be baked as well as the trap,* this whole section cannot be made out of glass. There is no problem