406 Vacuum Systems

7.5.4 The Liquid Gauge Family

The liquid gauge family is identified simply as vacuum gauges that have some liquid (usually mercury or a low-vapor-pressure diffusion pump oil) directly in contact with the vacuum. The amount of liquid movement is directly proportional to the force exerted on it, and the (measured) amount of movement is read as the vacuum. Because mercury has traditionally been used for vacuum measurement, the term "millimeters of mercury" is commonly used even with nonliquid gauges.

Low-vapor-pressure oils can be substituted for some operations, but calibrations for density must be made so that their measurements can be interpreted as millimeters of mercury, torr, or Pa. Unfortunately, vacuum measurements can take a considerable amount of time when using oil because it takes a long time for a film of oil to settle from the walls of a manometer.

Despite the limitations and problems that the liquid gauge family presents, perhaps the biggest problem with the liquid gauge family is that its members can be very difficult to keep clean because the liquid is in direct contact with materials in the vacuum system. As mercury becomes dirty or contaminated, it tends to stick to the walls of glass tubing, thus decreasing its accuracy. As oil becomes dirty, its density can change and may provide inaccurate measurements. Realistically though, any vacuum system whose upper range is 10"3 torr is not likely to be significantly affected by limited contamination.

On the other hand, a liquid trap is essential between any liquid vacuum gauge and the manifold itself. The simplest accident can cause hours of needless delay as the vacuum line is cleaned out. Normally a simple splash trap is sufficient (see Sec. 7.4.6). Vacuum systems that go below 1(T5 torr should consider to have a cold trap placed between their liquid vacuum gauges and vacuum systems. This placement will keep vapors from contaminating either side of the system.

When a vacuum system is first started and brought from atmospheric pressure to a vacuum state, parts within the vacuum system need to be outgassed. When liquids are outgassed, they boil. This process needs to be done slowly, otherwise the liquids will "bump" (boil) violently, possibly causing the mercury (or oil) to splash across the system. This splashing could waste some time minimally by trying to get the liquid back into the gauge, or more time by requiring the system to be cleaned up; maximally, an extended period of time could be wasted by breaking the system. Therefore, slowly open the stopcock to your liquid manometer for outgassing (Sec. 7.5.8 provides specific information on how to let the liquid outgas in a McLeod gauge). It is not necessary to open the stopcock all the way to full open. If the stopcocks to the liquid vacuum gauge are kept closed to atmospheric pressures, the outgassing process needn't be repeated. However, if the gauge is brought to atmospheric pressure for an extended time (about a day), gases can re-enter the gauge and the full outgassing process must be repeated.

7.5.5 The Manometer

Second only to the mechanical gauge as the easiest device to measure and read a vacuum (and decidedly easiest in construction) is the liquid manometer (see Fig. 7.37). A well-made mercury manometer, kept very clean, can measure vacuums of up to 10"3 torr. This sensitivity can be increased by up to 15 times if a liquid with less density, such as diffusion pump oil, is used. However, diffusion pump oil is far more difficult to keep clean and can require either (a) a very tall (and thereby impractical) column or (b) a manometer of very limited range. In addition, because of the strong surface tension between diffusion pump oil and glass, long waiting periods between readings are required as the oil settles into place.

The mechanics of a U-tube manometer are simple: "The difference between the levels of two interconnected columns of liquid is directly proportional to the difference between the pressures exerted upon them, assuming equal capillary effects on both tubes."46 In practice, one end of a manometer is attached to an unknown pressure, and the other end is attached to a reference pressure that is known. For all practical purposes, the known value needs only to be a much smaller order of magnitude than the unknown value.

By maintaining a manometer at a constant temperature, the liquid within will maintain a consistent density. From 0° to 30°C the density of mercury changes about 0.5%. Therefore, a consistent reading of accurate vacuums can only come from a consistent room temperature. Manometer reading techniques (see Fig. 7.38) are consistent and simple regardless of the design and type of manometer. All that is necessary is some basic subtraction, a good metric ruler, and a steady eye. The manometer is read by observing and measuring the difference of the mercury column heights within the manometer.

Fig. 7.37 Open manometers are used to measure pressures relative to atmospheric whereas closed manometers are used for pressures far smaller than atmospheric. From Vacuum Science and Engineering, Fig. 3-1, by C. Van Atta, © 1965 by McGraw-Hill, New York, reproduced with permission.

Fig. 7.38 In this example, take the measurement from the closed side (439 mm) of the manometer, and subtract from it the measurement from the side of the manometer connected to the vacuum system (426 mm) to obtain the vacuum reading (13 mm Hg). Remember, avoid parallax problems, and with mercury, read the top of the meniscus.

The choice of mercury for manometers is often a matter of convenience, or rather the acceptance of the least amount of inconvenience. Mercury has a rather high vapor pressure (10~3 torr), but this vapor pressure is also at the upper ranges of what can be read by a manometer. Low-vapor-pressure oils (as used in diffusion pumps) can be used, but these oils wet the walls of a manometer and can take a long time to settle before reading can be made. Mercury is fairly nonreactive and retains a limited amounts of condensable vapors.

One of the biggest reasons why mercury is used is because it is a very heavy liquid. Atmospheric pressure can only push mercury about 76 cm, while it pushes water some 30 feet. Manometers using lighter-density liquids can be so tall they become beyond inconvenient. Additionally, because mercury doesn't wet the glass walls of the manometer, there is never a time lag for reading as there is with oil.

Fortunately, the mechanics of reading manometers is generally irrelevant to the amount of mercury (or any other liquid) within the device. Reading manometers can be difficult because of parallax problems. This difficulty can be complicated by the fact that manometers lack lines encircling them like burettes. Carroll found that by placing graph paper behind a mercury manometer, it was possible to see problems of parallax by the line's reflection off the glass and mercury (see Fig. 7.39).

For very accurate manometer readings, the NIST (National Institute of Standards and Testing) has a compiled list of eight possible errors that can develop while reading manometers.48 These errors introduce extremely small variations and therefore are not included here. If you need to read a manometer to the sensi-