The passage from the old to the new vacuum technique may perhaps be said to be marked by Kaufmann's invention of a rotary mercury pump in 1905. In this pump, which needed a rough preliminary pump (or "fore pump," as it is called) to bring the pressure down to 2 cm. of mercury or so, the gas was trapped and driven out by mercury moving in two spiral glass tubes, rotating about an inclined axis (fig. 3). This pump enjoyed a very short popularity, as it was speedily replaced by the Gaede rotary mercury pump, described under the heading MODERN METHODS AND TECHNIQUE.
Measurement of Low Pressures.—The first successful spe cial device for measuring pressure too low to be correctly estimated by the ordinary U-tube manometer was the gauge invented by McLeod in 1874, and known by his name. It consists in principle of a bulb B, some 25o cc. in capacity, provided with a fine cali brated vertical tube C (fig. 4). The mercury surface can be adjusted to any desired height by the usual device of a mercury reservoir connected to the apparatus by a flexible tube. The bulb B is connected to the space where the pressure to be measured prevails. The mercury surface is raised, cutting off a known volume of the gas, at very low pressure, in B, and then forcing it into the fine tube C. The pressure which the amount of gas present exerts in this confined space is then observed. To avoid error due to capillarity in measuring the pressure a side-tube D is provided of exactly the same bore as C, and the difference of level in C and D measured. Assuming Boyle's law, and knowing the volume of the bulb B and of the graduated tube C, it is easy to calculate the original pressure in B, which is the pressure required.
This gauge is still a standard instrument and receives further reference under the heading of MODERN METHODS AND TECH NIQUE, where the more recent methods for measuring minute pressures are discussed.
It may be well to preface the description of modern work at low pressures with a note on the units used to express these pressures. A pressure of i dyne per square centimetre is generally known as a microbar, a pressure of I million dynes per sq. cm. being a bar. This nomenclature, which has been widely adopted and is used by the meteorologists, who express their pressures in millibars, is used in the present article, but unfortunately some writers on high vacua (notably Dushman) use the term bar to denote a pressure of I dyne per sq.cm. and call a pressure of million dynes per sq.cm. a mega bar. Some French writers call a pressure of 1 dyne per sq.cm. a barye. The reader must, there fore, be on his guard.
The pressure conventionally taken as that of one atmosphere is 76o mm. of mercury at o° C,
latitude and sea level. Our bar, dynes per sq.cm., is al most exactly 75o mm. of mercury at standard conditions. For con siderations of high vacua, it is often sufficiently accurate to take I bar as equivalent to I atmos phere, which is a great advantage of the unit.
General Considerations at Very Low Pressures.—As a pre liminary to the discussion of many features of modern high vacuum technique, it is necessary to appreciate that the physical t behaviour of gases at very low pressures is in certain respects quite different from that of gases at pressures above, say, a thousandth of an atmosphere. The molecules of a gas at a given pressure travel, on the average, a certain distance between im pacts with one another, this distance being known as the mean free path (see KINETIC THEORY OF MATTER) and the physical criterion which decides whether a gas behaves in what may be termed the normal way or in the low-pressure way is given by the length of the mean free path as compared with the linear dimensions of the vessels in which the gas is contained. Taking oxygen as an example, the mean free path of the molecule at atmospheric pressure is about .00001 cm., which is exceedingly small compared to even the finest capillary tubes used in gas manipulations. The mean free path being inversely proportional to the pressure, it is .01 cm. at a pressure of .76 mm. of mercury; I cm. at .0076 mm. of mercury; and at a pressure of .000076 mm. of mercury, which is easily attained with modern technique, the mean free path is a metre, which is large compared to the vessels generally used, especially compared to the diameter of connecting tubes. The thermal conductivity and the viscosity of gases, for example, which are independent of the pressure so long as the mean free path is much smaller than the dimensions of the vessel, vary markedly with the pressure at very low pressures, a fact which is utilized in the construction of certain low-pressure gauges described later. In general, the laws which gases obey when moving relative to the enclosing surfaces change completely at the very low pressures under consideration, the physical reason being that whereas at higher pressures the impacts of a gas mole cule with the solid walls of the vessel are very infrequent com pared to its impacts with other gas molecules, with high vacua a molecule makes very many more impacts with the walls than with other gas molecules. These considerations are of prime importance in connection with the so-called molecular pumps, and also lead to results on the passage of low-pressure gases through tubes which are essential in high vacuum technique.