Full Text
From Current Literature
INTRODUCTION OF SPECTRALLY PURE GASES INTO VACUUM SYSTEMS¹
It is very often necessary to introduce various gases into vacuum systems in as pure a state as possible and to be able to control the quantity of gas entering and the rate of its entry. In such cases, various kinds of ground stopcocks, mercury and other valves, are usually used. These devices, however, are unsatisfactory for a number of reasons well known to every vacuum technician: the gas is invariably contaminated by vapors of the grease, and also, owing to leakage of the stopcocks, contains mercury vapor; regulation of the rate of entry is possible, at best, within very narrow limits, etc.
The method, widely known, of introducing hydrogen through a palladium tube is the closest to the ideal method. Until now, however, it has been used only for hydrogen. As it turns out, this method can also be used for many other gases.
As early as 1904, Richardson, Nicol, and Parnell² gave a theoretical formula for the rate of diffusion of a gas through a metal. This formula has the form
\[ D = \frac{k}{dP^{\frac{1}{2}} e^{-\frac{b}{T}}}, \]
where \(D\) is the rate of diffusion (i.e., the quantity of gas in \(1\ \mathrm{cm^3}\) under normal conditions, penetrating through \(1\ \mathrm{cm^2}\) of a metal wall, with its thickness in \(1\ \mathrm{mm}\), in \(1\ \mathrm{sec}\)), \(P\) is the gas pressure in millimeters of mercury, \(d\) is the wall thickness in millimeters, \(T\) is the temperature of the metal in degrees Kelvin, and \(b\) and \(k\) are constants for the given metal–gas system. This equation may also be written in the form:
\[ D = \frac{k}{dP^{\frac{1}{2}} e^{-\frac{E_0}{2RT}}}, \]
where \(R\) is the gas constant \((\mathrm{cal}/\mathrm{mol}\cdot\mathrm{deg})\), \(E_0\) is the activation energy \((\mathrm{cal}/\mathrm{g\text{-}mol})\) for the given metal–gas system, and the remaining notation is retained as before. The applicability of this equation was tested experimentally for a number of metal–gas systems, and good agreement was found within the usual ranges of temperatures and pressures. At pressures on the order of several millimeters of mercury and below, however, deviations are observed. The condition of the surface and its previous history strongly affect the rate of diffusion. Thus, for example, “poisoning” the surface with gases greatly reduces \(D\). The presence of impurities in the metal acts in the same way. Conversely, loosening of the surface by etching in acids or by alternate oxidation and reduction can increase \(D\) hundreds of times. The latter phenomenon is caused not only by an increase in the surface area on which the gas can be adsorbed, but also by an increase in the “active” portion of the surface (i.e., the number of active sites on the surface). Table 1 gives the constants \(k\) and \(b\) for several metal–gas systems.
This method also makes it possible to introduce inert gases into a vacuum system. For them, however, one should use not metals but fused quartz \((\mathrm{SiO_2})\). Fused quartz can also serve as
and for introducing hydrogen, oxygen, nitrogen, and argon, but the diffusion rates of these latter gases are very small. Moreover, it turns out that for helium, neon, and hydrogen the diffusion rate does not depend on the duration of heating of the quartz, whereas for the other gases listed the diffusion rate rapidly decreases with heating time. This occurs as a result of the formation of a crystalline layer on the surface of the fused quartz. If this layer is removed with hydrofluoric acid, then
Table 1
| System | \(k\) | \(b\) | \(R\) (at \(1200^\circ\mathrm{K}\)) |
\(D\) (of the given system) / \(D\) (\(\mathrm{H}_2\)—Pd) (at \(1200^\circ\mathrm{K}\)) |
|---|---|---|---|---|
| \(\mathrm{H}_2\)—Ni | \(2.3\cdot10^{-2}\) | 7 500 | \(1.4\cdot10^{-2}\) | \(7.1\cdot10^{-3}\) |
| \(\mathrm{H}_2\)—Pt | \(1.4\cdot10^{-2}\) | 9 800 | \(1.2\cdot10^{-4}\) | \(6.4\cdot10^{-4}\) |
| \(\mathrm{H}_2\)—Mo | \(9.0\cdot10^{-3}\) | 10 000 | \(6.3\cdot10^{-5}\) | \(3.3\cdot10^{-4}\) |
| \(\mathrm{H}_2\)—Cu | \(2.3\cdot10^{-3}\) | 8 300 | \(7.4\cdot10^{-5}\) | \(3.9\cdot10^{-4}\) |
| \(\mathrm{H}_2\)—Al | 3.3 | 15 500 | \(2.7\cdot10^{-4}\) | \(1.4\cdot10^{-3}\) |
| \(\mathrm{H}_2\)—Fe | \(1.6\cdot10^{-3}\) | 4 820 | \(8.5\cdot10^{-4}\) | \(4.5\cdot10^{-3}\) |
| \(\mathrm{H}_2\)—Pd | \(4.0\cdot10^{-2}\) | 2 220 | \(1.9\cdot10^{-1}\) | 1 |
| \(\mathrm{O}_2\)—Ag | \(3.7\cdot10^{-2}\) | 22 600 | \(1.6\cdot10^{-8}\) | \(8.2\cdot10^{-8}\) |
| \(\mathrm{N}_2\)—Fe | \(4.5\cdot10^{-3}\) | 11 900 | \(7.0\cdot10^{-6}\) | \(3.7\cdot10^{-5}\) |
| CO—Fe | \(1.3\cdot10^{-3}\) | 9 350 | \(1.7\cdot10^{-5}\) | \(9.0\cdot10^{-5}\) |
| He—\(\mathrm{SiO}_2\) | \(1.3\cdot10^{-8}\) | 1 400 | \(3.2\cdot10^{-6}\) | \(1.7\cdot10^{-5}\) |
| Ne—\(\mathrm{SiO}_2\) | \(1.1\cdot10^{-9}\) | 2 400 | \(1.2\cdot10^{-7}\) | \(6.3\cdot10^{-7}\) |
the initial value of the diffusion rate is restored. When working with quartz, only transparent specimens should be used. Opaque quartz is so porous that it lets all gases through.
As is evident from Table 1, the diffusion rates of gases are small, but sufficient for most cases when pure gases are required.
In addition to the systems listed in Table 1 (the data of which represent a summary of results by many authors), other systems can also be used. Thus, for example, with hydrogen and deuterium one may use Ni, Pt, Mo, Cu, Fe, Al, Pd, and \(\mathrm{SiO}_2\) (quartz). The diffusion rate of deuterium proves to be approximately one quarter less than that of hydrogen. Oxygen can diffuse through Ag, Ni, and Cu. It is interesting to note that the rate of its diffusion through nickel does not depend on pressure if it is equal to or greater than 0.25 mm Hg. This is explained by the fact that under these conditions an equilibrium is established between oxygen and the NiO layer on the metal surface. Nitrogen diffuses through Mo, Fe, and Cr and does not pass through Cu and Ni. Carbon monoxide passes through Ni and Fe. There is reason to think that CO, when diffusing through Fe, passes in molecular form; in the case of nickel, however, carbon monoxide apparently decomposes into the corresponding atoms, which recombine on the other side of the wall.
The design of the instrument for obtaining gases by the method described is very simple. The instrument consists of a tube of the required material, wound with a wire heated by current. This tube (welded to a wider tube connected to the vacuum system into which the gas is introduced) is surrounded by a cylinder. The cylinder, filled with the desired gas at the required pressure, constitutes the source of the gas introduced into the vacuum system. In cases where helium and neon are required in spectroscopic quantities, the diffusion (quartz) tube may be placed directly in the atmosphere, since the diffusion rate of other gases through quartz is much lower.
N. Khlebnikov, Moscow
REFERENCES
- E. L. Jossem, Rev. Sci. Instr., 11, 164, 1940.
- Richardson, Nicol and Parnell, Phil. Mag., 8, 1, 1904.