1911 Encyclopædia Britannica/Vacuum Tube

VACUUM TUBE. The phenomena associated with the passage of electricity through gases at low pressures have attracted the attention of physicists ever since the invention of the frictional electrical machine first placed at their disposal a means of producing a more or less continuous flow of electricity through vessels from which the air had been partially exhausted. In recent years the importance of the subject in connexion with the theory of electricity has been fully realized; indeed, the modern theory of electricity is based upon ideas which have been obtained from the study of the electric discharge through gases. Most of the important principles deduced from these investigations are given in the article Conduction, Electric (Through Gases); here we shall confine ourselves to the consideration of the more striking features of the luminous phenomena observed when electricity passes through a luminous gas.

Methods of producing the Discharge.—To send the current through the gas it is necessary to produce between electrodes in the gas a large difference of potential. Unless the electrodes are of the very special type known as Wehnelt electrodes, this difference of potential is never less than 200 or 300 volts and may rise to almost any value, as it depends on the pressure of the gas and the size of the tube. In very many cases by far the most convenient method of producing this difference of potential is by means of an induction coil; there are some cases, however, when the induction coil is not suitable, the discharge from a coil being intermittent, so that at some times there is a large current going through the tube, while at others there is none at all, and certain kinds of measurement cannot be made under these conditions. Not only is the current intermittent, but it is apt with the coil to be sometimes in one direction and sometimes in the opposite; there is a tendency to send a discharge through the tube not only when the current through the primary is started but also when it is stopped. These discharges are in opposite directions, and though that produced by stopping the current is more intense than that due to starting it, the latter may be quite appreciable. The reversal of the current may be remedied by inserting in series with the discharge tube a. piece of apparatus known as a “rectifier” which allows a current to pass through it in one direction but not in the opposite. A common type of rectifier is another tube containing gas at a low pressure and having one of its electrodes very large and the other very small; a current passes much more easily through such a tube from the small to the large electrode than in the opposite direction. Sometimes an air-break inserted in the circuit with a point for one electrode and a disk for the other is sufficient to prevent the reversal of the current without the aid of any other rectifier.

There are cases, however, when the inevitable intermittence of the discharge produced by an induction coil is a fatal objection. When this is so, the potential difference may be produced by a battery of a large number of voltaic cells, of which the most convenient type, where more than a few milli-ampéres of current are required, are small storage cells. As each of these cells only produces a potential difference of two volts, a very large number of cells are required when potential differences of thousands of volts have to be produced, and the expense of this method becomes prohibitive. When continuous currents at these high potential differences are required, electrostatic induction machines are most generally used. By means of Wimshurst machines, with many plates, or the more recent Wehrsen machines, considerable currents can be produced and maintained at a very constant value.

The exhaustion of the tubes can, by the aid of modern mercury pumps, such as the Topler pump or the very convenient automatic Gaede pump, be carried to such a point that the pressure of the residual gas is less than a millionth of the atmospheric pressure. For very high exhaust ions, however, the best and quickest method is that introduced by Sir James Dewar. In this method a tube containing small pieces of dense charcoal (that made from the shells of coco-nuts does very well) is fused on to the tube to be exhausted. The preliminary exhaustion is done by means of a water-pump which reduces the pressure to that due to a few millimetres of mercury and the charcoal strongly heated at this low pressure to drive off any gases it may have absorbed. The tube is then disconnected from the water-pump and the charcoal tube surrounded by liquid air; the cold charcoal greedily absorbs most gases and removes them from the tube. In this way much higher exhaust ions can be obtained than is possible by means of mercury pumps; it has the advantage, too, of getting rid of the mercury vapour which is always present when the exhaustion is produced by mercury pumps. Charcoal does not absorb much helium even when cooled to the temperature of liquid air, so that the method fails in the case of this gas; the absorption of hydrogen, too, is slower than that of other gases. Both helium and hydrogen are vigorously absorbed when the charcoal is cooled to the temperature of liquid hydrogen.

When first the discharge is sent through an exhausted tube, a considerable amount of gas (chiefly hydrogen and carbon monoxide) is liberated from the electrodes and the walls of the tube, so that to obtain permanent high vacua the exhaustion must be continued until the discharge has been going through the tube for a considerable time. One of the greatest difficulties experienced in getting these high vacua is that even when all the joints are carefully made there may be very small holes in the tube through which the air is continually leaking from outside, and when the hole is very small it is sometimes very difficult to locate the leak. The writer has found that a method due to Goldstein is of the greatest service for this purpose. In this method one of the electrodes in the tube and one of the terminals of the induction coil are put to earth, and the pressure of the gas in the tube is reduced so that a discharge would pass through the tube with a small potential difference. The point of an insulated wire attached to the other terminal of the induction coil is then passed over the outside of the tube. When it comes to the hole, a very bright white spark may be seen passing through the glass, and in this way the leak located. The appearance of the discharge when the exhaustion is going on is a very good indication as to whether there is any leakage in the tube or not. If the colour of the discharge remains persistently red in spite of continued pumping, there is pretty surely a leak in the tube, as the red colour is probably due to the continued influx of air into the tube. Platinum is the only metal which can be fused through the glass with any certainty that the contact between the glass and the metal will be close enough to prevent air leaking into the tube. Platinum, however, when used as a cathode at low pressures “sputters,” and the walls of the tube get covered with a thin deposit of the metal: to avoid this, the platinum is often fastened to a piece of aluminium, which does not sputter nearly so much. Tantalum is also said to possess this property, and it has the advantage of being much less fusible than aluminium. This sputtering depends to some extent on the kind of gases present in the tube, as in monatomic gases, such as mercury vapour, even aluminium sputters badly.

Electrodeless Tubes.—As some gases, such as chlorine and bromine, attack all metals, it is impossible to use metallic electrodes when the discharge through these gases has to be investigated. In these cases “electrodeless” tubes are sometimes used. These are of two kinds. The more usual one is when tin-foil is placed at the ends of the tube on the outside, and the terminals of the induction coil connected with these pieces of foil; the glass under the foil virtually acts as an electrode. A more interesting form of the electrodeless discharge is what is known as the “ring” discharge. The tube in this case is placed inside a wire solenoid which forms a part of a circuit connecting the outside coatings of two Leyden jars, the inside coatings of these jars being connected with the terminals of an induction coil or electrical machine; the jars are charged up by the machine, and are discharged when sparks pass between its terminals. As the discharge of the jars is oscillatory (see Electric Waves), electric currents surge through the solenoid surrounding the discharge tube, and these currents reverse their direction hundreds oi thousands of times per second. We may compare the solenoid with the primary coil of an induction coil, and the exhausted bulb with the secondary; the rapidly alternating currents in the primary induce currents in the secondary which show themselves as a luminous ring inside the tube. Very bright discharges may be obtained in this way, and the method is especially suitable for spectroscopic purposes (see Phil. Mag. [5], 32, pp. 321, 445).

Appearance of the Discharge in Vacuum Tubes.—Fig. 15 b of the article Conduction, Electric (Through Gases) represents the appearance of the discharge when the pressure in the tube is comparable with that due to a millimetre of mercury and for a particular intensity of current. With variations in the pressure or the current some of these features may disappear or be modified. Beginning at the negative electrode k, we meet with the following phenomena: A velvety glow runs, often in irregular patches, over the surface of the cathode; this glow is often called the first negative layer. The spectrum of this layer is a bright line spectrum, and Stark has shown that it shows the Döppler effect due to the rapid motion of the luminous particles towards the cathode. Next to this there is a comparatively dark region known as the “Crookes’ dark space,” or the second negative layer. The luminous boundary of this dark space is approximately such as would be got by tracing the locus of the extremities of normals of constant length drawn from the negative electrode; thus if the electrode is a disk, the luminous boundary of the dark sphere is nearly plain over a part of its surface as in fig. 1, while if the electrode is a ring of wire (fig. 2) the luminous boundary resembles that shown in fig. 17 of the article Conduction, Electric (Through Gases). The length of the dark space depends on the pressure of the gas and on the intensity of the current passing through it. The width of the dark space increases as the pressure diminishes, and may, according to the experiments of Aston (Pro. Roy. Soc. 79, p. 81), be represented with considerable accuracy by the expression a+b/p or a+ cλ, where a, b, c are constants, p the pressure and λ the mean free path of a corpuscle through the gas. The thickness of the dark space is larger than this free path; for hydrogen, for example, the value of c is about 4.

Fig. 1.

Fig. 2.

When the current is so large that the whole of the cathode is covered with glow the width of the dark space depends upon the current decreasing as the current increases. In helium and hydrogen Aston (Pro. Roy. Soc. 80 A., p. 45) has detected the existence of another thin dark space quite close to the cathode whose thickness is independent of the pressure. The farther boundary of the Crookes dark space is luminous and is known as the negative glow or the third negative layer. Until the current gets so large that the glow next the cathode covers the whole of its surface the potential difference between the cathode and the negative glow is independent of the pressure of the gas and the current passing through it; it depends only on the kind of gas and the metal of which the cathode is made. This difference of potential is known as the cathode fall of potential; the values of it in volts for some gases and electrodes as determined by Mey (Verh. deuts. Phys. Ges., 1903, v. p. 72) are given in the table.

Cathode Fall
Gas  Electrode
Pt Hg Ag Cu Fe Zn Al Mg Na Na-K K
O2 369 . . . . . . . . . . . . . . . . . .
H2 300 . . 295 280 230 213 190 168 185 169 172
N2 232 226 . . . . . . . . . . 207 178 125 170
He 226 . . . . . . . . . . . . . . 80 78·5 69
Arg 167 . . . . . . . . . . 100 . . . . . .

The cathode fall of potential measures the smallest difference of potential which can produce a spark through the gas. Thus, for example, it is not possible to produce a's ark through nitro en with platinum electrodes with a potential difference of less tlian 232 volts, except when the electrodes are placed so close together that with a smaller potential difference the electric force between the terminals amounts to more than a million volts per centimetre; for this to be the case the distance between the electrodes must be comparable with the wave-length of sodium light.

When the current is small the glow next the cathode does not cover the whole of the surface, and when this occurs an increase in the current causes the glow to cover a greater area, but does not increase the current density nor the cathode fall. When the current is so much increased that the glow covers the whole of the cathode an increase in current must result in an increase of the current density over the cathode, and this is accomplished by a rapid increase in the cathode fall of potential. The cathode fall in this case has been investigated by Stark (Phys. Zeit. III, p. 274), who finds that its value K can be represented by the equation


where Kn is the normal cathode fall, f the area of the cathode, C the current through the tube, p the pressure of the gas and k and x constants.

The increase in the potential fall is much more marked in small tubes than in large ones, as with small tubes the formation of the negative glow is restricted; this gives rise to a greater concentration of the current at the cathode and an increase in the cathode fall. The intensity of the electric field in the dark space has been measured by many observers. Aston used very lar e plain cathodes and measured the electric force by observing the dejection of a small pencil of cathode rays sent across the dark space at different distances from the cathode. He found that the magnitude of the force at a point in the dark space was proportional to the distance of the point from the junction of the negative glow and the dark space. This law of force shows that positive electricity must be in excess in the dark space, and that the density of the electrification must be constant throughout that space. The force inside the negative glow if not absolutely zero is so small that no one has as yet succeeded in measuring it; thus the surface of this glow must be very approximately an equi-potential surface. In the dark space there is a. stream of positively electrified particles moving towards the cathode and of negatively electrified corpuscles moving away from it, these streams being mutually dependent; the impact of the positive particles against the cathode gives rise to the emission of corpuscles from the cathode; these, after acquiring kinetic energy in the dark space, ionize the gas and produce the positive ions which are attracted by the cathode and give rise to a fresh supply of corpuscles. The corpuscles which carry the negative electricity are very different from the carriers of the positive; the former have a mass of only 1/1700 of the atom of hydrogen, while the mass of the latter is never less than that of this atom. The stream of positive particles towards the cathode is. often called the Canalstrahlen, and may be investigated by allowing the streamito flow through a hole in the cathode and then measuring, by the methods described in Conduction, Electric (Through Gases), the velocity and the value of e/m when e is the charge on a carrier and m its mass. It has been found that this stream is somewhat complex and consists of—

α. A stream of neutral particles.

β. A stream of positively electrified particles moving with a constant velocity of 2×108 cm./sec., and having e/m=104. This is a secondary stream produced by the passage of α through the gas, and it is very small when the pressure of the gas is low.

γ. Streams of positively electrified atoms and perhaps molecules of the gases in tli)e tube. The velocity of these depends upon the cathode fall of potential.

The streams of negative corpuscles and positive particles produce different kinds of phosphorescence when they strike against a solid obstacle. The difference is especially marked when they strike against lithium chloride. The corpuscles make it phosphoresce with a steely blue light giving a continuous spectrum; the positive particles, on the other hand, make it shine with a bright red light giving in the spectroscope the red lithium iine. This affords a convenient method of investigating the rays; for example, the distribution of the positive stream over the cathode is readily studied by covering the cathode with fused lithium chloride and observing the distribution of the red glow. Goldstein has observed that the film of metal which is deposited on the sides of the tube through the sputtering of the cathode is quickly dissipated when the positive stream impinges on it. This suggests that the sputtering of the cathode is caused by the impact against it of the positive stream. This view is supported by the fact that the sputtering is not very copious until the increase in the current produces a large increase in the cathode fall of potential. The magnitude of the potential fall and the length of the dark space are determined by the condition that the positive particles when they strike against the cathode must give to it sufficient energy to liberate the number of cathode particles which produce, when they ionize the gas, sufficient positive particles to carry this amount of energy. Thus the cathode fall may be regarded as existing to make the cathode emit negative corpuscles. If the cathode can be made to emit corpuscles by other means, the cathode fall of potential is not required and may disappear. Now Wehnelt (Ann. Phys., 1904, 14, p. 425), found that when lime or barium oxide is heated to redness large quantities of negative corpuscles are emitted; hence if a cathode is covered with one of these substances and made red hot it can emit corpuscles without the assistance of an electric field, and we find that in this case the cathode fall of potential disappears, and current can be sent through the gas with very much smaller differences of potential than with cold cathodes. With these hot cathodes a luminous current can under favourable circumstances be sent through a gas with a potential difference as small as 18 volts.

The dimensions of the parts of the discharge we have been considering—the dark space and the negative glow-depend essentially upon the pressure of the gas and the shape of the cathode, and do not increase when the distance between the anode and cathode is increased. The dimensions of the other part of the discharge which reaches to the anode and is called the positive column depends upon the length of the tube, and in long tubes constitutes by far the greater part of the discharge. This positive column is separated from the negative glow by a dark interval generally known as the Faraday dark space; the dimensions of this dark interval are very variable—it is sometimes altogether absent.

The positive column assumes a considerable variety of forms as the current through the gas and the pressure are varied: sometimes it is a column of uniform luminosity, at others it breaks up into a series of bright and dark patches known as striations. Some examples of these are given in fig. 17 of Conduction, Electric (Through Gases). The distance between the striations varies with the pressure of the gas and the diameter of the tube, the bright parts being more widely separated when the pressure is low and the diameter of the tube large, than when the pressure is high and the tube small. The striations are especially brilliant and steady when a Wehnelt cathode covered with hot lime is used and the discharge produced by a number of storage cells; by this means large currents can be sent through the tube, resulting in very brilliant striations. When the current is increased the positive column shortens, retreating backwards towards the anode, and may, by using very low currents, be reduced to a glow over the surface of the anode. The electric force in the positive column has been measured by many observers. It is small compared with the forces which exist in the dark space; when the luminosity in the positive column is uniform, the force there is uniform; when the positive column is striated there are periodic variations in the electric force, the force being greater in the bright parts of the striation than in the dark.

Anode Drop of Potential.—Skinner (Wied. Ann. 68, p. 752; Phil. Mag. [6], 8, p. 387) has shown that there is a sudden change in potential between the anode itself and a point in the gas close to the anode. This change amounts to about 20 volts in air; it is thus much smaller than the cathode fall of potential, and it is also much more abrupt. There does not seem to be any region at the anode comparable in dimensions with the Crookes' dark space in which the drop of potential occurs.

The highly differentiated structure we have described is not the only way in which the current can pass through the tube. If a large Leyden jar is suddenly discharged through the tube the discharge passes as a uniform, continuous column stretching without interruption from anode to cathode; Goldstein has shown (Verh. deutsch. phys. Ges. 9, p. 321) that the spectrum of this discharge shows very interesting characteristics.  (J. J. T.)