US2622218A - Secondary-emission electron discharge device - Google Patents

Description

Dec. 16, 1952 D. A. JENNY 2,622,218

SECONDARY-EMISSION ELECTRON DISCHARGE DEVICE Filed Jan. 51, 1950 ATTORNEY Patented Dec. 16, 1952 SECONDARY-EMISSION ELECTRON DISCHARGE DEVICE Dietrich A. Jenny, Princeton, N. J assignor to Radio Corporation of America, a corporation of Delaware Application January 31, 1950, Serial No. 141,385

Claims. 1

This invention relates to improved electronmultipliers and to improved methods of and apparatus for prolonging the useful life of the secondary-electron emissive electrodes or dynodes of such devices.

Despite the many advantages of electron multiplier tubes such as very great gain and wide band operation, they have not come into widespread use as power devices because the active secondary emissive coatings on the dynodes are rapidly exhausted under intense bombardment.

Various explanations have been oifered for the relatively short life of the dynode coating material and various practices have been followed for prolonging it. According to one View, the secondary emitter coating is poisoned by evaporation products which reach it over straightline paths from the cathode. One pratcice based on this view is to block these paths with shields which can be circumvented by electrons if they are appropriately deflected, and another is to make the cathode coatings of special compositions and/or of limited areas to limit their rate of evaporation to the secondary emitter. However, the use of shields reduces the frequency range of all such devices without avoiding coating exhaustion in those used for power purposes.

According to another view, the coating decomposes under electron bombardment and one or more of its components evaporates. A prior art practice based on this View is to use special secondary-emitter coatings, such as denser coatings, which take longer to be decomposed. In support of the practice of using denser coatings, it has been supposed that such a coating will take longer to become exhausted because it contains more useful active material (inasmuch as more material can be crowded into the coating without increasing its size, and, in addition, the coating can be made thicker for a required value of front-to-back conductivity).

However, this practice has not been successful in providing a secondary emitter having a reliably long useful life under high current density conditions. The reason Why it has not been successful will be apparent from an analysis of secondary emission which is presented herein.

Accordingly, it is an object of the present invention to devise improvements in discharge devices using secondary electron emission in which the secondary emitter coatings can be depended upon to have long useful lives even under high current density conditions.

It is a further object of the present invention to devise a process for producing improved discharge devices of thekind set forth above.

It is a further object of the present invention to provide improved methods for operating discharge devices of the character indicated.

In general, according to the present invention these objects are attained by providing means for continuously reconstituting the secondary emissive coating during operation so that it will not become exhausted but will remain useful as long as any coating remains. This means includes a source for supplying an active gas which will combine with a free metal component of the coating as rapidly as it loses gas molecules due to electron bombardment.

In the drawing:

Fig. 1 is a longitudinal section through a secondary emission discharge device embodying the present invention;

Fig. 2 is a top view of the device of Fig. 1 with its envelope cover removed;

Fig. 3 is an enlargement of a portion of Fig. 1 to illustrate a secondary emissive coating; and

Fig. 4 is a fragmentary sectional view of a modification of the embodiment of Fig. 1.

The five-stage electron multiplier shown in Fig. 1 comprises an evacuated metal envelope ID consisting of a bottom 52 and a hollow cover 14 of the bathtub" type. Within the envelope the tube elements are combined in a rigid assembly I5 which is bolted to the bottom B2. In this assembly the secondary emitter electrodes are stacked one above the other and separated by mica shims to provide the multiplier stages with both D. C. isolation from each other and low impedance R. F. by-passes to ground. As is most apparent from Fig. 2, the assembly I? is rectangular in shape and is centrally positioned on the envelope bottom l2. At the bottom of the assembly is a base plate I8. This plate and a mica shim l8 by which it is insulated from the envelope bottom l2 have aligned central'openings 20, 22 through which an anode support rod [9 may extend in a manner to be described below. The right half of the base plate It (as shown in the drawing) is thicker than its ieft half. This causes the respective secondary emitter electrodes (or dynodes) which are stacked on the opposite ends of the plate It to be at staggered levels so that a zig-zag flow of electrons will occur, during operation, as represented by the dot-dash line 23.

The first, third and fifth dynodes (24, 26, 28 respectively) are stacked on the left half of base plate l6, each being insulated from its adjacent elements by two mica shims 39. Correspondingly, the second and fourth dynodes 32, 34

and a number of shims 38 are stacked on the a capsule 62, is formed of silver.

a right half of plate it. The dynodes are held tightly together between the envelope bottom 12 and a top plate 36 by a number of bolts 38. The bolts 38 are insulated from the dynodes and the top and base plates by dielectric sleeves, not shown, such as sleeves of ceramic material.

A recess 48 is formed on the under side of the top plate 35 across its central portion to provide space for an indirectly heated cathode t2 and a control grid 43. Cathode it comprises a nickel tube 44 supported at its ends by perforated mica covers (55 in Fig. 2). These covers are attached to the sides of the top plate 35 over the ends of its recess it. Two heater leads, '55, extend into the hollow nickel tube 6 3 from its opposite ends to a heater winding, not shown, which is carried bewteen them in the central portion of the nickel tube. These leads and this heater should be insulated from the inner surface of the tube M by a refractory insulating material such "as aluminum oxide.

An oxide primary emission coating 41 is formed on the portion of sleeve M which can be heated by connecting a source of potential to the heater leads at. This primary emissive coating 31 and the secondary emissive surface '35 of the first dynode 2d face toward each other with the control grid' l's mounted between them.

A secondary emissive coating 35 is formed on the central portion of the concave, inwardlyfacing surface of each dynode.

A collector electrode or anode :28 is provided for intercepting the secondary electrons from the fifth dynode 28. It is mounted to have a minimum of capacitance to ground so that even at high operating frequencies it will be free to float with respect to ground in accordance with the product of the varying output current and an appropriate, purposely-included load impedance. To this end the collector electrode 48 is supported in the manner shown in Fig. 1 so that in all directions it is well spaced from any surrounding conductive structure. For a similar reason the capacitance between the grid and the cathode as well as that between either of these elements and ground should be kept at-a minimum,

The collector electrode 38 is supported in the center of openings 29, 22 by the rod l9 and a glass bead 52 which are mounted within a hollow threaded fixture 53 to form a vacuum-tight coaxial output receptacle'fi l.

Respective leads for the cathode sleeve, the several dynodes and the grid and the two heater leads 46 are separately sealed through the --envelope bottom E2 to form individual terminal pins 55. As shown, each seal may consist of a metal tube silver soldered or welded over a hole in bottom I2, and a glass bead 5'! fused over the end of the tube with a lead sealed through it to form a terminal pin. Since the different leads are brought through the bottom [2 at positions on both sides of assembly l5 as well as beyond its ends, a number-of them do not appear in Fig. 1. However, a lead 58 for the final dynode 28 and a lead Gil for the second dynode 32 do appear therein.

The principal parts of the envelope Ill-may be made of any one of several metals which have appropriate mechanical properties (such as that of holding a hard vacuum). However, one part, Thisis done, for "reasons which will be more fully set forth "below, to make use of the fact that silver has the property of diffusing oxygen, and that its rate of diffusion can be controlled by varyin the temperature of the silver. To provide a means for varying its temperature, the capsule 52 is surrounded by a heater winding 84 which may be connected to an adjustable source of potential 66. In this way oxygenmay be controllably admitted into the envelope from the surrounding air atmosphere at the same time that other gases, such as nitrogen, are excluded.

Fig. 3 shows the condition of a preferred sec .ondary emissive coating 35 according to the resent invention after it has been activated. The preferred coating includes two distinct strata. One of thestrata, which may be called the surface layer (or"interaction layer), 68, is indispensable. It absorbs the bombarding electrons and emits the secondary electrons. The other, which may be called the backing layer, H3, is not essential. However, it is very useful as a reservoir of material into which the surface layer can progressively recede during operation (and from which its unexposed side can be replenished) as its exposed side is sputtered away. The backing layer should be conductive, so that replacement electrons can flow through it from the dynode to the surface layer to take the place of emitted secondary electrons.

The small squares, 69, shown in "Fig. 3, represent the free metal. It will be noted that in the preferred coating 35 there is more free metal in the backing layer'lil than in the interaction layer 63.

An important discovery of the present invention shows that the secondary emission ratio (the ratio of secondary electrons emitted to primary electrons received) can be controlled, and maximized, by controlling theratio of free metal to metal compound in the constituents of the surface layer. Moreover, the secondary emission ratio can be maintained at a high value, even when the coating is under extremely hard use, until it is entirely physically sputtered away, by continuously maintaining a proper ratio between "these constituents of the layer.

If there is either too little free metal or too much of it in the surface layer, the secondary emission ratio will be adversely affected. On the other hand, the ratio of free metal to metal compound is not critical in the backing layer except that there should be enough free metal in this layer for it to be a fairly good conductor.

The coatings 35 may be formed on the dynodes in accordance with any suitable prior art such as that of deposition by cataphoresis, or in accordance with any suitable future art. Moreover, they may be made of any satisfactory known secondary emissive material. Therefore, in accordance with usual practice, each coating 35 will include, as-at least one of its constituents,

a compound of a suitable metallic element and a suitable non-metal, usually a gas which may be either a gaseous element or a gaseous compound. Most of the suitable metallic elements are either alkaline earth metals or alkali metals or a few other metals, such as magnesium and beryllium, which are found in the first two groups of the periodic chart. However, not all of the metals found in the first two groups can be considered suitable since 'a few of them, silver, mercury, gold and copper have not yet been successfullycompounded to producegood secondary emitters. Nor are all of the suitable metals to be found in the first two groups. For example, thorium can be used for good secondary emitter coatings and the same promises to be true of aluminum.

. None of the suitable metals is a good secondary emitter in its pure state. To become useful as such, it must be combined with a suitable nonmetal. In general, a non-metal is usually suitable if it is normally gaseous at room temperature and if the suitable metal which is used has such an affinity for the non-metal in question that they readily combine to form a compound. Suitable non-metals comprise a group including gaseous elements such as oxygen, hydrogen and the halogens (fluorine, chlorine and iodine) and a great variety of gaseous compounds such as carbon dioxide, water vapor and carbon monoxide.

Apparently, it is generally true that, as initially formed, a secondary emissive coating contains substantially no free metal, 1. e., its freemetal-to-metal-compound ratio is zero. This is because the suitable metal constituents have a great afiinity for oxygen, and therefore any free metal which has not been converted into metal compound will become completely oxidized as the coating is subjected to air. I have discovered that it is because of this that some sort of activation has been, and is, necessary before such a coating becomes a good secondary emitter. In the past, activation has been accomplished by heating. This was suggested by primary emitter experience and empirically it proved to be useful.

As a result of my findings, it is now possible to understand what happens in activation by heating and. therefore the reason why it is only partially satisfactory.

It will be helpful to consider activation by heating before continuing with the disclosure of the present invention. During heating the metal compound forming the coating breaks down, for example, barium oxide separates into free metal barium and oxygen. More oxygen is liberated from the coating than metal vapor and therefore the free metal content rises. Relatively little recombination takes place since much of the oxygen is absorbed by the getter and other parts of the tube. As activation proceeds, the increasing free metal component of the coating becomes rather homogeneously diffused throughout the coating because of its high thermal energy. Therefore, the front-to-back conductivity of the coating is increased at the same time that the secondary emission ratio is improved. Up to this point, most of the results of heating are desirable.

In the practice of this method, it was observed that the use of more heat than necessary or the continuation of activation for a longer period than necessary did not deactivate the coating. As a result, it became an accepted practice to use activation by heating continuously during operation. However, it is now apparent that this practice shortens the actual life of the coating (as distinguished from its useful life). After activation by heating has caused the ratio of free metal to metal compound to reach a certain value, more free metal vapor will be liberated from the coating than oxygen. Therefore, excessive activation by heating, either as prolonged or extreme heating preparatory to operation or as continuous heating during operation, physically boils away the coating even though it does not deactivate it. In fact, this evaporation of free metal makes such heating entirely prohibitive for certain particular coatings. A good example is a coating which comprises magnesium. Though it affords a high secondary ratio when properly compounded, it is so volatile that if heated to a high temperature or if continuously heated it will soon be entirely vaporized. If the dynode which carries the coating comprises magnesium there will be suflicient vaporization to contaminate all of the interior elements of the tube as well as to destroy the coating itself.

In addition to the fact that the use of continuous activation by heating during operation shortens the life of a secondary emitter, it also tends to cause primary emission. Obviously, this is undesirable for a secondary emitter, since the secondary current must be controllable by electron bombardment alone.

After initial activation by heating (this has usually been done at temperatures between 800 C. and 1000 C. for periods of from 10 minutes to 1 hour), it has been customary to use continuing activation by heating at about 500 C. all during operation. It should be noted that if this were not done, the useful life of the coating under high current density conditions would be very short despite good initial activation. Continuous activation by heating is objectionable for additional reasons besides that of wasting away the coating(s).

Because of the necessity of maintaining the electrodes at such a high temperature during operation, it is often impractical to pass a varying signal through a tube which is being worked close to its rated output unless very fast-acting and critical temperature controls are used. Without such controls, a power tube which is already as hot as 500 0. might quickly burn up under certain signal conditions whereas under different signal conditions the secondary emitter coatings of the final stages might rapidly lose their efficiency. It would certainly be very expensive if not impossible to provide such controls. Moreover, most of the arrangements which have been used for providing dynode heating (for continuous activation) are actually incompatible with fast-acting, dynamically-controlled cooling. For example, in some arrangements, the dynodes have been thermally isolated from the rest of the discharge device so that heat provided by individual built-in heaters (similar to cathode heaters) would remain concentrated where needed.

It will be seen that continuing activation according to the present invention does not depend on maintaining a high temperature and therefore permits continuous use of a fixed generous amount of cooling adequate to dissipate heat generated during peak points in the operation of the device. In other words, the tube may be run cool whereby it is safe to operate it at a very high and widely varying output power level without danger of melting any of the dynodes and/or the anode.

It will be seen that the novel process for initial activation which is disclosed below is also temperature-independent. For this reason, it will be especially useful for activating tubes which are built without heaters and thermal isolation for their dynodes, but instead with heat conduction structures for cooling them.

Returning now to the disclosure of the present invention, the continuing temperature-independent activation of the present invention is accomplished, as has already been mentioned in a general way, by providing in the tube gas molecules which will combine with free metal in the surface layer of a secondary emitter at a rate which is comparable to that at which gas molecules are separatingoutfrom its metal-compound constituentis'). This con be done regardless of how the tube was activated in the first place. The gas which is provided is preferably but not necessarily the same kind that is separating out.

The initial temperature-independent activation of the present invention is accomplished by electron bombardment, preferably with the device under hard vacuum.

After the device shown in Fig. 1 is assembled, it is pumped down to a vacuum of about 10- mm. of mercury, and the tube dynodes are connected to sources of direct potential to cause a current flow through the device. At first, since the dynode coatings will contain practically no free metal, their secondary ratios will be very low. However, some zig-zag current will flow and there will be some electron bombardment of each dynode coating. This will cause a breakdown of the metal compound in its surface layer. Free gas molecules will escape and some of it will be absorbed by the getter and other elements of the tube.

Accordingly, a progression will be established in which the ratio of free metal to metal compound will increase; this will increase the secondary current flow; and this will accelerate the rate of breakdown and hence the rate of activation. At the same time there is a progressive improvement in the conductivity ofthe backing layer for the following reasons: Once the secondary emission ratio of any dynode is greater than one, a potential gradient will develop between the front and back surfaces of its secondary emissive coating due to: (l) the high front-to-back resistance v of'the coating, and (2) the loss of negative charge from its surface layer. This will cause the ionic transport of free metal particles from the front toward the back of the coating.

This activationprocess should be terminated after a period of from two to eight hours, when the rate of improvement has passed a maximum and decreased to zero, i. e., has leveled off. This point can be recognized by the fact that the anode current will have stopped increasing. By this time the ratio of free metal to metal compound in the surface layer will have attained the proper value for optimum secondary emission, and the backing layer will contain quite a large amount of free metal and will accordingly be quite conductive.

In the normal operation of the tube of Fig. 1, assuming that its secondary emitters have been activated in some suitable manner, the temperature of capsule 62 is adjusted to diffuse enough oxygen into the envelope so that the recombination thereof with the free metal constituent of the coating will offset the continuous breakdown of the compound. Some recombination will occur because the thermal velocities of the oxygen molecules will cause them to come into contact with the secondary emissive coatings. However, most of the recombination will occur because ionized molecules will be attracted to the coatings. Many of the oxygen molecules will be ionized in the zig-zag electron stream 23 and will then fall back along it in the opposite direction of movement of the electrons' Each positivelycharged ion will strike the most negative dynode which it sees (usually next back toward the cathode from the place where the molecule became ionized). This ionic recombining is very advantageous since it proceeds at the highest rate for. the. dynode which is losing the greatest number of gas molecules, 1. e., the dynode which is receiving the mostintense bombardment and. is providing the largest secondary current. The dense electron currents moving to and from this dynode will produce more ions in its vicinity than are being produced near any of the other dynodes. Consistent with this, I have found that in general if a proper amount of gas is provided for the continuous restoration of any dynode, this amount is substantially correct for all of them.

The zig-zag electron stream 23 is much narrower than the assembly [5 and only extends across small central region of each dynode. Therefore, ions formed in this stream are not significantly attracted by any part of the field of the envelope (assuming that the envelope is at a lower potential than the dynodes) which extends into the inter-dynode space. Accordingly, most of the ions move in directions to impinge upon the dynodes and therefore most of them accomplish the useful function explained above. For this reason, the gas pressure within the envelope may be as low as (and for certain gases lower than) 10- mm. of mercury at all times.

Fig. 4 shows a modification which is suitable where thegas to be provided within the envelope for continuous activation during operation comprises hydrogen instead of oxygen. In this modification the silver capsule (62) is replaced by a capsule 13, of a different material, e. g., palladium, which has the property of diffusing hydrogen. In addition, since hydrogen does not exist in substantial quantities in the atmosphere; a supply of this gas is confined about the outside of capsule 73 by a vessel E5. It should be noted that if the surface layer 68 includes an oxide, such as barium oxide, at the start of operation, it may end up with that constituent converted to barium hydride.

In general, the residual pressure which should be left after evacuation-pumping and the operating pressure which should be maintained during gas diffusion may be very nearly as low as a so-called hard vacuum of 10- mm. of mercury. The exact pressure will depend on the affinity of the free metal font-he gas which is provided during operation. For example, fluorine is very active, so that if used at all, it should be used at lower pressure than that suitable for oxygen. However, fluorine would be harmful to a cathode of the type now in use, and therefore might not be preferred at the present time, this also being true of chlorine.

An alternative or additional arrangement may be employed for providing gas molecules within the envelope. In it, the anode as will be formed of, or coated with, silver which contains a considerable amount of occluded oxygen. Though this sort of supply is exhaustible, it may last for a very considerable period of time. Moreover, inasmuch as the life of the secondary emitter coatings is not indefinite, but will eventually terminate when the coatings are finally physically sputtered away, it is not necessary that the supply of oxygen be inexhaustible. In the operation of the device, the amount of oxygen liberated from the anode will vary directly with the bombardment of this electrode. Therefore, the supply of gas will automatically change in the proper direction to cause the rate of reconstitution of the coatings to follow changes in their rate of breakdown. By providing adjustable anode cooling, as by blower 3! of Fig. l, or adjustable anode heating, as by a built-in heater winding (not shown) it will be possible to control the rate at which the anode diffuses the occluded oxygen by controllably influencing its average temperature. If desired, a tube element especially intended for the purpose, rather than the anode, may be used for containing the occluded oxygen. Moreover, wherever it is stored, the supply of oxygen need not necessarily be in the form of occluded free gas. For example, it may be bound in a compound and releasable therefrom as compound is decomposed by being heated.

Once the tube is in operation, thermal diffusion of free metal in the backing layer 10 will be relatively slow due to the low operating temperature. This fact, in combination with the continuous ionic transport of free metal from the front toward the back of this layer will tend to establish and maintain a reserve of nearly-pure, highlyconductive metal deep in the coating.

In testing one embodiment of the present invention in which the secondary emissive coatings included a mixture of magnesium oxide and barium oxide, it has been found that by continuously providing a proper amount of oxygen within the envelope it was practical to maintain an average secondary emission ratio of five for a period of one hundred hours with the final dynode emitting a current of higher than one ampere per centimeter squared. Moreover, since it was possible to operate the tube cold, no difliculty was occasioned due to evaporation of magnesium.

The non-metal constituent (of the compoundingredient of the secondary emissive coatings), which is described herein as being gaseous at room temperature and as having an afiinity for the metal constituent, includes all of the substances which are referred to as oxidizing agents in both the restricted and broad meanings of the expression. In the restricted meaning, these include the element oxygen and a number of compounds including oxygen, such as water vapor, carbon monoxide and carbon dioxide, which can provide an oxygen molecule to join with a molecule of the metal. In the broader meaning they further include a number of other substances any one of which has such a tendency to join with the metal that it can provide a molecule to displace an oxygen molecule already joined to the metal.

It is apparent from the foregoing that the present invention provides an improved discharge device, as well as a process for making such a device and a method for using it, in which the secondary emitter coatings can be depended upon to have long, useful lives even under high current density conditions.

What is claimed is:

1. A discharge device comprising a vacuum envelope containing a dynode having a secondary emissive coating including a, compound of a metal with a non-metal, and means for continuously reconstituting said coating by providing within the envelope in gaseous form a supply of a nonmetal, such as said first-mentioned non-metal, for which said metal has an aflinity. said means comprising a portion of said vacuum envelope which is formed of a material which has the property of selectively diffusing said last-ment-ioned non-metal.

2. A discharge device as in claim 1 in which said non-metal for which said metal has an affinity is oxygen, and said material forming a portion of the envelope is silver.

3. A discharge device as in claim 1 in which said means comprises means for controllably heating said portion of said envelope.

4. A discharge device as in claim 1, in which said non-metal for which said metal has an affinity is hydrogen, and said material forming a portion of the envelope is palladium.

5. A discharge device as in claim 1, in which said means includes a reservoir connected to said envelope around said portion and containing a supply of said non-metal for diifusion through said portion into said envelope.

DIETRICH A. JENNY.

REFERENCES CITED The following references are of record in the file of this patent:

UNITED STATES PATENTS Number Name Date 1,566,279 King Dec. 22, 1925 2,228,945 Bruining et a1 Jan. 14, 1941 2,242,644 DeBoer May 20, 1941 2,393,803 Nelson Jan. 29, 1946 2,497,911 Reilly et al Feb. 21, 1950

Images