The following artical originally appeared in "Electronics" magazine in October 1950.
Typical Ba flash getters for glass bulbs (upper row) and getter strip assemblies.
|By||WERNER ESPE||Tesla National Works Prague, Czechoslovakia|
|MAX KNOLL||Dept. of Electrical Engineering, Princeton University, and RCA Laboratories Div., Princeton, N. J.|
|MARSHALL P. WILDER||Tele-Radio, I?tc. Stamford, Conn|
THE USE of getter materials is based on the ability of certain solids to collect free gases by adsorption, absorption or occlusion.This effect is widely utilized in the field of electronics to shorten the exhaust period and to improve and maintain a high vacuum or the purity of an atmosphere of noble gases.
Bulk getters are sheets or wires of gas-absorbing metals, which usually are heated for this purpose by mounting them on hot electrodes af the tube. In some cases the heating is accomplished by a separate heating source.
Coating getters are generally applied to those electrodes of vacuum tubes, which during service are maintained continuously at tem-peratures between 200 and 1,200 C. Such getters usually consist of non-volatile metal powders that are sintered upon the electrode surfaces.
In the case of flash gettering, chemically active, comparatively volatile metals-mostly metals of the alkaline earth group are evaporated by heating their supports at the conclusion of the pumping process. The metal vapor before and during condensation reacts in-stantaneously with all other than noble gaseous residues and forms on all cold parts of the tube, particularly on the tube walls, the so-called getter mirror. This surface, because of its large area, is capable of binding chemically or physically gases that are liberated during the life of the tube. With respect to the mechanism of get-tering and the action of gases get-ter materials may be divided into two groups : corrosion type and solution type. From this aspect, barium (Ba) is the typical example of chemical corrosion by gases and zirconium (Zr) an example of solution of gases in a metal. Corrosion takes place if the oxide film is por-ous and incoherent, thus not preventing further oxidation, which usually happens only if the volume of the oxide is smaller than that of the metal to be oxidized. In this case, the sorption of gas from the surrounding atmosphere can con-tinue uninterruptedly, Oxides of the alkali and alkaline earth metals have less volume than the metal.
From a technical aspect the wide application of earth alkali metals for flashing getters in vacuum tubes is founded upon the fulfill-ment of the following requirements :-
|(1)||During the standard degas-sing procedure at 400 C the getter should have low vapor pressure (p < 10-2 mm Hg) .|
|(2)||The getter should be readily vaporized at its activation tempera-ture. This temperature range must be high enough so that the getter may be readily degassed prior to the flashing; on the other hand, it must be sufficiently low to avoid the evaporating, melting or loosening of electrode metals. For nickel sup-ports this establishes a temperature range of 600 to 1,000 C.|
|(3)||After flashing, the getter deposit on the tube glass wall must have a negligibly low vapor pressure (p < 10-7 mm Hg) assuming the operating temperature of the vacuum tube at 200 C.|
|(4)||Between ambient and oper-ating temperature the getter should be very active for all gases, espe-cially oxygen. The oxide film crea-ted must be porous and incoherent in order that the gases may diffuse without hindrance into the interior of the getter mirror and may be absorbed also by getter molecules in the volume of the getter material.The specific action and applicability of Ta, Cb, Zr, Tb, Ti, Al, Mg, Ba and P as getters for vacuum tubes are shown. They can be formed as wires, sheets, tubes and tablets directly or indirectly heated; they may be powder coatings on hot electrodes; or they can be flashed|
|(5)||The chemical compounds should be stable in order that during operating temperature or electron bombardment the absorbed gases are not expelled. For gas-fllled tubes, the absorbed residual gas should not be replaced during operation by the filling gas, for example Hg-vapor in tubes.|
The first requirement is satisfied by all earth alkali metals. As seen from Fig. 1, Mg is near the border line, and cannot be used in tubes with high operating temperature because of the danger of migra-tion of Mg atoms. Requirement 2 is met by all earth alkali metals but not by aluminum. Requirement 3 is met by all earth alkali metals with the exception of magnesium. The fourth requirement, with em-phasis on activity, is met by all earth alkali metals with the exception of magnesium and aluminum. In a similar manner requirement of an incoherent oxide layer is not met by magnesium and aluminum. Requirement 5 (stable reactions) is met by all earth alkali metals but only up to a temperature of 200 C (regarding the oxides created even up to higher temperatures).
The materials Sr, Ba and Ca, are the most suitable for flash
getters in high-vacuum electron tubes. Concerning the replacement
of ab-sorbed gases by mercury va-por, however, all earth alkali
metals fail with the exception of magne-sium, which is the reason
that the latter metal is used in mercury tubes.
There are few quantitative data that permit a comparison of the getter-efficiency of different metals for different gases. Table I shows such a comparison for flashed de-posits of Al, Mg, Th, U, misch-metal and Ba and the gases O2, H2, N2 and CO2. It exhibits Ba as the most efficient getter among the metals investigated. The higher efficiency of diffuse deposits is due to their much larger surface, re-sulting from their finely divided state. This phenomenon, known as dispersal gettering is illustrated by the black Ba deposit.
The practical choice of the proper getter material is a function of factors other than efficiency alone. The broad use of high-efficiency flash getters requires considerations of insulation, interelectrode capacitance, contact potential and secondary emission of vacuum-tube electrodes and insulators, which often suggest their avoidance. On the other hand, Ba flash getters are preferred for oxide-coated cathodes because the Ba does not poison the cathodes and in some cases will improve the BaO cathodes. Judicious use of shields and proper positioning of the flash getters avoids most of their disadvantages.
It has been known for a long time that almost all metals (after thorough degassing) are capable of adsorbing gases on their surfaces.
Certain metals are capable of incorporating gases, even noble
gases, by solution in their bulk volume. The classical example is
tantalum which, on account of this property, plays a predominant
role in the construction of high-power transmitting tubes. After
degassing in a high vacuum for several hours at a temperature of
1,600 to 2,000 C, tantalum is capable of ab-sorbing gases in
amounts up to several hundred times its own volume. The optimum
gettering temperature for tantalum appears to be in the
neighborhood of 1,000 C. At temperatures above 1,500 C, the
gettering action is reversed. The maximum getter effect is
secured, therefore, by dimensioning tantalum anodes so that
during normal service the electrodes operate at red to yellow
heat. Generally, the high price of tantalum sheets and wires
limits the use of whole tantalum electrodes to particularly
valuable tubes and suggests the coating of electrodes with
tantalum powder. This is mostly performed by applying very fine
tantalum powder on the surface of anode sheets in the finished
assemblies. They are sintered together while simultaneously
degassing these powders during the pumping operation by means of
high-frequency heating or by electron bombardment On account of
the high degassing temperature re-quired for tantalum, only
molybdenum or tungsten is suitable as a base metal for tantalum
powder. Tantalum should never be hydrogen-fired because of
embrittlement and consequent destruction by this gas.
The main disadvantages of tantalum are the high materialand the high temperature range required for proper degassing and subsequent gettering operation.
During recent years columbium getter pellets have been introduced to the vacuum technique. These pellets are approximately three to five millimeters in diameter and one to three millimeters high, and consist of oxide-free columbium metal.
The getter pellets must occupy a position in the tube where they can be heated to a high temperature during exhaust by either high-frequency induction or electron bombardment. The position of the pellet must be such that the temperature is maintained by either radiation or electron bombardment at approximately 500 C. This temperature is not critical but must be above 400 and less than 900 C. A temperature of 1,650 C is needed to outgas columbium pellets because at this temperature occluded and absorbed gases are expelled and col-umbium oxide is volatilized. A tem-perature lower than 1,650 C will not accomplish this expulsion of columbium oxide. The outgassing time may vary from a period of five minutes to a somewhat longer time. The preferred manner for supporting the columbium getter pellet is to weld a molybdenum wire to it. The temperature of out-
Zirconium has valuable gettering characteristics and has come
into wide use during the past decade. It forms very stable solid
solutions (or compounds) with such gases as O, N, CO and CO2.
Zirconium metal is cheaper than tantalum and requires somewhat
lower operating temperatures. Zirconium is available either in
solid metal form (sheets or wires) or it may be applied in the
form of a powder to base metals (molybdenum) as described above.
The proper outgassing temperature for zirconium lies between
1,000 and 1,700 C, which is attained by either direct or indirect
heating of the zirconium metal or of the base metal to which
zir-conium metal or zirconium powder has been applied. Wherever
it is deemed inadvisable to heat the getter material to this
temperature range, an outgassing temperature of 700 C must be
considered minimum for activating the surface of the zirconium
getter. While zirconium is effective as a getter from about 400 C
on, it is most active at temperatures up to 1,600 C if used, for
example, on molybdenum and carbon anodes.
The solubility of H2 in Zr at room temperature equals 1,500 times its own volume at 1 atmosphere. Sorption begins at 300 to 400 C and is completed at 500 C. As the temperature is increased, the metal frees H2, but at 850 C the H2 is again taken up during transition from alpha to beta Zr. Above 850 C, H2 is evolved. Sorption and desorption are reversible with decrease in temperature. Preliminary heating to a high temperature is a necessary condition for the sorption of H2 at lower temperatures. During a rapid pas-sage from a high temperature (above 1,200 C) to room temperature, a large amount of H2 is quickly taken up.
Oxygen as well as N, dissolve homogeneously in Zr. When a Zr rod covered with a thick white oxide layer is heated in vacuum, the metallic luster reappears. Water vapor is cleaned up between 200 to 250 C. Care must be taken that a part of the Zr getter re-mains at a low temperature (approximately 400 C) during operation in order to bind the H2, while another part must assume much higher temperatures (approximately 800 C) in order to absorb 02, N2 and other gases.
Zirconium metal in sheet form 0.002 to 0.005 inch thick is used on locations that can be properly out-gassed and which operate in the temperature range indicated above. Very often zirconium sheets are mounted to grid shields, cathode supports, and other structures, which during operation attain a temperature of 600 to 800 C. In small tubes zirconium sheets, cylinders, or ribbons are used for cathode supports, grid supports and radiation shields. Zirconium wire of 0.005 to 0.020-inch diameter can be mounted for direct heating, being heated whenever absorption of gas is required, or it may be mounted for the same purpose by winding zirconium wire around MO rods or other suitable structures.
ontinuous gas absorbers such as that shown in Fig. 2 provide a support for the Zr wire and are operated from 350 to 1,700 C, for example, in x-ray tubes in series or in parallel with the filament (the temperature being adjusted by proper length of the wire). They have also been used for shortening the degassing time of electrode systems during pumping.
Using a Zr wire spiral of fifteen turns on a 0.040 inch Mo mandrel, treated and outgassed for one hour at 1,700 C, a pressure of 5 x 10-4 mm Hg was reached in ten minutes instead of thirty minutes with the high vacuum pump alone. This auxiliary pumping getter can be used repeatedly even though exposed to atmospheric pressure between pumping cycles, no additional outgassing between cycles being required. In every case the Zr absorber maintained a higher vacuum at a considerably higher effective pumping speed than the 20-liter-per-second high-vacuum pump.
A convenient way of using zirconium is to spray the tube parts with fine zirconium powder. Such powders, of particle size between 1 and 8 microns, may be suspended in a temporary binder such as nitro-cellulose dissolved in amyl acetate. For high voltage tubes a permanent binder such as colloidal silicic acid has been used with success. Such a binder has the further advantage of not giving off gaseous products dur-ing outgassing and operation. The amount of binder is usually two to five percent. This mixture is sprayed on the electrode parts, which in turn are fired in vacuum in order to remove the binder or to convert the binder to a stable compound . Nickel electrodes (pre-heated at 1,000 C, operating at 200 to 500 C), molybdenum electrodes (preheated at 1,300 C in vacuum and operating at 800 C) and graph-ite electrodes can be satisfactorily coated with such mixtures. Quantitative data on the sorption of different gases at different temperatures by powdered Zr are shown in Table II. Other methods of applying zirconium powder to electrodes have been reported, such as a suspension mixture consisting of paraffin, naphthalene, xylene and methanol or deposition of zirconium powder by cataphoresis.
Zirconium hydride (ZrH4) may also be applied to MO, Ni, Fe or graphite anodes or grids as a paste, by spraying or cataphoretic precipi-tation, and reduced to pure Zr upon heating. This compound compared with pure Zr powder presents the advantage that at lower temperatures the zirconium is tied up and protected against oxidation or poisoning during seal-in and ex-haust. Then as the temperature is raised, metallic zirconium is formed, liberating its hydrogen completely in vacuum at about 800 C. Thus, for coating on carbonized plates, which liberate much ad-sorbed gas during exhaust, zircon-ium hydride may be preferable to the pure metal since not only is the combined zirconium protected from the evolved gases but furthermore the hydrogen that it liberates at higher temperatures apparently reduces the last traces of adsorbed oxygen in the carbon layer.
Zirconium has been used successfully in high-power transmitting tubes, especially tubes having thoriated tungsten filaments, small microwave tubes and gaseous discharge tubes. Zirconium is inert to mercury vapor.
Stable solutions or compounds of Zr are formed with most gases including water vapor, with the exception of hydrogen.
The chief disadvantage of zir-conium as a getter is that the optimum temperature for the sorption of hydrogen is too low for the effec-tive cleanup of oxygen, nitrogen and the oxides of carbon. If, there-fore, the zirconium-coated part is to operate at a temperature much in excess of 300 C, a supplementary lower temperature Zr getter or a getter of the barium or barium-magnesium type should be added to absorb the hydrogen. Whether this precaution is necessary or not depends upon the tube and the amount of water vapor or pure hydrogen found within it.
During World War II, thorium, thorium alloys and mixtures of these with other getter materials were developed in Germany for use in vacuum tubes. Thorium metal is manufactured by reduction of ThO2 with Ca. Powdered Th is very pyrophoric; it is inflammable by mere friction. Electrode parts were coated with thorium powder by cataphoresis and heated for two to three hours in a vacuum furnace. For wires the coating was 5 to 10 microns thick and for sheets 1 to 2.5 mg per sq cm. The heating temperature is about 800 to 1,000 C for nickel and iron and 1,500 to 1,600 C for graphite electrodes. Considerable gas absorption is re-ported to occur around 200 C but especially in the range from 400 to 500 c. This getter is therefore suitable for power tubes and very small tubes with high anode temperatures. If the anode is covered with thorium powder its surface finish is rough, resulting in increased emissivity by blackening.
One alloy of thorium is the getter called Ceto, which comprises a 20-percent mischmetal, (chiefly cerium) and 80 percent thorium. This powder mixture is sintered at approximately 1,000 C, and the bars are milled to powder again. It is very inflammable. Ceto getter powder is transformed with amyl ace-tate into a paste that is applied to the tube electrodes in the amount of 15 to 25 mg powder per sq cm and then sintered upon the base metal in a vacuum furnace. The degassing temperature of the Ceto getter is 800 'C and a marked getter action is exhibited from 80 to 130 C up, with an absorption maximum at 200 to 500 C.
Ceto getter material has a lower secondary emission than barium. It is used when it is desired to avoid or to reduce secondary emission that might arise from the use of Ba. It cannot be used above 600 C. Ceto bridges the gap between the low-temperature flashing getter such as Ba and the high-temperature non-volatile types (Th, Ta and Zr.) Pure thorium or compounds of Th and Zr are highly pyrophoric.
A recent addition to bulk getter materials is titanium, which can be used either as bulk or as coating getter. It is not pyrophoric and its getter properties are good, besides being easy to form and machine. Blackening of the parts can be obtained by a short heating in air. At present titanium sheet is more expensive than zirconium per unit of gas absorbed.
Aluminum in its pure state is not used as a flash-getter, chiefly because its vapor pressure is too low (see Fig. 1). It vaporizes sufficiently only above 1,300 C, which is much too high for the conventional base metals like Ni. On the other hand, aluminum plated on Fe to a thickness of about 15 microns shows a considerable coating-getter effect for traces of O2 being bound by the carbon content of Fe electrodes and released only slowly during the life of the tube. After heating at 700 to 800 C in vacuum, the aluminum forms a black compound with the Fe base (FeAl3 or FeAl5) . This blackening increases the total emissivity to a level that is equal to or greater than that of carbonized Ni anodes. Such Al-plated sheet-iron has been widely used in Europe for anodes in receiving and amplifier tubes with oxide-coated cathodes.
The trade name for sheet iron plated with aluminum on both sides is PZ-iron; plated with Al on one side and with Ni on the other-PN- iron.
Pure magnesium possesses many desirable properties for a getter material, such as availability in suitable forms, and being comparatively stable and volatile under vacuum at convenient temperatures around 500 C. Unfortunately, the gettering power of magnesium is not high because most gases are only physica!ly absorbed.
As a re-sult, magnesium by itself is not used in high-vacuum tubes. The only evidence of its use is in Hg-vapor-filled rectifiers and in certain types as a grid coating powder to reduce secondary emission.
In order to obtain a material of greater stability and safety in use than pure magnesium, the so-called Formier getter was developed, It consists of aluminum-magnesium alloy powder (55 percent Al, 45 percent Mg) which is applied suspended in a nitrocellulose binder and applied to tube parts as a paint. On account of the limited gettering powers of magnesium, Formier is used only when other types of getters with higher evaporation temperatures must avoided. Magnesium getters are difficult to degas, have little gas absorption up to temperatures of 175 C and absorb only oxygen. High vapor pressure precludes use in small tubes and at high operating temperatures.
The active ingredient of most flash getters is. barium, which is used in combination with aluminum, magnesium, tantalum, thorium, strontium or calcium. The getter is attached to the electrodes in the form of a pellet, or more frequently, to a special metallic support within the tube as shown in the top row of the accompanying photograph. It is mounted in such positions as to insure that the vapor stream produced is not splashed against such parts as the stem or the insulated lead-in wires. Shielding screens of metal, mica or ceramic materials are often provided to prevent this. The getter pellet must be attached to parts that during the pumping process may be readily heated to the evaporation temperature of the getter. This heating is performed by electron bombardment or, more frequently, by high-frequency induction from coils arranged outside the tube.
Flash getters of pure barium have the disadvantage that the un-protected barium reacts at room temperature with oxygen or with water vapor, thereby becoming in-active. This condition may be prevented by using: a protective layer or casing, alloys of Ba that are inert at room temperature or by generating the gettering material in the vacuum tube by a chemica1 reaction between stable Ba compounds and deoxidizing agents to form a reaction-type getter. Alloys of barium with magnesium and aluminum are relatively stable at room temperature and yield pure barium upon dispersal or flashing of the getter. The percentage of metals in standard alloys for getter tablets are: Ba 25, Mg 55, Al 20; Ba 37, Mg 37, Al 26 ; Ba 43, Mg 20, Al 37 (known by the trade name Kemet) . The tablets are mounted on nickel flags of various shapes.
Barium-magnesium alloys yield very little Ba metal in the flashing and are, therefore, seldom used in modern high-vacuum tubes. Ba-Al getters provide much larger amounts of Ba metal, which accounts for their wider use. A common disadvantage of both types of alloys is their rapid rate of deterioration upon exposure to air and the necessity of vacuum packing.
Another commercial assembly is shown in Fig. 3. Other types comprise short pieces of iron, nickel and copper-clad barium wires to be mounted on a support of Ni-sheet, which can be high-frequency flashed at 900 to 1,100 C. Trade names of these materials are Feba, Niba and Cuba.
Several x-ray tubes of European make use iron-clad barium
(Feba) getter made in wire form, 2 mm diameter by 15 mm long. The
getter is mounted within a miniature oven consisting of a ceramic
tube into which the getter just fits. A tungsten spiral heater is
wound on the outside of the ceramic tube. Care must be taken to
avoid the possibility of migration of the barium to active tube
elements. After sealing off the tube, the getter is flashed by
heating the spiral.
Examples of the reaction type Ba getters are the reduction of BaO by Al to Ba (Alba getter) and of BaCO3, or barium berylliate by Ta to Ba. In the case of BaCO3, a tantalum wire heater is coated with a mixture of BaCO3, and SrCO3, (SrCO3 prevents fusing of BaCO3) . At 800 to 1,100 C the carbonates dissociate to oxides and at 1,300 C the oxides react with Ta to form metallic Ba, whereby 40 percent of the theoretical Ba yield is obtained in the so-called Batalum process. Also barium berylliate (BaBeO2,) is stable in air and is used in a directly heated getter, shown in Fig. 4, formed in the shape of a trough from a 0.040 x 0.001 in. Ta ribbon, which holds approximately 2.5 mg of Ba and gives a Ba yield of 60 percent.
Another flashing-reaction-type getter is the so-called Bato getter, which is prepared by mixing an aluminum-barium alloy with iron oxide and thorium powder. Its purpose is to provide a source of heat in the getter pellet and in so doing aims at flashing Ba metal at a relatively low getter flag temperature. The source for the Ba is the Ba-Al alloy and the latent heat so derived forms an exothermic reaction between iron oxide and metallic thorium. The powder is formed into tablets, which are pressed into nickel cups and attached to special supporting members within the tube. Since it is important to store the Bato getter in a dry atmosphere, the getter flags are usually sealed into evacuated ampoules or cans. After the pumping process, the getter is evaporated by high-frequency.
Flash getters are outgassed at temperatures between 600 and
700 C, usually by r-f heating from the outside of the tube, and
flashed at temperatures beween 900 and 1,300 C. The barium vapor
condenses on the cold surface opposite the getter material,
usually on the envelope of the tube. The appearance of the
condensed getter deposit depends upon the vapor pressure in the
tube at the time of flashing. If the getter is vaporized very
slowly, the first barium atoms evaporated will absorb the gas
present so that the remaining getter is deposited in a very high
vacuum, exhibiting a shining mirror. If flashing is done very
rapidly, however, the getter deposits in a rather high vapor
pressure and the getter mirror will be discolored due to
dispersion of the Ba. If vaporization is carried out in the inert
atmosphere of a rare gas the condensed deposit will be black,
resulting in a dispersal getter. This condition does not mean
that the getter is contaminated, but merely that the deposit is
finely divided and therefore absorbs light. Such deposits
ex-hibit higher efficiency than the bright deposits indicated in
Barium reacts with atmospheric gases such as oxygen, nitrogen and carbon dioxide at room tempera-ture, as well as with hydrogen and carbon monoxide. This absorption at low temperature makes flash getters particularly valuable for tubes that do not attain high operating temperatures. If the volume of the vacuum tube is large, multiple getter strips or several getter pellets are employed.
Flash getters have the disadvantage that during flashing metallic vapor is produced, which may settle on insulating parts or build up a conducting layer on the glass envelope. Such layers may become charged during operation or represent interelectrode capacitance. Usually flash getters are inadvisable in high voltage tubes or in microwave tubes, the latter type having very close spacing and only short insulating surfaces . It should also be considered that the opaque mirror interferes with the cooling of electrodes by radiation. In high-voltage types like transmitting and x-ray tubes fast stray electrons may hit the mirror and evaporate sufficient getter material to cause a gas discharge followed by a short circuit and destruction of the tube.
Phosphorus has a comparatively high vapor pressure (indicated in Fig. 1) and for this reason it is not used in radio transmitting, x-ray, or other high-vacuum dis-charge tubes. Inexpensive and simple to handle, it is used for high-vacuum and gas-filled lamps, espec-ially for types of below 60 watts with voltages of 90 to 250 volts.
The outgassing, flashing and operating temperatures of typical phosphorus getters are given in Table III.
The authors wish to extend thanks to E. B. Steinberg (Remington-Rand) and E. A. Lederer (RCA) for assistance and suggestions in preparation of this manuscript.
|J.Evans 2000||CLICK HERE TO RETURN TO
THE VALVE TECHIE PAGE
15th September 2001