US3259779A - Improving efficiency of infrared radiation generation by alkali metal vapor lamps and prolonging their useful lives - Google Patents

Description

July 5, 1966 N, C, BEESE 3,259,779

IMPROVING EFFICIENCY OF INFRARED RADIATION GENERATION BY ALKALI METAL VAPOR LAMPS AND PROLONGING THEIR USEFUL LIVES Filed NOV. 17, 1951 J Mrk ze l l INVENTOR l /V. .B-'55- 1 g 4' Y :e0 40 55 55 .57

T/ME /fv H0025 ATTORNEY United States Patent 3 259 779 IMPROVING EFFHCIECY 0F INFRARED RADIA- TION GENERA'HDN BY ALKALI METAL VAPQR Lhl/.EIS AND PRQLONGING THEIR USEFUL L Norman C. Beese, Verona, NJ., assigner to Westinghouse Electric Corporation, East Pittsburgh, Pa., a corporation of Pennsylvania Filed Nov. 17, 1951, Ser. No. 256,916 1 Claim. (Cl. 313-221) This invention relates to discharge lamps and, more particularly, to such adapted to be modulated for signal purposes.

The principal object of my invention, generally considered, is an alkali metal vapor lamp which is operable at a higher than normal temperature in order to improve its eiliciency in the lgeneration of infrare-d radiations, vas well as prolong its useful .life by effecting regeneration of alkali metal employed from a decomposable compound thereof which forms during operation.

Another object of my invention is a cesium or rubidium Vapor lamp, whose operating life decreases inversely with power input, |but which may Ibe operated with improved efiiciency so as to regenerate the active metal from a compound thereof formed during operation, the lamp for that purpose being operated at a temperature higher than the decomposing temperature of the compound.

A further object of my invention is the production of a chemical lamp operable yat such a Ihigh temperature that ya compound of the active metal lformed during operation is decomposed, due to the high temperature in the lamp, so as to make available such metal for a longer period of time, as well as increase the efficiency of operation.

Other objects and advantages will become apparent as the description proceeds.

Referring to the drawing:

FIGURE 1 is an elevational view approximately to scale of a lamp embodying my invention.

FIGURE 2 is a graph illustrating how the efficiency of lamp operation is improved by higher current input.

FIGURE 3 is 4a diagram showing how the resonant spectral lines are not only increased in intensity when the lamp is operated at a higher current input, but also how certain non-resonant lines are greatly increased in intensity at the same time.

AFIGURE 4 is a diagram showing relative intensities of infrared output and operative lives of such a vapor lamp withvariations in the power input.

'FIGURE 5 is an elevational view of an improved form of electrode Which may replace those of FIGURE l.

Alkali metal vapor lamps, such as those described in my Patent No. 2,562,887, dated August 7, 1951, the continuation-in-part thereof Serial No. 672,205, filed May 25, 1946, now abandoned, and my application iiled simultaneously'herewith and entitled, Discharge Lamps .and Power Amplification of Signals Therefrom, Serial No. 256,915, now abandoned, and of which pend-ing applications the present application is a continuation-impart, operate satisfactorily with an arc current in accordance with the disclosures of said cases. For example, with a lamp such as disclosed in my patent referred to, the arc current may be 5 to 51/2 amperes. When cesium is used, the infra-red energy is concentrated almost entirely in the two resonance lines at 8521 and 8944 A.U. Throughout the useful life of about 500 hours the intensity of the resonance lines is practically constant.

At the end of such life, the excess cesium metal had disappeared, the bulb or envelope turned brown, and the output suddenly dropped t-o a low value. Under these conditions the lamp then operated as a practically pure ar- Fice gon arc lamp, assuming argon as the gas filling, with only weak infrared radiations. The cesium had combined with the glass bulb or with the glaze coating on the inside of the bulb, to form a cesium compound such as the silicate or oxide, which removed it from further use in the arc stream. At higher currents, for example, 6 or 7 amperes, the life was appreciably less, while the increase in infrared radiation was approximately proportional to the increase in arc current. At 10 amperes arc current one would expect a realtively short life and an infrared intensity twice as great as that of 5 amps.

When cesium vapor lamps were operatted between 4 and 10 amperes arc current and measurements made on the intensities of the emitted various infrared spectral lines, it was found that the two resonance lines increase quite uniformly with increase in current as was anticipated. The non-resonance lines between 8500 A.U. and 15,000 A.U. were Weak below 5 amperes arc current, but increased rapidly with increased current. At about 10 amperes, the non-resonance near infrared lines had a combined intensity quite comp-arable to ythe sum of the two resonance lines. In other words, at about 10 amperes the energy radiated lby the near infrared lines was about 4 times as great as that obtained at 5 amperes arc current. Tlhe eliciency `of the radi-ated energy at l0 amperes thus became about twice as great as at 5 ampenes FIGURE 3 shows the spectral distribution, as measured with a Perkin-Elmer infrared spectrometer of a cesium vapor lamp operating at 9.15 amperes arc alternating current. The relative intensities are galvano-meter deflections and not absolute values.

The sum yof the intensities of the infrared spectral lines throughout life -is shown by the graph 11 in FIGURE 4, for one of these lamps :burned at labout l0 amperes alternating current. The break in the curve near 40 lhours of life is the point at which the free cesium metal had disappeared from the lamp. The continued high output of radiation when the supply was maintained at about l0 amperes lalternating current was a pleasant surprise. This rta-diaiton is modulable and can be used for infrared communication. A probable explanation of this phenomenon is that Ia compound of cesium, probably, the monoxide, is formed in the lamp during operation and which is quite stable at a temperature of about 300 C. with 5 amperes arc current. Ten amperes arc current is enough `to raise the minimum bulb Wall temperature to approximately 500 C. which is high enough to decompose cesium monoxide into cesium and oxygen, thereby regenerating metallic cesium in the arc stream, but below the temperature at which the bulb or coating thereon softens. l

The cyclic process of formation and decomposition of cesium monoxide, using the bulb wall as au intermediate step, can explain the prolonged supply of cesium vapor at high-temperature high-current operation. Thermodynamic considerations show this process to be a reasonable explanation, since it is capable of giving a vapor pressure of cesium of the correct order of magnitude to produce the observed spectral intensities.

'Equation 1 shows how the decomposition of cesium monoxide may account for the high infrared output soon after the free cesium metal has disappeared. A gradual decrease in total output as burning is prolonged indicates that a more stable compound is formed that does not readily yield cesium vapor. Equation 2 cannot, by itself, account for the required cesium vapor pressure in a lamp having good infrared intensity, but cesium dioxide may be an end product that makes cesium unavailable for indefinite life.

The changes in bulb coloration indicate the formation of a chemical compound that dissociated and migrated at elevated bulb temperatures. At first the brown discoloration of the inner bulb occurred at the warmest parts of the bulb. With prolonged burning after the cesium metal had disappeared, the warmest part of the lamp became the clearest because of chemical decomposition. The lamp then operated as a chemical lamp, that is, in which a chemical compound is decomposed thermally to supply free metal at sufficient vapor pressure to generate radiation of the desired spectral lines.

With the knowledge of these chemical reactions and radiation phenomena that occur under high-temperature operating conditions, one can design small lamps of good eiciency with satisfactory life for small size communication equipment, and yet having .a long operating range. Even in cesium Vapor lamps designed specifically for light sources of high luminous efliciency, the above chemical equations define the equilibrium conditions that occur between cesium vapor and the cesium oxide compounds.

Although the foregoing is specific to cesium, yet similar comments apply to comparable lamps in which rubidium has been substituted for cesium. Rubidiu-m has resonance lines at 7800.3 .and 7947.6 A.U., and other strong nearinfrared lines at 10,082, 13,237, 13,444, 13,667, 14,754, 15,290 A U. While radiations of wave lengths longer than about 11,000 A.U. are invisible to cesium photoelectric cells, thalofide photoconductive cells, and infrared image tubes used on sniperscopes, yet the strong rubidium lines between 13,000 and 15,000 A.U. can be readily detected with a lead sulphide photoconductive cell having a response from .about 7500 A U. to about 33,000 A.U. For a secret communication system, a filter can be used with a rubidium lamp to screen off all radiations with wave lengths shorter than 12,000 A.U. and then a lead sulphide cell used for a detector.

Rubidium monoxide decomposes at 400 C., that is, at about the same `temperature at which cesium monoxide decomposes. Hence, the operation of rubidium vapor lamp at high current density and high bulb temperature, as described for cesium vapor lamps, is appropriate to obtain similar results.

Referring now to the drawing in detail, like parts being designated by like reference characters, one embodiment of the lamp itself, designated 12, is illustrated in FIG- URE 1. It comprises an inner generally-cylindrical, but spherically-ended envelope 13 of vitreous material such as glass, preferably that sold by Corning Glass Works as No. 705, containing an inert gaseous lilling and a small quantity of either metallic cesium or metallic rubidium, depending on the kind of lamp being manufactured. Inert gas pressures ranging from about cm. to about 30 cm. of mercury may be employed. It was found that the optimum pressure or argon gas was near 20 cm. of mercury, .and while it is ordinarily employed on account of low cost and availability, the heavy rare gases, krypton and xenon, serve to increase the efficiency of such devices. Mixtures of neon and argon have also been found satisfactory.

The selected alkali metal is distilled into the lamp before it is sealed off and there is a preferably visible excess of the liquid metal while the lamp is in normal operation. Lamps in accordance with specific embodiment here contemplated, that is, with diameters between 1 and 11/2 and 4 to 6 long, the alkali metal should be about .2 gram, as in the Patent No. 2,562,887, referred to. Such a lamp operates satisfactorily with a power consumption between 50 and 125 watts.

In each end of the lamp is mounted an oxide-coated filamentary electrode, designated 14 and 15, respectively. Electrode 14 is shown supported on leads 16 and 17 extending through the sealed portions 18 and 19. The other electrode is similarly constructed, that is, it has supporting leads and 21 passing throughout the sealed portions 22 and 23 of the envelope. For small lamps, such as here contemplated, filamentary electrodes without the ring electrodes disclosed ,in the preceding cases referred to, but with the auxiliary wire electrode 30 illustrated in FIGURE 6 of my application tiled simultaneously herewith, are considered desirable. Each. electrode 14 and 15 is desirably for-med as a coil of fine tungsten wire overwound upon a coil of tungsten filament and coated with alkaline earth oxide to make it eiciently emit electrons when heated.

FIGURE 5 illustrates a form of electrode 14a which may replace either of lboth of the electrodes 14 and 15. The electrode 14a is like those numbered 14 and 15 except that, in addition to the filamentary portion, there is a wire 30 formed of tantalum or zirconium and about .02" diameter. Such a wire takes the place of the toroidal or other auxiliary electrode structure, absorbing part of the energy of the discharge, disclosed in my patent No. 2,562,887 referred to, and has an outer portion 40 bent at an angle and connected with one of the leads, such as that designated 21a, which is desirably the same lead to which the straight or uncoiled portion 4of the filamentary electrode is connected. The other end of the wire 30 desirably extends beyond the filamentary portion of the electrode toward the other electrode, as illustrated, and is desirably bent to the curved form illustrated, in order to furnish an auxiliary electrode or anode-acting portion thereof, as well as a getter or clean-up member for removing residual gases. By virtue of the central or axial positioning of the auxiliary electrode 30, concentration of the light source is augmented.

A cesium or rubidium vapor lamp, like a sodium vapor lamp, must operate at `a temperature well .above ambient in oder to obtain a sucient vapor pressure of the metal. In order to conserve heat and make the lamp operate efliciently, it is mounted inside .an evacuated outer housing or inside of a double wall vacuum ask known as a Dewar flask. In the present embodiment, the inner envelope 13 with its associated parts is held inside an outer generally-cyclindrical envelope 24, of vitreous material such as glass, by means of supporting leads 25, 26, 27 `and 28 passing through the press 29 thereof. The leads 25 and 26 are extended between the inner and outer envelopes and connect respectively with the leads 16 and 17 as indicated at 31 and 32.

The envelope 25 which I propose, in accordance with the embodiment specifically described, is one of glass manufactured by the Corning Glass Works under the trademark Nonex. The leads 27 and 28, in turn, connect respectively with the leads 20 and 21, by means of flexible conductors 33 and 34. The leads 25, 26, 27 and 28 extend out of the outer envelope 24 and respectively connect with the outer contact members 35, 36, 37 and 38 projecting from the lbase 39.

The inner envelope 13 is held in place in the outer envelope 24, Vas by means of metal rings 41 and 42 having lugs apertured for the reception of the leads 25 and 26. The rings 41 and 42 are held on the leads 25 and 26 in any desired manner, such as that described in my Patent No. 2,562,887, previously referred to. Each ring 41 and 42 is of such an internal diameter that it seats the corresponding generally-spherical end of the inner envelope 13, `and has a plurality of pheripheral tabs which extend diagonally and frictionally engage the inner surface of the outer envelope 24, to thereby hold the inner envelope in fixed position with respect thereto.

After the inner envelope has been mounted on the supporting leads extending from the press portion 29 of the outer envelope and sealed, it is inserted in the outer envelope 24 and the latter evacuated and sealed. A conventional getter may be used in the envelope 24 to maintain a high vacuum throughout .the life of the lamp. The whole assembly may be supported by having the outer envelope formed with a pair of ridges 4S and 49 between which a clamp grips and is suitable secured to a support,

The material which I propose to use for the envelope 13 of my lamp is glass which is quite resistant to attack by alkali vapors. At present, I prefer such manufactured as No. 705 (now designated No. 7050 iu accordance with the new nomenclature) by the Corning Glass Works. This is a sodium borosilicate glass with a small percentage of alumina. It is designated as a tungsten sealing glass and sold under the trademark Pyrex, laboratory No. G-705-AJ. It has a softening point at 703 C., an annealing point at 496 C., la coeflicient of expansion of 46x10-7 between 0 and 300 C., and a density of 2.23. The logw of its resistivity at 350 C. is 6.77. Its power factor is .0033 (expressed as a decimal-not as percent), its dielectric constant 4.9, and its loss factor .0161 (expressed as a decimal-not as percen-t). The composition of this yglass is substantially as follows:

Percent SiOZ 67 12.203 22 A1203 21/2 Na20+K2O 7 Misoel. 11/2 To further improve the chemical stability of the lamp envelope, it is desirably interiorly coated with a glaze composed largely of boric acid, similar to the way sodium lamps are protected. This treatment makes the device a practical lamp of long life expectancy. The -above statement with regard to known glasses is made without prejudice to the possibility of a better glass Ibeing developed.

Cesium and rubidiu-m vapor lamps, like all alkali metal lamps, are dil'licult to start so provision must be made to have either the line voltage supply greatly in excess of the normal operating voltage, or some auxiliary external device must be used for starting. My Patent No. 2,562,- 887, previously mentioned, is referred to for instructions as to desirable methods of starting.

FIGURE 2 lis a graph showing how the eiciency of lamp operation is improved by increasing the input to an extent much higher than .previously considered normal. As will be seen, the output in infrared energy gradually increases along the curve 50 as the input increases up to 5 amperes representing the increase in the resonance energy from cesium vapor with increase in arc current, that is, the gradual building up of the energy output as 8521 and 8944 A.U. designated respectively 51 and 52 in FIGURE 3.

Beyond 5 amperes, however, the non-resonance lines at 13,589 A.U. and 14,695 A.U. respectively designated 53 and 54, and at other places, increase non-linearly, that is, -much faster than the input current, thereby causing the curve 50 to be concave upward, that is, to have its ordinates increase at a faster rate than its abscissas. Although these graphs of FIGURES 2 and 3 are specific to cesium, corresponding graphs may be constructed for rubidium. It will also be understood tha-t although a cesium photocell does not respond to radiations beyond about 11,000 A.U. yet, as previously explained, a lead sulphide oell may be employed for detecting all of the desired radiations.

The graph 11, as previously explained, represents the output of a cesium vapor lamp at 10 amperes, showing how it decreases with time. In the same way the graphs 55, 56 Iand 57 represent the outputs of a similar lamp operated at 7, 6, yand 5 amperes, respectively. AIt will be noted that whereas when operated at 10 amperes, the Alamp after breaking from its initial output line gradually tapers off, yet when operated at 7 amperes and below, the failure is substantially instantaneous and complete, due to the absence of regenerization of free metal, as previously explained.

From the foregoing, it will be seen that I have devised a method of operating alkali metal vapor lamps of the cesium and rubidium types, whereby the monoxide is regenerated to free metal due to the high operating temperature and complete and instantaneous failure is avoided.

Although preferrd embodiments of my invention have been disclosed, it will be understood that modifications may be made within the spirit and scope of the appended claim.

I claim:

A vapor lamp comprising a cylindrical envelope formed of borosilicate glass with a small percentage of alumina, designated as a tungsten sealing glass, with a softening point at 703 C., an annealing point at 496 C., a coeicient of expansion of 46 10'I between 0 and 300 C., and a density of 2:23, the loglo of its resistivity at 350 C. being 6.77, it power factor being .0033 (expressed as a decimal-not as percent) its dielectric constant being 4.9, and its loss factor being .0161 (expressed as fa decimal-not as percent), the interior surface of which is coatedwith a film composed largely of boric acid and protective against alkali metal vapor, electrodes in said envelope between which a discharge occurs upon the application of electrical energy thereto, and an inert gaseous iilling admixed with rubidium in said envelope.

References Cited bythe Examiner UNITED STATES PATENTS 1,830,312 11/1931 Compton 313-174 1,984,426 12/1934 Pirani et al. 313-174 2,018,815 11/1935 Spencer 313-174 2,042,261 5/1936 Kreit 313-174 2,053,501 9/1936 Spencer 313-208 2,05 6,926 10/ 1936 Kreift 313-221 2,104,073 1/ 1938 Druyvesteyn et al. 313-225 2,114,536 4/1938 Kirsten 313-344 2,116,689 5/1938 Rompe 313-221 2,562,887 8/ 1951 Beese 313-195 GEORGE N. WESTBY, Primary Examiner.

NORMAN H. EVANS, WILLIAM G. WILES, y

Examiners.

J. H. LINSCOTT, R. JUDD, Assistant Examiners.

Images