US2761086A - Electric discharge lamp - Google Patents

Description

Aug. 28, 1956 E. B. NOEL ET AL 2,761,086

ELECTRIC DISCHARGE LAMP Filed Aug. 29. 1952 l 20 50 40 50 LENGTH INCHES LENGTH'JNCH-Sw John I FaLconev dr.,

United States Patent Office 2,761,086 ELECTRIC DISCHARGE LAMP Application August 29, 1952, Serial No.. 307,082 6 Claims. (Cl. `S13-184) This invention relates to electric discharge lamps containing an ionizable medium including a metal vapor` and an inert starting gas. lt is more particularly concerned with high pressure `discharge lamps in the form of elongated and relatively slender tubes which, by reason of conventional mode of manufacture, may be somewhat irregular in diameter along their length.

'In the `design of photochemieal lamps suitable for photoprinting on sensitized paper, sometimes known as Whiteprinting, it is desirable to provide a lamp in the `form of an elongated tube having high radiation elliciency at 3650 A. with substantially uniform intensity of radiation per unit length of the lamp. The necessity for substantial uniformity arises by reason of the fact that in the whiteprinting process, the sensitized paper passes immediately below the lamp, moving in a direction transverse to the length of the lamp. If the radiant output is not substantially uniform per unit length of lamp, the printing will be uneven and will be streaked with darker bands corresponding to the regions of lower output. Thus a line source of substantially uniform` radiation is required.

In order to realize good eiiciency at 3650 A., the electric discharge is effected through mercury vapor at a relatively high pressure, for instance at 1A; to 5 `atmospheres. A highmercury pressure of course entails a high operating temperature so that it is necessary to utilize for the envelope of the lamp a quartz capable of withstanding the high temperatures involved. But when lamps were made having an envelope of commercial straight drawn quartz tubing and containing the usual filling of mercury with argon as a starting gas, it was found that the 3650 A. output of such lamps was extremely non-uniform along the length of the lamp. could be corrected by utilizing precision bore or sized quartz tubing. However such quartz tubing is` prohibitively expensive, costing for instance twice as much as the commercial drawn tubing whereof the cost is already a substantial part of the total cost of the lamp.

Accordingly the object of the present invention is to provide a new and improved lamp of the type discussed above and having substantial uniformity of radiant output perunit length thereof, without using precision bore or sized tubing for the envelope.

Another object of the invention is to provide a new and improved elongated source of 3650 A. radiation having substantial uniformity of radiant output along its length and having an efficiency superior to that heretofore realizable.

In the course of the studies which culminated in the present invention, we discovered that the non-uniformity in `radiant output per unit length in lamps made of commercial straight drawn quartz tubing was associated with variations in diameter of the `tubing along .thelength material such as t It was also found that the non-uniformity i Patented Aug. 28,' 1956 thereof. Ordinary commercial straight drawn quartz tubing may vary in diameter up to 20%, and nonuniformity was observed in tubing varying in diameter as little as about 5%. We further discovered that `the non-uniformity of 3650 A. output in such tubes was associated with the use of a starting gas such as argon. When the filling pressure of the starting gas was reduced, the non-uniformity also decreased, and, in a lamp con-k taining only mercury without any starting gas, it sub-` stantially disappeared. t It will be appreciated however that a commercially practical lamp requires a starting gas in order to have a reasonably' low starting voltage and short warm-up time. We further discovered that for a lamp having a given starting gas lling pressure, the non-uniformity became more pronounced as the load-` ing, thatis the wattage input into the lamp, was increased.

We have found that the non-uniformity of radiation is associated with variations in the partial pressures of the mercury vapor and of the starting gas along the length of the lamp. Throughout the whole length of the lamp, the partial pressure of each component is a substantial` fraction of thetotal pressure.` The total pressure in the lamp at any `one time is constant throughout the lamp; however `in the regions of smaller diameter where the 3650 A. radiation decreases, the partial pressure of the starting gas is a greater percentage of the total pressure, and the partial pressure of the mercury is proportionately less. For instance, the partial pressure of mercury may vary by as much as l to 2.5 as between regions of smaller and larger diameter. It was also observed that the spectral distribution of the radiation in the regions of high 3650 A. output shows a higher arc temperature, and conversely the arc temperature is lower in the regions of low 3650 A. output. Proceeding on these discoveries, it was theorized that the non-uniformity might be reduced by using a starting gas having a lesser thermal conduc-` tivity so as to reduce its cooling effect upon the arc in the regions of small diameter. fact that when krypton is used as the starting gas, the uniformity is improved and when finally Xenon isl substituted for the filling gas, a substantially uniform out.-

put is realized. Furthermore, with Xenon as the starting gas, not only is a uniform output achieved, butthe eiliciency of the lamp unexpectedly improves by as much as 10% beyond what has previously been considered thel maximum possible efliciency, namely the efliciency of the lamp containing only mercury vapor without any starting gas whatsoever.

For a fuller understanding of the invention and` of the advantages which may be realized through it, attention is now directed to the following description and accompanying drawings. The features of the invention believed to be novel will be more particularly pointed out in the appended claims.

In the drawings:

Fig. l is a side view of a form invention.

Fig. 2 is a sectional view of one end of the lamp showing the electrode and seal structure.

Fig. 3 is a graph illustrating typical variations in diameter of commercial straight drawn quartz tubing such as that of the lamp of Fig. l, along the length thereof.

Fig. 4 is a graph illustrating light output along the of lamp embodying our length of a lamp for various pressures of argon gas filling.

It was then found in` comprises an elongated slender tubular envelope 1 having thermionic electrodes 2, 2' sealed into. opposite ends thereof, and containing a small quantity of mercury represented by the droplet 3, together with a illing of an inert gas, which, in accordance with the invention, will include a substantial proportion ot' xenon. In Fig. l, the lamp is shown with a central portion of the envelope removed in order to facilitate the illustration: in an actual construction, the lamp may have an internal diameter of mm. and a length of 125 cm. The electrodes 2, 2"may be of any suitable type, preferably either activated with electron emissive material such as alkaline earth oxides or having other electron emissive materials such as thorium associated therewith. As il-lustrated in the drawing, the electrode comprises a tungsten rod 4 projecting into the end of the envelope which is somewhat reduced in diameter, and having wound about it a tungsten slip-over coil 5. A narrow strip or sliver of thorium metal (not visible in the drawing) is enclosed within the coil or helix 5, being laid alongside the tungsten rod 4. The other end of the tungsten rod 4 is welded to a tungsten foil 6 which makes a hermetic seal to the quartz, the quartz being collapsed about it at the end of the lamp.

' The invention is of particular value in the case of a tube or envelope 1 made of a vitreous material varying appreciably in diameter. For instance the tube may be made of commercial straight drawn quartz of approximately 10 millimeters inside diameter which, in accordance with ordinary commercial manufacture, may vary in diameter up to about 20%. Referring to Fig. 3, curve 11A illustrates the variation in diameter of a typical tube about 48 inches (125 cm.) long along the length thereof. The range of variation for this lamp is approximately 10%' and this is fairly typical of commercial drawn quartz tubing. It will be observed that the variations appear as short irregular jumps which are superimposed uponrelatively long sweeps producing a peak at 5 inches and another at 35 inches approximately. The long sweeping variations are the ones which apparently cause the non-uniformity of output; the short term irregular variations do not appear to have any noticeable effect upon the uniformity of the discharge. The 3650 A. radiation will exhibit a pattern corresponding in general to the long sweeping variations in diameter. As a rule, it appears that variations in diameter will tend to cause non-uniformity of output when the cycle of variations or distance between successive peaks or valleys exceeds 15 times the inside diameter of the tubing. The variations may be as large as of the internal diameter of the tubing, the effect on the non-uniformity varying of course with its magnitude. in precision bore or sized tubing, the variations do not exceed 1 to 2%, and non-uniformity of 3650 A. output then ceases to be a problem.

The curves of Fig. 4 illustrate the relative 3650 A. radiation for various pressures of argon as the starting gas. Curves 21, 22, 23 and 24 show the relative output for-25, 121/2, 6 and 0 millimeters of argon respectively. It will be observed that reducing thev pressure of the argon has two etfects: firstly the over-all output of the lamp increases, and secondly the non-uniformity becomes less pronounced, and substantially disappears in the case of 0 millimeter of argon, that is, when no starting gas is included in the lamp. It will be appreciated of course that as the argon pressure is reduced, the warm-up time increases. Thus in the case of the lamp without any starting gas, the lamp was started by heating it with a torch so as to raise the mercury vapory pressure sufiiciently to allow the establishment of a discharge. Of course in a lamp for commercial use, such an expedient is, impractical so that a reduction of the filling pressure of theA starting .gas does not afford an acceptable solution to the problem of reducing non-uniformity of output.

Referring to Fig. 5, curves 31, 32 and 33- illustrate the relative 3,650 A. radiation for this type ofv argon lled lamp under wattage inputs of 2200, 1400 and 1000 watts respectively. Although the non-uniformity de.- creases fairly rapidly as the wattage input is decreased, the total radiant output decreases at a rate which is greater than the decrease in wattage. Thus although it is possible to reduce the non-uniformity by reducing the Wattage input, no real solution is afforded because the eiliciency of the lamp drops more than can be tolerated.

The non-uniformity effects described above were first noted in lamps of the general conguration of that illustrated in Fig. l, comprising a quartz arc tube of approximately 10 mm. inside diameter and defining an arc gap approximately 125 crn. (48 inches) long. The nominal wattage, voltage, and current of this lamp are 1400 watts, 1675 volts and 1.0 ampere respectively. It will be appreciated that the actual wattage is somewhat less than the product of volts by amperes by reason of the nonsinusoidal waveform of the current. The quantity of mercury is such as to be all vaporized when these operating conditions are attained, in accordance with the teachings of U. S. Patent 2,247,176, Pirani et al. For the lamp described, the quantity of mercury is approximately milligrams. The starting gas iilling of these lamps previous to the invention was argon gas at approximately 25 millimeters pressure at room temperature. Our studies indicated that the non-uniformity of 3650 A. output under those conditions became appreciable when the variations in diameter of the tubing exceeded 2%, and in order to utilize commercial drawn quartz tubing, variations in diameter of as much as 20% had to be taken into account. These variations in diameter are significant when they occur over a cycle, that is over a length between successive maxima or minima exceeding 15 times the inside diameter.

In accordance with our invention, it was found that when xenon is employed as a starting gas, the 3650 A. radiation of the lamp becomes substantially uniform even at pressures of xenon as high as about 40 mm. of mercury. These results are illustrated in Fig. 6 wherein curve 41 shows the relative 3650 A. radiation for a lamp such as has been described and containing argon as the starting gas at 25 mm. pressure. When the same lamp is resealed with xenon at 25 mm. pressure, the relative 3650 A.

radiation becomes substantially uniform over the whole length of lamp as illustrated by curve 42. It will be appreciated that the curves of Fig. 6 were measured from another lamp than that illustrated by the diameter variation characteristic of Fig. 3. The lamp from which were measured the curves of Fig. 6, as will appear upon inspection, had a maximum diameter near the center and minimum diameters at the ends, such being the reverse of the lamps from which characteristics of Figs. 4 and 5 were measured. However the range of variations in diameter is substantially the same in all instances and like results are obtained.

In addition to the improvement in uniformity of 3650 A. output, the substitution of xenon for argon quite unexpectedly gives an appreciable increase in total light output over a mercury-argon mixture and even over mercury alone. The relative output for the same lamp not including any starting gas is illustrated by curve 43. It will be seen that the addition of 25 mm. of xenon actually gives an increase of approximately 10% in 3650 A. output over the theoretically optimum condition of no starting gas illustrated by curve 43.

Our further studies after the discovery of the remarkable improvements both as to uniformity and efficiency resulting from the use of xenon and forming the basis of the present invention, have indicated that the phenomenon of non-uniformity of radiant output in the prior art lamps is associated with variations in the partial pressures of the mercury vapor and ofthe lling gas throughout the length of the tube. To test this hypothesis, the partial pressure of the mercury was measured at the positions of: maximum and minimum intensity on a lamp containing 5 argon as the starting gas. The partial pressure of the mercury in a particular region of the lamp may be determined by cooling asmall area of the bulb wall in that region and measuring the temperature at which the mercury started to condense out. The condensation of the mercury may readily be determined by means of a photometer which measures the intensity of the radiation from the region in question; as soon as mercury begins to condense,`the intensity of radiation starts to drop. The measure of the condensation temperature of the mercury gives of course a measure of the mercury pressure, and

`the results indicated that the mercury pressure varied by as much as 2.5 to 1 as between regions of maximum and minimum diameter. These results were confirmed by measuring the spectral distribution of the radiation as between these different regions. The spectral distribution showed a considerably higher arc temperature in the regions of maximum diameter than in the regions of minimum diameter.

Our iindings point to the conclusion that in a relatively long and slender discharge tube having variations in diameter of the order described, nonuniformity of radiant output will occur inasmuch as the loading is such that the partial pressure of the mercury and the partial pressure of the starting gas are both substantial fractions of the total partial pressure, provided further that the said starting gas has a substantial cooling elfect upon the arc. In applying this rule, it must be kept in mind that the pressure of the starting gas increases directly with the absolute temperature of the lamp. For instance, when the lamp considered here is iilled with argon at 25 mm. at room temperature and operated at an average internal temperature of approximately 2700 C., the total pressure of the mercury vapor and argon is approximately eleven hundred millimeters. However at this temperature, the argon pressure has increased approximately l fold and becomes approximately 250 mm., the mercury vapor of course exerting the remaining 850 mm. of pressure. Thus in this specific instance, the argon filling exerts approximately one-quarter of the total pressure and the mercury exerts three-quarters thereof. In such a lamp, the nonuniformity of 3650 A. output is very pronounced whenever the tubing has variations in diameter of the order in question here. In general it appears that the phenomenon will occur under the stated condition when the starting gas exerts, during operation of the lamp, a partial pressure between and 75% of the total pressure exerted by the starting gas and the mercury vapor. Further experiments have indicated that non-uniformity effect may be just as pronounced with a mm. diameter lamp as with 10 mm. diameter lamp when both have the same percentage diameter variation. In` general, it appears that non-uniformity becomes a serious factor under the stated conditions with discharge lamps having an arc voltage drop and loading in excess of 4 volts and 5 watts respectively per centimeter length of the envelope.

In accordance with the invention, the undesirable nonuniformity characteristic may be substantially reduced or even eliminated under the stated conditions by using a starting gas such as xenon having a lower thermal conductivity than argon. Our experiments have conrmed that at least for the inert gases neon, argon, krypton and xenon (named herein in decreasing order of thermal conductivity) the non-uniformity is most pronounced with neon and least or substantially non-existent with xenon. Helium appears to provide an exception since being the lightest in atomic Weight and hence having the highest thermal conductivity, it would be expected that the non-uniformity would be most pronounced; in fact, the opposite is true and heliumgives a very uniform output although the efficiency of production of 3650 A. radiation is too low for a commercial lamp. The explanation for this may reside in the fact that helium has a very high diffusion coeicient so that it may beimpossible for the segregation of helium and mercury in the variousportions` of the tube to occur. i The theoretical explanation which has been proposed would also appear to indicate that radon having a higher atomic weight than xenon and hence a lower thermalconductivity, would be superior even to xenon in achieving a uniform output. However due to the scarcity of radon, its radioactive character, and its short half-life, it isnot suitable as a starting gas for a commercial lamp.

While a certain specific embodiment of the invention has been shown and described in detail, it will be understood that the invention is equally applicable to lamps falling within the stated ranges of parameters and the appended claims are intended to cover any such modifications Which come within the true scope and spirit of the invention.

i What we claim as new and desire to secure by Letters Patent of the United States is:

1. A high pressure gaseous electric discharge lamp forming a line source of radiation comprising an elongated tubular envelope having a length at: least l5 times its diameter and having diameter variations greater than 2% occurring along the length of said lamp with a cycle extending over at least l5 diameters and the inten nal diameter of said envelope -being in the range of approximately 10 to 25 millimeters, a pair of electrodes I sealed into opposite ends of the envelope, and an ionizable atmosphere therein comprising a starting gas and mercury vapor, the quantity of said starting gas being proportioned such that its partial pressure at the operating temperature of the lamp is a substantial fraction of the total pressure exerted `by the said starting gas and the mercury vapor, the said starting gas comprising predominantly an inert gas having a thermal conductivity less than that of argon whereby the radiant output is substantially uniform per unit length of lamp irrespectively of the said variations in envelope diameter.

2. A lamp as in claim l wherein the inert starting gas is predominantly xenon.

3. A lamp as in claim l wherein the quantity of the starting gas is such as to exert a partial pressure at the operating temperature of the lamp which lies in the range of 5% to 75% of thetotal pressure developed by the said starting gas and mercury vapor and the quantity of mercury is sufficient to produce at normal operating temperature a voltage drop between the electrodes in excess of 4 volts per centimeter with a loading of 5 watts per centimeter length of lamp.

4. A lamp as in claim 1 wherein the quantity of the starting gas is such as to exert a partial pressure at the operating temperature of the lamp lwhich lies in the range of 5% to 75 of the total pressure developed by the said starting gas and mercury vapor, the quantity of mercury is sufiicient to produce at normal operating temperature a voltage drop between the electrodes in excess of 4 volts per centimeter with a loading of 5 watts percentimeter length of lamp, and the starting gas is predominantly xenon.

5. A lamp as in claim 1 wherein the lamp is operable with a wattage input not less than 5 Watts per centimeter length thereof to develop a total pressure in the range of 1A to 3 atmospheres, the quantity of said starting gas is proportioned such that the partial pressure exerted thereby under normal operating conditions of said lamp is a substantial fraction of the total pressure of said starting gas and mercury vapor, and the said starting gas consists predominantly of xenon at a pressure in the range of a few centimeters whereby to realize a total radiant output exceeding that realizable without any starting gas in the envelope.

6. A lamp as in claim 1 wherein the lamp is operable with a Wattage input not less than 5 watts per centimeter length thereof and a voltage drop not less than 4 volts per centimeter length to develop a total pressure in the lamp of 1A to 3 atmospheres, the quantity of mercury being proportioned to be totally vaporized under the operating conditions vand the quantity of said starting gas being proportioned such that the partial pressure exerted thereby under normal operating conditions of said lamp is a substantial fraction of the total pressure of said starting gas and mercury vapor, and the starting gas consists predominantly of Xenon at a pressure in the range of a few centimeters whereby to realize a total radiant output in excess of that realizable without any starting gas in the envelope.

References. Cited in the le of this patent UNITED STATES PATENTS Zecher Feb. 10, 1931 Claude Oct. 6, 1931 Pirani June 24, 1941 Hillman July 18, 1944 Johnson f Nov. 28, 1944

Claims (1)

1. A HIGH PRESSURE GASEOUS ELECTRIC DISCHARGE LAMP FORMING A LINE SOURCE OF RADIATION COMPRISING AN ELONGATED TUBULAR ENVELOPE HAVING A LENGTH AT LEAST 15 TIMES ITS DIAMETER AND HAVING DIAMETER VARIATIONS GREATER THAN 2% OCCURING ALONG THE LENGTH OF SAID LAMP WITH A CYCLE EXTENDING OVER AT LEAST 15 DIAMETERS AND THE INTERNAL DIAMETER OF SAID ENVELOPE BEING IN RANGE OF APPROXIMATELY 10 TO 25 MILLIMETERS, A PAIR OF ELECTRODES SEALED INTO OPPOSITE ENDS OF THE ENVELOPE, AND AN IONIZABLE ATMOSPHERE THEREIN COMPRISING A STARTING GAS AND MERCURY VAPOR, THE QUANTITY OF SAID STARTING GAS BEING PRONORTIONED SUCH THAT ITS PARTIAL PRESSURE AT THE OPERATING TEMPERATURE OF THE LAMP IS A SUBSTANTIAL FRACTION OF THE TOTAL PRESSURE EXERTED BY THE SAID STARGING GAS AND THE MERCURY VAPOR, THE SAID STARTING GAS COMPRISING PREDOMINANTLY AN INERT GAS HAVING A THERMAL CONDUCTIVITY LESS THAN THAT OF ARGON WHEREBY THE RADIANT OUTPUT IS SUBSTANTIALLY UNIFORM PER UNIT LENGTH OF LAMP IRRESPECTIVELY OF THE SAID VARIATIONS IN ENVELOPE DIAMETER.

Images