US 2714684 A
Description (OCR text may contain errors)
Allg- 2, 1955 G. MEISTER ET AL 2,714,684
LOW PRESSURE FLUORESCENT AND DISCHARGE LAMPS Filed June 29, 1949 afwas-Pase; aP 50% Kr, 50%,? me zr/v/rf 2.5%/Ve eo aa /aa 40 20 a ATTORNEY United States Patent @nice 2,7lfi84 Patented Aug. 2, 1955 LGW PRESSURE FLURESCENT AND DESCHARGE LAMPS George Meister, Newark, and Thomas H. Heine, Montciair, N. l., assign-ors to Westinghouse Electric Corn poration, East Pittsburgh, Pa., a corporation of Penn- Sylvania Application .lune 29, i949, Serial No. 102,016 3 Claims. (Cl. S13- 112) This invention relates to low pressure discharge lamps and, more particularly, to such employing only mixtures of rare gases and mercury as the iilling material.
The principal object of our invention, generally considered, is to produce low pressure fluorescent and discharge lamps having a iilling of a mixture of mercury and rare gases of such a composition that fluorescence is elciently excited in the phosphor.
Another object of our invention is to produce a low pressure fluorescent or discharge lamp comprising a mixture 0f rare gases in which either argon or neon is mixed with krypton and mercury in the proportion to get good eiciency at practical pressures.
A further object of our invention is to produce a fluorescent or discharge lamp which has characteristics such as above outlined, whereby it is better than a lamp having only one rare gas admixed with mercury vapor because it is easier to start; it has improved operating characteristics, including a higher ultra-violet and corresponding luminous eiiiciency, at low temperatures, resulting in a decrease in the warm-up period for steady light; and it has a lower cost for the same output.
Other objects and advantages of the invention will `become apparent as the description proceeds.
Referring to the drawing:
Figure l is an elevational view, with parts in longitudinal section of a lamp embodying our invention.
Figures 2 and 3 are graphs showing the variations in ultra-violet output as the composition of the contained gas is changed, each curve having applied thereto a numeral indicating pressure in millimeters of the enclosed gas.
Figures4 and 5 are graphs showing the variations in fluorescent light output as the composition of the contained gas is changed, each curve having applied thereto a numeral indicating pressure in millimeters of the enclosed gas.
As is well known, mercury vapor admixed with an inert or rare gas, such as argon at low pressure is commercially `employed for the generation of ultra-violet radiations which may excite phosphors to give off visible radiations in iluorescent discharge lamps.
The electrical characteristics of such low pressure `mercury discharges are known. it has been shown that, at a given lamp temperature, pressure and current, as the atomic weight of the inert gas increases, the voltage gradient in the mercury discharge and the voltage drop at the electrodes decrease.
The efficiency of such a low pressure mercury discharge uorescent lamp depends upon the mercury pres- Sure and the wattage in the discharge. It has been shown that at a constant bulb temperature and a given current, the inert gas with the greater atomic weight and lower ionization `potential has a higher efficiency than `an inert gas of loweratomic weight and higher ionization potential. Thus, `of the gases tried, krypton was better than argon or neon.
The early data was mainly taken on the individual in position opposite the Vycor inert gases mixed with mercury. We have investigated the use of mixtures of inert gases with mercury for discharge lamps.
In order to obtain as much information as possible on one experimental lamp a special tube was designed in which the central portion consisted of a 5 inch Vycor (as identified in the Rentschler Patent No. 2,469,410, dated May l0, 1949) section of 1.5 inches inside diameter which was sealed to a 1.5 inch diameter Pyrex tube at each end. The Pyrex sections could be coated with phosphors. Independently heated oxide coated cathodes were used and spaced about 24 inches apart. In addition a water-jacket three inches in diameter was designed to fit around the lamp in which the central portion again was a 5 inch Vycor section sealed with White wax to a Pyrex section at each end. With this construction it was possible to measure the light output of the phosphors, the ultra-violet output of the arc discharge, and the light output of the arc, at dilerent gas pressures and controlled temperatures with corresponding voltages and power input.
The assembled tube with coated phosphor sections was sealed onto an exhaust system which had a cooling trap, surrounded by Dry Ice throughout the experiment. A liter reservoir for allowing proper diffusion of gas mixtures was provided into which could be introduced the spectroscopically pure inert or noble gases which are employed. The system also was equipped with a McLeod gauge to read pressures.
The lamp was exhausted and baked at. 475 C. for about one hour. The cathodes were processed until no more gas was liberated. The water-jacket was put into place so that the Vycor sections of the jacket and lamp Vycor sections coincided and was firmly fastened to keep it in a xed position. The annular space between the lamp and jacket could be iiushed With water at any desired temperature, which was read on a thermometer placed in the Water surrounding the lamp. A voltage stabilizer was used to control the voltage input on the filament transformers, the discharge transformer, and the ultra-violet meter.
The ultra-violet meter containing a tantalum cell to read the 2537 A. U. radiation was set in position opposite the "Vycor section. The photovoltaic cells were placed section and the phosphor sections. They were checked before and after a set of readings with a standard incandescent lamp, also using a voltage stabilizer in the circuit. All stands were securely fastened so as to keep all positions xed during a run.
The lamp was seasoned for several days before any readings were taken. This seasoning was done in mercury vapor alone, continuously exhausting during the entire period. During operation of the lam water was ushed up and down the tube until the water temperature was 45 C. All the data were subsequently taken at this temperature. The temperature was checked before and after each current reading. As each current setting was made, the ultra-violet output of the arc, the voltage, the visible light output of the arc, and the light output of the phosphor were read in that order. These zero gas pressure data were again obtained as a reference check after each change of inert gas or inert gas mixtures. After all the preliminary data were obtained, then this lamp was' operated at different inert gas pressures and mixtures of inert gases. The inert gases investigated were krypton, nec-n and argon and mixtures of ltryptonargon and krypton-neon. These gas mixtures were varied so as to obtain suilcient data to show any difference in characteristics and each one was diffused at least 16 hours to insure uniform mixing.
The following mixtures were used:
Argon Krypton Percent Percent All the data presented in the accompanying composition curves are for a constant current of 500 milliamperes and a constant temperature of 45 C. for different pressure parameters in millimeters.
Voltage characteristics For krypton-argon mixtures the lowest voltage is obtained with 2 millimeters gas pressure throughout the entire range of compositions from 100% krypton to 100% argon. Since the voltage varies directly as the ionization potentials of the inert gases, krypton has a lower voltage than argon. Also there is a more or less linear relation with Krypton-argon mixtures between 100% krypton and 100% argon. There appears to be a slight deviation from linearity as 100% krypton is approached above 1 mm. gas pressure. At 4 mm. gas pressure, voltage deviations from linearity are found at both pure krypton and pure argon.
In krypton-neon gas mixtures, no such voltage linearity is observed between pure krypton and pure neon, but instead the curves are concave upward for all pressures and are more or less parallel for 2, 3 and 4 mm. pressure. The voltage again is lowest at 2 mm. gas pressure for composition between 100% and 40% krypton. Beyond this composition to 100% neon, the curves cross and the lowest voltage is obtained with 1 mm. gas pressure.
The voltage in the voltage curves is highest with neon. This result coincides with the fact that the ionization potential of the inert gas is higher for neon than for either argon or krypton. Table ll, page 5 of paper entitled Electrical Characteristics of Low Pressure Discharges, by John W. Marden and the present joint inventor Meister, presented before the 33rd Annual Convention of the Illuminating Engineering Society, San Francisco, Calif. Aug. 2l-25, 1939, gives the drop in a tube .85 cm. in diameter with about 39.3 cm. arc length, as 228 volts in neon-mercury, 185 volts in argon-mercury, and 143 volts in Krypton-mercury, the gas pressures being 2 mm. and the current 70 milliamperes in each instance. The foregoing is a confirmation of the findings in that paper, to this extent. This means, that taking the voltage in Krypton-mercury as 100, the voltage in argon-mercury is 129 and that in neon-mercury 159. This is a further confirmation of the previous statement that the voltage varies directly as the ionization potentials of the inert gases, said potentials being 13.9 volts for krypton, 15.7 volts for argon, and 21.5 volts for neon.
Arc characteristics The arc efliciency-composition curves of the discharge at constant current and temperature for krypton-argon mixtures show a minimum in the visible light output at about 50% krypton-50% argon, and the curves are more or less parallelr for all pressures from l to 4 mm. However, while the voltage in the krypton-argon was lowest at 2 mm. gas pressure, it is seen that the visible light eiliciency is highest for l mm. gas pressure for all compositions between pure krypton and pure argon. Since it is known that at constant current and constant temperature, and at a given pressure, the voltage for krypton is lower than argon, it follows that the efiiciency must be higher for pure krypton than for pure argon, as was observed.
W ith Krypton-neon mixtures no such minimum is found,
l but again the visible lighteiciency is greatest for 1 mm. gas pressure from 100% krypton to 20% krypton-80% argon. Beyond this point there is a steep decline in the arc eiciency, so that at 100% neon the arc efficiency is the lowest for all the pressures shown up to 4 mm. gas pressure.
Ultra-violet characteristics The ultra-violet efficiency-composition curves at constant current and temperature for Krypton-argon mixtures, Figure 2, show a slight maximum at approximately 50% Krypton-50% argon at 2 mm. pressure, but this maximum seems to shift slightly at higher pressures.
The highest ultra-violet eiciency is obtained at approximately 2 mm. gas pressure for all compositions between pure krypton and pure argon. The ultra-violet elftciency is higher for pure krypton than for pure argon at all pressures. it is further to be noted that the ultra-violet etiiciency decreases rather rapidly towards pure argon, while only a slight decrease is observed towards pure lrrypton, so that at 100% krypton the ultra-violet eiciencies at 2, 3 and 4 mm. gas pressures are relatively close. The etliciencies at 1 and 4 mm. are substantially identical between pure argon and 50% Krypton-50% argon. In this 'instance only a curve for 5 mm. pressure is also plotted.
In Krypton-neon mixtures, Figure 3, this maximum is more pronounced and occurs at about krypton-25% neon and at all four gas pressures. Again the ultraviolet etficiency is highest at approximately 2 mm. gas pressure for all compositions between pure krypton and pure neon. The ultra-violet eciency for pure krypton is greater than for pure neon at all pressures.
It is seen that the curves for 2, 3 and 4 mm. pressures cross as they approach pure krypton, where the highest ultra-violet-eficiency appears to be at 3 mm. The actual eiciencies are very close for this narrow pressure range.
Fluorescent characteristics The uorescent output eiiiciency-composition curves at constant current and temperature for krypton-argon mixtures, Figure 4, show a maximum efficiency at about 50% krypton-50% argon at 2 mm. pressure. The highest uorescent eiiiciency is obtained at 2 mm. for all compositions between pure krypton and pure argon. This uorescent efficiency at 2 mm. pressure coincides reasonably Well with the ultra-violet efficiency curve at 2 mm. pressure. It was found that at this composition, where the fluorescent eiciency is a maximum, the visible light efliciency of the arc is a minimum.
The uorescent output efficiency-composition curves for Krypton-neon mixtures, `Figure 5, have their maximum at approximately 75 krypton-25% neon. These curves again agree reasonably well at the maximum point with the ultra-violet-eiiiciency composition curves.
It is further to be noted that while the maxima in the ultra-violet and uorescent efciency for these inert gas mixtures all occur at about 2 mm. gas pressure, the output eiciency of the arc for the krypton-argon and Krypton-neon gas mixtures is greatest at 1 mm. pressure. The only exception found was below 20% krypton-80% neon where the 1 mm. curve crosses the 2 mm. curve so that the are, ultra-violet and fluorescent efficiencies from this composition to neon occur in the vicinity of 2 mm. gas pressure. Thus the visible light output efficiency of the arc occurs at 1 mm. gas pressure, while the ultra-violet and corersponding uorescent efficiency occur at 2 mm. pressure, exactly where the voltage of the experimental lamp is a minimum.
These data indicate an optimum composition for inert gas mixtures where the ultra-violet production, mainly as mercury resonance radiation, is the most efficient.
An explanation for this result may be the presence of metastable atoms. It has been shown that the output of 2537 A. U. in a mercury-rare-gas discharge depends upon the concentration of metastable inert gas atoms. Output curves for the inert gases have their maximum at about the same pressure as that at which the mean life of the metastable atoms is a maximum. It is concluded that the collisions between metastable inert gas atoms and mercury atoms must be considerably greater than collisions between two inert gas atoms which cause the destruction oi' metastable states.
The experimental results were obtained using separately heated cathodes by which part of the cathode losses were supplied, but not counted against efficiency. Therefore, in practical lamps, in which the cathode losses are included, the eiciencies of the gas mixtures are less than the experimental curve shown by an amount depending on the percentage of argon-or neon added to krypton. However, the reduction is not great enough to overcome the increased eiciencies of the optimum mixtures, and efciencies comparable to krypton lamps results.
The following results of tests demonstrate a significant improvement in low temperature operation. These lamps with the improved gas filling have considerably lower bulb Wall and ambient temperatures at which striation disappears.
Striation Comparison-Elapsed time to produce steady light Ambient Temperature Striation persists indefinitely with Kr.
From this table it can be seen the improved gas filling eliminates more than 70% of the lamp warm-up period needed with regular krypton lamps before the light output stops iickering and becomes steady.
So far, the matter of voltage, arc, ultraviolet or fluorescent output only, have been discussed. We now come to the bearing that voltage has on fluorescent light output. A consideration of the voltages shows that they increase from pure krypton toward the pure argon or the pure neon end of the range, thereby effecting a corresponding increase in total output as the percentage of argon or neon is increased, even though the eiiciency is not at a maximum.
Since relative eiiiciencies are presented in the curves they had to be calculated from some wattage or voltampere value because the efliciency is the total output divided by the input in volt-amperes. All these tests as disclosed were made at 500 milliamperes current. Therefore, in order to calculate the eiciencies, some voltage Values had to be used. As stated in the speciication under Voltage characteristics the kryptonargon mixtures have voltages intermediate between 100% krypton and 100% argon and are more or less linear (and the voltage varies directly as the ionization potential of the inert gases, then as stated krypton has a lower voltage than argon or neon and neon has the highest voltage of those inert gases). Therefore, in the case of kryptonargon mixtures, the intermediate compositions have voltages in between those of the pure gases. Thus it can be seen that argon-rich mixtures will have the higher 30 between 40 e 5 mixture.
,5 ambient and bulb wall temperature.
113 mixtures, pure neon will have a relatively higher total output than pure krypton or argon. Here also a 30% Krypton-70% neon gas mixture has the same eiciency as pure krypton and therefore definitely would have a higher total output because the voltage at this composition is higher than for pure krypton.
From the relative ultraviolet and fluorescent efficiency curves it can thus be seen that the gas mixtures containing about 8090% argon or 80-90% neon when mixed with krypton will have a higher total output than pure krypton and will have an e'iciency approximately 95% as high as pure krypton.
Summary In this investigation of the krypton-argon and kryptonneon gas mixtures in the presence of mercury vapor at low pressure (about 9 microns, corresponding to a wall temperature of about C., which is believed to be the most eiicient operatingtemperature, althought approximately the same efficiency is obtained at temperatures and C., the corresponding mercury pressures being between 6 and 13 microns), it was found that maximum ultra-violet and fluorescent output efiieiency occurred at 2 mm. pressure for a 50% krypton- 50% argon mixture, and for a 75% krypton-25% neon This indicates that for best results, the proportion should range between Ll555% krypton and 55- 45% argon, or between 7080% krypton and 30*20% neon, with the pressure about 2 or between 2 and 3 mm. for an operating temperature near or at 45 C.; that is, be-
n tween 40 and 50 C.
Commercial discharge lamps, one of which is illustrated in Figure l, corroborate these experimental data. In addition they show the decided advantage over pure krypton lamps of operating without striations at a lower The lamp of said figure comprises an elongated translucent vitreous en.- velope ll, with heated iilamentary electrodes l2 and i3, one in each end portion, and containing the selected noble gas mixture and some mercury, indicated by the .j globule ltd. If a iluorescent lamp, the selected phosphor l5 is applied to the inner surface of the envelope.
Although the experimental lamp described contained a 3500" white phosphor composed of zinc beryllium silicate and magnesium tungstate, the phosphor for commercial use may be any one which eiciently uses and,
tions in the region of 2537 A. U., that is, resonance radiation, and consequently a violet response giving a good light output.
the mercury strong ultra- Another ex- CO ample of phosphors which may be employed are the halo 5 ciency and maintenace of lamps containing the preferred krypton-neon mixture with that of identical lamps containing krypton.
O Hours 100 Hours 300 Hours Gas Fill Watts L. E
Kr-25 Ne-Hg 25. 8 l, 453 Krypton-Hg 25. 0 1 363 www L. P. W. Watts L. L. P. W. Watts L. Li P. W.
Gas Fill Watts Lumens L. l?. W
Krypton-Hg 24.7 1, 296 52. 6 Kr-A-Hg 25. 4 l, 470 57. 8
If the lamp is used for bactericidal, or other purposes where ultra-violet light is necessary, the envelope should, of course, be formed of hightransmitting ultra-violet glass such as Vycor, Corex, those having Corning code Nos. 9740, 9741, 972, or other well-known glass for such purpose. Thus, the word translucent as here used is generic to ultra-violet transmitting, as will oe obvious. Likewise, the words discharge lamp are generic to uorescent lamp.
Although preferred embodiments of our invention have been disclosed, it will be understood that modifications may be made within the spirit and scope of the appended claims.
1. A discharge lamp comprising an elongated translucent vitreous envelope, an electrode in each end portion of said envelope, and a contained mixture of krypton and argon admixed With mercury vapor, the proportion of the krypton being between 45% and 55%, and that of the argon being between 55% and 45% 0i the gas mixture.
2. A discharge lamp for the generation of ultra-violet radiations comprising an elongated envelope of ultraviolet transmitting vitreous material, an electrode in each end portion of said envelope, and a contained mixture of krypton and argon admixed with mercury vapor, the proportion of the krypton being between 45% and 55% and that oi the argon being between 55% and 45% of the gas mixture.
3. A discharge lamp comprising an elongated translucent vitreous envelope, an electrode in each end portion of said envelope, and a contained mixture of krypton and argon at a pressure between 2 and 3 mm. of mercur admixed with mercury vapor, the proportion of the krypton being between 45% and 55%, and oi the argon being between 55% and 45% of the gas mixture.
References Cited in the file of this patent UNiTED STATES PATENTS 1,726,107 Hertz Aug. 27, 1929 1,929,369 Found Oct. 3, 1933 2,177,710 Gordon Oct. 31, 1939 2,182,732 Meyer Dec. 5, 1939 2,228,327 Spanner Jan. 14, 1941 2,355,258 Biggs Aug. 7, 1944 2,363,531 Johnson Nov, 28, 1944V 2,425,697 Hultgren Aug. l2, 1947 2,473,642 Found June 2i, 1949r