US2922065A - Incandescent lamp - Google Patents

Description

Jan. 19, 1960 G. MEISTER ErAL INCANDESCENT LAMP 4 Sheets-Sheet 1 Filed Jan. 20. 1956 idd All

a /oa zoo .soa

574/60 T2571? l/V Jan. 19, .1960 G. MEISTER ErAL INCANDESCENT LAMP Filed Ja'n. 2o. 195e 4 Sheets-Sheet 2 avra/26E 4E/6715.2 ...i

Jan. 19, 1960 G. MEISTER ETAL 2,922,055

INCANDESCENT LAMP Filed Jan. 20, 1956 4 sheets-sheet 4 T zgr 14- 15. 'Egjf United States Patent Oi INCANDESCENT LAMP George Meister, Newark, and Nicholas F. Cerulli, Caldwell, NJ., assignors to Westinghouse Electric Corporation, East Pittsburgh, Pa., a corporation of Pennsyl- Vania Application January 20, 1956, Serial No. 560,441

12 Claims. (Cl. 313-116) This invention relates to incandescent' lamps and, more particularly, to diffusing coatings for incandescent lamp envelopes and to a process for applying a diffusing coating to an incandescent lamp envelope and is a continuation-iu-part of application Serial No. 444,316, filed July 19, 1954, and now abandoned, titled Incandescent Lamp With Light Diffusing Coating and Method of Manufacture by the co-inventors herein.

Heretofore commercially-available incandescent lamps with a nely-divided, light-diffusing envelope coating have had a silica coating applied to the lamp envelope by methods as outlined in U.S. Patent No. 2,545,896 to Pipkin, by flushing processes, or as outlined in U.S. Patent No. 2,661,438 to Shand. In the hush-coating processes of the prior art, finely-divided silicon dioxide (silica) is suspended in a volatile solvent such as butyl acetate With a binder such as nitrocellulose to impart the desired coating viscosity. In such flush-coating processes the silica must be maintained substantially free from moisture, or such moisture will react with the nitrocellulose binder which is water insoluble and deleteriou'sly affect the resulting coating. The process for removing moisture from finely-divided silica before flush coating has entailed baking the silica at relatively high temperatures, for example, 825 C., or higher, and, as will hereinafter be explained, such baking of the silica deleteriously affects the maintenance of the finished or processed silica-coated lamps.

In the silica-coated lamps prepared by the process of burning organo-silicates to form a fume or smoke, as disclosed by Pipkin in bis patent, the resulting silica formed by the burning is quite inert with regard to moisture-repossessing characteristics. Further, the cost of organo-silicates is relatively high and the cost of commercially-available, iinely-divided silica is roughly one quarter that of the organo-silicate, which lower cost favors the flush-coating methods of the prior art which can use commercially-available silica. However, the solvents which are used in the llush method, butyl acetate, for example, are relatively expensive, which somewhat decreases lthe cost advantage realized through using a commercially-available silica in a flush-coating process.

In the silica-coated lamps as prepared by the process of spraying onto a heated bulb an alkaline-reacting silica aquasol carrying large silica particles, as disclosed in the Shand patent, the silica coating is relatively inert to moisture-repossessing ability, which apparently is attributable either to the structure of the silica which results from the method of preparing the silica gel or to the method of applying the coating to the bub. Silica aquasols are also relatively expensive as compared to commerciallyavailable silica since considerable processing is required, and in addition, silica aquasols containing some large silica particles must be used in relatively large amounts on the incandescent lamp envelope in order to achieve `adequate light dilfusion.

Lastly, the burned organo-silicate coatings cannot readily be applied to va clear-glass bulb and still provide a light diffusion which is equivalent to a burned organosilicate coating on an inside-frost type bulb. The same is true of the flush-applied, silica coatings of the prior art. lIf these prior-art silica coatings are applied to a clear glass bulb, the resulting coating must be relatively heavy, resulting in decreased coating transmission ehiciency, i.e. the light absorbed by thecoating is excessive.

While some of the prior-art silica coatings have a better coating light-transmission eiiiciency than others, all of the silica coatings of the prior art have a coated-envelope light-transmission eliiciency which is inferior to the envelop-e light-transmission eiciency for an insidefrost bulb.

It is the general object of this invention to provide a finely-divided, light-dilusng coating for an incandescent lamp envelope, which will result in improved performance for the completed lamp.

It is another object of this invention to provide a silicacoated incandescent lamp which has improved lumen maintenance.

It. is yet another object to provide an incandescent lamp having a clear-glass envelope coated With a linelydivided, light-scattering material, which coating will eiect filament coverage with a minimum of light absorption.

It is a further object to provide an incandescent lamp having a clear-glass envelope coated with a looselypacked silica, which silica coating will effect filament coverage with a minimum of light absorption.

It is a still further object to provide an incandescent lamp having a clear-glass envelope coated with a looselypacked silica, which silica coating will effect lament coverage with a minimum of light absorption and which coating will improve the lamp lumen-maintenance.

It is still another object to provide an incandescent lamp with a silica coating which is a very eiective light scatterer.

The aforesaid objects of the invention, and other objects Which will become apparent as the description proceeds, are achieved by providing a finely-divided, lightscattering material coating for an incandescent lamp envelope. If the coating material is silica, the light may be more effectively ditused by virtue of moisture which may be included in the silica, and the lumen maintenance ofthe lamp may be improved by rendering the silica a moisture getter. lf the coau'ng material is loosely packed and coated onto clear glass, the filament is covered or hidden with a minimum of light absorption by virtue of the eiective light-scattering properties of the coating material.

For a better understanding of the invention, reference should be had to the accompanying drawings wherein:

Fig. 1 is a `graph of silica moisture content vs. silica temperature for samples of silica powder cooled from various tiring temperatures;

Fig. 2 is a graph representing silica molsture loss vs. tiring temperature for a silica powder; t

Fig. 3 represents moisture repossesslng abilltles for Ivarious types of silica substances;

Fig. 4 represents moisture gain under 100% relative humidity conditions for samples of silica powders initially tired at Various temperatures;

Fig. 5 illustrates a silica-coated incandescent lamp;

Fig. 6 represents a lirst step in the electrostatic coating process;

Fig. 7 represents a coating-material vsmoke generator;

Fig. 8 illustrates the ,coating operation for electro- .statically applying -the coating materialV to the lamp envelope; h

Fig. 9 illustrates the bulb-lehring operation following 'the coating operation;

Fig. 1 represents the Ysealing-in operation following bulb lehring;

Fig. 1.1 is a lgraph representing observed brightness'for various types of silica-coated envelopes vs. distance from neck to topofbulb, i.e., the candle power distribution for silica-,coatedenvelopes;

Fig.V 12 ,is a` graph of relative light absorption vs. silica moisture content for a silica-coated, inside-frost incandescent `lamp envelope;

VFig. 131is aksketch of'an operating lamp having a clearn glass `envelope coated with silica, which silica hasV been ethyl orthosilicate, which silica lhas been deposited thereafter by an electrostatic process. I

Commercially-available silica is normally prepared by precipitating silica from a silicate by means of an acid, for example, by precipitating silica from sodium silicate by means of lhydrochloric acid. Such silica is substantially White, porous, generally amorphous and normally inherently spherical in configuration as far as the ultimate particles are concerned. By the descriptive term, Vgenerally amorphous, it is meant that X-ray diffraction patterns do not show sharply-defined lines. Also, the porous nature of the silica is another way of stating that the ultimate particles are loosely packed.

The precipitated silica will normally have a relatively high moisture content which may vary from 6 to 15% by weight, for example. However, the commerciallyavailable silica should possess or have at least 1.7% by weight of moisture when heated to 200 C. in order to provide an improved lumen maintenance for the processed lamp. Whether the moisture in the silica as received is absorbed or adsorbed is Vnot definite, but it is probable that the moisture is possessed by the silica as both absorbed and adsorbed moisture. Of course, if the silica has been baked or iired during processing, the moisture content vmay vary considerably. The following discussion of the moisture-possessing and repossessing properties'of silica is based on test data performed on silica which has not been baked or iired by the commercial supplier vduring initial Aprocessing or preparation orin any other Way subjected to a moisture-removing process which renders the silica relatively inert with respect to moisture repossession, hereinafter explained.

Extensive tests have been conducted on commerciallyavailable silica which has a moisture content of about 12% by weight, which 1,2% represents an average moisture content for such silicas. In order to determine the total moisture content of the finely-divided silica, as received, a sample is accurately weighed, then fired at about 1000 C. until no more weight loss is observed. The 1000 C. tired sample is then reweighed before it can regainl any moisture from the atmosphere. The weight difference in the silica before and after firing at 1000 C. represents the moisture content of the original sample and'may be called the loss on ignition, as it is known in the art.

The moisture-possessing and repossessing characteristics of silica are very unusual and extensive tests have been conducted on these characteristics, the results of which tests are graphically represented in Figs. 1, 2, 3 and 4, wherein moisture content is plotted vs. the silica temperature with the initial silica liringtemperatures also being indicated. -Infconductin-g the-tests which provided the .accumulated dataas represented in the curves of Figs. 1-4, .silica samples were heated to predetermined temperatures, forexample,` 'to 900 C., until a constant weight for the 900 C. red sample was obtained. This tiring temperatureof 900 C.;is represented at point (a) in Fig. `-l. 'The -samplev was V-weighed after ring and before it could `acquire `orf-repossess jvany moisture from the atmosphere. The sample was then allowed to cool slowly under normal room conditions of temperature and humidity (25 C. vand 30-50% relative humidity) and a gradual weightlgain vi/asiobservedV as the sample slowly cooled and acquired moisture" from the atmosphere. When thejsample reached room'ftemperature the test was stopped. The results of this 900 vC. testare illustrated in the lowermost curve (curve vA)' of Fig. 1 and, as shown, the 900 C. liiredsainp'le regained a total weight of only 1.15%. Similar tests'were conducted at .tiring temperaturesrof 625 C., '515 C., 335 Ci, 230 C., 155 l100 C. and 65 C. Y. In reach of these tests the silica" powders were lr'st maintained at the designated firing temperatures until no urtherf weight losses from the samples were observed. The samples were then allowed to cool slowly at normal room conditions of temperature and humidity until room temperature was attained, whereupon the tests were stopped. The results of these tests are also graphically represented in Fig. l by the curves I65B, 14C, D En MF, G99 and HH.

An analysis of the curves represented in Fig. l indicates that the higher the tiring temperature for silica powder, the les's Vthe moisture which will be resorbed as the silica cools down to room temperatures. Also, a silica powder which has been fired until no additional weight loss is observed (i.e., when steady-state conditions are obtained) has the ability to repossess only a certain limited amount of moisture. For example, a 5l5j C.vred powder on being cooled torroom temperature can repossess about 2.0% by weight of moisture. If the same 515 C. tired powder were co-oled only to C. in the absence of moisture, it would have the abilityto repossess 0.65% by weight of moisture. Referring to powders which are red at other than 515 C., assuming that the iired powders are maintained under substantially moisture-free conditions after `firing, the moisture-repossessing abilities at 110 C. are as follows: 0.40% by weight for 900 C. fired powder, 0.5% by weight for 625 C. red powder, 0.7% by Weight for 335 C. tired powder, 0.65% by weight for 230 C. fired powder and`0.4% by weight for C. -lired powder. It can thus be seen that as silica powder is tired at higher and higher temperatures it tends to lose it'smoisture-repossessing characteristics and when cooled to 110 C. the moisture repossession approaches a maximumfor silicapowders viired at from about 230 C. to .515 C.

There is villustrated in Fig. 2 acurve showing moisture `loss vs. tiring temperatures Vfor silicav powders which originally possessed approximately 12% Aby weight of moisture. `Ordinates on this curve were determined by abbauen pe D iring powders at the temperatures as indicated until no more weight loss was observed (i.e. when steady-state conditions were reached). As illustrated, the moisture loss generally follows three straight lines, from room temperature up to 200 C. tiring temperature it is theorized that the loss is primarily adsorbed moisture. From 200 C. to 825 C. it is theorized that the moisture loss is primarily absorbed moisture, and above approximately 825 C. there is substantially no moisture remaining in the silica.

The foregoing illustrations of the moisture-repossessing characteristics of silica may be brietly summarized by noting that when a silica, baked or otherwise exposed to a moisture-removing process, in accordance with the teachings of this invention, is maintained under substantially moisture-free conditions, such silica will have an ability to repossess additional moisture, and indeed will act as a moisture getter. At higher ring temperatures, however, the silica loses its moisture gettering characteristics.

Other materials which have been used to coat lamps to form diiusing coatings do not exhibit these moisturerepossessing characteristics to such a degree. For example, there are illustrated in Fig. 3 curves showing the vmoisture-repossessing abilities of sodium silicate, silica which is formed by burning ethyl orthosilicate, a silica .aquasol containing large silica particles, and for purposes of comparison, a 900 C. tired silica powder, a 625 C. tired silica powder and a 500 C. tired silica powder. As illustrated in curve l of Fig. 3 when heated to 500 C. and then allowed to cool to room temperature under normal room conditions, silica formed by burning ethyl silicate is relatively inert with regard to repossession of moisture, since such silica gains only 0.6% by weight of moisture. Sodium silicate when fired at 50.0 C. and cooled under the same conditions repossessed only 1.05 of weight by moisture, as illustrated in curve I of Fig. 3. An alkaline-reacting silica aquasol containing colloidal silica admixed with a limited amount of larger silica particles for purposes of light diffusion, and prepared as outlined in the aforementioned Shand patent, when red at 500 C. and allowed to cool to room temperature, will repossess only 1.2% by weight of moisture, as illustrated in curve K of Fig. 3. A 900 C. fired silica powder, when later tired at 500 C. and allowed to cool to room temperature will repossess only 1.05% by weight of moisture as shown in curve l of Fig. 3, i.e., it displays the same characteristics as sodium silicate. A 500 C. tired silica powder (curve L) and a 625 C. fired silica powder (curve M) when rered at 500 C. and allowed to cool at room temperature will respectively gain 2.0% and 1.3% by weight of moisture.

The inertness of the silica formed by burning ethyl silicate is attributed to the very high ame temperature of ethyl silicate (1320 kilocalories per mole liberated to form SiO2, CO2 and H2O) and it has been shown that the higher the temperature to which silica is exposed the more inert with regard to moisture repossession it becomes. The inertness of sodium silicate with regard to moisture-repossession ability is attributed to the fact that sodium silicate is an entirely different compound from silica, and cannot be expected to display the same physical properties. The relative inertness with respect to moisture repossession of the admixture of alkaline-reacting silica aquasol and large silica particles is attributed either to the silica-gel structure, or to the method of applying the coating wherein the gel has an apparent tendency to frit itself when sprayed onto a hot bulb, as outlined in the aforementioned Shand patent. The relative inertness of the 900 C. fired powder with respect to moisture repossession is attributed to a basic structural change in the highly-tired, finely-.divided silica powder, which structural change is not definitely understood. Even when a 900 C. red silica powder is allowed to remain at normal room temperatures and humidity for periods of months, it will repossess very little moisture.

As a further illustration of the moisture-repossessing characteristics of finely-divided silica, various tired samples were allowed to cool to room temperature while maintained under substantially moisture-free conditions. These tired samples were then exposed to humidity for two hours and the percent moisture gain in weight was measured. This percent moisture gain in weight is plotted vs. initial tiring temperatures in Fig. 4 and, as illustrated, a 900 C. red powder gains only about 0.7% by weight and the C. tired powder gains 14.5% by weight of moisture. It should be noted that on tiring silica, substantially all moisture is driven off at about 825 C. and tiring `at higher temperatures does not result in any further moisture loss. However, tiring at temperatures in excess of 825 C. does render the silica more inert to repossession of moisture and this is illustrated in Fig. 4 where the 1000 C. red powder is shown to be more inert toward moisture repossession than the 825 C. tired powder.

It is deemed proper to note that ever since the first days of the incandescent lamp, engineers have resorted to all conceivable means and mechanisms to remove all possible moisture from the finished lamp. This is because any moisture present in the finished lamp tends to set up the well-known, so-called water cycle with the tungsten filament during lamp operation. In this water cycle, the moisture reacts with the hot tungsten lament to form tungsten oxides and release atomic hydrogen. The tungsten oxides deposit on the envelope surface. The atomic hydrogen reacts with oxygen present in the deposited tungsten oxides to form a black tungsten deposit on the envelope and more water vapor, and so on, until the envelope is quite blackened and relatively opaque. This problem has existed with acid-etched bulbs, as well as silica-coated bulbs.

It has been found that byV using a silica coating in accordance with the teachings of this invention, there is provided a diusing coating which also acts as a moisture getter and which will inhibit the heretoforementioned water cycle. Thus, the lumen maintenance under normal-operating conditions for the silica-coated lamps of this invention is measurably improved over the lumen-maintenance under normal-operating conditions for all lamps of the prior art. The prior art has taught that all possible moisture should be removed from the processed lamp in order to produce the best possible normal-operation lumen-maintenance. In this case, however, removal of all moisture from the silica impairs its moisture-gettering action.

Silicacoated lamps have been prepared under the same conditions except that some of the lamps were coated with 900 C. red silica powder and other lamps were coated with 600 C. red silica powder, both powder rings being prior to the envelope coating step. Of course, after firing at the designated temperatures and during lamp processing, the red powders were maintained as moisture-free `as possible. The 600 C. fired silica coated lamps had a 70% normal-life lumen-maintenance which was appreciably greater than the 70% normal-life lumen-maintenance of the lamps coated with the 900 C. fired powder. It is significant to note that all lamps which were coated with the two powders were prepared under the same controlled conditions and the maintenance improvement can thus be attributed solely to the fact that the powders were processed by tiring at different temperatures before being coated on the l-amp envelope.

In explanation of the term 70% normal-life lumenmaintenance, which term is well-known through the lamp art, it is generally accepted that the lumen output of a lamp, when measured at 70% of its normal life, is an accurate indication of the performance which is to be expected throughout the life of the lamp. In control testsV for determining the 70% normal-life lumen-maintenance, the initial lumen output of the lamp is corrected for any variations in actual lamp life from the rated lamp life, as is customary in the lamp art.

The foregoing tests, and all normal-life tests herein referred to, were conducted on 100 watt lamps burned in a base-upward position, which is the usual service operating position for 100 watt lamps. The bulb temperatures for such lamps when burned in such ya position vary considerably from one portion of the bulb toanother, but the minimum bulb temperature as measured with a preheated thermocouple is approximately 110 C. In the following discussion the normal-operation minimum envelope temperatures .will be considered, since the silica coating at the coolest portion of the lamp envelope has the greatest moisture-repossessing ability, which moisturerepossessing `ability results in the improved normal-life lumen-maintenance, as will be further discussed.

` Referring now to the curves illustrated in Fig. 1, where the 900 C. tired powder is coated onto a bulb it `must necessarily be subjected to temperatures during lamp processing which kare suiciently below the deformation temperature of the soft glass envelope so that the envelope will not be damaged during processing. Assuming the coated lamp envelope is baked or lehred at about 450 C. during processing (this lehring temperature was employed in processing the lamps), a 900 C. fired powder will have the opportunity to gain or repossess approximately 0.15% by weight of moisture (e.g., note the diierences in values of moisture content between the ordinates of points (a) `and (b) on the curve A of Fig. 1, wherein -a 900 C. red powder on cooling to 450 C. can acquire about 0.15% by weight of moisture). On further cooling to 110 C. a 900 C. fired powder will have the Iability to Iaccumulate or repossess an additional 0.25% by weight of moisture (ie. the difference in ordinate values of points (b) and (c) on the 900 C. iired powder curve A). A 600 C. red powder, in contrast, on cooling from 450Y C. to 110 C. has the ability to `accumulate an additional 0.45% lby weight of moisture (e.g. note the ordinate diierences between points (d) and (e) obtained by extrapolation between the 625 C. and 515 C. tired powder curves). Since the lumen-maintenance for the 600 C. silica coated lamps was appreciably better than the lumen maintenance for the 900 C. silica coated lamps, which lamps were processed the same except for the silica powder, it would seem to follow ipso facto that the `addition-al moisture repossessing ability of the 600 C. silica, as compared to the 900 C. silica, is responsible for `the improved lamp lumen-maintenance characteristics.

It should be noted that the tenacity with which silica getters `and holds small amounts of moisture apparently increases greatly as the silica becomes more moisture hungry. Thus a relatively small increase in moisture gettering ability, expressed as a percent by weight, represents a very large increase in the tenacity with which silica getters and holds moisture.

It has been found that in order to have a 7 0% normallife lumen-maintenance which is appreciably better than the corresponding normal-life lumen-maintenance of prior art lamps, the silica coating must have a moisture content which is at least equivalent to the moisture content of a 625 C. red silica powder which has the ability to repossess an additional 0.4% by weight of moisture at minimum lamp envelope operating temperatures. The total moisture content for a 625 C. tired powder, before such powder is allowed to repossess any moisture is about 1.4% by weight. At 450 C., under normal room conditions, the total moisture content for this 625 C. tired powder will be about 1.55% by weight and if the coated envelope were lehred during lamp processing to about 450 C., under normal-room conditions, the total moisture content for this 625 C. ired powder would thus `be about 1.55 by weight. Assuming the lamp is tirped- 8 oi while thel moisture content of 625 C. silica is mailitained at not greater than 1.55% the silica will still possess an ability to accumulate an additional 0.4% by weight of moisture at 110 C., which is the normal-operation minimum envelopevtemperature for a 100 watt lamp, and which 0.4%` represents the moisture gettering abilities of the coated 625 C. tired powder when processed. into a iinished lamp. When higher red silicas are processed into a nished lamp their potential moisture gettering abilities are so limited as to minimize any maintenance improvement which may be realized through the so-called silica moisture gettering action. Since this moisture gettering action results in improved 70% normal-life lumen-maintenance, the silica coating must have at least 1.55% by weight of moisture and the ability to repossess an additional 0.4% by weight of .will thus show an increased moisture at normal-operation minimum envelope temperatures. This establishes a lower range for moisture content in the silica coating and the minimum moisture gettering ability which the silica coating must possess.

At the upper permissible moisture content limitation for the silica coating it would seem possible to use any silica which has been subjected just prior to lamp tip-olf `to temperatures which are reasonably in'excess of the normal-operation minimum lamp envelope temperature of C. For example, the lamp could be baked just prior to exhaust and tip-oil at a temperature of 200 C., which would result in a moisture' gettering ability for the silica coating of approximately 0.5% by weight at normal-operation minimum envelope temperatures (this ligure is obtained by extrapolation between the 230 C. and C. curves designated E and F in Fig. 1). This of course does not take into account the operation of lamps in enclosed or recessed type fixtures 'where the minimum jlamp envelope temperature may be as high as 225 C. If a coated 200 C. tired powder, having a 'moisture content of approximately 4.8% was exposed lto a temperature of about 225 C., as in a recessed lixture, the coating of 200 C. Itired silica would have the ability to give joi approximately 0.3% by weight of moisture, resulting in some lamp-blackening and decreased lumen maintenance for such applications. However, silica-coated lamps are now sold at a premium price and are normally intended to be used in ixtures where their esthetic, even appearance will be visible. Thus silica-coated lamps having a moisture content as high as 4.8% will still be Very acceptable for normal operation Where the minimum envelope temperature is 110 C., provided such silica also has the ability to repossess at least 0.4% by weight of additional moisture. Such lamps normal-life lumenmaintenance over the silica coated lamps of the prior art and this normal-life lumen-maintenance improvement will more than olset any increased blackening encountered in recessed or other hot-iixture applications.

It can thus be stated that in order to show an appreciable improvement in 70% normal-life lumen-maintenance, the silica coating must possess at least 1.55 by weight of moisture and not more than 4.8% by weight of moisture, and in addition must have the ability to possess at least 0.4% by weight of additional moisture at normal-operation minimum lamp envelope temperatures.

lf it is desired to process the lamp so that its performance in special-type recessed or other hot fixtures will be at least as good as the performance of the insideffrost lamps of the prior art, it is necessary to limit the moisture content of the silica coating to not more than 4.0% by weight, or otherwise expressed, the moisture content of the silica coating should be equivalent to the moisture content of about a 30.1 C. fired silica (extrapolating between curves D and E of Fig. l). Of course the'silica should still have the ability to repossms at least 0.4% by weight of additional moisture at normaloperation minimum lamp envelope temperatures if TABLE I 70% normal life maintenance Lamp type (indicated as a percentage of the initial lumens per watt) Inside-frost type bulb 93A 8 Silica-coated bulb (formed by burning ethyl orthosilicate) 93. 7 Silica-coated bulb (900 C fired powder coated by ush 92. 7 Silica-coated bulb (silica aquasol-large silica particle mixture sprayed on hot bulb) 92. Silica-coated bulb (450c C. fired powder) 95. 2

An electrostatic process is preferable for coating onto a bulb finely-divided silica which possesses limited and controlled amounts of moisture and an apparatus for electrostatically coating silica onto an envelope and further processing the coated envelope is illustrated in Figs. 6-9.

As heretofore noted, silica possessing substantial amounts of moisture cannot be coated onto a bulb by the ush methods of the prior art since the organic binders which are necessary to impart the desired viscosity to the coating composition are not water soluble and are deleteriously affected by any appreciable amounts of water in the silica. 'Ihus where flush methods are used to coat lamps, 825 C. to l000 C. fired powder should be used to substantially eliminate the moisture possessed by the silica.

ln Fig. is illustrated a silica-coated incandescent lamp 20 comprising a vitreous, light-transmitting envelope 22 carrying an internal coating of moisture-containing, ne- A ly-divided silica 24 and having a mount sealed to the neck thereof. A brass or aluminum screw-type base 26 is cemented to the neck to facilitate connection to a power source, as is usual. As is well known, the mount comprises a vitreous re-entrant stem press 28 having lead-in conductors 30 and 32 sealed there-through and supporting a refractory metal filament 34, such as tungsten, between their inwardly extending extremities. The envelope preferably contains inert gases such as nitrogen, argon, krypton, etc., or mixtures thereof, as is well-known, although the lamp may be a vacuum type, if desired.

In electrostatically applying the diffusing silica coating or any other finely-divided, light scattering coating to the inner surface of the unsealed bulb, the open-necked bulb is rst placed under and supported by a hollow lava chuck 36, as illustrated in Fig. 6, which chuck cooperates with the bulb cullet 38 and bulb neck 40 to support the bulb. While thus supported, the bulb is rotated either manually or by a belt drive (drive unit not shown) and heated by gas-air burners 42 to approximately 100 C. Because of the negative temperature coeicient of electrical resistance of glass, this heating renders the envelope substantially uniformly electrically-conductive. The heating temperature of 100 C. is given only by way of example and not by way of limitation since the temperature to which the glass is heated to render it substantially uniformly electrically-conductive is not particularly critical and may be varied considerably, according to the type of glass being heated, for example, temperature extremes of 70 C. to 300 C. have been used, although these temperatures are not intended to be limiting. It should be noted that most incandescent lamp bulbs are fabricated of the wellknown lime glass.

. As illustrated, the insulating lava chuck 36 of the electrostatic coating apparatus is aixed to a collar 43,'which l0 is insulated from the bulb by the chuck, to allow the bulli' to be rotated readily, either manually or automatically, during the heating and later steps of the process.

There is illustrated in Fig. 7 a smoke generator unit 44 for producing a smoke of nely-divided particles suspended in air, prior to electrostatic deposition of the powder. The smoke generator comprises generally a powder and smoke reservoir 46 having an outlet 48 at the bottom thereof through which the finely-divided material is admitted into a mixing venturi 50. Compressed air is admitted to the venturi through a pressure-regulating valve 52 and thence to the venturi where the nely-divided material is picked up and carried through a feed conduit 54 to cause the air-particle mixture to impinge upon a target 56 to break-up agglomerato/s which might have formed and to disperse thoroughly the coating material to form a smoke of finely-divided particles suspended in the air vehicle.

The powder before being placed in the smoke generator unit must be timely-divided and may be ground in an air-velocity type grinder such as marketed under the trademark Micronizer by Sturtevant Mill Co., Boston, Mass., or as marketed under the trademark Wheeler Mill by C. H. Wheeler Mfg. Co., Philadelphia, Pa. This: breaks up the overly-large particle agglomerates.

For the specific embodiment of the particle smoke nozzle, which will hereinafter be illustrated and described, it is preferable to maintain a particle-smoke pressure within the reservoir 46 between 6 and l2 pounds during coating to cause the particle-smoke to pass through the smoke nozzle at a desirable Velocity. To maintain the smoke pressure in the reservoir within these aforementioned preferred pressure limitations, an indicating gauge 58 is provided from which the operator may have a visual indication so that the pressure-regulating valve 52 may be manually adjusted to maintain the smoke pressure in the reservoir within the aforementioned preferred pressure limitations. Such pressure-indicating and pressureregulating valves are well-known. A smoke-nozzle conduit 60 connects the s-moke reservoir with the injector nozzle assembly 62, as shown in Fig. 8. To control the ilow of smoke to the nozzle assembly, a manually-operable butterily valve 64 is provided in the conduit 60.

The air compressor 66 which supplies air to the smokegenerator unit is preferably regulable between 2 lbs. and 20 lbs. output pressure and an air dryer 68, such as a well-known, aluminum-oxide type air dryer is provided in the output line of the compressor so that the particle powder may be maintained under substantially Watervapor-free conditions until it is forced into the uncoated bulb. A powder-driven agitator 70 (power source not shown) is provided near the base of the reservoir to agitato continually the nely-divided coating material to keep it in a finely-divided state by breaking up the largest particle agglomerates.

The particle powder which is introduced into the reservoir may be a commercially-available grade of silica, for example, which is reasonably pure. Many other finely-divided materials have also been successfully coated. While silica has been found to be the best from a light-scattering aspect, as hereinafter explained, the following materials have been found to be suitable for deposition by the electrostatic process as herein illustrated and described: alkaline-earth and magnesium titanates, stannates, zirconates, oxides, carbonates and silicates; alkaline-earth sulphates; titania; zirconia; zinc oxide; alumina; talcs; sodium or calcium alumino-silicates (zeolite); and zirconium silicate. Of course the foregoing materials should be timely-divided (preferably from .02 to slightly more than one micron average diameter ultimate-particle size) to effectively scatter the light and these particles should appear substantially white under reflected light in order not to absorb excessive amounts of light. The foregoing list of materials is by no means inclusive, but is only illustrative of the multitude of ma.

terials which may be electrostatically deposited. It should vbe noted that as the particles decrease in average, ultimate-particle size, some bluish tint can be seen in relected light from the coated material, i.e., in the coated larnp when it is unburned. This is apparently due to scattering similar to Rayleigh scattering, to which the blue inthe sky is attributed, V,although the coated material under reected light only .displays a very slight amount of this Rayleigh scattering and cannot be compared to the sky in color. It should Valso be'noted that any finely-divided, substantially-White material normally owes its white color to the fact that the material is actually transparent and the white color is attributed to the Ylight-scattening properties of the tine state of division. An example ,0f this is pure ice, which is transparent, and snow, or nely divided iCe.

ln order to control better the moisture content of commercially-available silica, if silica is to be coated, theV powder may be baked, if desired, before coating, although the moisture possessed by the silica may becarefully controlled by lehring or baking the coated bulbs kafter the coating operation and before sealing in and/ or exhaust. The only limitation to powder baking before coating is that the baking temperature should notexceed 625 C. so as to impair the potential moisture-gettering properties of the silica as processed into the lamps, as heretofore explaind.

in fig. 8 is shown v the coating operation for the bulb. rthe positive pole 72 of a high-tension, direct-.current source lis electrically connected to the gas-burner unit 42 and the negative pole 74.is electrically connected to a Vprolle 76 which projects through the hollow-lava chuck 36 into the lower extremities of the bulb neck. if desired these Vpolarities may be reversed with but little effect on the resultant coating. The magnitude of the applied DC. voltage is not particularly critical and may Vary between about 8 kv. and V25 kv., for-example, a specific example rbeing l kv.

The particle smoke injector nozzle assembly 62 is circumferentially disposed about the probe 7,6, and the nozzle assembly connects with the nozzle conduit 66 of the smoke generator. The air which is present in the bulb and the particle smoke which does not deposit on the bulb wall during coating passes through a return conduit 78 which is disposed about the nozzle assembly 62 and the nozzle conduit 60 and which discharges into a collecting hopper (not shown) so as to collect the uncoated particles for reprocessing and further use. A conduit support collar 30 supports the probeand nozzleconduit assemblies and may be positioned longitudinally with respect to the lava chuck and bulb neck either manually or automatically.

lt has been found that the particle smoke must be forced through the nozzles of the nozzle assembly in order to coat properly for when the nozzle assembly is removed and the particle smoke is passed through the nozzle conduit into the heated -bulb an imperfect coating results. lt is thus theorized Vthat the particles when passing through the nozzles are frictionally charged and the application of the high voltage merely attracts the particles carrying an opposite charge to the bulb Wall where the charge on the particles is dropped. In snpport of this, tests were conducted with silica where the applied uni-directionm potential was varied between 8 kv. and 25 kv. with no observed diierence in the amount of silica which was deposited'(i.e., the coating weight on the bulb wall) at the extremes of the 8 kv. to 251cv. applied voltage. Also, in the case of silica, for example, only about of the powder injected through the nozzle into the bulb is actually coated, the other 50% passing through the return conduit 7 S.

When the silica smoke, forV example, is forced through the nozzles, the particles carry a small net-negative charge i.e. there are more negatively-charged particles ythan positively-charged particles.` This is perhaps due to the stainless steel of which the nozzle is fabricated and apparently this is responsible for a slightly better powder deposition on the bulb when the bulb Wallis made positive `with respect to the probe, although veny satisfactory coatings can be obtained when the bulb wall is made negative with respect to the probe.

Charging of the particles through static friction before application of a high voltage to direct the charged particles to the bulb wall represents a general departure in mechanism from that mechanism which is generally accepted as customary in the electrical-precipitation art. Vin explanation of this, in the usualelectrical precipitation the particles are charged by means of gaseous ions ot electrons as opposed to a static-frictional charging through turbulence. The charged particles are then transported to the collecting electrode by the force exerted on the particles by the electric field, and the charged particles are then discharged at the collecting electrode. Thus the electrostatic deposition of this invention apparently differs from ythe conventional electrical precipitation in the particle charging means wherein substantially all particles within the field are deposited.

As a specic example for silica coating Ya bulb designed for a w. lamp, thenozzle-injector assembly has eight, evenly spaced nozzles, circumferentially disposed about the probe and each having a diameter of 46 mils. As heretofore noted, the preferred pressure in the smokegenerator may vary between 6 and l2 pounds. ln coating a bulb adapted for 100 watt operation, the butterfly valve 64'may be opened for about 2 seconds while applying a high Vt'ension'DC. of l5 kv. between the bulb wall and the probe. This will deposit approximately 50 mg. of silica onto the bulb. ln coating bulb sizes other than the size 'adapted for 100 watt operation, the number of nozzles which may be used may vary depending on the bulb size. Alternatively, fewer nozzles having a larger diameter or more nozzles having a smaller diameter may be used in the nozzle assembly to coat identical bulbs in order to achieve the same coating result. The nozzle assembly thus constitutes a diffusing orifice which projects charged particles into the heated bulb.

After being coated the bulb is baked or lehred while rotated on the lava chuck, as illustrated in Fig. 9. Bulb lehring is necessary to drive off moisture which may have accumulated during coating and in the case of a silica coating, lehring is necessary in order to render the silica coating as moisture possessive or acquisitive as possible. The lehring may be accomplished by gas-air burn ers 42, as illustrated, and the lehring temperatures may vary considerably depending on the prior processing of the silica coating powder and the conditions under which the processed lamp is intended to operate. For example, if a silica-powder is red before the coating operation at a temperature of about 500 C. for a suicient time to approach steady-State conditions with regard to moisture content, a bulb lehr of 350 C. for a period of l0 to 20 seconds will normally be suilicientfor the silica coating to have suicient affinity for moisture to provide an improved lumen-maintenance at normal-operation minimum envelope temperatures, provided the mount is sealed-in, lamp exhausted, gas-fill inserted and exhaust tube tipped-off while the bulb is still hot, thus preventing the silica coating from repossessing appreciable amounts of moisture from the atmosphere between the coating, sealing-in and tipping-olf operations.

In order to insure adequate` moisture-free conditions for the silica coating, particularly where it is desired to operate under high temperature conditions, it is desirable to lehr the coated bulb at from 400 C. to 500 C. for about l0 to20 seconds and even at this lehring temperature range it is desirable to simultaneously flush the coated lamp with hot, dry air, or other gas at a temperature of about 250 C., for example, to carry away all possible moisture. The air ush temperature is not para 13 ticularly critical and may vary from about 150 C. to the lehring temperature of the bulb.

If the silica powder has not-been baked or iired before the coating operation, higher lehring temperatures preferably are used while simultaneously flushing the bulb with hot dry air. For example, a bake or lehr of 450 C. for about 15 seconds is not considered excessive where a relatively moist, unti'red, silica coating must be activated tol impart thereto adequate moisture gettering ability for normal lamp operation.

Immediately/,following the lehring operation and while the. bulb is still hot, the mount is sealed in as illustrated in Fig. l0. It is desirable to flush the bulb with hot, dry nitrogen, or other inert gas, while sealing the mountrto the bulb neck in lorder to remove any moisture which may accumulate from the sealing res, which are normally provided by gas-air burners, as is usual. Such hot, dry-nitrogen flushing is preferablyaccomplished through the exhaust tube 80 in order to maintain a slight pressure within the bulb to force any moisture out of the neck.

Immediately following the sealing-in operation, and while the bulb portion of the envelope is still hot, the lamp is exhausted through theexhaust tube, the gas-lill inserted and the exhaust tube tipped-off, as is customary. It may be desirable to further bake the bulb on exhaust to insure that all possible moisture is removed to give the silica coating all possible moisture gettering ability. Baking on exhaust is not absolutely necessary, but is desirable, particularly where the processed lamp is to be operated in hot-recessed or other high-temperature-type fixtures.

After tipping-off, the lamp base is cemented to the neck and the lead-in conductors connected by well-known lamp basing techniques (basing operation not shown).

The main purpose of a lamp envelope diffusing coating is to diiuse and soften the light emitted by the incandescent iialment. Thus the performance of any ditusing coating can be evaluated by the amount of diffusion eected as compared to the percentage of light which is transmitted through the coated, light-diiusing envelope. There are illustrated in Fig. ll curves showing the observed brightness in candles per sq. cm. vs. distance from the neck of the bulb to the top measured in a plane perpendicular to the iilament. All laments used in these measurements were type CC6. lt should be noted that the observed brightness measurements represented in Fig. ll were integrated over a circular area of the lamp surface approximately 0.1 in diameter (roughly 0.008 sq.

in.). lt is obvious that by increasing the integrated area, hot spots will be smoothed out in the observed brightness measurements, or vice-versa, by decreasing the integrated area, hot spots can be intensified. Thus whenever brightness measurements, as illustrated in Fig. 11, are to be evaluated, the area over which such measurements are integrated should be indicated in order that the observed data may be given the proper evaluation. The curves of Fig. ll are identified as follows. Curve N represents a standard acid-etch, inside-frost bulb. Curve O represents an electrostatically-deposited silica coating on an acid-etched, inside-frost bulb. Curve P represents an electrostatically deposited silica coating on a clear-glass bulb. Curve Q represents a burned ethyl orthosilicate silica coating on an acid-etched, inside-frost bulb. Curve R represents a silica coating on an acidetched, inside-frost bulb, which coating is applied by spraying onto a hot bulb a silica aquasol containing large particles of silica. Curve S vrepresents a silica coating on an acid-etched, inside-frost bulb, which coating is applied by ushing a 900 C. tired silica powder onto the bulb.

As observed, the general shape of all of these curves, with the exception of the inside-frost bulb which is shown for purposes of comparison, is substantially similar and the maximum observed brightness for each type of coated lamp is given in the following table, designated Table II:

' and the results are given in the following table.

It will be observed that the maximum brightness for the electrostatically-applied silica coatings, both on clear and inside-frost type bulbs, and the maximum brightness for the burned ethyl orthosilicate coatings when applied to inside-frost bulbs are equivalent. The maximum brightness for the flush-coated lamps is comparatively greater, i.e., the diffusion effected by such coatings is somewhat less. The maximum brightness for the silica aquasol-large silica particle coatings on inside-frost bulbs approaches that of the electrostatically-applied silica coatings'.

As previously noted, the light-diiusion efficiency for a lamp envelope coating must be measured by the relative transmission eiciency as well as by the actual diffusion eiected by the coating. It is obvious that a very heavy silica coating will be an excellent ditusing means, although the coating transmission eciency may be suiciently low as to render such coatings relatively poor (i.e., excessively absorptive of light). Thus, in order to evaluate further the foregoing types of coatings, the light-transmission eciency for these coatings were tested In conducting the coatingtransmission-efficiency tests, opennecked bulbs were placed over a standard light source in a photometry sphere. A sensitive, linear-responsive photocell, shielded from direct radiation from the stand- Y ard source, indicated the transmitted-light intensity, which indication is relative to the intensity of the standard source. For example, the standard light source was energized within the sphere and the photocell output noted. This reading represents the value. The bulb whose transmission etliciency was to be measured was placed over the standard light source and the photocell output measurement noted. This photocell measurement was then corrected to indicate a percentage reading as compared to the intensity of the standard light source. The transmission eciencies for the heretofore-discussed various types of envelope diffusing coatings were as indicated below in Table III. Also indicated for purposes of comparison are the transmission eiiciencies for clearglass bulbs and inside-frost bulbs.

TABLE llI Transmission Lamp bulb type Eiiciency in Percent Standard light source 100 Clear-glass bulb 99 Acid-etch, inside-frost bulb 98. 9

97. 2 Burned ethyl orthosilicate on inside-frost bulb 95. 3 Silica aquasol-large silica particles sprayed on hot inside-frost bulb 95. 5 900 C. red silica powder flushed on inside-frost bulb 96. 7

It will be observed that as compared to a standard acidetched, inside-frost bulb, the electrostatically-applied silica coatings have a transmission efficiency which is 1.7% lower. An equivalent burned ethyl orthosilicate coating on an inside-frost bulb has a transmission etliciency which is 2.6% lower than an inside-frost type bulb. An equivalent silica aquasol-large silica particle,

hot-spot characteristics of the electrostatically-appliedsilica is attributable to either the low bulb density of the silica coating, as hereinafter elaborated upon, or the moisture possessed by the silica coating, which increases the light scattering and thus decreases the hot spot. Regarding the affect of moisture on light scattering, when the silica coating is substantially water saturated, by in- Jecting steam into the coated envelope, for example, the silica coating becomes relatively transparent and thus non-diffusing, since the gas interfaces between the individual particles are substantially filled with water. It was thus completely unexpected that a small quantity of moisture in the silica coating would increase the lightscattering properties of the coating; To measure this phenomenon a light source and a photocell were placed on opposite sides of a silica-coated lamp, in which test set-up the photocell reading indicated the hot spot or diffusing characteristics of the envelope coating, as integrated overa fairly large area. With the lamp between the light source and the photocell, the lamp 'envelope was broken at the base so that the silica would getter the moisture from the air (normal room temperature'and 35% humidity). Upon breaking the envelope, the photo'- cell reading rapidly decreased from 46 to 38' (arbitrary light-intensity units) as the silica gettered the moisture from the air. The differential in photocell readingv before and after the lamp isv opened depends, of course, on the weight of the coating material, i.e., the heavier the coating, the smaller the initial reading, but the percentage change in photocell reading is normally about as represented in the foregoing specific example. Since moisture is substantially transparent, the logical conclusion to be drawn from the foregoing is that the moisture increases the scattering properties of the silica, thereby decreasing the light transmitted directly through the lamp and thus decreasing the photocell reading. Further tests are represented in Fig. 12, wherein photocell reading is plotted vs. moisture content in the silica coating. As illustrated, when the moisture within the silica coating varied from about 3 to about 9%, the photocell reading varied from about 46 to about 38 arbitrary light units. Similar tests conducted on flush-coated silica (900 C. fired material) and on burned ethyl orthosilicate-applied silica showed substantially no change in photocell reading, further illustrating the inertness of' these substantially moisture-free materials with regard to moisture gettering. Thus, the

affect of comparativelylarge amounts of .moisture in the.

silica coating is to require less material for an equivalent hot-spot, assuming all other conditions are the same. In addition, this oers one, or a partial, explanation for the improved diffusion-transmission characteristics of the silica-coated lamps of this invention.

The electrostatically-applied silica may also be coated on clear-glass envelopes where the coating material will effect filament coverage with a minimum of light absorption. In explanation of the terms herein employed, light transmission, or conversely, light absorption is that amount of light which is transmitted or absorbed in passingV through the coated envelope, asconveniently measured with a photometry sphere, hereinbefore explained. A hot spot is that which the ey'e sees as a round glow in thecenter of the lamp, as best illustrated'in the usual well-known,V acid-etch, inside-frost type ofV bulb. Filament covering is defined as the ability to hide the .iilament per, se. For example, in an inside-frost type of lamp thelameut is not distinguishable as such, but the round glow or hot spot is quite intense. In contrast, when burned ethyl orthosilicate, flushed silica, or titania, for example, are coated onto clear glass, the hot spot is relatively small, but the filament is normally quite distinguishable, :that is, there is insuflcient filament coverage.

The well-known, acid-etch, inside frost weakens the glass envelopes, resulting in considerable bulb breakage, both in handling and in use. This objectionable acid etch may be dispensed with by electrostatically coating the silica onto a clear-glass envelope, while obtaining a relatively small hot spot, adequate filament coverage, and still maintaining a transmission efiiciency better than the prior-art, silica-coated lamps. These improvements are realized through the use of a loosely-packed, silica coating, as will be elaborated upon hereinafter.

`In Fig. 13 is illustrated a sketch of aV clear-glass envelope coated by burning ethyl orthosilicate. As shown, the hot spot is minimized, but the filament shines through clearly. In Fig. 14 is illustrated a sketch of a clear-glass envelope coated with titania. Here again there is only a very small hot spot, but also very little filament coverage since the filament shines through clearly. In Fig. 15 is illustrated a sketch of an inside-frost envelope with no additional coating. As shown, the filament itself is covered, that is, the inside frost does effect filament coverage, but there is an intense hot spot. In Fig. 16 is illustrated a sketch of a clear-glass envelope coated with loosely-packed silica `in accordance with this invention. As shown, the filament is covered and the hot spot is minimized. It should be noted that the coated envelopes illustrated in Figs. 13, 14 and 16 all have a light absorption substantially as given in Table III, with the light v absorption for the titania-coated envelope being slightly fio' greater than that for the silica-coated envelopes. In Figs. 13 and 14, the filament is shown as a negative of its actual appearance, sinceV in a sketch it is impossible to represent the contrast displayed by the filament shining throug the coated envelope.

As one explanation of the improvement realized from a loosely-packed coating, filament coverage is primarily attributed to light scattering. The formula for light scattering is:

Noria- For a full treatment of the above formula, see

I National Bureau of Standards Applied Mathematics Series A4.

A denserl packing of the coating material apparently increases the eective particle size (r) to substantially more than the wavelength of light, thus deleteriously affecting the scattering to which filament coverage is attributed.

Another explanation of the effective filament coverage of the loosely-packed coatings is that a low density' or loose packing provides more particle-gas interfaces which increase the light scattering. Whatever the explanation, a material which is coated on a clear-glass envelope to a low density, as specified hereinafter, will effectively cover the filament without excessive loss in transmitted light.

As noted before, the average ultimate-particle size does not appear to enter into the phenomenon of filament coverage to any measurable degree, since finely-divided, averag'e ultimate-particle sizes of from 0.02 to slightly over 1 micron have been explored, with no measurable effect with regard to altering the filament-coverage properties ofthe coated rrra'terial.` For best results with regard to obtaining an even appearance, however, the particles 117 should fall within the foregoing range and preferably should lie toward the upper end of the range, i.e., toward the larger particles which approximate the wave-length of light.

Following is a table, designated Table IV, listing coating bulk-densities for various diierent materials, as coated onto a 100 watt-size envelope, In obtaining these readings, lamps were broken under carefully controlled conditions (room temperature and lil-12% humidity) so that results would be consistent.

TABLE IV [Extremes as indicated are due to mean variations from lamp to lamp] N'rE.Silica true density used in calculation was 2.2 gms/c1113; titania true density, 4.2 gms/cm3.

In the following table, designated Table V, are given the minimum thicknesses for the coating material encountered in the lamp testing as used in determining the densities given in Table IV.

TABLE V [Extremes as indicated are due to mean variations from] amp to lamp Coating Weight Coating Thickness (gms/cm.2 times 5) (in microns) Material Max. Mn. Aver. Max. Min. Aver.

Silica-electrostatically applied 60 22 36 90 35 57 Burnedethylorthosilicate. 29 25 27 28 20 24 Silica aquasol-admixed large silica particles. 130 105 118 37 33 35 Silica applied by flush technique on clear glass (very heavy coat-8% absorption) 82 7G 79 29 23 26 Silica applied by ush technique on inside- Y l frost glass (4.3% absorpi tion 59 55 57 23 10 21 Electrostatically applied silica which is later steamed 96 44 70 49 19 34 Titania (electrostatically applied as a very heavy coat with high absorption) 181 89 135 5l 20 40 The important figue given in the foregoing Tables IV and V is the density ratio, that is, the coating material bulk density divided by the true material density. This ratio eliminates the differences between true material densities, e.g., silica and titania, which have about a 2:1 difference in true material density which aliects the coating material bulk density a corresponding amount.

It is the coating material bulk density, otherwiseexpressed as a density ratio, which eiects superior filament covering on clear glass. For example, silica may be electrostatically deposited on a clear-glass envelope in order to eect filament covering with a minimum of light absorption. This silica-coated lamp may then be steamed, that is, the silica coating may be wetted down with a cloud of steam, which densities the silica coating greatly, see Table IV. The amount of silica in this lamp coating has not changed by virtue of being densied, but the bulk density of the coating material has almost quadrupled. This causes the lament to shine through the coating, similar to the burned ethyl orthosilicate-coated envelope shown in Fig. 13. Of course, lamps coated with a high density-ratio coating material can be made to hide the lament when coated onto clear glass. For example, flushed silica-coated lamps can have the silica applied sufliciently heavy to clear glass to hide the lament, but the light absorption in such lamps is usually about 8%. The same is true of the burned ethyl silicate, silica-coated lamps. Such a heavy coating, however, approaches the status of a decorative coating rather than a competitive product, since the electrostatically-deposited silica-coated lamps of this invention normally have less than half the light absorption of these other coatings, where tilament coverage is effected with a clear-glass envelope.

The preferred coating material density, in the case of silica, is from about 0.055 to 0.075. Above a coating density of about 0.094 (density ratio of 0.043), the total weight of the powder needed to cover the filament becomes relatively great, thus resulting in excessive light absorption. Very low density coatings have also been investigated, down to a silica-coating bulk density of .029 (density ratio of 0.013). At such a low density ratio, a very small amount of material will eiiect filament covering, but the lamp assumes the appearance of an insidefrost type, regarding the formation of a hot spot, even with larger amounts of coating material. Also, at coating weights which provide a somewhat satisfactory appearance, the coating material must be very thick, as far as coatings are concerned (in the order of i60-180 microns), and ilakes olf very easily. In order to preserve a good appearance for the coated clear-glass envelope it is preferable that the density ratio4 for the coating material be at least about 0.018, which in the case of silica represents a coating bulk density of about 0.043.

It has also been observed that to effect filament coverage with a minimum of light absorption not only must the material density ratio bebetween 0.043 and 0.018, but the .average coating material thickness must be atleast 35 microns, and in practice the average coating material thickness is preferably from 45 to 70 microns. At the minimum permissible average coating material thickness of 35 microns the coating material density ratio should be toward the upper end of the permissible coating-material density range to effect filament coverage. For example, a coating material thickness of 35 microns and a silica coating bulk density of 0.048 (density ratio 0.022) will normally not effectively cover the filament. With the thinner coatings, that is, those approaching the minimum 35 microns thickness, it has been found that the product of coating thickness, in microns, and the coating density ratio should be at least about 0.9. In other Words, if the coating thickness is 40 microns, the coating material density ratio should be at least 0.0225, in 0rder to eiectively cover the filament when using clear glass.

In practice it is usually desirable to coat a clear-glass envelope with a relatively small amount of coating material. In some cases, however, when coating to a density ratio within the prescribed range of 0.018 to 0.043, it may be desirable to increase the coating thicknessA as much as possible. This will result in increased light absorption, but will provide the lamp with a very even appearance, as in a decorative lamp. Of course, the advantages of very efficient light scattering because of the low density ratio of the coating are still realized, even though the coatingris quite thick. In experiments it has been found that the coating thickness may be as great as about microns. With such a relatively thick coating, however, there is some tendency for the coating to flake olf, under shock, especially under production conditions.

The best method for depositing light-scattering material on clear glass to this low bulk density is by an electrostatic process as outlined hereinbefore. If silica is used, the silica coating may or may not act as a moisture getter, depending on whether or not it has been fired to very high temperatures before deposition. Normally, of course, the silica coating will act as a getter, since firing the silica at very high temperatures before thepowder is coated not only decreases the lamp lumen maintenance, but is a very inconvenient and time consuming procedure.

Conversely, the silica coating may be a moisture getter, as hereinbefore defined, and yet not effect filament coverage when coated on clear glass to a weight whichY will absorb only a minimum of light. Such a coating may beY applied by an electrostatic process and then steamed to wet down and densify the coating. n

In order 'to achieve a low coating-bulk-density, a drycoating process is normally required. The electrostaticdeposition process represents such a dry-coating process wherein the vehicle or carrier for the light-scattering material is air, in contrast to a flush process wherein the vehicle is a volatile liquid, which causes the deposited material to mat down and density. Even when silica is deposited by fuming ethyl orthosilicate, the products of combustion are silica, carbon dioxide and water, which Water apparently serves to mat down and densify the silica coating. In illustration of this, ethyl orthosilicate wasburned and the silica collected, dried and electrostatically deposited onto clear glass. A sketch of an envelope coated with this electrostatically-deposited silica is shown in Fig. 17 and, as observed, this low-bulkdensity silica coating effects filament coverage, While the same material when burned directly onto a clear-glass envelope does not effect filament coverage, see Fig. 13.

While the electrostatic deposition process is preferred for coating material to a low-bulk-density, other dry deposition methods may be used to achieve ythe same effect, such as the centrifugal force method disclosed-in Patent No. 2,336,946 to Marden and Meister, one of the coinventors herein. The processv as disclosed in this Marden and Meister patent may be used to deposit silica, for example, to a low-bulk-density, which silica coating may or may not be a moisture'getter, depending on the processing of the silica-before deposition.

Finely-divided, light scattering materials other than silica can also be electrostatically deposited on normallytransparent bulbs toV a density ratio as specified to effect filament covering. Some 150 different finely-divided, light-scattering materials have been experimentally coated to date and the best found so far, other than silica, are some metallic oxides and carbonates.Y Of these, zinc oxide, magnesia and alumina most nearly approach silica in performance although Vsilica is the preferred coating "materiaL The main objection to most of the materials coated experimentally is that they have too high a coating bulk density and thus are not as effective in covering the filament when coating a normally-transparent envelope and titania is an example of this, see Fig. V14. Other materials which electrostatcally coat to a relatively high coating-density (and a relatively high coating-densityJatio) are the alkaline-earth zirconates, titanates and stannates. If these materials could be coated to a low density ratio, as specified, they too should effect filament coverage, -but this has not been experimentally achieved.

An additional advantage of the electrostatic deposition process is that of cost, as the cost of electrostatically applying a silica coating,l for example, to an incandescent lamp envelope is much less than the cost of applying silica Vcoatings, to incandescent lamp envelopes by the methods of the prior art. YFor example, the cost of ethyl orthosilicate is roughly four times as much as `thecost of commercially-available silica, and only about 30% of the organo-silicate is available as silica. Thus the c'ostrof silica per pound as procured from burning ethyl orthosilicate is roughly thirtecen times that of commercially-available silica. It should be noted that in the 20 electrostatic' process, one pound of silica will coat approximately 4,000, 10D-watt typeenvelopes and the silica which is not coated may be reclaimed, if desired.

Another advantage of the electrostatic coating process as compared ,to the processes of the prior art is the evenness of coating which is realized. When a silica coating, for example, is properly applied by the electrostatic process there will be substantially ino variations in the coating which are detectable by the nakedY eye. The silica aquasol-large silica particle coating when sprayed onto a hot bulb will produce some very large imperfections which are especially visible when the lamp is not lighted. The flush method and burned ethyl ortho-silicate method of applying a silica coating may produce large variations in thickness of coating. In the case of the flush methods, the vehicle has a tendency to run and to deposit the silica unevenly. In the case of the burned ethyl orthosilicate, the flame apparentlyrhas a tendency to be uneven, which produces corresponding uneven portions in the lamp coating.

It should be understood that all of the finely-divided coating materials of this invention are intended for light scattering and thus should possess such physical and chemical characteristics that their light-scattering properties are not deleteriously affected by lamp manufacturing schedules or lamp operating conditions.

It will be recognized that'the objects of the invention have been achieved by the provision of a finelydivided, light-scattering material coating for an incandescent lamp envelope,rwhich coating will result in improved performance for the completed lamp. In the case of a silica coating, the normal-life lumen-maintenance for the completed lamp may be improved and the material may be coated onto normally-transparent bulbs to effect filament coverage with a minimum of light absorption. Materials other than silica will also elit'ect this filament coverage with a minimum of light absorption.

As a possible alternative embodiment, the electrostatically-applied silica may be steamed to increase adherence. Such steamed silica coatings will not effect filament coverage'on normally-transparent glass without excessive light absorption, but the silica may still act as a getter. In this case, a slightly longer and hotter bulb lehr is required to render the silica moisture hungry, eg., a 475 C. lehr for about 15 seconds. Also, the electrostatically-applied silica may be neck steamed, that is, only the neck of the envelope may be steamed to increase adherence of the silica coating for the glass at this point, since imperfections arising from insufficient adherence most often occur at the neck of the lamp. In such a case, the neck-coated silica will be densified, while the Vrest of the envelope will be coated to a low bulk density to effect filament coverage and maximum diffusion with a minimum of light absorption. It should be understood, that when calculating the coating bulk density for such a neck-steamed embodiment, only that part of the envelope which is most effective in scattering the light, namely, all of the envelope excluding the steamed neck, should be used in calculating the coating bulk density.

While in accordance with the patent statutes, one best embodiment has been illustrated and described in detail, it is to be particularly-understood that the invention is not limited thereto or thereby.

We claim:

1. An incandescent lamp having a light-transmitting envelope and carrying on the internal surface of said envelope a coating of finely-divided, porous, generallyamorphous silica, said silica coating possessing from 1.55% to 4.8% by weight of moisture, and said silica coating at normal-operation minimium lamp envelope temperatures constituting a getter for at least 0.4% by weight of moisture.

2. An incandescent lamp having a normally-transparent envelope and a coating of finely-divided, substantially white material coated on the interior surface thereof, said material coating having an average coating thickness of 21 at least 35 microns, a coating material bulk density to true material density ratio of from 0.018 to 0.043, with the product of coating thickness in microns and the coating density ratio being at least about 0.9.

3. An incandescent lamp having a normally-transparent envelope and a coating of iinely-divided, generally-amorphous silica coated on the interior surface thereof, said silica coating having an average coating thickness of at least 35 microns, a coating material bulk density to true material density ratio of from 0.018 to 0.043, with the product of coating thickness in microns and the coating density ratio being7 at least about 0.9.

4. An incandescent lamp having a normally-transparent envelope and a coating of nely-divided, generallyamor phous silica coated on the interior surface thereof, said silica coating having an average coating thickness of about 4570 microns and a material bulk density of from 0.055 to 0.075 gms/cm.

5. An incandescent lamp having a normally-transparent envelope and a coating of finely-divided, generally-amorphous silica coated on the interior surface thereof, said silica coating having an average coating thickness of at least 35 microns, a coating material bulk density to true material density ratio of from 0.018 to` 0.043, with the product of coating thickness in microns and the coating density ratio being at least about 0.9, said silica coating possessing from 1.55% to 4.8% by Weight of moisture, and said silica coating at normal-operation minimum lamp envelope temperatures constituting a getter for at least 0.4% by weight of moisture.

6. An incandescent lamp having a normally-transparent envelope and a coating of nelydivided, generally-amorphous silica coated on the interior surface thereof, said silica coating having an average coating thickness of from 45-70 microns, a coating material bulk density of from 0.055 to 0.075 gms./cm.3, said silica coating possessing from 1.55 to 4.8% by weight of moisture, and said silica coating at normal-operating minimum lamp envelope temperatures constituting a getter for at least 0.4% by weight of moisture.

7. An incandescent lamp having a normally-transparent envelope and a coating of finely-divided, generally-amorphous silica coated on the interior surface thereof, said silica coating having an average coating thickness of at least 35 microns, a coating material bulk density to true material density ratio of from 0.018 to 0.043, with the product of coating thickness in microns and the coating density ratio being at least about 0.9, and said silca coating at normal-operation minimum lamp envelope tem peratures constituting a getter for at least 0.4% by weight of additional moisture.

8. An incandescent lamp having a light-transmitting envelope and carrying on the internal surface of said envelope a coating of finely-divided, porous, generallyamorphous silica, said silica coating having been baked at temperatures of at least 200 C. and not exceeding 625 C., and said silica coating at normal-operation minimum lamp envelope temperatures constituting a getter for at least 0.4% by'weight of additional moisture.

9. An incandescent lamp comprising in combination: sealed envelope means formed of light-transmitting material; visible-light-generating means enclosed by said envelope and comprising a tungsten filament adapted to be heated to incandescence by electrical energy; minute quantities of moisture enclosed by said envelope and tending to react with said lament when incandesced; and lightscattering means comprising finely-divided, porous, genorally-amorphous silica carried as a coating on the interior surface of said envelope and at normal-operation minimum lamp envelope temperatures constituting a get ter for at least 0.4% by weight of moisture; whereby the normal-life lumen maintenance of said lamp is improved.

10. An incandescent lamp comprising in combination: sealed envelope means formed of light-transmitting material and having alilling of inert gas; visible-light-generating means enclosed by said envelope and comprising a tungsten filament adapted to be heated to incandescence by electrical energy; minute quantities of moisture enclosed by said envelope and tending to react with said lament when incandesced; and light-scattering means comprising finely-divided, porous, generally-amorphous silica carried as a coating on the interior surface of said envelope and at normal-operation minimum lamp envelope temperatures constituting a getter for at least 0.4% by weight of moisture; whereby the normal-life lumen maintenance of said lamp is improved.

11. An incandescent lamp comprising in combination: sealed envelope means formed of light-transmitting material and having a filling of inert gas; visible-light-generating means enclosed by said envelope and comprising a tungsten filament adapted to be heated to incandescence by electrical energy; minute quantities of moisture enclosed by said envelope and tending to react with said filament when incandesced; and light-scattering means comprising finely-divided, porous, generally-amorphous silica possessing from 1.55% to 4.8% by weight of moisture carried as a coating on the interior surface of said envelope, and at normal-operation minimum lamp envelope temperatures constituting a getter for at least 0.4% by Weight of additional moisture; whereby the normal-life lumen maintenance of said lamp is improved.

12. An incandescent lamp comprising in combination: sealed envelope means formed of light-transmitting material and having a filling of inert gas; Visible-light-generating means enclosed by said envelope and comprising a tungsten lament adapted to be heated to incandescence by electrical energy; minute quantities of moisture en closed by said envelope and tending to react with said lament when incandesced; and light-scattering means comprising finely-divided, porous, generally-amorphous silica possessing from 1.55% to 4.0% by weight of moisture carried as a coating on the interior surface of said envelope, and at normal-operation minimum lamp envelope temperatures constituting a getter for at least 0.4% by Weight of additional moisture; whereby the normal-life lumen maintenance of said lamp is improved.

References @ted in the file of this patent UNITED STATES PATENTS 1,552,128 Ettinger et al. Sept. 1, 1925 1,698,845 Gustin Jan. 15, 1929 1,830,165 Gustin Nov. 3, 1931 2,006,850 Wein July 2, 1935 2,107,352 Teves et al. ..-u Feb. 8, 1938 2,426,016 Gustin et al Aug. 19, 1947 2,438,561 Kearsley Mar. 30, 1948 2,538,562 Gustin et al Jan. 16, 1951 2,545,896 Pipkin Mar. 20, 1951 2,626,874 Pipkin Jan. 27, 1953 2,661,438 Shand Dec. l, 1953 FOREIGN PATENTS 145,398 Australia Feb. 5, 1948

Claims (1)

1. AN INCANDESCENT LAMP HAVING A LIGHT-TRANSMITTING ENVELOPE AND CARRYING ON THE INTERNAL SURFACE OF SAID ENVELOPE OF COATING OF FINELY-DIVIDED, POROUS, GENERALLYAMORPHOUS SILICA, SAID SILICA COATING POSSESSING FROM 1.55% TO 4.8% BY WEIGHT OF MOISTURE, AND SAID SILICA COATING AT NORMAL-OPERATION MINIMUM LAMP ENVELOPE TEMPERATURES CONSTITUTING A GETTER FOR AT LEAST 0.4% BY WEIGHT OF MOISTURE.

Images