Updated 23-XII-2018

(James) Arthur Graves

This article was written by fellow lamp engineer and collector Edward J. Covington, and originally appeared on his own website of biographical sketches of persons involved in the lamp industry. Following his passing in February 2017, and with kind permission of his family, Ed's words have been preserved here in the hope of maintaining access to his writings for the benefit of subsequent generations.

Art Graves was an engineer at GE's NELA Park headquarters in Cleveland, Ohio, and was famously responsible for the development of a remarkable new getter for incandescent lamps, based on Phosphorus Pentanitride - more commonly known as P3N5. Still today it remains the getter of choice among incandescent lampmakers worldwide. He was also responsible for the development of leadwires having improved mechanical properties, for the support of filaments and discharge tubes within lamps. A summary of incandescent lamp getter developments leading up to the P3N5 innovations is presented below.

Phosphoric Anhydride and Red Phosphorus
The use of chemicals in the incandescent lamp was found to be necessary from the very earliest days of manufacture in order to avoid detrimental interaction of residual gases with the filament, as well as to reduce the time required to exhaust the lamp. The situation was described by John W. Howell in 1925 in the following words3:
"The problem of lamp exhaustion has been with us from the beginning. The chief difficulty in exhausting a lamp is taking care of the moisture which adheres to the surface of the glass. The amount of this moisture is very considerable, and it can be liberated by heating the glass. As the glass is heated hotter and hotter, even up to the softening temperature of the glass, more and more moisture is liberated, so during or immediately before exhaustion the lamp must be heated to a temperature considerably higher than it will attain under any condition of use. In practice we heat the lamp as hot as practicable and to at least 300°C. After the moisture is driven from the glass, the problem is to get rid of it. During the first fifteen years of the history of the lamp the exhausting was done on mercury pumps which would not pump out this water vapor, so it was absorbed by phosphoric anhydride in a glass cup placed to bring the drier as close as possible to the lamp; but even then the absorbing action took place through the exhaust tube, which was of small diameter and about 2-1/2 in. long. This absorption took considerable time, and exhausting an ordinary lamp in this way took a half an hour. The modern way of getting rid of this water vapor and other residual gases is by chemicals. An Italian engineer - Malignani - discovered that phosphorus vaporized in the lamp under proper conditions precipitated all water vapor and all other gases remaining in small quantities in a lamp after it had been exhausted to a vacuum less than one millimeter pressure. He painted the inside of the exhaust tube with red phosphorus and exhausted the lamp on a fast mechanical pump. When the vacuum was under one millimeter the lamp was lighted to high incandescence and a blue glow appeared all through the bulb. The connection with the pump was then closed and the phosphorus heated, driving a lot of phosphorus vapor into the bulb while the bulb glow filled the bulb. The blue glow instantly disappeared and a good vacuum resulted. The exhaust tube was then sealed off from the bulb. By this method lamps were exhausted in about one minute, the lamps being thoroughly heated before they were put on the pump. In the present day practice the phosphorus is applied as a coating on the filament. The lamp is exhausted on a highly developed rotary pump to a pressure less than one-tenth of a millimeter of mercury. The lamps are well heated beforehand, but are not lighted up during exhaustion. After the lamps have been sealed off and based they are lighted up brightly. A blue glow appears in the lamp which immediately disappears, leaving a good vacuum of less than one-thousandth of a millimeter pressure of mercury."
The picture above (unfortunately no longer available) shows the exhaust arrangement used by Edison in 18801. The "drying" tube, which contained the phosphoric anhydride, can be seen located near the fork with the sealed lamps. At a later date the drying tube was moved closer to the lamps so that the water was removed in a shorter time. The process using phosphorus that was mentioned above by John Howell was patented (U.S. 537,693) by Arturo Malignani on Apr 16, 1895. A more extensive discussion of Malignani's development can be found in Howell's book4.

Phosphorus Pentanitride
Although one might assume that no new getters would be developed after so many years, it turned out that in the 1960s work was performed that resulted in an improved getter for both vacuum and gas-filled lamps. In his patent for the use of pure crystalline phosphorus pentanitride (P3N5), Art Graves5 provided a description of the prior art. It is instructive to quote some of that description here:
"For many years past it has been customary in the art to effect the clean up of residual gases in incandescent electric lamps by means of phosphorus. It has been customary to introduce the phosphorus within the lamp in such a position that upon incandescing of the filament, the phosphorus vaporizes or sublimes and reacts with and cleans up residual gases within the lamp envelope. In the case of gas-filled lamps, it is the practice to use the red phosphorus alone, whereas in the case of vacuum lamps, which are usually those of lower wattages such as 25 watts or less, it is customary to mix the red phosphorus with a material such as cryolite.

"In the case of incandescent lamps, it is customary to apply the getter composition directly to the filament, usually a helical coil or coiled-coil of tungsten wire...The lamp is then exhausted through a tubulature communicating with the interior of the envelope, filled with gas in the case of a gas-filled lamp and the tubulature sealed off.

"During the sealing-in and evacuation of the lamp it is customary to apply heat which raises the glass envelope to a temperature in the range of about 200 to 600°C. depending on the composition of the glass of the envelope, for the purpose of eliminating surface adsorbed and absorbed gases. As a result of the heating, the red phosphorus getter on the filament is subject to oxidation and vaporization at a temperature of around 200°C., the temperature of the filament being near or above that figure during sealing in. The loss in active getter material can result in variation in life and maintenance of incandescent electric lamps employing this type of getter.

"Moreover, variable conditions of atmospheric humidity can have an even more drastic effect upon the lamp quality due to deterioration of the red phosphorus and its ability to clean up the increased water vapor content in the lamp due to the presence of excess water vapor on the lamp parts and on the machine parts..."
Graves pointed out that during particularly humid weather a very large percentage of lamps could be rejected when the lamps were made with red phosphorus getter in the manner established in the past. The new getter was made from a commercial product, crystalline P3N5. Graves said:
"It can be concluded that the pure crystalline P3 N5 is clearly superior to red phosphorus because of the following advantages. It is not affected by the elevated temperatures used in lamp manufacture. It is inert and does not hydrolyze or react with the suspension vehicle. There are no fire hazards in lamp plants occasioned by dried getter. The number of rejections by the sensitive leak detector...due to getter deficiencies are greatly reduced which obviously results in substantial savings. The P3 N5 has less effect, if any, on the crystalline structure of the tungsten filaments. It avoids the occasional problem of lamp burn outs as a result of faulty flashing. Since presumably white phosphorus is one of the products of decomposition, it can be concluded that there is more residual gettering action, that is, continued gettering action during the life of the lamp. White phosphorus is effective as a getter at much lower temperatures than is red phosphorus."
Graves did comparative testing of P3N5 versus red phosphorus as well as no getter to determine relative amounts of residual gases that were present after processing. This was performed by means of a mass spectrometer analysis. The gases carbon dioxide, water, methane, carbon monoxide and hydrogen were measured. The amounts of these gases decreased in the following order: no getter, red phosphorus, P3N5.

Photomicrographs of 60-watt coils taken after flashing indicated a "superior microstructure of recrystallized tungsten filaments when P3N5 is used than when red phosphorus is. Graves also performed lamp burning testing in what are called "slumper cans", which simulate confined burning fixtures. These are essentially formed from a standard food can which has been cut fully open at one end, and which has a smaller opening at the opposite end which is affixed to the lamp holder. As a result the lamp is heated very greatly by reflections from the inner surface of the can, and by reducing the normal convective cooling of the bulb. At the end of 150 hours burning the lumen output of the P3N5 lamps exceeded the output of those lamps with red phosphorus and those without getter.

Improved P3N5 Processing
In U.S. Patent 3,679,285 Edward G. Zubler described a process for increasing the degree of thermal decomposition of P3N5 during processing6. This was accomplished by having the getter in intimate contact with a catalyzer. The catalyzer could be selected from the following metals: tungsten, molybdenum, rhenium, nickel, titanium, zirconium, hafnium, vanadium, niobium, tantalum, rhodium, palladium, osmium, iridium or platinum. Zubler found that tungsten was preferred in tungsten-halogen lamps.

The Element Phosphorus
The element phosphorus continues to be a valuable part of processing in the manufacture of lamps. As early as 1882 two patents apparently were granted in England2 - one to Van Cleve for the use of phosphorus pentachloride and one to Cherrill for the use of phosphorus. After about 87 years it was still possible to use that element for the improvement of the incandescent lamp.

  1. "Edison's Electric Light", Francis R. Upton, Scribner's Monthly, Feb 1880, pp.531-544.
  2. "Phosphorus in Incandescent Lamps", Reginald A. Fessenden, Electrical World and Engineer, Vol.38 No.12, Sep 21 1901, p.473.
  3. "Lamp Developments in America", John W. Howell, Electrical World, Vol.85 No.8, Feb 21 1925, p.395.
  4. "History of the Incandescent Lamp", John W. Howell & Henry Schroeder, publ. The Maqua Company, Schenectady, NY, 1927, pp.126-127.
  5. "Getter for Incandescent Lamps and Similar Devices", James Arthur Graves, U.S. Patent 3,475,072, Oct 28 1969; filed May 19 1967.
  6. "Gettering Processes for Electric Lamps and Similar Devices", Edward G. Zubler, U.S. Patent 3,679,285, Jul 25 1972; filed Dec 28 1970