Inventing 20th Century Lamps:
Bracketed information [xxx] does not appear on the label.
[SL10 - Section #2 introduction label]
Step 2: Invention
"Genius is ninety-nine percent perspiration and one percent inspiration"
Whatever the percentages, it's probably the same for inventors today as for Edison.
But circumstances have changed. Work is more often done in groups in large
laboratories; scientific training is more important; equipment is more expensive. See
some of the differences and similarities in six recent lighting inventions, displayed in the
cases behind this panel.
[CT3 - header panel: Tungsten Halogen]
Tungsten Halogen: working in a modern industrial laboratory
[L11: Information Label for Case Study #1 - Tungsten Halogen]
[G-15: Time line on counter in front of case]
"So just out of pure frustration I suddenly decided to make a really bad
Edward Zubler, former GE engineer, 1996
Zubler had been trying to get a better and better vacuum, so that only the halogen
gas would be in the lamp. But results just got worse. Letting in some air improved
things, leading to the conclusion that small amounts of oxygen and carbon somehow are
essential--a process that is still being studied. Efficacy: up to 32 lumens per watt
Zubler was one of several highly trained scientists and engineers who worked on a
process to improve the tungsten lamp by adding a halogen gas inside the bulb.
Teamwork is commonplace in a modern industrial laboratory--partly because it is easy to
call on different experts for different phases of the project, partly because the company
doesn't want to have all of the knowledge held by one person.
Elmer Fridrich, seen here in a photo (A) taken around 1980, began the experiments
that led to a successful visible-light lamp after participating in the development of a
quartz heat-lamp. (1,2, and 3)
Other members of the development team are seen in the 1959 newspaper clipping
(B): Frederick Mosby (seated), Alton Foote, Stanley Ackerman, & Edward Zubler
(standing, left to right).
[tungsten halogen information]
[L12 - combined information and credit label]
Halogen in the lamp cleans evaporated tungsten from the walls, as seen in the
Crystals of tungsten growing around the filament support (D), show how sensitive
the lamp is to contamination. Much of Mosby's work focused on reducing contaminants
so the tungsten would re-deposit more evenly on the filament.
- Experimental tungsten-halogen lamp, about 1955, [1996.0147.02] from Elmer G.
- Quartz heat-lamp, about 1959, [1997.0388.18] from General Electric Lighting Co.
- Experimental tungsten-halogen lamp, about 1955, [1996.0147.03] from Elmer G.
- Elmer G. Fridrich, about 1980, from Elmer G. Fridrich
- Tungsten halogen development team, 1959, from Elmer G. Fridrich, ©General
- Lamp before and after action of the halogen cycle, about 1960, from Edward Zubler,
- Crystal growth in a contaminated lamp, about 1960, from Edward Zubler,
[CT4 - header panel: Metal Halide]
Metal Halide: the value of scientific training
[L13: information label for Case Study #2 - Metal Halide]
[G-21: Time line on counter in front of case]
"I'd taken everything from topography to matrices, everything you could possibly
mention in the field of mathematics, because you need it in the field of physics."
Gilbert Reiling, former GE engineer, 1996
Gilbert Reiling (B), after working on several government projects at GE's research
laboratory in Schenectady, got permission to play with an old concept--using sodium to
balance the blue color of a mercury vapor lamp. His extensive mathematical training
allowed him to calculate that in the cooler area near the wall of the arc-tube, a halogen
(like iodine or bromine) would combine with the sodium and prevent it from attacking the
Reiling's initial patent application was denied on grounds that it was covered by a
1912 patent by Charles Steinmetz. This lamp (1) was made following Steinmetz's
instructions; Reiling then demonstrated to the patent examiner that his lamp was
It should be noted that similar work was being done at this time--apparently
Sylvania then brought out a lamp with scandium iodide that had even better color
characteristics. As with other metal halide lamps, convection currents tended to push
the arc upwards in the middle when the lamp was horizontal, so the tube was curved to
prevent the arc from hitting the wall. (2)
[metal halide information]
[L14 - credit label]
- Patent-demonstration mercury lamp, 1964, [1996.0084.03] from Gilbert H. Reiling
- Sylvania metal-halide lamp with curved arc-tube, about 1972, [1998.0005.24] from
OSRAM SYLVANIA Inc.
- U.S. patent #1,025,932 to Charles Steinmetz, 1912
- Gilbert H. Reiling demonstrating a metal-halide lamp, 1962, from Gilbert H. Reiling,
- Spectral readings from Reiling's report, 1960, from Gilbert H. Reiling,
- Experimental lamps from Reiling's report, 1960, from Gilbert H. Reiling,
- Color comparison between mercury-vapor and metal-halide, 1962, from Gilbert H.
Reiling, ©General Electric
[CT5 - header panel: High Pressure Sodium]
High Pressure Sodium: studying materials
[L15 - information label for Case Study #3 - High Pressure Sodium]
[G-29: Time line on counter in front of case]
"He had been fascinated by aluminum oxide."
William Louden, former GE engineer, speaking of Joseph Burke,
Knowing about materials can make a big difference. It was important to Edison as
he first tried different kinds of metals and then tried many varieties of carbon for his
filament. When Joseph Burke (A) came to GE's Research Lab in 1954, he began
researching the properties of ceramics, specifically aluminum oxide. This material turned
out to make a good container for highly reactive sodium and it could withstand high
pressure. At high pressure the yellow spectral lines of sodium widen, resulting in better
color than low-pressure sodium lamps. (See interactive unit, behind you to your left.)
Several people contributed further developments: Robert Coble added magnesium
to make the aluminum oxide 95 percent transparent (1); William Louden worked on the
end seals; Kurt Schmidt tried cesium, sodium and other metals as light sources (B). The
lamp was announced in 1962 as "Lucalox" (for translucent aluminum
oxide), though it wasn't until 1968 that a truly practical lamp was marketed.
The efficacy of high pressure sodium lamps is about 100 lumens per watt, but the
color is quite yellow. Increasing the pressure improves the color, but reduces the
efficacy to about 40 lumens per watt.
[L16 - combined information and credit label]
Other manufacturers quickly saw the advantages of the new light source and
introduced their own models. In the Sylvania experimental lamp, (3) note the heater
alongside the aluminum oxide tube. The Duro-Test example has two tubes (4); if one
goes out, the other will light.
- Polycrystalline-alumina arc-tube body, about 1970, [1996.3042.34.01] from Elmer
- GE 35w "Lucalox" lamp, about 1989, [1992.0553.11] from Lawrence Berkeley
- Sylvania experimental HPS lamp, about 1971, [1998.0005.21] from OSRAM
- Duro-Test dual arc-tube HPS lamp, 1996, [1997.0062.17] from Duro-Test Corp.
- Joseph E. Burke, 1958, from the Hall of Electrical History of the Schenectady
- Kurt Schmidt and William Louden, 1962, from William Louden, ©General
- Temperature gradients along an HPS arc-tube, 1968, from William Louden,
- Worker assembling HPS lamps, about 1969, from William Louden, ©General
- Make-shift fixture to demonstrate effect of reflected light on arc-tube, about 1964,
from William Louden, ©General Electric
[CT6 - header panel: Compact Fluorescent]
Compact Fluorescent: the challenge of manufacturing
[L17 - information label for Case Study #4 - Compact Fluorescent]
[G-48: Time line on counter in front of case]
"We didn't license that bridge weld - it was a very significant piece of
Steve Goldmacher, marketing division, Philips, 1996
The major "scientific" problem with the compact fluorescent lamp was finding
phosphors that would last long enough when exposed to much greater radiation density
in the smaller tube. There also were engineering problems associated with production:
how to make a small ballast to fit in the body of the lamp, how to dissipate heat, how to
construct in a small space the long tube that was still needed for electrical efficiency,
how to keep the cost down. The Philips company solved the "long tube" problem by
creating bridges between short tubes (3); other companies bent a longer tube.
Philips marketed the first true compact fluorescent lamp in 1980 (1).
Westinghouse quickly followed (2).
Efficacy of the compact fluorescent lamps at first was about 40 lumens per watt. In
later versions (with electronic instead of magnetic ballasts) efficacy can be above 80
lumens per watt.
[L18.1- information label]
Inventors during the 1970s tried many ways of making compact fluorescent lamps.
Displayed here are a few of those designs. Lamps using some of the concepts shown
here are now starting to appear on the market (see the Competition section on the wall
Donald Hollister, an independent inventor, and John Anderson of GE, designed
lamps that did not use electrodes (4, 5). They both encountered problems with heat and
complex electronics. Edward Hammer's (GE) "spiral lamp" (8) was difficult to
mass-produce, as were "partition" designs from several inventors (7). Leo Gross's and
Merrill Skeist's (Spellman Electronics) "magnetic arc-spreader" also involved complex
[L18.2 - credit label]
- Philips "SL-18", European version, about 1981, [1992.0553.08] from Lawrence
- Westinghouse "Econ-Nova" lamp, about 1981, [1997.0389.24] from Philips Lighting
- Philips "PL-7/9" modular lamp, about 1983, [1999.0324.02] from the U.S.
Department of Energy
- Experimental "Litek" electrodeless lamp, about 1979, [1992.0466.01] from the U.S.
Department of Energy
- Experimental "SEF" electrodeless lamp, about 1979, [1998.0050.07] from the
General Electric Corporate Research & Development Lab.
- Experimental "magnetic arc-spreading" lamp, about 1978, [1998.0050.16] from the
General Electric Corporate Research & Development Lab.
- Experimental "partition" lamp, about 1978, [1998.0050.19] from the General Electric
Corporate Research & Development Lab.
- 7 - Experimental "spiral" lamp, about 1976, [1997.0212.01] from Edward E. Hammer
- Diagram of the "SL-18" lamp, 1981, from the U.S. Department of Energy,
©Philips Lighting Co. (Note: Original part descriptions added to photo
- Louis Vrenken and Johan B.J. van Overveld with experimental lamps, 1981, from
Philips Lighting Co.
[CT7 - header panel: Silica Carbide]
Silica Carbide Incandescent: the lone inventor
[L19- information label for Case Study #5 - Silica Carbide]
[G-36: Time line on counter in front of case]
"Ninety percent of my time goes into building and designing equipment."
John Milewski, founder Superkinetic Inc., 1996
It's still possible to be a lone inventor--but it's not easy. A lot of time is spent making
equipment that might be provided in a company laboratory. And significant resources
are needed for development, patent applications, promotion.
John Milewski had experience as an engineer using silica fibers to reinforce
materials. In 1987 his son Peter used these fibers--which could be heated to very high
temperatures-- as lighting filaments for a science fair project. The son's project took
third place. But it also produced U.S. patent number 4,864,186, and the father was
stimulated to continue experiments in his living room using surplus equipment (B).
Among that equipment were the items you see here: an optical pyrometer (4), and a
vacuum bell-jar with filament mounted in the holder (3).
In 1991 and 1992 he obtained financial support from the Electric Power Research
Institute and in 1993 from a program sponsored by the National Institutes of Science and
Technology with the Department of Energy. As this exhibition is being prepared in 1999
the practicality of the silica-carbide lamp is still uncertain, but experiments continue.
[silica carbide information]
[L20 - information and credit label]
Early experimental silica-carbide lamps (1, 2). The one on top has an intact
filament, but it is very small and difficult to see. The one on the bottom is the lamp seen
in this photo (D). Note that the bases have been salvaged from ordinary commercial
- Experimental silica carbide filament lamp, 1989, [1992.0554.01] from John V.
- Experimental silica carbide filament lamp, 1989, [1992.0554.03] from John V.
- Bell-jar with filament mount used in home lab, [1996.0358.04] from John V. Milewski
- Optical pyrometer used in home lab, about 1987, [1996.0358.06] from John V.
- John, Sr. and Peter Milewski, 1988, from Drs. John and Peter Milewski
- Home lab of the Milewskis, about 1989, from Drs. John and Peter Milewski
- Chart of experimental results for science fair, 1987, from Drs. John and Peter
- Silica carbide lamp in operation, 1988, from Superkinetic, Inc.
[CT8 - header panel: Sulfur Lamp]
Sulfur: serendipity in a non-lighting company
[L21 - information label for Case Study #6 - Sulfur Lamp]
[G-42: Time line on counter in front of case]
"You were allowed to make a fool of yourself."
Michael Ury, engineer at Fusion Lighting, 1996
What Ury meant was that it's nice to be able to try something that seems silly and
not have everybody laugh at you. Sometimes it can to lead to an unexpected but
In 1972 four physicists and Ury (an engineer) founded Fusion Systems, where they
hoped to develop products based on their experience in high-energy physics, with
microwaves in particular. They were successful with an ultraviolet lamp (energized by
microwaves) for drying inks in high-speed industrial processes.
In 1980 Ury and Charles Wood (one of the physicists) tried activating sulfur in one
of the ultraviolet tubular lamps to produce visible light. The effort failed and was quickly
abandoned. Meanwhile various improvements were made to the ultraviolet system,
including use of a spherical rotating source lamp. In 1990, looking for ideas for new
products, Ury remembered the work with sulfur and decided to try it in a spherical
rotating arrangement. The result was a very intense light with good color properties.
To diffuse the light over a large area they found just what they needed in a prism
light guide that had been recently invented by Lorne Whitehead at the University of
British Columbia (B).
Ury, on the left (A), poses with Lee Anderson, Lighting Program Manager at the
Department of Energy. The DOE Program provided funding and technical assistance,
and helped in promoting the new lamp. A commercial version was available in 1996.
[sulfur lamp information]
[L22 - combined information and credit label]
The tubular, ultraviolet lamp (1) lay at the core of the company's main product in
Eureka! This test result (C) showed that the new lamp worked.
Lamp 2 contains sulfur and is one of the original experimental lamps. Lamp 3
contains selenium, and lamps 4 and 5 have varying proportions of selenium and sulfur.
Note that there are no electrodes. Power is supplied by an external microwave source.
It's output is 135,000 lumens (equal to about 75 standard 100-watt incandescent lamps).
Efficacy: 160 lumens per watt of microwave energy, 98 lumens per watt for the total
- Electrodeless ultra-violet lamp, about 1980, [1996.0359.03] from Fusion Lighting
- Early experimental sulfur bulb, 1990, [1992.0467.01] from Fusion Lighting Inc.
- Experimental selenium bulb 1996, [1996.0359.09] from Fusion Lighting Inc.
- Experimental sulfur / selenium bulb, 1996, [1996.0359.05] from Fusion Lighting Inc.
- Experimental sulfur / selenium bulb, 1996, [1996.0359.08] from Fusion Lighting Inc.
- Michael Ury, Lee Anderson, and the five demonstration lighting units, 1994, from
- Experimental lighting system in Fusion System's plant, 1993, from Fusion Lighting
- Photocopy of computer read out showing spectra of successful experiment, 1990,
from Fusion Lighting
[GL120 photomural on end wall]
"If I have to tell you what to do I don't want you."
John Anderson, former GE engineer, quoting his boss in 1951, six
months after he started work at the GE laboratory, 1996
At Menlo Park, the head of the laboratory (Edison) was its principal thinker; others
were there primarily to carry out his ideas. In the industrial research laboratories that
emerged after 1900, the head of the laboratory was an administrator; other engineers
and scientists developed ideas, though often in response to particular problems. It was
the function of the administrator to define problems, to determine which ideas should be
pursued, and to provide resources for their pursuit.
Life-test room at Duro-Test plant, about 1972, from Duro-Test Corporation