A History of Field Emission
Displays
by
Jeffrey A. Hart
Indiana University
Stefanie Ann Lenway
University of Minnesota
Thomas Murtha
University of Minnesota
Third Draft
September 1999
The research for this paper
was funded by the Alfred P. Sloan Foundation.
We would like to thank Ivor Brodie, Francis Courreges, Ted Fahlen, Tom
Holzel, David Mentley, Dev Palmer, Robert Pressley, and Charles Spindt for
comments on previous drafts. Craig
Ortsey provided research assistance.
Please do not cite or quote without the written permission of the
authors.
Introduction
It
has taken more than three decades for field emission displays (FEDs) to go from
idea to commercial product. The purpose
of this essay is to trace the history of FEDs and to assess the future
prospects for the successful commercialization of FEDs.
In
1968, Charles A. "Capp" Spindt at the Stanford Research Institute
(now called SRI International) had the idea of fabricating a flat display using
microscopic molybdenum cones singly or in arrays (FEA’s). This development was
the enabling technology the concept for using FEA’s in a matrix addressed
display (FED) conceived by the SRI team of which Capp was a member, and
patented by Crost, Shoulders and Zinn in 1970 (US Patent 3,500,102). However
SRI was unable to obtain funding for developing this concept in the decades of
the seventies and early eighties, and the initiative for developing the
technology moved to the Laboratoire
d'Electronique de Technologie et de l'Informatique (LETI), a research arm
of the French Atomic Energy Commission, in Grenoble. LETI picked up on the technology and publicly demonstrated an
operating display in 1985. The SRI team
was finally funded by Boeing and Commtech International (a venture capital
patnership) to develop a full color display and were able to demonstrate the
first color FED in 1987. The technologies developed at SRI were licensed to a
start-up company called Coloray.
Coloray was not successful but some key members of the staff were able
to join Silicon Video Corporation (SVC), a company which subsequently became
Candescent Technologies Corporation (CTC).
The technologies developed at LETI became the basis for a firm called
Pixtech, which developed a new strategy of building multiple partnerships for
the manufacturing and commercialization of FEDs. Firms like Raytheon, Futaba, Candescent, Motorola, and Micron
Display developed their own FED programs with variants of the original Spindt
cathode. Of these firms, only
Candescent and Micron were not partners with PixTech. By the middle of 1999,
the race for commercialization was reduced to one between Candescent and the
PixTech alliance.
The Fowler Nordheim Law
The
Fowler-Nordheim Law explaining field emission as a quantum effect became the
basis for research on FEDs. A potential
barrier at the surface of a metallic conductor called the "work
function" binds electrons to the material.
For an electron to leave the material, the electron
must gain an energy which exceeds the work function. This can be accomplished in a variety of ways, including thermal
excitation (thermionic emission), electron and ionic bombardment (secondary
emission), and the absorption of photons (photoelectric effect). Fowler-Nordheim emission or field emission
differs from these other forms of emission in that the emitted electrons do not
gain an energy which exceeds the material work function.
Field emission occurs when an externally applied
electric field at the material surface thins the potential barrier to the point
where electron tunneling occurs, and thus differs greatly from thermionic
emission. Since there is no heat
involved, field emitters are a "cold cathode" electron source.[1]
The
original work of Fowler and Nordheim resulted from efforts to understand the
causes for the interruption of transatlantic radio broadcasts from the Rocky
Point, New Jersey, transmitter of the Marconi Company in the 1920s. These interruptions resulted from arcs in
vacuum tubes in the radio equipment.
Careful examination revealed that the arcs occurred at sites on the
tungsten filament of the tubes where there was a bump or a whisker. B.S. Gossling, an employee of the Marconi
firm, published a paper to this effect in 1926. He noticed that the arcs occurred only when certain currents and
voltages were reached.[2]
Ralph
H. Fowler was Professor of Physics at Cambridge University[3]
and Lothar W. Nordheim had a research position there in 1928 when the two
published an article that explained the arcing at Rocky Point in terms of
quantum mechanics. The Fowler-Nordheim
equation contained in the article linked the voltage and current of the field
emission to the surface area of the emitter, the average work function of the
emitter surface, and the magnitude of the electric field at the emitting surface.[4] One of the reasons for the arcs experienced
at Rocky Point was the very small surface area of the emitter (the bump or
whisker) combined with the high intensity of the electric field at that point.
It
took eight more years for experimentalists to adequately test the theory of
Fowler and Nordheim. According to Ivor
Brodie:
Gossling tried to make electrolytically polished
spherical points, but these did not quantitatively verify the Fowler-Nordheim
theory, again due to whisker growth under high fields. The theory was not quantitatively verified
until Mueller grew single-crystal points…in 1936.[5]
Erwin
W. Mueller not only helped to verify the Fowler-Nordheim equations, he also
contributed to later work on FEDs by accidentally inventing the field electron
microscope (FEM). Again according to
Brodie:
Mueller was the first man to truly resolve atoms
with his field ion microscope -- he once told me how he discovered the effect
by accidentally reversing the field on a field electron emission microscope and
momentarily seeing an atomic resolution image.
Afterwards he wondered if he was dreaming since he could not repeat the
experiment. But by perseverance and a
clearer understanding of what was occurring, he finally achieved his goal.[6]
The
scanning electron microscope (SEM) now used for inspecting the surface of
semiconductor devices in all world-class integrated circuit laboratories and
plants is based on the same principles as the FEM discovered in 1936 by
Mueller.[7] The work on electron microscopes also led
to the development of electron-beam (e-beam) lithography, which made it
possible to write patterns smaller than one micron on a thin-film device before
it was possible to do this with photolithography.
The
theoretical work of Fowler and Nordheim and the experimental work of Mueller
paved the way for the work of Spindt and his colleagues at SRI and for all
subsequent efforts to fabricate field emission displays.
The Work of Spindt and His
Colleagues at SRI
There
had been many attempts by researchers to improve upon the cathode ray tubes
that were used for television displays and scientific instruments (like
oscilloscopes), and particularly to make these displays flatter, thinner, and
more energy efficient. Some of the
earlier work on flat CRTs focused on creating an array of thermionic emitters
of electrons inside a flattened vacuum tube, but these efforts did not pan out.
|
Figure 1. Charles A.
“Capp” Spindt |
|
|
|
Source: Society for
Information Display web site at http://www.sid.org. |
In
1961, Kenneth R. Shoulders at the Stanford Research Institute (now called SRI
International) had the idea of micro-miniaturizing cathode ray tubes by using
field emission as the source of electrons and using electron beams for
micro-patterning the emitters.[8] Before leaving SRI, Shoulders hired Capp
Spindt and gave him the mission of developing a method for manufacturing a
field emission display like the one outlined in his 1961 article.
Spindt's
idea was to use thin-film deposition and etching techniques together with
e-beam lithography to build conical emitters (he called them triodes) on a
silicon substrate. In 1968, Spindt
published the results of his efforts to fabricate an array of polycrystalline
molybdenum (Mo) microtips on a substrate.[9] Subsequent work with Ivor Brodie and others
was funded by a variety of U.S. government agencies (particularly the Army
Research Office, the Naval Research Laboratory and NASA).[10] One of the early problems in this research
program was sharpening the tips of the cathodes in order to reduce the voltage
required to start the tunneling effect.
The sharper the tip, the lower the voltage needed. Jules D. Levine at RCA's Sarnoff Laboratory
discovered that it was possible to sharpen the tips by growing a layer of
copper oxide on fully-formed copper tips and then etching away the oxide layer
leaving tips composed of a few atoms of copper.[11] Levine received a patent on this process in
1975.
|
Figure
2. Structure of a Field Emission Display |
|
|
|
Source:
Micron Display at http://www.micron.com/mdt/tech/tech.htm
on September 15, 1999. |
At
the beginning of the research on Spindt cathodes, it was not unusual for FEDs
to require gate voltages of 400V and anode voltages of 1,000V or more. Such high voltages often resulted in arcing,
especially if there were any gas molecules between the cathode and anode. The
arcing would damage the cathode surface even though it did not necessarily
damage the cathodes themselves. Arcing
might result, however, in the formation of metallic whiskers on the cathodes
that could eventually destroy the device.
The better the vacuum, the lower the likelihood of arcing; but it was
not practical to build a display that required an almost perfect vacuum. Therefore, tip sharpening using oxidization
techniques was a very important step forward in making it possible to
manufacture FEDs.
By
1972, the SRI group had solved enough of the engineering problems to
demonstrate to their satisfaction that it was feasible to manufacture FEDs.[12] Patents for their fabrication methods were
issued to Spindt, Shoulders, and L.N. Heynick in 1973 and 1974.[13]
However, it was not until1985 that they were able to find backers for the next
stage of the research program, namely the development of a fully operational
prototype display. The work was initially funded by Boeing through Commtech
International (a venture capital company formed to capitalize on SRI’s
inventions headed by Dr. Gerald Marxman).
A
12-inch diagonal flat display for individual passenger usage was required for a
new aircraft that was then on the drawing boards. The initial prototypes showed promise but due to the availability
of competitive technology Boeing ceased support after two years. At this point
Commtech decided to form a new company called Coloray (President Charles
“Chuck” Anthony) to exploit the SRI FED technology, with an investment of $3
million (the maximum allowed by Commtech’s charter). The company elected to
acquire a pilot production facility and intended to find partners to finance
putting the plant into operation. In the business climate at the time this
strategy proved unsuccessful and Coloray after some adventures was finally
dissolved. Some key members of the staff joined another display company,
Silicon Video Corporation (SVC), which subsequently became Candescent
Technologies.
However,
they could not find any backers for the next stage of the research program: the
building of a pilot production line. It
took them ten years to build prototypes, but by this time, a better-funded
program at LETI had moved ahead of the SRI group to solve some key
technological problems connected with manufacturing FEDs. At this time, the
well-funded program at LETI had moved ahead of the SRI group to solve some key
technological problems connected with manufacturing FEDs.
Silicon Emitters vs.
Molybdenum Microtips
Many
different technologies have been investigated to overcome the real and imagined
problems with the Spindt technology, Silicon based devices being the most
prominent due to the availabilty of silicon fab lines for integrated circuits. One
approach to FEDs was to use a p-n junction in a semiconductor device
"biased into avalanche" as a low voltage "cold" electron
source.
In such a device electrons in the depletion layer,
either existing or generated by impact ionization within the same layer, gain
energy from the high field in such a region; and, if the junction is
sufficiently shallow, many of them can acquire the necessary kinetic energy to
reach the surface and win the surface work function, thus escaping into vacuum.[14]
In
1969, Williams and Simon published a paper on getting field emission from
forward biased silicon p-n junctions.
This work set off a series of experiments on using silicon instead of
molybdenum as cold cathodes.
In
1981, Henry F. Gray and George J. Campisi of the Naval Research Laboratory
(NRL) patented a method for fabricating an array of silicon microtips.[15] The process involved thermal oxidation of
silicon followed by patterning of the oxide and selective etching to form
silicon tips. The silicon tips could
act as cold cathodes for a FED.
In
December 1986, Gray, Campisi, and Richard Greene demonstrated a silicon field
emitter array (FEA) that was fabricated on a five-inch diameter silicon wafer
at the International Electron Devices Meeting but did not formally published
their results.[16] At this time, Gray and his colleagues were
more interested in demonstrating the benefits of using FEAs as switching
devices (like transistors) or amplifiers than as displays, but the possibility
of using them for displays was held out as an additional potential benefit.
The
arguments that Gray made in favor of FEAs was that they might permit faster
switching speeds than semiconductors and that they had a greater probability of
surviving extreme conditions such as those experienced in war or natural
disasters. FEAs could be made in either
a vertical or a planar structure. They
could be constructed using a variety of materials: silicon, gallium arsenide,
molybdenum, nickel, platinum, iridium and conducting carbides. The devices did not require an extreme
vacuum to operate but there were problems of getting uniformity in FEAs that
had to be overcome before they could be used in commercial displays.[17]
After
1988, Gray became a major supporter of federal funding for further research on
FEDs and FEAs. Government funding of
subsequent research on FEAs was motivated primarily by the desire of the
government to replace thermionic cathodes in microwave amplifier tubes. As it turns out, the operating parameters
required for microwave tubes are much more demanding than those for display
applications.[18]
In
September 1995, Gray and David S.Y. Hsu announced a new process, using chemical
beam deposition techniques to fabricate thin-film-edge field emitters. Instead of using molybdenum or silicon
cones, as before, the new process created vertical freestanding
platinum/lithium/platinum triangular shaped wedges with very sharp tips.[19]
Research
on silicon emitters continues, but -- to our knowledge -- apart from Micron’s
(now defunct?) use silicon emitters for camcorder displays, there are no major
investments in this technology for the purpose of fabricating displays.
The Work of Robert Meyer and
His Colleagues at LETI
In
1983, Robert Meyer of LETI began a program of research on fabricating arrays of
Spindt cathodes using molybdenum tips.
By 1988, the LETI research program had born fruit. Meyer patented a method for fabricating an
array of molybdenum microtips with a lateral resistive layer. The lateral resistive layer helped to
prevent arcing between the cathode and anode and therefore to improve the
performance and prolong the life of the FED.
It also evened out the electron emissions from the individual arrays
across the display. The LETI
researchers were able to demonstrate that FEDs fabricated with lateral
resistive layers could be manufactured with the desired uniformities and had
the expected lifespans needed for commercial displays.
However,
using emitters with a base diameter of one micron, approximately 80V was needed
to switch the emitters. This was much
higher than the voltage needed to switch pixels in liquid crystal displays. Therefore, integrated circuit drivers for
LETI's FEDs would be very expensive and the device itself would not be portable
because of its high power requirements.
Tom
Holzel of Raytheon was following Meyer's work and recommended in 1988 that
Raytheon Corporation invest in FEDs.
Raytheon decided to license the patents that SRI had obtained on the
basis of the work of Spindt and his colleagues. A venture capital company named Commtech International had
purchased the patents from SRI in 1985.
However, Commtech and Raytheon could not work out a deal. Commtech would not assume any of the risk in
commercializing the SRI technologies.
Raytheon decided to work on its own.
The Commtech technology was sold to Coloray (see below).[20]
|
Figure 3. Jean Luc
Grand-Clement |
|
|
|
Source: PixTech |
Meyer
tried to sell LETI's technology to Raytheon in 1988, asking $25 million for
it. Raytheon opted not to take this
offer, so the technology was sold instead in 1992 to Jean Luc Grand Clement, a
former executive of Motorola in France who had already been involved in three
previous startup firms, including, most notably, a firm called European Silicon
Structures (ES2). After selling his
interest in ES2, Grand Clement began working with Apple Computer's venture
capital fund for Europe. His primary
responsibility was to review opportunities for investment in Europe. He took some Apple people to meet with the
people at LETI who were working on FEDs in October 1991.
PixTech
Apple
was not interested in manufacturing displays, but Grand Clement was
intrigued. He put together a short
proposal for venture capital financing of a new display company called
PixTech. A VC firm called Advent
International became one of his earliest investors. Advent was convinced that PixTech would be able to produce FEDs
using the LETI patents without infringing the patents owned by SRI. PixTech raised capital in incremental
amounts: starting at $3 million from Advent in late 1991 and then rising to
$10, $22 and $35 million from other investors.
Over half of the funds came from U.S. investors.[21] Pixel International was formed as a French
company in June 1992; PixTech, Inc., was formed as a U.S. corporation in 1993.
Using
these funds Grand Clement purchased exclusive rights to all 16 of the patents
that were owned by LETI in early 1992.
He convinced Raytheon to become a partner of PixTech so that it would
then have access to LETI's patents. Tom
Holzel moved from Raytheon to become Vice President for Marketing and Sales of
PixTech in July 1995. Motorola, Texas
Instruments, and Futaba also joined the PixTech alliance for the same
reason. TI pulled out in March 1996
when it abandoned its efforts to manufacture FEDs.
All
knowledge and patents in the PixTech alliance are held in common. Everything is shared, but only within the
partnership. The partners meet quarterly
and request help on specific problems.
They split up tasks and then report back on their progress at subsequent
meetings. The meetings are rotated to
various locations convenient for the partners.[22]
PixTech
set up an R&D facility in Montpellier, not too far from LETI in
Grenoble. The equipment in their pilot
production plant in Montpellier was "off the shelf" -- a small amount
of the equipment was custom developed within the PixTech alliance. The plant used a conventional semiconductor
photolithography process, with an MRS 2 micron stepper. They also used a Holtronics holographic mask
system at LETI for photolithography of microtip emitter holes. The facility was leased from IBM
France. Production of FEDs commenced in
November 1995 and the first samples of monochromatic (green) FEDs produced by
Futaba were sold in December of that year.
These displays were 5.2 inch diagonal, 1/4 VGA displays with 70 foot
Lamberts (fLs) of brightness.[23] PixTech opened an R&D facility in Santa
Clara, California, in February 1996.
The
displays made initially by PixTech used Mo cathodes 1.2 microns tall. There were 2,000 emitters per pixel. The anode voltage was 350V; the gate voltage
was 80V. Grey scale was obtained by
varying the gate voltage by about 30V.
The phosphors used on the anode of the Futaba monochromatic displays
were ZnO:Zn phosphors, just like those used in fluorescent tubes.
Production
yields in the first half of 1996 were erratic. There were problems with the
lithography equipment. PixTech's approach involved the use of high-voltage
drivers for the cathode combined with low anode voltages. The high cathode voltages created problems
of arcing and expensive drivers while the low anode voltages meant that PixTech
FEDs could not be as bright as CRTs (which relied on high-voltage phosphors)
unless better low-voltage phosphors could be developed.
The
first markets that PixTech targeted were avionics, medical devices, and motor
vehicles. These markets were interested
in smaller displays, in the 4 to 8.5 inch diagonal range. Customers in these markets needed very
bright displays that were readable from wider angles than were available with
LCDs
|
Figure
4. PixTech’s Field Emission Display |
|
|
|
Source: PixTech web site at
http://www.pixtech.com. |
The
decision of Texas Instruments to end its work on FEDs in March 1996 was a blow
to the alliance. TI's marketing manager
for flat panel products said there was "…too much R&D going into
liquid crystal technology based on the revenue streams being generated for the
technology not to have made some significant improvements."[24] The decline in prices for TFT LCD displays
that followed the ramp up of Samsung's first major factory also influenced TI's
decision. Nevertheless, PixTech and its
remaining partners persisted. For
example, PixTech concluded a deal with Nichia Chemical of Japan in March 1996
to develop new low-voltage phosphors (to make brighter displays).[25]
PixTech
hired Richard Rodriguez away from IBM in April 1996 to improve yields. He had previously directed an IBM plant in
Corbeil, France. Rodriguez became
Executive VP and Chief Operating Officer in March 1997.[26]
In
May 1996, PixTech unveiled a 10.5-inch FED color display at the annual meeting
of the Society for Information Display in San Diego, California. It also showed an advanced 6-inch color
display with a brightness of 100fL and 160 degrees viewing angle. There was also a brighter 5.2 inch
monochrome display with a brightness of 200fL at this show.
In
September 1996, PixTech signed a marketing agreement with an unnamed Japanese
manufacturer. It later announced that
the Japanese firm involved in this agreement, Sumitomo Corporation, would be
the exclusive distributor of PixTech products in Japan.[27]
|
Figure 5. Picture of a Color FED |
|
|
|
Source: PixTech web site
at http://www.pixtech.com. |
In
November 1996, PixTech announced that it had concluded a memorandum of
understanding with Unipac Corporation in Taiwan to use Unipac's high-volume TFT
LCD manufacturing facility for the production of FEDs on a contract basis. In January 1997, UMC, Unipac's parent firm,
said that it would purchase 1.1 million shares of PixTech for $5 million of
PixTech's. PixTech would sign a display
foundry agreement with Unipac in June 1997.[28] Also in January, PixTech was given a
"Display of the Year" award from Information
Display magazine for its 5.2-in. monochrome display.
In
April 1997, PixTech announced that it would lead a consortium of European
companies to develop FED technology under a $3 million grant from the European
Union. They would work with SAES
Getters and Rhone-Poulenc to develop getters and low voltage phosphors
respectively. PixTech also received a
zero-interest loan of $2 million from the French government to develop FED
displays for multimedia applications.[29]
Dieter
Mezger replaced Jean Luc Grand Clement as President of PixTech in March
1998. Mezger had previously worked for
Texas Instruments and VLSI Technology.
Grand Clement retained his position as Chairman and CEO of the firm.[30] In the second quarter of 1998, Zoll Medical
Corporation introduced a portable defibrillator that incorporated an FED. During this same period, the first FEDs came
off the production line at Unipac.[31]
PixTech
purchased Micron Display for $16.8 million of 7.1 million shares of PixTech
stock in March 1999. Micron would own
30 percent of PixTech's equity as a result and, as part of the deal, Micron
would transfer $4.35 million in cash and an unspecified amount in liabilities
to PixTech. PixTech said that it would
use the Micron display facility in Boise, Idaho, for the co-development of
15-inch displays with a major unnamed Japanese company. David Cathey, formerly Vice President and
General Manager of Micron Display, would become an observer on the PixTech
board of directors.[32]
With
the purchase of Micron, it was now theoretically possible for PixTech and its
partners to fabricate a wide range of sizes of FEDs for a variety of potential
applications. The combination of
government research subsidies with private venture capital financing was enough
to get PixTech started, but it took a careful strategy of alliances and
partnerships to take the firm to the brink of successful
commercialization. We shall see below
that a similar pattern emerged for PixTech's main rival, Candescent.
Pixtech's Partners:
Raytheon, Texas Instruments, Futaba, and Motorola
Raytheon
As
discussed above, Raytheon's interest in FEDs started in 1988 with Tom Holzel's
recommendations to invest in the new technology. Raytheon is primarily a supplier of products using advanced
technologies to the U.S. Department of Defense. Like most other defense contractors, Raytheon began looking for
new dual-use technologies to get into in the late 1980s as the Cold War wound
down. Raytheon focused its efforts in
this area on high brightness FEDs for military avionics applications. It did not produce displays for civilian
applications.
Texas
Instruments
Texas
Instruments, along with Raytheon and Futaba, was one of the three initial
partners in the PixTech alliance. TI's initial interest in FEDs stemmed from
its desire to have an alternative to TFT LCDs for notebook computers. In 1990, TI was paying around $1500 per
display for TFT LCDs. When PixTech
board members suggested that equivalent FEDs would eventually cost around $500
per unit, TI got interested in FEDs.
The FED project at TI was put under the control of the analog IC
division of the firm that was responsible for digital signal processors
(DSPs). There was no long-term budget
for the project. It was reviewed on a
year-by-year basis. TI built an R&D
laboratory, and PixTech transferred its proprietary technology to the lab, but
apparently the lab had difficulties using the PixTech technologies to build FED
prototypes. In October 1995, a new
manager, Bruce Gnade, was brought in to direct the project. Gnade later went on to become the head of
the High Definition Systems (HDS) Program at DARPA. By the end of 1995, the price of 12.1-inch TFT LCDs had descended
to less than $500, so TI management decided to pull the plug on the FED program
early in 1996.[33]
Futaba
Futaba,
like Raytheon and Texas Instruments, was one of the three initial partners in
the PixTech alliance. Futaba had
experience with manufacturing monochrome vacuum fluorescent displays (VFDs) for
small devices like watches and was interested in scaling up its work on
VFDs. VFDs use low-voltage phosphors,
so the fit with PixTech's FED technology was good. Futaba was the first member of the PixTech alliance to produce a
commercial FED. As mentioned above,
this was a 5.2-inch diagonal, 1/4 VGA display with 70 foot Lamberts (fLs) of
brightness. Production began in
November 1995 and the first units were sold in December 1995.
One
of Futaba's contributions to the development of FED technology was its work on
"getters." A getter is a
chemical substance that is introduced into the cavity between the FED anode and
cathode to reduce cathode current fluctuations caused by residual gases in the
cavity. Good getter technology, in
short, in necessary to obtain uniformity in performance and to prolong the
expected lifetime of the display.[34]
Futaba
demonstrated prototypes of color FEDs at a number of international conferences
after 1997, but had not introduced any of these displays to the marketplace as
of mid 1999.
Motorola
In
1993, Motorola set up a corporate R&D facility to do research on FEDs. In 1995, the company set up the Flat Panel
Display Division of the Automotive Energy Components Sector (AECS). A six-inch glass wafer pilot production line
was built in Chandler, Arizona, in 1996.
Later that year, construction of a second-generation 275,000 square-foot
facility was begun in Tempe, Arizona.
In 1998, construction was completed.
The new plant was capable of producing up to 10,000 units per
month. It was designed to produce
multiple displays on 37cm by 47cm glass substrates, a standard size for
second-generation TFT LCD plants.
Motorola worked with the following suppliers:
Nikon
-- for production steppers
MRS
-- for development steppers
Lam
Research -- for dry etch equipment
Ulvac
-- for vacuum technology
Balzers
-- for sputtering equipment
Problems
of ramping up production occurred soon after the completion of the plant. The main problems had to do with sealing the
glass tubes, getting the right spacers to get uniform distances between the
anode and the cathode, and discovering the right getters for dealing with
residual gases in the cavity.[35]
In
May 1999, the manager of the Flat Panel Display Division of Motorola, Pete
Shinyeda, said that Motorola would delay the ramp up of its FED production
facility in order to solve some basic technology problems. Shinyeda said that the main problem was the
limited lifetime of color displays (the firm was unable to achieve its goal of
a 3,000-hour lifetime). On May 10,
Motorola announced that it laid off 120 of its 200 employees at the Tempe site
and that the staff would be pared down to a "small research team."[36]
Micron Display
Commtech
International sold the 14 SRI patents to Coloray Display Corporation of
Fremont, California, in 1989.
Micron
Display Technology, Inc. (MDT) was formed in January 17, 1992 as a subsidiary
of Micron Technology, Inc. On the day
of its founding, MDT purchased a small stake in Coloray, but apparently it was
not enough to keep Coloray afloat. In
1992, Coloray filed for bankruptcy under Chapter 11. It was purchased after reorganization by Scriptel Holding in
April 1994. Scriptel was a pen-based
digitizer company headquartered in Columbus, Ohio.[37] Coloray closed its doors for good in ______.
MDT
secured a DARPA contract for development of FEDs in 1993. It demonstrated 0.7-inch prototypes in
1994. MDT developed a new manufacturing
technology for FEDs involving chemical mechanical polishing (CMP) of the
cathode substrate. CMP makes it possible to manufacture cold cathode arrays
precisely and reliably without the need for lithography.
The process begins with the deposition of the tips,
which are then covered by a dielectric layer, which in turn is covered by the
conductive film that will constitute the FED's gate. All you see at this point is a planar film having many bumps, the
bumps having been created by the buried tips.
The trick involved is that chemical-mechanical
polishing removes material much faster at the bumps than anywhere else. As the bumps are polished down, the surface
becomes flat, the polishing nearly stops.
The polishing is self limiting because the removal rate is dependent on
the local pressure between the polishing pad and the conductive layer. A wet chemical etch is then used to remove
the insulator surrounding the tips.[38]
The
process was patented in ____.
Micron
also developed some new spacer technology for both high and low voltage FED
applications.[39]
At
the end of 1996, MDT began to produce .55-inch diagonal monochrome FEDs for
military viewfinders and thermal imagers.
Kaiser Opto-Electronics purchased MDT's FEDs for head-mounted displays
under a contract with DARPA.[40] These displays are used on rifles and other
hand-held weapons. They can detect both
movement and heat. Another application
of these displays was for firefighting: they were used on devices to detect hot
spots in burning buildings.
A
new 50,000 square foot facility in Micron's hometown, Boise, Idaho, with 30,000
square feet of clean room became operational early in 1997.[41] This new facility was initially designed to
produce 12-inch displays for the Abrams M-1 tank commander station. David Cathey, president of MDT, said that
the CMP technology would permit MDT to scale up to displays as large as 17
inches. GVC Corporation of Taiwan
invested $4.5 million in MDT around this time.[42]
After
struggling for two years to manufacture large-area FEDs at the Boise facility,
Micron Technology decided instead to sell it to PixTech in exchange for an
equity stake in the latter firm.
PixTech would use the facility to produce large-area FEDs for
televisions and workstations.
Candescent
With
the sale of Micron Display to PixTech in March 1999, Candescent Technologies
Corporation became the main competitor to the firms in the PixTech
alliance. Candescent Technology
Corporation grew out of Paul Lovoi's idea of using multi-layer ceramic material
as an enabling element for flat CRTs.
Earlier efforts at Source Technology had involved the use of etchable
glass as a grid above a toaster wire cathode to fabricate a monochromatic tube,
but Source was unable to solve fragility and focusing problems, even with thick
self-supporting face and backplates.
Lovoi, co-founder and president of InTA, a metal ceramic research and
development company, proposed that recent developments in multi-layer and
zero-shrink ceramics could solve both the electrical and mechanical problems
connected with flat CRTs.
Lovoi,
Darryl Wilburn and Lowell Nobel approached Marko Slusarczuk at DARPA in 1989
with a request for $1 million to fund development of a flat CRT based this
multi-layer ceramic technology and using a hot-wire cathode. He expressed considerable interest, but such
a project did not fit with the work of any of the three founders' companies.
Robert
Pressley had just sold his previous startup, XMR, a manufacturer of industrial
excimer lasers, to Amoco Corporation.
Pressley was approached by Lovoi, Wilburn, and Nobel to form a new
company called Silicon Video Corporation (SVC) to develop flat CRTs. Pressley had a background in flat CRTs,
having worked on the idea when he was employed at the RCA Sarnoff Labs in
Princeton, New Jersey. Pressley, along
with Capform Ventures, decided to invest $400,000 in the new company. Slusarczuk negotiated a contract with SVC
to build a flat CRT using funds from the Army.
Pressley
recruited Robert Duboc from Coloray, Ted Fahlen, VP of R&D at XMR, and
Chris Spindt, son of Capp Spindt with a newly minted Ph.D. from Stanford, to
work at SVC. These individuals built a
prototype hot cathode display with 80 lines per inch that operated on 10 watts
of power.
A
second round of financing was secured from Hewlett-Packard and Carl Berg, a
Silicon Valley venture capitalist and real estate executive. The role of
Hewlett-Packard here was crucial.
Richard Hackborn at H-P had backed the relationship with SVC because he
felt that HP should be a dominant player in displays. Displays would be very important to the future of the company, he
thought. Hackborn had led the
successful program at H-P to partner with Canon on the development and
manufacturing of ink-jet printers. He wanted
to do something novel in displays, while avoiding the expense and difficulty of
setting up a TFT LCD plant. His
philosophy was "if the enemy has superior forces and occupies the high
ground, do something more novel than stage a direct uphill attack."[43]
When
it became clear in 1992 that LCD power consumption was coming down and that hot
cathodes would not be efficiency competitive with LCDs for portable notebooks,
SVC shifted its efforts toward implementing cold cathodes. After extensive modeling and experience with
ceramics and electron beam focusing, it became apparent that a truly
competitive thin cathode ray tube must have the following attributes:
1.
The
display must operate at a high anode voltage, both the efficiency and lifetime
of phosphors increase rapidly with voltage.
It should not generate x-rays, so the voltage should be no more than 6
to 8 KV. The existing P22 CRT phosphors are adequate, they took twenty years to
optimize in the beginning of color television and any new and better ones are
unlikely in any reasonable time. (This conclusion came out of Pressley’s experience
with phosphors at RCA Sarnoff Labs.)
2.
The
front and back glass should be supported with thin, weakly conductive ceramic
walls spaced between the pixels. The
ceramic thermal expansion must match the glass, the electrical resistivity must
be high enough to minimize leakage from the anode, but must be low enough to
avoid any space charge buildup. (This came out of Lovoi’s ceramics experience.)
3.
The
switching voltage on the Spindt type emitters must be 10 volts or less to
minimize switching losses. (This was
based on Fahlen's modeling.) This, in
turn, required gate diameters to be about 0.1 microns using standard molybdenum
Spindt emitters, or a tip coating of some very low work function material on
1.0 micron diameter emitters. The small
diameter approach recognized that these emitters had to be made without any
high-resolution semiconductor lithography. (This conclusion was based on
Spindt’s modeling.)
4.
Field
emission is noisy. For a satisfactory
display, there should be a large number of emitters for each pixel. At the
desired resolution of about 100 lines per inch, the small gate approach gave
room for some 3000 emitters. This gave
a statistical reduction in the inherent noise to a satisfactory level. (Spindt
modeling)
5.
The
emitted electrons come off the tip with a lateral velocity and must be focused
to obtain satisfactory performance with a high voltage display. This can be done with an additional
electrode on the cathode surface. (Spindt modeling)
6.
The
structure must incorporate a vertical resistor behind each emitter to limit
runaway emission and resultant damage. (Spindt and Fahlen planning).
7.
The
thin structure must be evacuated, sealed, and gettered to have a long life.
(Duboc ideas and structures)
A
cold cathode tube appeared to be feasible if a low cost way could be found to
make the sub-micron gates. SVC made
this gate project their highest priority.
A variety of approaches based upon ion tracking in insulators demonstrated
the possibility making these sub-micron Spindt emitters without expensive
steppers. A specific procedure using
ion tracking invented by John Macaulay was shown to work. The rest of the requirements then required
more perspiration than invention and Silicon Video proposed to HP that this
display could meet all of HP’s notebook requirements.
After
reviewing the results of the FED work at SVC, Hackborn convinced HP to provide
additional funding to SVC. Not only
did HP provide money, it also gave SVC technical help from HP Labs and helped
SVC later to obtain more financing from venture capitalists.
While
the molybdenum tips are sufficient for a competitive display, any reduction of
the tip work function improves the overall system efficiency. There were occasional reports of “negative
work function” in some forms of diamond emitters, but a larger problem was how
to measure and or determine the work function of various materials and
coatings.
Lincoln
Labs had suggested that DARPA fund a theoretical analysis of cold cathode
emitters for displays. DARPA asked
Silicon Video to supervise this and compare and evaluate all known approaches
to building FEDs as an extension of SVC’s first DARPA grant. Duboc led this evaluation and after some six
months, SVC determined that the best direction was to proceed with the
molybdenum tip approach as none of the diamond type surfaces were yet
reproducible or had any definable path to low cost, large area displays.
.
SVC
continued to require equity investment and approached, among many others, J.H.
Whitney. Harry Marshall, a general partner
at the venture capital firm of J.H. Whitney, did not convince Whitney to
invest, but left Whitney to become an investor in SVC in 1992. Prior to working
at J.H. Whitney, where he was a principal in 15 technology start-ups, Marshall
was a senior partner at Hambrecht and Quist, a well-established Silicon Valley
investment firm. In 1993, Marshall came
on board as President and CEO of SVC.[44] He led the fund raising for the C and later
rounds of financing.
In
May 1994, Advanced Technology Materials Inc. (ATMI) and SVC received a $1.1
million award from DARPA for optimizing methods for building FEDs using diamond
microtips. It was hoped that the depositing of thin films of diamond materials
on cold cathodes would lower power consumption. This work resulted in significant intellectual property, but
still no manufacturable diamond coated tips, due to the difficulties of working
with the diamond materials.
In
the spring of 1994, Compaq Computers invested in SVC. [45] Wyse technology of Taiwan also invested $1.8
million in March 1994.[46]
In
September 1994, SVC leased a disk drive manufacturing facility in south San
Jose previously occupied by Dastek and started to install a display fabrication
line based on 4” semiconductor fabrication equipment. In October 1994, SVC
received a $22 million grant from the Technology Reinvestment Program (TRP)
administered by DARPA to further develop its FEDs. The TRP funded about a third
of the total of $67.2 million SVC received, which also included matching funds
from the State of California and private investments.[47]
By
January 1995, the San Jose plant was ready to begin producing prototype FEDs on
4-inch glass circles. In March 1995, the first prototypes were actually
produced, but these displays were not "sealed tubes." It was not
until June that the first sealed tubes came off the line. SVC first showed its
3-inch prototype display publicly at Display Works in San Jose in January 1996.
Lithography
equipment for the plant was purchased from Dainippon Screen. SVC received assistance
from the US Display Consortium (USDC) in purchasing holders, scribers, and
markers, but the firm had some difficulty purchasing tracking equipment
(Dainippon would not build them in the United States).
SVC
continued working very closely with one of its major partners, Hewlett-Packard,
to complete this FED tube technology and characterized it with a new name, "Thin CRT."
Unlike
the PixTech FED, which required about 80V on the gate for switching pixels, the
SVC display required only 10V. Instead of using emitters that were about 1
micron tall and gate holes with a 1-micron diameter (as in the PixTech design),
the SVC device used emitters that were 0.2 microns tall with 0.15 micron gate
diameters. Instead of using 200
emitters per pixel, the SVC device used >1,000 emitters per pixel. The anode
voltage for the SVC device was around 5,000V as compared with 350V for the
PixTech device. The gap between the anode and the cathode was 50 mils (as
compared with the 4 mils of the PixTech device). This meant that there had to
be a way of focusing the electrons emitted from the cathode before they reached
the anode and that spacing between anode and cathode had to be highly uniform.
SVC developed its own "spacer" technology using ceramic strips that
were 2 mils thick and 50 mils tall. Using this distinctive design, it was
possible for SVC to build a display that could use regular high-voltage (P22)
CRT phosphors -- a real advantage because of their lower cost and longer
expected lifetime as compared with the low-voltage phosphors used by PixTech. [48]
Manufacturing
ThinCRT’s in the San Jose plant initially involved 250 separate process steps
and 80 different tools. About 80 percent of the tools were the same as those
used in a TFT LCD plant. The faceplate (anode), spacers, and cathode were
processed separately and then assembled. By June 1996, SVC was sealing about
15-20 tubes per week. One primary goal for the San Jose plant was to reduce the
number of process steps from 250 to around 110 and increase the yield. SVC hired David Bergeron from IBM in January
1995 to be Vice President in charge of manufacturing technology. Bergeron
brought with him extensive experience in semiconductor manufacturing. He had
worked previously on the IBM-Siemens-Toshiba joint venture for DRAMs and on the
joint venture between IBM and Motorola to make a microprocessor for Power PCs.
Bergeron helped to formulate a "concurrent engineering" approach at
SVC -- which meant that the firm would work on product and process technologies
simultaneously.
A
second fabrication line was installed in the San Jose plant to produce
ThinCRT’son 32cm by 34cm glass plates (commonly called "mother glass"
because more than one display could be produced on a single sheet of glass).
SVC continued to use a stepperless lithographic process in order to increase
throughput. After processing the mother glass, SVC would cut up the panel into
smaller pieces for further processing as 5.3” diagonal engineering prototypes.[49]
Accu-Fab Systems provided assembly equipment for the prototypes. These
prototypes would be sold as engineering kits to companies interested in
developing FEDs for final products.[50]
One
goal for the second San Jose plant was to reduce the number of process steps
from 250 to 110. SVC hired David
Bergeron from IBM in January 1995 to be Vice President in charge of
manufacturing technology. Bergeron
brought with him extensive experience in semiconductor manufacturing. He had worked previously on the
IBM-Siemens-Toshiba joint venture for DRAMs and on the joint venture between
IBM and Motorola to make a microprocessor for Power PCs.[51]
Bergeron helped to formulate a "concurrent engineering" approach at
SVC -- which meant that the firm would work on product and process technologies
simultaneously.
--------------------------
SVC
Rounds of Financing
1st:
1990, $1 million from DARPA; $400,000 from Robert Pressley and Capform
ventures.
2nd:
1992, unspecified amount from Hewlett-Packard and Carl Berg.
3rd:
March 1994, $10 million: New Enterprise Associates & Affiliates and
Sevin-Rosen Management; new investors owned 40% of the stock
4th:
May, 1995, $17 million:
5th:
May 1996; $55 million: Bankers Trust, Berger Associates, BKP Partners, Compaq
Computers, Citicorp, Hewlett-Packard,
J.P. Morgan Investment Management, New Enterprise Associates &
Affiliates, and 21st Century Communication Partners
6th:
May 1998: $125 million: sale of senior subordinated convertible debentures
other
earlier investors:
Bayview
Investors
Farmers
Insurance and Affiliates
Oppenheimer
& Affiliates
Sierra
Venture Affiliates
Century
Financial Partners, Inc.
John
Govett & Co.
Leach
Capital
Rock
Rimmon Securities
-------------------------------
On
June 22, 1996, SVC changed its name to Candescent Technologies Corporation
(CTC). The press release for the name change explains it as follows:
The name, taken from the Latin word candere, "to shine, highlights the
most prominent and appealing features of the company's novel Thin CRT display
technology -- brightness and vibrant colors that are comparable to the cathode
ray tubes found in televisions and desktop computer monitors. "Candescent" was chosen to convey
the ideas and innovation that are necessary to compete with liquid crystal displays
(LCDs), the dominant display technology found in notebook computers today.[52]
The
press release mentioned the CTC strategy of targeting the notebook display
market in competition with TFT LCDs.
This strategy was criticized by a number of observers as unrealistic,
given the growing competition and declining prices of TFT LCDs, especially
after the entry of Korean firms in 1995.[53] Nevertheless, senior managers of CTC
continued to assert their intention to compete in notebook markets eventually
on the basis of the expected superior performance of Thin CRTs for multimedia
applications.
In
May 1997, CTC and Schott Corporation announced a joint development and supply
agreement under which Schott would establish a new high-volume glass processing
facility to supply glass for Candescent's Thin CRT fabrication facility in San
Jose. As part of the agreement, Schott
purchased ___ percent of the equity of CTC for ___ dollars and CTC agreed to
use Schott as its primary supplier of glass.
The two firms agreed to co-develop glass processing and finishing
technologies.[54]
On
May 18, 1998, CTC announced that it had raised $125 million through the sale of
"Senior Subordinated Convertible Notes." The seven percent notes were due in the year 2003 and were
convertible into common stock at a conversion price of $13 per share. All of
the money would be used in financing a high-volume manufacturing facility in
San Jose. The new 340,000 square-feet production facility would cost around
$400 million and would be capable of producing 12-15 inch displays on 59cm by
67cm glass substrates for notebook computers.
It would not be completed until late 1999 or early 2000.[55]
In
November 1998, CTC announced an agreement with Sony Corporation for joint
development of a 14-inch diagonal FED by the year 2000. Both companies pledged to spend $50 million
on this effort. Most of the work would
be performed at CTC's plants. A team of
six Sony engineers were sent to San Jose to begin the work, with some
additional staff dedicated to the project in Japan.
In
short, by mid 1999 Candescent was the only viable competitor to PixTech and its
allies in FEDs.
Conclusions
Key
elements necessary for surviving the very difficult process of converting a new
and promising technology into a commercial product are:
n
access
the right technologies at the right time
n
adequate
control over the necessary intellectual property
n
a
combination of strong management skills and technical excellence in senior
managers
n
access
to patient capital during the period before products can be commercialized, and
n
a
convincing and viable strategy for managing product and process technologies so
that key investors can remain patient.
PixTech
and its partners and Candescent used rather different designs for products and
pursued different marketing goals.
PixTech opted for fast commercialization of high gate, low anode voltage
FEDs, while Candescent opted for somewhat slower commercialization of low gate,
high anode voltage devices. PixTech
never explicitly targeted the notebook market, while Candescent did so from its
very beginning.
PixTech
was an international alliance from the start with its base in France but strong
links to both Japanese and North American firms. Candescent had a more nationalistic focus, but did not limit
itself to working only with other U.S. firms, even though a sizeable proportion
of its R&D spending was financed by the Department of Defense.
The
key similarity between the two firms was their ability to early combine
government funding (from the French government in the case of PixTech; from the
U.S. government in the case of Candescent) with venture capital to assure that
they would be able later to move to high-volume production. Neither firm became a "hot house
flower" dependent solely on government subsidies.
We
do not know, of course, whether either firm will survive in the long run. Nevertheless, the history of the FED can
provide useful lessons about competition in high technology industries in a
globalizing world economy.
List of References
Bergeron,
David, "Out of the Lab Into the Fab: Successful Technology Introduction
into Manufacturing," presentation at the USDC Business Conference, Display
Works '96, San Jose, California, February 6-8, 1996
Brodie,
Ivor, "Keynote Address to the First International Vacuum Microelectronics
Conference, June 1988: Pathways to Vacuum Microelectronics, " IEEE Transactions on Electron Devices,
v. 36, No. 11 (July 1988), part 2, pp. 2637-40.
Brodie,
Ivor, "Physical Considerations in Vacuum Microelectronics Devices," IEEE Transactions on Electron Devices,
v. 35, No. 11, pt. 2 (November 1989), pp. 2641-2644.
Brodie,
Ivor, and Charles A. Spindt, "Vacuum Microelectronics," Advances in Electronics and Electron Physics,
vol. 83 (1992), pp. 1-106.
Brodie,
Ivor and Charles A. Spindt, "Vacuum Microelectronics," in P.W. Hawkes, ed., Advances in Electronics and Electron Physics (New York: Academic
Press, 1992).
Busta,
Heinz H., "Review: Vacuum Microelectronics - 1992," Journal of Micromechanical Microengineering,
v. 2 (1992), p. 43.
Cathey,
David A., "Field-Emission Displays," Information Display, No. 16 (October 1995),
Dance,
Brian, "Europe's FPD Development Offers a Chance to Compete," Semiconductor International, v. 229
(July 1995),
Derbyshire,
Katherine, "Beyond AMLCDs: Field Emission Displays?" Solid State Technology (November 1994),
55-65.
Ellis,
Roger K., "Field Emission Displays," Seminar Lecture Notes, Society for Information Display, Hynes
Convention Center, Boston, Mass., May 12 and 16, 1997, v. 1 (May 12), pp. M-8/3
to M-8/35.
Fowler,
R.H. and L.W. Nordheim, "Electron Emission in Intense Electrical
Fields," Proceedings of the Royal
Society of London, v. 119 (1928), p. 173.
Ghis,
A., Robert Meyer, P. Rambaud, F. Levy, and T. Leroux, ""Sealed Vacuum
Devices: Fluorescent Microtip Display," IEEE Transactions on Electron Devices, v. 39, No. 10 (October
1991), pp. 2320-2322.
Gomer,
R., Field Emission and Field Ionization
(New York: American Institute of Physics, 1993).
Gossling,
B.S., "The Emission of Electrons under the Influence of Intense Electric
Fields," Philosophical Magazine,
v. 1, No. 3 (March 1926), pp. 609-635.
Holland,
C.E., Charles A. Spindt, Ivor Brodie, J.B. Mooney and E.R. Westerberg,
"Spindt Cold Cathode Vacuum Fluorescent Display," presented at the
First International Vacuum Microelectronics Conference, Williamsburg, Virginia,
June 13-15, 1988.
Iannazzo,
S., "A Survey of the Present Status of Vacuum Microelectronics," Solid State Electronics, v. 36, no. 3
(1993), pp. 301-320.
Itoh,
Junji and Seigo Kanemaru, "Adding 'Intelligence' to Field-Emission
Displays," Information Display,
v. 14, No. 8 (August 1998), pp. 12-14.
Kesling,
W. Dawson, Field-Emission Device Modeling
and Application to Flat-Panel Displays, Ph.D. Dissertation, U.C. Davis,
1995.
Levine,
Jules D., "Field Emitter Displays," presentation made at a conference
on Advanced Flat Panel Display Processing and Research, San Jose Convention
Center, June 21, 1995.
Levine,
Jules D., "Benefits of the Lateral Resistor in a Field Effect
Display," Journal of Vacuum Science
and Technology, v. 14, No. 3 (May 1996), p. 2008.
Lucas,
Mark, "Lithography for FED Production," Information Display, v. 13, Nos. 4&5 (April/May 1997), pp.
20-23.
Marshall,
Harry, "Candescent Technologies: Corporate Update," presentation at
the USDC Business Conference, Display Works '98, San Jose, California, January
20, 1998.
Meyer,
Robert, A. Ghis, Ph. Rambaud, and F. Muller, "Development of a Matrix
Array of Cold Cathode Emitters on Glass Substrate for Flat Panel Display
Applications," presented at the First International Vacuum
Microelectronics Conference, Williamsburg, Virginia, June 13-15, 1988.
Meyer,
Robert, "Recent Developments on Microtips Display at LETI," Technical
digest of IVMC '91, Nagahama, Japan.
Meyer,
Robert, "Color Field Emission Display: State of the Art and
Prospects," Eurodisplay '93,
Strasbourg, France, pp. 189-192.
Meyer,
Robert, A. Ghis, P. Rambaud, and F. Muller, "Microtips Fluorescent
Display," Japan Display, 86
(1986), p. 513.
Nicolaescu,
Dan, "Vacuum Microelectronics Devices: Computer Aided Modeling and
Design," National Institute of Microtechnology, Bucharest, Romania,
accessed via the Internet at http://www.imt.ro/InternationalProjects/internal_proj/vmrom.htm
on April 16, 1999.
Schwoebel,
Paul R. and Charles A. Spindt, "Glow Discharge Processing to Enhance
Field-Emitter Array Performance," Journal
of Vacuum Science and Technology, v. 12, No. 4 (July/August 1994), pp.
2414-2421.
Shoulders,
Kenneth R., "Microelectronics Using Electron Beam Activated Machining
Techniques," in F.L. Alt, ed., Advances
in Computers ( New York: Academic Press, 1961), pp. 135-293.
Smith,
Robert T., "Electronics Developments for Field-Emission Displays," Information Display, v. 14, No. 2
(February 1998), p. 12.
Spindt,
Charles A., "A Thin Film Field Emission Cathode," Journal of Applied Physics, v. 39, No. 7
(1968), pp. 3504-3505.
Spindt,
Charles A. and Ivor Brodie, "Molybdenum Field-Emitter Arrays," International Electron Device Meeting
'96, pp. 289-292.
Spindt,
Charles A., Ivor Brodie, L. Humphrey, and E.R. Westerberg, "Physical
Properties of Thin-Film Field Emission Cathodes with Molydbenum Cones" Journal of Applied Physics, v. 47, No.
12 (December 1976), pp. 5248-5263.
Spindt,
Charles A., Christopher E. Holland, A. Rosengreen, and Ivor Brodie, "Field
Emitter Arrays for Vacuum Microelectronics," IEEE Transactions in Electron Devices, v. 38, No. 10 (1991), pp.
2355-2363.