Transistors and Cold Fusion Part 1
(Originally Published May-June,
1999 In Infinite Energy Magazine Issue #25)
by Jed Rothwell
Much of this paper is based on the book
Crystal Fire,1 a good introduction to the history
The history of transistors teaches many lessons about
how cold fusion might develop and what should be done to help it
Transistors are physically similar to cold fusion
devices. In fact, some of the earliest experimental transistors
were immersed in electrolyte with a counter electrode to neutralize
the surface barrier.2 Transistors and cold fusion cathodes
are both small, low- temperature, solid state crystalline devices
that replace large, hot, plasma or vacuum-state devices triode
vacuum tube amplifiers and tokamak reactors. Many of the specific
hurdles overcome by early transistor researchers are directly applicable
to cold fusion, especially problems with reproducibility, contamination,
and materials. George Miley3 and others think that commercial
cold fusion cathodes may be fabricated by using modified semiconductor
manufacturing equipment, especially thin-film or electroplating
apparatus. Cold fusion cathodes do not require precise placement
of components the way integrated circuits do, but they do require
precise control of materials and composition, extreme cleanliness,
automated production, and packaging in plastic or ceramic containers
to exclude contamination.
Transistors and cold fusion were developed after decades
of theoretical speculation, false starts, and precursor devices.
Paneth and Peters conducted cold fusion experiments with palladium
in the 1920s; J. E. Lilienfield received a patent for a semiconductor
field-effect amplifier in 1930, which probably would not have worked.
Development of both transistors and cold fusion was delayed for
years because there was no broad theory to guide research, and basic
questions remained unanswered. Until 1952 it was not clear whether
the transistor effect occurs in the bulk of the material or on the
surface, and this is still a major unanswered question in cold fusion.
Fleischmann believes the effect occurs in the bulk, but most other
scientists say it happens at the surface or near surface layers.
They point to evidence such as damage found near the surface and
the fact that helium and tritium are generally found in the effluent
gas, not trapped in the metal lattice.
Cold fusion has been far more controversial than semiconductors,
and more difficult to replicate, but semiconductors did cause controversy
during the first twenty-five years of development. In 1931, Wolfgang
Pauli said: 4
I don't like this solid state physics . .
. though I initiated it . . . One shouldn't work with semiconductors,
that is a filthy mess; who knows whether they really exist.
Problems with Contamination and Reproducibility
Difficulties with contamination have always plagued
the semiconductor industry, and they are major problem in cold fusion.
Minute quantities of impurities, called dopants, must be added to
silicon to make a transistor. Other impurities must be rigorously
excluded, because they poison the reaction. It is likely that similar
dopants enhance the cold fusion effect.
Before there was a theory, the only way to begin learning
how to make an effective semiconductor amplifier was to look at
samples of solid state material that did things similar to amplification,
like rectifiers and photovoltaic converters. Subtle differences
in materials were important. Copper oxide rectifiers worked best
when made with copper from a particular mine in Chile. In 1938,
a piece of silicon that acted as a photovoltaic device was discovered
fortuitously. The chemical makeup of this sample was investigated
with the best mass spectroscopy available at the time. Since cold
fusion scientists have no effective theory, they should concentrate
on examining and analyzing materials from effective cathodes.
One of the myths spread by opponents of cold fusion
is that soon after things are discovered, they become easy to reproduce.
Transistors were extremely difficult to reproduce for many years.
One scientist recollected, "In the very early days the performance
of a transistor was apt to change if someone slammed a door." In
the mid-1950s transistors cost $16 apiece compared to the $3 vacuum
tubes they were designed to replace. Integrated circuits were even
worse. In the 1980s, after three decades of the most intense high-tech
R&D in history and hundreds of billions of dollars of investment
in transistor technology, more than half of the computer chips coming
off production lines in the U.S. were defective and had to be scrapped.
(Production rates were better in Japan.) In the late 1990s, 10 to
20% of chips fail.5 I will have more to say about this in Part 2.
Riordan and Hoddeson describe contamination problems
seven years after intense semiconductor research began: 6
For several years during the early to mid-1950s,
Shockley (and others) spoke of a mysterious class of substances
that he dubbed "deathnium," which somehow crept into semiconductors
and acted as traps for holes, gobbling them up and further shortening
their already-too-brief lifetimes. After much consternation and
head scratching, trace atoms of copper were finally identified as
one of the culprits. They were thought to have found their way from
laboratory doorknobs to germanium surfaces on the unwashed hands
of unwitting technicians!
An Educated Guess About Crystal Grain
Most metals are made up of randomly oriented
microscopic crystals (also called crystallites or grains). Metal
with many small grains is harder than metal with fewer, larger grains.
A blacksmith heats iron to reduce the number of grains by melting
them together, making the iron malleable. Then he pounds it with
a hammer to increase the number of grains, making it more amorphous
or polycrystalline. Some materials, such as gems, are made of a
single large crystal. In 1948 it was discovered that a transistor
made from a single crystal of germanium, with no grain boundaries,
worked much better than amorphous germanium, because grain boundaries
interfere with the transport of electrons. Cold fusion works best
the other way. Recent work by De Ninno et al. shows that
cold fusion cathodes work better with small grains, each no larger
than 50 microns across, apparently because the grains load more
deuterium without fracturing from the buildup of mechanical stress.7
De Ninno concludes that variations in grain size and shape cause
large differences in performance, and this factor alone may explain
all of the variability in performance. This did not surprise experienced
electrochemists. During ICCF7, the NEDO Japanese researchers described
how they had made great efforts to fabricate large grain cathode
metal. Robert Huggins heard the presentation and said to me, "That's
a splendid effort but it is just the opposite of what they should
be doing." Huggins prepared his first successful cold fusion cathodes
in 1989 by pounding the metal with a hammer to make it more amorphous.8
At Bell Labs in 1948, Gordon Teal made an educated
guess that a single crystal germanium would be needed for predictable,
effective transistors. Teal thought he could develop a process to
fabricate single crystals of far greater purity than any then in
existence. William Shockley, the head of transistor research, was
headstrong and did not appreciate how important this was. At first
he paid no attention to Teal's work. Teal had to pursue the project
at night with "bootlegged" equipment. He would unplug his apparatus
every day, roll it into a closet, and work on his official assigned
task instead, which is how many cold fusion experiments are performed.
In 1949 Teal gave a sample of his germanium crystals
to a chemist in the newly-formed semiconductor research group. The
chemist tested the electron and hole mobility of the crystal and
found that performance was twenty to one hundred times better than
that of the conventional, polycrystalline samples. "As word of this
success percolated through the semiconductor group, Shockley finally
began to sit up and take notice of Teal's work. By late 1949 he
had to admit he had been wrong. . . . Soon Bell Labs would have
an entire group devoted to growing single germanium crystals."9
A few years later Bell Labs developed zone refining, which made
materials 1000 times purer than any previous technique. This alone
would have been sensational, even without transistors. It was one
of many breakthroughs needed to make transistors practical. The
transistor was not "one" innovation; it was the culmination of a
series of related and directed innovations triggered by the 1948
breakthrough. You could not make a practical transistor without
ultra-pure single crystal germanium (later, silicon), but no one
thought to make ultra-pure single crystal germanium until the first
crude, unreliable prototype transistor was demonstrated. No one
investigated small grain palladium in detail until many years after
1989. Cold fusion will not succeed until many other related, directed
innovations are undertaken to support it, which will cost huge amounts
The early history of transistors teaches many
lessons about how science works and what to expect in the early
stages of groundbreaking research.
Prototype inventions are often crude, unreliable,
and makeshift. The first point contact transistor was held in place
with a spring fashioned from a paper clip, which literally pushed
a point down to keep it in contact with germanium. Kilby described
the first integrated circuit: "It looked crude, and it was crude."
Riordan and Hoddeson add, "These prototypes were extremely awkward
realizations of the much more sophisticated ideas he penned into
his notebook two months earlier. But the first prototype of an important
technological idea is often crude witness the first transistor."10
Some experiments are too crude. Shockley was a theoretician,
not an experimentalist. One day in 1940, a scientist named Wooldridge
found him fiddling around in the lab with a piece of oxidized copper,
which "had apparently been cut out of some very old copper back
porch screen with very dull scissors." Shockley was trying to position
wires so they would barely touch the green oxide coating. He hoped
to adjust the voltage applied to the mesh to control the current
flow. In other words, he was trying to make a crude transistor.
Wooldridge later wrote: "So here he had the three elements of a
transistor, these two wires and the copper screen. Of course, he
was orders of magnitude away from anything that would work!"11
Most of the 1989 "replications" of cold fusion were equally laughable.
Critical electrochemical parameters like loading and open circuit
voltage were not measured; materials were not analyzed before or
after the run. This is flailing in the dark, not science. You might
be orders of magnitude away from anything that will work, or you
might be on the verge of success. You have no way of knowing, and
even if you do achieve success, you will be unable to replicate
Scientists must pay close attention to inexplicable
phenomena which may look like instrument errors at first. They should
not dismiss weird, marginal, unexpected, anomalous phenomena. In
the 1920s, silicon crystal radio detectors developed a bad reputation.
"Variability, bordering on what seemed the mystical, plagued the
early history of crystal detectors and caused many of the vacuum
tube experts of a later generation to regard the art of crystal
rectification as being close to disreputable."12 In 1938 a scientist
named Becker at Bell Labs tried to measure the conductivity of silicon
rods. When the probe was moved from one spot to another, conductivity
varied wildly. One of the rods was "so erratic that no consistent
values could be reported." The rod was put aside and forgotten for
a year until Russell Ohl examined it carefully and discovered that
it acted as a photovoltaic cell, converting light into electricity.
This was the first time anyone had seen the photovoltaic effect
in silicon. It was the real start of AT&T's research in semiconductors.
Ohl later said that Becker's career suffered:13
He had that active silicon in his department,
in his hands, and he didn't find it . . . That is what you are up
against in research. You've got to watch for things like that, for
something unusual. If that happens, you have got to learn to recognize
Not eloquent perhaps, but it expresses an essential
truth which many scientists pay lip service to, while in practice
they ignore. Many scientists have ignored evidence for cold fusion.
In a few cases they have actively tried to bury it.
You must be careful not to fool yourself. After the
triumphant in-house demonstration of the first functional transistor
in 1948, one of the observers cautioned the discoverers: "Look boys,
there's one sure test of an amplifier, that you aren't kidding yourselves.
An amplifier, if fed back on itself with a proper circuit, will
oscillate. This shows that it is really producing power more
than you put into it."14 The next day, Brattain performed
the feedback test and confirmed that the device did oscillate. Many
cold fusion scientists have failed to perform simple calorimetric
tests which would reveal that their devices are "really producing
power" and other simple tests such as autoradiographs, which confirm
that the effect really is nuclear.
Scientists should cooperate. While manufacturing
radar sets and other electronics during World War II, AT&T learned
the value of close cooperation between theoreticians, experimentalists,
and production line workers. The economic boom and postwar office
space crunch helped to prolong this cooperation. Bell Labs rushed
to hire many new scientists while it completed a new laboratory.
When Bardeen joined Bell Labs in 1945 "office space was extremely
scarce . . . So employees were being asked to double up . . . Bardeen
didn't mind; he liked the company of experimentalists. Here was
an opportunity to glance over their shoulders and talk about the
data as they collected it."15 This spirit of cooperation
was essential to the rapid development of transistors.
Success Was an Accident
Success in research is often the unlikely result
of a series of accidents. Consider some of history's might-have-beens.
Gordon Teal worked at night on his "bootleg" crystal-growing experiments
and during the day on his regular assignment. His wife grew upset
at this overwork, and asked him to cut back. She might have prevailed,
or he might have grown discouraged and burned out on his own. Or
he might have missed the opportunity to show the chemist his single
crystal sample. Shockley might have remained characteristically
obstinate, continuing to ignore Teal's research. This one oversight
by Shockley might have held back the development of transistors
for years. Thousands of technical decisions and choices must be
made in the course of developing a commercial product, and each
might be a wrong turn or a dead end. That is why research must be
done by many different independent laboratories, at different corporations
and universities. One person or one funding agency committee cannot
be placed in charge. One person, no matter how brilliant, may guess
wrongly and lead the whole project into a dead-end. Competing ideas
must be tested, even ideas the experts consider crazy.
In their Epilogue, Riordan and Hoddeson describe the
mix of personalities and institutions needed to bring about the
None of these men [Shockley, Bardeen and
Brattain] could have invented the transistor alone. But their lives
intersected at a unique American institution during a peculiar moment
in history to make it possible, even likely. Nothing on the scientific
landscape at the time compared with Bell Labs. It combined intellectual
power equal to that of the nation's best science departments with
technical resources and manpower that none of them could come close
to matching. When these tremendous resources became focused on developing
practical products based on wartime advances in semiconductor technology,
something big had to happen . . . Each man's shortcomings were compensated
by the others in this multidisciplinary environment. With his single-minded
focus on "trying simplest cases first," Shockley would never have
conceived the unwieldy point-contact gadget that opened the door
to the transistor . . . . . . Almost as important as the transistor's
invention are the techniques of crystal growing and zone refining,
which allow one to fabricate large single crystals of ultrapure
silicon and germanium. Without these crystals, the industry would
This is contradictory. The mix of personalities was
unlikely, the postwar boom was a "unique moment in history" which
we hope will never be repeated (if it takes a war to trigger such
a moment), yet "something big had to happen." The invention was
unlikely yet inevitable. Was the transistor truly inevitable? Where
would we be without it? Is any innovation inevitable and unstoppable?
I will examine these issues in Part
- M. Riordan, L. Hoddeson, Crystal Fire: The Birth
of the Information Age, (Norton, 1997).
- Riordan and Hoddeson, ibid., p. 130 Private
- L. Hoddeson, E, Braun, J. Teichmann, and S. Weart,
Out of the Crystal Maze, (Oxford Univ. Press, 1992), p.121.
- S. Lohr, "Suiting Up for America's High-Tech
Future," New York Times, December 3, 1995.
- Riordan and Hoddeson, ibid., p. 219.
- A. De Ninno, "Material Science Studies Aimed at
Improving the Reproducibility of the Heat Excess Experiments,"
Proc. ICCF7, p. 103.
- M. Schreiber et al., "Recent Measurement
of Excess Energy Production in Electrochemical Cells Containing
Heavy Water and Palladium," Proc. ICCF1, Table 1, 2 and 3, p.
- Riordan and Hoddeson, ibid., p.179.
- Riordan and Hoddeson, ibid.,
- Riordan and Hoddeson, ibid., p. 86.
- Riordan and Hoddeson, ibid., quoting F.
Seitz of AT&T, p. 92.
- Riordan and Hoddeson, ibid., quote from
Ohl, p. 96.
- Riordan and Hoddeson, ibid., p. 140. Note,
however, that an amplifier does not literally produce power. It
transfers energy from the power supplies to the output, increasing
or decreasing power depending on the control current. All energy
originates in the electric power supplies. With cold fusion the
energy originates in the cell. Cold fusion cells sometimes resemble
amplifiers. When the cathode current density increases, deuterium
loading usually increases, and after a while the cold fusion power
may increase proportionally, assuming other control parameters
remain the same.
- Riordan and Hoddeson, ibid.,
- Riordan and Hoddeson, ibid., p. 280