You are on page 1of 366

Electron Multiplication through Secondary Electron

By Bruce Darrow Gaither

In this era of energy shortages we have all daydreamed about owning a device which
would take the energy that we have and multiply it. Imagine how happy we would be if
you could simply plug in a device which would double your electricity. Numerous
researchers, writers and inventors have sought to do just that.

But scientists have cautioned that the law of conservation of energy dictates that energy is
never created nor destroyed, only converted from one form to another. None of these
schemes, they say, would ever work. They are just perpetual motion machines.

The purpose of this book is to discuss advances in electronics and materials science
which have made things possible which were not contemplated when the laws of
Thermodynamics were postulated decades ago.

Secondary electron emission is a well-known process. It is that effect which causes
additional electrons to be emitted when a substance is bombarded by a stream of
electrons. This secondary emission effect was discovered a century ago, and it has found
application in a variety of devices which are in use today.

Advances in electronics and the development of new materials have revealed new
methods and substances which make this secondary electron emission even more
effective. In fact, today the impossible is possible—one electron at a time.

If a beam of electrons is aimed at a target electrode coated with a given substance then
electrons are emitted from that target. The number of electrons emitted from the target
which has been bombarded is compared to the number of primary electrons in the
original electron beam. The electrons emitted after bombardment by primary electrons
are called secondary electrons.

The material’s propensity to emit electrons after bombardment is called the “secondary
electron emission coefficient”. That is expressed as the number of secondary electrons
divided by the number of primary electrons.

A secondary electron coefficient of less than 1 means that the substance does not emit as
many electrons as it is bombarded with. A coefficient of greater than 1.0 means that the
substance emits more secondary electrons than bombarded it

We are interested in this book in those materials which exhibit a high coefficient of
secondary electron emission. The goal is to perfect a device which will emit more
electrons than you started out with. Some devices have been invented to minimize or
eliminate secondary electron emission, but those are not within the scope of this book.

Years of testing and research have resulted in well-known coefficients of secondary
electron emission for scores of materials. Research first focused upon basic substances
such as copper, steel, silver, gold and other metals. Then more exotic combinations and
alloys were tested, such as oxides and alkalai metal combinations. Some metals and some
combination coatings worked much better than others.

In recent years miniaturization and nanotechnology as well as chemical vapor deposition
of thin films have come up with even more effective secondary electron multipliers.
In fact today’s plasma televisions and LED lights are examples of such substances and

A reference to various charts is now possible to determine at a glance exactly what the
secondary electron emission coefficient is for these materials.

But the devil is still in the details.

Two variables are of importance as to the materials. One variable is the angle at which
the primary electrons impact the target. In general it has been found that a better result is
obtained when the primary electrons graze the target material at an angle rather than
hitting straight-on. But the physics of the secondary emission process and the atomic and
molecular structure of the target materials mean that the optimum angle is different for
different materials. The second main variable is the voltage of the beam of primary
electrons. Some substances max out their secondary emission coefficient at very low
voltages and some materials reach their highest secondary emission only at much higher
voltages. This is believed to be due to the depth to which the primary electrons impinges
the material and the amount of tunneling into the substance.

Therefore care must be be paid to the research of the variables in the individual research
papers and patents and there is no magical formula which will work for all materials. The
configuration of the device used in the secondary emission process will need to vary to
meet the characteristics of the target material.

Once the primary electrons have impacted the target material and secondary electrons are
released, then the secondary electrons do not form a focused beam. Instead they tend to
just sit there in a cloud of electrons. Meanwhile the primary electrons will usually bounce
off the target at an angle equal to the angle with which they make incidence to the target.
This means that the kinetic energy of the primary electrons is greater than that of the
secondary electrons after impact. In other words the secondary electrons are indeed a
larger number of electrons—but they are of low voltage.

However numerous methods exist to increase the voltage or kinetic energy of the
secondary electrons.
One process which is common in devices which employ secondary electron emission is
that of multiple impacts upon this emissive target material. The great inventor, Philo
Farnsworth, was the first to devise methods to facilitate these multiple impacts upon
emissive materials. He called his devices “multipactors” because of the multiple impacts
they made with secondary electron emissive material. Thus, if a target electrode had a
secondary electron emission coefficient of 2 then the number of primary electrons would
double when they hit the target electrode. If there were two successive impacts then the
primary electrons would double, and then that total would double again—or be four times
the original primary electrons put into the device.

If the primary electrons were somehow sent through a series of 8 target electrodes then
the multiplication factor would become astronomical, and each of the impacts would
result in an exponential increase of electrons based upon the coefficient of secondary
electron emission from that material.

So various devices were designed and perfected to make the primary electrons impact
numerous electrodes one after another. One branch of these devices is employed by
photomultiplier tubes. Many of these devices are capable of multiplying the primary
electrons one hundred million times. Thus minute electric currents can be sensed and
multiplied so that they can register on scientific equipment. But numerous other
configurations and devices are in use today.

One method is to bounce the electrons off of two opposing electrodes over and over again,
like a game of ping pong. Another configuration would be to have the electrons strike
electrodes arranged inside a circular tube so that they impact coated electrodes over and
over again. A third method is that of forming a cascade of specially-coated electrodes and
having the primary electrons bounce off off each successive electrode until they all come
out the end.

Another genre of devices are called “channel” devices. In these designs the primary
electrons are sent down a waveguide or tunnel of some sort and the entire length of the
device is coated with the emissive materials. The electrons keep bouncing off the walls of
these guides until they reach the end and the repeated impacts result in a high
multiplication of the primary electrons.

One of the axioms of electricity is that current will not conduct very well when exposed
to the atmosphere because the gas acts as an insulator. Therefore most secondary electron
emission devices were made in the form of vacuum tubes. The electricity goes through
the vacuum without loss and then the impacts upon emissive material have the desired

However secondary emission and multipactors have been made into semiconductors and
chips. These use the process of “avalanche” multiplication in many instances, where the
electrons hit the emissive substance and are then multiplied and pass through a solid state
stack of materials. Sometimes the semiconductors include a tiny vacuum space and they
act in the same way as a vacuum tube.
However there is a snag or two for the use of secondary emission in chips. The first
problem is “space charge”. That is the effect of completely filling up a given space with
electrons. One you saturate the evacuated space with space charge then an equilibrium
state is achieved and the primary electrons will no longer multiply as desired. As you
might guess, the greater the area the more electrons will fit into the space before the
space charge saturates the area. The space charge, then, has been found to diminish as to
the 4/3 power of the area of the evacuated space. This means that, for instance, if you
triple the size of vacuum space then that would result in 3x 4/3 power= 12/3 power (or
the 4th power). Then a tripling of space would end up in shrinking the space charge by the
4th power. The bottom line is that bigger is better.

The second drawback to micro multipactors is that the vacuums must be higher than in
vacuum tubes, and this is hard to achieve. Also many devices use sharp points as
electrodes because more electrodes will emit from sharp points than from blunt shapes. In
the micro world though the sharpness of the sharp point has to be correspondingly
sharper. The finer the point on the electrode the harder it is to fabricate and fit into the
layered semiconductor devices.

What this book hopes to achieve is not the simple multiplication of electrons to provide
light or brightness but to generate electricity on a larger scale. The aim is not to build a
generator station for thousands of people but to scale the multipactor devices to work
with individual appliances and vehicles. Thus the size and rated capacities of the
components in the proposed multipactors must be designed to be in the range of home
current up to the amount of voltage and current required to power an electric car.

At this point the discussion of secondary electron emission must include some of the
math and physics. Don’t let your eyes glaze over. Everybody knows a little bit about
electricity—and it is pretty simple. But there is a hazy horizon on the amount of
knowledge of the basics of electricity. The terms are VOLTAGE, AMPERAGE and
POWER. The easy rule of thumb is that VOLTAGE x AMPS = POWER.

You need to throw in the RESISTANCE into this formula—but for now we will stick

OK—so we will calculate one AMP. An Ampere involves the amount of “charge”, which
is calculated in terms of a COULOMB. A Coulomb is −6.24151 × 1018 electrons.
So the process of secondary electron emission results in a lot of electrons.
The secondary electrons are not moving very much after they are multiplied.
So they have low voltage—but they DO have AMPERAGE because of the presence of
lots of electrons.

The purpose of this analysis is to point out that we have low volts and high amps from
secondary emission. When you remember volts x amps = power then you can see that we
have to have just a high enough voltage to meet the requirements of modern electrical
There are numerous well-known devices which can act as VOLTAGE MULTIPLIERS.
These devices will increase the voltage, but only at the expense of a proportional
decrease in the amperage.

The end product of these multipactors can be made usable, therefore, by running these
high amperage currents through a voltage multiplier. You just fine-tune the voltage
multiplier to give the right mix of volts and amps.

In short, secondary electron emission creates NEW electrons. We put the new electrons
to use by stepping up the voltage to required levels.

In this way the laws of conservation of energy are negotiated. The multipactor creates
new electrons and THEN the energy is converted from one form into another. But there is
indeed a creation of new electrons in multipactors.

I will leave it to the scientists and inventors in their respective papers and patents to
describe the manner in which the secondary electrons are created and how the
multipactor devices are designed.

The point of my analysis is simply that the multipactors will create new electrons and the
new electrons can be made usable through voltage multipliers.

One of these voltage multipliers is a Cockroft-Walton circuit. Modern electronics has
manufactured numerous cheap transistor devices that you could get at Radio Shack or
electric supply houses. A Cockroft-Walton circuit is simply a ladder of diodes and
capacitors (pennies apiece) which double the voltage at each step of the ladder. So a
multi-step ladder creates a multiple doubling of the original voltage. Some of the older
designs apply a step-up transformer to do the same thing.

A simple Cockroft-Walton multiplier would look like this

So we see that there is a problem with voltage in secondary electron emission. The
inventors have figured out a method to use voltage to their advantage in the multipactor
devices. They apply the rule that opposite charges attract. This accelerates the cloud of
secondary electrons so that they will impact the next target with it’s emissive coating.
The electron is a negative charge. So the inventors manipulate the sluggish cloud of
negative charge by providing a positive electrode to put it into motion.
Some designs will make the next electrode one with a positive charge, so when the
primary electrons strike the first target and are multiplied then the cloud of secondary
electrons is accelerated right up to the second electrode coated with emissive materials.
When they make a long chain or cascade of these target electrodes then they give each of
them a successively higher positive charge so that the ever-increasing cloud of secondary
electrons is accelerated one step at a time in the desired direction.

Other designs use an electrode which is positioned between the first impact target and the
second and they give that intermediate electrode a positive charge to accelerate the cloud
of secondary electrons in the desired direction. This intermediate electrode might be in
the form of a screen or grid or a tube. The positive charge, in every instance, attracts the
opposite charge on the electrons and they are suddenly accelerated through the holes in
these intermediate electrodes and then the electrons continue with their increased voltage
until they impact the coated electrode. This step may be repeated again and again.

The positive charge on these attracting electrodes is often provided by using Cockroft-
Walton circuits. So either a single or a multi-step CW circuit may be used to multiply an
initial small current to give a charge bias of increasing strength to a series of attracting
electrodes. Oftentimes the CW circuit contains “taps” which tap the current at a certain
step in that multiplying step ladder. The step would then have one voltage level to apply
to the attracting electrode, and then the next step would have a higher voltage which
could be tapped at that level and applied to the next attracting electrode, and so on.

Going back the purpose of this analysis again: we are trying to get as many electrons as
possible out of the multipactor. So the gameplan is to select the coating material for
electrodes which has the highest secondary electron emission coefficient. Then the
voltage at which the primary electrons must be accelerated to achieve the optimal
secondary emission must be applied. The spatial requirements are important too because
we want the right angle and the right depth for the impact zone. So we get the highest
electron multiplication at each step. Then we take that level of electron multiplication and
exponentially multiply it by the number of impacts in the multipactor device.

Some devices, as aforestated, simply bounce the electrons back and forth between two
opposed electrodes. In these designs the electrons are moving at the speed of light, so
they hit the opposite electrode in a known length of time. Then they bounce back to the
original electrode. The desired effect is to have but one cloud of secondary electrons
bouncing back and forth, and not a lot of different clouds. Therefore the two electrodes
are given opposite charges, positive and negative, and these charges are sequentially
reversed so that the electron cloud always moves away from the first electrode after they
have been multiplied and then toward the target electrode for more multiplication. Since
we know the distance between the two electrodes and because the speed of light is known,
then we can determine the FREQUENCY at which the electric charge is reversed on
these electrodes. So, take the speed of light and divide it by the distance between the
electrodes. Say, 186,000 miles per second divided by 6 inches.
The resulting frequency is in the range of billions of cycles per second.
There are modern oscillator chips which cost pennies which can do that.

The point here is that we take the secondary electron coefficient, and let’s say that this is
2 for the sake of argument. Then we apply the frequency of the impacts on these emissive
electrodes—and that is perhaps one billion times per second. In this example we would
then obtain 2 to the one billionth power!

Are you beginning to get the picture?

If we make the device the right size so that the space charge does not saturate the vacuum
then we can generate sufficient electrons so that we can step up the voltage and step
down the amperes to achieve the desired power characteristics for our electric appliance
or motor.

For the purposes of our last example we have a secondary emission coefficient of 2, or a
doubling of the primary electrons at each impact with the electrode with the emissive
coating. But what if the secondary emission coefficient were 10…or 100…or even 1000?
Just apply the math and you can see the possibilities of these multipactors.

Attached to this anthology is one of the latest research papers from Korea where
scientists have obtained a new record for the secondary emission coefficient: 22,000!

Thus reason dictates that the proper coating must be selected for the electrodes. Then the
rest of the components must be selected and positioned so that the size, frequency and
angle of impact are optimal.

I think I heard somebody say, “Hey, Einstein—it still has to be hooked up to electricity to
start up and to power the attracting electrodes. What about that?”

The answer lies in the principle of feedback and self-oscillation. We know that many
oscillators are known to exhibit the characteristic of self-oscillation. Once you get them
going then they tend to keep on oscillating on their own. This process works in
multipactor-oscillators. It just takes a little electricity to get them started and then the
internal processes take over and they self-oscillate, producing electrons without the input
of outside electricity.

Many electronic devices apply the principle of feedback, especially in audio devices. We
can remember Jimi Hendrix hitting a note on his guitar and then holding the guitar in
front of his amplifier. The amp’s sound creates a feedback loop with the guitar and a
sound is created which is self-sustaining without the additional input of playing another
note. Numerous transistors work with feedback loops to take the electrical output of the
device and split that output and send part of it back to the original input where it is again
amplified. So the coupling of the output to the input wires is what is required.
So using either feedback or self-oscillation or both a multipactor device can be fabricated
so that it will have self-sustaining output of electrons.

That still leaves us the positive bias charge that is placed upon the attracting electrodes to
accelerate those sluggish clouds of secondary electrons.

Again, we simply split the output signal and loop part of it back to the accelerating
electrodes, and this is the positive charge remember. So the negative charge goes back to
feedback the input and the other loop goes to the voltage multiplier. A Cockroft-Walton
multiplier can be either positive or negative in charge—you simply reverse the
connection between the diodes and capacitors and it multiplies the positive charge.

Therefore, we could use batteries to start up the multipactor and then apply common
electronics components and devices to split the output and loop it back to the input and
bias the positive electrodes. Then the battery can be shut off, and even recharged while
the multipactor runs on self-sustaining current.

That guy who used Einstein’s name like a dirty word again wants to voice his opinion,
“Hey, genius, this stuff is a bunch of hooey! How do we know this would work?”

How do we know?
Because of TELEVISION.
These multipactor devices were invented by Philo Farnsworth when he invented
television. Just one glance at this super-egghead fellow should give you the answer. This
guy was a super-brain and he just NEEDED to have special vacuum tubes to strengthen
the broadcast signal of television from remote locations to make the picture tubes bright
enough to see—so he simply invented multipactors to multiply that weak input signal.
Check out that cranium!
If these multipactors work then why didn’t Farnsworth take over the whole world?
The reason is related to the laws of business and not the laws of physics. Philo
Farnsworth saw the value of television and his multipactors but he had an independent
streak which caused him to form his own Farnsworth Television company with which he
intended to put RCA and GE out of business. Instead they put Farnsworth out of business
by using monopoly tactics. But Philo Farnsworth applied his principles based upon
secondary electron emission to the point that he invented a nuclear fusion reactor before
he was through.

The heyday of vacuum tubes was filled with imitators of every sort. There is even an
International Patent classification which contains only “Farnsworth Tubes”.

Since Farnsworth’s day the vacuum tube was supplanted by the Japanese transistor and
then the Silicon Valley semiconductor chip. Nobody makes vacuum tubes anymore and
the vacuum tube multipactor concepts have been lost in the world of microelectronics.
But even today secondary electron emission is applied in the plasma television sets where
scores of little holes and dots are brightened by electron multiplication. Other areas such
as scintillation counters and electron detectors and night vision goggles use the process,
often in the solid-state configuration. There exists an offshoot applying vacuum tubes—
the sector called PHOTONICS which use vacuum tubes to multiply light into electronic

As stated above, there are several basic methods of achieving multiple impacts of
The BACK AND FORTH method:

Secondary Emission Coefficients

These graphs and excerpts were developed over a period of time. The more ancient the
research the lower the coefficients. As newer and newer materials were invented and
tested there is a general trend toward higher and higher coefficients. I would respectfully
call your attention to the source material in the following sections for detailed analyis of
the methodology and results of individual studies and devices with various emissive

Attention should be paid to the voltage required to obtain a certain coefficient of
secondary electron multiplication. The graphs are not in parallel so they are slightly
different pictures. But they should give a general idea of how much electron
multiplication could be obtained by a particular substance.
The following chapters will discuss individual studies and patents. Some of these
resource documents contain excellent discussion of historical development of the
secondary electron emission devices. It is of note that secondary electron emission was
first discovered about a century ago, and the first patent for a vacuum tube as applied for
in 1919.

The discussion also includes mention of “work factor” as an indicator of secondary
emission coefficient. The lower the work factor the higher the coefficient.

Another area of interest is that of “negative electron affinity” as an explanation for
secondary electron emission. In short, the term affinity implies that a particular substance
either likes or rejects electrons. The materials with negative electron affinity then are
predisposed to not like negatively-charged particles and thus reject them when

Treatises on vacuum tubes have been consulted and quoted in pertinent part. Patents are
inserted to this anthology to examine their significance at particular points in time.
Various studies on the individual materials exhibiting secondary emission.

Finally, I include several of my own designs for multipactor devices to power electrical
appliances and motors.
1. Introduction

The photomultiplier is a very versatile and without need for additional signal amplifica-
sensitive detector of radiant energy in the tion. Extremely fast time response with rise
ultraviolet, visible, and near infrared regions times as short as a fraction of a nanosecond
of the electromagnetic spectrum. A schemat- provides a measurement capability in special
ic diagram of a typical photomultiplier tube applications that is unmatched by other
is given in Fig. 1. The basic radiation sensor radiation detectors.
is the photocathode which is located inside a
vacuum envelope. Photoelectrons are emit-
ted and directed by an appropriate electric EARLY DEVELOPMENT
field to an electrode or dynode within the The development (history) of the photo-
envelope. A number of secondary electrons multiplier is rooted in early studies of secon-1
are emitted at this dynode for each imping- dary emission. In 1902, Austin and Starke
ing primary photoelectron. These secondary reported that the metal surfaces impacted by
electrons in turn are directed to a second cathode rays emitted a larger number of elec-
dynode and so on until a final gain of trons than were incident. The use of secon-
perhaps 106 is achieved. The electrons from dary emission as a means for signal
the last dynode are collected by an anode amplification was proposed as early as
which provides the signal current that is read 1919.2 In 1935, Iams and Salzberg 3 of RCA
out. reported on a single-stage photomultiplier.
The device consisted of a semicylindrical
photocathode, a secondary emitter mounted
on the axis, and a collector grid surrounding
the secondary emitter. The tube had a gain
of about eight. Because of its better frequen-
cy response the single-stage photomultiplier
was intended for replacement of the gas-
filled phototube as a sound pickup for
movies. But despite its advantages, it saw
only a brief developmental sales activity
PHOTOELECTRONS before it became obsolete.
Multistage Devices
Fig. 1 - Schematic representation of a photo-
multiplier tube and its operation In 1936, Zworykin, Morton, and Malter,
all of RCA4 reported on a multistage
photomultiplier. Again, the principal con-
For a large number of applications, the templated application was sound-on-film
photomultiplier is the most practical or sen- pickup. Their tube used a combination of
sitive detector available. The basic reason electrostatic and magnetic fields to direct
for the superiority of the photomultiplier is electrons from stage to stage. A photograph
the secondary-emission amplification that of a developmental sample is given in Fig. 2.
makes it possible for the tube to approach Although the magnetic-type photomultiplier
“ideal” device performance limited only by provided high gain, it had several dif-
the statistics of photoemission. Amplifica- ficulties. The adjustment of the magnetic
tions ranging from 103 to as much as 108 field was very critical, and to change the gain
provide output signal levels that are com- by reducing the applied voltage, the
patible with auxiliary electronic equipment magnetic field also had to be adjusted.

Photomultiplier Handbook

Another problem was that its rather wide Snyder 8 and by Janes and Gloverg, all of
open structure resulted in high dark current RCA. The basic electron-optics of the cir-
because of feedback from ions and light cular cage was thus well determined by 1941
developed near the output end of the device. and has not changed to the present time
For these reasons, and because of the although improvements have been made in
development of electrostatically focused processing, construction, and performance
photomultipliers, commercialization did not of the 931A product.
follow. The success of the 931 type also resulted
from the development of a much improved
photocathode, Cs3Sb, reported by Gorlich 10
in 1936. The first experimental photo-
multipliers had used a Ag-O-Cs photocath-
ode having a typical peak quantum efficien-
cy of 0.4% at 800 nm. (The Ag-O-Cs layer
Fig. 2 - Magnetic-type multistage photomul-
was also used for the dynodes.) The new
tiplier reported by Zworykin, Morton, and Cs3Sb photocathode had a quantum effi-
Malter in 1936. ciency of 12% (higher today) at 400 nm. It
was used in the first 931’s, both as a
photocathode and as a secondary-emitting
material for the dynodes.
The design of multistage electrostatically
focused photomultipliers required an
analysis of the equipotential surfaces be-
tween electrodes and of the electron trajec- PHOTOEMITTER AND SECONDARY-
tories. Before the days of high-speed com- EMITTER DEVELOPMENT
puters, this problem was solved by a Photocathode Materials
mechanical analogue: a stretched rubber
membrane. By placing mechanical models of Much of the development work on
the electrodes under the membrane, the photomultiplier tubes has been concerned
height of the membrane was controlled and with their physical configuration and the
corresponded to the electrical potential of related electron optics. But a very important
the electrode. Small balls were then allowed part of the development of photomultiplier
to roll from one electrode to the next. The tubes was related to the photocathode and
trajectories of the balls were shown to cor- secondary-emission surfaces and their pro-
respond to those of the electrons in the cor- cessing. RCA was very fortunate during the
responding electrostatic fields. Working 1950’s and 60’s in having on its staff, prob-
with the rubber-dam analogue, both J.R. ably the world’s foremost photocathode ex-
Pierce5 of Bell Laboratories and J.A. pert, Dr. A.H. Sommer. His treatise on
Rajchman6 of RCA devices linear arrays of Photoemissive Materials1 1 continues to pro-
electrodes that provided good focusing prop- vide a wealth of information to all
erties. Although commerical designs did not photocathode process engineers.
result immediately from the linear dynode Sommer explored the properties of
array, The Rajchmann design with some numerous photocathode materials-par-
modifications eventually was, and still is, ticularly alkali-antimonides. Perhaps his
used in photomultipliers-particularly for most noteworthy contribution was the
high-gain wide-bandwidth requirements. multialkali photocathode (S-20 spectral
response). This photocathode, Na2KSb:Cs,
First Commercial Devices is important because of its high sensitivity in
The first commercially successful the red and near infrared; the earlier Cs3Sb
photomultiplier was the type 931. This tube photocathode spectral response barely ex-
had a compact circular array of nine tends through the visible, although it is very
dynodes using electrostatic focusing. The sensitive in the blue where most scintillators
first such arrangement was described by emit.
Zworykin and Rajchman.7 Modifications Bialkali photocathodes were also de-
were later reported by Rajchmann and veloped by Sommer and have proven to be


better in some applications than the Cs3Sb APPLICATIONS DEVELOPMENT
photocathode. Thus, the Na2KSb photo- Astronomy and Spectroscopy
cathode has been found to be stable at higher
temperatures than Cs3Sb and, in addition, Early applications of the photomultiplier
has a very low dark (thermal) emission. It were in astronomy and spectroscopy.
has been particularly useful in oil-well- Because the effective quantum efficiency of
logging applications. Another bialkali pho- the photomultiplier was at least ten times
tocathode, K2CsSb, is more sensitive than that of photographic film, astronomers were
Cs3Sb in the blue and is, therefore, used by quick to realize the photomultiplier tube’s
RCA to provide a better match to the advantage. Furthermore, because the output
NaI:Tl crystals used in scintillation count- current of the photomultiplier is linear with
ing. incident radiation power, the tube could be
used directly in photometric and spec-
trophotometric astronomy. The type 1P28, a
Dynode Materials tube similar to the 931 but having an
The first secondary-emission material ultraviolet-transmitting envelope was par-
used practically by RCA was the Ag-O-Cs ticularly useful in spectroscopy. The size and
surface. But with the development of the shape of the photocathode were suitable for
Cs3Sb material for photocathodes, it was the detection and measurement of line spec-
found that this material was also an excellent tra and the very wide14 range of available gain
secondary emitter. Other practical secondary proved very useful.
emitters developed during the early years of
photomultiplier development were MgO:Cs Radar Jammer
(often referred to as “silver-magnesium”) A totally unexpected application for the
and BeO:Cs (“copper-beryllium”). new photomultiplier tube occurred during
In the early 1960’s, R.E. Simon12 while World War II. The development of radar for
working at the RCA Laboratories developed detecting and tracking aircraft led to the
his revolutionary concept of Negative Elec- simultaneous need for wideband electronic-
tron Affinity (NEA). Electron affinity is the noise sources as radar jammers. Although
energy required for an electron at the
other sources of noise were tried, the
conduction-band level to escape to the
photomultiplier proved to be most suc-
vacuum level. By suitably treating the sur- cessful. The advantage of the tube was its
face of a p-type semiconductor material, the high gain (107) and wide band width (several
band levels at the surface can be bent hundred MHz). As a noise source the tube
downward so that the effective electron af-
was operated with a non-modulated input
finity is actually negative. Thermalized elec- light source and with high gain. The output
trons in the conduction band are normally amplifier photoelectric shot noise was
repelled by the electron-affinity barrier; the “white” and thus indistinguishable from
advantage of the NEA materials is that these natural noise sources. This application of
electrons can now escape into the vacuum as photomultiplier tubes resulted in production
they approach the surface. In the case of of thousands per month compared with
secondary emission, secondary electrons can previous production measured in only hun-
be created at greater depths in the material dreds per year.
and still escape, thus providing a much
greater secondary-emission yield. In the case Scintillation Counting
of photoemission, it has been possible to A proliferation of photomultiplier designs
achieve extended-red and infrared sen- followed the invention of the scintillation
sitivities greater than those obtainable with counter shortly after World War II.15,16 The
any other known materials. The first prac- photomultiplier tubes were designed with
tical application of the NEA concept was to semitransparent photocathodes deposited on
secondary emission. An early paper by an end window which could be coupled
Simon and Williams l3 described the theory directly to the scintillator. The principal
and early experimental results of secondary- scintillator used, NaI doped with thallium,
emission yields as high as 130 at 2.5 kV for was discovered by Hofstadter17. Much of
GaP:Cs. the development work on photomultiplier

Photomultiplier Handbook

tubes during this period was reported by depending upon the individual signals from
RCA and its competitors in the biannual each of the photomultipliers. Counting is
meetings of the Scintillation Counter Sym- continued until several hundred thousand
posium. These symposia were reported fully counts are obtained and the organ in ques-
in the IRE (and later the IEEE) Transactions tion is satisfactorily delineated. The location
on Nuclear Science beginning with the of each scintillation is represented by a point
meeting in Washington, January 1948. The on a cathode-ray-tube presentation.
scintillation counter became the most impor- The Computerized Axial Tomographic
tant measurement instrument in nuclear (CAT) scanner was introduced to this coun-
physics, nuclear medicine, and radioactive try in 1973. The device uses a pencil or fan-
tracer applications of a wide variety. beam of X-rays which rotates around the pa-
Headlight Dimmer tient providing X-ray transmission data
During the 1950’s, RCA collaborated with from many directions. A scintillator coupled
the General Motors Company (Guide-Lamp to a photomultiplier detects the transmitted
Division) on a successful headlight dimmer. beam-as an average photomultiplier cur-
The photoelectric headlight dimmer-first rent-and a computer stores and computes
made available only on Cadillacs and the cross-section density variation of the pa-
Oldsmobiles-basically used a tube similar tient’s torso or skull. The photomultipliers
to the 931A, but redesigned and tested to the are1/ or 3/4-inch end-on tubes which
auto manufacturer’s particular require- couple to the scintillator, commonly BGO
ments. The optical engineering problem was (bismuth germanate). Each unit is equipped
to sense the oncoming headlights or tail- with as many as 600 photomultipliers.
lights being followed without responding to PHOTOMULTIPLIERS AND SOLID-
street and house lights. Vertical and horizon- STATE DETECTORS COMPARED
tal angular sensitivity was designed to match
the spread of the high beams of the automo- In some applications either a photomulti-
bile. A red filter was installed in the optical plier or solid-state detector could be used.
path to provide a better balance between sen- The user may make his choice on the basis of
sitivity to oncoming headlights and to tail- factors such as cost, size, or previous ex-
lamps being followed. The device achieved a perience. In other applications, the choice
remarkable success, probably because of the may be dictated by fundamental properties
novelty, and thousands of photomultiplier of the photomultiplier or the solid-state
tubes were used. But today, one rarely sees a detector. A discussion follows of some of the
headlight dimmer. common applications favoring one or the
other detector with reasons for the choice. A
Medical Diagnostic Equipment summary presents the principal considera-
tions the user must apply in making a choice
In recent years two medical applications in an application for which he requires a
have used large numbers of photomultiplier photodetector. This information should be
tubes and have spurred further develop-
ments and improvements. The gamma cam- particularly useful to the designer who is not
era18 is a sophisticated version of the scin- well acquainted in this field.
tillation counter used medically for locating Photomultiplier Features
tumors or other biological abnormalities. A The photomultiplier is unique in its ability
radioactive isotope combined in a suitable to interface with a scintillation crystal and
compound is injected into the blood stream not only count the scintillations but measure
or ingested orally by the patient. The their magnitude and time their arrival. Most
radioactive material disintegrates and gam- scintillators emit in the blue and near ultra-
ma rays are ejected from preferential loca- violet. This spectral output obviously favors
tions such as tumors or specific organs. A the photomultiplier having a photocathode
large crystal intercepts the gamma rays and with high quantum efficiency in the short
scintillates. Behind the crystal are photo- wavelength range. On the other hand a sili-
multiplier tubes, perhaps 19, in hexagonal con p-i-n diode is relatively poor in this part
array. The location of the point of scintilla- of the spectrum but does best in the red and
tion origin is obtained by an algorithm near infrared. The most important factor,


probably, is the gain of the photomultiplier As a result of increasing concern about en-
which permits the measurement of the very vironment, pollution monitoring is becom-
small signals from individual scintillations ing another important application for photo-
with a good signal-to-noise ratio, limited multiplier tubes. For example, in the moni-
primarily by the statistics of the number of toring of NOx the gas sample is mixed with
photoelectrons per pulse. Finally, the short O 3 in a reaction chamber. A chemilum-
rise time of the photomultiplier using fast inescence results which is measured using a
scintillators permits time-of-flight measure- near-infrared-pass filter and a photomulti-
ments to be made in nuclear physics. plier having an S-20 spectral response. Al-
though the radiation level is very low, NO
Although the CAT scanner equipment can be detected down to a level of 0.1 ppm.
also uses photomultipier tubes to detect the The advantage of the photomultiplier in this
scintillations in bismuth germanate (BGO) application is again the high gain and good
crystals, the situation is somewhat different signal-to-noise ratio (the photomultiplier is
from the scintillation counting applications cooled to 0°C to reduce dark-current noise)
discussed above. In the CAT scanner the even though the radiation spectrum is ob-
X-rays produce a broad band of pulse served near the threshold of the S-20 spectral
heights and no attempt is made to single out range.
and detect single scintillation events. The In another pollution-monitoring applica-
photomultiplier is used in an analog mode to tion, SO2 is detected down to a level of 0.002
detect the level of radiation incident on the ppm. Here, the sample containing SO2 is ir-
crystal. In the CAT scan operation the radiated with ultraviolet and the excited SO2
typical machine scans the patient in a few molecules fluoresce with blue radiation that
seconds and the level of irradiance from the is detected with a combination of a narrow-
crystal onto the photomultiplier is relatively band filter and photomultiplier. Very weak
high so that only a relatively low gain photo- signals are detected and again it is the high
multiplier is required. Furthermore, the gain, good signal-to-noise ratio and, in addi-
speed of response requirement for the tion, good blue sensitivity which makes the
photomultiplier is relatively modest-per- detection and measurement of small contam-
haps a few hundred microseconds. Still, the inations of SO2 possible.
principal advantage of using a photomulti- Spectroscopy is one of the very early ap-
plier in this application for the detection of plications for photomultipliers. The wide
the radiant signal is its good signal-to-noise range of radiation levels encountered is
ratio. This ratio is very important to the pa- readily handled by the approximately loga-
tient because a reduction in its signal-to- rithmic gain variation of the photomultiplier
noise ratio would have to be made up for with voltage. At very low signal levels, the
with an increased X-ray dose. Nevertheless, signal-to-noise capability of the photomulti-
there is interest and development activity plier is essential. Because photomultiplier
aimed at replacing the photomultiplier with spectral response (with quartz or ultraviolet-
silicon p-i-n detectors. Two factors could transmitting-glass windows) covers the range
favor the alternate use of a silicon cell: (1) a from ultraviolet to near infrared, the
better scintillator (BGO is almost an order of photomultiplier is the logical choice for spec-
magnitude less sensitive than NaI:Tl; (2) a troscopic applications, except in the infrared
faster scanning machine (a very desirable region of the spectrum.
technological advance because is would
minimize effects of body motions). Both of Photocell* Features
these factors would result in a larger Because of their small size and low cost,
photocurrent and could bring the signal level CdSe and CdS type photocells are the logical
for the silicon detector to the point where the selection for applications such as automatic
fundamental signal-to-noise ratio from the exposure control in photographic cameras or
X-ray source would not be degraded. Such various inspection and counting require-
developments may be anticipated because ments.
*“Photocell” is used here to indicate a photosensitive
the silicon detector would also have the ad- device in which the charge transport takes place through
vantage of smaller size and perhaps lower a solid as compared with “phototube” in which the
cost. charge transport is through a vacuum.

Photomultiplier Handbook

Many p-i-n silicon cells are used in com- diode can be of the order of 100, but the sen-
bination with lasers or LED’s (light emitting sitive area is small-about 0.5 square
diodes). Here, one of the principal advantages millimeter.
of the silicon cell is its good response in the Sensitive Area. Photomultiplier tubes are
near infrared out to 1100 nm. In combina- made in a variety of sizes so that many dif-
tion with the Nd:YAG laser emitting at 1060 ferent optical configurations can be accom-
nm, the silicon cell is used widely in laser modated. The largest photocathode area
ranging and laser tracking. A similar ap- available in commercial RCA photomulti-
plication utilizes an LED emitting near 900 plier tubes has a nominal diameter of 5 in-
nm with a silicon cell for automatic ranging ches and a minimum useful area of 97 square
for special camera equipment. Size and in- centimeters. By way of contrast, the 1/2-inch
frared sensitivity are again the important side-on photomultiplier has a projected pho-
qualifications. tocathode area of 0.14 square centimeter.
A rapidly growing application for photo- Silicon p-i-n diodes are available with sen-
cells is for fiber-optic communication sitive areas generally not larger than 1 square
systems. LED’s are coupled to the fibers and centimeter; and avalanche silicon cells, 0.005
the detector may be a p-i-n diode or, for a square centimeter. In many applications, a
better signal-to-noise ratio, a silicon ava- fairly large area is required, e.g., coupling to
lanche diode. The qualifying attributes for a cathode-ray tube or a large scintillator.
the choice of detector are size, near infrared This requirement generally indicates the use
sensitivity, adequate speed of response, and of a photomultiplier tube. Silicon cells are at
good signal-to-noise ratio. an advantage when the source is small for
Smoke detectors now use large numbers of direct coupling or for lens imaging.
LED’s and p-i-n silicon cells. Again size, Temperature. Photomultipliers are gener-
cost, and infrared sensitivity are the impor- ally not rated for operation at temperatures
tant qualifications. higher than 75° C. Exceptions are photomul-
Characteristics Comparison Summary tipliers having a Na2KSb photocathode. This
Spectral Response. Photomultipliers can bi-alkali photocathode can tolerate temper-
be obtained with good spectral sensitivity in atures up to 150° C or even higher for short
the range 200 to 900 nm. Silicon cells have cycles. In oil-well logging measurements this
rather poor blue sensitivity, but are excellent consideration is important. Photocathode
out to 1100 nm. In general, then, the photo- sensitivities and gain change very little with
multiplier is to be preferred for applications temperature, but dark current does increase
involving the shorter wavelengths, although rapidly. Dark currents at room temperature
other factors may override this considera- are of the order of 10 - l5 ampere at the pho-
tion. tocathode and double about every 10° C.
Speed of Response. If very fast response is Silicon cells are rated from - 50 to 80° C.
required, the photomultiplier is usually the Sensitivities are also relatively independent
best choice of a detector. Photomultipliers of temperature. But dark current which may
are available with rise times (10 to 90%) of 1 be 10- 7 ampere at room temperature, also
or 2 nanoseconds using a 50-ohm load. The tends to double about every 10° C.
inherent rise time of silicon cells may be in Signal-to-Noise Ratio. At very low light
the range 10 to 20 nanoseconds, depending levels, the limitation to detection and
upon the area of the cell. However, because measurement is generally the signal-to-noise
of the cell’s capacitance, the effective rise ratio. One way of describing the limit to
time is much longer depending upon the detection is to state the Equivalent Noise In-
choice of load resistance. For example, with put (ENI) or the Noise Equivalent Power
a 1-megohm load resistance, the rise time (NEP). The NEP is the power level into the
may be of the order of 20 microseconds. A device which provides a signal just equal to
fairly large load resistance must be chosen to the noise. Most often the bandwidth is
maintain good signal-to-noise characteristics specified as 1 hertz and the wavelength of the
for the silicon cell. Silicon avalanche photo- measurement is at the peak of the spectral
diodes can have rise times as short as 2 responsivity. ENI is the same type of specifi-
nanoseconds. Gain for an avalanche photo- cation except the unit instead of power may


be luminous flux. 2. J. Slepian, U.S. Patent 1, 450, 265,
For a photomultiplier such as one used for April 3, 1923 (Filed 1919).
spectroscopy, the NEP at room temperature 3. H. E. Iams and B. Salzberg, “The
at 400 nm is about 7 x 10-16 watts, or the secondary emission phototube,” Proc. IRE,
EN1 is about 7 x 10 - l3 lumens. Both Vol. 23, pp. 55-64 (1935).
specifications are for a l-hertz ban 4. V.K. Zworykin, G.A. Morton, and L.

For a p-i-n silicon photocell, the NEP at 900 Malter, "The secondary-emission multipli -
nanometers may be of the order of 2 x er-a new electronic device,” Proc. IRE,
10- watts, or the EN1 of 1.5 x 10- l1 Vol. 24, pp. 351-375 (1936).
lumens. Both values are for a l-hertz band- 5 . J . R . P i e r c e , “Electron-multiplier
width. Thus, the photomultiplier is clearly design,” Bell Lab. Record, Vol. 16, pp.
superior in this category. Also it should be 305-309 (1938).
pointed out that the silicon diode must be 6. J.A. Rajchman, “Le courant residue1
coupled into a load resistance of about 5 dans les multiplicateurs d’electrons elec-
megohms in order to avoid noise domination trostatiques,” These L’Ecole Polytechnique
from the coupling resistor. Unfortunately, Federale (Zurich, 1938).
this large resistance then increases the effec- 7. V.K. Zworykin and J.A. Rajchman,
tive rise time of the silicon device to about “The electrostatic electron multiplier, Proc.
100 microseconds. The NEP of a silicon ava- IRE, Vol. 27, pp. 558-566 (1939).
lanche photodiode is about 10-14 watt at 8. J.A. Rajchman and R.L. Snyder, “An
900 nanometers or the ENI is 8 x 10 - l3 electrostatically focused multiplier
lumens, both for a l-Hz bandwidth. The phototube,” Electronics, Vol. 13, p. 20
lumen in these descriptions is that from a (1940).
tungsten source operating at 2856 K color 9. R.B. Janes, and A.M. Glover, “Recent
temperature. Peak emission for such a developments in phototubes,” RCA Review,
source is near 1000 nm and thus closely Vol. 6, pp. 43-54 (1941). Also, A.M. Glover,
matches the spectral peak of the silicon “A review of the development of sensitive
devices. phototubes,” Proc. IRE, Vol. 29, pp.
Gain. A photomultiplier can have a gain 413-423 (1941).
factor, by which the fundamental photo- 10. P. Gorlich, “Uber zusammengesetzte,
cathode signal is multiplied, of from 103 to durchsichtige Photokathoden,” 2. Physik,
10 8. Silicon avalanche photodiodes have a Vol. 101, p. 335 (1936).
gain of about 100. Silicon p-i-n diodes have 11. A.H. Sommer, P h o t o e m i s s i v e
no gain. The high gain of the photomulti- materials, John Wiley and Sons; 1968.
plier frequently eliminates the need of 12. R.E. Simon, Research in electron
special amplifiers, and its range of gain con- emission from semiconductors, Quarterly
trolled by the applied voltage provides flex- Report, Contract DA 36-039-AMC-02221
ibility in operation. (E) (1963).
Stability. Photomultiplier tubes are not 13. R.E. Simon and B.F. Williams,
noted for great stability although for low “Secondary-electron emission,” I E E E
anode currents and careful operation they Trans. Nucl. Sci., Vol. NS-15, pp. 166-170
are satisfactory. When the light level is (1968).
reasonably high, however, the very good 14. M.H. Sweet, “Logarithmic photomul-
stability of the silicon p-i-n cell is a con- tiplier tube photometer,” JOSA, Vol. 37, p.
siderable advantage. The silicon cell makes a 432 (1947).
particularly good reference device for this 15. H. Kallmann, Natur u Technik (July
reason. In fact, the National Bureau of Stan- 1947).
dards has been conducting special calibra- 16. J.W. Coltman and F.H. Marshall, “A
tion transfer studies using p-i-n silicon photomultiplier radiation detector,” Phys.
diodes. Rev., Vol. 72, p. 582 (1947).
17. R. Hofstadter, “Alkali halide scintilla-
REFERENCES tion counters,” Phys. Rev., Vol. 74, p. 100
1. H. Bruining, Physics and applications (1948).
of secondary electron emission, (McGraw- 18. H.O. Anger, “Scintillation camera”,
Hill Book Co., Inc.; 1954). Rev. Sci. Instr., Vol. 29, pp. 27-33 (1958).

2. Photomultiplier Design

PHOTOEMISSION In the energy diagram for a metal shown
The earliest observation of a photoelectric in Fig. 3, the work function represents the
effect was made by Becquerel in 1839. He energy which must be given to an electron at
found that when one of a pair of electrodes the top of the energy distribution to raise it
in an electrolyte was illuminated, a voltage to the level of the potential barrier at the
or current resulted. During the latter part of metal-vacuum interface.
the 19th century, the observation of a
photovoltaic effect in selenium led to the METAL
development of selenium and cuprous oxide
photovoltaic cells.
The emission of electrons resulting from
the action of light on a photoemissive sur-
face was a later development. Hertz dis-
covered the photoemission phenomenon in
1887, and in 1888 Hallwachs measured the
photocurrent from a zinc plate subjected to
ultraviolet radiation. In 1890, Elster and ENERGY
Geitel produced a forerunner of the vacuum
phototube which consisted of an evacuated
glass bulb containing an alkali metal and an
auxiliary electrode used to collect the 92CS - 32289
negative electrical carriers (photoelectrons)
emitted by the action of light on the alkali Fig. 3 - Energy mode/ for a metal showing
metal. the relationship of the work function and the
Fermi level.
Basic Photoelectric Theory
The modern concept of photoelectricity
stems from Einstein’s pioneer work for According to the quantum theory, only
which he received the Nobel Prize. The one electron can occupy a particular quan-
essence of Einstein’s work is the following tum state of an atom. In a single atom, these
equation for determining the maximum states are separated in distinct “shells”; nor-
kinetic energy E of an emitted photoelec- mally only the lower energy states are filled.
tron: In an agglomeration of atoms, these states
are modified by interaction with neighboring
(1) atoms, particularly for the outermost elec-
Eq. (1) shows that the maximum energy of trons of the atom. As a result, the outer
the emitted photoelectron mv2/2 is propor- energy levels tend to overlap and produce a
tional to the energy of the light quanta hv continuous band of possible energy levels, as
shown in Fig. 3.
must be given to an electron to allow it to The diagram shown in Fig. 3 is for a
escape the surface of a metal. For each temperature of absolute zero; all lower
metal, the photoelectric effect is character- energy levels are filled. As the temperature is
increased, some of the electrons absorb ther-
ed in electron-volts. mal energy which permits them to occupy

surface s c i e n c e
ELSEVIER Applied Surface Science 111 (1997) 251-258

Secondary electron emission studies
A. Shih *, J. Yater, C. Hor, R. Abrams
Natal Research l_xlboratory, Washington. DC 20375, USA
Received 14 June 1996;revised 28 July 1996; accepted 23 August 1996


Secondary-electron-emission processes under electron bombardment play an important role in the performance of a
variety of electron devices. While in some devices, the anode and the grid require materials that suppress the secondary-elec-
tron-generation process, the crossed-field amplifier (CFA) is an example where the cathode requires an efficient secondary-
electron-emission material. Secondary-electron-emission processes will be discussed by a three-step process: penetration of
the primary electrons, transmission of the secondary electrons through the material, and final escape of the secondary
electrons over the vacuum barrier. The transmission of the secondary electrons is one of the critical factors in determining
the magnitude of the secondary-electron yield. The wide band-gap in an insulator prevents low-energy secondary electrons
from losing energy through electron-electron collisions, thereby resulting in a large escape depth for the secondary electrons
and a large secondary-electron yield. In general, insulating materials have high secondary-electron yields, but a provision to
supply some level of electrical conductivity is necessary in order to replenish the electrons lost in the secondary-electron-
emission process. Our secondary-emission study of diamond demonstrates that the vacuum barrier height can have a strong
effect on the total yield. The combined effect of a large escape depth of the secondary electrons and a low vacuum-barrier
height is responsible for the extraordinarily high secondary-electron yields observed on hydrogen-terminated diamond

1. Introduction the tube walls to prevent RF vacuum breakdown. On
the other hand, high secondary-electron-emission
Secondary-electron-emission processes under
materials are desirable for grids in electron multipli-
electron bombardment play an essential role in vac-
ers and for cathodes in crossed-field devices, which
uum electronic devices. The materials used in the
is the area of our interest.
devices may need to be judiciously selected in some
Crossed-field devices, e.g., magnetrons and
cases to enhance the secondary-electron emission
crossed-field amplifiers (CFA), have established a
and in other cases to suppress the emission. In
long history of applications in radar systems. Their
microwave and millimeter wave power tubes, low
advantages include high efficiency, low cost, low
secondary-electron-emission materials are desirable
voltage, and compactness. Consequently, they also
for depressed collectors in order to ensure a high
find wide applications in microwave cooking, indus-
efficiency in the energy conversion. Low-emission
trial processing and radiation therapy equipment [1].
materials are also sought for coating the grids and
More recent applications are found in the generation
of plasma for precision etching and in highly-effi-
Corresponding author. Fax: + 1-202-7671280;e-mail address: cient electric lighting. Exotic applications [1] are being explored for the generation of gigawatt-power

0169-4332/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved.
PII S01 69-4332(96)00729-5
252 A. Shih et al. /Applied SurJace Science 111 (1997) 251-258

microwaves using a magnetically-insulated-line- dence (but not the magnitude) of the emission, ex-
oscillator (MILO) concept, and for beaming power cept in the simple case of A1, which is close to the
(wireless power transmission) using an electroni- assumed free-electron picture [9]. We will discuss
cally-steerable-phased-array-module (ESPAM) con- the secondary-electron-emission process in a mostly
cept. A good knowledge of secondary-electron-emis- qualitative manner, as done by Jenkin and Trodden
sion processes are crucial to the design of these [10], and will support the discussion with our obser-
devices. In some applications, e.g., AEGIS CFA's, vations. The discussion of the secondary-emission
only limited materials are available which have suffi- process is organized according to the distinct steps
ciently high secondary-electron yields to satisfy the used in all of the theories. The final step, which
device's needs. involves the overcome of the vacuum barrier, was
One of the major current challenges to crossed- considered to play only a minor role in the sec-
field devices (CFD's) is noise reduction. The re- ondary-electron-emission process. However, the sec-
quirement on the signal-to-noise ratio becomes more ondary-electron-emission behavior observed on dia-
stringent in radar applications, which demand high mond demonstrates the important role of the vacuum
resolution and fast response for small and close-to- barrier.
gether targets in a cluttered environment. In mi-
crowave-oven applications, a potential interference
with the neighboring band allocated to the new 2. Primary electron penetration and internal sec-
'wireless' services calls for a reduction in the side- ondary electron generation
band noise of the magnetrons. At present, CFDs are
The theories treat secondary-electron emission as
much noisier than coupled-cavity traveling-wave
occurring in three distinct steps: (l) production of
tubes, although in principle they should have compa-
internal secondary electrons by kinetic impact of the
rable noise characteristics [2,3]. Both experimental
primary electrons, (2) transport of the internal sec-
and theoretical efforts are pursuing noise reduction
ondary electrons through the sample bulk toward the
in CFDs with vigor. A thorough understanding of the
surface, and (3) escape of the electrons through the
secondary-electron-emission processes is essential to
solid-vacuum interface.
the success of these efforts. The secondary-electron-
The primary electrons are assumed to travel in a
emission characteristics of the cathodes are found to
straight-ahead path, slowing down through collisions
have a major effect on the signal-to-noise ratio [4,5].
with electrons and ions and transferring kinetic en-
In particular, numerical simulations and experimental
ergy to internally generated secondary electrons.
results have demonstrated that a very high electron
Most of the theories treat the energy loss according
emission (primary or secondary) would cause the
to the 'power law',
transition of the CFD to a low noise state [5].
Secondary-electron-emission is a complex pro- dE A
cess, and theoretical treatments are numerous. Earlier dx E" (1)
treatments have been reviewed by Dekker [6] and by where E is the energy of a primary electron at a
Hachenberg and Brauer [7]. More recent theoretical depth x, and A is an arbitrary constant. N(x), the
developments are summarized by Devooght et al. [8]. number of the secondary electrons produced in the
While the earlier models produced an overall agree- layer dx, is assumed to be equal to the energy loss in
ment with experimental observations as good as the the layer d E divided by the average excitation en-
more elaborate later models, the main problem was ergy B. Thus,
the lack of justification for the simplifying assump-
tions [8]. These theories were able to predict most of N(x)dx = - ~ (2)
the important characteristics of the secondary-elec-
tron emission, such as the secondary-electron yield A straight forward derivation [ 10] leads to
as a function of primary-electron energy and the ,
energy distribution of the secondary electrons. All of N(x) = B( R - x)"/"+ ' (3)
the theories predict well only the functional depen-
A. Shih et al. / Applied Surf'ace Science I l l (1997) 251-258 253

E~ +1
....... ...............................................
0 = ° oo
R (4)
(n+ 1)a
1.20 4 0°
where R is the maximum penetration depth and E o : 22.5 °
is the initial energy of the primary electrons. .?/ 0°

The best fit for the value of n was found to be II 0.80 .i;
0.35, as determined by electron transmission mea-
surements in A1203 [11] as well as by fitting the 0.40
'reduced yield curves' [6] taken from many materi-
als. A quantum mechanical calculation [7] derived an
0.00 "
. . . . i . . . . i . . . . i . . . . i

approximate value of n = 0.39, which is fairly close 500 1000 1500 2000

to 0.35. The value obtained for n is valid over the Primary Electron Energy (eV)
energy range between 300 eV and 3 keV for the Fig. 2. Changes in the secondary-electron yield with incident
theory and 300 eV and 7 keV for the experiments. angle 0. Normal incidence is at 0 = 0°. The data were taken on a
Eq. (3) reveals the increasing importance of sec- clean molybdenumsample.
ondary-electron production near the end of the pri-
mary-electron path and Eq. (4) points out that the process or by a diffusion process which involves a
penetration depth of the primary electrons increases large number of scattering events. Some later models
with increasing energy. The former fact is clearly explicitly take into account the electron-cascade pro-
shown in Fig. 1, which can be understood on a cess, which is the electron multiplication that occurs
physical level by considering a simple description of during the slowing-down of the internal secondary
the primary electrons moving through the solid. At electrons. In all of the models, the escape of the
high primary energies, the high-velocity electrons internal secondary electrons is described by an expo-
have a relatively short time to interact with the nential decay law with a characteristic escape depth
lattice electrons, and the internal yield per unit length X s•
is low. As the primary electrons lose energy, the Based on the penetration-depth (R) and escape-
interaction time increases and so does the yield. The depth (X~) concepts, the shape of the secondary-elec-
combined effect is that as the primary-electron en- tron yield curve as a function of the primary energy
ergy increases, the internal secondary electrons origi- is easily explained [10]. At very low primary ener-
nate deeper beneath the surface. gies for which R << X~, the internal secondary elec-
The transport of secondary electrons through the trons escape efficiently, and due to the small primary
material is modeled by either a single-scattering energy only a few secondary electrons are created.
Since the number of internal secondary electrons
generated increases with primary energy, the ob-
served yield also rises with primary energy. At very
high primary energies for which R >> X~, the expo-
nential nature of the escape process causes the de-
crease in the number of internal secondary electrons
z that escape to be more rapid than the increase in
I..- generation of internal secondary electrons. Conse-
quently, the observed yield decreases with primary
energy. The yield has a maximum value at a primary
13. energy for which R ~ X~. The resulting energy de-
, , , I , , • ' • - - ' • • - ' . . . . . . . .
pendent yield curve is bell shaped, as commonly
DEPTH R observed on most materials. An example is shown in
Fig. 1. The production rate of internal secondary electrons as a Fig. 2 of yield curves taken on a molybdenum
function of distance along the primary-electron path. sample.
254 A. Shih et al. /Applied SurJace Science I 11 (1997) 251-258

A similar qualitative discussion can be made about conduction-band minimum. The electron affinity is
the variation in yield with incident angle. If the typically on the order of an electron volt for insula-
primary beam is incident at an angle 0 to the tors. Since there are few conduction electrons in
surface, the maximum penetration depth is reduced insulators, the secondary electrons lose energy
by a factor cos0 relative to that at normal incidence through the excitation of valence electrons into the
( 0 = 0), assuming a straight-ahead path for the pri- conduction band. The wide band gap prevents sec-
mary beam. Consequently, the yields are larger at ondary electrons with kinetic energy less than E~,p
more oblique angles, and the enhancement in yield is from participating in such electron-electron colli-
more apparent at the higher primary energies for sions. For these electrons, electron-phonon and elec-
which R >> X~. The family of yield curves shown in tron-impurity collisions are mainly responsible for
Fig. 2 were taken at various 0 on a clean molybde- the energy loss. Because of the absence of electron-
num sample [12]. E .... , the energy at which the electron scattering, the secondary electron loses much
maximum yield occurs, increases with 0: E,..... = less energy as it moves through material and the
490, 520, 550, 580, and 660 eV for 0 = 0, 22,5.40, escape depth becomes large. Therefore, in general,
50, and 60 °, respectively. Since E,,,~,x occurs at an the yields are high in insulators [6]. For example,
energy for which R cos0 = X~, a larger 0 results in a 6 ..... is 6.8 for NaC1 and 25 for single-crystal MgO
larger R. Since the penetration depth R and the [6].
primary energy E are related by Eq. (4), a larger 0 When the secondary-electron-yield coefficient ex-
results in a larger Era, x as well. ceeds one, the number of emitted secondary elec-
trons exceeds the number of arriving primary elec-
trons. Consequently, there is a net electron flow from
3. Transport of internal secondary electrons the emitting sample into vacuum, and some level of
electrical conductivity is necessary in the sample in
In general, a primary electron with a reasonably order to replenish the lost electrons. Otherwise, sam-
high energy generates many internal secondary elec- ple charging will decrease the secondary emission
trons. However, high secondary-electron yields are yield to 1. Different approaches have been used in
not always observed because most of the internal practice to provide electrical conductivity. In oxi-
secondary electrons lose enough energy through col- dized beryllium, only a top 10-20 A thick layer is
lisions with electrons and ions to fall below the composed of BeO, beneath which metallic Be sup-
vacuum level. The energy-loss mechanisms for inter- plies the electrical conductivity. M g O / A u cermet is

nal secondary electrons differ in metals and insula- composed of a mixture of Mg and Au micro-crystal-
tors. In metals, the secondary electrons lose energy
by interacting with conduction electrons, lattice vi-
brations, and defects. The kinetic energy of a sec-
ondary electron must be at least E F + 05 when it 10,0
reaches the surface in order to escape. E F and 05 are
the Fermi energy and the work function of the metal,
respectively, and the minimum escape energy E v + 05
is typically about 10 eV. This large minimum escape
energy and the high collision probability due to the
il 4.0 ::"
large number of conduction electrons result in the
small secondary-electron yields found with metals.
The maximum yield 6r~,~ for a metal is on the order
of unity, and it varies between 0.5 (for Li) and 1.8 0.0
(for Pt) [6]. 0 500 1000 1500 2000 2500 3000

In insulators, the minimum kinetic energy for a Primary Electron Energy (eV)
secondary electron to escape is the electron affinity Fig. 3. Effect of electrical conductivity on the secondary-electron
X, the difference between the vacuum level and the yield from diamond.
A. Shih et al./Applied Surface Science 111 (1997) 251-258 255

lites of about 50 ,~ diameter, with the Au particles Our study reveals that two of the key reasons for the
providing the conducting path [13]. In diamond, extremely high secondary-electron yields from H-
appropriate dopants are used to provide the neces- terminated diamond surfaces are: (1) the presence of
sary electrical conductivity. Fig. 3 compares the a wide-band gap in diamond (5.47 eV) which allows
secondary-electron-yield curves from N-doped and low-energy secondary electrons to have large escape
B-doped diamond samples. N is a deep donor in depths, and (2) a very low or even negative electron
diamond, and at room temperature N-doped diamond affinity at the surface which permits a large concen-
samples are insulators. Fig. 3 shows that the sec- tration of secondary electrons residing near the bot-
ondary-electron yields from the N-doped diamond tom of the conduction band to escape. These premises
are very low. In contrast, B is a shallow acceptor in are supported by the analysis of the secondary-elec-
diamond, and therefore B-doped diamond samples tron-yield data and the secondary-electron-energy-
are conductive. The three B-doped samples studied distribution spectra taken on diamond samples as
here have electrical resistivities between 50 f~ cm presented below.
and 170 k l ) cm, and all three as-received samples All of the diamond samples were grown by chem-
showed comparably high yields. The yield curve ical vapor deposition (CVD) on Si substrates and
taken from one of the samples is shown in Fig. 3. then lifted off in a free-standing form. All samples
were acid cleaned to remove surface impurities such
as Si and non-diamond carbon. Most of the diamond
4. Escape of secondary electrons over the vacuum samples that we studied were subjected to a 30-min
barrier hydrogenation process at 20 Tort hydrogen pressure,
600 W microwave power, and 800°C sample temper-
Most of the theories do not explicitly treat the ature. Hydrogenated diamond samples are H-
escape process at the solid-vacuum interface. A phe- terminated [ 19].
nomenological approach by Bouchard and Carette Fig. 4 shows an energy distribution curve (EDC)
[14] found that the potential barrier at a solid-vacuum of the secondary electrons emitted from a H-
interface controls the final shape of the secondary- terminated diamond sample. The measurement tech-
electron energy distribution. This model and other nique has been discussed in an earlier paper [20].
theories do not regard the vacuum-barrier height as The energy of the secondary electrons is measured
having a strong impact on the magnitude of the relative to the Fermi level E F of the sample. The
yield. For example, Jenkins and Trodden [10] argued
that "most of the secondary electrons have energies
greater than 10 eV. For this reason, we should expect ... \
that changes of the order of a volt in the work
function of the surface would have only a small
effect on the secondary emission coefficient". Fur- ~'\ / Ec + Egap
thermore, they provided experimental evidence of Na
adsorption on W, which reduced the work function
by nearly 3 eV but increased the yield by only 60%. xlO
However, the results of secondary-electron emission
studies on diamond disagree with the conventional
wisdom. We will show below that hydrogen-termina-
tion of diamond lowers the vacuum level by 1.45 eV ~ * . . . . . . .
0 5 1'0 " ' 1'5
but enhances the secondary-electron yield by more ENERGY (eV above EF)
than an order of magnitude.
Recent investigations have found very high sec- Fig. 4. Secondary-electronEDC of an as-receivedhydrogenated,
B-doped, diamond sample. The data was taken at a primary
ondary-electron yields from hydrogen-terminated di- electron energy of 1000 eV. The sharp peak is larger in spectra
amond samples [15-18]. Malta et al. [18] found a taken from samples with a saturated hydrogen coverage or in
yield as high as 86 from a (100) diamond surface. spectra taken with higher primary electron energies.
256 A. Shih et al. /Applied Surface Science 111 (1997) 251-258

diamond sample used in the illustration is a B-doped 3
sample, which has a Fermi level above but near the
valence band maximum E v. The conduction band
minimum E c is an energy Eg,p above E v, where
Egap, the band gap energy, is 5.47 eV for diamond.
The position of E~ in Fig. 4 was determined by a
method described previously by Shih et al. [21].
Photoemission EDCs from diamond have been
shown to contain a high concentration of "quasi
thermalized secondaries in the lowest conduction
band" [22]. This high concentration of low-energy
electrons is also present in the electron-stimulated 5 10 15
secondary-electron distribution from diamond. They ENERGY (eV above EF)
are called quasi-thermalized electrons because the Fig. 5. Change in the secondary-electronEDC of a H-terminated,
energy width, although narrow (FWHM > 0.5 eV), B-doped, diamond sample after a 1000°C heating. Curve A is the
is much larger than that expected for fully thermal- EDC taken from the H-terminated sample, and curve B is the
EDC taken from the sample after the heating. The 1.45 eV rise in
ized electrons (0.025 eV). These electrons are piled
the vacuum level due to the heating prevents the high-concentra-
up just above E c as showed in Fig. 4. They represent tion of low-energy secondary electrons from being emitted.
a large fraction of the total emitted secondary elec-
trons, and in some cases over 90% of the emitted
electrons are contained in the sharp peak. When the electrons because of the absence of electron-electron
vacuum level falls near or below E~, these electrons scattering. The 'turn-off' of electron-electron scat-
can escape, resulting in a high secondary-electron tering is demonstrated in Fig. 4. In the enlarged part
yield. The vacuum level is ordinarily at the energy of the EDC, an abrupt change in the emission inten-
where the emission onset occurs, except in the case sity occurs at the energy E c + Egap. Below E c + Egap,
of a negative electron affinity (NEA) when the vac- only electron-phonon and electron-impurity scatter-
uum level can be below the emission onset energy• ings can occur, resulting in a large escape depth and
Fig. 4 shows that the emission onset occurs below E~ relatively high emission intensities. Above Ec + Egap,
for the H-terminated diamond sample. This observa- electron-electron scattering occurs which results in a
tion indicates the presence of a NEA on at least part small escape depth and a drop in the emission inten-
of the surface• sity.
H-termination of diamond is essential for the low
electron affinity and is consequently crucial to the
high secondary-electron yield• In Fig. 5, the sec-
ondary-electron EDC from a H-terminated diamond oo .,f~'J
sample is compared with that from the same sample
after heating at 1000°C for 15 min. The heating is
sufficient to desorb a large fraction of the surface
~- 40 ~/" with H-termination
hydrogen. Accompanying the hydrogen desorption, ./
the vacuum-barrier height increases by about 1.45
eV and is too high for the quasi-thermalized elec-
~ 20
trons to escape, resulting in a disappearance of the
sharp peak in the EDC. Simultaneously, the sec- diamond
10 ~ after H-desorption
ondary-electron yield drops precipitously. Fig. 6
shows the secondary-electron yield curves taken on 500 I000 1500 2000 2500 3000
the same sample before and after heating• Primary Electron Energy (eV)
The other important factor for the large yield is Fig. 6. A large reduction in the secondary-electron yield from
the large escape depth of the low-energy secondary diamond occurs with hydrogen desorption.
A. Shih et al./Applied Surface Science 111 (1997) 251-258 257

The quasi-thermalized electrons constitute a dom- 5. Summary
inant portion of the secondary electrons emitted from
H-terminated diamond samples. The low kinetic en-
Secondary-electron-emission processes have been
ergy of these electrons may be responsible for the
discussed using a three-step process: (1) generation
unusually large escape depth which is revealed by
of internal secondary electrons during the slow-down
the unusually high Emax. Earlier we discussed that
of primary electrons, (2) transport of the internal
the maximum secondary-electron yield occurs at an
secondary electrons through the material, and (3) the
energy for which R = X~. A large X~ means a large
final escape of the electrons over the vacuum barrier.
R, which in turn means a high Emax. Emax occurs at
A consideration of the penetration depth of the pri-
about several hundred eV for most metals and at
mary electrons and its relation to the escape depth of
about 1000-2000 eV for most wide-band-gap mate-
the internal secondary electrons explains the depen-
rials [7]. For H-terminated diamond, the yield contin-
dence of the secondary-electron yield on the energy
ues to rise over the 0 to 3000 eV energy range used
and the incident angle of the primary electrons. An
in the measurements (see Fig. 6).
examination of the scattering events experienced by
B impurities are introduced in the CVD diamond
the internal secondary electrons explains the differ-
samples to provide the required electrical conductiv-
ence in yield found for a metal and an insulator. The
ity. Since electron-phonon and electron-impurity
presence of a wide band gap allows a large escape
scatterings are the main energy-loss mechanisms for
depth for the low-energy secondary electrons, but it
the low-energy secondary electrons, an excessively
also depletes the number of conduction electrons
high impurity concentration may reduce the escape
available to sustain the emission process. Examples
depth and the yield. Fig. 7 compares the yield curves
of measures employed to supply electrical conductiv-
taken on diamond samples with different B-doping
ity were given.
levels. The higher yield curves are typical for low or
Diamond provides an opportunity to explore a
moderately B-doped diamond samples which have a
possible novel secondary-electron-emission process;
resistivity between 50-170 k12 cm and are transpar-
the novel characteristics are manifested by the dis-
ent. The lower yield curve is taken from the highly
tinctive sharp feature in the secondary-electron EDC
B-doped sample which appears black and opaque
and by the exceedingly high yields. The final step,
and has a resistivity between 50-170 l-I cm. We
which involves the overcome of the vacuum barrier,
have not studied any sample with a resistivity be-
was previously considered to play only a minor role
tween 170 ~ cm and 50 k f~ cm.
in the secondary-emission process. However, a high
concentration of quasi-thermalized secondary elec-
trons is present above the conduction-band minimum
of diamond. Hydrogen-termination on diamond sur-
faces lowers the vacuum barrier height to near or
below the conduction-band minimum which releases
a flood of low-energy secondary electrons and boosts
o 50 the total yield tremendously. The low electron affin-
40 ity at the surface and the large escape depth of the
~ J High B concentration secondary electrons are responsible for the extraordi-
II 3o / ~ - narily high secondary-electron yields from H-
co 20 terminated diamond samples.
The potentially rapid degradation in the yield
from H-terminated diamond due to electron and ion
°0 . . . . 500""' ;O'Od " ;5'00' " 20'0O ' ' 25OO ' 30'00 impingement prevents the wide application of dia-
Primaw Electron Energy (eV) mond in actual devices [16]. However, the knowl-
Fig. 7. Effect of the B impurity level on the secondary-electron edge gained from the secondary emission study of
yield from diamond. diamond points to the possibility of using other high
258 A. Shih et al. / Applied Surface Science 111 (1997) 251-258

secondary-yield materials, namely wide band-gap [7] O. Hachenberg and W. Brauer, in: ed. L. Marton, Advances
materials with low electron affinities. in Electronics and Electron Physics (Academic Press, New
York, 1959) p. 413.
[8] J. Devooght, A, Dubus and J.C. Dehaes, Phys. Rev. B 36
(1987) 5093.
Acknowledgements [9] A. Dubus, J. Devooght and J.C. Dehaes, Phys. Rev. B 36
The diamond samples studied were supplied by (1987) 5110.
[10] R.O. Jenkins and W.G. Trodden, Electron and Ion Emission
Drs. P. Pehrsson and J. Butler of the NRL Chemistry (Dover, New York, 1965) p. 54.
division. This work is supported in part by [11] J.R. Young, Phys. Rev 103 (1956)292.
NSWC/Crane Division and in part by the Vacuum [12] A. Shih and C. Hor, IEEE Trans. Elect. Devices 40 (1993)
Electronics Initiative. 824.
[13] V.H. Ritz, A. Shih and B. Sobocinski, Surf. Interface Anal.
18 (1992) 514.
References [14] C. Bouchard and J.D. Carette, Surf. Sci. 100 (1980) 251.
[15] G.T. Mearini, I.L. Krainsky, Y.X. Wang, J.A. Dayton, Jr., R.
[1] D. Chernin and Y.Y. Lau, ed., Proc. of the First Int. Work- Ramesham and M.F. Rose, Thin Solid Films 253 (1994) 151.
shop on Crossed-Field Devices, August 15-16, 1995, Ann [16] G.T. Mearini, I.L. Krainsky and J.A. Dayton, Jr., Surf. Int.
Arbor, Michigan. Anal. 21 (1994) 138.
[2] T.E. Ruden, G.E. Dombroski, D. Hobbs and G. Boles, Proc. [17] D.P. Malta, J.B. Posthill, T.P. Humphreys, R.E. Thomas,
of the First Int. Workshop on Crossed-Field Devices (1995) G.G. Fountain, R.A. Rudder, G.C. Hudson, M.J. Mantini and
66. R.J. Markunas, Mat. Res. Soc. Syrup. Proc. 339 (1994) 39.
[3] G. Ya. Levin, L.A. Semenov, A. Ya Usikov, Yu.A. Belov, [18] D.P. Malta, J.B. Posthill, T.P. Humphreys, M.J. Mantini and
Sov. J. Comm. Tech. Elect. 37 (1992) 128. R.J. Markunas, Mat. Res. Soc. Syrup. Proc. 416 (1996) 311.
[4] J.Z. Ye, R. MacGregor, C. Chan and T.E. Ruden, Proc. of [19] B.D.Thoms, P.E. Pehrsson and J.E. Butler, J. Appl. Phys. 75
the First Int. Workshop on Crossed-Field Devices (1995) (3) 1804 (1994).
202. [20] A. Shih, J. Yater, C. Hor and R. Abrams, IEEE Trans. Elect.
[5] H.L. McDowell, Proc. of the First Int. Workshop on Devices 41 (1994) 2448.
Crossed-Field Devices (1995) 236. [21] A. Shih, J. Yater, P. Pehrsson, J. Butler, C. Hor and R.
[6] A.J. Dekker, in: Solid State Physics, Advances in Research Abrams, Mat. Res. Soc. Symp. Proc. 416 (1996) 461.
and Applications, eds. F. Seitz and D. Turnbull (Academic [22] F.J. Himpsel, J.A. Knapp, J.A. Van Vechten and D.E. East-
Press, New York, 1958) p. 251. man, Phys. Rev. B 20 (2) (1979) 624.
European Laboratory for Particle Physics

Large Hadron Collider Project LHC Project Report 472


V. Baglin, I. Collins, B. Henrist, N. Hilleret and G. Vorlaufer

The secondary electron emission of surfaces exposed to the impact of energetic electrons contributes
significantly to the electron cloud build-up. For the prediction of the consequences of this effect the
measurements of the secondary electron yield carried out at CERN are an important source of
information. New experimental results concerning the total secondary electron yield for very low
primary electron energy (between 5 eV and 50 eV) will be also given in the case of as received copper.
Furthermore the energy distribution of the re-emitted electrons is drastically influenced by the primary
electron energy. The ratio of the number of reflected electrons to the total number of re-emitted electrons
has been measured and its variation with the primary electron energy will be shown. As a consequence
of these new experimental data, a numerical approximation to express the secondary electron yield as a
function of the primary electron energy will be given for the low incident electron energy region
(E < 50 eV). It has been shown that the decrease of the secondary electron yield due to the electron
bombardment could reduce sufficiently the consequences electron cloud effect. To understand further
the origin of this decrease, the results of experiments showing the variation of the electron induced
desorption yield with the incident electron dose will be compared to the concomitant reduction of the
secondary electron yield.


Administrative Secretariat
LHC Division
CH - 1211 Geneva 23

Geneva, 24 June 2002
The measurements of the secondary electron yield versus electron energy have been carried out
in two separate systems described previously1,2. They consist in an electron gun delivering a
focused primary electron beam of energy variable between 45 and 3000 eV. This beam hits the
surface of the sample under normal incidence at a circular spot (approximately 3 mm diameter). A
conical cage, coaxial to the primary beam, collects the secondary electrons. The deflection plates
and lens of the electron gun are driven by a computer program which optimises their settings on a
dummy sample to maximise the transmitted primary current. For the measurements short pulses
(30 ms) of low intensity (some nA) are used to reduce the electron dose received by the sample
(10 nC/mm2 for a complete energy scan). The vacuum system is bakeable and is evacuated by a
turbomolecular pump. After bake out the pressure reached in the system is in the low 10-10 Torr
In EPA1,3 the second system used was based on the same principle as the laboratory system, the
measurement procedure (beam optimisation, current pulses and data handling) was identical.











0 500 1000 1500 2000 2500 3000
Energy (eV)

Figure 1: Variation of the average secondary electron yield versus electron energy for 25 as received
copper samples
For the energy measurements, a 4 grids hemispherical energy analyser has been used. Its
energy resolution is limited to some eV and the energies given do not take into account any
correction of contact potential between the filament gun and the sample. The dose effect was
studied using various procedures. Initially (1978) the electron dose was delivered to the sample
using the measurement gun. The yield was continuously measured at the bombardment energy.
Because of the rapid destruction of the expensive gun filament, this method was abandoned and a
rustic flood gun was added to the laboratory system allowing to irradiate completely the sample

surface. In EPA the electron dose was accumulated by biasing the sample to a variable positive
voltage that attracted the photoelectrons created by the synchrotron light.
A change in the bias voltage modifies the mean energy of the incident electrons as well as their
amount. As the sample bias voltage changes the collection efficiency of the sample, the photon dose
needed to condition the sample are not directly comparable when the sample bias is modified. The
dose is determined by integrating the current collected on the sample. The main difference between
these two methods to accumulate the dose is that in EPA any artefact linked to the presence of a hot
filament for the production of the impinging electron has been eliminated.

23-01-01/ A.R.
31-01-01/ A.R.
23-03-01 /5.6 e-4
31-01-01/ 5.6 e-4
1.0 31-01-01/ 1.01 e-2
23-03-01 / 9.6 e-3




0 5 10 15 20 25 30
Energy (eV)

Figure 2: Variation of the average secondary electron yield versus electron energy for copper samples
and energies lower than 30 eV
Figure 1 shows the mean secondary electron yield (S.E.Y.) measured on 25 as received copper
samples cleaned following the LHC recipe. The average maximum is 2.06 (σ = 0.16) for an energy
of 271 eV (σ = 25). The bars are used to display the spread between the minimum and maximum
S.E.Y. at each energy for these 25 samples. The low energy part of the δ (E) curves was measured
recently and is displayed in Figure 2 for various electron doses (cf. next section). The two curves
labelled 23/03 AR and 31/01 AR show the SEY as a function of the energy for electron energies
between 4 eV and 30 eV. The SEY at 4 eV lies between 0.6 and 0.8. Below 10 eV the as received
curves tend to indicate a constant value of the SEY, independent of the electron energy.

The variation of the SEY with the electron dose is an effect leading, when not properly taken
into account, to underestimated secondary electron yield. This brought us to modify the
experimental set up and procedure in order to decrease as much as possible the electron dose
received by the sample during the measurements4. In Chamonix X 1 it was proposed to take profit
of this effect to obtain the decrease of the copper SEY necessary to operate LHC without “electron
cloud effect”.
2.6 DELTA MAX 23-03
ELECTRON GUN measured at 500 eV







1.E-08 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02
Dose (C/mm2)

Figure 3: The variation of the secondary electron yield with the incident electron dose
In Figure 3, the variation of the SEY of as received copper is plotted against the incident
electron dose for various measurements made using various procedures and under various
experimental conditions:
• The curve noted electron gun was obtained in 1979 using an electron gun on a copper
sample. It gives the SEY of copper at 500 eV energy versus the electron dose. The
impinging electrons had an energy of 500 eV. The area irradiated by the beam was a circle
of approximately 2 mm diameter.
• The curve note “delta max EPA” gives the variation of the maximum SEY against the dose
of electron collected with a bias of 99 V (i.e. 99 eV impinging electron energy).
• The 2 curves noted ‘delta max 23-03” and “delta max 31-01” were obtained this year using
a flood gun irradiating the whole sample with an electron energy of 500 eV.
These curves show that an electron dose between 8x 10-4 and 2x 10-3 C/mm2 is necessary to
reach a SEY lower than 1.3. The minimum of the SEY (close to 1.1) is obtained for an electron
dose close to 10-2 C/mm2.(fully conditioned sample).

In Figure 2 the low energy part of the two later curves (23-03 and 31-01) are also shown. The
reproducibility is very good, the SEY are continuously decreasing below 10 eV and are close to 0.2
at an energy of 4 eV.
The preceding measurements have shown the peculiar behaviour of the SEY as the primary
electron energy decreases. These peculiarities could be related to the change in the shape of the
secondary electron energy distribution. This distribution was studied using a 4 grids hemispherical
energy analyser. The energy axis is shifted by an unknown amount (< 3eV) due to the unknown
contact potential between the filament and the sample. It must be emphasised that, because of the
large amount of incident electrons needed for this type of measurement, the data presented here are
related to a copper close to the “conditioned state”.






0.6 EPA 99 eV
0.4 EPA 800 eV
EPA 350 eV
0.2 LAB GUN 500 eV

0 500 1000 1500 2000 2500 3000

Figure 4: Variation of the secondary electron yield versus the primary electron energy for 6 fully
conditioned (10-2 C/mm2) copper
In Figure 5 the energy distribution of secondary electrons emitted by a copper sample are
shown using two normalised axis: the abscissae are normalised to 1 at the incident electron energy,
the ordinates are normalised to 1000 for the maximum emitted intensity. On this graph the
increased importance of the reflected electrons at low primary electron energy is striking when the
curves measured at 10 eV and 550 eV are compared. A second fact, also reported in the literature5,
is illustrated in Figure 6 which shows a slight shift towards lower energy in the position of the
“true” secondary electron peak when the incident electron energy is raised.

The “true”secondary electron peak is found around an energy of 2 eV. Using the data partially
presented in Figure 5 the ratio between the number of reflected electrons and the total number of
emitted electrons has been calculated for the five energies considered (10, 30, 100, 300, 550 eV).
This ratio is given in Figure 7 as a function of the primary electron energy.
Usual expressions given by M. Furman6 or J.J. Scholtz7 can be used to numerically express the
variation of the true secondary electron yield δs with the primary electron energy (Ep). Both
formulae give good fits to the measured curves for a given energy range. The simple expression
given by M. Furman produces a reasonable fit in the low primary energy part (Ep< 1000 eV) which
is the most interesting for LHC. In the low energy region (<100 eV) the importance of the reflected
electrons becomes more important as it can be deduced from Figure 5. For this reason the fit
formula proposed here uses the Furman formula corrected for the contribution of reflected
electrons. The reflected electron contribution has been assessed from the previously shown
experimental data on the secondary electron energy distribution for copper. Furman’s relation for
the true secondary electron yield is:
s×( p )
δs = δ MAX E MAX
E s
s −1 +  p 
 E MAX 
δMAX, s and EMAX are 3 parameters used to obtain the best fit to the experimental.


Ep= 10 eV
Ep=30 eV
Ep= 100 eV
Ep=300 eV
500 Ep=550 eV





0.00 0.20 0.40 0.60 0.80 1.00

Figure 5: Normalised secondary electron energy distribution for conditioned copper

Ep= 10 eV
1000 Ep=30 eV
Ep= 100 eV
900 Ep=300 eV
800 Ep=550 eV

E max 550
Emax 10
0 2 4 6 8 10

Figure 6: Secondary electron energy distribution for copper ( below 10 eV)





0 100 200 300 400 500 600

Figure 7: The ratio between the reflected and the total number of re-emitted electrons in the case
of copper

The numerical value used to fit the experimental data presented in Figure 1 are given in
Table 1.

Table 1: Fit parameters for the true secondary yield (M.Furman formula)

δMAX 2.03 1.13
EMAX 262 318
s 1.39 1.35

To introduce the correction due to the reflected electrons two approximations could be made
depending on the accuracy needed. Of course as it appears in Figure 7, the correction for reflected
electrons is only significant at energy lower than 300 eV and amounts to more than 10% for
energies below 100 eV. A relation 7 allows to fit the experimental curve given in Figure 7:
ln( f ) = A0 + A1 × (ln(E p + E0 )) + A2 × (ln(E p + E0 ))2 + A3 × (ln(E p + E0 ))3
To obtain the best fit in the low energy part (below 300 eV), the following constants
has been used:
A0 = 20.699890, A1= -7.07605, A2= 0.483547, A3= 0, E0=56.914686
(Curve labelled FIT II low energy).
For use up to higher primary electron energy (2000 eV), the following coefficients
should be used:
A0 = 0.300207076, A1= 0.044915014, A2= -0.155498672, A3= 9.50318 x 10-4, E0=0
(Curve labelled FIT II)
A simplified exponential relation of the form:
f = R0 × exp(−E p / w)
can also be used below 100 eV using the following numerical constants:
R0= .64438713, w=43.2268304. (Curve labelled EXP FIT)
The accuracy of the various approximations can be appreciated from Figure 7.
The various formulae used to account for the reflected electron contribution at low energy
combined with Furman’s formula have been checked against the measured value of the total
secondary electron yield (δt) in the two cases of as received copper and fully conditioned copper
(dose = 10-2 C/mm2). To calculate δt, the following formula was considered:
δt = δS + δR ,
δR = f × δt ⇒ δt = δS + f × δ t
Hence: δt = δ s ×
(1 − f )
The results of the two fits are compared to the experimental results on the two Figures 8 (as
received case) and 9 (fully conditioned case). In both cases the agreement with the measured
secondary electron yield is good between 1000 and 100 eV incident energy. For energies greater
than 300 eV, the contribution of the reflected electrons can be neglected and M. Furman formula
used without correction. The low energy part (Ep < 100 eV) of the two graphs 8 and 9 is expanded
on the two graphs 10 and 11 to compare the results of the two fitting formulae to the experimental
results. Above 20 eV incident energy, both formulae give the same results. Below 20 eV (i.e. when
the reflected contribution accounts for more than 25% of the total number of secondary electrons)
the exponential fit gives increasingly underestimated value.









0 200 400 600 800 1000

Figure 8: The numerical expressions for the various contributions to the secondary electron yield
compared to the experimental measurements in the case of as received copper.



CU EXPOSED TO 0.01 C/mm2



0 200 400 600 800 1000

Figure 9: The numerical expressions for the various contributions to the secondary electron yield
compared to the experimental measurements in the case of fully conditioned copper




0.8 ON 25 SAMPLES)




0 20 40 60 80 100

Figure 10: The numerical expressions for the various contributions to the secondary electron yield
compared to the experimental measurements in the case of as received copper (Ep < 100 eV)

CU EXPOSED TO 0.01 C/mm2







0 20 40 60 80 100
Figure 11: The numerical expressions for the various contributions to the secondary electron yield
compared to the experimental measurements in the case of fully conditioned copper (Ep < 100 eV)

Measurements of the secondary electron yield of as received copper have been published in
references 8,9. They are compared to the results obtained at CERN for as received copper in
Figure 12. The agreement between CERN measurement and Hopman measurements is good. The
measurements made at SLAC gives a maximum secondary electron yield, which is significantly
smaller than those obtained in the two other labs. The different measurements obtained for a
primary electron energy in the vicinity of 1000 eV are close to each other.
A dose dependence curve has also been given in reference 8. This curve is compared with a
typical curve obtained at CERN in Figure 13. As noticed in Figure 12, the initial yield is much
smaller than what is measured at CERN and the yield decrease proceeds at a slower rate. The fully
conditioned state was apparently not obtained in that reference 8.











0 1000 2000 3000 4000

Figure 12: The variation of the secondary electron yield versus the primary electron energy as
obtained in three different laboratories

B. HENRIST CERN 31/01/01






1.E-09 1.E-07 1.E-05 1.E-03 1.E-01

DOSE (C/mm2)
Figure 13: Comparison of the dose dependence of the secondary electron yield as measured at CERN
and at SLAC
During the electron bombardment of a surface, leading to the above mentioned decrease of the
secondary electron yield, a significant decrease of the electron induced desorption yield is also
observed10,11. The desorption yields data, obtained in a separate system, on an as-received copper
sample, are presented together with the data concerning the secondary electron yield in Figure 14 as
a function of the number of electrons impinging per unit area. To allow a better comparison, both
data are normalised to 1 in the initial non-bombarded state. These data allow the calculation of the
amount of gas released during the conditioning of a copper surface by integrating the product of the
desorption yield and the electron dose.
Table 2: Total number of molecules released per unit area during processing for the main desorbed
QUANTITY (cm-2) 6x10 16 8x10 14 8x10 15 8x10 14 8x10 15 3x10 14

Figure 15 shows as a function of the total number of molecules released per unit surface area,
the secondary electron yield ratio and the desorption yield ratio for three molecular species. These
were chosen (for the sake of clarity) to represent the main desorbed gases of the two types:
hydrogen and carbon containing gases, (H2, CO and C2H6).

1.E+00 1


1.E-02 H2

1.E-03 C2H6
1.E-04 0.1
1E+12 1E+14 1E+16 1E+18 1E+20

Figure 14: The variation of the normalised desorption and secondary electron yields with the electron
These data show that the reduction of both the secondary electron yield and the desorption
yield are two processes evolving in parallel during the irradiation of a surface by electrons. The
upper limit of the total number (the mantissa being rounded to the next integer) of molecules
released during the conditioning of a copper surface is given in Table 2 for the main desorbed
gases. Hydrogen is the main released gas, its predominance is established during the initial part of
the conditioning when desorption yields are the highest and hydrogen the most abundant species,
carbon monoxyde and dioxyde come in second position. These results are in good agreement with
those concerning desorption yields published in the reference 11.The total number of molecules
removed from an as-received copper surface during its conditioning is smaller than 1017 molecules
per cm2 i.e. less than 100 monolayers. The number of molecules released per unit surface area
during the conditioning process is given in Figure 16 as a function of the final secondary electron
yield achieved. Although both the desorption yield and the secondary electron yield evolve in
parallel under electron bombardment, it should not be concluded that the cleaning of the surface is
the origin of the decrease of the secondary electron yield. For example a clean copper surface (e.g.
in situ glow discharge cleaned) has a secondary electron yield higher than a conditioned surface
(see Figure 4, curve labelled: “Argon G.D. Copper”).

1.E+00 1.0






1.E-04 0.1
1E+12 1E+13 1E+14 1E+15 1E+16 1E+17
AREA (mol/cm )

Figure 15: The normalised desorption and secondary electron yields as a function of the total number
of desorbed molecules

1.E+18 H2-23/03

1.E+17 CH4-23/03




1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
Figure 16: The quantity of desorbed molecules per unit area as a function of the final secondary
electron yield

The conditioning of as received copper surfaces has been studied in various experimental set-
ups and coherent results have proved its efficiency to reduce the secondary electron yield. Based on
measurements of the secondary electron energy distribution, a numerical expression has been given
which allows to calculate the fraction of reflected electrons with a good agreement to the
experimental data between 4 and 500 eV. The Furman’s expression has been used and the relevant
parameters are given to calculate the true secondary electron yield as a function of the primary
electron energy. Combining both formulae permits to compute the total number of electrons emitted
when the energy of the incident electrons is given. Using data obtained for electron induced
desorption it has been shown that the quantity of molecules removed from a surface during
processing is smaller than 1017 Using these data, it is now possible to better
estimate the amount of gas released during the beginning of the LHC operation when multipacting
can occur because of the electron cloud effect. Although the phenomenon of conditioning has been
obtained reproducibly on many samples, in different experimental set-ups, the exact mechanism
leading to this effect is not properly understood. This is of course not a comfortable situation as the
LHC operation at nominal intensity relies on this effect. Further studies are going on to try to
elucidate the main physical parameters responsible for this beneficial effect.

V. Baglin, B. Henrist, N. Hilleret, E. Mercier, C. Scheuerlein: Proceedings of the X workshop on LEP-SPS
performance Chamonix , 130, 2000.
V. Baglin, J. Bojko, O. Gröbner, B. Henrist, N. Hilleret, C. Scheuerlein, M. Taborelli: 7th Europen Accelerator
Conference, Vienna, 217-220, 2000.
V. Baglin, I.R. Collins, O. Gröbner, C. Grünhagel, B. Henrist, N. Hilleret, B. Jenninger: Proceedings of the
XI workshop on LEP-SPS performance Chamonix , 141, 2001.
G. Arnolds-Mayer, N. Hilleret: Advances in Cryogenic Engineering Materials, 28, 611-621, 1982
R. Bindi, H. Lanteri, P. Rostaing, J. Phys. D: Appl.Phys. 13,267, 1980.
M. A. Furman, CERN LHC Project Report 180, 1998.
J.J. Scholtz, D. Dijkkamp, R.W.A. Schmitz, Philips J. Res. 50, 375-389, 1996.
R. E. Kirby, F. K. King SLAC-PUB-8212, 2000.
H.J. Hopman, J. Verhoeven: Applied Surface Science, 150, 1-7, 1999.
F. le Pimpec: Thèse Université de Paris VI, 2000.
J. Gomez-Goni, A.G. Mathewson, J. Vac. Sci.Technol. A 15, 6, 3093-3103, 1997


Secondary electron emission from magnesium oxide on multiwalled
carbon nanotubes
Won Seok Kim
BK21 Physics Division and Center for Nanotubes and Nanostructured Composites, Sungkyunkwan
University, Suwon 440-746, Korea
Whikun Yi, SeGi Yu, Jungna Heo, Taewon Jeong, Jeonghee Lee, Chang Soo Lee,
and J. M. Kim
NCRI, Center for Electron Emission Source, Samsung Advanced Institute of Technology, P.O. Box 111,
Suwon 440-600, Korea
Hee Jin Jeong, Young Min Shin, and Young Hee Leea)
Department of Physics and Center for Nanotubes and Nanostructured Composites, Sungkyunkwan
University, Suwon 440-746, Korea
共Received 22 March 2002; accepted for publication 10 June 2002兲
We have investigated effects of electric fields on the yield of secondary electron emission 共SEE兲
from the primary electron bombardment on magnesium oxide 共MgO兲 covering vertically grown
multiwalled carbon nanotubes 共MWCNTs兲. We observe that the yield of SEE increases up to at least
22 000 at a special condition. The strong local field generated by the sharp tip of vertically grown
MWCNTs accelerates secondary electrons generated by primary electrons. This eventually gives
rise to so called Townsend avalanche effect, generating huge number of secondary electrons in a
MgO film. Emission mechanism for such a high SEE will be further discussed with energy spectrum
analysis. © 2002 American Institute of Physics. 关DOI: 10.1063/1.1498492兴

Secondary electron emission 共SEE兲 with bombardment was deposited on the vertically grown MWCNTs using elec-
of primary electrons plays an important role in vacuum de- tron beam deposition. Figure 1共a兲 shows the scanning elec-
vices. For instance, electron multiplier, microchannel plate, tron microscope 共SEM兲 image of MgO-coated MWCNTs,
and electron gun require the SEE materials with a high am- where CNTs are encapsulated by MgO only on the top area.
plification yield. In general, insulators are good candidates The inset shows a typical transmission electron microscope
for high SEE. The secondary electrons generated by the pri- 共TEM兲 image, where the thickness of MgO film at the side-
mary electrons in a magnesium oxide 共MgO兲 film move to wall is about 50–300 Å and the thickness at top of the CNT
the surface with relatively weak electron–electron scatter- tip about 500–2000 Å, although the nominal thickness of
ings due to the absence of free electrons in insulator, and MgO film from an electron-beam evaporator was 3000 Å.
finally escape from the surface if they have enough energy to This suggests that relatively small amount of MgO was de-
overcome the work function of the materials.1 Single crystal posited on CNTs compared to the thin film on a flat surface.
MgO, for instance, has a SEE yield of about 25 at best.2 On Figure 1共b兲 shows a schematic diagram of our apparatus
the other hand, the porous MgO produces high SEE yield of setup to measure the SEE. The MgO-coated CNT films were
about 1000 under the high electric field.3 However, for a bombarded by the primary electrons (I p ) which generates the
given yield, it is always desirable to look for the condition in secondary electrons (I s ) in the MgO film. The I p was mea-
which the lowest field is used. Recently a MgO film was sured by applying a positive bias of 200 V to the sample
deposited on randomly oriented carbon nanotube 共CNT兲 holder in order to avoid the leakage current to the chamber
powder, where relatively large SEE of maximum 15 000 at a wall. The I p was fixed at 233 nA during the experiment,
backbias of 1400 V was obtained.4 Yet the SEE obtained was while its kinetic energy was varied. The yield of secondary
strongly dependent on the sample positions and not repro- electron emission 共␦兲 is then defined as I s /I p ⫽(I p ⫹I t )/I p
ducible either. In this report, we introduce a systematic ap- ⫽1⫹I t /I p , where I s can be measured by the Faraday cup.5
proach to reproduce high SEE using MgO on vertically In practice, I s can be easily obtained by measuring I p and I t
grown multiwalled carbon nanotubes 共MWCNTs兲. We ob- 共the current supplied from the substrate兲, instead. The nega-
served an unusually high SEE yield of greater than 22 000 tive backbias to the sample decelerates the incident primary
共beyond the limit of a detector兲 at a backbias of 850 V, which electrons, while accelerating the secondary electrons to es-
was strongly related to the MgO film thickness. cape from the surface. In case of no primary electron, no
MWCNTs were grown by thermal chemical vapor depo- secondary electrons are generated. Note that the electric field
sition on the Ni-coated Si substrate using a C2 H2 gas at generated by the negative backbias is small such that no field
650 °C. 5 The average diameter and length of the grown emission currents are observed. The charge replenishment is
CNTs were 300 Å and 20 ␮m, respectively. A MgO thin film achieved by the current through the sample from the I t .
Figure 2共a兲 shows the yield as a function of net primary
Author to whom correspondence should be addressed: electronic mail: electron energy (⌬E), which is defined by the energy differ- ence E p -eV t , where E p and V t are primary electron energy

0003-6951/2002/81(6)/1098/3/$19.00 1098 © 2002 American Institute of Physics
Downloaded 30 Jul 2002 to Redistribution subject to AIP license or copyright, see
Appl. Phys. Lett., Vol. 81, No. 6, 5 August 2002 Kim et al. 1099

FIG. 1. 共a兲 The SEM image of MWCNTs covered with a MgO film. 共b兲 The
schematic diagram for SEE experimental setup.

and the negative bias applied to the substrate, respectively.
The nominal MgO thickness was 3000 Å. The yield is rela-
tively small for low E p over the whole range of ⌬E up to
900 eV, whereas this value shows an abrupt increase near
⌬E of 250 eV for the E p of 1000 eV. The existence of the
threshold voltage suggests that the SEE mechanism may be
related to an avalanche effect.7 With further increase of the
E p the yield was too large to measure by our multimeter FIG. 2. 共a兲 The yield for three different primary electron energies as a
whose maximum current is limited to 5 mA. The yield does function of the net primary electron energy. 共b兲 The yield as a function of
not increase with low ⌬E and the maximum peak is ob- measuring time with different nominal MgO film thicknesses at a fixed
primary electron energy of 1100 eV and the backbias of ⫺850 eV, and 共c兲
served near 200 eV. We measured the SEE by varying the the yield as a function of the backbias for various net primary electron
thickness of the MgO film. Figure 2共b兲 shows the yield as a energies.
function of time with different nominal MgO film thick-
nesses at fixed E p of 1100 eV. The yield becomes stabilized
after long measurement time. The very thin film of 1000 Å CNTs gives the yield of up to 800 at best at very high E p of
gives a low yield. The MgO film with a nominal thickness of 1500 eV and the CNTs without a MgO film gives a yield of
3000 Å gives rise to a maximum yield of 22 000, which is less than one. Figure 2共c兲 shows the yield as a function of the
beyond the detection limit of our multimeter. The SEE yield backbias for fixed net E p . The yield increases abruptly with
drops at larger MgO thickness. This may indicate the exis- increasing the backbias, indicating again an avalanche effect.
tence of an optimum penetration depth. The TEM image, as The values of on-set backbias are large at low and large net
shown in the inset of Fig. 1共a兲, shows the actual thickness of E p . The minimum on-set backbias exists at net E p of 250 eV.
the MgO film of about 500–2000 Å on the top of the CNT. We next measured kinetic energies of the emitted sec-
In general, the maximum yield is obtained when the penetra- ondary electrons by an electron energy analyzer 共VG Sci-
tion depth is about five times the escape depth in insulator.6 ence, Clam IV兲. Figure 3共a兲 shows the distribution of the
The escape depth of MgO is known as 60–200 Å,7 which kinetic energy of the secondary electrons in terms of three
implies the penetration depth of the MgO film to be 300– different backbiases at fixed E p of 1000 eV. The small peak
1000 Å. This is in good agreement with our observations of at 1000 eV, independent of the backbias, indicated by K 1 ,
the TEM image. We emphasize that the MgO film without results from the elastic backscatterings of the primary elec-
Downloaded 30 Jul 2002 to Redistribution subject to AIP license or copyright, see
1100 Appl. Phys. Lett., Vol. 81, No. 6, 5 August 2002 Kim et al.

drop across MgO film is considered to be about 500 V, i.e.,
corresponding to the voltage difference between K 3 and K 5
in Fig. 3共a兲. The average field strength is 500 V/d
(500– 2000 Å)⫽(2.5– 10)⫻107 V/cm, where the thickness
共d兲 of the MgO film on top of the CNT was measured from
the TEM image. This value is the same order of magnitude
of the field strength corresponding to the on-set field of ava-
lanche effect in typical metal oxides. This field may launch
the cascade emission of secondary electrons, but still cannot
explain such a high amplification factor. Thus, a higher back-
bias is required to ensure the avalanche effect in an MgO
insulator. We note that without CNTs such a high amplifica-
tion factor could not be observed. This proves that the small
diameter of CNTs enhances the local field strength near the
tip of nanotubes, since the typical diameter of MWCNTs is
about 100–300 Å. The field enhancement factor from a CNT
film is usually an order of 103,8 and the field enhancement
factor of a MgO coated CNT film was found to be 700–
1200, from the I – V curve of the field emission experiment.
Therefore, the local field near the tip can be enhanced by a
factor of 103 . This high local field will accelerate the primary
electrons arrived near the tip and generate a cascade emis-
sion of secondary electrons. This process results in the
charge depletion on the MgO surface and induces an addition
local field across the MgO film.3 The charge depletion was
replenished by the I t . In order to have avalanche phenom-
FIG. 3. 共a兲 The energy distribution for the secondary electrons in terms of enon to occur, the primary electrons should reach the MgO
three different backbiases at a fixed primary electron energy of 1000 eV. 共b兲 film near the nanotube tip and therefore optimum net energy
The schematic diagram of the potential drop of the secondary electron emis- for the primary electrons is expected, as observed from our
sion spectrum under a backbias of ⫺800 V.
results. Our data were fully reproducible with repetition of
the measurements. The variance of the SEE yield over the
trons. The electrons generated at the MgO surface (K 2 ) at a entire sample was within 10%.
backbias of ⫺800 V are negligible within kinetic energies of In summary, we have investigated the yield of SEE from
800–1000 eV. The secondary electrons start emitting from MgO/MWNTs. The SEE yield was measured for various
below 800 eV, where the sharp peak (K 3 ) at 800 eV is at- MgO film thicknesses, primary electron energies, and back-
tributed to the field emission. The broad peak (K 4 ) from 320 biases. The SEE yield was achieved to be much greater than
to 800 eV is considered to originate from the SEE that are 22 000. The electron energy analysis revealed that this high
generated through the whole MgO film layer. This trend of yield originates from an avalanche phenomenon due to the
broad energy distribution is very similar for other backbiases strong local field generated by the sharp CNT tip. We em-
except the position of the on-set kinetic energy (K 5 ) at phasize that this high SEE is obtainable over the sample with
around 320 eV. The voltage drop across the MgO film is good reproducibility. We expect our approach to be applied
shown in the diagram of Fig. 3共b兲. Although the backbias of, for various vacuum electronic devices.
for instance, ⫺800 V is applied between the chamber wall
and the substrate, it is only ⫺480 V that is applied across the This work was supported by the KOSEF through CNNC
MgO film with a voltage drop of 320 V between the chamber at SKKU and MOST through NRL and CRI 共J.M.K.兲.
wall and the MgO surface. The position of the on-set kinetic
energy is dependent on the geometry, i.e., the separation dis-
tance between the MgO surface and the chamber wall. 1
A. Shih, J. Yates, C. Hor, and R. Abrams, Appl. Surf. Sci. 111, 251 共1997兲.
In general, the electric field within MgO is considered to 2
A. J. Dekker, Solid State Phys. 6, 251 共1958兲.
be small, since the dielectric constant of MgO (␧⫽9.8␧ 0 ) is 3
H. Jacobs, J. Freely, and F. A. Brand, Phys. Rev. 88, 492 共1952兲.
W. K. Yi, S. Yu, W. T. Lee, I. T. Han, T. W. Jeong, Y. S. Woo, J. H. Lee,
about ten times larger than that of vacuum. However, the
S. H. Jin, W. B. Choi, J. N. Heo, D. R. Jeon, and J. M. Kim, J. Appl. Phys.
electric field of MgO in this experiment may be enhanced 89, 4091 共2001兲.
under the bombardment of primary electrons due to the es- 5
H. J. Jeong, Y. M. Shin, S. Y. Jeong, Y. C. Choi, Y. S. Park, S. C. Lim, G.
cape of secondary electrons of the surface from MgO. We S. Park, N. S. Lee, J. M. Kim, and Y. H. Lee, Chem. Vap. Deposition 8, 11
now estimate the strength of the electric field across the MgO 6
H. Seiler, Z. Angew. Phys. 22, 249 共1967兲.
film. For instance, at a backbias of ⫺800 V, large secondary 7
K. Kanaya, S. Ono, and F. Ishigaki, J. Phys. D 11, 2425 共1978兲.
electrons are obtained as shown in Fig. 2共a兲. The potential 8
P. G. Collins and A. Zettl, Phys. Rev. B 55, 9391 共1997兲.

Downloaded 30 Jul 2002 to Redistribution subject to AIP license or copyright, see

Qiong Wu

Submitted to the faculty of the University Graduate School
in partial fulfillment of the requirements
for the degree
Doctor of Philosophy
in the Department of Physics,
Indiana University

September, 2008

Accepted by the Graduate Faculty, Indiana University, in partial fulfillment of

the requirements for the degree of Doctor of Philosophy.

Doctoral Committee __________________________
(Shyh-Yuan Lee, PhD)

(David V. Baxter, PhD)

(Mark H. Hess, PhD)

(Richard J. Van Kooten, PhD)

August 21, 2008

Copyright ©2008 by
Qiong Wu

To my parents.


The work covered in this thesis could not be done without the help and

support from a great number of people. It is a pleasure to convey my

gratitude to them all in this humble acknowledgment.

In the first place, I would like to record my gratitude to my advisor Dr.

Shyh-Yuan Lee for his supervision, advice, and guidance from the very early

stage of my accelerator career as well as giving me the opportunity and

encouragement of finding this project. His diligence and passion in science

always inspired me as a student, and I believe the benefit I received from this

will continue on. He opened my mind to this distinguished field, and led me

through the hardest time at the beginning. I am honored to become one of his

students, and would like to give him my deepest respect.

After two and half years of course study in Indiana University, I moved

to Collider-Accelerator Department of Brookhaven National Laboratory (BNL)

looking for a PhD research topic. Dr. Ilan Ben-Zvi’s diamond secondary

emission group caught my interest. After I joined the group, all group

members offered the most generous help to my day-to-day research and

learning experience.

Dr. Ben-Zvi, as my local advisor in BNL, always sat with me with

patience whenever I am puzzled with the project. As a fully experienced and

genuinely talented scientist, he would then analyze the problem along with

me and lead me to a solution. I am deeply indebted to him for his detailed

editing of this document and of all the guidance he has provided.

I am also very grateful to Xiangyun Chang for his careful explanation to

all my questions and bright thoughts to this project. The working experience

with Xiangyun is very pleasant and profitable.

I would like to give my special thanks to Andrew Burrill, David Pate,

Triveni Rao, and John Smedley. As members in the group, they taught me so

much knowledge that a qualified experimentalist should know, little by little.

It is my honor to work with them and learn from them.

I would also like to acknowledge the people who have contribute in this

project: Craig Rhein, Joseph Saetta, John Walsh, Harrold Dorr, and Richard


I also wish to thank the other three scientists on my Doctorate committee:

Dr. David Baxter, Dr. Mark Hess, and Dr. Richard Van Kooten for agreeing

take time to review my thesis. They have given me many good suggestions

during my defense and in the thesis.

Finally, I would like to end this acknowledgement with my sincere

appreciation to my family. Their support and understanding is what led me

to this point and what made me the way I am.


Qiong Wu

Measurements and Studies of Secondary Electron Emission of Diamond

Amplified Photocathode

The Diamond Amplifier Photocathode (DAP) provides a very promising new

approach to provide high-average-current, high-brightness electron source for


High purity Chemical Vapor Deposition diamond films are used as the amplifier

of the electron beam. Primary electrons are provided by a traditional photocathode

and are bombarded onto tens of nanometer thick metal coating and into the diamond

sample. Within 1 micron travel in the diamond, the primary electrons generate

secondary electrons by collision on the order of two magnitudes increase in number.

Secondary electrons are accelerated through the diamond and will emit into vacuum

through surface with hydrogen termination. The electrons emitted should have very

low thermal emittance, for the electrons are constrained to the bottom of the

conduction band. The entire sample preparing process includes severe chemical

etching, metallization coating, and hydrogenation. Measurements are done with

specific equipments, and the quality of preparation is controlled by the Atomic Force

Microscope and electron or photon spectroscopy.

This thesis covers all aspects of this project, and will focus on the physics of

electron transfer within and out of the diamond sample. The measurements of the gain

are taken and compared under different conditions for obtaining the highest

amplification. The experiments have already demonstrated the secondary electron

gain of over 200 in the diamond and over 70 for emission into the vacuum. The

diamond will also act as a vacuum barrier, protecting the photocathode from

contamination by the accelerator vacuum. The emittance measurement is carefully

designed to reach the precision of 0.1eV. Theoretical calculations and computational

simulation are developed to fit with our experimental results.


1 Introduction ................................................................................................. 1

1.1 The Diamond Amplified Photocathode (DAP) Project .................................... 1

1.2 Advantages of diamond for amplified photocathodes ...................................... 5

1.2.1 Wide Band Gap ............................................................................................................ 5

1.2.2 Best Rigidity ................................................................................................................. 6

1.2.3 Highest Thermal Conductivity ..................................................................................... 7

1.2.4 Very high Mobility and Saturation Velocity ................................................................. 8

2 DAP Design................................................................................................. 10

2.1 Vacuum ........................................................................................................... 11

2.2 Temperature and current ................................................................................. 11

2.3 Energy bands and NEA surface in the diamond ............................................. 12

3 Preparation of diamond sample ............................................................... 15

3.1 Acid Etching ................................................................................................... 15

3.2 Metallization ................................................................................................... 19

3.3 Hydrogenation ................................................................................................ 20

3.4 Brazing ........................................................................................................... 20

4 Diamond laser ablation ............................................................................. 22

4.1 Laser system ................................................................................................... 23

4.2 Sample preparation and characterization........................................................ 23

4.3 Laser ablation of polycrystalline diamond ..................................................... 27

4.3.1 Ablation in vacuum .................................................................................................... 27

4.3.2 Ablation in rich O2 environment ................................................................................ 34

4.3.3 UV exposure ............................................................................................................... 35

4.4 Laser ablation of single crystalline diamond .................................................. 36

5 Transmission mode measurement of Diamond Secondary electron yield

..................................................................................................................... 40

5.1 Measurement setup ......................................................................................... 40

5.1.1 Primary electron source ............................................................................................. 40

5.1.2 Diamond holder and system setup ............................................................................. 42

5.1.3 Signal collection circuit ............................................................................................. 45

5.2 Experimental results ....................................................................................... 46

5.3 Laser Detrapping ............................................................................................ 53

5.3.1 Current Density .......................................................................................................... 57

5.3.2 Primary Energy .......................................................................................................... 59

5.3.3 Repetition Rate ........................................................................................................... 61

5.4 Energy lost in metal coating and estimation of energy needed for each e-h

pair .................................................................................................................. 63

5.5 Recombination of the charges in diamond ..................................................... 66

6 Emission Mode Measurement of Diamond Secondary Electron Yield 71

6.1 Measurement setup ......................................................................................... 71

6.2 DC beam or continuous long pulse measurements......................................... 73

6.3 Short-pulse measurements .............................................................................. 75

6.3.1 Delay time dependence .............................................................................................. 76

6.3.2 Temperature dependence............................................................................................ 79

6.3.3 Laser effect ................................................................................................................. 83

7 Thermal emittance measurement design ................................................ 87

7.1 Methods .......................................................................................................... 88

7.2 Limitations...................................................................................................... 89

7.3 Comparison of lenses ..................................................................................... 92

7.4 Shielding ......................................................................................................... 94

7.5 Alignment ....................................................................................................... 95

7.6 Conclusion ...................................................................................................... 98

8 Summary and conclusion .......................................................................... 99

List of Tables:

Table 1.1: Band gap energy list @300K [4]. ............................................................................. 6

Table 1.2: Thermal conductivity of various materials [6, 7]. .................................................... 8

Table 2.1: Electron affinity of carbon and hydrogen. .............................................................. 13

Table 5.1: Energy needed for create one electron hole pair. .................................................... 66

Table 6.1: Average radii used in theoratical model calculation.Error! Bookmark not defined.

List of Figures:

Figure 1.1: A complete DAP system. ........................................................................................ 3

Figure 1.2: Diamond conventional unit cell. ............................................................................. 5

Figure 2.1: DAP Design. (Photo on the left is taken by J. Grimes) ......................................... 10

Figure 2.2: Field and energy bands in the diamond with NEA surface. .................................. 14

Figure 3.1: Near-edge x-ray absorption fine structure (NEXAF) comparison of diamond

before and after acid cleaning. ....................................................................................... 17

Figure 4.1: AFM scanning pictures of typical single crystal diamond sample bare and

metalized surfaces. Top two are taken by optical camera, and the structure on the

bare diamond surface is the image of the holder of the diamond. The scanning

cantilever width of the AFM is 0.04mm. The middle and bottom pictures are AFM

scanning topography and feedback circuit signal of the bare diamond surface.

Sample number HID8. Sample scanned after ultrasonic cleaning in ethyl alcohol. ...... 25

Figure 4.2: Raman/photoluminescence spectra of different diamond sample lattice. ............. 26

Figure 4.3: 532nm ns pulse laser ablated polycrystalline diamond under optical

microscope (left) and Scanning Electron Microscope (SEM) (right). (These pictures

are taken by J. Smedley and T. Rao).............................................................................. 28

Figure 4.4: 266nm ns pulse laser ablated polycrystalline diamond under optical

microscope (left) and Scanning Electron Microscope (SEM) (right). (These pictures

are taken by J. Smedley and T. Rao).............................................................................. 29

Figure 4.5: AFM image of polycrystalline diamond 266nm laser ablated area center. ........... 30

Figure 4.6: NEXAFS trace of unablated and ablated regions of diamond sample and

reference trace for amorphous carbon. (These curves are obtained by J. Smedley) ...... 31

Figure 4.7: 213nm ns pulse laser ablated polycrystalline diamond under optical

microscope. (This picture is taken by J. Smedley and T. Rao) ...................................... 32

Figure 4.8: AFM image of polycrystalline diamond 213nm laser ablated area center. ........... 33

Figure 4.9: Raman spectra of samples following ablation with different O2 pressures.

(These curves are obtained by J. Smedley).................................................................... 35

Figure 4.10: FTIR spectra of before and after UV exposure of ablated area. Different

thickness of graphite and AC theoretical absorption curves are also shown as

reference. (These curves are obtained by J. Smedley) ................................................... 36

Figure 4.11: Optical microscope picture of 213nm laser ablated single crystal diamond.

(This picture is taken by J. Smedley and T. Rao) .......................................................... 37

Figure 4.12: AFM image of single crystal diamond 213nm laser ablated area edge. .............. 38

Figure 4.13: AFM image of single crystal diamond 213nm laser ablated area center. ............ 38

Figure 5.1: EGG-3101 thermionic electron gun from Kimball Physics. ................................. 41

Figure 5.2: Sketch of electron gun pulsing trigger. ................................................................. 42

Figure 5.3: Diamond holder structure. (Created by Chong-Jer Liaw)..................................... 43

Figure 5.4: Simple sketch of the transmission mode measurement principle. Either side of

the high-voltage supply can be grounded. ..................................................................... 45

Figure 5.5: Primary electron signal with pulse width of 200μs. .............................................. 47

Figure 5.6: Comparison of low field (top: 0.3MV/m) and high field (bottom: 3MV/m)

secondary electron signal of single crystal diamond with 0.3mm thickness in

transmission mode. ........................................................................................................ 49

Figure 5.7: Single crystal diamond window gain with primary electron energy ranges from

4keV to 8keV. Primary currents are shown in the legend. ............................................. 51

Figure 5.8: Secondary electron signal of 10μs pulse width with different duty cycles. No

laser beam applied. Primary electron peak current density is 0.03A/m2 and the field

applied in the diamond is 0.033MV/m. ......................................................................... 54

Figure 5.9: Secondary electron signal of 10μs pulse width with different duty cycles. Laser

beam applied (532nm, <5mW). Primary electron peak current density is 0.03A/m2

and the field applied in the diamond is 0.033MV/m. .................................................... 56

Figure 5.10: Gain comparison of with and without laser detrapping under different current

densities. Primary electron energy fixed at 6keV........................................................... 58

Figure 5.11: Gain comparison of with and without laser detrapping under different primary

electron energies. The primary current density is fixed at 0.04A/m2. ........................... 60

Figure 5.12: Gain comparison of with and without laser detrapping under various pulse

repetition rates. The pulse width for each measurement is fixed at 10μs. Primary

current density fixed at 0.036A/m2 and energy fixed at 6keV. ...................................... 62

Figure 5.13: Maximum gain of different primary electron energy. ......................................... 65

Figure 5.14: Diamond electron band structure [27]. ............................................................... 67

Figure 6.1: Simple sketch of emission mode measurement principle. Either side can be

grounded. ....................................................................................................................... 71

Figure 6.2: Emission mode gain with DC primary beam. ....................................................... 74

Figure 6.3: First pulse gain under room temperature (23°C) with different field and

effective energy for each pulse. Effective energy is the energy of the primary beam

excluding the energy loss in the metal coating. ............................................................. 77

Figure 6.4: Secondary signal with different delays for the primary electron after applying

the DC field. .................................................................................................................. 78

Figure 6.5: Emission signal of the diamond amplifier at 200°C. ............................................ 80

Figure 6.6: Emission signal of the diamond amplifier at -140°C. ........................................... 82

Figure 6.7: Room temperature (23°C) measurement with laser applied (532nm laser,

<5mW). .......................................................................................................................... 84

Figure 6.8: High temperature (200°C) measurement with laser applied (532nm laser,

<5mW). .......................................................................................................................... 85

Figure 6.9: Low temperature (-140°C) measurement with laser applied (532nm laser,

<5mW). .......................................................................................................................... 86

Figure 7.1: Setup design for thermal energy measurement of diamond secondary emission.. 88

Figure 7.2: Acceleration of the secondary electrons by applying a high voltage across the

diamond and anode. ....................................................................................................... 89

Figure 7.3: Aberration and space charge effects on beam spot size on screen. ....................... 91

Figure 7.4: Comparison of solenoid and two Einzel lenses. ................................................... 93

Figure 7.5: Comparison of solenoid and two quadrupoles. ..................................................... 94

Figure 7.6: Solenoid with different shielding. ......................................................................... 95

Figure 7.7: The offset magnetic field to the center of the electron beam and the waist.

FWHM difference according to the offset. Thermal energy of the beam is 0.1eV. ....... 97
Introduction 1


1.1 The Diamond Amplified Photocathode (DAP) Project

Over many years, electron accelerators have served in a large variety of

applications. Besides the contribution of electron colliders such as LEP in the precise

measurements of important values in the Standard Model, scientists and engineers

have made possible high brightness, constant power, and short wavelength light

sources to be possible. Synchrotron Light Sources and Free Electron Lasers (FEL)

provide the largest work platforms with high X-ray energy and brightness for biology,

chemistry, material science, medical science, and much more. Many applications

require continuous current, high intensity electron beams with narrow spatial and

temporal distributions.

Traditionally electrons are produced in thermionic cathodes. Thermal cathodes

rely on thermal energy overcoming the electrostatic forces restraining the charge

carriers of certain materials. The emitted current is limited by the temperature of the

cathode and the electric field. In recent years, photocathodes became widely used as

an electron source. The negatively charged electrode covered with a layer of light

sensitive material will emit electrons when struck by photons with energy above the

material‟s band gap. This type of cathode can provide high brightness electron beams
Introduction 2

and low emittance when combined with the high electric field available in an RF gun.

Metallic materials such as magnesium and copper are popular in RF guns that

operate with low average current [1, 2]. These metallic materials are easy to prepare,

robust and low in cost. The most important parameter for photocathodes is the

quantum efficiency (QE), which is the percentage of photons hitting the photoreactive

material surface that will produce an e-h (electron-hole) pair. The QE of the metallic

photocathodes is only in the order of ~0.1%. Semiconducting photocathodes, such as

GaAs, can provide a much higher QE (in the order of 10%) with a much larger photon

energy range than metallic materials. But they are very prone to contamination so that

the lifetime of semiconducting photocathodes is limited. Only a few hours of

continuous emitting are available at the regime of about 100 milliamps, which is of

interest for FEL and ERL [3]. The maximum current a photocathode can deliver

depends upon the power of input laser and its QE, which is wavelength and material


The Diamond Amplified Photocathode (DAP) is a new development for electron

sources that promise a high current, low emittance and high stability electron beam

with a long lifetime. The DAP constitutes two main components, a traditional

photocathode and an amplifier. The photocathode will provide a primary electron

beam as a laser illuminates the surface material. These primary electrons will be

accelerated to a few keV under a DC electric field applied between the photocathode
Introduction 3

and a thin metal layer coated on the amplifier window. The primary electrons will

penetrate through the metal coating into the amplifier material and produce secondary

electrons with amplification of more than 100. The secondary electrons will be

emitted into vacuum and accelerated by the RF field of the gun. Diamond is uniquely

suited to be the amplifier‟s material as will be explained later on.

Figure 1.1: A complete DAP system.

Figure 1.1 shows the schematic drawing of the DAP system with diamond

window as the amplifier. Diamond is ideal for this application, and its advantages will
Introduction 4

be discussed in the next section.

The scenario inside the diamond window after primary electrons penetrate

through the metal coating is as follows: Within about 100nm travel in the diamond,

the primary electrons generate secondary electrons and holes by collisions. The

number of secondary electrons depends on the primary electron energy, but it is

typically on the order of two orders of magnitudes larger than the number of primary

electrons. Thus each primary electron will form a plasma cloud of electrons and

holes along its range. The electron and holes will recombine within the plasma with

the presence of three-body scattering or diffuse and merge into the metal coating to

recombine within 1ns unless separated by an electric field. If the RF field of the gun is

carefully adjusted at the right phase to accelerate the electrons in the plasma cloud,

portion of the secondary electrons will be able to transfer across the diamond bulk.

The number of electrons that leave the plasma cloud is related to the RF electric field

amplitude and the energy of primary electrons (will be discussed in Chapter 5). As the

secondary electrons reach the opposite surface, they will exit into the vacuum through

the diamond surface with hydrogen termination. The electrons emitted are “cold”,

i.e. they have a very low thermal emittance, for the electrons are constrained to the

bottom of the conduction band.
Introduction 5

1.2 Advantages of diamond for amplified photocathodes

The material of the amplifier is carefully selected to gain the highest Secondary

Emission Yield (SEY). Among all the materials, diamond has many properties which

allow it to provide the best performance as an amplifier of electron beams.

1.2.1 Wide Band Gap

Diamond is the prototype material for the structure shown in Figure 1.2. The

conventional unit cell is face-centered cubic, with a side length a0 approximately

equal to 3.567Å (0.3567 nm), and two atoms per unit cell at room temperature.

Figure 1.2: Diamond conventional unit cell.

Diamond has a band gap of 5.47eV, which is very wide compared to other

materials. Table 1.1 shows the minimum band gap energy of several common
Introduction 6


Table 1.1: Band gap energy list @300K [4].

Material Band Gap Energy [eV] Breakdown field[MV/m]
Diamond (C) 5.47 (indirect) 2150
Silicon (Si) 1.11 (indirect) 30
Silicon carbide (SiC) 2.36-3.23 (indirect) 100-500
Germanium (Ge) 0.67 (indirect) 0.1
Indium arsenide (InAs) 0.354 (direct) 0.04
Gallium arsenide (GaAs) 1.43 (direct) 60
Gallium nitride (GaN) 3.4 (direct) 500

Materials with larger band gaps will have the ability of tolerant stronger electric

fields. According to Table 1.1, diamond can hold up to more than 2000MV/m electric

field before breakdown. High electric field is required for preventing recombination

and for increasing the drift velocity of the secondary electrons. As primary electrons

create e-h pair plasma along their penetration path, the higher the applied field, the

more secondary electrons can be pulled out before they diffuse into the metal coating.

Diamond is the only material that will have negative electron affinity (NEA)

with hydrogen termination on the surface, due to the large band gap. This will be

discussed further in Section 2.3.

1.2.2 Best Rigidity

It is well known that diamond is the most rigid material. This rigidity allows it to
Introduction 7

be prepared into thin films of only a few microns.

In principal, a thick diamond is desired for good thermal conductivity and

strength. But for RF applications, the electron acceleration phase limit in the RF gun

must be considered.

Calculations have been done by others for the single-cell RF gun designed for

RHIC electron cooling. The optimized initial phase is less than 35° and the frequency

of RF cavity is ~700MHz. The RF phase for secondary electrons generated near the

backside of the diamond should be at least 5°. The time allowed for secondary

electrons to pass through the diamond is less than 30°, i.e. ~120 ps. The saturation

velocity of electron transfer in diamond bulk is ~2×105m/s [5]. Thus the thickness of

the diamond window needs to be about 30 microns. For operation of a self-supporting

film with a diameter of the order of 1cm at this thickness the rigidity of diamond is a

great advantage.

1.2.3 Highest Thermal Conductivity

The thermal conductivity of diamond is much higher than even copper, which is

the most widely used thermal conducting material. The diamond is subjected to

several heat sources during amplification of the current:

 The power deposited by primary electrons bombardment.
Introduction 8

 The power deposited by secondary electron transportation.

 The power deposited by RF shielding on the metal coating.

 The resistive heating by the electron replenishment current flowing

through the metal coating.

The heat produced must be quickly dissipated to ensure the temperature of the

whole system does not increase to an unsafe limit. Simulations in X. Chang‟s Ph.D.

thesis showed that the diamond secondary emission cathode can easily handle 100W

of heat power without wide range temperature change [6].

Table 1.2: Thermal conductivity of various materials [7, 8].

Material Thermal conductivity [W/(m·K)] Temperature [K]
Diamond 1000 273
Silver 429 300
Copper 386 279
Aluminum 237 293
Stainless Steel 16.3 296
Water 0.6 293
Air 0.025 293

1.2.4 Very high Mobility and Saturation Velocity

The drift velocity of electrons in diamond is in the order of 1×105m/s under an
Introduction 9

electric field of a few megavolts per meter as mentioned above. High mobility is very

important for high frequency applications. The electrons under high frequency RF

field will only have a very small period of time to transfer through the amplifier in

case to be at the right phase when emitted into the gun. Along with high rigidity, the

diamond amplifier is easily capable of responding to 10GHz pulses.
DAP Design 10


The design of the DAP includes five main components designed to carry out the

process of electron emission described in Chapter 1:

 Diamond amplifier with its metal coating

 Niobium substrate

 Sapphire vacuum cell

 Photocathode

 Supporting window

Figure 2.1 shows the relative size and configuration of the DAP. Individual

issues of the DAP are discussed below.

Figure 2.1: DAP Design. (Photo on the left is taken by J. Grimes)
DAP Design 11

2.1 Vacuum

Photocathode performance and lifetime can be very sensitive to certain gas

density. An example is the cesium used in coating of a GaAs photocathode to provide

Negative Electron Affinity (NEA). Cesium is very reactive and can be easily oxidized

and lose its effect. The DAP capsule will protect the photocathode surface from

contamination during transportation.

A 30μm thick diamond window with 10mm diameter can hold up to 1Atm

pressure difference. The vacuum inside the DAP can be maintained even during

transfer of the assembly in air.

In addition to the diamond window rigidity, the brazing points must be vacuum

tight [9].

2.2 Temperature and current

The selection of niobium and sapphire as amplifier substrate and supporting of

the vacuum is based on their thermal expansion as well as their thermal and electrical


The DAP is designed as an electron source for both normal and

superconducting guns, thus it must be able to withstand repeated cryogenic cycles to

as low as a few degrees Kelvin. The preparation of the DAP involves diamond
DAP Design 12

brazing, which is a high temperature process. Considering the above extreme thermal

conditions, the material that is joined with the diamond needs to have a thermal

expansion coefficient as close to diamond as possible to provide minimum stress.

Good thermal conductivity is also important in the thermal processes.

As secondary electrons are emitted from the diamond, replenishment current is

needed to keep the whole system neutral. Thus the joined material must also be

electrical conductive.

Niobium with a thermal expansion coefficient of 7.3×10-6/K, electrical

conductivity of 6.6×10-6Ω-1m-1, and thermal conductivity of 53.7W/(m·K) is selected

according to the above criteria [9].

Along the same line, the cell must be a high thermal conductive insulator to

provide a good thermal path and electric standoff. Sapphire with thermal conductivity

of 35W/(m·K) at 300K, electrical conductivity of 1×10-14Ω-1m-1, and thermal

expansion of 5.6×10-6/K at 293K is very suitable for these purposes.

2.3 Energy bands and NEA surface in the diamond

Diamond is a material that easily demonstrates NEA [10]. The electron affinity

(EA) of an atom or molecule is the energy required to detach an electron from a singly

charged negative ion, i.e. the energy change for the process. Most metallic atoms have
DAP Design 13

more negative affinity than nonmetallic atoms.

Table 2.1: Electron affinity of carbon and hydrogen.

Atom Electron Affinity [kJ/mol] Reference

Hydrogen 72.77 [11]

Carbon 121.78 [12]

Lithium 59.62 [13]

Oxygen 141.004 [14]

Cesium 45.51 [15]

Potassium 48.38 [15]

The EA of a surface is defined as the vacuum energy level Evac minus the

conduction band minimum energy level Ec:

EA  Evac  Ec (2.1)

The Fermi levels of the diamond and of the termination materials (hydrogen or

alkaline elements) are aligned. Since the termination material has a relatively low

work function, the vacuum level can be lower than the bottom level of the diamond‟s

conduction band. This will allow the secondary electrons to escape into the vacuum.

As listed in Section 1.2.1, diamond has a very large band gap of 5.47eV. As
DAP Design 14

shown in Figure 2.2, the energy barrier from Fermi level to conduction band of

diamond is higher than the work function of most of the metallic elements including

hydrogen, which makes diamond very easy to form NEA. It is reported that the

electron affinity of the hydrogenated diamond surface can be as low as -3.4eV [10, 16,


Figure 2.2: Field and energy bands in the diamond with NEA surface.

Photoemission and reflective mode secondary electron emission of hydrogenated

diamond has been observed and reported [18-20]. For the application as current

amplifier, secondary electrons need to transfer through the entire diamond bulk and

only the non-recombined electrons can be emitted from the opposite surface.
Preparation of diamond sample 15


3.1 Acid Etching

The purpose of acid etching is to eliminate the various atoms and non-diamond

carbon layer present on the surface of the diamond sample. These elements and

structures will define the Fermi level of the area they occupied, thus will add

uncertainty to the experimental results.


 Prepare two work surfaces, one inside and another outside the fume hood. A hot

plate should be placed inside the fume hood

 Prepare large glass containers that can be tightly closed for storing liquid waste

 Prepare a plastic bag to collect solid waste such as wipes

 Clean all the glassware to remove contaminants

 Mount the diamond samples in the sample holder

 Position two beakers with deionized water near the hot plate in the fume hood

 Wear Personal Protective Equipment (PPE) for safety.

 Ensure the ventilation system in the fume hood is on
Preparation of diamond sample 16

 Place the chromic acid inside the fume hood. Chromic acid is hygroscopic, keep

it tightly closed, away from moisture or any source of water

 All the etching must be done under the fume hood. Keep hot liquids in the fume

hood untill they are cool. After etching, make sure to dispose of the chemical

waste in the pre-labeled glass containers. Follow the steps below:

o 15 min of saturated CrO5 in H2SO4 (heated so vapors are visible)

o Deionized (DI) water rinse

o 2 min NH4:H2O (1:10) Ultrasonic

o DI water rinse

o 2 min HCl:H2O (1:10) Ultrasonic

o DI water rinse

o 5 Min H2SO4:H2O2 (5:1) boiling

o DI water rinse

o 2 min NH4:H2O2:H2O (1:1:4) boiling

o DI water rinse

o 2 min HCl:H2O2:H2O (1:1:4) boiling

o DI water rinse

o DI water Ultrasonic

o DI water rinse
Preparation of diamond sample 17

Figure 3.1: Near-edge x-ray absorption fine structure (NEXAF) comparison of

diamond before and after acid cleaning.

In NEXAF, the photoelectron from core level excited by an input x-ray photon is

captured, and the consequent fluorescent photon or Auger electron or an inelastically

scattered photoelectron may also be measured. All information is combined to

determine all the final states of the photoelectron that are consistent with conservation

rules. Detailed analysis of the NEXAFS data yields information about the orientation
Preparation of diamond sample 18

and the binding of the surface. The curves in Figure 3.1 have been normalized

according to the photon flux, and the spectrum can be analyzed by comparing the

peak locations to Ref [21]. Between the two normalized curves, the acid etched

sample (blue curve) shows sharper diamond peaks and weaker non diamond peaks,

which indicates the possibility for detecting a photoelectron from pure diamond bond

is larger, i.e. the acid etching has eliminated the surface remnant for a certain degree.

The graphite peak around 280eV and second resonance peaks of iron around

350eV in Figure 3.1 gets smaller or vanishes completely after acid etching. The acid

etching would partially remove the amorphous carbon and eliminates iron on the


The strong acids used in the etching will create oxygen terminations on the

diamond surface. The oxygen-carbon bond is very strong, with a bond energy

D°298(C≡O) = (1076.4±0.7)kJ·mol-1 [22]. The oxygen termination will protect the

purification of the surface from contamination until we heat it up to 800°C or more.

After acid etching, the sample is kept in a clean environment at 1% humidity for

further experimentation. Each sample is etched and stored separately for

identification. We keep track of the history of each sample, so the metal coating

migration into the diamond due to electron bombardment and lifetime of the sample

can be evaluated.
Preparation of diamond sample 19

The migration of metal coating layer into the diamond bulk has been observed on

samples which has been used extensively, i.e. the thin metal coating on the surface

cannot be removed by acid etching following extended electron bombardment. In this

case, the diamond sample must be replaced.

3.2 Metallization

The signal collection highly depends on the details of the metallization of the

primary side of the diamond sample. Primary electrons must penetrate the metal

coating before entering the diamond for amplification. The metal coating will be in

contact with the niobium substrate to provide a replenishment current to the diamond.

All the electrons will lose part of their energy in the metal, and only the remnant

energy will contribute in generating secondary electrons. Thus the metal coating has

certain aspects that affect the performance of the amplification.

Clearly the selection of the coating material is important. The metal–diamond

contact should be able to allow holes easily escape from the diamond, and strong

adhesion to survive the extreme thermal process. The latter requirement prevents the

usage of metals which do not interact with carbon once applied, such as gold. The

former requirement therefore demands metals with EA close to that of the diamond to

form an ohmic contact on the diamond surface. In fact, at this specific surface, the

free carrier should be holes alone, which indicates that the surface will act as a p-type
Preparation of diamond sample 20

semiconductor, so metals with high work functions form the best contacts. In the

experiments discussed in later chapters, the metal coating is a ~50nm platinum on top

of ~30nm titanium, where the titanium forms a good bond with the diamond while the

platinum provides protection of the titanium from oxidation.

3.3 Hydrogenation

Hydrogenation is the process of applying hydrogen termination to the diamond

surface. Before applying hydrogen atoms, the diamond surface must be freed from

other elements by breaking their bonds at a high temperature, and usually this is done

with local temperature of 800°C or higher. After exposing the bare diamond surface,

the environment needs to be hydrogen atom rich. This can be done by flowing

hydrogen gas through a hot filament or gas cracker, or applying a hydrogen plasma

system. The hydrogenation in the experiments discussed in the following chapters was

performed with pure hydrogen flowing through a thermal gas cracker (MANTIS

MGC75) at ~0.5 cubic centimeters per minute.

3.4 Brazing

The configuration of the DAP as shown in Figure 2.1 can finished in two steps:

brazing and cold welding.
Preparation of diamond sample 21

The contacts between diamond and niobium, niobium and sapphire are produced

by high temperature vacuum brazing with an alloy of titanium, copper, and silver

(TiCuSil) as the brazing material. Ticusil is specially designed to join refractory

metals and ceramic/diamond. This brazing technique is carried out by J. Grimes and

described in detail in his MS thesis [9].

Cold welding is planned to be used to join the sapphire vacuum cell and the

supporting window of the photocathode. Considering that the photocathode material

is easily contaminated, it is important to keep it continuously under ultra-high vacuum.

The best solution would be sealing the supporting window onto the sapphire in-situ at

the preparation chamber of the photocathode. Cold welding is chosen to avoid a high

temperature process that could ruin the performance of the photocathode.
Diamond laser ablation 22


The diamond thickness for amplifier application is critical, because together with

the timing of the primary pulse, it defines the phase of the secondary emission pulse.

In addition, a very smooth surface is required to prevent spreading the electrons in

time, since the electron velocity is much slower in the diamond than in vacuum. For

the same reason, a highly parallel geometry is required for the diamond plate. The

ideal thickness to be used in a 700MHz electron gun for launch phase of 35° would be

about 30μm. This is not achievable with mechanical polishing due to the stress

embedded onto the surface during this process. Reactive Ion Etching (RIE) is one of

the possible methods to obtain such a dimension, but this is a time consuming process

if significant thickness reduction or large aspect ratios are required. The speed of the

RIE thinning of the diamond that is commercially available is about 10μm/day.

As an alternative method, laser ablation was introduced for shaping the Chemical

Vapor Deposition (CVD) diamonds. The advantages of this method, beyond thinning

the diamond to tens of microns with 10 times faster speed than RIE, include the laser

ablation can also engrave precise patterns onto the diamond surface or into the bulk as

Diamond laser ablation 23

4.1 Laser system

Two lasers, both operating at a fundamental wavelength of 1064nm with pulse

durations of 10ns and 30ps were used for the ablation. The ps laser is an actively and

passively mode locked Nd:YAG oscillator followed by pulse selector and double pass

amplifier, (Leopard series, Continuum Lasers) operating at 1064nm, capable of

delivering 125mJ at 10Hz in a pulse duration of 60ps. The wavelength of this laser is

converted to 266 nm by two successive doubling crystals to provide up to 13mJ

energy in 30 ps. Only the 266nm radiation of this laser was used to ablate the sample.

The energy on the target can be varied by changing the pump power of the amplifier.

The ns laser is a Q switched Nd:YAG capable of delivering up to 400mJ in a 20

ns pulse duration at 10Hz repetition rate. Two doubling crystals were used in series to

generate up to 200mJ of 532nm and 40mJ of 266nm radiation with pulse durations of

14ns and 10ns respectively. The energy of the laser beam on the sample was adjusted

by introducing appropriate filters in the beam path.

The beam from either of these lasers is focused using a nominal 15cm lens to

result in a 150μm focal spot size on the target. The position of the sample at the focus

was identified by the size of the laser beam back reflected from the sample.

4.2 Sample preparation and characterization

The diamonds used in experiments are all single crystal, ultra-high-purity, highly
Diamond laser ablation 24

polished and flat thin films produced by CVD. The size of the diamond samples used

in the experiments is typically 4.0mm × 4.0mm × 0.3mm, and the surface orientation

is [100]. Due to growth rate limitation, the other two orientations, [110] and [111] are

not available in ppb purity level. The impurity concentration in these diamond

samples are all within a few ppb. Both sides are polished to surface roughness (Ra)

<15nm, and the flatness is controlled to under 5 fringes @ 632.8nm over the entire


Atomic Force Microscope (AFM) scanning results of the diamond surface is

shown in Figure 4.1.

High purity and small thickness single crystal diamonds contain few electron

trapping centers. Impurity atoms, grain boundaries, and stress existing due to

imperfect lattice structure are possible potential wells for electrons. Secondary

electrons have a certain possibility being trapped by these potential wells. Electric

field shielding is produced due to these trappings, which are discussed in Ref [6].
Diamond laser ablation 25

Figure 4.1: AFM scanning pictures of typical single crystal diamond sample bare and

metalized surfaces. Top two are taken by optical camera, and the structure on the bare

diamond surface is the image of the holder of the diamond. The scanning cantilever

width of the AFM is 0.04mm. The middle and bottom pictures are AFM scanning

topography and feedback circuit signal of the bare diamond surface. Sample number

HID8. Sample scanned after ultrasonic cleaning in ethyl alcohol.
Diamond laser ablation 26

Figure 4.2: Raman/photoluminescence spectra of different diamond sample lattice.

In Raman/photoluminescence spectra above, electronic grade polycrystalline

(100μm grain size) sample HIE5 (green) and detector grade single crystal HID3 (red)

show negligible nitrogen impurities (<1ppm), while in optical grade polycrystalline

(10μm grain size) sample HIO7 and HIO8 (blue and black) the silicon and nitrogen

contents are significant. The optical grade samples also show large

photoluminescence background, which is likely due to the abundant randomly

oriented crystalline grain boundaries. In the secondary emission measurements, only

detector grade samples are selected. This simplifies the motion of electrons drifting

through the diamond, reduces the probability of trapping, and provides the best
Diamond laser ablation 27

rigidity and thermal conductivity.

Polycrystalline (electronic grade) and single crystal (detector grade) CVD

diamonds from Harris International were used for laser ablation.

In other samples, these baseline measurements were done on unablated regions

on the ablated samples. The sample is then mounted onto a holder that is suspended in

a vacuum commercial 2 ¾ inch vacuum cube held at pressures < 1μtorr. The cube is

mounted on a dual axis, motorized translation stage to facilitate movement of the

sample in a plane perpendicular to the laser beam and perform raster scanning. We

characterized all the ablated samples using optical and atomic force microscopes,

Fourier Transform Infrared Spectroscopy (FTIR) and Raman spectroscopy, Near Edge

X-ray Absorption Fine Structure (NEXAFS) and x-ray crystallography where


4.3 Laser ablation of polycrystalline diamond

4.3.1 Ablation in vacuum

Two identical electronic grade diamonds, labeled as HIE6 and HIE8, with

dimensions of 5 mm diameter, 0.16 mm thick and optical finish were exposed in

succession to 132J of 532nm and 90 µJ of 266 nm at 10 Hz repetition rate, and the

pulse duration was tuned to be at 10ns or 30ps for both wavelengths. The sample was

pumped down as described above to better than 1μTorr. The samples were moved in a
Diamond laser ablation 28

raster pattern in both x and y directions 10 times with a scan step of 50 µm for 70

minutes to ablate 1x1 mm2 on the sample. Since these two samples were identical, in

some cases, data from one sample before processing and from another after

processing are compared to study the effect of the process.

The ablation with 532nm ~10ns pulsed laser was quite destructive,

non-uniformly removing large regions of the diamond surface, although some raster

scan pattern is observable, as shown in Figure 4.3.

Figure 4.3: 532nm ns pulse laser ablated polycrystalline diamond under optical

microscope (left) and Scanning Electron Microscope (SEM) (right). (These pictures

are taken by J. Smedley and T. Rao)

Figure 4.4 shows the images of polycrystalline diamond after ablation with

266nm ~10ns pulse duration laser. Both images show clearer raster scan pattern as
Diamond laser ablation 29

well as some dark areas, and the surface damage caused by 266nm laser is limited to

small craters scattered on the ablated area.

Figure 4.4: 266nm ns pulse laser ablated polycrystalline diamond under optical

microscope (left) and Scanning Electron Microscope (SEM) (right). (These pictures

are taken by J. Smedley and T. Rao)
Diamond laser ablation 30

Figure 4.5: AFM image of polycrystalline diamond 266nm laser ablated area center.
Diamond laser ablation 31

Figure 4.6: NEXAFS trace of unablated and ablated regions of diamond sample and

reference trace for amorphous carbon. (These curves are obtained by J. Smedley)

AFM image shows the ablated area has about 2μm roughness. The clear trace of

the raster scanning is unacceptable for thinning the diamond amplifier, where the

roughness should be controlled to under 1μm.

NEXAFS from the unablated and ablated sections of the diamond are shown in

Figure 4.6. NEXAFS is a highly sensitive surface technique that can clearly delineate

surface impurity from the bulk impurity and can identify different forms of carbon

unequivocally. A reference curve for the amorphous carbon is also shown in the same

plot. In comparison to the unablated region, the ablation process has increased the
Diamond laser ablation 32

relative strength of the peak at 285eV while reducing the strength of the peaks at

289eV and beyond 292eV. The strong similarity between the reference plot and the

NEXAFS signal from ablated section of the diamond imply that the darkened region

is predominantly amorphous carbon (AC). The AC adheres on the surface of the

diamond, and can be easily removed chemically or in an ozone environment.

To form a smooth surface after the ablation, the raster scan size is decreased to

5μm to overlap the adjacent ablation traces. The laser is changed to 213nm, and the

pulse duration is decreased to ~30ps to deposit a comparable energy to the diamond

sample as ablations above. 266nm photon is corresponding to 4.7eV energy, which is

below the band gap of the diamond where 213nm photon is corresponding to 5.8eV

energy, which is just above the band gap. Thus 213nm laser is opaque to diamond.

Figure 4.7: 213nm ns pulse laser ablated polycrystalline diamond under optical
Diamond laser ablation 33

microscope. (This picture is taken by J. Smedley and T. Rao)

Figure 4.8: AFM image of polycrystalline diamond 213nm laser ablated area center.

The fine lines shown in Figure 4.7 within the ablated area are grain boundaries of

the polycrystalline diamond, which is same as the well defined sharp-angled pattern

shown in Figure 4.8. The area within the sharp angle has more material left after

ablation than the rest of the area, which is differently oriented. This indicates the

energy absorbed by one orientation is different from another orientation. So to obtain

a smooth surface, single crystal diamond must be applied.
Diamond laser ablation 34

4.3.2 Ablation in rich O2 environment

The vacuum chamber was backfilled with oxygen at base pressures of 14.2Torr,

0.9Torr, 0.1Torr and <0.01Torr using a leak valve and diamond was ablated at each of

these pressures. Optical transmission measurements using 623.8nm HeNe laser and

optical constants of amorphous carbon indicate that the thickness of the non-diamond

carbon layer is ~130nm and does not depend strongly on the oxygen pressure. The

Raman spectra for the ablated surfaces are shown in Figure 4.9. As can be inferred

from the data, the presence of oxygen did not make a difference to the formation of

the non-diamond carbon on the ablated surface.
Diamond laser ablation 35

Figure 4.9: Raman spectra of samples following ablation with different O2 pressures.

(These curves are obtained by J. Smedley)

4.3.3 UV exposure

The ablated surface of HIE8 was exposed to the radiation from a mercury arc

lamp (Jelight, GLF-12-SRC) for 6 hours at a distance of 5cm from the lamp.
Diamond laser ablation 36

Figure 4.10: FTIR spectra of before and after UV exposure of ablated area. Different

thickness of graphite and AC theoretical absorption curves are also shown as

reference. (These curves are obtained by J. Smedley)

In the region of wave numbers from 800cm-1 to 4000cm-1, the IR transmission

has changed from 65% before the UV radiation to 94% after the radiation. The UV

radiation in air would create a reactive ozone environment. Ozone would then form

CO2 at the presence of non-diamond carbon and left the surface cleaned.

4.4 Laser ablation of single crystalline diamond

The dark areas and uneven surface on ablated areas of polycrystalline diamond
Diamond laser ablation 37

could have been due to the grain boundaries. A single crystal was ablated with 213 nm

ps laser. This energy is above the band gap of the diamond, so the diamond is opaque

to the wavelength and all the energy will be deposited at the surface.

A single crystal diamond sample (HID13) was ablated using a laser pulse of

100µJ energy at a wavelength of 213nm and pulse duration of 30ps at a repetition rate

of 10Hz. The sample was raster scanned with 5μm step size. The optical microscope

pictures of the ablated and unablated regions are shown in Figure 4.11.

Figure 4.11: Optical microscope picture of 213nm laser ablated single crystal

diamond. (This picture is taken by J. Smedley and T. Rao)
Diamond laser ablation 38

Figure 4.12: AFM image of single crystal diamond 213nm laser ablated area edge.

Figure 4.13: AFM image of single crystal diamond 213nm laser ablated area center.
Diamond laser ablation 39

The roughness in Figure 4.13 is about 100nm, which is tolerable for the amplifier

application. Much less dark areas are created in this ablation process, and the raster

scanning pattern is hard to distinguish from the AFM images. From Figure 4.11, the

optical microscope image shows that there is much less surface non-diamond carbon

coverage than Figure 4.3 and Figure 4.4, and it does not show the grain boundaries as

Figure 4.7.

Laser ablation with energy above the diamond band gap is proved to be a

feasible method of thinning the diamond with submicron roughness with only a very

small amount of non-diamond carbon left on the surface. A RIE process can be added

to refine the surface of the laser ablated surface. The combination of both processes is

ideal for thinning the diamond sample to a few tens of microns with surface roughness

in nanometer scale.
Transmission mode measurement of Diamond Secondary electron yield 40


Transmission mode measurements are designed to study the secondary electron

generation and transport in the diamond. The emission into the vacuum uses a

different experimental setup and will be considered in Chapter 6.

5.1 Measurement setup

5.1.1 Primary electron source

Primary electrons are generated from a thermionic electron gun source Model

EGG-3101 commercially purchased from Kimball Physics. The cathode material is

high-brightness single-crystal lanthanum hexaboride (LaB6) with small spot size

option, which can provide 10μA peak current maximum and 100μm diameter spot

size minimum with optimum working distance. The gun has up to 10keV output

electron energy. The electron energies selected for experiments are limited to 8keV to

allow a safety margin.
Transmission mode measurement of Diamond Secondary electron yield 41

Figure 5.1: EGG-3101 thermionic electron gun from Kimball Physics.

Primary electrons are focused by an Einzel lens structure panels inside the gun.

With the available electron blanking option, the primary beam from the gun can

be pulsed. For short pulse output, an electric circuit is designed to trigger the blanker

inside the gun. Using the delay function of each output of the Stanford DG535 Digital

Delay and Pulse Generator shown in Figure 5.2, two successive pulses can be sent out

with programmed timing. These two successive pulses can be delivered when needed

to study the electron or hole trapping in diamond.
Transmission mode measurement of Diamond Secondary electron yield 42

Figure 5.2: Sketch of electron gun pulsing trigger.

5.1.2 Diamond holder and system setup

During the measurement, a sample holder is used for three purposes: Supporting

the diamond surface perpendicular to the primary beam; insuring electrical contact to

the primary electron impact surface of diamond in order to provide a replenishment

current; and providing an electric field to accelerate the secondary electrons and allow

measurement of the secondary current.
Transmission mode measurement of Diamond Secondary electron yield 43

Figure 5.3: Diamond holder structure. (Created by Chong-Jer Liaw)

In transmission mode measurements, the diamond is coated with selected thin

metal films, identically on both sides. A stainless steel tube is pressed onto the

primary electron impacting surface (primary surface), and the inner diameter of the

tube is slightly less than the metal coating outer rim. . The tube is then connected to

electrical ground through a current measurement resistor. The voltage on the resistor

is typically 10s of millivolts and therefore the primary surface can be considered

effectively grounded. Good electrical contact is ensured between the metal coating

and the tube to provide replenishment currents. The primary side of the diamond is

grounded through the oscilloscope during all studies and the signal is measured on

this low voltage level.

On the other side of the diamond, a copper anode is in good electrical contact

with the metal coating. The anode is connected to a high voltage power supply, and it
Transmission mode measurement of Diamond Secondary electron yield 44

can be set to any voltage between 0V to 5000V. The limit of 5000V is governed by the

potential breakdown, affecting the vacuum level of the system and for safety

considerations. Electrical break down through the vacuum might happen if the voltage

is exceeds 5000V, which corresponds to more than 15MV/m.

Ceramic spacers and washers are used as insulators given their good

performance in ultra-high vacuum. The use of the ceramic spacers also makes the

insulation path along the insulator‟s surface very long and the insulator‟s surface is

not directly exposed to the electron beams. These are very important considerations to

improve the insulation ability.

A copper block of relatively large mass with a stainless steel tube connected is at

the very back of the holder. The entire stack of components is screwed down onto this

block. The long tube which extending out of the system can provide heating and

cooling for different experimental conditions. Liquid nitrogen can be fed through the

tube and bring the system down to less than -140°C. A cartridge heater is also

available to slide down the tube to bring the system up to greater than 200°C.

A leak checking is done on the whole assembly, and the vacuum can reach

2E-9Torr after baking.

The principle of the transmission mode experiment is shown in Figure 5.4.
Transmission mode measurement of Diamond Secondary electron yield 45

Figure 5.4: Simple sketch of the transmission mode measurement principle.

Either side of the high-voltage supply can be grounded.

VDC is the applied high voltage on the anode. Current I measures the

replenishment current as the secondary electrons transits through the diamond. The

SEY of the amplifier is calculated by comparing the ratio between the replenishment

current and the primary current.

5.1.3 Signal collection circuit

According to the sample holder design for transmission mode measurement, high

voltage is applied to the copper anode for secondary electron acceleration and

collection. As the secondary electrons move away from the injection surface to the

opposite electrode and leave the diamond, replenishment current enters the diamond
Transmission mode measurement of Diamond Secondary electron yield 46

from the ground electrode and can be detected by the oscilloscope connected to the

holder. The signal collection circuit needs to have a fast response and good sensitivity

to small signals. The signal cable is kept short to reduce the distributed capacitance.

5.2 Experimental results

The electrode of the primary electron side of the holder is designed with a

tubular shape (similar to a Faraday Cup) to collect most of the secondary electrons

that are generated by the primary electrons on the metal coating surface and escape to

vacuum. It is found that more than 90% of those electrons are captured by the

electrode when the primary beam is focused in that hole. The primary electron current

is measured by focusing the beam inside the hole and applying no field on the


With the experiment setup discussed in Section 5.1.2, pulsed secondary electron

signals are measured by oscilloscope. The signal is then compared with the primary

electron signal. The ratio between the secondary current and primary current will give

the SEY under the given applied electrical field in the diamond.
Transmission mode measurement of Diamond Secondary electron yield 47

Figure 5.5: Primary electron signal with pulse width of 200μs.

Under a low applied field, the secondary electron drift velocity is not saturated.

The electron thermal energy at equilibrium is also smaller than that at higher applied

field and therefore under a low applied field these electrons have a larger probability

of being trapped. The trapped electrons will form an electric field, opposing the

applied field. The opposing field causes the succeeding electrons to experience a

weaker accelerating field. Therefore, fewer electrons will be able to leave the plasma

region and reach the anode, i.e. we observe a decrease of the signal or SEY. If the

pulse is adequately long, the effective field will decrease continuously in amplitude as
Transmission mode measurement of Diamond Secondary electron yield 48

more and more electrons are trapped. Eventually, the equilibrium state would be

reached when the trapped electron field fully cancels the applied field, and no more

secondary electrons will be able to leave the plasma range. Efforts to reduce trapping

will be further discussed in Section 5.3.

By using high-purity single crystal diamonds, the trapping centers in the

diamond bulk are limited. At a high applied field situation, the field decrease due to

the trapping is small compared to the applied field. The decrease of amplitude in each

pulse will be negligible.
Transmission mode measurement of Diamond Secondary electron yield 49

Figure 5.6: Comparison of low field (top: 0.3MV/m) and high field (bottom: 3MV/m)

secondary electron signal of single crystal diamond with 0.3mm thickness in

transmission mode.

The SEY or gain can be calculated by:
Transmission mode measurement of Diamond Secondary electron yield 50

gain = secondary current signal amplitude / primary current signal amplitude 1 (5.1)

The sign of the additional term equal to 1 added at the end of Equation (5.1) is to

reflect the negative polarity of the primary current signal.

Gain vs. field inside the diamond sample can be plotted as shown in Figure 5.7,

and this plot is a good tool for explaining the performance of the diamond amplifier

Transmission mode measurement of Diamond Secondary electron yield 51

Figure 5.7: Single crystal diamond window gain with primary electron energy ranges

from 4keV to 8keV. Primary currents are shown in the legend.

The plasma formed as each primary electron penetrates through the diamond

comprises electron-hole pairs. These will be separated under the applied field. With

no applied field the electrons and holes will all recombine with three-body scattering

or diffuse into the metal coating, leading to no gain. The electrons that make it

through the plasma region without recombination (or trapping) and arrive at the

opposite surface will contribute to the gain. According to each curve in Figure 5.7, the
Transmission mode measurement of Diamond Secondary electron yield 52

gain started with a small value under low electric field in diamond, which shows that

the recombination is significant. With the increase of the applied field, more electrons

can move away and the time needed for total separation decreases, thus the gain

increases. As the applied field increases, the charge carriers reach the saturation drift

velocity and the gain saturates. The gain under saturation only depends upon the

primary electron energy. Detailed explanation of recombination will be discussed in

Section 5.5.

With 8keV primary electron energy, the saturated gain is nearly 250. This

indicates that the diamond amplifier has the capability of increase the input current by

two orders of magnitude. The minimum applied field needed for reaching saturated

gain is less than 0.5MV/m, which corresponding to 150V on the anode in our setup.

The signal in a pulse dropping continuously for low field is due to trapping of the

electrons, and this can be avoided by detrapping methods that will be discussed in the

next Section. With a relatively fast-response circuit, the value of the peak in the

beginning of the signal corresponds to the number of electrons that are pulled out of

the plasma. Thus the value of the signal used for calculating the gain is taken at the

peak of the pulse in soon after its onset.
Transmission mode measurement of Diamond Secondary electron yield 53

5.3 Laser Detrapping

Due to the trapping mechanism diamond can easily have excessive electrons held

within potential wells. These electrons will create electric field with opposite polarity

to the applied field. The trapping centers can be impurities, grain boundaries,

distortion of the crystal lattice, etc., and can be decreased by using high purity single

crystal diamond samples.

However, during sample preparation process, the diamond is subjected to high

temperature, surface etching and metallization. These aggressive procedures may

introduce stress or impurities into the sample, and introduce trapping centers.

The probability of electrons being trapped by the trapping centers is constant,

given the large number of potential trapping centers. Thus, as current flows into the

diamond, the number of trapped electrons will increase with time. Figure 5.6 shows

the decrease of gain within the first pulse. Under a continuous pulse mode with high

repetition rate, the succeeding pulses will all be affected by the electrons trapped

during the preceding pulses. The electrons also have the possibility of escape from the

trapping centers due to applied electric field and thermal activity. So the escape

probability increases with increasing anode voltage and the time with current off

between pulses. If the applied field is high enough and/or time between pulses is long

enough to make the number of electrons released from the trapping centers

comparable with the number of electrons that are being trapped in one pulse, the
Transmission mode measurement of Diamond Secondary electron yield 54

signal will reach equilibrium. The equilibrium state is also dependent of the current

density of the primary beam. Higher current density will result in more electrons

trapped in the diamond.

Figure 5.8: Secondary electron signal of 10μs pulse width with different duty cycles.

No laser beam applied. Primary electron peak current density is 0.03A/m2 and the

field applied in the diamond is 0.033MV/m.

Detrapping of the electrons in a transmission mode is made possible by applying

a laser beam to the system during the experiment. Photons will excite the trapped
Transmission mode measurement of Diamond Secondary electron yield 55

electrons. If the applied field is set to zero, these exited electrons will move under the

field of the trapped electrons. Eventually all the trapped electrons would be


Most of the trapping centers in our diamond samples are rather shallow. Photons

of 532nm (green) wavelength, which is 2.34eV in energy, are applied to observe a

significant change in gain under relatively low field. A laser pointer with <5mW

power is used to provide the detrapping light, and a slot was opened on the holder

along the path of the light to allow the photons to reach and interact with the diamond.
Transmission mode measurement of Diamond Secondary electron yield 56

Figure 5.9: Secondary electron signal of 10μs pulse width with different duty cycles.

Laser beam applied (532nm, <5mW). Primary electron peak current density is

0.03A/m2 and the field applied in the diamond is 0.033MV/m.

By comparing the corresponding curves in Figure 5.8 and Figure 5.9, it is

obvious that laser beam played an important role in increasing the signal by releasing

the electrons from the trapping centers. Both figures show that the smaller the duty

cycle, i.e. more time between pulses, the larger the signal. The following sections will

discuss laser detrapping under different conditions as well as field dependence.
Transmission mode measurement of Diamond Secondary electron yield 57

5.3.1 Current Density

The total number of trapped electrons is determined by the product of probability

of trapping and the charge density within the trapping region in the diamond. Without

laser detrapping, for higher primary current density, more electrons will be trapped.

The signal observed after equilibrium is much lower than the first pulse unless the

applied field is extremely high. The time required for the system to reach equilibrium

varies with current density, but is within the range of 1 minute under all of our

experimental conditions.

The primary beam energy and spot size is fixed using the control system of the

gun. The peak of the first pulse is recorded as the total number of electrons created,

which should be the maximum gain under such applied field. The beam is pulsed at

10Hz repetition rate and duty cycle of 0.01%.

The applied field is increased from 0MV/m to above 1.8MV/m, and signals

under the equilibrium state are recorded for gain calculation both with and without the

laser beam.
Transmission mode measurement of Diamond Secondary electron yield 58

Figure 5.10: Gain comparison of with and without laser detrapping under different

current densities. Primary electron energy fixed at 6keV.

The transverse distribution of the primary beam is near uniform, and the

diameter is fixed at 3mm. So the highest current density of the 4 cases should be

5μA/(π × 2.25mm2) ≈ 0.7A/m2. The secondary electron current density is the product

of the primary current density and the gain. With the same primary electron energy,

the maximum gain under specific applied field should be the same.
Transmission mode measurement of Diamond Secondary electron yield 59

Figure 5.10 shows the significant effect of laser detrapping in all 4 different

primary currents, or current densities since the spot sizes for all case are the same. The

gain curve with laser detrapping is very close to the first pulse gain curve, which is

the maximum limit, even under a very small field and very high current density.

Compared with the maximum, the gain obtained without laser detrapping shows a

large decrease in value under low field and even under high fields when primary

current density is high. Saturation of the gain curve can be reached without laser

detrapping, but it would cost for a much higher applied field.

5.3.2 Primary Energy

Primary electron energy determines the saturated gain as discussed in Section 5.2.

With the current density fixed this time, but changing the primary electron energy

from 5keV to 8keV, and the secondary electron current density changes

correspondingly. The laser still brings the equilibrium gain close to the maximum gain

at each applied field.

To reach equilibrium state, the set up without laser detrapping requires a much

longer time for each data point. In the measurements, gain curves without laser were

only taken for 6keV and 7keV. 5keV and 8keV still show the same trend.
Transmission mode measurement of Diamond Secondary electron yield 60

Figure 5.11: Gain comparison of with and without laser detrapping under different

primary electron energies. The primary current density is fixed at 0.04A/m2.

The maximum gain increases with the primary electron energy, and so does the

secondary electron current. As shown in Figure 5.11, without laser detrapping, the

gain requires about 1MV/m of applied field to saturate in the 6keV case and even

more in 7keV case, but with the laser, it only requires <0.5MV/m. The gain with laser

is much closer to the expected value than the gain without laser before reaching
Transmission mode measurement of Diamond Secondary electron yield 61


5.3.3 Repetition Rate

If the secondary current density is the same by fixing the primary electron energy

and primary peak current density, the gain under low field will be determined by the

time lapse between pulses. For a more comprehensive study compared to Figure 5.8

and Figure 5.9, Figure 5.12 below shows the laser effect on gain under different duty

Transmission mode measurement of Diamond Secondary electron yield 62

Figure 5.12: Gain comparison of with and without laser detrapping under various

pulse repetition rates. The pulse width for each measurement is fixed at 10μs.

Primary current density fixed at 0.036A/m2 and energy fixed at 6keV.

Laser detrapping is essential to maintain a high gain under high repetition rates.

In the experiment for the measurement above, the pulse width is kept at 10μs

continuously. Without the laser, the decrease in signal is rather trivial for the duty

cycle of 1E-4. However, it became noticeable for duty cycle of 1E-3, and
Transmission mode measurement of Diamond Secondary electron yield 63

unacceptable for 1E-2. For a duty cycle of 1E-2, the repetition rate would be 1000Hz,

which is still very low compared with the frequency in most operating electron guns.

Laser detrapping shows significant increase in gain in transmission mode

measurements. With the correct selection of wavelength, the trapped electrons can be

released by photon-electron interaction. The laser used for detrapping is 523nm in

wavelength, which is corresponding to 2.34eV. This energy is less than half of the

band gap of diamond, which is ~5.7eV, and this ensures that all secondary electrons

are created by the primary electrons rather than by photons. Other wavelengths have

not been test quantitatively yet.

5.4 Energy lost in metal coating and estimation of energy needed for each e-h

Primary electrons lose their partial energy in the metal coating, and the amount

lost depends on the metal variety and thickness. Only the remaining energy

contributes to the value of maximum gain in all curves in Figure 5.7.

According to Ref [24, 25], the energy needed to create one e-h pair is fixed. The

saturated gain should be a linear function of the primary energy. The intercept on

horizontal axis in Figure 5.13 should be the energy lost in the metal coating Elost. The

data of the maximum gain is averaged over data from 1MV/m to 2.33MV/m. For a

coating of ~50nm Pt on top of ~30nm Ti, the energy lost is Elost =3.29 ± 0.15keV.
Transmission mode measurement of Diamond Secondary electron yield 64

From the linear model approximation of the incident electron energy loss [26],

the stopping powers for electron energy of ranged from 4keV to 8keV for Pt and Ti


E E
 12MeV  cm2 / g  20MeV  cm2 / g ,
 Pt xPt Ti xTi

where ρPt=21.45g/cm3 is the density of Pt and ρTi=4.506g/cm3 is the density of Ti, and

ΔxPt≈50nm and ΔxTi≈30nm are the thickness of the two layers of metal respectively.

Then the total energy loss in the metal coating calculated from this model is:

Etotal  EPt  ETi  1.556keV .

The discrepancy between the linear model and the transmission measurement result is

over 100%, which is very big. This is due to the thickness measurement of the metal

coating is not accurate, and should be verified.
Transmission mode measurement of Diamond Secondary electron yield 65

Figure 5.13: Maximum gain of different primary electron energy.

The energy needed to create one e-h pair is:

primary electron energy-energy lost in metal
Ee-h  (5.2)
maximum gain

Transmission mode measurement of Diamond Secondary electron yield 66

Table 5.1: Energy needed for create one electron hole pair.

Primary Energy [keV] Energy [eV]
4 19.80±1.29
5 19.25±1.61
6 20.06±1.82
7 19.63±3.67
8 20.33±2.60

The average is 19.82 ± 4.93eV. Former simulation approaches reported smaller

values of 14.66eV and 17eV [24, 25]. The difference is possibly due simplification of

the models and system error of the experiment.

5.5 Recombination of the charges in diamond

Diamond is an indirect band gap material, and its electron band structure has

been published in 1966 by W. Saslow et. al. as shown in Figure 5.14 [27].
Transmission mode measurement of Diamond Secondary electron yield 67

Figure 5.14: Diamond electron band structure.

Figure 5.14 shows a calculated band structure of the diamond which agrees with

experiment. The figure covered a primitive cell of the reciprocal lattice in the

frequency domain, or Brillouin zone, and each line indicates a different energy band.

The vertical axis is the energy of the electron with 0eV as the Fermi level in diamond,

and the horizontal axis is the value of wave vector k in the reciprocal lattice. The zone

center is Γ. L and X are boundary points at (2π/a)(½, ½, ½) and (2π/a)(100). Λ and Δ

are two axes along ΓL and ΓX directions respectively. The subscript labeled the

number of energy band.
Transmission mode measurement of Diamond Secondary electron yield 68

The band gap of the diamond is:

Γ25 - Δ1 = 5.4eV minimum value.
Γ25 - Γ15 = 7.3eV

Electrons present in the conduction band quickly settle into the energy minimum

of that band. For indirect band gap materials, electrons in this minimum cannot rejoin

the valence band by simply losing energy, or radiative recombination. They require

some source of momentum allowing them to observe both the laws of conservation of

energy and of momentum and fall into the valence band. The only way to promote

this recombination is to simultaneously emit (or absorb) a phonon that compensates

for the missing momentum vector, or a three-body collision with the impurities in the

diamond. However, such a combined transition has a very low probability.

Another possible path for the secondary electrons to recombine is to move to the

metal layer by diffusion. The secondary electrons and holes diffuse under Brownian

motion once they are created. Any electron or hole crossing the diamond-metal

interface will be recombined immediately in the metal layer and the defect rich

diamond near the metal boundary. Given that the plasma is formed tangent to the

interface and all charge particles are moving with random walks, the electrons and

holes are merged into the metal continuously through this diffusion process.

To illuminate the process, let us apply some numbers. Assume the SEY of the

primary electron is 100, and initial RMS radius of the plasma after thermal relaxation
Transmission mode measurement of Diamond Secondary electron yield 69

is 10nm. The density of the charged cloud under this assumption would be:

ne (t  0)  nh (t  0)  n0  2.4 1025 / m3 .

The Debye radius of this cloud should be:

 k T 
D   2 B   4.11010 m  0.41nm
 q N0 

under room temperature, which is more than 20 times smaller than the dimension of

the cloud. Therefore the charges can be considered as plasma at the initial state.

The plasma frequencies for both electron and hole clouds are:

n0e2 n0e2
 p ,e  and  p ,h  ,
 0 me  0 mh

where me and mh are effective masses of electron and hole inside the diamond. The

mobility of the two charge carriers are comparable and both are in the 10 5m/s scale

[29, 30], therefore it is reasonable to assume the effective masses are similar and both

close to the rest mass of electron. Plug in the values for the equation above, the

plasma frequencies should both be ~2.8×1014Hz.

For applied DC (or slowly varying) field, the attenuation constant is:

 plasma
  540nm ,
2 p

which is much larger than the RMS radius of the plasma.

This indicates that the applied field penetrates through the full length of the
Transmission mode measurement of Diamond Secondary electron yield 70

electron and hole clouds at the initial state without much decrease in the amplitude.

The plasma would start to separate under applied electric field.

As the electric field pulls the electron cloud away from the interface to prevent

loss of the electrons, the diffusion process, on the contrary, will keep pushing

electrons into the metal to be recombined. The eventual gain measured in the

experiment is the compromise of the two motions. High electric field can accelerate

the electrons to their saturation velocity away from the incident surface within a short

period of time, so that fewer electrons are lost. Therefore, the gain is small under low

electric field, and increases rapidly with the field amplitude until the drift velocity of

the electrons can reach the saturation velocity. The gain will not change as much after

the electron drift velocity saturates within a short distance. This explains the gain

curve shape in Figure 5.7.
Emission Mode Measurement of Diamond Secondary Electron Yield 71


6.1 Measurement setup

The emission mode measurement system reproduces the setup of the

transmission mode measurement discussed in Chapter 5, with the exception that we

introduced a vacuum gap between the copper anode and the diamond surface, and the

termination of this surface is changed to hydrogen instead of a metal coating.

Figure 6.1: Simple sketch of emission mode measurement principle. Either side can

be grounded.
Emission Mode Measurement of Diamond Secondary Electron Yield 72

With hydrogen termination, the NEA surface will allow the secondary electrons

to escape into vacuum and get collected by the anode. An amplifier circuit was

designed and added for signal collection. With this improvement, the response time

decreased from ~1μs into ~20ns.

The emitted electrons can be measured by an oscilloscope connected to either the

anode or cathode. In the anode connection measurements, the anode is grounded and

the metal coating on the incident side is negatively biased. The effective energy of the

primary electrons contributed to the emission is calculated as

Eeffective  E primary - U metal - Elost ,

where Eprimary is the primary electron energy at the exit of the gun, Umetal is the

negative bias voltage on the metal coating, and Elost is the energy lost in the metal


The grounding of the anode is designed for application of a phosphor screen and

a CCD camera. If a focusing device is applied between the anode and the phosphor

screen and the distance is carefully chosen, the thermal emittance of the diamond

amplifier can be measured by imaging the beam spot size on the CCD camera. The

design for this measurement will be discussed in Chapter 7.
Emission Mode Measurement of Diamond Secondary Electron Yield 73

6.2 DC beam or continuous long pulse measurements

The trapping issue in emission mode measurement is more problematic due to

the critical process of hydrogenation. For the hydrogenation process, the diamond is

heated to above 800°C to break the carbon-oxygen bond and leave free dangling

carbon bond on the surface. Hydrogen gas is broken into atoms (or “cracked”) and

impacts the diamond surface. We use a commercial thermal gas cracker (Mantis

MGC-75) to carry out this process which leads to create negative affinity surface

(NEA) on the diamond. During the formation of carbon-hydrogen bonds, it is likely to

have residual dangling carbon bonds on the surface without any hydrogen attached.

The dangling bonds are capable of trapping the secondary electrons transferred to the

hydrogenated surface. The field applied to the diamond during experiments is about

1MV/m, and to cancel such field, the trapped electron density only needs to be

0.5C/cm2, which is about 1 trapped electron per 10,000 conventional unit cells on the

surface. This is a very small number considering the occupation of hydrogen coverage

during CVD diamond [111] surface growth can be as low as 86% [33].

With a continuous primary electron beam, sufficient time is allowed for all

possible trapping events to occur. The signal measured on the anode side comprises

the trapping and releasing of the electrons in the diamond. The signal detected is

purely caused by emitted electrons and has no capacitive coupling components.
Emission Mode Measurement of Diamond Secondary Electron Yield 74

Figure 6.2: Emission mode gain with DC primary beam.

The signal is much smaller than the transmission mode measurements, and green

laser had no effect on the signal. Since the dangling bond state energy is more than

1.3eV [34, 35], the green laser detrapping is not as effective in emission mode,

because the electrons are trapped much deeper by dangling bonds on the surface than

the trapping under the surface caused by impurities and stress in lattice. The signal

will increase with better hydrogenation, which will reduce the number of dangling

bonds on the surface.
Emission Mode Measurement of Diamond Secondary Electron Yield 75

6.3 Short-pulse measurements

Despite the small signal in the DC beam mode, the diamond performance is

significantly better in short-pulse measurements. As in transmission mode

measurements, electrons in the beginning of the first pulse applied to the diamond will

experience a strong, unshielded applied field, and have much higher possibility to

emit. The gain in the beginning of the pulse is much higher than the value measured

under equilibrium state.

In real applications, the applied pulse would be very short. As an example, the

proposed e-cooling facility in Brookhaven National Laboratory has a frequency of

703MHz. The pulse from the source must be much shorter than 1ns. Therefore the

performance of the diamond amplifier with short pulses is more important and


For short pulse measurements, the diamond was first set at neutral. Shortly after

(<3s) the electric field was applied, a train of pulses of 40μs was sent to impact the

diamond from the electron gun. With the fast response signal collection circuit, only

the SEY of the first pulse was recorded. This ensures that the front segment of the

pulse experienced the field as applied, and the change of the signal due to trapping

and shielding of the field is reflected later in the pulse.
Emission Mode Measurement of Diamond Secondary Electron Yield 76

6.3.1 Delay time dependence

During pulse measurements, the diamond is first bombarded with continuous

primary electrons and no electric field is applied. Any previously trapped charge will

cause attract opposite species to transport through the diamond and recombine. This

will „clean up‟ the diamond electrically and achieve a neutral state after a very short

Emission Mode Measurement of Diamond Secondary Electron Yield 77

Figure 6.3: First pulse gain under room temperature (23°C) with different field and

effective energy for each pulse. Effective energy is the energy of the primary beam

excluding the energy loss in the metal coating.

High purity single crystal diamond is a good insulator; however free

charge-carriers inside the diamond that move according to the applied field, as in the

application of diamond detectors. There are also few trap centers in the bulk to trap

the charge carriers and therefore cause shielding to the applied field. In pulsed

measurements, the time delay of the first pulse after applying the electrical field is

critical to the amplitude of the gain.
Emission Mode Measurement of Diamond Secondary Electron Yield 78

Figure 6.4: Secondary signal with different delays for the primary electron after

applying the DC field.

Figure 6.4 shows the difference between two delay times, <1s and 3s. For all

four applied field strengths, the first pulses with longer delay time have smaller gain

at the beginning of the pulse and decreased under approximately the same rate with

the shorter delay time pulses. The gain at the beginning of the pulse is mainly defined

by the existing field in the diamond when the pulse is applied, and the decrease
Emission Mode Measurement of Diamond Secondary Electron Yield 79

afterwards shows trapping by the diamond as discussed above.

Notice that the delay time effect is significant here on the scale of seconds,

which is very long compared to the eventual use of the diamond amplifier. Thus the

data above with long time scale is just to show the exaggerated difference, but this

will probably not be an issue in the real application.

6.3.2 Temperature dependence

As the diamond samples are pure single crystal diamonds, the trap centers‟

density in the bulk is very small. This is demonstrated in the transmission

measurements where the signals are much higher than that in the emission mode.

There is another mechanism of trapping and possibly this is the dominant reason for

the signal‟s decrease with time, that is the effect of the dangling bonds. The dangling

bonds in emission measurements are serious obstacle for achieving a large gain.

Before the primary electron would reach the diamond, the dangling bonds of the

surface atom would capture electrons from nearby carbon atoms. This will generate a

hole in the valence band. The hole moves to the cathode surface under the influence

of the applied electric field. The time it takes the holes to traverse the thickness of the

diamond is under one nanosecond, and this will lead to charging up of the emitting

face and thus to a decrease of the field of the diamond. This will result in a gain

decrease due to the density of the dangling bonds.
Emission Mode Measurement of Diamond Secondary Electron Yield 80

Nevertheless, the density of ionized dangling bonds is a function of temperature

in materials, and this will be critical because the diamond amplifier will be operated at

a low temperatures in actual applications [36].

Figure 6.5: Emission signal of the diamond amplifier at 200°C.

In Figure 6.5, the diamond sample was heated to 250°C and kept under this

specific temperature for more than 3 hours to remove the H2O molecules attached on

the hydrogenated surface. The heating procedure leads to H2O removal from the

H-terminated surface of the CVD diamond as described in reference [37]. The NEA
Emission Mode Measurement of Diamond Secondary Electron Yield 81

surface of the diamond should be improved after the heating cycle. To ensure the

monolayer of hydrogen is not covered by atoms from the residual gas, the sample was

reheated to 200°C and kept under such temperature during the experiment.

Comparing with Figure 6.3, which shows room temperature (23°C) results, the gain at

the beginning of the pulse in the high temperature (200°C) environment is much lower.

Also in high temperature signals, the decrease of the gain is much faster than that

decrease observed at room temperature. This indicates that the trapping effect on the

surface of the diamond is more significant under higher temperature. Higher

temperature with lower SEY is consistent with the dangling bonds density positive

dependence over the temperature.

Under the same conditions, liquid nitrogen cooled system should give higher

gain than the above two cases.
Emission Mode Measurement of Diamond Secondary Electron Yield 82

Figure 6.6: Emission signal of the diamond amplifier at -140°C.

For the gain results shown in Figure 6.6, the sample and holder are cooled by a

liquid nitrogen cooling loop, and the temperature of the whole system is calibrated to

be at -140°C. The gain in this case is higher than Figure 6.3 under high field. With

high SEY, the system can still maintain the gain for a fairly long time. This indicates

that the field does not decrease as fast as under higher temperatures, namely many

more electrons are emitted within the pulse. Hence, to obtain the best performance of

the diamond amplifier, it should be operated with the ambient temperature as low as

Emission Mode Measurement of Diamond Secondary Electron Yield 83

Thus with proper hydrogenation, the diamond amplifier will achieve reasonable

gain (>50) in nanosecond or shorter pulses under low temperature.

6.3.3 Laser effect

In transmission measurements, the laser helped improve the gain of the diamond.

But under emission mode, the main concern comes from the dangling bond on the

hydrogenated surface. The photon‟s energy is above the energy gap between the

dangling bond and the top of the valence band. So the laser increases the ionization

probability of the dangling bonds. The photons also will heat up the surface

temperature by photon-electron interaction, and this will also increase the dangling

bond ionization by a small degree.

Figure 6.7, Figure 6.8, and Figure 6.9 are three emission temperature

measurements with laser applied and only temperature difference. All figures are

plotted with the same scale for easier comparison. All three figures show much lower

gain and faster falling edge than the corresponding temperature with no laser applied

in Sections 6.3.1 and 6.3.2. Therefore trapping centers considered under transmission

mode are less effective to the signal in emission mode, and photons create more

dangling bond ionization. The shielding from the dangling bond ionization is more

serious under emission. This is reasonable considering the high purity and careful

surface treatment of the diamond samples.
Emission Mode Measurement of Diamond Secondary Electron Yield 84

Figure 6.7: Room temperature (23°C) measurement with laser applied (532nm laser,

Emission Mode Measurement of Diamond Secondary Electron Yield 85

Figure 6.8: High temperature (200°C) measurement with laser applied (532nm laser,

Emission Mode Measurement of Diamond Secondary Electron Yield 86

Figure 6.9: Low temperature (-140°C) measurement with laser applied (532nm laser,


Focusing on the shape and amplitude of the signal, it is noticeable that the higher

the temperature, the less the difference between with and without laser results. This

also proves the dangling bond charging density is highly dependent on the

temperature. The thermal energy brought by photons is very small due to the low

power of the laser. Hence under higher temperature, the dangling bond density is

defined mainly by the heating source applied instead of the laser.
Thermal emittance measurement design 87


The ultimate emittance of a laser-photocathode gun is limited by the thermal

emittance. As laser shaping techniques improve, this limit is becoming more

important. The thermal emittance of positive electron affinity photocathodes in terms

of the thermal electron energy is a large fraction of an electron volt, depending on the

type of photocathode. The diamond amplified photocathode, being a negative electron

affinity (NEA) cathode, promises to deliver a very small thermal emittance.

Theoretical expectations as well as measurements in the reflection mode point

towards a temperature less than 0.1eV. This has never been measured in a

transmission mode. Therefore the aim is to measure the thermal emittance of

hydrogenated NEA diamond.

The thermal energy of the secondary emission beam depends on the

characteristic of the diamond sample, the properties of the hydrogenated surface and

the electric field in the diamond, and yet it is expected that the thermal energy is not a

strong function of the field. The measurement system should have the ability to

measure this small emittance.

The design is shown in Figure 7.1. The beam coming out of the diamond NEA

surface is accelerated towards an anode. The anode has a small hole, to collimate an
Thermal emittance measurement design 88

output beam. The output beam is focused on a screen by a solenoid. The main idea of

the measurement is to compensate the strong linear defocusing effect at the anode by

applying a linear focusing lens and measure the waist spot size downstream. The

waist spot size is only a function of the thermal emittance and is less affected by the

other components. Therefore, we can compare the measurement with simulation [38].

It can also be calibrated with known thermal emittance electron source such as a

thermal cathode.

Figure 7.1: Setup design for thermal energy measurement of diamond secondary


7.1 Methods

Common to all possible focusing schemes for thermal energy measurement is the

application of a small aperture on the anode plate to limit the beam size, such that the
Thermal emittance measurement design 89

beam size after a drift distance is measurable. In addition, the field applied for

acceleration is constrained by the need to obtain a good gain in the diamond amplifier.

The schematic layout of the cathode and anode is shown in Figure 7.2.

Figure 7.2: Acceleration of the secondary electrons by applying a high voltage across

the diamond and anode.

7.2 Limitations

The aperture in the anode applies a defocusing force on the electron beam, which

is mainly a linear force with a small component of nonlinear aberration. A larger

aperture will lead to a stronger aberration, and will cause an increase in the beam spot
Thermal emittance measurement design 90

size on the screen. Space charge is another issue in increasing the spot size. A small

aperture in the anode also leads to a lower current with smaller space charge effect,

increasing the accuracy of the measurement. On the other hand this will make the

imaging more difficult. From simulation results in Figure 7.3, we choose the upper

limit for each effect.
Thermal emittance measurement design 91

Figure 7.3: Aberration and space charge effects on beam spot size on screen.
Thermal emittance measurement design 92

The upper limit of the aperture and current is determined by a choice leading to a

small change in the spot size. From the simulation above, an aperture radius of 100μm

and a current of 10μA are chosen.

7.3 Comparison of lenses

The criterion of choosing the focusing lens is the growth rate in beam waist spot

size with respect to thermal energy of the cathode. A large growth rate as a function of

temperature will minimize the system errors. Results from simulation of different

kinds of focusing lenses are shown in Figure 7.4 and Figure 7.5. All curves in both

plots are simulated with same distance from anode to lens (2cm) and from lens to

screen (25.5cm) for comparison. In both plots, the solenoid has a faster growth rate

compared to other lenses. In Figure 7.4, the growth rate of a 0.5cm aperture Einzel

lens is comparable with a solenoid, but the nonlinear effect due to small aperture is

significant, and this will result in large systematic error for spot size measurement.

The simulation also shows that the closer the lens is to the anode and the further is the

screen location, the better the resolution.
Thermal emittance measurement design 93

Figure 7.4: Comparison of solenoid and two Einzel lenses.
Thermal emittance measurement design 94

Figure 7.5: Comparison of solenoid and two quadrupoles.

From above simulation results, it can be concluded that solenoid is the final choice
of focusing lens.

7.4 Shielding

All components in the measurement setup have low permeability. Shielding the

solenoid is very important to prevent a magnetic force on the electrons during

acceleration, which will reduce the resolution. Figure 7.6 shows the performance of a

solenoid with and without shielding, and compares with hard edge field.
Thermal emittance measurement design 95

Figure 7.6: Solenoid with different shielding.

Shielding material of 1mm thick with relative permeability μr~20,000 around the

solenoid coil will decrease the magnetic field at the anode to <1.5Gauss. Solenoid

without shielding gives the worst resolution (green), while the resolution of solenoid

with shielding of μr=20,000 (blue) is comparable with hard edge ideal case (red).

7.5 Alignment

The setup design is based on highly accurate alignment of the anode pinhole and

the solenoid center axis. It is very important to have the secondary electron beam,
Thermal emittance measurement design 96

after passing through the anode pinhole, symmetrically distributed around the

solenoid center axis in the transverse direction. This avoids distortion of the waist spot

on the screen.

During the alignment, a laser beam is passed through the system downstream of

the electron path. Along the laser path, the anode pinhole and three additional

250μm diameter pinholes are used to conduct the alignment. Two are mounted on the

center of the flanges at each end of the system. The third is located at the center of

the start point of the solenoid field. With the laser beam passing all four pinholes,

the offset of secondary electron beam away from the solenoid center axis is

maintained to be less than 150μm.
Thermal emittance measurement design 97

Figure 7.7: The offset magnetic field to the center of the electron beam and the waist.

FWHM difference according to the offset. Thermal energy of the beam is 0.1eV.

This procedure also ensures that the solenoid axis is perpendicular to the anode

surface, and the tilt angle can be controlled to better than 5mrad.

Figure 7.7 shows the simulation of the beam waist FWHM change due to a

different offsets of the electron beam for 0.1eV thermal energy. Thus the error caused

by offset after alignment should be controlled within 15%. A quadrupole lens may be

added to the drift section for additional correction so as to lower the error even more.
Thermal emittance measurement design 98

7.6 Conclusion

With the design of the solenoid lens based measurement system shown in Figure

7.1 and the calibrated simulation plot of FWHM waist spot size vs. thermal energy, a

CCD camera with pixel size ~5μm should be able to analyze the beam waist spot size

down to 50μm diameter. Thus, we should be able to determine the thermal energy of

diamond secondary emission to about 0.01eV.

Careful alignment and shielding is very important in minimizing the systematic


The limit of thermal energy measurement from aberration of aperture of anode

and space charge effect is 100μm in radius and 10μA for secondary electrons


By changing the focusing strength of lens, the waist of electron beam will be

imaged on screen at a specific location. For good resolution, lens should be as close to

the anode as possible, and the screen should be relatively far away from the lens.
Summary and conclusion 99


The Diamond Amplified Photocathode (DAP) is a novel approach to generating

electrons. By following the primary electron beam, which is generated by traditional

electron sources, with an amplifier, the electron beam available to the eventual

application is increased by 1 to 2 orders of magnitude in current.

Diamond has a very wide band gap of 5.47eV which allows for a good negative

electron affinity with simple hydrogenation, diamond can hold more than 2000MV/m

field before breakdown. Diamond also provides the best rigidity among all materials.

These two characters offer the capability of applying high voltage across very thin

diamond film to achieve high SEY and desired emission phase. The diamond

amplifier also is capable of handling a large heat load by conduction and

sub-nanosecond pulse input.

The preparation of the diamond amplifier includes thinning and polishing,

cleaning with acid etching, metallization, and hydrogenation.

The best mechanical polishing available can provide high purity single crystal

diamond films with no less than 100μm thickness and <15nm Ra surface roughness.

The ideal thickness for 700MHz beam is ~30μm, which requires further thinning with

RIE or laser ablation. RIE can achieve atomic layer removal precision and roughness
Summary and conclusion 100

eventually, but the time consumption for this procedure is very significant. Laser

ablation proved that with <266nm ps laser beam, the ablation process on the diamond

can easily achieve removing a few microns per hour from the surface and <100nm

roughness. For amplifier application, laser ablation is an adequate and efficient

process to make ultra thin diamond wafers following mechanical polishing.

Hydrogenation will terminate the diamond surface with monolayer of hydrogen,

and form NEA so that secondary electrons in the conduction band can escape into the

vacuum. The method is using hydrogen cracker to strike hydrogen atoms onto the

bare diamond surface to form H-C bonds.

Two independent experiments were carried out to determine the transport of the

electrons within the diamond and their emission at the surface. In transmission mode

measurements, the diamond amplifier was coated with metal on both sides, so results

simply depend only on the electron transport within the diamond. The SEY for this

mode provides one secondary electron per 20eV energy, which gives the gain of more

than 200 for 4.7keV (effective energy) primary electrons under 2MV/m. Laser

detrapping can help the signal maintain the gain with 10μs pulse and duty cycle of

1.67×10-7. In emission mode measurements, in which the diamond is prepared as in

the actual application, the SEY is ~20 for 700eV (effective energy) primary electrons

under 1.21MV/m.

The electric field applied and the primary electron energy is limited by the
Summary and conclusion 101

experiment setup, but the results show good trend toward large gain under high field.

Thermal emittance of the diamond secondary emission is critical for the beam

application. A careful design is setup to measure with very fine precision and accuracy

of 0.01eV.
Reference 102


[1] R. Kuroda, Y. Hama, S. Kasahiwagi, H. Kawai, F. Nagasawa, M. Washio, H.
Hayano, J. Urakawa, and X. J. Wang, European Particla Accelerator
Conference, Paris, France, 2002.
[2] X. J. Wang, T. S. Rao, K. Batchelor, I. Ben-zvi, and J. Fischer, Nuclear
Instruments & Methods in Physics Research Section a-Accelerators
Spectrometers Detectors and Associated Equipment 356 (1995) 159.
[3] J. Grames, P. Adderley, J. Brittian, D. Charles, J. Clark, J. Hansknecht, M.
Poelker, M. Stutzman, and M. Surles-Law, 2005 Particle Accelerator
Conference, Knoxville, TN, 2005.
[4] L. M. Wang, 25th International Conference on Microelectronics, Belgrade,
Serbia and Montenegro, 2006.
[5] F. Nava, C. Canali, C. Jacoboni, L. Reggiani, and S. F. Kozlov, Solid State
Communications 33 (1980) 475.
[6] X. Chang, Ph.D Thesis, 2005.
[7] J. R. Olson, R. O. Pohl, J. W. Vandersande, A. Zoltan, T. R. Anthony, and W.
F. Banholzer, Physical Review B 47 (1993) 14850.
[9] J. Grimes, Master of Science Thesis, 2007.
[10] I. L. Krainsky and V. M. Asnin, Applied Physics Letters 72 (1998) 2574.
[11] K. R. Lykke, K. K. Murray, and W. C. Lineberger, Physical Review A 43
(1991) 6104.
[12] M. Scheer, R. C. Bilodeau, C. A. Brodie, and H. K. Haugen, Physical Review
A 58 (1998) 2844.
[13] G. Haeffler, D. Hanstorp, I. Kiyan, A. E. Klinkmuller, U. Ljungblad, and D. J.
Pegg, Physical Review A 53 (1996) 4127.
[14] D. M. Neumark, K. R. Lykke, T. Andersen, and W. C. Lineberger, Physical
Review A 32 (1985) 1890.
[15] J. Slater, F. H. Read, S. E. Novick, and W. C. Lineberger, Physical Review A
17 (1978) 201.
[16] Z. Zhang, M. Wensell, and J. Bernholc, Physical Review B 51 (1995) 5291.
[17] J. Furthmuller, J. Hafner, and G. Kresse, Physical Review B 53 (1996) 7334.
[18] A. Laikhtman, A. Hoffman, R. Kalish, Y. Avigal, A. Breskin, R. Chechik, E.
Shefer, and Y. Lifshitz, Applied Physics Letters 73 (1998) 1433.
[19] A. Shih, J. Yater, P. Pehrsson, J. Butler, C. Hor, and R. Abrams, Journal of
Applied Physics 82 (1997) 1860.
[20] J. E. Yater and A. Shih, Journal of Applied Physics 87 (2000) 8103.
Reference 103

[21] J. Stöhr, NEXAFS spectroscopy, Springer, Berlin ; New York, 1996.
[22] G. Christian, R. Stranger, S. Petrie, B. F. Yates, and C. C. Cummins,
Chemistry-a European Journal 13 (2007) 4264.
[23] X. Chang, I. Ben-zvi, A. Burrill, J. Grimes, T. Rao, Z. Segalov, and J.
Smedley, Particle Accelerator Conference, Albuquerque, NM, 2007.
[24] B. Ziaja, A. Szoke, D. van der Spoel, and J. Hajdu, Physical Review B 66
[25] D. Dimitrov, R. Busby, J. Bruhwiler, J. Cary, I. Ben-zvi, T. Rao, X. Chang, J.
Smedley, and Q. Wu, Particle Accelerator Conference, Albuquerque, NM,
[27] W. Saslow, Bergstre.Tk, and M. L. Cohen, Physical Review Letters 16 (1966)
[28] D. Dimitrov et al, unpublished results.
[29] S. Adachi, Handbook on Physical Properties of Semiconductors, Springer,
New York, NY, 2004.
[30] L. Reggiani, S. Bosi, C. Canali, F. Nava, and S. F. Kozlov, Physical Review B
23 (1981) 3050.
[31] J. Isberg, A. Lindblom, A. Tajani, and D. Twitchen, Physica Status Solidi
a-Applications and Materials Science 202 (2005) 2194.
[32] M. Nazare and A. Neves, Properties, Growth and Applications of Diamond,
INSPEC, Inc., Stevenage, UK, 2000.
[33] Y. F. Zhang, Z. G. Zhou, J. W. Sun, and G. G. Chen, Physics Letters A 214
(1996) 76.
[34] A. S. Barnard and I. K. Snook, Journal of Chemical Physics 120 (2004) 3817.
[35] G. Bussetti, C. Goletti, P. Chiaradia, and T. Derry, Europhysics Letters 79
[36] Z. E. Smith and S. Wagner, Physical Review B 32 (1985) 5510.
[37] G. Piantanida, A. Breskin, R. Chechik, O. Katz, A. Laikhtman, A. Hoffman,
and C. Coluzza, Journal of Applied Physics 89 (2001) 8259.
[38] Q. Wu, I. Ben-zvi, A. Burrill, X. Chang, D. Kayran, T. Rao, and J. Smedley,
Particle Accelerator Conference, Albuquerque, NM, 2007.
Proceedings of EPAC 2006, Edinburgh, Scotland THPCH174

J. Y. Zhai#, C. X. Tang, S. X. Zheng, Tsinghua University, Beijing 100084, P. R. China

Abstract gun with CVD diamond cathode is presented, including
A Multipactor Electron Gun (MEG) is developed for the gun description, cathode preparation and the high
the high power microwave generation in the Accelerator power experiment. The aim of this study is to select and
Lab of Tsinghua University. This paper presents the recent prepare the appropriate cathode material in order to make
experimental results of the S-band MEG using hydrogen- the MEG beam current as high as possible.
terminated and CsI-terminated CVD diamond cathodes.
The gun design, cathode preparation and high power THE MULTIPACTOR ELECTRON GUN
experiment are described. An electron beam with 5 μs The MEG is composed of a cylindrical RF cavity
macro-pulse, 10 Hz repetition rate, greater than 900 mA working in TM010 mode, a secondary electron emission
beam current was obtained. cathode and its position adjuster, a secondary electron
emission grid with transmittance of 40 %, a waveguide
INTRODUCTION and a RF window (the picture of the MEG is shown in
Multipactor is a resonant secondary electron emission Figure 1).
discharge frequently observed in accelerator structures,
couplers, RF windows et al. It is usually undesirable and
will be suppressed in most cases. However there are a few
applications based on this effect, such as the multipactor
electron guns (MEG) [1-6], microwave frequency
multipliers [7], radar receiver protectors [8], multipactor
ion pump [9] and flat thin display [10].
The MEG can be seen as an improvement of the
thermionic RF gun. By substituting the thermionic
cathode with a secondary electron emission cathode, and
the beam exit hole with a secondary electron emission
grid, the back bombardment of the thermionic RF gun is
used to form the multipactor process. The electron beam
is produced, amplified and bunched simultaneously in the
cathode-grid gap. When the current in the cavity reaches a Figure 1: Photograph of the Multipactor Electron Gun.
steady level by space charge and beam loading effects, the
outgoing electron beams will be the saturation emission The gap distance between the cathode and the grid is
current. adjustable; as a result, the natural frequency of the cavity
The advantages of MEG are: 1) short duration, high changes rapidly. Since no frequency tuner is designed, the
current electron beams due to self-bunching mechanism MEG usually works in detuning conditions.
of multipactor; 2) fast response, simple structure, long RF parameters are measured before hot tests. At the gap
lifetime and tolerance to contamination due to cold distance of 2.73 mm, the natural frequency is 2806 MHz,
cathode. The MEG can form compact and efficient high the unloaded quality factor is 1972, the unloaded coupling
power microwave generator systems. It can also be used is 3.3, the shunt impedance is 0.175 MΩ.
in the small industrial and medical accelerators. By
replacing the grid with a two-surface hydrogen- CAHTODE PREPARATION
terminated single crystal thin diamond film, the MEG will The preparation of high performance secondary
produce electron beams of even larger current density and electron emission material is the key point of making an
good beam quality, which may be used as a high MEG. According to multipactor dynamics, the MEG has
brightness injector. This scheme is similar to the two operation statuses [12]: 1). Small current (~ 100 mA)
secondary emission enhanced photoinjector [11]. status corresponding to low impact energy (< 2 keV). 2).
The first MEG was designed in 1969 by Gallagher [1], large current (~ 1 A – 10 A) status corresponding to high
and then different kinds of such RF guns were developed impact energy (>5 keV). The MEGs in references [1] and
[2-5]. Among these guns, only the Micro-Pulse Electron [2] worked in the first status, while the MPGs in [3] and
Gun (S-band) reached a high macro-pulse current density [4] and the MEG in this study worked in the second one.
of 10 A / cm2, but the cathode material was not mentioned When MEG works in large current status, it has to use
in the literature [3]. In this paper, a multipactor electron materials having high second cross-over point energy,

*Work supported by the National High-Tech Research and Development Program of China

07 Accelerator Technology 3203
T26 Subsystems, Technology and Components, Other
THPCH174 Proceedings of EPAC 2006, Edinburgh, Scotland

which is corresponding to high escape depth of the 3002 MHz, peak output power 3.1 MW), circulators,
material. Diamond is one of the best cathode materials for directional couplers, vacuum pumps and other necessary
MEG. parts. The emission electron beam is collected by a
The excellent secondary electron emission property of faraday cup with a bias voltage of +50 V. The forward RF
CVD diamond films is due to the negative electron power from the directional coupler and the macro-pulse
affinity formed by the hydrogen or CsI termination at the current from the faraday cup are measured by a Tektronix
film surface [13]. The secondary electron emission yield 7404B oscilloscope. The repetition rate is 10 Hz. The
is also dependent on the crystal quality which is mainly pressure remained ~ 10-4 Pa during the experiment. At the
determined by the thickness of the CVD films [14]. distance of 2.5 mm, the waveforms are shown in Figure 3
The cathodes used in the MEG are polycrystalline and 4.
diamond films on Mo substrate grown by microwave
plasma assisted chemical vapour deposition (MPCVD).
The growth parameters are shown in Table 1.

Table 1: Growth Parameters of CVD Diamond Films
Gas Supply 1 % CH4 12 hrs.
Substrate Thickness 0.06 mm
Substrate Temperature 870 °C
Microwave Power 1700 W
Boron Doping (B2H4) 0.2 %

The Boron doping is to provide conductivity for the
diamond films. The hydrogenation procedure involved Figure 3: Beam current of MEG with hydrogen
exposure to H2 plasma for 30 minutes while the cesiation terminated CVD diamond cathode (bottom) and RF
procedure involved exposure to CsI vapour. Figure 2 power delivered to the cavity (top), both on a 1μs / div.
shows the SEM image of the film. The thickness of the time scale.
film is about 5 μm, and the grain size is about 0.5μm.

Figure 2: SEM image of the CVD diamond film. Figure 4: Beam current of MEG with CsI terminated
CVD diamond cathode (bottom) and RF power delivered
to the cavity (top), both on a 1μs / div. time scale.
When the RF power in the cavity (or the corresponding The macro-pulse current of the hydrogen terminated
gap voltage) grows to the susceptive region and the and CsI terminated CVD diamond cathode is 920 mA and
transient beam loading is just appropriate, the multipactor 600 mA respectively. The emission current density of the
discharge will occur [15]. The method of the high power hydrogen terminated diamond cathode is 1.2 A / cm2, and
experiment is to adjust the gap distance in order to fit the the micro-pulse current density is estimated to be 12 A /
cathode material requirement and in the same time change cm2 with the duration of 30 ps.
the power level of the RF source to initiate the The emission current decreases rapidly both for
multipactor process. hydrogen terminated and CsI terminated CVD diamond
The high power experiment system is composed of an films. This is caused by the secondary electron emission
E2V MG5349 magnetron (frequency range 2993 MHz - yield declination due to continuous high current density
electron bombardment. The emission may be restored or

3204 07 Accelerator Technology
T26 Subsystems, Technology and Components, Other
Proceedings of EPAC 2006, Edinburgh, Scotland THPCH174

enhanced if the surface is exposed to hydrogen or CsI [10] Dijk R, Vissenberg G, Zwart S. A flat thin display
environment [16-18]. This hydrogen gas feed experiment with RF electron generation. Journal of the Society
is now in progress in our lab in order to maintain the for Information Display, 2005, 13 (3): 221-225.
performance and extend the lifetime of the diamond [11] Chang X Y, Ben-Zvi I, Burrill A, et al. Study of
cathode. But the graphitization of the diamond films by secondary emission enhanced photoinjector. In:
the hydrogen ions may be a new problem [18]. Proceedings of 2005 Particle Accelerator Conference.
We are also investigating the other robust cathode IEEE, 2005. 2711-2713.
materials such as OFHC copper, Cu-Al-Mg alloy [19], [12] Zhai J Y. Initial experimental study of the
Pd-Ba alloy and MgO films [20], which may be used in multipactor electron gun. Master’s Thesis. Beijing:
MEG for the small industrial and medical linear Tsinghua University, 2006.
accelerators. [13] Yater J E, Shih A. Secondary electron emission
characteristics of single-crystal and polycrystalline
CONCLUSIONS diamond. Journal of Applied Physics, 2000, 87(11):
A multipactor electron gun has been designed, built and
[14] Dvorkin V V, Dzbanovsky N N, Suetin N V, et al.
tested in the Accelerator Lab of Tsinghua University, and
Secondary electron emission from CVD diamond
a 920 mA pulse current, 5 μs pulse length electron beam
films. Diamond and Related Materials, 2003, 12:
has been obtained using the boron doped hydrogen
terminated CVD diamond film cathodes. The cathode
[15] Kishek R A. Interaction of multipactor discharge and
lifetime is under further investigation.
RF structures: [PhD thesis]. Ann Arbor: University of
Michigan, 1997.
ACKNOWLEDGMENTS [16] Bekker T L, Dayton J A, Gilmour A S, et al.
The authors are grateful to Prof. W. Z. Tang and his Observation of secondary electron emission from
colleagues at the University of Science and Technology diamond films. IEEE IEDM Tech. Dig, 1992, 92:
Beijing for preparing the CVD diamond films. 949-952.
[17] Mearini G T, Krainsky I L, Dayton Jr J A.
REFERENCES Investigation of diamond films for electronic devices.
Surface and Interface Analysis, 1994, 21: 138-143.
[1] Gallagher W J. The multipactor electron gun. In:
[18] Mearini G T, Krainsky I L, Wang Y X et al.
Proceeding of IEEE, 57(1), 1969. 94-95.
Fabrication of an electron multiplier utilizing
[2] Liska D J. Multipactoring electron gun for high duty
diamond films. Thin Solid Films, 1994, 253: 151-156
linacs. In: Proceedings of IEEE, 59(8), 1971. 1253-
[19] Zhai J Y, Tang C X, Zheng S X, et al. Multipactor
experiment on a Micro-Pulse Electron Gun. High
[3] Mako F M. Electron gun. European Patent
Energy Physics and Nuclear Physics, 2006, 30(2):
Application, EP 0809271 A2, 1997-11-26
[4] Len L K, Mako F M. Self-bunching electron guns. In:
[20] Zhai J Y, Tang C X, Zheng S X. Experimental study
Proceedings of the 1999 Particle Accelerator
on the beam dynamics of Micro-Pulse Electron Gun.
Conference. IEEE, 1999. 70-74.
High Energy Physics and Nuclear Physics, 2006,
[5] Guharay S K, Len L K, Mako F M. High-current
30(Suppl. I): 99-101.
micro-pulse electron guns and accelerator
applications. In: Proceedings of the 2001 Particle
Accelerator Conference. IEEE, 2001. 2084-2086.
[6] Galdetskiy A V, Sazonov B V. Theoretical and
experimental investigation of cold emitters for
alenotron amplifiers. In: Proceedings of the 12th
International Conference on Microwave and
Telecommunication Technology (CriMiCo 2002).
IEEE, 2002: 172-173.
[7] Mako F M. Electron gun having multiple transmitting
and emitting sections. US Patent, 6633129, 2003-10-
[8] Carlisle T P. X band high power multipactor receiver
protector. IEEE Transactions on Microwave Theory
and Techniques, 1978, 26(5): 345-347.
[9] Yokoo K, Ono S. High-speed ion pump with a
multipactor cathode-the multipactor ion pump.
Vacuum, 1981, 32(5): 265-268.

07 Accelerator Technology 3205
T26 Subsystems, Technology and Components, Other
Proceedings of 2005 Particle Accelerator Conference, Knoxville, Tennessee

Xiangyun Chang, Ilan Ben-Zvi, Andrew Burrill, Steven Hulbert, Peter Johnson, Jorg Kewisch,
Triveni Rao, John Smedley, Zvi Segalov, Yongxiang Zhao, BNL, Upton, NY 11973 U.S.A.

Abstract photocathode are accelerated to a few thousand electron-
The secondary emission enhanced photoinjector volts by the DC field applied between the photocathode
(SEEP) is a promising new approach to the generation of and a diamond window. The primary electrons penetrate
high-current, high-brightness electron beams. A low the thin metal film deposited on the backside (left side) of
current primary electron beam with energy of a few the diamond window and produce in the diamond a large
thousand electron-volts strikes a specially prepared number of secondary electrons. The Secondary Electron
diamond window which emits secondary electrons with a Yield (SEY) can be more than 100. The secondary
current two orders of magnitude higher. The secondary electrons drift through the diamond under a high RF
electrons are created at the back side of the diamond and electric field and emerge into the accelerating proper of
drift through the window under the influence of a strong the “gun” through a Negative Electron Affinity (NEA) [1]
electrical field. A hydrogen termination at the exit surface surface of the diamond.
of the window creates a negative electron affinity (NEA) Most of the work on Secondary Electron Emission
which allows the electrons to leave the diamond. An (SEE) from diamond has been in the reflection mode
experiment was performed to measure the secondary where the primary and secondary electrons are on the
electron yield and other properties. The results are same side of the sample with no external electric field.
discussed in this paper.. But for SEEP application the electron must go through
INTRODUCTION In general, diamond may has the following defects:
foreign substitutional atoms, e.g., B, N and P, foreign
The creation of high average-current, high brightness interstitial atoms, e.g., H and Li, self-interstitial
electron beams is a key technology for a large number of interstices, vacancies and grain boundaries, which are
accelerator-based systems, such as ultra-high-power Free- formed by group of interstices.
Electron Lasers (FELs), Energy-Recovery Linac (ERL) The defects generate states with energy levels in the
light sources, electron cooling of hadron accelerators, and band gap. When charge carriers (electrons or holes) move
many more. though the lattice, they can be captured by the defects.
Fig.1 shows a schematic layout of such a device. Once the carriers are captured by defects, they can be
The laser light illuminates on a high-QE photocathode, released by thermal excitation or light. The release time
such as CsK2Sb. The primary electrons produced on the can be much longer than the capture time.

Figure 1: The schematic of the secondary emission enhanced photo-injector

* Work performed under the auspices of the U.S. Department of
Energy and partially funded by the US Department of Defense

2251 c
0-7803-8859-3/05/$20.00 2005 IEEE
Proceedings of 2005 Particle Accelerator Conference, Knoxville, Tennessee

Trapping creates space charge and partially
• Polycrystalline diamond:
compensates the external field (field shielding). It also
We first measured a Polycrystalline diamond sample
generates dark currents and heat. The trap concentration
with thickness of 200µm and diameter of 6mm provided
in the diamond must be small for the diamond amplified
by Harris International. The aluminum coating on the
cathode application.
front side has a diameter of 4mm and a thickness of
It is expected that the single crystal diamond with high
20nm, were the back side (left side) coating has a
purity would have the best electron (and hole)
diameter of 4mm and a thickness of 50nm. The diameter
of the SS plate holes is 2.5mm, to enable easy alignment
and good contacts with the coatings.
THE EXPERIMENTAL SETUP The primary electron energy ranges from 0eV to 6keV
The test setup is schematically shown in Fig.2. The and the current ranges from 0A to 5µA. No electron or
diamond is coated on both sides with aluminum. The hole transmission was measured.
stainless steel (SS) plates, which are isolated from one • Sample 1 diamond:
another by kapton films are pressed against the aluminum Another measurement was done on a diamond sample
films. They are mounted onto a sample holder and placed provided by the Apollo Company with the same
in a vacuum system. The electrodes are connected to the dimensions as the above sample. The maximum
SS plates and are led out of the vacuum so that one can transmitted hole current measured (V1 is positive while
apply voltages on the aluminum films and measure the V2 is grounded) is listed in Table 1.
currents. The primary electrons are emitted from a Table 1: Hole transmission in a single crystal
0~5keV electron gun. The primary electrons strike the diamond. Eeff is the effective primary electron
front side (right side) of the diamond with current ranging energy, Ip is the primary current and F is the
from 0 to 10µA. The minimum spot size of the primary electric field across the diamond
electron beam is less than 1mm in diameter.
Eeff (keV) Ip (µA) F (MV/m) Gain (I2/Ip)
8 0.1 20 >200
8 0.5 20 ~30
6 8.5 10 2~3

At low temperature (80K) the gain was reduced by
more than 3 orders of magnitude. The reason is
presumably that the sample quality is not good enough,
i.e., there are too many traps and grain boundaries in the

Figure 2: Schematic of the electron (or hole) Figure 3: SEM picture of the cross-section of the sample 1
transmission experiment. V1 and V2 are floating power diamond.
supplies ranging from -5kV to +5kV independently.

0-7803-8859-3/05/$20.00 2005 IEEE 2252
Proceedings of 2005 Particle Accelerator Conference, Knoxville, Tennessee

Fig.3 is the Scanning Electron Microscope (SEM) 600
picture of the cross-section of the sample 1 diamond. It’s

Electron transmission gain
clear that the substrate side (bottom) is different from the
growth side (top). The grain size is very small (<1µm) on 400 5keV
the substrate side and even on the growth side the grain 4keV
size is not very big. The small lumps on the surface (~ a 300
few µm) may also indicate small grains. 200 2keV
We got high electron transmission gain (I2/Ip~200)
under the following conditions: Ep=4keV, V1=80V,
Eeff≈4keV, Ip≈5nA and F=20MV/m. 0

• Natural diamond 0 5 10
Gradient (MV/m)
The natural diamond was purchased from a Russian
company. It has a rectangular shape, 3x2.64mm2 and a
thickness of 160 microns. The diamond is a type 2a with a
modest nitrogen concentration of about 60 ppm (1.3x10-19 Figure 5: Electron transmission through the natural
/cm3). No other contaminant atoms were reported diamond for constant primary current of 20 nA at 80K
although we suspect there may be other contaminants too. and primary energies of 2, 3, 4 and 5 keV
The diamond was checked with a microscope and was
verified to be of good crystal quality. The diamond was Fig.5 shows the plots of electron transmission gain for
coated on both faces with aluminum 30 nm thick. constant primary current of 100nA at a diamond
Fig.4 shows four plots of electron transmission gain temperature of 80K. Even at low temperatures we still get
through the diamond for constant primary current of 20 very high gain. Similar to the room temperature case, at
nA at room temperature and four primary energies: 1, 2, 3 high current through the diamond, we observe a higher
and 4 keV. It is obvious that the maximum gain decreases gain than expected just by SEY from the primary
fast as we decreases the primary electron energy. At an electrons. We are not sure what the exact mechanism for
electron energy of 1 keV the gain disappears because the that is. Other details of the curve also will require further
low energy electrons are totally absorbed in the aluminum investigation, such as the clear bend in the current vs.
coating. One also notices that the high gain curves do not field curve.”
show the saturated gain property at high field gradient. The transmission through the natural diamond was
The slope of the curve at high field gradient increases as sensitive to the primary beam position and spot size. This
the current through the diamond increases. This may led to large variation of the electron transmission between
possibly due to ionization of impurity atoms, leading to a sets of experiments as seen in the room temperature and
larger SEY. low temperature measurements.
We also measured the hole transmission by changing
300 the polarity of the field across the diamond. No hole
transmission was measured. This is because the Nitrogen
250 impurity is a deep trap for holes while a shallow trap for
electron transmission gain

200 4k,20nA
2k,20nA The measurements of charge carrier transmission
100 1k,20nA through diamond films are presented. To have better
transmission properties, a single crystal pure diamond is
50 desired. Single crystal requirement seem more important.
With better quality diamond, we expect that the electron
0 transmission through diamond will be appropriate for
0 2 4 6 8 SEEP.
Gradient (MV/m)
[1] J. E. Yater, A. Shih, and R. Abrams, “Electron
Figure 4: Electron transmission through the natural transport and emission properties of C(100)”,
diamond for constant primary current of 20 nA at room Physical Review B, V56, R4410.
temperature and primary energies of 1, 2, 3 and 4 keV

2253 c
0-7803-8859-3/05/$20.00 2005 IEEE
Nuclear Instruments and Methods in Physics Research A 610 (2009) 253–257

Contents lists available at ScienceDirect

Nuclear Instruments and Methods in
Physics Research A
journal homepage:

Investigation of the secondary emission characteristics of CVD diamond films
for electron amplification
J.S. Lapington a,, D.P. Thompson a, P.W. May b, N.A. Fox b, J. Howorth c, J. Milnes c, V. Taillandier d
Space Research Centre, Univeristy of Leicester, Leicester LE1 7RH, UK
Department of Chemistry, Univeristy of Bristol, Bristol BS8 1TS, UK
Photek Ltd., St. Leonards on Sea, East Sussex TN38 9NS, UK
Commissariat a l’Energie Atomique, SPN Bruyeres-le-Chatel, 91297 Arpajon Cedex, France

a r t i c l e in fo abstract

Available online 12 June 2009 Chemical vapour deposition (CVD) diamond offers great potential as a low-cost, high-yield, easily
Keywords: manufactured secondary electron emitter for electron multiplication in devices such as photomultiplier
CVD diamond tubes. Its potential for high secondary electron yield offers several significant benefits for these devices
Dynode including higher time resolution, faster signal rise time, reduced pulse height distribution, low noise,
Secondary electron emission and chemical stability.
Photomultiplier We describe an experiment to characterize the secondary emission yield of CVD diamond
Detector manufactured using different processes and process parameters and discuss the degradation of
secondary electron yield and experimental difficulties encountered due to unwanted electron beam-
induced contamination. We describe techniques utilized to overcome these difficulties, and present
measurements of secondary yield from CVD diamond dynodes in reflection mode.
We discuss the application of CVD diamond dynode technology, both in reflection and transmission
mode, to advanced high-speed imaging and photon-counting detectors and describe future plans in this
& 2009 Elsevier B.V. All rights reserved.

1. Introduction The performance of in vacuo secondary electron gain devices
such as PMTs and MCPs is fundamentally limited by the number
Despite the widespread success of solid-state imaging devices of secondary electrons produced by a primary electron impacting
such as the ubiquitous CCD, there are many fields where the the dynode surface. Since the gain progression is geometric, a
sensitivity to detect single-quanta accurately in time and space is technology providing improved dynode emission with the
required, a regime where signal amplification is essential. Despite stability to use at high signal currents would produce a step
advances in solid-state avalanche gain technologies, the physics of advance in the performance. Boron-doped diamond film depos-
solid-state semiconductors limit the ultimate response time and ited using chemical vapour deposition (CVD) techniques has been
signal-to-noise ratio (SNR) compared with the fundamental shown to exhibit significantly enhanced secondary electron yield
limitations for in vacuo electron gain devices, such as the (SEY), especially at higher voltages, compared to traditional
photomultiplier tube (PMT) and microchannel plate (MCP). dynode materials [2] (see Fig. 1), and its application offers
PMT and MCP detectors are utilized in many fields of science, significant advantages:
from space-based astronomical UV [1] and X-ray observations, to
microscopic analysis and diagnosis of in vivo biological processes.
1. Enhanced SEY: Diamond, along with several other high band-
PMTs are the workhorse detector for photon-counting applica-
gap semiconductors, can exhibit negative electron affinity
tions, from medical imaging as PET scanners to detection of exotic
(NEA) with suitable surface preparation, a major factor
particles as neutrino detectors. MCP detectors have the capability
enhancing SEY. However, unlike other higher band-gap
to achieve picosecond event timing resolution, and their planar
materials (e.g. AlN, Ga1yAlyN), synthetic CVD diamond is easy
geometry lends itself to high-resolution imaging, with formats
to produce [3].
exceeding 4000  4000 pixel2.
2. Simplified design: Higher SEY results in a lower number of
dynodes being required for a given gain, a distinct advantage in
terms of device size, cost, and complexity.
 Corresponding author. Tel.: +44 116 252 3498; fax: +44 116 252 2464. 3. Enhanced timing: Boron-doped (p-type) diamond surfaces
E-mail address: (J.S. Lapington). produce a narrow energy distribution of low-energy secondary

0168-9002/$ - see front matter & 2009 Elsevier B.V. All rights reserved.
254 J.S. Lapington et al. / Nuclear Instruments and Methods in Physics Research A 610 (2009) 253–257

Fig. 1. SEY coefficients of three common dynode materials (inset—Photonis PMT
Handbook) cf. CVD diamond for different surface terminations [2] and data (grey
squares) from our research programme.
Fig. 2. A schematic diagram of the SEY characterization apparatus.

electrons [4] allowing improved confinement of trajectories
with improved time resolution and reduced background [5].
4. Narrow pulse height distribution (PHD): The Poisson statistics of
high-gain diamond dynodes intrinsically produce a narrower
PHD with improved SNR, allowing detector operation at lower
gain and better energy resolution for scintillation counters, PET
scanners, etc.
5. Low noise: Its wide band gap (5.47 eV) precludes the thermal
excitation of electrons at typical operating temperatures cf.
materials with comparable SEY (e.g. GaP—2.26 eV). Conversely,
it has lower thermal noise than traditional dynode materials at
high temperatures.
6. Large-area: CVD deposition offers a low-cost, large-area coat-
ing technique and is applicable to shaped substrates c.f.
expensive materials such as GaN and GaP, which can only be
grown on lattice matched substrates e.g. sapphire. Inkjet
printing of diamond nanoparticles and subsequent growth
into patterned films has also been demonstrated.
7. Stability: CVD diamond has a stable SEY, little affected by long-
term storage. Its SEY remains high after exposure and it can be
reactivated [6].

One potential drawback of diamond’s higher gain is the
increased voltage required per stage. However, diamond’s smaller
pulse height variance and low noise promote operation at lower
gain with a smaller number of stages, thus alleviating the overall
voltage requirement.

2. Experimental setup

The SEY characterization apparatus is a bakeable stainless-
steel vacuum system with a centrally mounted, electrically
insulated conductive arm supporting up to four 1 cm2 samples
and a small Faraday cup (SFC) for beam current calibration. Fig. 2
shows a schematic diagram of the apparatus and Fig. 3a shows a
photograph of the arm with three samples and the SFC aperture.
The sample support arm can be moved laterally by 725 mm
and rotated 3601 around its axis. Lateral motion is used to position
one of the samples or the Faraday cup in the path of a beam of
electrons generated by an electron gun and rotation allows the
beam incidence angle to be varied or the samples placed out of
Fig. 3. (a). A photograph of the sample arm with three samples. The SFC is the
sight of the beam. The ELS5000 electron gun, manufactured by
aperture between sample 1 and 2. The montage below is the false colour SEY
PSP Vacuum Technology [7], produces a focussed electron beam image produced by scanning the electron beam. Fig. 3b is an SEY image showing
with a spot size of o50 mm at 1 mA, working distance of the SFC aperture (the black hole in the middle).
J.S. Lapington et al. / Nuclear Instruments and Methods in Physics Research A 610 (2009) 253–257 255

15–75 mm, and beam energy of up to 5 keV. The beam can be degraded rapidly over time-scales of minutes as a result of
electrostatically deflected in the xy plane by 75 mm. exposure to the electron beam. Evidence in the form of visible
Secondary electrons produced by the electron beam impinging contamination, known as electron beam-induced contamination
the sample are collected by a stainless-steel cylinder, the large [8], from areas subject to long-term exposure, pointed strongly to
Faraday cup (LFC), which surrounds the arm. The LFC is coaxial the presence of organic contamination. This occurred in two
with the support arm and is mounted on an insulating support at forms, both highly structured, suggesting interaction of the beam
its closed end. The electron beam enters through a small aperture with adsorbed contaminants on the sample surface, and as diffuse
in the side of the LFC. The LFC can (a) be biased and, (b) has shadows whose geometry suggested either the presence of a
extended length to minimize the solid angle of the external large-angle component in the beam distribution or isotropic
chamber seen by emitted secondary electrons from the sample. evolution of contaminants from a remote source back along the
Both measures serve to ensure secondary electrons from the beam line. At this stage the vacuum system was cleaned and the
sample are captured by the LFC. Both the sample support arm and pump set replaced with an oil-free system comprising an Edwards
the LFC are electrically insulated allowing them to be biased and Maglev turbo and scroll pump, though the ultimate vacuum
their respective collected currents to be measured using a Keithley achieved was not appreciably lower.
6514 Electrometer and 6485 Picoammeter, both operating as Another mechanism for SEY degradation is hydrogen deso-
precision ammeters. These can be interchanged, however, only the rption from the diamond surface, which causes the disappearance
electrometer input can be biased. of NEA [9]. Since both electron beam-induced contamination and
Two independent methods can be used to calibrate the hydrogen desorption are produced by exposure to the electron
electron beam current, (i) the beam can be collected completely beam and are sensitive to total beam flux, they are difficult to
in the SFC, the sample current equalling the beam current, distinguish, but their effect can be alleviated if the total flux can
confirmed by a zero LFC current, or, (ii) the beam current can be be reduced. This approach was used to enable measurement of the
calculated from the sum of sample and LFC currents. The former nominal SEY of an initially ‘‘clean’’ material. The overall flux was
method requires the electron beam to be sufficiently well reduced by both pulsing the beam and lowering the beam current.
focussed to completely pass through the SFC aperture of LabView was used for the software platform to automate the
0.5 mm. Focussing was carried out by scanning the beam in two measurement procedure, which was used to control the electron
dimensions across the SFC aperture using the beam deflection gun HV, the xy beam deflection, the beam focus, and to initialize
capability, and imaging the aperture by means of the measured and take synchronised measurements from the Keithley am-
SEY, calculated using the sample and LFC currents. The SEY was meters, process the data, and save and display results. Several
then given by modes of operation allowed: (1) automated beam focussing, (2) xy
area beam scan and SEY image capture, (3) energy scan at a single
SEY ¼ (1) point to produce SEY versus beam energy data, and (4) other
Isample þ ILFC
calibration procedures. The low-noise Keithley ammeters and
The focus was adjusted to achieve zero SEY over the aperture, good electrical insulation allowed electron beam currents in the
the expected result when the beam is fully collected in the SFC. 1–10 nA range to be used even when several hundred volt biases
Elsewhere, the phosphor-bronze shim from which the SFC was were applied to the LFC. The HV was typically pulsed for 0.3 s with
constructed produced an SEY of 1.0. This technique also allowed the ammeters taking measurements every 10 ms. The time profile
the beam profile to be calculated by deconvolution of the circular of measurements was used to select a window where the currents
SFC pinhole from the imaged SEY (see Fig. 3b). had reached a stable regime.
These modifications drastically reduced the rate of SEY
degradation with total beam charge. Fig. 4 shows a plot of the
3. Results variation of SEY for a single 100 mm spot illuminated by the
electron beam versus total deposited charge. The contribution of
Initially the system was pumped with a Leybold turbo pump the two SEY degradation mechanisms to this data is uncertain,
and rotary backing pump, reaching an ultimate pressure of however, we have qualitative proof that, despite the 
107 mbar. Early on in the experiment it was noted that the decontaminated vacuum system and oil-free pump set, electron
measured SEY from any sample material, unless already very low, beam-induced contamination is still present. This is evidenced by


Secondary Electron Yield





0.00 0.50 1.00 1.50 2.00 2.50
Extracted Charge (mC)

Fig. 4. A plot showing the degradation of SEY with increasing electron beam charge.
256 J.S. Lapington et al. / Nuclear Instruments and Methods in Physics Research A 610 (2009) 253–257

Fig. 5, which shows a photograph of the SFC after prolonged showed no trace of actual elongation of the beam distribution that
exposure to the electron beam, producing a diagonal dark stain. this deposit implies. This would be apparent as elongation of the
However, concurrent SEY imaging of the area around the SFC SFC aperture in the SEY image owing to convolution of the beam
and aperture distributions.
With the apparatus modified to reduce the SEY degradation
effects, we measured the SEY of a variety of diamond samples
from various sources, manufactured using different methods
(CVD, inkjet-printed nanoparticles), dopants and dopant concen-
trations, and differing crystallinities. Fig. 6 shows a plot of the best
results obtained for boron-doped, CVD diamond coated on silicon
and molybdenum substrates. The best SEY value so far obtained of
45 for B-doped diamond on silicon at 2.4 keV is the highest
reported for hydrogen-terminated material and a strong driver to
apply this technology to electron multiplication. However, in the
presence of degradation due to electron beam-induced
contamination, these must be regarded as preliminary.
Fig. 5. A photograph of the SFC before (left) and after (right) prolonged However, measurements of variation of SEY with angle of
illumination with the electron beam. incidence show unexpected behaviour (see Fig. 7). Typically, SEY



Secondary Electron Yield


30 Sample A
Sample B





0 0.5 1 1.5 2 2.5 3
Electron Beam Energy (keV)

Fig. 6. Measured SEY versus beam energy data for hydrogen-terminated B-doped CVD diamond deposited on molybdenum (sample A) and silicon (sample B) substrates.

0 deg
15 deg
30 deg
45 deg
60 deg
Secondary Electron Yield

75 deg






0 0.5 1 1.5 2 2.5 3
Electron Beam Energy (keV)

Fig. 7. Measured SEY versus energy for hydrogen-terminated B-doped CVD diamond deposited on molybdenum, with the electron beam at various angles of incidence.
J.S. Lapington et al. / Nuclear Instruments and Methods in Physics Research A 610 (2009) 253–257 257

is expected to increase with angle of incidence (defined as the guaranteed electron collision at every dynode, ideally suited for
angle with respect to the surface normal). This is due to the imaging applications. It is technologically more demanding;
reduced depth of penetration normal to the surface at higher challenges being mechanical support of very thin films or
angles, resulting in an increase in secondary electron escape incorporation of internal drift fields.
probability. Either, the high granularity of the surface plays a role
in reversing this trend or, as has been suggested by other authors,
the secondary electron escape mechanism from polycrystalline 5. Conclusions
diamond is different, utilizing transport along grain boundaries
[10]. We have obtained SEY results from B-doped hydrogen-
terminated CVD diamond comparable or better than reported
elsewhere. However, we are still encountering electron beam-
4. Future developments
induced contamination effects within the apparatus and are
continuing to develop counter measures to this limitation, while
We have plans to investigate the application of diamond as
in parallel, attempting to establish the relevance of SEY reduction due
secondary electron multiplication dynodes in actual detectors
to hydrogen desorption. The scale of the latter is the more serious
operating in two modes:
issue, being unavoidable in a detector context in reflection mode.
Successful demonstration of high SEY from a transmission dynode
1. Transmission: primary electrons enter through one surface of a will circumvent this problem since hydrogen desorption from the
thin film of diamond, and secondary electrons exit through the input surface should not affect the NEA of the output surface.
other. Primary track lengths at typical energies coupled with
the minority carrier diffusion length requires (a) film thick-
nesses o200 nm, or (b) thicker films with an internal drift field
to increase the secondary diffusion length.
[1] J.S. Lapington, S. Chakrabarti, T. Cook, J.C. Gsell, V.T. Gsell, Nucl. Instr. and
2. Reflection: diamond is deposited on an open conductive wire Meth. A A513 (2003) 159.
mesh or electroformed grid. Primary and secondary electrons [2] J.E. Yater, A. Shih, R. Abrams, J. Vac. Sci. Technol. A 16 (1998) 913.
enter and exit through the same diamond surface. Coated mesh [3] P.W. May, Phil. Trans. R. Soc. London A 358 (2000) 473.
[4] J.B. Miller, G.R. Brandes, J. Appl. Phys. 82 (1997) 4538.
dynodes (not CVD diamond) are already used in some imaging [5] Jon Howorth, Photek Ltd., private communication.
PMTs. [6] G.T. Mearini, I.L. Krainsky, J.A. Dayton Jr., Surf. Interface Anal. 21 (1994) 138.
[7] PSP Vacuum Technology, Macclesfield, UK, /www.pspvacuum.comS.
[8] A.E. Vladar, M.T. Postek, Microsc. Microanal. 11 (Suppl. 2) (2005) 764.
The transmission technique is conceptually superior, providing [9] J.B. Cui, J. Ristein, L. Ley, Phys. Rev. B59 (1999) 5847.
a linear detector design with minimum charge spread and [10] J.E. Yater, A. Shih, J.E. Butler, P.E. Pehrsson, J. Appl. Phys. 96 (2004) 446.

Multipactor discharge on metals and dielectrics: Historical review
and recent theories*
R. A. Kishek†,a)
Institute for Plasma Research, ERB#223, University of Maryland, College Park, Maryland 20742
Y. Y. Lau,b) L. K. Ang, A. Valfells, and R. M. Gilgenbach
Department of Nuclear Engineering and Radiological Sciences, University of Michigan, Ann Arbor,
Michigan 48109-2104
~Received 18 November 1997; accepted 12 January 1998!
This paper reviews the history of multipactor discharge theory, focusing on recent models of
multipactor accessibility and saturation. Two cases are treated in detail: That of a first-order,
two-surface multipactor, and that of a single-surface multipactor on a dielectric. In both cases,
susceptibility curves are constructed to indicate the regions of external parameter space where
multipactor is likely to occur, taking into account the dependence on surface materials, and the
effects of space charge and cavity loading. In the case of a dielectric, multipactor is found to deliver
about 1% of the rf power to the surface. The two cases are contrasted in light of experimental
observations. © 1998 American Institute of Physics. @S1070-664X~98!92205-4#

I. INTRODUCTION used for the protection of sensitive receivers.
In this paper, we review recent advances at the Univer-
Multipactor discharge is a resonant vacuum discharge sity of Michigan in the theoretical understanding of multi-
frequently observed in microwave systems such as radio fre- pactor. The discharge is modeled from first principles, taking
quency ~rf! windows,1–6 accelerator structures,7,8 microwave into account both space charge forces and interaction with
tubes and devices,6,9–12 and rf satellite payloads.13–15 The the structure ~loading and detuning!. To our knowledge, this
discharge can occur for a wide range of frequencies, from the is the first theory that includes this interaction between the
MHz range to tens of GHz, and in a wide array of geom- multipactor discharge and the surrounding rf structure. Our
etries. The underlying mechanism behind the multipactor theory employs a realistic model of secondary electron emis-
discharge is an avalanche caused by secondary electron sion for various surface materials. To put the recent theories
emission. Primary electrons accelerated by rf fields can im- in their proper context we start with a brief historical review
pact a surface and release a larger number of secondary elec- of multipactor theory ~Sec. II!. Afterwards we describe the
trons, which may in turn be accelerated by the rf fields and recent theories on accessibility and saturation of multipactor
made to impact again, releasing even more electrons, and so discharge. The presentation is divided into two sections: one
on. The discharge can take place on a single surface or be- dealing with two-surface multipactor ~Sec. III!, and another
tween two surfaces. concentrating on the special, yet commonplace, case of
Multipactor is usually undesirable. It can dissipate sub- single-surface multipactor on dielectrics ~Sec. IV!. The two
stantial amounts of energy fed into microwave cavities, thus cases are contrasted in the conclusion.
lengthening the conditioning process. It can detune a micro-
wave signal ~with detrimental consequences for space com-
munications systems!. The discharge also heats the surface, Multipactor has been observed as early as 1924 by the
possibly increasing noise levels in crossed-field devices and Guttons ~see Ref. 20!, but was not identified until Farnsworth
perhaps causing damage. Through various mechanisms mul- a decade later, who coined the name ‘‘multipactor’’ ~from
tipactor may result in window breakage, destroying the ‘‘AC Electron Multiplier’’! for an amplifier he invented
vacuum. In some circumstances ~see Ref. 12!, multipactor based on the discharge ~see Ref. 21!. Observations of multi-
may even induce vacuum breakdown. It is worth noting that, pactor in gaseous breakdown experiments in 1930s and
despite all of the above, there is a growing number of appli- 1940s led to many early theories. Most notable are those of
cations for the discharge. For example, the inherent resonant Daniellson, and of Henneburg et al., in 1936 ~see Ref. 20!.
bunching in multipactor has been explored for use in electron The latter derive the resonance condition on the transit time
gun technology.16–18 The capability of multipactor to induce for electrons emitted with zero initial velocity, and also iden-
breakdown is under study for application in plasma display tify single particle phase focusing and stability of the dis-
technology, while its capability to dissipate energy has been charge.
Perhaps the first systematic study of multipactor came
from Gill and von Engel of Oxford University in the 1940s,22
*Paper kWeaI2-6 Bull. Am. Phys. Soc. 42, 1949 ~1997!.

Invited speaker. in which they experimentally outlined part of the region sus-
Electronic mail: ceptible to multipactor and, in conjunction, advanced a
Electronic mail: theory that recognized the sensitivity of the solution to a

1070-664X/98/5(5)/2120/7/$15.00 2120 © 1998 American Institute of Physics
Phys. Plasmas, Vol. 5, No. 5, May 1998 Kishek et al. 2121

nonzero emission velocity of secondaries. In order to avoid
the complications of a distribution of random emission ve-
locities, Gill and von Engel introduced the ad hoc assump-
tion that a parameter k, equal to the impact velocity of the
primaries relative to the emission velocity of secondaries, is
constant. There is no physical basis for this assumption. Fur-
thermore, Gill and von Engel’s attempt at using the theory to
interpret their experimental results was flawed, and it was
not until Hatch and Williams reformulated the theory in the
1950s to explain their own multipactor experiments20 that
reasonable agreement was obtained. Since Hatch and Will-
FIG. 1. Dependence of secondary electron yield on impact energy.
iams retained the constant ‘‘k’’ assumption of Gill and von
Engel, the modified theory became known as the ‘‘constant-
k’’ theory and for decades remained the classic theory on the until recently, few satisfactory theories have been
accessibility of multipactor,23,24 because of its utility in con- developed.40,41
structing susceptibility curves. Neither the trajectory computations nor the accessibility
More recently, Vaughan21 has promoted an alternative to theories say anything of the saturation level of the multipac-
the constant-k theory, based rather on that by Henneburg tor current or the time scale over which multipactor evolves.
et al., and derived from first principles. Vaughan’s theory Many early researchers have suspected that debunching due
replaces the baseless assumption of constant k with the more to space charge forces can saturate the multipactor,1,16,17,22
realistic assumption of a monoenergetic nonzero initial ve- but it was not until Vaughan in 1988 that anyone attempted
locity. Other researchers19,25–30 have since then adopted to analytically estimate the level of saturation.21 Mako18 and
Vaughan’s theory. Riyopoulos et al. extended it to include a Riyopoulos27 also assumed space charge debunching of the
crossed-magnetic field. In Sec. III we use this theory in de- multipactor to be the cause of saturation, although each
riving the susceptibility curve, and extend it to include the treated this effect somewhat differently. None of these satu-
effects of materials and of cavity loading. ration theories accounted in any way for the important pro-
In the decades following the rf breakdown experiments cesses of loading and detuning of rf structures by the multi-
of Hatch and Williams, multipactor proliferated in all types pactor. In Sec. III B we shall look at the first theory to
of rf devices, and its suppression became a major concern. explore this loading effect in resonant structures, while in
Many of the advances in understanding multipactor and the Sec. IV we shall look at the first complete theory on acces-
suppression techniques, however, remained proprietary to in- sibility and saturation of multipactor on dielectric windows.
dustries, and hence relatively little survives in the open
literature. Currently, there are three general approaches III. TWO-SURFACE MULTIPACTOR ON METALS
for suppressing multipactor: conditioning of the
In this section, we address the subject of a multipactor
components by the discharge,6 coatings and other
discharge between two metallic surfaces. For simplicity, we
‘‘surface treatments,’’6,8,11,31,32 and geometrical modifi-
assume two parallel plane surfaces, separated by a gap of
cations.10,21,33–35 Conditioning is a lengthy process that de-
width D. A perpendicular rf electric field, of the form
pends on the discharge affecting the surface characteristics (V g0 /D) sin (2p ft1u), where f is the frequency of the rf
enough to quench itself. In general, surface treatments are and V g0 is the amplitude of the voltage across the gap, ac-
aimed at reducing the effective secondary electron yield to celerates electrons inside the gap. Upon impact, secondaries
inhibit the growth of multipactor, but unfortunately cannot are released with a monoenergetic nonzero initial velocity of
be applied in all situations and are likely to degrade over a few eV ~in Sec. IV we further include a random distribu-
time.21 The most promising method so far has been modify- tion of emission velocities!. The secondary electron yield, d,
ing the geometry ~e.g., moving the location of a rf is a function of impact energy and angle of the primary
window21,35!, with the goal of altering the trajectories of electron.42 In this paper, we adopt Vaughan’s empirical for-
electrons so as to eliminate multipactor. mula for the yield.43 The dependence on impact energy is
Despite the wealth of experimental data, the advances in illustrated in Fig. 1 where the yield as a function of impact
multipactor theory moved at a much slower pace. The avail- energy is determined by the two material-dependent param-
ability of more powerful computer platforms led to the emer- eters E max and d max . These two parameters determine two
gence of many computer codes to model multipactor and aid energies for which the yield is unity, called the first and
in designing multipactor-free structures.8,33,34 Yet even to- second cross-over points, E 1 and E 2 , respectively ~shown in
day, the multipactor codes are limited to calculating single- the figure!. Clearly, multipactor grows only for impact ener-
electron trajectories subject to a given rf field. Only recently gies in between E 1 and E 2 , where d .1.
are particle-in-cell ~PIC! codes being applied to
A. Susceptibility curves
multipactor.18,26,27,36 The accessibility theories have also
been extended to various types of geometries, such as For the multipactor discharge to be sustained, electrons
single-surface1 or coaxial37–39 geometries. Single-surface just released from one plate must be accelerated strongly
multipactor on dielectrics poses a special problem,1–5 and enough to reach the other plate, and do so at a time when the
2122 Phys. Plasmas, Vol. 5, No. 5, May 1998 Kishek et al.

field has reversed ~or is close to doing so! in order that the
secondary electrons released from the impact also be accel-
erated. This consideration leads to a condition on the transit
time of the electrons,20,21,25 namely, that the transit time
should be near an odd number, N, of 21 rf cycles for multi-
pactor to take place, where the number N is called the order
of the multipactor. This transit time condition ~or synchrony
condition! relates the rf voltage and frequency to the geom-
etry, allowing us to map the regions of external parameter
space in which multipactor is possible. Such maps are called
susceptibility curves in the literature ~e.g., Fig. 3, which will
be described in greater detail below!. Typically, the gap volt- FIG. 2. Universal curve to determine the upper limiting voltage, V gmax , for
age is plotted against the product of the frequency and gap a first-order (N51), two-surface multipactor assuming a general nonzero
monoenergetic emission energy E 0 ~in eV!. Here, f is the frequency in GHz
separation ( f D). For each order of the multipactor, a pair of
and D is the gap separation in cm.
lines defines the region in which multipactor is possible.
Given the frequency and the geometry, the curve immedi-
ately provides us with the voltage range for multipactor Reformulating the results of Gill and von Engel, we
(V gmin ,Vgmax). For zero emission velocity in a parallel plate show in Fig. 2 the correction to Eq. ~1b! introduced by the
geometry, the voltage boundaries scale as ( f D) 2 and hence nonzero initial velocity.19 Note that this is valid only for a
they appear as straight lines of slope 2 on a log–log plot.21,30 first-order (N51), two-surface multipactor. With this graph,
The voltage boundaries V gmin and V gmax can be derived it becomes a simple matter to calculate the upper bound
by a simple integration of the force law, imposing the transit V gmax in the case of a nonzero initial velocity. First, one uses
time condition x(t5N/2 f )5D. This results in an equation Eq. ~1b! to calculate V gmax0 for a zero initial velocity. Next,
relating the gap voltage amplitude to the external parameters one locates the point on the graph in Fig. 2 that corresponds
f and D, as well as the launch phase, u, and the emission to the given frequency, gap separation, and monoenergetic
energy, E 0 , of secondary electrons. The orbits are stable nonzero initial velocity, then reads off the correction factor.
only for a narrow range of launch phases, thus defining the Multiplying this correction factor by V gmax0 gives the desired
voltage boundaries. The minimum voltage, V gmin , is ob- upper bound. Note from Fig. 2 that the importance of a non-
tained at the maximum stable phase u 5 u m zero initial velocity decreases for higher ( f D) products. This
5arctan (2/N p ). The maximum voltage, V gmax , is obtained is a direct result of the higher voltages and, hence, the higher
at the maximum negative phase that allows an electron impact energies at higher ( f D) products. Figure 3 is an ex-
launched with nonzero initial velocity to reach the other ample of a susceptibility curve showing the resulting voltage
plate. A negative phase is possible if the emission velocity is boundaries for a first-order multipactor as pairs of diagonal
nonzero and the electric field changes sign before the elec- lines, one for each of a number of monoenergetic initial ve-
tron hits the plate of origin. The maximum negative phase is locities, v 0 .
that which results in the electron reversing, going back to the In addition to the boundaries on the voltage presented
plate of origin, and reversing again just before touching that above, another factor ~which has been largely ignored in the
plate. literature thus far! further constrains the existence of two-
Although such curves that include the dependence of surface multipactor. This originates from the consideration
V gmin or V gmax on the emission energy, E 0 , have been that the secondary electron yield must be greater than unity
published,21 no explicit formulas for them have thus far been
provided. For easy reference, we display, in physical units,
V gmin in Eq. ~1a!44 for general emission energy E 0 , and
V gmax in Eq. ~1b! only for the special case E 0 50 ~therefore
we call it V gmax0!:30

22480~ f D ! 2 2 ~ N p f D ! A44960E 0
V g min5 , ~1a!
A~ N p ! 2 14
V g max 05 ~ f D !2, ~1b!

where the voltages are in volts, the frequency f is in GHz,
the gap width D is in cm, and the secondary-electron emis-
sion energy E 0 is in eV. There is no simple closed-form
solution of V gmax for nonzero E 0 , but it is possible to solve FIG. 3. Susceptibility curve for a two-surface multipactor showing the ef-
fects of nonzero emission velocity and surface materials. This example is
for it numerically. Gill and von Engel had solved the result- constructed for oxygen-free copper ~d max51.3; E max5600 eV! and alumina
ing transcendental equations, expressing the maximum nega- ~d max56.5; E max51300 eV!, assuming a monoenergetic emission energy,
tive phases in terms of their ‘‘k’’ parameter.22 E 0 52 eV ~—!; 5 eV ~-•-!; and 10 eV ~...!.
Phys. Plasmas, Vol. 5, No. 5, May 1998 Kishek et al. 2123

for the avalanche to occur, and the impact energy has to be in
the vicinity of E 1 ~Ref. 28!, even for voltages in the range
(V gmin ,Vgmax). Since E 1 depends on the surface materials,
these surface materials provide additional conditions on the
existence of multipactor.30 The first such condition is that the
starting impact energy cannot be below E 1 ~or else d ,1!.
The second condition on the impact energy arises from dy-
namic considerations. Our extensive simulations ~with Q in
FIG. 4. Model of a two-surface multipactor and interaction with rf circuit.
the range 10–1000 and d max51.2! show that if the starting
impact energy is greater than about 1.33E 1 ~depending on
many particulars!, then cavity loading alone is insufficient in
saturating the multipactor, and the discharge appears spo- The motion of the electron sheet is driven by the modified rf
radically. It is possible in such a situation for the multipactor voltage across the gap, as well as by the evolving image
to saturate by space charge debunching, as investigated by space charge forces.
Vaughan21 and in greater depth by Riyopoulos.27 Unfortu- We have later extended this model to account for the
nately, the latter papers completely neglect the effect of load- mutual repulsion between the space charges through the ad-
ing, which would in fact be significant in the intense space dition of a second electron sheet.29 The two-sheet model has
charge. revealed a novel phase-focusing mechanism caused by the
We now proceed to translate these boundaries onto the dependence of the secondary electron yield on impact energy
susceptibility curve. Examples for oxygen-free copper and ~Fig. 1!. This phase-focusing mechanism results in the
alumina are provided in Fig. 3. Ignoring any nonzero initial charge density of one sheet growing at the expense of the
velocity of secondaries, the impact energy can be related to other in our two-sheet model, thus focusing the electrons into
the driving voltage ~in the absence of multipactor! by solving a narrow bunch despite their mutual repulsion. This demon-
the force law for the impact energy. When the gap voltage is strates the adequacy of the single-sheet model, at least for the
at the upper voltage boundary ~V g0 5V gmax0 and u 50! elec- regime Q*10, d max&1.5 that we have studied.
trons experience the maximum possible impact energy. Set- Extensive simulations based on our one-dimensional
ting that equal to E 1 gives the first boundary, the solid ver- ~1-D! model show that the interaction of the multipactor with
tical line on the susceptibility curve ~see Fig. 3!. Similarly, the rf circuit, a combination of detuning and de-Q-ing, is the
the lower voltage boundary, at V gmin and u min , results in the dominant saturation mechanism for resonant rf structures
minimum possible impact energy. Setting that to 1.33E 1 (Q*10). The impact energy in the steady state is equal to
gives the second boundary, which we represent with the dot- the first cross-over point energy, E 1 , of the surface material.
ted vertical line in Fig. 3. The boundaries imposed by a given The simulations have been augmented with an analytical
material are then30 derivation of the steady-state power dissipated in multipactor

~for a resonantly driven, first-order discharge, the most pre-
E1 4 E1
0.15•N , f D,0.17 N1 , ~2! dominant case!.30 The formula shown in Eq. ~3! gives the
100 Np2 100 fraction of rf power consumed by the steady-state multipac-
where f is in GHz, D in cm, and E 1 in eV. These suscepti- tor as a function of external parameters. This agrees well
bility curves, derived from a dynamic theory, show the quali- with the simulation results and predicts that anywhere from
tative features of the experimentally observed data.19 1%–50% of the input rf power can be dissipated in the mul-
P m 2B1 AB 2 24AC m̄
B. Saturation 5 • , ~3!
Pd 2A r̄
Once the conditions for existence of multipactor are sat- where

isfied, any seed electron can trigger the discharge. The ava-
lanche of electrons cannot go on unchecked, however, and 1 1 p 2 2
A5 ~ m̄ 2 n̄ 0 ! 2 1 2 2 ~ m̄ 1 n̄ 0 ! ,
sooner or later a saturation mechanism will set in. As shown 4 2 4 p
in Sec. II, prior researchers have proposed space charge
forces to be the ~only! saturation mechanism.16–18,21,27 While B5 ~ m̄ 2 2 n̄ 20 ! ,
this is a likely mechanism in some cases, the loading of rf p
structures by multipactor can play a significant role, espe-
cially in resonant structures. We have proposed a model of
the multipactor discharge that includes such interaction with
C5 F 1 p 2
2 ~ m̄ 1 n̄ 0 ! 1 ~ m̄ 2 n̄ 0 ! 2 2r̄ 2 .
2 4 4
the structure.28 In this model, which is illustrated in Fig. 4, a Here, P m is the power consumed by the steady-state multi-
single infinitesimally thin electron sheet travels in one di- pactor, P d is the drive power, m̄ 5( A2E 1 /m)/2p f D is the
mension inside a gap between two parallel plates. Through normalized impact velocity corresponding to the first cross-
Ramo’s theorem,45 the motion of the electrons induces a wall over point, n̄ 0 5 n 0 /2p f D is the normalized emission veloc-
current, I m (t), which in turn loads the rf structure, repre- ity of the secondaries ~assumed monoenergetic!, and r̄
sented in our model by a RLC circuit with a quality factor Q. 5eV g0 /m(2 p f D) 2 is the normalized steady-state gap volt-
2124 Phys. Plasmas, Vol. 5, No. 5, May 1998 Kishek et al.

age in the absence of multipactor. Note that the power dissi-
pated in a multipactor discharge in the steady state is essen-
tially a function of just two external parameters: m̄ , which is
related to the surface material, and r̄, which is related to the
energy supplied to the structure.
Although Eq. ~3! is simplified by assuming a resonantly
driven cavity, our model is capable of simulating off-
resonance drive, and has been used to derive a more general
formula for the latter case.46 Since multipactor is a resonant
discharge, it can only exist for a narrow band of frequencies
around the resonant frequency of the cavity. The width of
this frequency band is found to be proportional to the gap
voltage amplitude and inversely proportional to the Q of the
The slow response time and the large amount of energy
stored in a highly resonant cavity can lead to difficulties in
FIG. 5. Model of a single-surface multipactor in a parallel rf and normal dc
powering such a cavity, considerably lengthening the condi-
electric fields.
tioning time. Simulations with Q;1000 show that once the
voltage reaches a multipactor-prone region, it locks in as the
multipactor avalanche dissipates all the additional power
energy higher than the emission energy. Their impact energy
supplied.30 These simulations corroborate experimental ob-
and angle, both of which determine the yield, depend on the
servations since, in practice, it typically takes hours or days
phase of the rf and the random initial velocity and angle. The
for the surface to become sufficiently conditioned to quench
Monte Carlo simulation tracks a macroparticle over a large
the multipactor and allow the voltage to rise to its full design
number of transits, and for every transit assigns it a random
initial velocity and angle according to realistic
Gopinath et al. subsequently used a 1-D PIC code to
distributions.40 Over a long time, we can see an average
simulate multipactor in a resonant cavity.36 Their more so-
trend of growth or decay in the charge density of that mac-
phisticated simulations verify the predictions of our simple
roparticle, depending on the external parameters of the simu-
model, namely that cavity loading by the multipactor can
lation: E rf0 , E DC , the frequency, and the material used.
result in saturation with an average impact energy in the
With the aid of such Monte Carlo simulations, we were
vicinity of the first cross-over point. Moreover, they have
able to construct a universal susceptibility curve ~Fig. 6!
concluded that a full distribution of initial velocities pro-
showing the regions susceptible to multipactor for any com-
duces the same results as a monoenergetic velocity that is the
bination of external parameters and for any dielectric mate-
average of such a distribution,36 thus validating our assump-
rial. In Fig. 6, the dc electric field is plotted on one axis and
the rf electric field on the other. The fields are normalized to
the frequency of the rf and also to E max , a property of the
IV. SINGLE-SURFACE MULTIPACTOR ON surface material. The figure displays a set of boundaries, a
DIELECTRICS pair for each value of d max0 ~the value of d max for normal
incidence! of the surface material, inside of which multipac-
In contrast to metal surfaces, dielectric surfaces are ca-
pable of supporting a distribution of fixed charges. The pres-
ence of fixed charges creates dc electric fields which change
the relationship between the electron motion and the rf. In
this section, we focus on a situation where the rf electric
field, E rf0 sin(2p ft1u), is parallel to a dielectric surface ~as
is usually the case near transmission windows!.1,40 We as-
sume that the surface has a net positive charge which sets up
a restoring dc electric field, E DC ~Fig. 5!. In addition, we use
Monte Carlo methods to simulate a realistic distribution of
initial velocities for the secondaries. As will be seen, the
susceptibility curve and the saturation mechanism in this
case are very different from those of the two-surface multi-
pactor discussed in Sec. III.
A. Susceptibility curve
Electrons emitted from the surface with a random initial FIG. 6. Universal susceptibility curve for a single-surface multipactor on a
dielectric as a function of the dc and rf electric fields, shown for various
velocity and angle will be attracted back to the surface by the values of d max0 ~from the innermost boundaries, d max051.5, 2.0, 2.5, 3.0,
dc electric field ~Fig. 5!. During transit, they gain energy 6.0, and 9.0!. This assumes a distribution of emission energies for secondary
from the rf electric field and so impact the surface at an electrons, peaked at an energy E om /E max050.005.
Phys. Plasmas, Vol. 5, No. 5, May 1998 Kishek et al. 2125

tor grows. This curve can be quite useful in designing mi-
crowave components to avoid multipactor. Furthermore, as
will be shown shortly, this curve can be used to predict the
saturation level of multipactor and the amount of energy dis-
sipated, and therefore the heating of the dielectric due to
Physically, such a curve is simple to understand. The
electrons gain their energy from the rf. Hence, if the rf elec-
tric field is too low ~high!, the average impact energy of
electrons will become below E 1 ~above E 2 !, and ~in both
cases! the yield will drop to below unity. Hence the lower
boundary corresponds to the first crossover point, and the
upper boundary corresponds to the second crossover point.
Since the dc electric field is the restoring force, increasing it
reduces the transit time and, consequently, the energy gained
from the rf. Therefore the rf field must linearly increase with
an increasing dc field to maintain the same yield. Finally, as
FIG. 7. Evolution of a single-surface multipactor on a dielectric. Point A
d max0 increases, the range of impact energies for which d designates an initial combination of electric fields that permit multipactor
.1 widens, and so does the multipactor region. growth. At saturation, point B is reached.

B. Saturation ~only weakly dependent on the details of the distribution of
In constructing the susceptibility curve ~Fig. 6! we have emission energies! and is of the order of 0.5%–1% of the rf
assumed that the charge on the window, and therefore the dc power.41 Yet this amount can be sufficient to overheat the
electric field, is static and unchanging. In reality, the second- window and lead to breakage, and the results of this theory
ary emission process of the multipactor itself creates addi- compare favorably with experimental measurements.35,41
tional surface charging. A fully dynamic model has to ac-
count for such variations in the surface charge density, as V. FURTHER DISCUSSION AND CONCLUSION
well as for the changing rf field due to the loading by the A. Comparison of multipactor on dielectrics to that
multipactor. Such a model has been very recently on metals
developed41 and used to predict the dynamic evolution and
It is instructive now to contrast the two cases of two-
saturation of multipactor on a dielectric. The loading of the rf
surface multipactor on metals, and single-surface multipactor
is accounted for using a transmission line model. The surface
on a dielectric. Although in both cases the discharge is found
charge responsible for the dc electric field is now broken into
to saturate with an impact energy in the vicinity of the first
two parts: a small fixed charge necessary to initiate the mul-
cross-over point, E 1 , the saturation mechanisms are quite
tipactor, and a time-varying part that is created by the mul-
different. Resonant cavities are quite sensitive to loading and
tipactor secondaries leaving the surface. Space charge forces
detuning by the multipactor, which eventually lead to its
between the multipactor macroparticle and the surface are
saturation. The large amount of energy and slow response
included in this model.
time make the passage of the cavity voltage through narrow
It is evident that there are two mechanisms responsible
multipactor bands very difficult, as multipactor can readily
for saturation ~Fig. 7!. Loading of the rf changes only the rf
grow to a large extent and ‘‘lock in’’ the voltage. Dielectric
electric field and is indicated on the susceptibility diagram as
windows, on the other hand, are much less sensitive to such
a movement along the vertical line from A to C in Fig. 7.
loading and saturate primarily by space charge forces. The
The space charge forces and the dynamic growth of the sur-
proximity of the charges to the dielectric surface, unlike the
face charge density correspond to a change in the dc electric
two-surface configuration, further aids the space charge
field rather than the rf, and are indicated by a movement on
forces. The most striking difference, however, is the absence
the horizontal line from A to B. Since saturation is a combi-
of the resonance condition on the transit time in the single-
nation of these two effects, it must occur at the lower bound-
surface case. This greatly widens the parameter space for
ary, where the impact energy averages around the first cross-
multipactor and makes the dielectric surface much more sus-
over point. Simulations have revealed that the effect of the
ceptible. Even though the power dissipated by multipactor on
loading is very small in this case ~the dielectric window is a
a dielectric window is relatively small, the poor heat conduc-
nonresonant structure!, and therefore saturation occurs near
tivity of dielectrics possibly makes it sufficient to destroy the
point B. These simulations, which follow the dynamic evo-
lution, confirm the validity of the susceptibility curves de-
rived from the kinematic model ~Fig. 6!.
B. Summary
It is now straightforward to estimate the surface charging
of the dielectric, as well as the power deposited on the di- In light of the historical development of multipactor
electric surface, by the multipactor. The remarkable result is theory, the theories presented in this paper survey the first
that this power is independent of most external parameters attempt on multipactor susceptibility curves and saturation
2126 Phys. Plasmas, Vol. 5, No. 5, May 1998 Kishek et al.

from dynamic considerations, taking into account the inter- IEEE Trans. Microwave Theory Tech. MTT-42, 558 ~1994!.
action with the surrounding rf structure and the material
W. J. Gallagher, Proc. IEEE 94 ~Jan. 1969!.
W. J. Gallagher, IEEE Trans. Nucl. Sci. NS-26, 4280 ~1979!.
properties of the surface. Computer simulations have re- 18
F. Mako and W. Peter, IEEE Proceedings of the 1993 Particle Accelerator
vealed the time evolution of the discharge, the dependence Conference 2702 ~Institute of Electrical and Electronics Engineers, Piscat-
on materials, a new phase-focusing mechanism among mul- away, NJ, 1993!.
tiple electron sheets, and the relative importance of the vari- R. A. Kishek, Ph.D. dissertation, University of Michigan, Ann Arbor, MI,
ous saturation mechanisms. The results of this study are well 1997.
A. J. Hatch and H. B. Williams, J. Appl. Phys. 25, 417 ~1954!; theory later
in agreement with experimental observations. modified in A. J. Hatch and H. B. Williams, Phys. Rev. 112, 681 ~1958!.
J. R. M. Vaughan, IEEE Trans. ED-35, 1172 ~1988!.
E. W. B. Gill and A. von Engel, Proc. R. Soc. London, Ser. A 192, 446
In the development of these theories, we have benefited 23
S. Brown, Basic Data of Plasma Physics ~American Institute of Physics,
from conversations with C. M. Armstrong, J. Benford, C. K. New York, 1959, reprinted 1994!, p. 202–221.
Birdsall, M. Brennan, K. Brown, D. Chernin, W. Getty, V. P. 24
A. L. Gilardini, J. Appl. Phys. 78, 783 ~1995!.
Gopinath, E. Haebel, J. Haimson, J. P. Holloway, S. Isa-
S. Riyopoulos, D. Chernin, and D. Dialetis, Phys. Plasmas 2, 3194 ~1995!.
S. Riyopoulos, D. Chernin, and D. Dialetis, IEEE Trans. Electron Devices
gawa, T. Lavine, A. Lombardi, F. Mako, R. Miller, R. ED-44, 489 ~1997!.
Poirier, R. Rimmer, S. Riyopoulos, J. Shiloh, J. J. Song, G. 27
S. Riyopoulos, Phys. Plasmas 4, 1448 ~1997!.
Thomas, J. P. Verboncoeur, J. Wang, and P. Wilson. 28
R. Kishek and Y. Y. Lau, Phys. Rev. Lett. 75, 1218 ~1995!.
This work was supported by the Multidisciplinary Uni-
R. A. Kishek and Y. Y. Lau, Phys. Plasmas 3, 1481 ~1996!.
R. A. Kishek, Y. Y. Lau, and D. Chernin, Phys. Plasmas 4, 863 ~1997!.
versity Research Initiative ~MURI!, managed by the Air 31
J. W. Noe, Nucl. Instrum. Methods Phys. Res. A 328, 291 ~1993!.
Force Office of Scientific Research and subcontracted 32
D. Proch, D. Einfeld, R. Onken, and N. Steinhauser, IEEE Proceedings of
through Texas Tech University, by the Department of Navy the 1995 Particle Accelerator Conference ~Institute of Electrical and Elec-
Grant N00014-97-1-G001 issued by the Naval Research tronics Engineers, Piscataway, NJ, 1995!, p. 1776.
C. M. Lyneis, H. A. Schwettman, and J. P. Turneaure, Appl. Phys. Lett.
Laboratory, by the Northrop Grumman Industrial Affiliates
31, 541 ~1977!.
Program, and also by the Department of Energy. 34
C. Pasotti, P. Pittana, and M. Svandrlik, Proceedings of the IEEE 1997
Particle Accelerator Conference Vancouver, to be published ~Institute of
1 Electrical and Electronics Engineers, Piscataway, NJ!.
D. H. Preist and R. C. Talcott, IRE Trans. Electron Devices ED-8, 243
~1961!. Robert Alan Rimmer, Ph.D. dissertation, Lancaster University, Lancaster,
J. R. M. Vaughan, IEEE Trans. Electron Devices ED-8, 302 ~1961!. UK, 1988.
A. S. Gilmore, Microwave Tubes ~Artech House, Norwood, MA, 1986!, p. 36
V. P. Gopinath, J. P. Verboncoeur, and C. K. Birdsall, ‘‘Multipactor elec-
474. tron discharge physics using an improved secondary emission model,’’
S. Yamaguchi, Y. Saito, S. Anami, and S. Michizono, IEEE Trans. Nucl. submitted to Phys. Plasmas.
Sci. 39, 278 ~1992!. 37
R. Woo and A. Ishimaru, J. Appl. Phys. 38, 5240 ~1967!.
Y. Saito, S. Michizono, S. Anami, and S. Kobayashi, IEEE Trans. Electr. 38
R. Woo, J. Appl. Phys. 39, 1528 ~1968!.
Insul. 28, 566 ~1993!. 39
E. Somersalo, P. Yla-Oijala, and D. Proch, IEEE Proceedings of the 1995
J. Tuckmantel, C. Benvenuti, D. Bloess, D. Boussard, G. Geschonke, E. Particle Accelerator Conference ~Institute of Electrical and Electronics
Haebel, N. Hilleret, S. Juras, H. P. Kindermann, J. Uythoven, C. Wyss, Engineers, Piscataway, NJ, 1995!, p. 1500.
and M. Stirbet, IEEE Proceedings of the 1995 Particle Accelerator Con- 40
R. A. Kishek and Y. Y. Lau, Phys. Rev. Lett. 80, 193 ~1998!.
ference ~Institute of Electrical and Electronics Engineers, Piscataway, NJ, 41
L. K. Ang, Y. Y. Lau, R. A. Kishek, and R. M. Gilgenbach, ‘‘Power
1995!, p. 1642. deposited on a dielectric by multipactor,’’ to appear in IEEE Trans.
Linear Accelerators, edited by P. M. Lapostolle and A. L. Septier ~North Plasma Sci. ~1998!.
Holland, Amsterdam, 1970!, p. 917. 42
C. K. Birdsall and W. B. Bridges, Electron Dynamics of Diode Regions
K. J. Kleman, Proceedings 1993 Particle Accelerator Conference ~Insti- ~Academic, New York, 1966!; also O. Hachenberg and W. Brauer, Adv.
tute of Electrical and Electronics Engineers, Piscataway, NJ, 1993!, p. 924.
Electron. Electron Phys. XI, 413 ~1959!.
J. R. M. Vaughan, IEEE Trans. Electron Devices ED-15, 883 ~1968!.
J. R. M. Vaughan, IEEE Trans. Electron Devices ED-36, 1963 ~1989!;
M. Yoshida, S. Isagawa, Yo. Takeuchi, M. Ono, M. Sato, K. Takata, and
verified experimentally in A. Shih and C. Hor, IEEE Trans. Electron De-
H. Baba, Proceedings 6th Symposium on Accelerator Science and Tech-
vices ED-40, 824 ~1993!.
nology, Tokyo, Oct. 1987 ~Iomico, Tokyo, 1987!, pp. 126–128.
We remark that Eq. ~1a! is not exact for nonzero initial velocity, because
E. G. Schweppe, R. Bachmor, and E. Demmel, IEEE Proc. of the 1993
Particle Accelerator Conference ~Institute of Electrical and Electronic En- in such a case the maximum stable phase u m is no longer given by the
gineers, Piscataway, NJ, 1993!, p. 1178. simple expression arctan (2/N p ) derived for zero emission velocity, but is
12 nevertheless sufficiently accurate.
F. Hohn, W. Jacob, R. Beckmann, and R. Wilhelm, Phys. Plasmas 4, 940
S. Ramo, Proc. IRE 27, 584 ~1939!; also W. Shockley, J. Appl. Phys. 9,
P. F. Clancy, Microw. J. 77 ~March, 1978!. 635 ~1938!.
A. D. Woode and J. Petit, Microw. J. 142 ~January, 1992!. 46
August Valfells, R. A. Kishek, and Y. Y. Lau, Phys. Plasmas 5, 300
N. Rozario, H. F. Lenzing, K. F. Reardon, M. S. Zarro, and C. G. Baran, ~1998!.
Pres. PAC'09 Conf., Vancouver, BC BNL 90592-2009-CP
4-8 May (2009)


R. Busby, D.A. Dimitrov, and J.R. Cary
Tech-X Corporation
Boulder, CO 80303

I. Ben-Zvi, X. Chang, J. Keister, E. Muller, T. Rao, J. Smedley, and Q. WU
Brookhaven National Laboratory
Upton, NY 11973-5000


*This manuscript has been authored by Brookhaven Science Associates, LLC under
Contract No. DC-AC02-98CHI0886 with the U.S. Department of Energy. The United
States Government retains, and the publisher, by accepting the article for publication,
acknowledges, a world-wide license to publish or reproduce the published form of this
manuscript, or allow others to do so, for the United States Government purposes.

This work was prepared as an account of work sponsored by an agency of the United
States Government. Neither the United States Government nor any agency thereof, nor
any of their employees, nor any of their contractors, subcontractors, or their employees,
makes any warranty, express or implied, or assumes any legal liability or responsibility
for the accuracy, completeness, or any third party's use or the results of such use of any
information, apparatus, product, or process disclosed, or represents that its use would
not infringe privately owned rights. Reference herein to any specific commercial
product, process, or service by trade name, trademark, manufacturer, or otherwise, does
not necessarily constitute or imply its endorsement, recommendation, or favoring by the
United States Government or any agency thereof or its contractors or subcontractors.
The views and opinions of authors expressed herein do not necessarily state or reflect
those ofthe United States Government or any agency thereof

R.Busby, D. A. Dimitrovl, J. R. Cary, Tech-X Corp., Boulder, CO 80303, USA
I. Ben-Zvi, X. Chang, J. Keister, E. Muller, T. Rao, J. Smedley, Q. Wu, BNL, Upton, NY 11973, USA

Abstract the number of primary electrons. The free charged carriers
also experience inelastic scattering involving the emission
The Relativistic Heavy Ion Collider (RHIC) contributes
or absorption of phonons (lattice excitations). If an exter-
fundamental advances to nuclear physics by colliding a
nal field is applied in the diamond, the free electrons and
wide range of ions. A novel electron cooling section, which
holes will start drifting in opposite directions. Large num-
is a key component of the proposed luminosity upgrade
bers of the secondary electrons are transported through the
for RHIC, requires the acceleration of high-charge elec-
diamond and emitted from one of its surfaces. Diamond
tron bunches with low emittance and energy spread. A
emission surfaces are prepared to have a negative electron
promising candidate for the electron source is the recently
affinity (NEA) in order to increase the emission. The elec-
developed concept of a high quantum efficiency photoin-
trons are emitted into the accelerating cavio/ of an electron
jector with a diamond amplifier [1]. To assist in the devel-
opment of such an electron source, we have implemented
To optimize the performance of the DAP concept, it is of
algorithms within the VORPAL particle-in-cell framework
considerable interest to understand the generation of secon-
for modeling secondary electron and hole generation, and
daries and their transport in diamond samples with different
for charge transport in diamond. The algorithms include
characteristics (thickness, concentration and type of impu-
elastic, phonon, and impurity scattering processes over a
rities, surface effects at both metal coated and NEA sur-
wide range of charge carrier energies. Results from simula-
faces) and how a diamond amplifier couples to the overall
tions using the implemented capabilities will be presented
operation of the complete photoinjector system. To address
and discussed.
this problem, we have been developing models, within the
VORPAL [4] computational framework, for simulation of
INTRODUCTION secondary electron generation and charge transport in dia-
mond. Initially, we implemented models for the inelastic
A new concept for a photo-cathode with a diamond am- scattering leading to generation of secondary electrons and
plifier was recently proposed [1] and is currently under ac- holes [5]. Here, we report on our work to enable the sim-
tive development [2, 3]. The diamond amplified photo- ulation of charge transport in diamond. The ability to sim-
cathode (DAP) has demonstrated the potential to address ulate both secondary electron generation and charge trans-
the need for high peak and average current, high brightness port allows us to investigate electron gain from diamond in
and low thermal emittance electron beams in current and a transmission mode setup (D. A. Dimitrov et al. in these
future accelerator-based systems. Moreover, the DAP has proceedings).
negligible contamination problems, requires two orders of
magnitude less laser power, and is expected to have a long
lifetime in comparison to existing photo-cathodes. MODELING SECONDARY ELECTRON
The idea for the DAP operation consists of generating GENERATION AND TRANSPORT
first a beam of electrons using a conventional photocath- The short time scale (of the order of 100 fs) dynamics of
ode. These are the primary electrons that are accelerated electrons in diamond with medium range of energies (from
to energies of about 10 keY before impacting a diamond about 50 eV to 10 keV) was treated by considering elastic
window. When these primary electrons enter into the dia- and inelastic scattering processes, with the latter leading
mond, they experience inelastic scattering processes lead- to generation of secondary electrons and holes [6, 7, 8].
ing to the generation of secondary electrons and holes. Any For the inelastic scattering, we implemented first the Ash-
secondary electron or a hole that is produced with a suf- ley model described by Ziaja et al. [6, 7] and considered
ficiently high energy can also participate in such inelastic secondary electron generation effects with it [5]. Recently,
processes, generating more electron-hole (e-h) pairs. These we also implemented the TPP-2 optical model for the in-
inelastic scattering processes continue until none ofthe free elastic scattering that produces scattering rates in better
electrons and holes has sufficiently high energy to gener- agreement [8] with the rates from band-structure calcula-
ate additional e-h pairs. However, the total number of pro- tions [9] at low energies (f'.J 10 eV). For the elastic scat-
duced secondary electrons is generally much larger than tering in the medium energy range we used the rates given
* We are grateful to the US DoE for supporting this work under the
in Ref. [8] but our implementation [5] is based on methods
DE-FG02-06ER84509 grant. from semiconductor device simulations [10]. The scatter-
t ing rates for the elastic and the inelastic (calculated with
the TPP-2 model) processes are shown in the top plot of tic phonons and the three optical phonons considered are
Fig. 1. shown in the bottom plot of Fig. 1 at T = 300 K. The
9 and f type rates are for transitions between parallel and
perpendicular valleys, respectively. Decreasing the temper- '" ,.
ature of diamond reduces these rates [11]. For comparison,
the rates for scattering of electrons from charged impuri-
ties (using the model from Ref. [11]) at two concentrations
is also shown together with the electron-phonon scattering
rates. The data plotted in Fig. 1 shows that in the medium
electron energy range, the rates for secondary electron gen-
eration and for elastic scattering are several orders of mag-
nitude larger than the electron-phonon and charged impu-
rity scattering ones. However, at low electron energies (10

eV and lower) the secondary electron generation rates are
comparable to the electron-phonon rates. Note that mod-
« •. imp. deDII. 8e2O
eling of electron transport with energies close to the bot-
•• , imp. deDII. 1.64e20 tom of the conduction band does not include elastic scat-
:.: : : : : : ~t tering [9] from the periodic potential of the carbon ions as
_. g3Abs
_ g3Emit done for the medium energy range [6].
_. t3Ab.
- f3Emit We have implemented a general Monte Carlo algorithm
t2Emit to handle all scattering processes (it also includes null col-
lisions [10] to increase the efficiency). In between scatter-
ing events, the particle-in-cell algorithms in VORPAL [4]
provide the capability to self-consistently move particles
interacting with electromagnetic fields.

1~.01 0.1
Energy (eV)
In this Section, we discuss representative results from
simulations with the implemented models on electron"
Figure 1: The top plot shows the inelastic scattering rates transport after secondary electrons and holes are generated
for generation of secondary electrons and holes (calculated using one of the models for inelastic scattering. We used
from the TPP-2 model) together with the elastic rates in the the Ashley model for the results here and primary electron
medium energy range. The bottom plot shows the rates for energy of 2.7 keY. We reported effects related to the sec-
electron-phonon (low energy inelastic) and impurity (elas- ondary electron generation with this model in a previous
tic) scattering processes. study [5] and present only new results from the transport
simulations here.
The free (conduction band) electrons in diamond also ex- The domain was split into a three-dimensional grid, typ-
perience low-energy inelastic scattering with phonons (lat- ically of 24 x 24 x 24 cells, with a cell edge size of 0.1
tice excitations) and elastic scattering from charged impuri- J.1m. For the simulations described here, a small time step
ties. Charge transport cannot be simulated accurately with- (usually 8 x 10- 18 s) was used during the creation of the
out including the electron-phonon scattering. These inelas- secondary electrons (to resolve the mean free time for these
tic processes lead to additinal relaxation of the the electron processes) which happens during the first few hundred fem-
energies to reach the drift state in an applied field. At low toseconds. Phonon scattering was not enabled during this
fields, free electrons have energies close to the bottom of time interval. At the end of this interval, the VORPAL data
the conduction band and drift with temperature effectively was dumped, the input file modified to increase the time
equal to the temperature of the diamond crystal (a highly step (for faster simulation runs while still resolving the
desirable property for generation of beams with low ther- mean free times of the low energy scattering processes),
mal emittance). and to enable scattering of the free charged carriers with
For the low-energy inelastic scattering with optical and phonons. The total propagation time of these simulations
acoustic phonons we implemented Monte Carlo algorithms was approximately 20 ps.
based on the models given by Jacoboni and Reggiani [11]. We specifically studied the dependence of the drift ve-
In diamond, these models approximate the energy bands locity obtained with the phonon scattering models on the
with a parabolic dependence on the electron wave vector applied external field magnitudes' in a range values from
and have ellipsoidal constant energy surfaces. Both in- 0.05 to 10 MV/m. This is a range of interest for use in
tervalley and intravalley electron-phonon processes are in- diamond amplified photo-cathodes. Our goal was to ver-
cluded. The rates for the emission and absorption of acous- ify that the models can describe the transport of electron
clouds with drift velocities in agreement with experimental framework for simulation of secondary electron generation
measurements for this range of applied fields. and transport in diamond. We presented results from trans-
port simulations to determine the electron drift velocity de-
pendence on the applied field. These simulations test the
implemented electron-phonon scattering models. The drift
0.8 velocity obtained from VORPAL agrees well with experi-
~ mental data and more detailed band-structure calculations
'=0. 0 .6 for the fields of interest. This gives us confidence that
• the electron-phonon models we implemented in VORPAL
.s 0.4
are adequate to simulate electron transport in diamond for
studying DAP properites. In future developments, we will
~ investigate how to incorporate models of electron emission
~ 0.2
from diamond surfaces with negative electron affinities and
effects related to trapping of electrons in diamond.
6 10 12
t Ips)
[1] I. Ben-Zvi, X. Chang, P. D. Johnson, 1. Kewisch, and T. Rao.
Figure 2: The drift velocity evolution at two applied fields Secondary emission enhanced photoinjector. C-AD Acceler-
indicate the rate of convergence to the (time) average values ator Physics Report C-A/AP/149, BNL, 2004.
(given in the legend). [2] X. Chang, 1. Ben-Zvi, A. Burrill, S. Hulbert, P. D. Johnson,
1. Kewisch, T. Rao, 1. Smedley, Z. Segalov, and Y. Zhao.
The relaxation behavior of the drift velocity to steady Measurement of the secondary emission yield of a thin di-
state from simulations at two field values is shown Fig. 2. amond window in transmission mode. In C. Horak, editor,
The results show that it takes of the order of a few pi- 2005 Particle Accelerator Conference. IEEE, number 2251-
coseconds to reach the steady state. The relaxation time 3,2005.
decreases when increasing the applied field. [3] X. Chang, I. Ben-Zvi, A. Burrill, J. Grimes, T. Rao,
Z. Segalov, 1. Smedley, Q. WU Recent Progress on the Dia-
mond Amplified Photo-cathode Experiment In 2007 Particle

• '. Accelerator Conference. IEEE, pp. 2044-6, 2007.
[4] C. Nieter and 1. R. Cary. VORPAL: a versatile plasma simu-

• lation code. 1. Comput. Phys., 196:448-473, 2004.

i 0
• [5] D. A. Dimitrov, R. Busby, D. L. Bruhwiler, J. R. Cary, 1. Ben-
Zvi, T. Rao, X. Chang, 1. Smedley, Q. Wu 3D Simulations of
~ Secondary Electron Generation and Transport in a Diamond
~ • Amplifier for Photocathodes In 2007 Particle Accelerator
Conference. IEEE, pp. 3555-7,2007.

• 0.1 1 10
[6] B. Ziaja, D. van der Spoel, A. Szoke, and J. Hajdu. Auger-
electron cascades in diamond and amorphous carbon. Phys.
Electric Field Magnitude (MV/m) Rev. B, 64(214104-1/8), 2001.
[7] B. Ziaja, A. Szoke, D. van der Spoel, and J. Hajdu. Space-
time evolution of electron cascades in diamond. Phys. Rev. B,
Figure 3: The obtained average electron drift velocity vs 66:024116-1/9, 2002.
applied field from VORPAL simulations of electron trans- [8] B. Ziaja, R. A. London, and 1. Hajdu. Unified model of
port in diamond agrees well with experimental data. secondary electron cascades in diamond. 1. Appl. Phys.,
97:064905-1/9, 2005.
The dependence of the average electron drift velocity [9] T. Watanabe, T. Teraji, T. Ito, Y. Kamakura, and K. Taniguchi
on the applied field magnitude obtained from the VOR- Monte Carlo simulations of electron transport properties of
PAL simulations with the models we implemented for the diamond in high electric fields using full band structure 1.
electron-phonon scattering is shown in Fig. 3. These val- Appl. Phys. 95:4866-74, 2004.
ues agree well with experimental data and the more de- [10] M. Lundstrom. Fundamentals of Carrier Transport. Cam-
tailed, band-structure Monte Carlo simulation results given bridge University Press, second edition, 2000.
by Watanabe et ale [9].
[11] C. Jacoboni and L. Reggiani The Monte Carlo method for
the solution of charge transport in semiconductors with ap-
SUMMARY plications to covalent materials Rev. Mod. Phys., 64:645-705,
We described here the overall modeling capabilities
we have implemented within the VORPAL computational
Oscillator-Generator Powered by
Secondary Electron Emission
By Bruce Darrow Gaither,
Tulsa, Oklahoma

The instant invention discloses a generator of electricity which is
produced by a self-oscillating electron multiplier wherein the startup
power is replaced by secondary electron emission. It is a purpose of
the present application to present an oscillator which will amplify the
initial input of electrons through the use of secondary emission
coatings on the electrodes. It is further an objective of the invention
to apply part of the electrons so multiplied as feedback to the input of
the vacuum tube, so that the oscillatory multiplication of electrons
will continue under self-power and without any further input of
electricity from outside sources.

In 1902, Austin and Starke reported that the metal surfaces impacted
by cathode rays emitted a larger number of electrons than were

Numerous subsequent inventions applied the secondary emission
phenomena to various uses, such as film sound, signal amplification,
scintillation counting, radar jamming and a multiplicity of other

In the 1930’s radio tubes began to apply secondary emission
materials in vacuum tubes to amplify signals and multiply electrons.
The technique of coating electrodes in tubes with materials which
exhibit a high secondary emission coefficient continues to today, and
it is upon this foundation that the present invention depends.

When an electrode coated with emissive material is bombarded with
electrons then it emits additional electrons in a multiplicative manner.
Early devices depended upon metal oxides of various sorts and those
materials exhibited secondary electron emission in the range of 4 to 8
fold multiplication of initial electron input.
Since the 1930’s new materials and configurations have been tested
and modern secondary emission materials can exhibit secondary
emission coefficients as high as 60-100 fold multiplication.

Thus, classic tube designs have been given a new vitality by the
advances in materials and component parts. A higher secondary
emission coefficient results in an additional multiplication of initial
electron input, and that means that a greater portion of the multiplied
electrons can be diverted from the output of vacuum tubes and
reapplied to the input of the tube, so that outside power sources can
be turned off once the device is started.

The present device improves upon the configuration disclosed by the
Farnsworth Multipactor Oscillator and Amplifier. In the various
devices employing that configuration cathodes and anodes were
coated with secondarily emissive materials. Cathodes were placed on
each end of the tube and a positively-biased anode was placed in the
middle. The cathodes on each end were then continually reversed in
polarity, so that one end was negative while the other was positive,
and then they were reversed in charge. This change in polarity was
accomplished at a frequency which resulted in initial electron input
being multiplied as it bounced back and forth between the electrodes.

The instant invention constitutes an improvement upon the classic
configuration because the modern materials allow the oscillator to be
used as a generator of electricity while the device is being powered by
it’s own internally-produced current.

Wound around the outside of the vacuum tube is an electrical coil.
This serves to focus the stream of electrons as they near the midpoint
of the tube where the positively charged anode is located. If the anode
is in the form of a ring then the focused stream of electrons are
attracted in the direction of the anode but their focus causes them to
miss the ring and instead pass through the bullseye and then continue
toward the second electrode.

The anode may also be formed of a conductive substance in the form
of a wire mesh. The electromagnetic lens is still used. In either
configuration of the anode, the electrodes which impinge upon the
anode may be extracted from the tube and applied to output. In either
ring or grid anode configuration there is provided ample room for the
electrons to pass through and for the anode to be permeable to

The relationship between the travel time for the electron clouds of
secondary electrons and the frequency of the oscillation of the
cathodes is important. The electron movement and the frequency of
the alternating polarity to the cathodes needs to be in phase so that
the oscillations of the tube itself will continue. So the time of the
cycle of the electricity applied from the outside to the two cathodes
must be twice the time of flight for the cloud of electrons.

The cloud of electrons is not, of course, in one location at any time.
They occupy an amorphous space between the electrodes. Thus the
problem of phase control between the travel time of the cloud and the
frequency of the applied voltage can never be exact. But the frequency
of the movement of the electron cloud tends to predominate over the
applied voltage to a certain extent. Nevertheless it is important to
keep the two frequencies in phase as much as possible.

It should be noted that the electrons, once they pass all they way
through the tube and strike the other cathode, multiply and produce
another cloud of electrons which are thereupon accelerated toward
the first cathode. The cycle of endless multiplication results in a flow
toward the anode due to the increased electron density, and the
anode will tend to extract electrons in each half cycle in an amount
equal to the excess of secondary of primary electrons.

Eventually space charge will build up to the point that an equilibrium
level will be reached within the tube.

Electrostatic beam focusing by using a coil wound around the middle
of the tube will tend to lessen the space charge by minimizing the
spread of the electrons as they near the anode in the center.

Philo Farnsworth patented many secondary emission devices,
including oscillators. He remarked that in a tube where the two
cathodes were 5.5 centimeters apart, and applied frequency to the
cathodes was 30 megahertz and 100 megahertz that the secondary
emission oscillator required input power of 350 volts and 800 volts

As stated in US patent 2,168,052, “It is now well known that
multipactors or electron multipliers can be made to generate self-
sustaining oscillations, methods of causing such oscillations in
various modes having been described in the patents of Philo T.
Farnsworth, No. 2,071,516 granted February 23, 1937, and No.
2,128,580 granted August 30, 1938, among others. Circuits utilized in
causing these tubes to generate such oscillations are, in general, of
great simplicity, usually involving merely a parallel resonant circuit
connecting two of the elements of the tube, together with a source of
direct current and potential for supplying the primary energy. In
general, within reasonable limits, ease of starting oscillation and
stability of oscillation when started in tubes of this character
increases with the size of the tube, with the voltage applied to the tube,
and with the current drawn.”

As a general design consideration a smaller transit distance for the
electrons makes it easier to accelerate the electrons, while a larger
oscillator makes space charge saturation less of a factor. So any
configuration of the instant invention would involve an optimal
combination of these two competing design considerations.

A simple battery can be used to provide the needed power for the
present invention, with a Cockroft-Walton voltage multiplier being
interposed between the battery and the tube in order to step up the
voltage to the required level.

This kind of ladder voltage multiplier is composed of a series of
capacitors and diodes. These electronic components are cheap and
readily available on the open market. The well-known configuration
tends to double the voltage at each step, so that it is referred to as a
ladder. Each succeeding stage doubles the voltage from the previous
stage and simultaneously cuts the previous amperage in half. Atomic
accelerators used this kind of voltage multiplier to achieve high
voltages for their early-day experiments. They are designed to take a
low voltage and to step it up to a high voltage. But the battery will
supply the power necessary to energize the Cockroft-Walton circuit
which then makes a bias of positive polarity for the anode in the

The regime of the oscillator is to use low voltage for the cathodes and
high voltage for the positive anode.

The oscillator multiplies power, meaning that the number of electrons
are increased, and the amperage increased. Part of the multiplied
output of the oscillator is used to recharge the battery, part of the
output is used, if necessary to power the frequency resonant circuit
which flip-flops the cathode polarity. The device tends to be self-
oscillating once it begins operating, and the mere maintenance of the
anode in positive bias tends to attract free electrons without input of
additional electrons.

The type of oscillator circuits used by the Farnsworth inventions can
be replaced by modern crystal oscillators or ceramic oscillators, which
are ubiquitous and cheap today.

Thus, new coating materials result secondary emission coefficients
ten or twenty times higher than the materials used in the original
Farnsworth oscillators. In addition, semiconductors and transistors
replace the old fashioned circuits in the prior art. Therefore the
current osciallator tube is an improvement over the prior art.


In the drawings, Figure One illustrates the basic configuration of the
oscillator tube. The first cathode, 1, is negative in the initial cycle of
the oscillator. The second cathode, 3., is located on the other end of
the vacuum tube, and is charged positively on the first cycle. The
anode, 2, is located in the middle of the tube and it is always charged
with a positive bias.

A battery or other current source is located at the bottom of the
diagram in Figure One. The input current travels from the battery to a
clock frequency transistor or chip which impresses the current with a
frequency of either fixed or variable timing. Then the clock frequency
is fed to a flip-flop chip or transistor which serves to divide the
frequency into two outputs, one at the top of the cycle and one at the
bottom. So a positive charge is applied to one cathode while a
negative charge is sent to the other cathode. Then, at the applied
frequency, the polarity of those cathode inputs is alternated. One
cathode is positive while the other is negative at all times. Electrons
are emitted from the negative cathode and are attracted and
accelerated toward the positively-charged anode, 2, and the other

Figure Two shows a modification of the present device in which
output from the oscillator tube is reapplied to the input sequence.

Figure Three is a diagram of the basic oscillation cycle, in which
cathode, 1, is negative and cathode 3, is positive at step one, and then
the poles reverse and the cathode, 1, is positive while cathode, 3, is
negative in the second half-cycle. Electrons are emitted from the
negative cathode at each step. When the electrodes so accelerated
impact the other cathode with sufficient velocity then secondary
electrons are emitted. So a cloud of say fifty or sixty electrons bounces
off the second cathode and that cathode is then attracted toward the
positive anode, 2.

After the first cloud of electrons hits the cathode then that cloud is no
longer accelerated. The multiplied cloud is another cloud and it is the
cloud which goes back in the other direction. So the original cloud of
electrons merely drifts without acceleration and they eventually orbit
toward the positive anode, 2, where they are extracted and sent to

Farnsworth sometimes found it convenient to make a tube with five
electrodes instead of three, with one anode and four cathodes. The
only difference was that there was an additional cup placed behind
the active cathodes to help catch the stray electrons after impact. But
the oscillator tube worked the same way. Said prior art is disclosed in
Figure Four and constitutes one of the possible modifications of the
current invention, equipped however with improved emissive
coatings and miniaturized components.


US 2168052 SNYDER
Self-Charging Cellphones and PDA’s
regenerated by Secondary Electron
By Bruce Darrow Gaither,
Tulsa, Oklahoa

The proliferation of innumerable handheld devices such as cell
phones and personal desk assistants has created a need for a
longlasting method of powering said devices. Presently batteries have
been employed for this purpose. Microwave wireless systems or solar
power have also been proposed as means to recharge the batteries.
The present invention discloses a method and apparatus for first
recharging batteries, and secondly to power the device during usage
and simultaneous recharging of the battery.

The principle of secondary electron emission has been known since
Austin and Starke reported that the metal surfaces impacted by
cathode rays emitted a larger number of electrons than were incident
in 1902. Certain materials exhibit an electron multiplication effect
when they are bombarded with high velocity electrons.

Radio and television tubes starting in the 1930’s employed secondary
emission materials and numerous devices and applications using
electron multiplication have been devised, such as radiation detectors,
photomultipliers, well logging devices, night vision goggles and digital
tv display devices.

The different kinds of secondary electron emissive materials have
been developed and these various materials multiply the bombarding
electrons at their own coefficient of emission, and the voltage used to
accelerate the bombarding electrons is known to be a major variable
in these devices.

Also, the secondary emission materials exhibit differing emissive
characteristics depending upon the spectra of radiation with which
they are bombarded.
Figure One below details the known characteristics of various
secondarily emissive materials as to the wavelength of the light to
which the materials respond with electron multiplication. Thus, some
materials are known to multiply electrons at the wavelength which
corresponds to visible light. But some materials are not emissive to
visible light, but they multiply electrons which are in the infrared
light spectrum.

The applications of secondary emissive materials to night vision
goggles provides the principle upon which the current invention is
based. These goggles apply light which strikes the emissive surfaces in
the infrared range. The human body is known to emit infrared
radiation due to its’ heat. When the heat image is received the
electrode multiplies that signal until it can be made into a visible light
image and transmitted to the goggle eyepieces.

The instant device employs the same characteristic of the human
body to emit infrared radiation. When the device is held in the hand
or placed in a pocket near the body it becomes irradiated with
infrared radiation. The emissive materials on the electrode then
multiply that signal, and then the multiplied electrons generate
electricity to power the device and to recharge the battery while it is
being used by an individual.

The initial electrons gathered from the infrared source need to be
accelerated to high enough velocity to maximize the secondary
emission, so a battery is used to power an anode with a positive
charge bias. A voltage multiplier, such as a Cockroft-Walton circuit
can increase the voltage of the battery to a high enough bias to attract
the incoming electrons. When the electrons strike the emissive
surface then they are gathered and the effect is a multiplication of
CURRENT. Part of the electrons so gathered goes to recharge said
battery and part of the current goes to meet the power requirements
of the handheld device.

Thus the cell phone or PDA will not run short of charge while it is
being held by the individual, and the battery will recharge by the self-
produced current from electron multiplication.
Figure One, Photomultiplier Handbook, Burle Industries, Table 1
By Bruce Darrow Gaither
Tulsa, Oklahoma


2,135,645 Farnsworth 11/1938
2,174,487 Farnsworth 09/1939
2,553,687 Quam 12/1950
3,128,408 Goodrich 04,/1964
3,478,213 Simon 11/1969
3,699,404 Simon 10/1972
4,041,342 Catanese 08/1977
4,410,902 Malik 10/1983
4,602,190 Evankow 07/1986
5,117,149 Fijol 05/1992


A small initial current is sent through an electron multiplier which is self-energized.
Once the multiplier is in operation the input current is shut off and said multiplier then
supplies electricity to a second electron multiplier which is a multistage device. Multiple
steps of electron multiplication occur until sufficient electricity is obtained to energize an
electric motor which powers an automobile.

The instant disclosure relates to the subject of electric cars. Usually the electric
automobile is powered by an array of 12 volt batteries, but the current invention improves
the state of the art by providing for a power system based upon secondary electron
emission similar to the photomultipliers manufactured by Hamamatsu Photonics and the
classic Multipactor tubes invented by Philo Farnsworth.

An electric motor is a device which is well known to those acquainted with the art.
Usually such motors will have requirements for electricity which can be expressed in
terms of volts and amps, and any combination of batteries, alternators or electric
generators (whether powered by gasoline or electricity) will satisfy the power
requirements of the electric motor which turns the wheels of the car.

In this instance an initial current is sent through a current multiplier which is scaled to a
degree which is designed to dovetail with the requirements of the electric motor. Thus,
the hereinafter-described invention is meant to be a general description of the invention
and it is not meant to foreclose all possible permutations or variations of the invention
which may be configured to meet the requirements of any particular electric motor, but
this disclosure is meant to include all possible variations thereof without limitation by the
illustrations or descriptions hereinafter used by the inventor.
There is presented a current multiplier which may be fabricated in the form of a vacuum
tube or a semiconductor device or a combination of the two.
The initial electric current for this device may be visible light, as in the manner of the
photocathodes of Hamamatsu, or a current from a battery or generator. The current is
conveyed through one or more cathode/anodes and is multiplied thereby until the current
meets the needs of the electric motor of the vehicle.

Thus, a small current initiates the process and a multiplication process occurs which
increases the number of electrons to the desired level, and then the mutliplied electrons
are finally subjected to transformers or voltage regulators to more closely guarantee a
constant current of the desired quantity and quality for the use of the electric motor.

It has been long known that certain substances exhibit a "secondary electron emissivity"
when they are bombarded with electrons. When these substances are exposed to a stream
of electrons then they emit electrons on their own. Some of these elements in an amount
greater than unity. Thus, when an element or alloy emits electrons in a ratio greater than
1.0 (electrons emitted/electrons inserted) it is called "greater than unity". In the days of
radio tubes a number of such substances were studied carefully and their secondary
emissivity charted. Substances such as Cesium and Magnesium were found to emit
secondary electrons in a ratio of up to 8 times the number of initial electrons by which
they were bombarded. A long list of other substances, alloys and layered combinations of
metals and other elements have been described.

Based upon this data the golden age of radio employed vacuum tubes which depended
upon the process of secondary emission. These radio tubes found a number of
applications in other fields of early electronics, including the field of television. Philo T.
Farnsworth, the inventor of television, designed a number of devices employing
secondary emission and which featured multiple impacts of electrons upon anodes within
a vacuum tube. He termed these devices "multipactors" for the multiple impacts upon the
secondary emissive materials. In one such device Farnsworth calculated that he could
produce over 4 times the initial number of electrons on each successive impact, and that
if he fabricated a vacuum tube containing multiple such impacts in succession then he
could generate 4x4x4x4...etc. Thus, the multiplicative effect of successive cathode/anode
combinations with each step being higher than unity, finally produced a huge
multiplication of the initial current.

Then Farnsworth invented a device which took part of the output of the secondary
emission cathodes and anodes and then reapplied it to the input of the device, so as to use
a regenerative feedback. This device resulted in a self-sustaining multipactor, which,
once it is started, would continue to multiply electrons without the addition of any
external power. He noted that in a cold-cathode configuration wherein light was used to
trigger a photo-emission to start this device that the input would be of a variable and not a
constant or predictable level, and that the output would vary according to the input. But if
a hot cathode or a battery-powered or Ac-powered input were used to initiate the process
that the level of input would be constant and the multiplication effect would be linear.
The design of these Farnsworth multipactor tubes is included in the Figures below and
the aforestated Farnsworth patents are incorporated herein by reference as if fully set
forth herein.

In the instant device the output desired is to be of a constant and predictable level, so
there is provided a non-photonic initial input in some configurations. In those
configurations of the current invention wherein light is to be used to stimulate photo-
emission to trigger the multiplication, then additional devices and circuits must be
employed to assure the predictable and constant linear multiplication required to power
the electric motor.

As the state of the art progressed most of the radio tubes were replaced with
semiconductor devices which performed the same function. This has likewise been the
case with secondary emission devices. The multipactor tube was microengineered and
placed on a semiconductor chip. The state of the art in photomultipliers (PMTs) has also
evolved into smaller and more efficient devices, but due to the characteristics of cesium
which deteriorates in the atmosphere, and because the atmosphere is insulative and not
conductive to electricity, the evolving devices must still be enclosed in a vacuum
envelope, whether they are in tubes or semiconductor chips.

Heating occurs in the multiplication of electrons by secondary emission. Sufficient
cooling must be provided in the form of air or water cooling.
These elements are not discussed or included in this patent disclosure. In some
configurations the target electrode may need to be charged to a positive potential that
exceeds the potential of the emitting electrode so that the electrons will be attracted at the
target with sufficient speed to penetrate the secondary emissive material sufficiently to
initiate electron multiplication. But secondary emissive materials of a very low work
function or of a negative electron affinity will tend to negate the need of charging the
target electrode to a positive potential. In any event, a battery can be used to provide the
initial current for this target charging, and the potential can be regulated by using a
Cockroft-Walton circuit to produce a sufficient potential as desired for the target

It is for this reason that the device, though self-energized, cannot be termed a perpetua
mobilia device. The self-energization is a function of the heat generated and low work
function of the specialized materials, and is not violative of the laws of thermodynamics.
The well-recognized principle of secondary emission over unity merely provides
scintillation to the initial tube once it is already in operation.

Einstein's law is that theorem which describes the work function involved in the process
of photoemission. In the work which won Einstein the Nobel Prize

he determined the amount of kinetic energy, E, of an emitted photoelectron.

Where hv is the energy quanta of the light and phi is the amount of energy requried for
the electron to escape the surface of the metal.
Thus the energy emitted by the secondary emission substance is equal to the energy of
the light MINUS the "work function" required for the electron to escape the metal.

Therefore, it is desirable for the "work function" to be minimized in order for more
energy to be liberated by the secondary emission material.
The best coatings for the cathode/anode combinations, then, is one which has a very low
work function.

But there was devised an effect which made this work function process much more
efficient by making a NEGATIVE subtraction, which in effect ADDED to the energy
quanta of the light, as Einstein described.

In the early 1960’s, R.E. Simon while working at the RCA Laboratories developed his
revolutionary concept of Negative Electron Affinity (NEA). Electron affinity is the
energy required for an electron at the conduction-band level to escape to the vacuum
level. By suitably treating the surface of a p-type semiconductor material, the band levels
at the surface can be bent downward so that the effective electron affinity is actually
negative. Thermalized electrons in the conduction band are normally repelled by the
electron-affinity barrier; the advantage of the NEA materials is that these electrons can
now escape into the vacuum as they approach the surface. In the case of secondary
emission, secondary electrons can be created at greater depths in the material and still
escape, thus providing a much greater secondary-emission yield. In the case of
photoemission, it has been possible to achieve extended-red and infrared sensitivities
greater than those obtainable with any other known materials. The first practical
application of the NEA concept was to secondary emission. Simon and Williams An
early paper by described the theory and early experimental results of secondary emission
yields as high as 130 at 2.5 kV for GaP:Cs.

So Gallium Phosphide, treated with Cesium was the subject of this initial discovery of
Negative Electron Affinity. A number of other substances were thereafter discovered to
exhibit this negative electron affinity and therefore exhibit extremely elevated levels of
secondary emissivity multipaction.

Many of these multiplication levels, expressed in terms of "Gain" increase as the amount
of input electricity is increased. So increasing the input to the cathode/anode pairs will
further enhance the multiplication effect.

Various means then exist whereby the gain from the initial input current can be controlled.
In addition to the control of the input levels of electricity devices extrinsic to the
secondary emission devices can fine tune the current produced so that the final product is
of the levels desired to power the electric motor of the car. Such devices may include
transformers, Cockroft-Walton circuits or other ladder circuits, voltage regulators,
magnetic amplifiers or operational amplifiers might be applied in this fine-tuning process.
Sometimes the current may pick up a ripple in this multiplication process and need to be
smoothed by other devices or circuits, but these devices are all well-known to those
acquainted in the art and need not be discussed here.

This invention discloses a module which is designed to produce electric current sufficient
to power an electric motor for an automobile. Such a module includes a power source, an
array of tubes and/or semiconductor devices, and other extrinsic devices which do not
employ secondary emission but which serve to modulate and regulate the signal which
ultimately results from the secondary devices and which extrinsic devices are employed
to shape the current to meed the needs of the electric motor.

Although a solar cell or photocathode could be used to provide the initial jumpstart for
the device, the preferred embodiment should use a battery in the range of 12-15 volts.
This will assure a reliable and constant amount of electricity which will make it easier to
produce a linear amplification of the intial input.

The first of the array of secondary emission devices shall be of the type described by
Philo Farnsworth in his Self-Energized Alternate Current Multiplier, US Patent 2,174,487
(1937). In this disclosure Farnsworth describes that it is easier to start this device than it
is to get it to stop.

Thus, once the self-energized vacuum tube gets warmed up it will continue to multiply
electrons on its own and the power source can be turned off in order to save the batteries.
The batteries, then, can be recharged by the ultimate output of the module while the
secondary emission process of multiplication is carried on using internally-produced
power. In order to insure that the electrons will not scatter before they hit their target
electrodes, a magnetic focusing device may be provided in the space between the emitter
and the target electrode, and said magnetic device may be either a permanent magnet or
electromagnet whose purpose is to deflect and confine the beam of electrons.

Once the self-energized tube is multiplying electrons then that output us conveyed to the
input of secondary emission devices employing negative electron affinity and which
multiply electrons at a much higher rate of gain.

In the manner of the multistage multipactor of Philo Farnsworth, US Patent 2,135,615, a
multistage process is then initiated. In this part of the process the materials employed
shall be of the type similar to the Gallium Phosphide (Cs), which can also include
Gallium Arsenide or Gallium Nitride, or other materials which shall not be described here,
including synthetic diamond.

Using Gallium Phosphide as an example, and applying a calculation based upon a gain of
100 to 1 for each step we can easily determine that the first step would produce 100 times
as many electrons as are input from the self-energized electron multiplier. Then the
second step would be 100 times 100, or 10,000 times as many electrons. A third step
would be 100 times 10,000 or a multiplication of 1 million from the initial input of the
self-energized electron multiplier device. The graphs and formulas provided from the
Hamamatsu PMT handbook demonstrate that it is common for such devices to exhibit
gains in the range of 106 to 108.

At multiplication levels of this degree then, small variations in the signal will likewise be
multiplied and can result in undesirable effects.

Therefore it is possible that an extrinsic device or circuit be interposed between the self-
energized device tube and the negative electron affinity device in order to modulate and
regulate the output of the Farnsworth device and assure, as much as possible, that the
input of the higher multiplication device be constant and smooth. A voltage regulator or
transformer or smoothing circuit can be employed at this stage. At lower levels such an
intermediate modulation extrinsic device might not need to be employed.

The operation of many electron multipliers and photomultipliers requires that an
accelerating electric field be applied to the device so that the stream of electrons will hit
the target electrodes with sufficient speed to penetrate the electrode and cause secondary
emission. In these devices a battery is sometimes used to provide a biasing charge to the
target electrode to attract the incoming electrons of opposite charge. This biasing voltage
should either meet or exceed the charge on the incoming electrons.

The greater the number of multiple electrodes the more involved the accelerating voltage
apparatus will be. In the case of a single electrode with coating of very low work function
or of negative electron affinity the accelerating voltage field may be foregone completely.
Where the voltage is required the current from a battery may pass through a Cockroft-
Walton circuit and be tapped at the appropriate step of the ladder to supply the voltage
required to each particular electrode. Thus, a separate kind of voltage multiplier such as a
CW ladder is used to step up the voltage from a battery in order to charge the electrodes
in the electron multiplier tube. If this accelerating voltage is required in the instant
invention then the output of the electron multiplier tube can be used to recharge said
battery so that the accelerating voltage can be applied as needed indefinitely and without
the need for external power. Thus, the present invention applies the prior art involving a
self-energized multipactor tube which then provides power for a second tube, with part of
the power of the secondary tube being used to either power a battery for the accelerating
voltage CW circuit, or the initial self-energized multipactor can have part of its output
diverted to power the CW circuit, and part to recharge the battery, and the remainder to
supply the second electron multiplier tube.

The selection of materials, then, is vital to the need for accelerating voltage. If a material
is used which has a very high secondary emission gain then the device will be able to
divide the current output therefrom and apply it to the accelerating voltage requirements.
Also the desired level of multiplication is known from the specifications of the electric
motor to ultimately be powered. This means that, depending upon the materials and the
motor, the second tube in the present invention may not need to be a multistage tube, and
the problem of accelerating voltage may be eliminated or lessened to a large degree. If
there is no accelerating voltage then there is no need for a separate power source for the
Cockroft-Walton circuit which might be unnecessary.

It is clear that such levels of current can be obtained by a module containing the
described array of secondary emission devices that the requirements for a simple electric
motor might be obtained with considerable ease. The ultimate output of the secondary
emission devices then will need to be adjusted so that the voltage and amperage
requirements of the electric motor be matched as perfectly as possible. The module
described herein, then, will employ well-known devices which may step down the
voltage or step up the amperage and change the frequency of the current to meet the
needs of the motor.

Current dividers will also be employed so that the ultimate current from the secondary
emission devices be divided into separate paths in order to first power the motor, and then
to recharge the battery, and also to power the headlights, windshield wipers of the
automobile. Rectifiers may be employed to convert between AC and Dc and the current
may be shaped and scaled to suit the needs of the battery charger and the auto electric
devices respectively.

The current invention represents an improvement over the current state of the art in
electric automobiles because it will produce a small and lightweight and versatile power
supply as compared with the bulky and heavy battery arrays employed at this time. The
savings in space and weight will enable the electric car to be smaller and more spacious
to the passengers and allow more carrying space. It will also constitute an improvement
in that it will not need to be recharged every 40 miles like a single battery would, and it
will not need to be plugged into Ac current overnight to recharge like the plug-in cars
require. Instead the instant invention will be completely self-contained and self-energized,
with only an initial battery for start up being employed, and that battery being recharged
after self-energization occurs.

Such a secondary emission module will be capable of powering larger electric motors,
thus advancing the art as to trucks and cargo vehicles and other work vehicles of heavier
weights, and will constitute an improvement over the golf cart type of electric cars.

Materials selected for their secondary electron emission properties may vary widely
depending upon the degree of electron multiplication needed and the cost and difficulty
in fabrication and deposition of these materials upon electrodes. In addition,
microdevices employing this principle sometimes create difficulties in the formation of
sharp points at the micro level and in the creation of vacuum spaces on the
semiconductors due to the fact that extremely high vacuum levels, beyond the normal
vacuum levels which would suffice for regular tubes. Thus, a trade-off of design
characteristics is indicated when it comes to the selection of materials for the secondary
emission of electrons.
Many materials, especially metals, exhibit secondary emission of electrons. As the graphs
indicate in the figures below, silver, magnerisum, copper, calcium, cesium, gold and
others are commonly-known elements. Next amalgams or alloys of the emissive materials
are found to multiply the initial input of electrons at elevated levels.

A silver base which is then oxidized and coated with a thin layer of cesium was the
preferred combination at one time due to the fact that it would be both photo-emissive
and cathode-emissive.

The principle of negative electron affinity and the development of new materials and
processes has expanded the range of possibilities and further elevated the gain of
secondary emission possible. Much higher levels of secondary emission have been
attained through the use of Gallium Phosphide treated with Cesium, and also with Gallim
Nitride and Cesium. Llacer & Garwin of the Stanford Linear Accelerator have predicted a
1000 fold gain for Gallium Arsenide at primary voltages of 20,000 volts. A complete
rendition of these newer materials is unnecessary here.

Recent advances have resulted in processes for the creation of thin films and coatings
involving synthetic diamonds through the chemical vapor deposition of carbon upon
substrates. Likewise research into carbon nanotubes have created another branch of
secondary emission materials. Each of these materials carries with it a separate process
for fabrication and assembly.

So there is no single material combination which is proper in this design, and this
invention is meant to include all such secondary emission materials and negative electron
affinity elements and processes for fabrication and assembly, whether stated herein or not.

But the preferred configuration is to use the tradition cesium oxide coated with cesium
for the first tube in the module due to the well-defined and predictable secondary
emission performance in this three electrode configuration, and for the negative electron
affinity materials to be used in the second tube in the module. The reasoning here is that
the first tube is to be used to create a steady and reliable power source which is
internally-generated and self-energized, and that the steady output of the first tube is then
to be applied to the second tube. Otherwise, a configuration where another power source
could apply current to a single multistage multipactor tube using negative electron
affinity materials would be likewise generative of electricity and would suit the purposes
of this invention. So, the instant invention is not meant to be limited to a two-tube design
and it may also include a single tube configuration or any other combination involving
any number of tubes in cascade design.

In the drawings:

FIGURE 1 is a chart showing the secondary emission yield versus
primary electron energy for various metals. It shows that Cesium
and Magnesium are the highest, with a secondary gain of about 8.
FIGURE 2 is a short discussion of Einstein's Law, which describes
the energy of an emitted photoelectron. The relevant fact in this
discussion is that the energy emitted by a metal will be the
initial energy minus the work function.

FIGURE 3 is a chart which shows the secondary emission of newer,
materials which exhibit a Negative Electron Affinity. Thus Gallium
Phosphide treated with Cesium has a gain which increases with the
increase in input boltage, with said gain approaching 100 for this
material, and comparing with Magnesium Oxide, a material with a low
work function, but not a negative affinity material.

FIGURE 4 is a graph which shows the gain of photomultipliers which have
a negative electron affinity. The gain increases as the increase in
supply voltage, reaching a peak of from 10^6 to 10^8--or one million to
one hundred million fold multiplication due to secondary emission.

FIGURE 5 is a discussion of the methodology used for the figures used in
Figure 4.

FIGURE 6 is a picture of a 1931 Pierce Arrow, a classic body design which
can be retrofitted with an electric motor powered by the present
invention. Nikola Tesla retrofitted a 1931 Pierce Arrow with an electric
motor, powered by a secondary emission tube array and module. This body type,
known as a Silber Arrow, has been chosen as the initial recipient for
the instant invention's secondary emission module.

FIGURE 7 is a diagram and description of the 1938 Multipactor patent
made by Philo Farnsworth. This figure discusses the benefits of
a multistaged series of secondary emission cathode-anodes, and
explains how a million-fold multiplication can be obtained. The second tube of the instant
invention, the secondary emission module, is to be of similar design and configuration,
but without limitation as to the number of stages of multipaction employed and without
reference to the particular materials to be used in these electrodes, as more fully
discussed above.

FIGURE 8 is a diagram of the Farnsworth Self Energized multipactor and
a description of the patent. This design takes advantage of the multiplication
of greater than unity of the secondary emissive materials, such as Cesium coated Silver
Oxide, to feedback the multiplied current to power the multipactor once
it has been started. This design has one cathode and two anodes.
Firgure 8a is a diagram and description of a multipactor oscillator by Farnsworth which
explains that the multipactor design is capable of generation of large amounts of power,
and that self-sustaining operation of the device will result from back-coupling the output
of the tube to the input, and that large amounts of power can be drawn down off of the
intermediate electrodes before the electron emissions reach the final electrode.

FIGURE 9 is a flow chart which depicts the present invention. It shows that a battery
provides the initial supply, which then passes through a self-energized multipactor.
Then the power goes to a second multiplying stage by flowing through a negative
electron affinity device, which has even greater multiplicative power.

FIGURE 10 is a perspective view of the instant invention. This figure depicts a
two-tube configuration. The power source initially actuates a self-energized
multipactor. The output of this self-energizing device then passes through a device which
contains a regulator and modulator. The signal is then input into the second tube.
The second tube contains the newer negative electron affinity materials. Once again the
current is multiplied, to an even higher degree, and then an ultimate current is obtained,
which may then be fine-tuned using a regulator and divider to dovetail the ultimate
current to the end desired. Thus the first tube provides a regular and constant current
which is self-generating, and the second tube makes use of this internally-generated
current to provide the bulk of the needed current multiplication, and finally the
current is adjusted to fit the end desired--here an electric motor for a car.

FIGURE 11 is a diagram of tube one of the secondary emission module, showing a three-
electrode configuration, and featuring one cathode and two anodes placed coaxially
thereinside the cathode. A variable capacitor and a choke coil are used to control the tube.
Figure 11 also contains a diagram of the circuit for the electric current and electrons of
the module.

What I claim is:

1. A power tube module to provide the electricity requirements of an electric motor
in an automobile or truck or other vehicle.

2. A power tube module for vehicles in which one tube employs the principle of
secondary electron emission.

3. A power tube module where the first tube using secondary emission has a three
electrod configuration, with one cathode and two anodes, placed in a concentric or
coaxial arrangement.

4. A power tube module wherein the first tube is self energized by virtue of
regenerative feedback from one of the anodes and the fact that the initial start-up
current is multiplied by the principle of secondary electron admission.
5. A power tube module wherein the first tube is self-energized after start-up and the
remaining output of said first tube is then regulated and modulated to assure
constancy and quality to signal and then said remaining output is applied to the
second tube in the module.

6. A power tube module wherein the first tube is self-energized and the remaining
output thereof is applied to the second tube, which then applies the principle of
negative electron affinity to further optimize the process of secondary electron
emission so that a multi-stage multipactor or Farnsworth tube is configured to
multiply the initial input to a still greater degree.

7. A power tube module wherein the first tube powers the second tube which then
multiplies the electron input to a great degree and then the output of the second
tube is further regulated and modulated so that it will produce an electric current
which closely meets the requirements of whichever electric motor is to be used to
power the electric vehicle.

8. A power tube module wherein the electrical current to initiate the process of
secondary emission multiplication is provided by a battery, which battery can be
switched off after the initial warmup of the first tube, and which battery can be
recharged by the ultimate output of the second tube, and which power tube
module can be operated for most of the time through the electrons internally
produced and multiplied by said secondary electron emission principle.
Figure 8a

by Bruce Darrow Gaither,
By Bruce Darrow Gaither
Tulsa, Oklahoma

The present invention discloses a circular accelerator employing
secondary emission.
Numerous electron multipliers are known to the art, with most of
them configured in a linear accelerator form, with some in undular,
spiral, semicircular or other unorthodox shapes. Heretofore, few
designs have been claimed which involve a secondary emission
electron multiplier employing a circular accelerator.

This device is of a similar design to that of a cyclotron, with a circular
or toroidal vacuum chamber being the conduit in which charged
particles are accelerated and extracted. Such cyclotrons have been
reported to generate electricity in the millions of volts and to require
complicated and expensive containment and accelerating magnets
and electrostatic devices.

The instant invention is designed to generate electricity in a much
smaller amount, such as the hundreds of volts. Consequently it is an
improvement on the prior art of cyclotrons.
The present invention is designed to provide a sufficient amount of
current to power an electric auto, home or business. The
containment and accelerating devices in the present disclosure are
then much smaller and less expensive and complicated than those
required for a large-scale cyclotron.

The operation of the uniform-field cyclotron [E. 0. Lawrence, Science
72, 376 (1930)] is based on the fact that the gyrofrequency for non-
relativistic ions [Eq. (3.39)] is independent of kinetic energy.
Resonance between the orbital motion and an accelerating electric
field can be achieved for ion kinetic energy that is small compared to
the rest energy….A cyclotron has constant magnetic field magnitude
and constant rf frequency. Beam energy is limited by relativistic
effects, which destroy synchronization between particle orbits and rf
Therefore, the cyclotron is useful only for ion acceleration. The virtue
of cyclotrons is that they generate a continuous train of beam
micropulses. Cyclotrons are characterized by large-area magnetic
fields to confine ions from zero energy to the output energy
Principles of Charged particle Accelerators, Stanley Humphries,
Chapter 15.

Secondary Electron Emission has been exhibited in cyclotrons from
their inception, but this kind of secondary emission is caused by the
unintended bombardment of the containing structure by the
accelerated charged particles, and is not usually the result of plan or
design. The instant disclosure details a device in which the inside of
the toroidal channel is intentionally designed to multiply electrons
through secondary electron emission. The inside of the accelerating
channel is coated with materials which exhibit an enhanced
secondary electron emission coefficient due to their negative electron

In 1902, Austin and Starke reported that the metal surfaces impacted
by cathode rays emitted a larger number of electrons than were

Numerous subsequent inventions applied the secondary emission
phenomena to various uses, such as film sound, signal amplification,
scintillation counting, radar jamming and a multiplicity of other

In the 1930’s radio tubes began to apply secondary emission
materials in vacuum tubes to amplify signals and multiply electrons.
The technique of coating electrodes in tubes with materials which
exhibit a high secondary emission coefficient continues to today, and
it is upon this foundation that the present invention depends.

When an electrode coated with emissive material is bombarded with
electrons then it emits additional electrons in a multiplicative manner.
Early devices depended upon metal oxides of various sorts and those
materials exhibited secondary electron emission in the range of 4 to 8
fold multiplication of initial electron input.
Since the 1930’s new materials and configurations have been tested
and modern secondary emission materials can exhibit secondary
emission coefficients as high as 60-100 fold multiplication.

Materials such as gallium phosphide, gallium nitride, gallium
arsenide, magnesium coated with cesium, silver coated with cesium,
carbon nanotubes and synthetic diamonds, to name a few materials,
have demonstrated a low work function so as to exhibit enhanced
secondary electron emission when bombarded with accelerated
electrons. The instant device can operate with a coating of any of the
above-mentioned materials, or others not directly discussed here, and
said invention is not limited to construction or fabrication with any
one kind of emissive coating. For the sake of illustration, however,
this application shall use carbon nanotubes or synthetic diamonds in
a thin film as the exemplary emissive coating material.

When an electron encounters or bombards one of these emissive
materials at a sufficient velocity then it liberates additional electrons
from that material. The key, then, is to accelerate the electron so that
it attains sufficient velocity to trigger the secondary emission.

In this device a circular accelerator is used to attain that electron
velocity. An electron, being negatively charged, will be attracted to a
positively charged electrode.

The present invention involves the use of two electrodes inside an
evacuated toroidal channel. The two electrodes are semicircular in
shape and they are separated by a small gap. The accelerator principle
calls for one electrode to be negative and the other positive, and for
the polarities to reverse on a periodic basis. So the negative electrode
will release an electron which will be attracted toward the other
electrode due to its positive charge. When it bombards the coated
electrode the emissive coating causes electron multiplication, and
many more electrons are released. Then the polarity changes and the
multiplied electrons are then attracted to the other electrode, which is
now positive. They again bombard the coated electrode at sufficient
velocity to cause further multiplication.

This cycle repeats itself at a sufficient frequency to assure high
velocity in the ever-multiplying electron cloud.
A magnetic field is provided on the top and bottom of the device, so
the electron population is contained in those directions and the
electrons have nowhere to go but in the direction of the circulation
around the inside of the accelerating toroid. As they travel their mass
increases and the electrons rotate in a constantly increasing spiral
toward the outside of the chamber. When the electrons reach the
outer wall of the device they are extracted and the resulting current
can be applied to output or as regenerative feedback to the accelerator.

Although the multiplication of electrons may be accomplished by
coating the entirety of the toroidal channel with secondarily emissive
materials, or by coating portions of the cathode dees with emissive
materials, a third embodiiment of the present invention is to coat a
grid or wire mesh with the materials exhibiting a high secondary
emission coefficient. This wire mesh is arranged to be perpendicular
to the flow of electrons around the channel and is supported by a ring.
This secondarily emissive wire mesh is charged positive throughout
the cycle, and it is a mesh which is permeable to the flow of the
electrons to a sufficient extent to attract electrons, to multiply
electrons which impact the mesh, and to allow the vast majority of
electrons to pass through the mesh without impact on any particular
orbit. During the course of time, however, the accelerated electrons
will impact said mesh and be multiplied. The outer edge of said ring
support structure can be used to extract current from the tube and
divert it to the outside of the vacuum envelope to a work circuit for
use in output or to be applied to input for the electrodes.

The multiplication of electrons causes an amplification of current,
which is expressed in amperes. The rotation of the electrons causes an
increase in kinetic energy, which is expressed in volts. Thus, the
present invention amplifies both amperes and volts. Power is a term
which is defined as amperes multiplied by volts. Therefore, the
instant disclosure results in an exponential amplification effect.

Although such a configuration could be used to achieve extremely
high voltages, the aim of the device is merely to produce sufficient
power to satisfy the requirements of an electric automobile or home
or business. Several hundreds of volts is sufficient to meet that need.
Therefore, expensive and cumbersome containment and accelerating
apparatus is not necessary in the current invention, and the objective
of this invention can be achieved by a device which is of relatively
small size.


US 2630549 SZIKLAI
DE 687265 BRIOT