1
+
I OFT
ORNLP
2911
.
.
-
-
4125
-
EEEEEEEE
MICROCOPY RESOLUTION TEST CHART
NATIONAL BUREAU OF STANDARDS - 1963
II
.
...
1.
'
"
"'
.. motor . Jika pemenuhi il
..
o
..
l en
Trimmer
3
.
"
BLANK PAGE
ווין
.
......
TDW TWT
o
n nep-2011
CITI PRICES
CASH PRICES
H.C. $97.09; MN.65
Mae ? 1 1967
HIGH CURRENT DC ION BEAMS*

I
BALLASID FOR ADDOUNCLONI
George G. Kelley
Oak Ridge National Laboratory
Oak Ridge, Tennessee
I SULLA SCIRO APATAON
Stimmáry
of the ion emission surface at the source, are
· common to both fields, however,
The beam systems which have been developed
for use in high energy injection experiments in
thermonuclear research are described. An at-
teinpt is made to compare the quality of these
beams (not the relative mei'it of the injection
systems). Probléms peculiar to the production
of intense de beams are discussed.
Evaluation of Performance
Introduction
The true emittance and brightness of a bea
depend only on random motions of the particles.
Wroe has shown that when the brightness is
measured properly directly after extraction, it
has the value expected from the temperature of
the ions in the source--much higher than would
be inferred from previous measurements. Since
this quantity is invariant in a steady beam for
rays at small angles to the axis (so that the
axial and transverse motions are uncoupled) any
decrease in measured brightness must be due to
time dependent effects. These effects are like
to be important, it at all only in space charge-
neutralized beams. Improvements in performance
must come then from improved beam optics.
In crder to make a rough comparison of
the performance of the various systems to be
described in this paper, we shall use an ef.
fective brighness which is the ratio of curren
density to an estimate of the solid angle of thi
beam envelope, normalized to a constant ac-
celerating voltage and constant mass. It is de
fined as follows:
B
At the present time intense dc ion beams
are used both in isotope separation and in exper-
iments leading to controlled thermonuclear power.
This paper will deal only with the latter appli-
cation. In one kind of experiment a plasma con-
taining very energetic ions is built up by
trapping previously accelerated particles in a
éritably shaped maguetic field. The major de-
vices of this kind are the OGRA'S at the Kircha-
tov İnstitute in Moscow, the PHOENIX experiments
at the Culham Laboratory in England, the MMII
device at Fontenay-aux-Roses, France, the ALICE
expériments at Lawrence Radiation Laboratory in
Livermore, California, and the DCX experiments
at the Oak Ridge National Laboratory in Oak
Ridge, Tennessee.,2 The injectors used in
thésé expériments will be described briefly, and
problems peculiar to the production and handling
oi dc beams will be discussed.
The interaction between this work and the
development of preinjectors for high energy ac-
celerators has been considerably less than might
be expected at first thought. Several factors
have been responsible for the separation of
these two fields. In the first place, there has
been a different point of view. A preinjector
must be tailored to a device whose ion optical
properties have been carefully worked out. The
preinjector must produce the maximum possible
beam in a particular region of phase space. On
the other hand, the thermonuclear devices have
requirements which depend primarily on a partic-
ular spatial limitation of the beam. Usually
all of the current entering the magnetic trap is
useful. In the second place, preinjectors work
with short pulses of current reducing the prob-
léms of heat dissipation but preventing the
buildup of an electron population to reduce the
dispersive force of space charge. The emittance
figure found using short beam pulses does not
apply when the beam remains on long enough to
become neutralized several hundred us at tyui-
cal pressures). Detailed measurements are diffi-
cult on intense dc beams. In view of the differ-
ence in both the nature of the problems to be
Bolved and in the methods used for evaluating
performance, the amount of interaction is under-
720 AI
V ha wa (h1 + ha)(wa + wa)/12
for ribbon beams and
B
=
900 AI
v.de2.(da + d2)2/12
for beams of circular cross section where A 12
the mass in am, I is the current in ma, V is
the accelerating voltage in MeV, hi and wi ars
the height and width of a defining aperture, ..
distance L from the exit aperture, which has s
height and width he and wa. For circular bear
di and de are the corresponding diameters.
Dimensions are in cm. This quantity differs
from the true brightness as defined by van Ste
. bergen* by the assumption that the rays at all.
points of the exit aperture diverge at the
*maximum angle defined by this aperture and
another limiting aperture. The number grossly
underestimates the true brightness since there
is evidence that beams are in general much mor
highly organized. The values obtained should
not be taken too seriously since they depend
*

WY
..
.
_
B
=
Lvaluation of Performance
beams (not the relative merit of the inject!on
systems). Problems peculiar to the production
The true emittance and brigatness of a beal
nt intense de beams are discussed.
depend only on random motions or the particles.
Wroe bas shown that when the brightness is
Introduction
measured properly directly after extraction, it
bag the value expected from the temperature of
At the present time intense de ion beams
'the ions in the source--much higher than would
are used both in isotope separation and in exper be inferred from previous measurements. Since
Iments leading to controlled thermonuclear power. this quantity 16 invariant in & steady beam for
This paper will deal only with the latter appli rays at small angles to the axis (so that the
cation. In one kind of experiment a plasma con axial and transverse motions are we coupled) any
tainag very energetic ions is built up by
decrease in measured brightness must be due to
trapping previously accelerated particles in a 'time dependent effects. These effects are 11.ke
sul tably shaped magnetic field. The major de-
to te mportant, if at all only in space charged
vices of this kind are the OCPA's at the Kircha neutralized beans. Improvements in performance -
tói Institute in Moscow, the PHOENIX experiments must come ther from improved bean optics.
at the Culham laboratory in England, the MMII
In order to make a rough comparison of
déviée at Fontenay-aux-Roses, France, the ALICE the performance of the various systems to be
described in this paper, we shall use an ef-
Livermore, California, and the Day experiments · Pective brightness which is the ratio of curren
at the Oak Ridge National Laboratory ir Oak
density to an estimate of the solid angle of the
Ridge, Tennessee.,2 The injectors used in
beam envelope, normalized to a constant ac-
these experiments will be described briefly, and celerating voltage and constant mass. It is de
problems peculiar to the production and handling fined as follows:
of dc beams will be discussed.
The interaction between this work and the
720 AI
development of preinjectors for high energy ac-
celerators has been considerably less than might . V ba wa (h. + ha)(wa + wa)/L2
be expected at first thought. Several factors
have been responsible for the separation of
for ribbon beams und
these two fields. In the first place, there has
been a different point of view. A preinjector
must be tailored to a device whose ion optical
properties have been carefully worked out. The
Vod22. (dz + da)2/16
preinjector must produce the maximum possible
beam in a particular region of phase space. On
the other hand, the thermonuclear devices have
for beams of circular cross section where A 1.
requirements which depend primarily on a partic the mass in ami, I is the current in mA, V is
ular spatial limitation of the beam. Usually
the accelerating voltage in MeV, hi and wi are:
all of the current entering the magnetic crap is the height and width of a defining a perture,
useful. In the second place, preinjectors work distance L from the exit aperture, which has a
with short pulses of current reducing the prob height and width h2 and wa. For circular beer:
léms of heat dissipation but preventing the
di and dz are the corresponding diameters.
buildup of an electron population to reduce the Dimensions are in cm. This quantity differs
dispersive force of space charge. The emittance. from the true brightness as defined by van Ste
figure found using short beam pulses does not bergen* by the assumption tliat the rays at all
apply when the beam remains on long enough to
points of the exit aperture diverge at the
become neutralized (several hundred us at typi "maximum angle defined by this aperture and ..
cal pressures). Detailed measurements are diffi another limiting aperture. The number grossly
eult on intense de beams. In view of the differ underestimates the true brightness since there
enče in both the nature of the problems to be
is evidence that beams are in general much mor
Bolved and in the methods used for evaluating
highly organized. The values obtained should
performance, the amount of interaction is under not be taken too seriously since they depend
stakable. Some very important problems, such
sensitively on the convergence angle which is
aß eolum design, and the control of the shape not well known and which should really be a
Research sponsored by the U.S. Atomic Energy Commission under contract with Union Carbide Corporat
B
=
-
900 AI
*
.

.
,
BLANK PAGE
weighted average across the exit apertures.
Also they do not measure the relative perits
of the systems as injectors because other factors
are lavolved,
Crossed Field Extraction
Ceneral
1 beam of neutral atomic hydrogen.5,6,7 figurd
I shows the essential features of these
injectors. Some spreading of the beams is be-
lieved due to fluctuations arising from insta-
dilities in the lan source. This dirriculty !
cannon to all of the son sources wed. It is
dealt nito at some length by Artemenkov et al.
and by Kistemaker et al. The Russian workers
state that by proper adjustment of the gas rie
to the source, the modulation of the beam cas
reduced to 15 to 20%.
he wlue of the effective brightness of
tham OCRA-I berm was calculated using reported
nalues of 155 mi or molecular ions at 160 :V
through a 36 coa aperture at a reported con.
vergence angle of 5°. It is
The injectors used in the OCRA and ALICE
experiments obtain ions i'rom the soirce plama
by diffusion across a strong magnetic field. In
the source Itsell, an arc runs along magvetic
field lines from a heated filament. There are
apertures derining the electron stiroam to pro-
duce a ribbon of plaswa which is located very
close to the exit aperture. The plasma density
in these crossed-field aources 18 insufficient
to prevent the passage of neutral gas in the
source through it, and the resulting gas erri.
ciency 18 rather low (approximately 20% or less).
B
.
20 x 155 x 2
0.16 x 36 x 7.9 x 10-3
• 4.9 x 106
pumping are required to prevent excessive pres-
sure in the beam region.
The size of the limiting aperture has not beer.
reported.
The OCRA'S
ALICE
The injectors for the OCRA devices are
alike except for the addition ut a neutralizing
cell and analyzing magnet in CGRA-II to provide
The crossed field sources at LRL are used
to produce neutral beams at COD siderably lower
energy. A typical arrangement is shown in
-------...
--..
.•
BEAM
BEAM

TARGET
"
Y
SOURCE
MAGNET
QUADRUPOLE
MAGNET
(o)
METERS
TARGET

BENDING
MAGNET
ION SOURCE

BAFFLES
BAFFLES
Me
int
ALLT
!
LUSSA
B.
AL
:
BEAM
QUADRUPOLE
MAGNET
HION STOPPER
SOURCE MAGNET
(b)
METERS

BLANK PAGE
11
-
...
....
.
.
.
wer
DO
aliu w woulu
the OGRA-I beam was calculated using reporte
nlues of 155 mA of molecular ions at 160 XV
through a 36 con aperture at a reported con-
vergence angle of 5. It is
Z
the source itsell, an arc runs along magnetic
field lines from a heated filament. There are
aperturcs defining the electron stream to pro-
duce a ribbon of plasua which is located very
close to the exit aperture. The piasma density
1o these crossed-field sources 1. innuriciant
to prevent the pasiege of neutral gas in the
source through it, and the resulting gas effi-
ciency is rather low (approximately 20% or 1388).
Large pumps and long beam fuths with differential
pumping are required to prevent excessiva pres-
mure la the beam l'egion.

B.
Izox 155 x 2
0.16 * 36 x 7.9 x 10-5
• 4.9 x 10
The size of the limiting aperture has not be
reported.
The OORA'S
ALICE
The injectors for the OCRA devices are
alike except for the addition of a neutralizing
cell ar analyzing m et in OSRA-II to provide
The crossed field sources at LRL are use
to produce neutril beams at considerably lowe
wargy. A typical arrangement is shown in
.... --
.
....
.
.
A
M
BEAM
TARGET
SOURCE TW
SOURCE
MAGNET
QUADRUPOLE
MAGNET
(o)
METERS
TARGET
BENDING
MAGNET
ION SOURCE
BAFFLES
BAFFLET
BEAM'
-QUADRUPOLE
MAGNET
*ION STOPPER
SOURCE MAGNET'
Fundinn
(6)
METERS
Fig. 1 - The OGRA Injectors. (a) OCRA-I. The beam is limited by a 6 x 6 cm aperture at the target,
which is actually the point of entry into the magnetic trap. (b) OCRA-II. The smallest aperture,
where the beam 18 smallest, 18 4.5 x 10 cm. The maximum Ho current through this aperture 18 26 ma.
The scales shown are approximate.
BLANK PAGE
FINN
BOTO
H
:
15
:

M
SWEEP MAGNET
BAFFLE
length of the extraction gap. The ions are ac
-SOURCE MAGNET COILS
swaed to be emitted at zero velocity. The tot
-IRON MAGNETIC SHIELDS
current then in a beam is
BAFFLE
I - Apieg : 5.44 x 10-6 v3/2
22/r2
-BEAM
SOURCE
TARGET . The denominator is seen to be a factor depende
on the geometry of the extraction gap. If the
gap 18 much smaller than the diameter of the
beam, the extraction field is very nonuniform
METERS
with respect to radius and the beam is badly
-. --.. torted. Thus for a given degree of distortio..
the total current depends only on v3/2. In
other words the perveance is fixed. A ribbon
Pg. 2 - The ALICE Injector.
beam suffers this same restriction across its
width, but its area can be much greater. The
Figure 2. The source has an arc cross section
above formula can be used to predict the axi..
of 0.125 in x 0.3 in and an exit slit of
current of ions from the source that can be
0.102 in x 3 in. A typical extraction gap is
handled at a given extraction voltage without
0.06 in. More recent design uses an ion source beam 1088 to the extractor. A correction must
slit of 0.170 in x 2.5 in. This source produces . be made for the presence of the extractor hole
50 mA equivalent current of 20 keV Hlo through a
however. It has been found empirically that i
2.0 x 5.1 cm baffle 21 ft from the source.
formula
The effective brigitness calculated from this
information and the assumption that the beam
: width remains constant is
A2/2 (z + x)2/r2
I
=
5.44 x 10-8 a vo/2
-
.
.
B
720 x 74
0.02 x 2 x 5.1 x 4 x 10.2/3602
= 8.3 x 108
predicts the maximum current usually to better
than 5%.12 The graph shown in Figure 4 is
simply a plot of the predicted current for a
beam in which the extraction gap Length is equal
to the diameter of the extraction aperture.
· 1200
Extraction in Axial Magne: tc Fields
General
1000
.
-
The PHOENIX experiments at Culham and the
DCX's at Oak Ridge use a beam derived from ions
from a duoplasmatron source. 10 This type of
: source has the advantage of considerably greater
gas efficiency because of the high density and
: high degree of ionization of its plasma, but it
is at some disadvantage at low extraction
voltages because it is adapted only to the pro-
duction of axisymmetric beans. The crossed .
field sources produce ribbon beams whose total
output can be increased by increasing the length
of the extraction slit. There is a fundamental
limitation, however, with axisymmetric beams.
The current density in a beam due to space
charge limitation of the current at the surface
of the plasma is given by
URRENT (MA).
T
!
J = 8.44 x 10-8 v12.
... (3)
A2/2 22
0
50
200
100 150
BEAM ENERGY (keV)
where j is the current density in amp
V 18 the extraction voltage, A is the mass number
in amu, and z is the spacing between electrodes
1n om. This formula applies in the case of a
beam of infinite cross section or in Pierce
geometry. 12. It 18 a good approximation in gen-
eral 1f the beam diameter is comparable to the
Fig. 3 - Maximum Current Expected from an Axia
Field Source without Gridded Apertures.

'
..'
SUM
BLANK PAGE
.
.............................
.
*
.
.-.,-
.-
.
.
.
.
length on mass in these leases permits separa
tion of mass species by apertures. In all ca
when it is desired to obtain a beam of the .
highest possible current density, the lens
should be as close to the source as possible.
Both the PHOENIX injectors and the DCX injecto
use solenoidal magnetic lenses. In every casa
the ion source is in the stray field of this
lens, and special attention must be paid to til
field which exists at the plasma emission sur
face. Compensating coils are required to
optimize performance. This subject will be d:
cussed in more detail below.
Fluctuations in the current from these
sources usually can be held below 1,0% peak to
peak by proper adjustment of gas flow and mag-
netic field.
Somewhat more current than indicated can be ob
tained, but the quality of the beam deteriorates
rapidly with increased current. Grids at the
source aperture and the extraction electrode can
eliminate this restriction, but heat dissipation
problems make their use difficult at low ex-
traction voltage and impossible above about 20
XV. The source gas efficiency is impaired by
the use of grids because of recombination of th
source plasma on the grid structure.
When a duoplasmatron 18 operated at high
gas efficiency, the plasma density at the source
anode aperture 18 about 1044 ions/cms and the
ions have an outward directed velocity of
greater than 10 eV. The result is a plasma
flow equivalent to a current density of up to
100 A/cm2. At the same time, the extraction
current density is determined by space charge
(Equations 1 and 2), and in most cases, 18 less
U cases, 18 less .
than about 3 A/cme. The plasma then must expand
as it leaves the aperture until it has formed
itself into & configuration in which the current
is space-charge limited at all points on its
surface. In order to prevent ion emission from
the sides of the resulting bubble, a cup is pro-
vided to expose only the expanded front surface :
of the plasma to the extraction field. Tae
control of the shape of this surface probably
18 the biggest problem in beam technology. The
amount of expansion allowed and the strength of
the extraction field which is most deisrable,
depends on the relative importance of the initial
diameter of the beam and the effect of space
charge forces in it. When the beam is acceler-
ated to very high energy as rapidly as possible, ..
it usually is best to use the maximum possible
electric field consistent with high voltage
technology and the smallest possible initial
beam diameter. At low energies a larger initial
beam may produce a greater intensity at a dis-
tance from the source.
For values of perveance up to about
2 x 10-10 A/v3/2, corresponding to a current
of 100 mA at 80 KV, electrostatic focusing
by mearis of a unit:otential lens can be used.
Higher perveance beacs require the use of mag-
netic lenses. The dependence of the focal
. PHOENIX
PHOENIX
The PHOENIX-II injector13 has a beam patul
shown in Figure 4. It can pass 50 mA of Hz*
at 40 keV through a 4.6 cm diameter aperture
300 cm from the source exit. The convergence
angle is 0.02 radians. When the source is ad-
justed to produce an H2' beam, 200 mA at 56 |
can be passed through the system. In actual u
the Hz* beam 18. neutralized in a vapor cell to
produce 20 keV Ho particles. The Ha* beam and
The H,+ beam require 200 mA and 800 mA respec-
tively, of total power supply drain, and source
arc currents of 10 A and 24 A respectively. T!
effective brightness for the molecular ion bear
is
900 x 50 x 2
= 2.3 x 107
0.04 x 4.62 x 12.62/1852
.
.
B
The Oak Ridge Beams
The beam path for the DCX-1.5, DCX-3, and
INTEREM experiments at Oak Ridge is shown in
Figure 5.15 The beam at 240 cm corresponds to
an effective brightness of
CALORIMETER

COILS
BEAM STOP
8 cm dia
BEAM STOP
4.6 cm dia
ION
SOURCE
e
autonome instrumenata i m
i
nistrimin licemente
w
i
nnende
----------- 185 cm
-
BLANK PAGE
...........
......
wirin...
..
.
.......
......
M
ersi..
ions have an outward directed velocity of
Fluctuations in the current from these
I greater than 10 eV. The result is a plasma
sources usually can be held below 10% peak to :
flow equivalent to a current density of up to
peak by proper adjustment of gas flow and mag-
100 A/cm2. At the same time, the extraction
netic field.
current density is determined by space charge
(Equations 1 and 2), and in most cases, 18 less PHOENIX
than about 3 A/cme. The plasma then must expand
as it leaves the aperture until it has formed
The PHOENIX-II injector13 has a beam path
itself into a configuration in which the current shown in Figure 4. It can pass 50 mA of Ha*
16 Space-charge limited at all points on its .. at 40 keV through a 4.6 cm diameter aperture :
surface. In order to prevent ion emission from
300 cm from the source exit. The convergence :
the sides of the resulting bubble, a cup is pro-.. angle is 0.02 radians. When the source is ad-
vided to expose only the expanded front surface. justed to produce an lid beam, 200 mA at 56 ||
of the plasma to the extraction field. The
can be passed through the system. In actual ull
control of the shape of this surface probably
the Hat beam 18. neutralized in a vapor cell to
is the biggest problem in beam technology. The produce 20 keV H. particles. The Ha* beam and
amount of expansion allowed and the strength of
the Hi+ beam require 200 mA and 800 mA respec-
the extraction field which is most deisrable,
tively. of total power supply drain, and source
depends on the relative importance of the initial arc currents of 10 A and 24 A respectively. The
diameter of the beam and the erfect of space
effective brightness for the molecular ion beam
:charge forces in it. When the beam is acceler-
i at 'd to very high energy as rapidly as possible, .
900 x 50 x 2
: it usually 18 best to use the maximum possible
= 2.3 x 107 |
electric field consistent with high voltage
0.04 x 4.62 x 12.62/1852
technology and the smallest possible initial
beam diameter. At low energies a larger initial
beam may produce a greater intensity at a dis-
The Oak Ridge Beams
tance from the source.
For values of perveance up to about
2 x 10-10 A/v3/2, corresponding to a current
· The beam path for the DCX-1.5, DCX-3, and
of 100 mA at 80 kV, electrostatic focusing
INTEREM experiments at Oak Ridge 18 shown in
Figure 5.15 The beam at 240 cm corresponds to
by means of a unipotential lens can be used.
an effective brightness of
Higher perveance beams require the use of mag-
netic lenses. The dependence of the focal
is
B
=
cm
CALORIMETER,

COILS
BEAM STOP
8 cm dia
BEAM STOP
4.6 cm dia
ION
SOURCE
C
+ V
YO
..
.....
.
-80 cm -
185 cm -
Fig. 4 - The PHOENIX Injector.

BLANK PAGE
{
_
900 x 66 x 2
-
.
• 2.0 x 10
: 0.04 x 10.2 x 19/1842
for molecular ions at 40 keV. These ions are
dlssociated in a magnesium vapor cell to produce
La beam of 20 keV neutral hydrogen.
Other experiments at Oak Ridge (DCX-2) use
bydrogen ion beams at 600 keV. 4- The power sup-
viy for accelerating these beams has taps at ..
150 kv, 300 kV, and 450 kV. The first 150 kV
18 supplied directly to the extraction gap. The
gap spacing is about 8 mm. Acceleration is done.
as quickly as possible and the beam enters a
solenoidal magnetic lens directly below the
column. Mgure 6 shows the arrangement. This
system has produced a beam of 350 mA total h
gen ion curront and bas produced 100 mA of
ions through the duct shown. It has been in
operation since 1960.16
example, 40 A of arc current vare needed in a
particular rather efficient duoplasmatron to
produce 1 A of proton output. Cooling of the
source then is a major problem, particularly i
tha' axdal field case, since most of the power
produced on an extremely small area at the end
· sion aperture of the anode. The original duo.
plasmatron of von Ardenne could not be operato
continuously at arc currents of greater than
· about 5 A. The power was dissipated on a
tungsten insert and the heat was conducted frc:
this insert to an iron anode. At Oak Ridge ve
finind that most of the temperature drop was
across the joint between the insert and the res
body of the anode. In fact, it turned out the
ferramagnetic material was not needed in the
anoder and that solid copper could withstand
. at least 50 A of arc--the improved heat con.
ductivity more than oft setting the difference:
in melting temperature. Improved cooling was
Deeded also in the intermediate electrode.
sources of this type are described elsewhere.
Dissipation of the energy of the beam it:
also presents a problem. Power densities can
easily be pushed beyond the capabilities of the
best available beat transfer techniques. The
greatest power density can be handled by the
technique shown in Mgure 7. The beam
terminates on copper tubes in which are swaged
spiral ribbons cr incanel. Centrifugal force
Problems of Power Dissipation
Ion sources for the production of contim-
ous currents of large fractions of l ampere re-
quire source arc currents of many amperes. As an

SOURCE
EXTRACTING
ELECTRODES-

EPOXY SKIRTS
FOCUSSING
MAGNET
Lo
Q2 RADIANS, 7.118 in
3.8-cra
APERTURE
117 cm
-
Lin 100 in., ose
3.8-cm
APERTURE
240 cm
RETRACTABLE
TARGET
ANALYZED
BEAM
TARGET
32 in., . 2 in.
3.8-cm
APERTURE
MAGNETICALLY SHIELDE:
TARGET
O, d. lin.

BLANK PAGE
Fig. 1 - The OCRA Injectors. (a) OGRA-I. The beam 18 limited by a 6 x 6 cm aperture at the target,
which is actually the point of entry into the racentin tuer. (b) OOPA-IT. TH ...?.662 ",
6!!. ...... ... .
.......
quire source arc currents of many amperes. As an best available beat transfer techniques. The
greatest power density can be handled by the
technique shown in Mgure 7.18 The beam
terminates on copper tubes in which are swage
SOURCE
spiral ribbons of Inconel. Centrifugal force

EXTRACTING
ELECTRODES

EPOXY SKIRTS
FOCUSSING
MAGNET
-200
0.2 RADIANS, 1*118 in
e-cm
APERTURE
117 cm
_
Lis100 in..
3.8-cm
APERTURE
240 cm
RETRACTABLE
TARGET
ANALYZED
BEAM,
TARGET
-232 in., ** 2 in.
3.8-cm
APERTURE
- MAGNETICALLY SHIELDEC
TARGET
0, de fin.
Fig. 5 - Injectors Simulating the INTEREM and
DCX-3 Geometries. 140 ma and 82 ma of 40 keV
Hg*.dons are delivered to the targets at 117 cm
and 240 cm respectively.
Fig. 6. The DCX-2 Injector. The maximum beam
through the duct is 100 ma of H2* ions at 600 }

BLANK PAGE
causes the water to flow in a thin film on
the inner surface of the tubes at very high
velocity. When the target is placed so that the
beam strikes the tubes at a mazing angie, pover
densities of more than 6 kW/cm can be sustained.
In gote of the very large amount of power
carried by these dc beams, in many cases the
loss to the accelerating electrodes 18 so small
that they require no deliberate cooling. The
electrodes usually are of stainless steel.
Problems of Voltage Breakdown
In all of the systems used for producing in-
tense dc beams, there are both electric and
magmetic fields within the vacuum system. Many
of the problems of voltage standori are caused
by the presence of the magnetic field. In the
crossed field sources, electrons formed in the
extraction gap follow a trochoidal path in the
rields and are intensified by cascading. These
electrons cause damage to the source supports
unless some means are provided to get rid of
them. The axisymmetric sources can have problems
too, but of a different kind. Unless great care.
18 taken in design, there will be electrostatic
potential wells for electrons constrained to :
'follow magnetic linee of force. In such a
situation the electrons drift around the axis of
symmetry, under the influence of the forces due
to the crossed fields. A cascading process
occurs until there is sufficient plasma to cause
breakdown. Figure 8 shows a design in which
electron trapping regions are avoided. The
prototype of the source itself is due to the
Bukumi group under Demirkhanov.20 This group.
has done other excellent work with dc beams.'
It is specially suited to the production of
large currents (hundreds of ma) of molecular
ions. This source and accelerator are used with
the system of Figure 5.
:. Voltage cleanup time is reduced by desigr
the electrodes to have a minimum amount of sur :
lace at a high electr.c field. It also seems
belo to avoid situations where particle flight
paths coincide with electric field lines. In
L surfaces, caxial
cylinders, and concentric spheres should be
avoided. When the surface area is small, we :
have found that many materials, among them cor-
per, stainless steel, nickel, and platinum, bu
not aluminum, will withstand at least 2 .MV cm
at voltages up to 80 kV after cleanup. · At hij
total voltage, the CERN group has found titan.
to be particularly good.?
The terminal equipment for the productio:
high current de beams is considerably larger
that required for beams of low average power. :
In some cases voltages are supplied through ic
transmission cables. As a result, transient ..
currents during a discharge can be much greatü.
and may have more serious effect. These proba
and same solutions are described el.sewhere. 15
It is interesting and perhaps surprising that
damage to source and electrodes 18 not a serio
problem in most cases. In the Oak Ridge wor?:
breakdowns are detected electronically, but
power is removed through standard commercial 1
quick-acting disconnects, or in the case of th:
600 kV (11) supply, by means of vacuum switche
in the primary. Damage is done only when there
is an abnormal flood of backstreaming electrons
such as produced by a vacuum failure or in the
absence of an electron suppressor, to be dis-
cussed below.'
Space Charge Neutralization
• If the positive ion beam 18 allowed to be
electrons formed in the beam and by secondary
emission rom a target are kept out of regions
where they can be pulled from the beam by elec
tric fields, the beam charge will be neutral-
ized. 19,22 The degree of neutralization will
depend on how effectively the electrons are cc
tained. It is easy to calculate the time requ

.
m
..f

..
VN
1
main con
i wir uns
ALONS
mm
C

:,
'
BLANK PAGE
en stor role in the
-
1
.
*
.
-
.
-
--
--
-
--
'
.
.
*
-
-
-
electrodes usually are of stainless steel.
per, stainless steel, nickel, and plaüinwi,
not aluminum, will withstand at least 2.MV/cm
Problems of Voltage Breakdown
at voltages up to 80 kV after cleanup. At' bii.
total voltage, the CERN group has found titan:
In all of the systems used for producing in.. to be particularly good. 14
tense de beams, there are both electric and
: The terminal equipment for the productior
magnetic fields within the vacuum system. Many
high current dc beams is considerably larger
of the problems of voltage standoff are caused
that required for beams of low average power. ;
by the presence of the magnetic field. In the
In some cases voltages are supplied through 1c**
crossed field sources, electrons formed in the
transmission cab.les. As a result, transient ...
extraction gap follow a trochoidal path in the
currents during a discharge can be much greater
fields and are intensified by cascading. These and may have more serious effect. These prob
electrons cause damage to the source supports . , and some solutions are described elsewhere. 15
unless some means are provided to get rid of
It is interesting and perhaps surprising that
them. The axisymmetric sources can have problems damage to source and electrodes is not a serio
too, but of a different kind. Unless great care... problem in most cases. In the Oak Ridge wory.
is taken in design, there will be electrostatic breakdowns are detected electronically, but
potential wells for electrons constrained to : power is removed through standard cominercial
:'follow magnetic lines of force. In such a
quick-acting disconnects, or in the case of th:
situation the electrons drift around the axis of 600 kV (11) supply, by means of vacuum switche
: symmetry, under the influence of the forces due in the primary. Damage is done only wher, thero
to the crossed fields. A cascading process
is an abnormal flood of backstreaming electrons
occurs until there is sufficient plasma to cause such as produced by a vacuum failure or in the
breakdown. Figure 8 shows a design in which
absence of an electron suppressor, to be dis- .
electron trapping regions are avoided. The
cussed below.
prototype of the source itself is due to the
Sukumi group under Demirkhanov.20 This group
Space Charge Neutralization
has done other excellent work with dc beams.
It is especially suited to the production of
. If the positive ion beam is allowed to be
large currents (hundreds of ma) of molecular
a trap for electrons, or in other words, if th.
ions. This source and accelerator are used with electrons formed in the beam and by secondary
the system of Figure 5.
emission from a target are kept out of regions
where they can be pulled from the beam by elec
tric fields, the beam charge will be neutral-
ized. 15,22 The degree of neutralization will
depend on how effectively the electrons are co
tained. It is easy to calculate the time requ
-
-
-
-


more on any
**
**
----
-
LI
-
--
-
1
-
-
-
-
.
-
--
PUTERE DATE
actione
-
•
-
-
-
-
-
-
-
..
..
-.-
:-
-
-
Y
-:-
in obratom
DOTA ANTE
Ar
WALIMTOOLS
7
-
+---
-.-.-
+
.
-
- RO
-
-
-
.
:
-

.
.
-
.
3
-
•
WT
11-
1.
....
.....
at al alto
----
RAKLIOLO GATH OF SOU WITH ACONEL_SPICAL EIR'DAL PLACE
Fig. 8. Beam Accelerating System in which Elec
tron Trapping Regions have been avoided.
Fig. 7 - Target used for High Intensity DC Beams.
: 3.4
P
I
V
to
BLANK PAGE
.
.
.
-
.
.
..
..
for thio neutralization process assuming that
electrons come from collisions with the back-
ground yas. It is simply the beam density
unit volume, or
(6)
focused by a unipotential lens uperated at &
uegative voltage with respect to the beam dri
chamber so that electrons are confined to th.
beam drift region. In the picture on the le!
electrons are being drained from the beam by
positive target bias. The beam on the right
space-charge neutralized.
Waen there is a beam cross-over, the neu.
tralization fails to allow the ions to follca
straight flight paths for reasons not comple:
understood. Figure 10 is a photograph made :
Oak Ridge of a helium ion beam focused by a
solenoidal magnetic lens in the region of a
cross-over and beyond. The radial current di
tribution in this beam was studied by means
a differential calorimeter probe. We found :
beyond the cross-over, most of the beam curri.
was flowing in a hollow cone having a diver:
where ab 18 tine beam particle density, no is the
background density, 18 the ionization cross
section of the background gas, and vb is the
beam velocity. This time constant is about
240 us for a 100 kV proton beam passing through
hydrogen at a pressure of 2 x 10-torr. It
depends only very weakly on energy, being about
340 us at 20 keV.
· Electrons are lost most readily in the ac-
celerating gaps. This loss is avoided auto-
matically by the use of a solenoidal magnetic
lens for focusing. It also can be avoided by
deaccelerating the ion beam before J.t caters
the field free region. The accel-decel arrange-
ment also 18 used in the OGRA and ALICE crossed-
field sources, but in this case only electrons
formed in the vicinity of the extraction gap are
affected. The transverse magnetic field pre-
vents migration of electrons from the drift
space into the vicinity of the source.
The effect of neutralization is shown ir
Figure 9.2 These beams both are 50 mA at 60 keV
of total hydrogen ion current. They have been
incoming beam. There was also an intense cc:!
ponent in the form of a collimated rod on ax:
We believe that the central rod is caused by
trapping of ions in the potential well of a :
virtual cathode formed by the electrons osci:.
lating radially in the symmetrical space cha.
field. A 15 V potential difference can trap
30 keV ions which would otherwise be divergi:
at an angle of less than 1t . · This phenomeno
is particularly unfortunate since, if the ior.
formed the kind of cross-over expected from
geometrical optics, the current density of a
beam could be increased by collimating it just
beyond the cross-over.



.
..'
.
.
.
.
P...
...
-
-
-
- -
wniowardeer
BLANK PAGE
-
-
-
-
-
که نا نا نادا لمب iota با: 64
المنار،
راه من
، مد،
. -
..د:
beam velocity. This time constant 18 about
240 us for a 100 kV proton beam passing through
bydrogen at a pressure of 3 x 10° torr. It
depends only very weakly on energy, being about
340 us at 20 keV.
:. Electrons are lost most readily in the ac-
celerating gaps. This loss is avoided auto-
matically by the use of a solenoidal magnetic
lens for focusing. It also can be avoided by
deaccelerating the ion beam before it erters
the field free region. The accel-decel arrange-
mcat albo 16 used in the OCRA and ALICE crossed-
field sources, but in this case only electrons
formed in the vicinity of the extraction gap are
affected. The transverse magnetic field pre-
vents migration of electrons from the drift
space into the vicinity of the source.
.
The effect of neutralization is shown in
Figure 9. These beams both are 50 mA at 60 keV
of total hydrogen ion current. They have been
CIOSS-OVET'únic UCjuidio bobl swine... mais
tribution in chis beam was studied by means
a differential calorimeter probe. We found :
beyond the crn88-over, most of the beam curr.
was flowing in a hollow cone having a diver:
angle about twice the convergence angle of ti.
incoming beam. There was also an intense cc:
ponent in the form of a collimated rod on ax:
We believe that the central rod is caused by
trapping of ions in the potential vell of a :.
virtual cathode formed by the electrons osci:
lating radially in the symmetrical space cha.
field. A 15 V potential difference can trar
30 keV ions which would otherwise be divergi:
at an angle of less than 1t. This phenomen
1.8 particularly unfortunate since, if the lor.
formed the kind of cross-over expected from
geometrical optice, the current density of a
beam could be increased by collimating it just
beyond the cross-over.
.
formed



.
.
. .
.
.
. .
.
.
.
(b)
Fig. 9 - A 50 mA 60 keV Hat Beam; (a) without
Space Charge Neutralization, (b) with Space
Charge Neutralization.
fig. 10 - Behavior of an Intense Beari beyond
Cross-Over. The beam is composed of hellum 1
at 30 keV.

BLANK PAGE
"
.
.
!!!
W
iny
,
H
.:
--
...
.
111
Cor.trol of the Emission Surface
The exceptionally high value of effective
brightness for the Livermore crossed-field source
may be due to the fact that the ion emission
sur face is controlled by the electron beam in
the source and the parallel magnetic field, and
is flat. Tao X19yuwetric sources do not permit
a direct control of the plasma emission surface.
The problem was reviewed by Gabovich. Later
Gabovich et al. allowed a considerable expansion
of the source plasma before extraction. 2 Books
et al. 22 found that the resulting plasma surface
had a central convex bulge, and that this bulge
could be eliminated by making tiae walls of the
expansion chamber conical. When a strong
. magnetic field is present, however, the situa.
tion again is difficult. Experiments made using
& small auxiliary coil around the expansion
chamber to control the magnetic field at its
surface yielded greatest beam intensity when the
coil was adjusted to produce zero field at the
surface. Another smaller optimum was found when
the sense of the auxiliary coil was reversed a
a nlue walch produced i parallel megetic che
at the surface. In the null field case the
symmetry of the extracted beam is disturbed
asily by the effect of small assymmetries in
the coil, or by small stray magnetic fields.
. Evidence of dirriculties at the plasma cu
tace is found in the fact that in axisymmetris
systems whenever the current is maximized thr.
an aperture, the current density is sharply
peaked on axis. Migure 11 shows a example o:
this effect. These scans were made using the
differential calcrimeter probe shown in Figuri:
12.31 New equipment recently designed will
make it possible to study the angular distribu
tion of the beam rays as well as the intensit:
distribution. It will be possible then to te:
whether the fault is dominirarm emission at ti.
plasma surface, variations in the radius or
curvature of the plasma surface, or a combinat.
of both.
wasion che hated by make and that the surface
Conclusion

The performance of some ion beam systems
used in thermonuclear research has been descr!!
In all cases average current densities obtaine ..
normalized to constant convergence angle have
been less by more than two orders of magnitude
than should be obtainable with perfect optics.
1
.
-CELL AT 470°C

Hº 20-KOV EQUIVALENT
CURRENT
TOTAL HO THROUGH CELL
CELL COOL
H-PARAMETERS
ADJUSTED FOR
MAXIMUM CURRE
FOR THIS RADIUS
TOTAL BEAM 220 mA
CURRENT DENSITY (MA/cm2)
Hot 40-keV-
CURRENT
E
savaduouw
KOV EQUIVALENT CURRENT
TOTAL HAY THROUGH CELL
Hat 40-KOV CURRENT
TOTAL BEAM 220 mA
2
O
Q5
1.0 1.5 2.0
BEAM RADIUS (cm)
2.5
3.0
PROBE POSITION (cm)
BLANK PAGE
.
I
.:
W
::
.-
W y
TAW"
..
"
"
S
pr,
. :
.
.
...
*
- TYTTIST
WICK
**

**
.
.
die Wuual... ".com dor...... ..........
Qabovich et al. allowed a considerable expansion
of the source plasm before catruction. Books
et al. 4* found that the rerulting planne surface
had a central convex bulge, and that this bulge
could be eliminated by making the walls of the
expansion chambar conical. When a strong
aumetic field is present, however, the situa-
tion agin is difficult. Experiments made using
a small auxiliary coll around the expansion
chamber to control the mugoretic field at its
surface yielded createst beam intensity when the
coil was adjusted to produce zoro field at the
surface. Another smaller optimum was found when
peaked an axis. Figure ll shows an example o!'.
this effect. Thes0 SCDS were made using the
dirfarantial calorimeter probe shown in Figure
12. Pew equipment recently designed will
make it possible to study the angular distribu
tion of the beam rays as well as the intensit:
distribution. It will be possible then to tei!.
whether the fault is nominirarm emission at ts.
plasma nurface, variations in the radius or..
curvature of the plasma surface, .or a combinat:
of both.
Conclusior
•
180
The performance of some ion beam systems
used in thermonuclear research has been descril'
In all cases average current densities obtaino:..
normalized to constant convergence angle have
been less by more than two orders of magnitude.
than should be obtainable with perfect optics.

-CELL AT 470 °C
*
400

*
.
.
.
-
.
-HO 20-KOV EQUIVALENT
CURRENT
.
-.
TOTAL HO THROUGH CELL
CELL COOL -
-PARAMETERS
ADJUSTED FOR
MAXIMUM CURRE
FOR THIS RADIUS
TOTAL BEAM ?20 mA
CURRENT DENSITY (MA/cm,
Hat 40-kev
CURRENT
80
milliamperes
KEV EQUIVALENT CURRENT
.
TOTAL HY THROUGH CELL
- K&V CURRENT
TOTAL BEAM 220 mA
-
20
0
0.5
1.0
2.5
3.0
PROBE POSITION (cm)
1.5 2.0
BEAM RADIUS (cm)
Fig. ll - Radial Current Distribution in a Beam
with Parameters adjusted to maximize Current ,
through a 4.2 cm Diameter Aperture.
Fig. 12 - Total Current as a Function of Radius
for the Current Distribution in Fig. ll. .
.
i
.
.
..
.
..
.
6
!
.
.
"
...
.,
IFE.
.
!
L
.
.
11
..
::
JA
::
:
.
.
.
2
YTY.

T'
N
BLANK PAGE
References

#
-
-
-
-
BEAM
Proceedings of the IAEA Conference on Plas:
Physics and Controlled Nuclear Fusion Re.
search, Salzburg, 1961, J. Plasma Phys.
and Thermonuclear Fusion, Part 1, 1962
Supplement.
-
-
-
-
-
-
WONOWYWNO!
.
.
Proceedings of the Conference on Plasma
Physics and Controlled Nuclear Fusion Re-
search, Culham, CN-21/112 (1965).
.
-
-
VIEW A-A
-
-
(3) 8. Wroe, Brookhaven National Laboratory
Accelerator Dept. AGS Internal Repts.
HW-1, HW-2, and HW-3, (1965-66).

-Ve-in.-THICK WATER-
COOLED TARGET
TOP PLATE
(4) A. van Steenbergen, "Recent Developments
In High Intensity Ion Beam Production and
Preacceleration," IEME Transactions on liuc
· Science, NS-12, No. 3, June, 1965.
(5) P. M. Morozov and L. N. Pilgunov, "A Sourc
of Molecular Hydrogen Ions for the 'OGRA'
Apparatus," Sov. Phys. Tech. Phys. 8, 347
(1963).
DIRECTION
OF SCAN
74-in.- THICK WATER-
COOLED TARGET
BOTTOM PLATE
12.15-mm? EXPOSED AREA
Fig. 13 . Differential Calorimetric Probe used
for Obtaining Results shown in Fig. 11.
. L. Bezbatchenko, V. V. Kuznetsov,
N. P. Malakhov, and N. N. Semashko, "Injec.
tion of an Ion Beam into the Magnetic Mirr
Machine OGRA;" Plasma Phys. (J. of Nuc.
Energy Part ć) 6, 301 (1964).
(7)
.
L. I. Artemenkov, N. I. Klochkov, V. V. Kuz
netzov, V. M. Kulygin, N. P. Malakhov,
P. A. Mikhin, D. A. Panov, V. S. Svishcher,
and N. N. Sema shko, "An Injector of Fast
Hydrogen Atoms," Proc. VII Int. Conf. on
Ionization Phenomena in Gases, Belgrade,
1965.
(8)
This effect can be seen by substituting kr/v
for (dz + d2)/L, the square of the convergence
angle, into the formula for effective brightness.
flere kI is the ion temperature in eV at the
emission surface, which is probably less than
I eV. The conclusion is also born out by the
experiments of Wroe and others.,26 Fluctuations
in the ion current emitted 1'rom the source affect
intensity since they prevent the beam from being
space-charge neutralized at all times, but in
nost cases the shape of the plasma emission
surface and spatial nonuniformity are probably
sore important factors.
The presence of electron trapping regions
when there is a magnetic field in an accelerating
Column has a marked effect on the voltage holding
apability of the column. These regions can be
eliminated by careful design.
: The MMII Injector has not been discussed
since it produces an inward-directed flow of
Lons from an annular source. It does not pro-
Juce'an ion beam.
J. Kistemaker, P. K. Rol, and J. Politiek,
"Some Plasma-Physical Aspects of Mono- and
Duo-Plasmatron Ion Sources," Nuc. Instr. 8:
Methods 38, 1 (1965).
..
(9
F. J. Gordon und C. C. Damm, "High Intensit
Source of 20-KeV Hydrogen Atoms," R.S.I. 31
963 (1963). See also Lawrence Radiation
Laboratory Progress Repts. UCRL-14285 and
UCRL-50002.
(10) M. von Ardenne, Tabellen der Electronen-
physik, Ionenphysik, and Übermikroskopie,
VEB Deutscher Verlag der Wissenschaften,
Berlin (1956).
(11) J. R. Pierce, Theory and Design of Electror.
Beams (D. Van Nostrand Company, Inc., New ..
York, 1954), 2nd edition...
(12) G. G. Kelley, N. H. Lazar, and 0. B. Morgan
"A Source for the Production of Large DC
Ton Currents," Nuc. Instr. and Methods 10,
263 (1961).

BLANK PAGE

D. P. Hammond and D. R. Sweetman, "The
Phoenix-II Injector and Burial Lime Commis-
sioning," Proc. 4th Symposium on Engineering
Problems in Thermonuclear Research, Free-
catti, 1966.
(28) See papers presented at the 1966 Linear
Accelerator Conference by Sluyters (Brook-
haven), Vosicki et al. (CERN), end Faure
et al. (Saclay).
(14) J. Haguenin and R. Dubois, "Measurement of a
High Gradient Accelerator Tube Model-
Investigation of the Properties of Titanium
Electrodes," CERN, 65-23.
(15) 0. B. Morgan, G. G. Kelley, and R. C. Davis, .
"The Technology of Intense DC Ion Beams,
R.S.I. (to be published).
(16) Thermonuclear Div. Semian. Progr. Rept.
July 31, 1959, ORNL-2802. See also Thermo-
nuclear Div. Progr. Rept. February 1, 1961-
(17) Thermonuclear Proj. Semiann. Rept. January
31, 1960, ORNL-2926, p. 60.
LEGAL NOTICE
This report ms prepared as an account of Government sponsored worteMatthwes the United
tour
pround or implied, with repect to the NCC-
racy, completeness, or unstulpeus of the taformation contained han this report, or to the wo
A. Makes nay warranty or representation,
of any information, apparitus, method, or procese declared to the report may in
B. Asmmes any liabilities with respect to the wood, or for duman mondeling
uw of any information, appunto, method, or process dlaclound to the reporte
As und in thise above, “per non acting on behalf of the Comunitetom" that
ployee or contractor of the Commission, or employna of mua contractor
such employee or contractor of the Commissjon, or employee of much
disseminatos, or provides accou to, any information permint to bene
with the Commission, or wo employment with such contractor.
privately owned rights; or

(18) W. R. Gambill, R. D. Bundy and R. W. Wans-
brough, "Heat Transfer, Burnout and Pressure
Drop for Water in Swirl Flow Tubes with In
ternal Twisted Tapes," Chem. Eng. Symp.
Series 57, 32 (1961).
:19) Thermonuclear Div. Semiann. Progr. Rept.
April 30, 1966, ORNL-3989, p. 89.
:20) O. F. Poroshin and J. Coutant (unpublished).
21) R. A. Demirkhanov, H. Freulich, v. V. Kur.
sanov, and T. I. Gutkin, "A Collection of
High Energy Accelerator Papers from U.S.S.R."
BNL Report 767 (C-36), 1962, 224.
22) G. G. Kelley and O. B. Morgan, "Space Charge
Neutralized Ion Beams," Physics of Fluids
4, 1446 (1961).
23) Thermonuclear Div. Semiann. Progr. Rept.
October 31, 1964, ORNL-3760, p. 64.
24) M. D. Gabovich, "Extraction of Ions from
Plasma Ion Sources and Primary Function of
Ion Beams, " Trans. in Instr. and Exp.
Techniques No. 2, 195 (1963).
25) M. D. Gabovich, L. L. Pasechnik, and
L. I. Romanyuk, "Plasma Penetration Boundary
and Plasma Focusing," Sov. Phys. Tech. Phys.
6, 61 (1961).
B. Brooks, P. H. Rose, A. B. Wittkower
and R. P. Bastide, "Production of Low Di-
vergence Positive Ion Beams of High Intensi-
ty," R.S.I. 35, 894 (1964).
27) Thermonuclear Div. Semiann. Progr. Rept. :
October 31, 1966, p. 88.
TE
w
END

DATE FILMED
0 / 21 / 167







tit