Author(s)
Han, Hwang-Jin.
Title
Physical processes in hollow cathode discharge sources.
Publisher
Monterey, California. Naval Postgraduate School
Issue Date
1989
URL
http://hdl.handle.net/10945/27208
This document was downloaded on May 04, 2015 at 22:54:29
..vi-oOCte
NAVAL POSTGRADUATE SCHOOL
Monterey
,
California
THESIS
Hin^5
PHYSICAL PROCESSES IN HOLLOW CATHODE DISCHARGE
uy
Han, Hwang- J in
December 1989
Thesis Advisor;
Richard C. Olsen
Approved for public release: distribution is unlimited
T2A7903
CLASS'?. CA"'0".
0!"
PAGE
'H.S
REPORT DOCUMENTATION PAGE
i£?ORT SECURITY CLASSi^.CAriON
RESTRICTIVE
lb
MARKINGS
Unclassified
icCuRlTY CwASS^rlCATlON
DEC-ASSiPiCATION
iRrORiVlING
MA.ViE
/
AUTHORITY
Approved for public release;
distribution is unlimited
DOWNGRADING SCHEDULE
ORGANIZATION REPORT NUMa£R(S)
Of PERFORMING ORGANIZATION
6b OFFICE
and
SYMBOL
7a.
NAME OF MONITORING ORGANIZATION
applicable)
Naval Postgraduate School
{City, Stace.
MONITORING ORGANIZATION REPORT NUMBER(S)
5
(If
IDDRESS
DISTRIBUTION /AVAILABILITY OF REPORT
3
61
Naval Postgraduate School
ZIP Code)
7b.
ADDRESS
Monterey, California 93943-5000
NAME OF FUNDING/ SPONSORING
ORGANIZATION
ADDRESS
(City. State,
and
ZIP Code)
Monterey, California 93943-5000
8b OFFICE
(If
(Gty, State,
SYMBOL
9,
PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
applicable)
and ZIP Code)
10
SOURCE OF FUNDING NUMBERS
PROGRAM
PROJECT
NO.
ELEMENT NO.
TASK
WORK
NO
ACCESSION NO
UNIT
TITLE (Include Security Classification)
Physical Processes in Hollow Cathode Discharge
PERSONAL AUTHOR(S)
TYPE OF REPORT
3b TIME
Master' Thesis
COVERED
FROM
U. DATE OF REPORT
TO
(Year, f^ontti, Day)
15
PAGE COUNT
1989 December
61
SUPPLEMENTARY NOTATION
The views expressed in this thesis are those of the author and do not
fleet the official policy or position of the Department of Defense or the U. S. Government
COSATI CODES
FiElD
GROUP
18
SUBJECT TERMS (Continue on reverse
if
necessary and identify by block number)
SUB-GROUP
Hollow Cathode, Plasma Source, Ion Beam Source
ABSTRACT (Continue on
reverse
if
necessary and identify by block number)
The hollow cathode is an effective source of dense, low energy plasma.
Hollow cathodes
find use in ion beam sources for laboratory and space applications.
They can also be used
independently for satellite charge control, and ion beam neutralization. A heaterless hoi
-low cathode design was tested with argon gas used as a propellant.
This thesis work inve
-stigated the device properties, that is, the emission currents as a function of discharge
current, propellant flow rate and other physical parameters.
Starting behavior was a main
point of the investigation.
The results of these experiments were compared with studies of
the conventional hollow cathode.
D.STRiBUTlON/AVAiLASiLlTY OF ABSTRACT
Qu^CLASS'FiED/UNLiMl'ED
a. NA.-.IE
21.
SAiYE AS RPT
1473, 84
ABSTRACT SECURITY CLASSIFICATION
D OTIC USERS
Unclassified
22b TELEPHONE (Include Area Code)
OF R£S?0NS;8lE iNOiViOUAu
Professor
)FORM
D
Richard Chrlstoper Olsen
MAR
83
APR
eO'tion
All
may be used
(408)
until
exhausted.
other editions are obsolete
22c.
OFFICE
SYMBOL
61 Os
646 - 2019
SECURITY CLASSIFICATION OF THIS PAGE
S
U.S.
Goytminent
Prlnliin 0"ice:
UH— 60S143
Unclassified
Approved
for public releause; distribution is unlimited.
Physical Processes in Hollow Cathode Discharge Sources
by
Han,
Hwang—Jin
Major, Republic of Korea Army
B.S., Republic of Korea Military Academy, 1981
Submitted
in partial fulfillment of
the
requirements for the degree of
MASTER OF SCIENCE
IN PHYSICS
from the
NAVAL POSTGRADUATE SCHOOL
December 1989
ABSTRACT
The hollow cathode
Hollow cathodes find use
They can
is
in ion
an effective source of dense, low energy plasma-
beam
sources for laboratory and space applications.
also be used independently for satellite chcirge control,
neutr2dization.
A
as the propellant.
heaterless hollow cathode design
and ion beam
was tested with argon gas used
This thesis work investigated the device properties, that
is,
the
emission currents as a function of discharge current, propellant flow rate and other
physical parameters.
Starting behavior was a main point of the investigation.
results of these experiments were
compared with studies
cathode.
Ill
The
of the conventional hollow
ct
TABLE OF CONTENTS
I.
INTRODUCTION
1
II.
BACKGROUND
4
HI.
A.
HOLLOW CATHODE PHYSICS
4
B.
PREVIOUS EXPERIMENTAL RESULTS
7
C.
HEATERLESS HOLLOW CATHODE
18
EXPERIMENTAL EQUIPMENT AND PROCEDURE
A.
B.
EXPERIMENTAL EQUIPMENT
25
Chamber
1.
Discharge
2.
Electrical Circuit
26
3.
Measuring Equipment
29
4.
Vcicuum System
29
25
PROCEDURE
29
1.
Vacuum System
2.
Starting and Shutting
29
Down
the
Plasma
Source
IV.
31
(A).
Standard Hollow Cathode
31
(B).
Spectra-Mat Hollow Cathode
33
EXPERIMENTAL RESULTS
A.
25
34
STANDARD HOLLOW CATHODE
34
1.
Flow Rate Dependence
34
2.
Temperature Dependence
34
3.
Time Dependence
36
IV
B.
V.
SPECTRA-MAT HOLLOW CATHODE
Mode
Dischcirge
38
38
1.
Idle
2.
Extraction of Discharge
40
3.
Discharge Failure
41
CONCLUSION
46
LIST
Table
2.1
Vcilues of Coefficients
A
OF TABLES
and B
for various
21
gaises
Table 2.2
Mininum Sparking
Table 4.1
Data
for
Potentials
Spectra-Mat Cathode
VI
21
40
LIST
OF FIGURES
Fig. 1.1
Hollow Cathode Schematic
Fig. 2.1
Discharge Initiation Data for Cathode with Rolled
2
Foil Dispenser
9
Fig. 2.2
Small Orificed Hollow Cathode
Fig. 2.3
Variation of
Minimum
10
Discharge Voltage with
Pressure for Several Sepcirations
Fig. 2.4
Keeper Voltage Dependence on
the Operation
Fig. 2.5
Time
14
Keeper Voltage Dependence on
the
Fig. 2.6
13
Mass Flow Rate
Volt
14
Ampere Discharge Chaxact eristics
of
Cesium
Hollow Cathode for Different Flow Rate
Fig. 2.7
Compcirison between D.C. Ignition Voltage and
Pulse Ignition Voltage
Fig. 2.8
16
17
Relation between Pulse Ignition Voltage and
Mercury Flow Rate
17
Fig. 2.9
Entire Test Log of the Cathode
18
Fig. 2.10
Paschen's
Law (Breakdown
voltage
of the reduced electrode distance
Vb
as a function
P*D)
20
Fig. 2.11
Spectra—Mat Hollow Cathode Appziratus
24
Fig. 3.1
General Experimental Arrcingement
25
vn
Fig. 3.2
Electrical Circuitry for the
HC-252 Hollow
Cathode
Fig. 3.3
27
ElectricaJ Circuitry for the Spectra—Mat
Hollow Cathode
28
Vacuum System
Fig. 3.4
Major Parts
Fig. 3.6
Relation between Propellant Flow Rate and
of Varian
30
Chamber Pressure
30
Fig. 4.1
Discharge Voltage vs Flow Rate
35
Fig. 4.2
Discharge Voltage vs Heater Current
36
Fig. 4.3
Idle
Mode
Discharge Current vs Keeper Biasing
37
Potential(Vk)
Fig. 4.4
Idle
Mode
Dischcirge Current for Different
Flow Rates
Fig. 4.5
Idle
Mode
39
Discharge Current for Different
Biasing Keeper Voltage
39
Damaged Cathode
Fig. 4.6
Picture of
Fig. 4.7
Broken Tip
Fig. 4.8
Detailed Digram of Disassembled Spectra—Mat
of
Surface
Ceramic Insulator
Hollow Cathode
43
44
45
Vlll
ACKNOWLEDGEMENTS
I
wish to express
my
gratitude and appreciation to thesis advisor, Professor
Richcird Christopher Olsen «ind second reaider, Professor S.
instruction, guidance
I
and
Gnanalingam
for the
friendly advices throughout this study.
wish to thank the numerous scientists
this work, particularly. Dr.
Paul
J.
who
contributed figures and data for
Wilbur and Dr. Daniel E.
Siegfried,
Colorado
State University.
Finally,
for their love
many thanks
to
my
wife,
and being supportive
for
Kyoung—Sook and my
two and
IX
son, Frederick Teut,
hali years in Monterey, California.
INTRODUCTION
I.
One motivation
for the study of g2LS discharge technology
beams
plasma
for
reliable
implementation
devices
provide
ion
Applications
in
is
the
include
bombcirdment thrusters
Hollow cathode gas discharge
hollow cathode.
important
an
capability
electron
source
a source
as
the
for
low
of
cind the neutralize! for ion thruster
ion sources
filament
beams.
in
plasma.
in
ion
Ion thrusters,
vacuum
processing
Large area ion beams are extracted from electron bombardment—type
whose ionizing electrons are typically supplied by a hot refractory metal
cathode.
difficulties
energy
chamber
dischcirge
developed for space applications, axe finding increasing use
applications.
to produce a
One popular and
laboratory aind space applications.
the
is
in
However, short filament
maintaining
constant
filaiment
lifetimes
(several
tens
emission,
electron
of
and
hours),
filament
breakage, ane significant undesirable features of using this type of cathode in space
fhght or production processing equipment.
Hollow cathodes
offer substantially longer lifetimes
However, cathodes axe intrinsically complex devices and
than filzLment cathodes.
for a specific application
require a great deal of testing and pcircimetric optimization before reliable operation
can
be
cissured.
These technical challenges have limited their application
industrial ion sources.
is
Fig. 1.1 details a hollow cathode
easy to fabricate, has demonstrated long
beam and plasma
life
which does operate
operation, and
may
in
reliably,
be used for ion
sources. [Ref. 2]
The hollow cathode
tantalum, covered on
its
consists
of an
outer
downstream end by an
refractory
orifice
metal tube, usually
plate usually
made
of
1
PLATE
ORIFICE
—^
X
^KEEPER
INSERT--.
FLOW
^.^ANODE
I
GAS
—
i
CATHODE^
ll
1
TUBE
D. C.
POWER
SUPPLY
C.
POWER
SUPPLY
-
z-
Fig. 1.1
thoriated tungsten.
insert
Hollow Cathode Schematic
The cathode
also normally
incorporates a refractory metal
either coated or impregnated with such chemicals as
baxium compounds,
which aid the emission process by reducing the work function of the
The cathodes used
in
ion
thrusters
typically
insert surface.
have inner diameters of a few
millimeters and orifice diameters of a few tenths to one millimeter.
length
is
usually a few tube diameters.
The
electron current
is
The
collected by an
anode biased positive with respect to the cathode. Hollow cathodes generally
a small secondary anode, called a keeper, which
heater
is
process.
is
in.-sert
utilize
used to initiate the discharge.
A
normally used to heat the cathode as an aid to etaiting the emission
However, the discharge
and the heater can be turned
off or
is
self-sustaining (self—heating) once established
turned down.
As mentioned above, there has been an
ion
bombardment
interest in hollow cathodes for use in
thrusters as the electron source for the discharge chcimber and as
Because operating requirements for the
the neutralize! for the thruster beam.
thruster dictate long lifetimes and stable operation for this component,
become
of
prime importance to understand physical phenomena taking
it
pldtce in
has
the
hollow cathodes used in these devices.
With the
realization that electric propulsion systems will probably be applied
at first to the station
operate
a thruster
sufficient.
and
is
it
It
keeping mission,
without
it
has become appso^ent that the ability to
deterioration
must also be capable
for
thousands of hours
of rapidly starting
is
no longer
from cold thousands of times,
therefore important to identify the parameters governing this process, so
that problem areas can receive attention.
In the case of the electron
thruster, the ability to initiate the discharge
hollow cathode.
For
on demand
this reason, this aspect of hollow
is
largely
bombardment
dependent on the
cathode operation has been
studied in conjunction with the fundamental investigations mentioned above.
In this investigation,
than
was
thought
phenomena.
it
initially
In particular,
was found that such characteristics are more complex
from
consideration
under any one
required was not reproducible, but
fell
parameters
The
process
of
it is
other
set of conditions,
gaseous
breakdown
the initiation voltage
within a range, the magnitude of which
depended strongly on temperature and flow
initiation
of
rate.
In
deciding upon
suitable
therefore necessary to balance these quantities. [Ref.
6]
objective of this study has been to gain a better insight into the physical
hollow
cathode
operation.
Towards
this
goal,
an
experimental
investigation was undertaken to measure plasma properties and other pertinent
physical parameters, and to observe the starting behavior under several conditions.
BACKGROUND
II.
HOLLOW CATHODE PHYSICS
A.
Ignition of the standaird hollow cathode begins with the activation of the
heater power supply, which heats the cathode to approximately lOOO^C, followed by
the introduction of propellant into the cathode as
supply
is
then activated, the gas breaJis
down
shown
in Fig.
electriccJly
1.1.
The keeper
emd an arc discharge
Stable hollow cathode arcs require a copious source of electrons which the
ignites.
insert provides
by the mechanism of
scenario, the insert
must be heated
a region large enough such that
in
field
enhanced thermionic emission.
In this
to a temperature of approximately lOOO^C over
combination with the electric
field
generated by a
nearby, dense plasma, the insert emits enough electrons to maintain a stable arc.
[Ref. 10]
Daniel
E.
Siegfried,
Colorado
State
University,
provided
the
current
understanding of the physical processes which taie place inside the hollow cathode.
He explained
as follows in "Phenomenological
Cathode Operation".
Model Describing
Orificed,
Hollow
[Ref. 7]
The cathode orifice maintains a high neutral density inside the cathode
by restricting the propellant follow and it also provides a current path to the
downstream dischaxge. The electron emission comes uniformly from a narrow
(a2mm) band on tne downstream end of the insert.
The electrons are
produced at the surface of the insert by field—enhanced, thermionic emission
(the very strong electric field is a consequence of a very dense plasma and the
resulting potential drop across a very thin plasma sheath).
The electrons
produced at the insert surface are accelerated across the plasma sheath by a
potential of 8 to 10 volts. Since the mean free path for inelcistic collisions of
these energetic electrons is on the order of the internal cathode diameter, the
"ion production" region can be idealized to be the volume circumscribed by
the emitting region of the insert. The dense internaJ plasma is established by
the ionization taking place in this region. Ions produced in this volume diffuse
Bohm
The electrons strike the insert surface with
the temperature necessary to provide the
required electron emission.
The emission surface temperature, however, is
determined not only by the emission current but also by the local plasma
out of
it
at the
velocity.
energy to heat
sufficient
it
to
properties.
The plasma properties in the ion production re^on axe coupled into the
problem by the energy balance at the insert surface m the following manner.
The plasma properties determine the ion flux and therefore the energy input to
the emission surface. For a given emission current, the surface temperature is
determined by the energy balance which demands that the thermal losses from
the surface due to electron production, radiation and conduction are balanced
by the energy input from the ion flux. The plasma properties also affect the
required emission temperature because they determine the magnitude of the
electric field—enhancement in the emission process.
Therefore, for a given
emission current the surface temperature and plasma properties must be
consistent to the extent that they satisfy the energy balance at the surface.
All cathode surfaces which contact
the plasma receive ion currents
proportional to the Bohm velocity and the plasma density adjacent to the
surface.
Electron emission, on the other hcind, can be assumed to come only
from the
band on the downstreami end of the insert. The total emission
current from the cathode is equal to the sum of the ion currents to the various
cathode surfaces and the current of the emitted electrons.
Certain aspects of this phenomenological model can be expressed
analytically in a simple form which will allow comparison with experimental
results.
The plasma density adJ2w;ent to a particulax surface (n) can be
calculated based on the Bohm condition using
2mm
n
]
=
ev
I
B ohm
i
KTe"
(2.1)
1/2
m,
where
Ii
is
the ion current to the surface,
A
K
is
the surface area, Tg
is
the
Maxwellian electron temperature (°K), and
is Boltzmann's constant.
For
an electron emitting surface the measured current to the surface is determined
by both collected ions and emitted electrons, so that the total current density
to the surface
is
!,=].
total
+
1
where
is
j
the ion current density and
(2.2)
]
e
is
j
the electron emission current
e
i
_
The ratio of ion to electron currents can be estimated from an energy
balance on the emitting surface. In such a balance the ion heating power is
equated to power conducted and radiated from the surface plus the power
required to boil off electrons. The equation describing this is
density.
j^+q
e
=
+V.-<^)
j.(V
1
c
1
s
(2.3)
where
^
the effective work function of the surface, q
is
away from the
surface,
V
is
the thermal heat flux
the potentiaJ drop across the plasma sheath,
is
V
c
is
i
the ionization potential, and
is
4>
the work function of the surface material
(
a
s
material property ).
Equation (2) euid (3) can be
electron emission current density from the surface.
=
jg
- aq
]
t^-t^i
+
1
where a =
(
V^
+
Vj
—
)"
^g
.
combined to give the
(2.4)
a^^
In general the thermal loss is a function of the
temperature and the cathode thermal design. Most of the thermal loss
due to radiation from the outer surfcice of the msert to rather cold external
surfaces, and can be estimated from
surfaice
is
q ^ e<7T^
(2.5)
e is the emissivity (^0.5 for tantalum),
a is the Stefan^oltzmann
constant, and T is the surface temperature.
Emission from the surface is
assumed to be given by the Schottky equation for field—enhanced, thermionic
emission
where
e
where A^
=
The average
120
A/cm^K^ and
effective
^i-1
=Ayexp[
j
i<rT
KT
L
(2.6)
J
the other parameters are as previously defined.
work function ^^
is
given by
\=*.-[^Y'
L
O
(2.7)
J
where
to is the permitivity of free space and E, the electric field adjacent to
the surface, can be estimated using
E =
is
^
^"^
=-
^
V
of 4/3 comes from Child's Law considerations and the sheath
estimated as one Debye length (Ai-^). [Ref. 16]
Here the factor
thickness
-4^=-
This model provides an estimate of the insert temperature and this
criticaJ pcLTciraeter in
is
a
This
determining both the cathode lifetime and performance.
can be done for example, by picking an electron temperature and the plasma
These properties have been measured experimentally and found to be
potential.
rather insensitive to operating conditions with typical values of
volts respectively for a cathode operating at a few
—
0.8
eV and ^
8.0
amperes of discharge current.
Using these vaJues together with a specified surface work function (^) and the
desired emission current
(j
),
Equations (1) through (8) can be solved to provide
total
the
emission surface
geneicdly be
cin
iterative one requiring
Typical results are:
B.
temperature (T).
T =
1000°C. [Ref.
an
Note that the solution scheme would
initial
guess of a value for either
je
or T.
7]
PREVIOUS EXPERIMENTAL RESULTS
Much
of the published
work on standard hollow cathode
in the 1970's
and
1980's was done at Colorado State University, as illustrated by the work of Siegfried
[Ref.
3,7,8]
physical
and Willicims
phenomena
[Ref.
inside the
31,32,33,34].
This work was concerned with the
hollow cathode.
As part
determined search for other experimental results was made.
United States was done, or sponsored by,
1960's. [Ref. 17,35,36]
During
NASA
of this thesis work,
The
initial
work
Lewis Research Center
in
a
the
in the
this decade, parametric characteristics of the hollow
cathode were closely observed and described.
Many
similar investigations with hollow cathodes have been conducted by
other countries
U.S.S.R. [R^f.
-England
19],
China
[Ref. 5,6,13,14,20,22,23,24,25,26],
[Ref. 28]
and Japan.
[Ref. 29,30]
Germany
(Ref. 21,27],
Englsuid
1.
Work by
the RoyaJ Aerospace E8tablishment(RAE) and at Mullaid
Laboratory, provides a comprehensive look at behavior of hollow cathode discharges,
with
a variety
of
This investigation considered starting
cathode geometries.
behavior as a function of temperature, flow rate, voltage, geometry of the orifice and
dispenser,
and barium
Discheirge initiation
availability.
experiments using the
keeper electrode were done for a tubular insert cathode, a rolled
cathode, a curved orificed cathode and
non—bciriated cathode
foil
dispenser
For a given
design.
cathode and fixed flow rate and temperature, the voltage necessary to start a
Above the upper
discharge falls randomly between two limits.
will
limit,
a discharge
always occur, while below the lower limit one can never be obtained.
As
temperature and flow rate are increased, these limits approach each other until at
sufficiently
high values, they merge and behavior becomes reproducible.
design of a thruster system,
upper limit
it
is
desirable to choose these parameters so that the
always exceeded from these studies.
is
In the
Fig. 2.1
shows one typical
result.
[Ref. 6]
Further
experiments
at
RAE
physical processes operating in cathodes.
cathode are illustrated in Fig.
2.2.
provide
The
The cathode
basic
information
erosion or
orifice of
diamond
the
basic design features of this hollow
tip
was a tungsten disk
and 3.5mm diameter) electron—beam welded into a tantalum tube.
with a central
on
(1mm
thick
was provided
It
between 100 and 350/im diameter formed by either spark
drilling.
A
stainless—steel flange at the upstream end of the
cathode wa^ provided for mating with other components.
To
initiate
molybdenum
the discharge a keeper
disk with a
2mm
central hole,
was used.
and
it
This consisted of a thin
was spaced about
1mm
from the
••
c
o
1
/
>
o-
^
o
-5
A
r^
S.
c
Oi
~
f
fl
o
"O
— w
o ° z
•£
"5
„
5
_
•
«
c ^
* w
^
O
f7>
rs«
^— " '
•
__
^
•"
_— --""
/..J
/
^-'''
/
• ••
,
/
/
^3
•"o
CD
*
_
/
3"^
o^0»
-'
/
/
c •
!•
**
/
tt
^
>M
/
•»
o
'
/
^
-V-•
1
/
^"
'
1
/
/
/
•-
•
q:
/
/
•»
5a =
C * "O
»;;,;;£
f-i
tj
<-»
r
/
^^
"^
_
^-
%
^^
a.
5t
^\
/
/
3
1
"•
"
<•
O.
15
c —
•
^
5
^
^^
/
—^
»/
^/
v^*^^^
^^
• • « K
\
^^^^^^
«•
•
•
\
•
--•'
•-
,t_2
I
/
^^^y^''^
^
o
o
2
,— -"^
.'
«
•
-.-
1
«
•
1
*
O
o
—4
"^
^
•
•
• • •
•
•
•
•
.
1
•
•
•
•
/
•
•
1,a.
*
*
O
o
CO
o
•
1
8
•«
1
S
r»
<»f A
Fig. 2.1 Dischajge Ignition
•B«l|OA
Data
for
1
1
S
M
e
^
—
o
-.
o
uofivtliui •6j««t9«!a
Cathode with Rolled Foil Dispenser
[Ref.6]
6
•»
•—
Q.
^
cathode
tip.
stainless steel
From
The anode,
or
beam simulator
was normally a disk of
electrode,
whose distance from the cathode could be varied.
the result of these experiments, the following physical explanation of
hollow cathode behavior can be deduced.
emission mechanisms that
that thermionic emission
may account
is
The
results obtained
for the observed behavior.
suggest certain
seems certain
It
normally necessary to initiate the discharge from this
form of hollow cathode, but the
site of this
emission was not established.
However,
modification to include an internal auxiliary electrode showed that starting by field
emission
is
quite feasible, even in the absence of electron-emitting coatings.
thermionic emission
is
STAINLESS
STEEL
FLANGE
Thus,
not always required.
HEATER WIRE IN
STAINLESS STEEL
TUBE
TUNGSTEN
DISC
ORIFICE
^V
8.
WELD
/^S
•
m
9
'
•
•
'•
• • •
•••
VAPOR FLOW
» • ••. •< •
x:
HEATER EMBEDDED
TANTALUM
^
TUBE
EITHER ALUMINA
ULTRA TEMP
^5
IN
OR
516
mm
Fig. 2.2 Small-Orificed
Hollow Cathode
Once breakdown has occurred and a current
is
drawn
to
the
anode,
thermionic emission cannot possibly account for the high current densities obtained.
One
possibility is field-enhanced emission caused
by sufficiently high electric
fields
across the space charge sheath separating the walls of the cathode from an
internal
10
This
plasma.
effect wais
This
potential barrier.
work function
an electric
is
field
shown by Schottky
is
to be due to a reduction in the surface
produced by the external accelerating
decreased and the emission thus increased.
E
amount (eE/4Teo)
is
,
e is the electron charge.
It
is
I^'
=
E
if
reduced by an
follows therefore that
the saturated emission current obtained without an external field and
saturated current with an external field
the
Analysis shows that
applied to an emitter the work function
where
field so that
Is'
if Is is
is
the
then
exp(eE/4Teo)'/^
l^
^^.g)
KT
This
is
the Schottky equation of field emission. [Ref. 37]
The conditions within the hollow cathode
T
that
much
10'^
will
it
higher owing to the larger pressure and the presence of bcirium.
that the sheath
'd
Xj.
^ 10
potential of 5V.
at
unknown, but
is
possible
be close to the value measured outside(4xl0^ <*K) and that ne will be
cm~^ and assuming
Then
are
is
of the order of
- [/'''
L
4t ne
Taking ng ^
a Debye length Xr^ where
'^'-
J
e"^
(2.10)
-
the electric field gradient
G
^ 4x10
This results from a large increase
in
the ion density in the sheath
cm and
V/cm
for a plausma
the negative electrode, probably to values several orders of magnitude greater
than the prevailing ion density
in
the plasma. [Ref. 14]
Another emission mechanism that could be very
cathode
is
effective
the release of electrons by the impact of excited atoms.
11
in
the hollow
High yields are
to be expected
when
the excitation energies are not
function of the emitting surface.
number
of
processes
the
in
than a typical
discharge,
dimension.
orifice
and ion—neutral mean free paths are
By
considering the emission and absorption of
atoms back—scattered towcird the cathode by
cathode surface
in
current density, and further evidence for
mode
minimum
increased.
its
all of
the
a small area
capable of providing the required
is
existence
is
provided by the ability of
From
Fig. 2.3,
dischaige voltage approached a value close to
This corresponds to the
producing emission.
will
probable that
discharges to operate at potentials considerably lower than the
maximum
mercury atoms to the 4.9 eV metastable
It
is
collisions with ions reach
ionization potential of the propellant( mercury).
the
it
excited states and thus cause electron emission. [Hef. 14]
This emission mechanism, therefore,
stable spot
work
such as coUisional excitation or charge
resonance radiation, von Engel and Robson showed that
of the
greater than the
Excited propellcint atoms can be formed by a
transfer, provided that the electron—neutral
less
much
6V
it
can be seen that
as the pressure was
of the cross section for excitation of
level,
and these are very
effective at
[Fief. 14]
be noted that the preceding mechanism
is
not in any
on the presence of an alkali metal within the cathode, and
it
way dependent
should therefore
operate successfully in the absence of the triple carbonate coating.
It
was, in fact,
found that a discharge could be run without this coating, but the voltage required
was considerably higher than normal. This suggests that a mechanism requiring the
coating normally operated
in
conjunction with that dependent on metastable atoms.
The hollow cathode mechanism
preceding treatment would suggest.
is
undoubtedly
Although
it
more complex than the
seems hkely that,
in
some
either or both of the emission processes discussed so far are dominant, others
necessary to explain
all of
the data satisfactory. [Ref. 14]
12
cases,
may
be
16
cm
2.8
K
N\
)2
'min
(V)
10
1.8
cm
\
B
\\
\^ ^
V
0.8cm—
6
9
3
i
10
PRESSURE
Fig. 2.3 Variation of
Minimum
13
12
11
1
4
(torr)
Discharge Voltage with Pressure for Several
Sepairations. (Ref. 14]
2.
Germany
The
research work on hollow cathodes as plasma bridge neutralizers
for ion thrusters started at the University of
Groh and H.W. Loeb were concerned about the
tests with oxide coated rolled—foil inserts
A
with increasing operation time.
voltage raises rapidly in the
first
and then remains constant
at
ampere.
life
showed
time of the cathodes.
ain
Duration
increase of the keeper voltage
neutralizer system with impregnated insert
investigated in a shortened duration
ignitions after exposure to air.
S.E Walther, K.H.
Giessen at 1970.
The
test
of about
result
is
was
1,000 hours including some
graphed
in
Fig. 2.4.
The keeper
100 hours from 18 to 20 volts
about 20 volts.
The keeper
current
is
fixed at 0.3
Moreover, the dependence of the keeper voltage on the mass flow rate was
13
ecorded after 1,000 hours operation as shown in Fig.
resulting in lower keeper voltages at very small
2.6.
mass flow
The curve
got flatter,
rates. [Ref. 27]
30
m
s
5i1
Ij^s 300
UKe
UJ
I
o
>
'">'''"
20
It
I
u uu
^
u
o
'
t
I
I
I
o
It
It
about 20
a:
LU
ignitions
I
:
t
a
ma
ma
Ignition
Ignition after
exposefing tooir
10
1
ma =
0.014
SCCM
X
600
^00
200
600
1000
OPERATION TIME hrs
,
Fig. 2.4
Keeper Voltage Dependence on the Operation time
[Ref. 27]
30
after
UJ
§20
o
>
"Ae
ar>.
1000 hrs. operation.
/
o—o-
'>--x--
after
2hrs operation
UJ
a.
LU
10
5
10-15
MASS FLOW
Fig. 2.6
RATE, m*a(i
25
20
ma =
0.014
Keeper Voltage Dependence on the Mass Flow Rate
14
SCCM)
[Ref. 27]
'
U.S.S.R
3.
In 1988, parametric investigation of the hollow cathode for ion
thrusters was presented in the U.S.S.FL
This paper presented the result of
parametric investigations of a cesium hollow cathode with diameter of 5
2.6
shows typical voltage—current characteristics
current
is
of the discharge.
seen under the different mass flow rates.
mm.
Fig.
Saturation of
Appeaxcince of abnormal
resistance in a discharge gap can be predicted as follows;
—
emission current limitation as a result of a virtual cathode
discharge gap;
— discharge
carrier at
an anode
current limitation as a result of the lack of discharge
surfawre
and a positive anode decrease
of potential [Ref. 19]
3
mg/s
9x10
_2
+3
c
10
mg/s
Qi
1.03xl0~^mg/s
u
o
to
2
a
mg/s
1.1x10
<—
"o
03
1.16xl0~^mg/s
•H
10
20
30
Discharge voltap;e (V
Fig. 2.6
)
Voltage—Ampere DischcLrge Characteristics of Cesium Hollow Cathode
[Ref. 19]
15
in
a
China
4.
Pulse ignition charcict eristics for a hollow cathode for an electron
bombaxdment mercury
ion thruster
was presented
pulse igniter with positive pulse output of 0.1
ignition voltage of the hollow cathode
in
kV —
China
6
kV
in 1984.
A
high—voltage
wais developed.
was measured as a function
pulse repeat frequency and mercury flow rate respectively.
A
The
pulse
of pulse width,
comparison between
D.C. ignition voltage and pulse ignition voltage was also made. The pulse ignition
voltage dropped with the increcise of the pulse width.
the pulse repeat frequency
D.C.
ignition
was increased.
Fig. 2.7
voltage and pulse ignition voltage.
between pulse ignition voltage and mass flow
5.
Fig.
2.8
shows the relation
rate. [Ref. 28]
10,000 hours neutralize! hollow cathode endurance test was run by
The
the Electrotechnical Laboratory, Japan in 1984.
fabricated with the
same process
installed in a small
anode was
shows the comparison between
J^an
A
was
Also, the voltage dropped as
as one for
set before the
cathode.
mercury ion thruster).
liquid nitrogen trap
It
and a virtucJ
Parametric change tests and spectral analysis
No
were carried out every 1,000 hours.
observed after the 10,000 hours operation.
negligibly small,
ETS—III(5cm
vacuum chamber with a
was
tested hollow cathode
severe degradation of the cathode
Change
of the propellant flow rate
compared with the beginning flow
test log of the cathode. [Ref. 30]
16
rate.
Fig. 2.9
is
was
also
shows the entire
(V)
5x10^
L.C.
pulse 1.5us
.
4x10^
\
3x10^
2x10^
-
\^
-
m= 105ma
1 ma = 0.014 SCCM
\
\
\
'
1x10^
*
t
1
10 00.
1050
t
1100
1
1200
1150
Cc
Fig. 2.7
)
Comparison between D.C. Ignition Voltage and Pulse Ignition Voltage
[Ref.28]
xlO
(V)
tc=1157
T =1.
4
5
f=20
1
ma =
C
^s
H
z
0.014
SCCM
0.
20
40
60
80
100
ina(
Fig. 2.8 Relation
equ)
between Pulse Ignition Voltage and Mass Flow Rate
17
[Ref. 28]
•
40
o
h
•
'c
o
n
30
>
•
>
,
20
8
•
8
•
8
8
8
2
S
«
•
•
•
O
S
8
•
O
•
O
•
o
o
•
o
8
(^
8
S
.
•
8
01
1000
C
4000
3000
2000
8000
7000
6000
5000
TIME (Hours)
TIME (Hours)
40
»
30
8
s
•
8
•
•
o
•
o
8
20
10000
9000
TIME (Hours)
8000
Fig. 2.9 Test
Log
of the
Cathode
[Ref. 30]
HEATERLESS HOLLOW CATHODE
C.
With
inert gas (for exainple, Argon), the cathode heater
prevent condensation.
is
not needed to
Further, hollow cathodes in laboratory inert gas ion thrusters
have been started without a hollow cathode heater by flooding the cathode during
ignition.
Although
it
has been demonstrated that reliable heaters aie possible, some
view them as a failure prone component which
is
sensitive to fabrication procedures.
Ultimately, heaterless ignition of ion thruster hollow cathodes should contribute to
more
reliable ion thruster designs with
a lower parts count.
Heaterless inert gas ion thruster hollow cathodes were investigated with the
aim
of reducing ion
Before
the
hollow
thruster complexity and increasing ion thruster reliability.
cathode can ignite
18
without
a
heater,
the
propellant
must
breakdown
low
voltage
"arc"), while
(not
a heater. (Note that "ignition" meains establishing a
electrically without
(10
40 V)
to
high
current
(>1 A)
"breakdown" implies the onset of an
necessarily
an
Thus,
arc).)
is
it
electrical discharge in
important
mechanisms govern the heaterless breakdown
discharge
electrical
to
an
some mode
understand
first
of propellant.
(i.e.,
what
In this investigation,
Paschen's law served as the model of electrical breakdown.
Theory
1.
For clearer understanding, the derivation of Paschen's law
presented.
The breakdown voltage
is
briefly
is
C+ln(P*D)
where P
is
the pressure aind
D
is
the distance between the electrodes.
shows Vb as a function of (P*D), and the constants
given in Table 2.1.
For large values
of
linecirly
vises
For small values
and B
for several gases are
(P*D) because the logarithmic term
(P*D) the numerator
in
equation (2.11) decreases
with decreasing (P*D), but ln(P*D) decreases faster, with result that
when (P*D)
minimum) whose
(P*D)„,„
It
of
A
(P*D) the breakdown voltage Vb according
to equation (2.11) rises nearly linearly with
varies slowly.
7,
Fig. 2.10
is
value
=
lowered.
is
Hence there
minimum
a
is
found from dVb/d(P*D)
=
(2.72/A)ln(H-l/7) and (Vb)n>in
follows that, for example, the lowest
gases and cathodes for which
B/A
is
19
is
Vb(called Paschen's
0, viz.
=
B(P*D)„in
breakdown voltage
small and 7
Vb
is
(2.12)
to be expected for
large (Fig.2.10). Table 2.2
shows
the
minimum
values for a
equation (2.12).
number
For instance,
for
of gases.
The
general trend
a given cathode
in
often smaller and 7 larger than in molecular gases
Again B
is
small for the raxe gases and so
is
is
Law (Breaikdown
and thus (P*D)min
is
10^
mm
A
is
larger.
Vb- [Ref. 15]
Hg cm
voltage Vb as a function of the reduced
electrode distance P*D). [Hef. 15]
20
agreement with
rare gases, the constant
,
Fig. 2.10 Paschen's
in
TABLE
Gas
VALUES OF COEFFICIENTS A AND B
VARIOUS GASES. [REF. 15]
2.1
1
A-
cm
mm
Hg
N»
Hz
12
Air
15
20
13
12
3
20
cm
Ar
He
Hg
TABLE
2.2
Hg
342
139
365
466
290
180
34(25)
370
V m in
(Volts)
He
Ne
Ar
X/P
100-600
20-1000
100-800
500-1000
150-1000
100-600
20-150(3-10)
200-^0
MINIMUM SPARKING POTENTIALS
Cathode
Fe
150
244
265
275
450
330
295
420
425
520
330
335
N2
O2
Air
H2
Ft
CO2
Hg
Hg
?
Fe
Hg
Na
Hg
Fe?
W
21
FOR
Range of validity
mm
5.4
CO2
H2O
Gas
V
B
IN (2.11)
[REF
15]
(P*D),in
(mm Hg cm)
2.5
3
1.5
0.75
0.7
0.57
1.25
0.5
1.8
C!2
7
0.04
Law has been experimentally
Paschen's
D
gases)and
satisfied
(parallel
geometry) (Fig. 2.10); however, neither
plate
standard
the
in
hollow
nonplanar and the pressures
cathode
orifice plate
remain the same,
the
i.e.,
cire
cathode
geometry
because
P
(static
criteria
is
geometry
the
is
nonstatic in the breakdown region between the
Nevertheless, one would expect the trends to
and the keeper.
when breakdown voltage
hollow cathode, perhaps a Paschen
MMHG*CM
established for well defined
is
plotted as a function of
minimum
value characteristic of well defined
P*D
could be found
P*D
for
near the
Examining
cases (Fig. 2.10).
the standard hollow cathode under this assumption, with reasonable estimates of
and D, the
(0.001
P*D
MMHG)
product
*
is
seen to be well below this characteristic value
(0.15CM)
=
0.00015
MMHG*CM).
P
(P*D =
Hence Paschen's law
well suited to the conditions in a stcindard hollow cathode.
1
is
not
Paschen's theory would
qualitatively explain the experimental observation that for heaterless ignition of
hollow cathodes high flow rates are required.
increasing
level
P through
satisfied
by
increased flowrate brings the breakdown voltage
open
the
circuit
voltage of the
igniter
heaterless ignition at reasonable igniter supply voltages
electrical
breakdown
minimum breakdown
P*D
Increasing the
in
heaterless
down
to the
Demanding
supply.
(<1KV)
hollow cathodes should
product by
implies that
occur
voltages, typically on the order of 200 to 400
V
near Paschen
for
most gases.
(Fig. 2.10)
Under Paschen's Law, the breakdown voltage
cathode can be lowered
in
a
number
of
ways
(e.g.,
for
the
hollow
lengthening D, increasing P,
"seeding" the propellant with a low ionization potential material,
2.
heaterless
etc.). [Ref. 10]
Hardw8u-e
Heaterless inert gas ion thruster hollow cathodes were investigated
22
with the aim of reducing ion thruster complexity aind increaising ion thruster
One
reliability.
of the heaterless
This design
[Ref.18]
thesis as the
alternative
is
cathode
is
the design invented by Aston, 1981.
manufactured by Spectra—Mat
Spectra—Mat cathode. (Fig. 2.11)
the
to
refractory
Inc.
is
referred to in this
This cathode design provides an
The Spectra—Mat model
cathode.
raetaJ filcLment
and
modifies the original design by Aston by including a tungsten dispenser.
Porous
tungsten with a formula of barium oxide dispersed throughout the matrix
essential
form of dispenser cathodes. The claimed performance
hollow cathode
device
is
is
the
Spectra—Mat
for the
a starting time of approximately ten seconds after which the
is
capable of emitting several amperes of electron
With argon, a flow
requirements are low.
rate
of
Gas flow
current.
3-^ seem (standard cubic
centimeters per minute) will support a cathode emission current level of 5—10
amperes.
Lower gas flows can be used
for
smaller electron emission current
requirements such as ion becmi neutralizer applications.
characteristic
of
hollow
cathodes
requires
that
the
An
anode
regulating with a compliance voltage of about 80—100 volts.
inherent
supply
The
operating
be
fast
current
emission
current response of the Spectra—Mat hollow cathode
means that anode power supply
Most
trcinsistor regulated laboratory
response times less than
power supplies
1
msec are required.
satisfy this requirement.
There are several advantages claimed for the Spectra—Mat hollow cathode.
Some
of these
advantages axe
— Much
longer cathode
hollow cathode and so
the ionizing
listed here.
is
life.
The
electron emitting surface
shielded from sputtering damstge by the
is
within the
50—100
volt ions in
plasma
— Lower
power consumption and operating temperatures. This results
undesirable substrate and process chamber heating.
23
in less
— Less
is
subject to
A
chance of ion beam and substrate contamination.
much
evaporation because of
less
it's
lower operating temperatures.
unique feature of the Spectra—Mat hollow cathode
cathode to be placed
in cin idle
chamber plasma
beam
or ion
is
mode where
The hollow cathode
the cathode
being produced. This feature
is
is
is
the ability of the
on but no discharge
especially useful in
continuous operations, such as ion etching, ion sputter deposition, ion milling and
ion implantation.
substrate
erosion,
is
In these applications, the ion source can be pulsed on as each
put in place.
This minimizes the chamber heating and ion sputter
promoting a much cleamer substrate environment.
Fig. 2.11
new
[Ref.
1]
Spectra—Mat Hollow Cathode Apparatus.
24
III.
A.
EXPERIMENTAL EQUIPMENT AND PROCEDURE
EXPERIMENTAL EQUIPMENT
1.
Discharge
Fig.3.1
arrangement.
is
Chamber
the
picture
that
illustrates
There are four main components
in the
the
general
vacuum chamber.
cathode, keeper, anode and Langmuir probe.
Fig. 3.1 General
ExperimentcJ Arrangement.
25
experimental
They
are
Two
hollow cathodes (a standard design manufactured by Ion—Tech
and the Spectra—Mat cathode) are mounted
cathode can be used at a time.
to each other.
Only one
The standard cathode was mounted
so that the
parallel
distance between the cathode tip and the keeper could be varied.
of the thesis
by Park. [Ref.SS]
Both
This
is
the subject
of the cathodes are connected to copper tubes
by swaigelock connectors and receive axgon
ga.s
The copper
through these tubes.
tubes were disconnected in the middle and they were replaced by tygon tubing for
electricaJ isolation.
The Spectra—Mat hollow cathode has the keeper
in its
body,
but a similar appearing external anode.
A
Lcingmuir probe has been placed in the chamber to measure the
electron temperature, electron current and
plasma density.
a stainless steel ball (diameter 9.614mm) and
bai.
This baj
is
it is
The Langmuir probe
is
connected by a thin stainless steel
covered with ceramic to insulate
it
from the plasma
in
the
chamber.
The anode,
and to take the
or collector,
is
designed to collect the discharge current
role of the electric field of the space
copper grid which surrounds the inside wall of the
2.
environment.
The anode
is
a
jcir.
Electrical Circuit
Fig.3—2 illustrates the electrical circuitry for the Ion—Tech( standard)
hollow cathode. As expected, High voltage (^300V)is applied to
Right after the ignition, two circuits
exists.
through the very high resistor(llOkQ).
supply
is
stcirt
the dischcirge.
However, only a low current can flow
Therefore only the low voltage, high current
active after the initial igniter.
Fig.3—3
illustrate
the
heaterless hollow cathode.
26
electrical
circuitry
for
the
Spectra—Mat
Fig. 3.2 Electrical Circuitry for the
27
Ion-Tech Hollow Cathode.
Fig. 3.3 Electrcal circuitry for the
28
Spectra—Mat Hollow Cathode
Measuring Equipment
3.
Two
Varicin
pressures in the rough
to
Type 0531 thermocouples were used
vacuum range and Varian 880RS
measure the
to
ionization gauge
was used
measure the high vacuum pressures.
Fluke 85 multimeter, Fluke 75 multimeter and Keithley 195A
DMM
were used to measure the anode to keeper current, the keeper to cathode current
and keeper to cathode voltage respectively.
HP
noise
model 120B oscilloscope was usually used
and the plasma
to
watch the system
oscillation.
HS—lOS
Hasting mass flow transducer and Nail flow meter are used to
measure the argon gas flow
rate,
(unit:
SCCM—standard
cubic centimeters per
minute)
Vacuum System
4.
Fig.3—4 shows the major parts of the Varian system.
system consists of two pumps.
A
Rotary Vane Oil—Sealed Mechanical
3
for
rough pumping; pressure range 760 torr to 10
vacuum pumping;
a high
range
pressure
10
torr.
torr
rate
and
vacuum chamber
Fig. 3.5
pressure.
Pump
A Turbo pump
to
10
experimental system, base pressure without propellant flow
valves to the gas supply open.
This vacuum
is
torr.
is
is
used
used for
For
this
2.8x10"^ torr with
shows the relationship between argon flow
Chamber
pressure
increases
linearly
by
increasing the flow rate as expected.
B.
Procedure
1.
Vacuum System
In order to start this experiment, the Bell Jaj should be evacuated to
_6
the order of 10
torr.
29
'
'
VACUUM
CHAMBER
HIGH VACUUM
VALVE
TURBO
PUMP
Fig. 3.4
MECHANICAL
PUMP
Major
Variam
Psurts of
Vacuum
System.
12.00 -
-e-
10.00
l_
-
•
•-
-1
o
•
,_
o
o
o
o
OX
ff
/
Z
/
'/
/
:
8.00 ^
a
y
• •
/
H
t
-'./'
'
tr
«;
6.00 -
D
3
c
y'
;
in
(U
>4
*
i_
j^ t
CL
^
4.00
-
/
/^
<u
£>
j/i
)
E
{v/
:
o
JZ
u
2.00
T
u.uu
0.00
1
1
1
1
1
1
r
1
4.00
2.00
6.00
8.00
10.00
Gas Flow Rate (SCCM)
Fig. 3.5 Relation
between Propellaint Flow Rate and Chamber Pressure
30
It
usually takes several hours to reach the desired pressure range of the
pumping
chamber and up
to 24 hours
for operating the
vacuum system
to assure complete outgassing.
vacuum
The procedure
as follows;
is
Turn on the cooling water and open the nitrogen gas
(A)
bottle valve (Nitrogen bottle valve
is set
to 2.5
— 6.0
psi as the regulated pressure)
(B)
Place switch marked "Turbo
(C)
Switch on
(D)
Push "Start" on
Pump"
to "Off" position
power to turbo controller and ionization
gauge
start,
by
turbo
TC2
pump
should not.
Let system
Houghing pump should
controller.
pump down
to 100 milli torr as indicated
gauge on ionization gauge panel.
With
(E)
Pump"
Switch "Turbo
depressed)
controller in
switch to
"Low Speed"
(i.e.,
"On" when 100
Low
speed button
milli torr reached
"Acceleration" and "Leak" indicators will
(F)
light.
Both
should go out and "Normal" indicator will light within 6 minutes
system to
(G)
Switch on ionization gauge to
When
cathode has been exposed to atmosphere, allow vacuum
pump down
2.
at least overnight before
Starting and shutting
(A)
pump
the plasma source
As mentioned above,
has been exposed to atmosphere)
The gas
lines
make
sure the
(2)
lA increments
for the initial start
is
8A
(if
—
it
takes 1—2 hours to
to
4A
for
10 minutes.
waiting 10 minutes between each increase.
31
cathode
sealed.
Set the heater current
to
up
vacuum system has pumped down
should be flushed twice with argon
out the gas lines once the argon cylinder
increase in
attempting to start plasma.
Standard Hollow Cathode
(1)
overnight.
down
reaxi pressure.
Then,
Start the argon flow at 3--5
(3)
current to 9A.
Set the
Wait several minutes.
increase the flow rate to 5 or 6
—
SCCM. Wait
When
SCCM.
3.0
Do
bypass valve 1/4 turn.
will overload.
again.
If
does not
discharge
nothing happens, reduce flow
not leave the bypass valve open or the
When
plasma source
the
reduce the heater current to
6A and
start,
flow stabilizes, quickly open and close the
voltage and rapid increase in current will occur.
When
vacuum system
ignites.
starts,
a sudden
this happens,
drop
in
immediately
adjust the flow rate to the desired level. (2
—4
range works best)
Change paxameters as
(6)
best.
(If
Repeat as necessary until a stable discharge
(5)
SCCM
Increase the heater
anode voltage to 2300V.
(4)
rate back to 2.5
SCCM.
Rapid changes
in current (i.e. 0.2
To
(7)
shut
-»
required.
2A) may cause
down
loss of
Slow changes work
plasma.
the plasma source, Switch off the power
supply and reduce the heater current slowly to OA. (Opposite to the initiation)
Allow argon gas to flow
Do
(8)
(It is
SCCM.
at 2
not switch off
vacuum system
best to wait at least one hour to allow complete cooling).
occur, close gate valve
vacuum
and shut
in the Bell Jar while
(9)
off
cathode
To
argon flow immediately.
while cathode
If
power
This will maintain a
cools.
restart
SCCM
hot.
loss should
plasma source when plasma
loss occurs while
changing parameters, bring up the heater current to 9A immediately and
propellant flow rate to 3
is
and wait.
Discharge usually auto—stcu:ts quickly.
Restarts are normally quicker and easier than the
32
set the
initial daily start.
(B)
Spectra—Mat Heaterless Hollow cathode
(1)
The
(2)
Increase the flow rate to 6
(3)
Turn on the power supply and increase the
(4)
Wait 30 seconds.
initial
stage
is
the
same
as that of the
Ion—Tech
cathode.
SCCM.
voltsige
slowly to 270V.
Do
close bypass valve 1/4 turn.
vacuum system
is
If
(6)
When
nothing happens quickly open and
not leave bypass valve open or you will cause
plasma
to overload.
If
still
does not start, try again when pressure
stable.
the
plasma source
starts,
a sudden
drop
in
voltage (to ^ lOV) and rapid increase in current (to ^ 1.5A) will occur and voltage
regulated
mode changes
(6)
better.
Rapid change
to current regulated
Change parameters as
in current
(7)
mode.
To
may
shut
required.
Slow change works
cause loss of plasma.
down
slowly to zero. (opposite to the initiation)
33
the plasma source, decrease the current
Allow argon gas to flow
at
^ 2
SCCM.
IV.
The purpose
make
EXPERIMENTAL RESULTS
of this investigation
is
optimum
to find the
paicimeters that
the hollow cathode initiation certainly and operation properly.
Not only the
continuous, long—life operation but also proper, quick and reliable starting condition
is
important
the
for
hollow
cathode
behaviors
St anting
discharge.
conditions vary according to the following parameters;
running
propellant flow rate(m),
biasing potential between cathode and keeper(Vk), cathode tip temperature(T) or
heater current(Ih), keeper spacing(D) and time of turning on and
A.
off(t).
STANDARD HOLLOW CATHODE
1.
Flow Rate Dependence
It
(heater current
was
temperature slowly to about 1300<>C
sufficient to increase the
8A) and
to hold
steawiy for
it
about
1
hour.
Discharge initiation
could then be accomplished by passing a sufficient flow rate of cirgon through the
cathode while applying a potential
At
data could be taken.
first,
Vk
Vk wa5
to the keeper.
increased,
2.
Fig. 4.1
Vd became
this condition, several
slowly increased at constant flow rate until
This process was repeated
discharge occurred at a voltage Vddifferent flow rates.
From
shows the
result.
From
many
times at
this result, as flow rate
smaller, discharge occurred fcister and
was
more predictably.
Heater (Temperature) Dependence
Vk was
a potential Vd-
Changing the
increased at constant temperature until discharge occurred at
This was repeated at different heater current (temperatures).
heater
current
was
used.
34
Fig.
4.2
shows the heater current
dependence.
From
Fig.4.1 and Fig. 4.2, as either(flow rate or temperature)
increased,
Vd became
current 5
SCCM
With the flow
smaller and more predictable.
rate
was
and heater
and 8 A, the discharge could often be initiated at voltages as low
as40 V.
40.00
-.
38.00
\
•
en
bo
\
(Volts)
<u
34.00
T
:
1
'-
. 32.00 i
D>
;
i1
O
-c
:
m 30.00
Q
-.
:
^
28.00 i
26.00 ^ Till
0.00
III!
>
1
T
2.00
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
T
1
1
1
1
1
1
T
4.00
6.00
8.00
Gas Flow Rate (SCCM)
Fig. 4.1 Discharge Voltage vs
35
Flow Rate
1
1
1
1
1
1
10.00
13.50
0.00
I
I
1
1
1
1
I
1
I
1
1
1
1.00
1
1
1
1
1
1
1
I
[
I
1
1
1
1
f
1
1
1
1
I
1
1
1
1
1
1
1
1
1
1
1
I
I
1
1
1
1
1
I
2.00
3.00
4.00
Heater Current (Amperes)
5.00
Fig. 4.2 Dischau-ge Voltage vs Heater Current
Time Dependence
3.
On
closer exanmi nation,
as unpredictable as at first thought.
hours,
Vd was
contrast,
it
Vjc
was
initiation
was not
After the discharge had been off for several
generally high at given values of flow rate and temperature.
was considerably lower
had been switched
of
was found that discharge
it
also
off.
if
Vk was
The time taken
dependent
In
reapplied shortly after the discharge
for the discharge to strike after application
upon the recent history
of
cathode.
the
In
this
experiment, the values of temperature (heater current) and flow rate were held
constant, the discharge was extinguished and a
before a fixed value of
from the cathode
Vk was
orifice;
microamperes. (Fig. 4.3)
this
This
applied.
At
known time was allowed
this stage,
was accompcinied by a
is
a very
idle
faint
mode
glow emerged
current of several
the thermionic emission current.
36
to elapse
The luminosity
gradually increased, as did the current, until a discontinuous rise to several hundred
milliamperes indicated discharge ignition.
The maximum
approximately
of
constant
given
at
values
flow
the surface migration of barium are responsible.
essential
is
switched
is
as
is
implying that sufficient barium
orifice,
Once the discharge
there.
mechanism
applicable, adequate thermionic emission
from areas close to the cathode
must be available
baxium
is
The
that chemiccJ changes or
the initiation
If
current was
temperature.
auid
phenomena suggest
relatively long times involved in these
discussed in the b2u;kground section
rate
mode
idle
off,
would appear that
it
gradually lost from the emitting zone,so that, after reapplication of Vk,
a finite time
is
The
required for replenishment.
situation
is
undoubtedly extremely
complex, and no attempt has been made to ascertain the nature of the chemical and
surface processes taking plcice.
would be reasonable to assume, however, that the
It
geometry and position of the dispenser have by no means been optimized.
0.00
0.00
I
Fig. 4.3 Idle
t
I
1
Mode
T
I
I
I
I
I
I
I
I
I
1
I
I
I
I
I
I
I
I
I
I
I
I
I
'
I
I
I
I
300.00
200.00
100.00
Cathode to Keeper Voltage (Volts)
I
I
I
I
'
I
I
I
400.00
Discharge Current vs Keeper Biasing Potential
37
SPECTRA-MAT HOLLOW CATHODE
B.
To
find out the
optimum parameters
for
Spectra—Mat hollow cathode to
Flow
operate properly, similar experiments were attempted.
potential
were the main parameters.
rate
and biasing keeper
Unlike the standard hollow cathode, the
When
Spectra—Mat hollow cathode does not have a heater to activate the cathode.
a sufficient biasing keeper potentiaJ(^ 315V) wais given, a very small discharge
occured within the hollow cathode.
mode
This
is
the so called idle
discharge begcin in a short time(<10 sees).
second, outer anode must be
1.
Mode
Idle
During
mounted and a
discharge.
this experiment,
In order to extract a current,
Fig. 4.4
shows these two
flow rates.
When
the idle
mode was
idle
levels of idle
mode was
mode
found, lower level
discharge for different
higher level, extraction of discharge was easier.
level is insufficient for ctctivation of externcJ dischsu^ge.
migration of the barium to the surface of the cathode.
also
depends on the biasing keeper voltage.
this
dependence.
Fig. 4.5
Idle
mode
To
discharge current
shows one typical example
explainable
mode
by
discharge cam be seen.
field
enhanced
emission.
approximately one order of magnitude higher current
presumably due to the more
This
idle
Note
in
idle
mode
that
effective fields to the cylindriccJ capacitor
presented by the Spectra—Mat cathode.
38
discharge
this
mode.
of
Above
Discharge current has the transition point around 295V.
this range, the inside idle
reach
That helps
the higher level, hollow cathode should be turned on for several hours.
be
a
biausing potential applied.
two kinds of
level.
should
Idle
Discharge
amd higher
That means lower
mode
design
This
is
which
is
2.00
HIGH LEVEL
(n
Q.
E
<
-
1.50 -
c
1.00
I
0)
I-
o
J=
.^
o
LOW LEVEL
0.50
00
1
1
1
1
I
1
1
1
ri
I
'
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
10.00
8.00
6.00
Gas Flow Rale (SCCM)
Mode Discharge
Fig. 4.4 Idle
I
I
4.00
2.00
0.00
for Different
Flow Rate
V)
9
SCCM
8
SCCM
7
SCCM
G
SCCM
5
SCCM
A
SCCM
3
SCCM
2
SCCM
^ 1.50
Ql
E
<
-
1.00 H
l_
D
o
4)
i_
° 0.50
o
'U
0.00
'I
M
M
M
M
50.00
100.00 150.00 200.00 250.00 300.00 350.00
Cathode to Keeper Voltage (Volts)
I
I
I
I
I
I
I
Fig. 4.5 Idle
Mode
I
I
1
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
r
I
I
I
I
I
I
I
I
I
I
Discharge Current for Different Biasing Voltage
39
Extraction of Discharge
2.
Discharge extraction was unhke that found for the standard hollow
cathode.
If
discharge extraction was easy and relible, the Spectra—Mat hollow
cathode would be highly recommended.
took a long time for the
first
Unfortunately,
beginning attempt.
was not easy
it
at all.
It
Instead of the heater power
supply, the Spectra—Mat hollow cathode needs smother power supply to bias the
anode to keeper potential.
taken.
The problem was
That compensate each other.
that the status of the discharge
Several data could be
was not stable and did not
run long.
There were two kinds of mode
for this extraction.
For the
first
the external glow was seen, but the collected discharge current was small.
the result at the
bottom of Table
around 10 minutes.
this
mode
4.1.
This
mode was somewhat
However, without collecting
stable.
sufficient external
mode,
This
It
is
lasted
current(l—2A),
is useless.
The second mode more
cleaily
matched our expectations.
Some data
were taken as follows.
TABLE
4.1
DATA FOR SPECTRA-MAT CATHODE
Flow Rate
(SCCM)
high
3.75
2.96
5.59
The range
76.5
60
25
307
of values of discharge voltage
discharge fluctuated too
External
Iiiternal
Vci(V)
much and
Ii(A)
Vd(V)
Ie(A)
1.22
99.1
1.3
1.1
1.3
1.1
99
99
0.026
60.2
0.03
and current
is
cicceptable.
the values were unpredictable.
40
1.3
However, the
That means that
Spectra—Mat hollow cathode
not as good as a standard hollow cathode for most
is
electron and ion emission purposes.
Discharge Failure
3.
In the beginning of these experiments with the original Spectra—Mat
cathode, the disch«irge was very difficult to start.
to a very high value, (>10
10
torr, the
SCCM),
When
the flow rate was increased
the chamber pressure increased to the order of
discharge would start momentarily.
The
status was very unstable.
Ultimately, the cathode was destroyed by overheating and arcing.
There seems to
be three possible reasons for the discharge failure.
First, the
The hollow cathode might have been contaminated by
time.
dust,
cathode had been exposed to the atmosphere for too long
oil,
To
etc during this period.
too
much water
vapor,
prevent this kind of problem, mechanical work
should've been finished completely aind checked several times before opening the
The hollow cathode should be
cathode shipping containers.
chamber when not
A
initial
attempts
in
vacuum
in use.
second possibility of the problem
at
stored
ignition,
is
surface damage.
During the
both flow rate and biasing potential between cathode
and keeper were too high. The flow rate was increased because the discharge started
in
only
that
consumption
condition.
of
barium oxide.
lower flow rates, or
A
flaws.
That might cause the surface damage and too much
in idle
third
Further operations should have been attempted at
mode.
possible reason for the discharge fciilure
The cathode might have had uneven
is
manufacturing
surfaces from the beginning.
It
observed that the discharge location was not always at the front of the cathode
Sometimes discharge occurred around 2—3
41
cm
back from the cathode
tip.
was
tip.
Fig. 4.6
and Fig. 4.7 shows the damaged cathode surface spot and broken part of ceramic
The damaged
insulator respectively.
This
is
spot
is
the plaice where the discharge occured.
unexpected and the main reason for the ultimate discharge
Fig.4.8
cathode.
is
failure.
a detailed dicigram of disassembled Spectra—Mat hollow
This old cathode was disassembled after discharge failure for better
understanding
of
the
geometrical
That
structure.
understanding the inner structure of the cathode.
was
very
helpful
for
After several other attempts to
recover the cathode operation, barium oxide was recoated over the surface of the
cathode
tip.
Liquid barium oxide was used for this work.
but the results were erratic.
helped the discharge,
outer anode was mounted in front of the cathode
Biasing the external voltage helped the bright ignition of discharge.
tip.
cathode acted
Still
One
It
like
a standard hollow cathode.
Bieising voltage
The
was high.(Kl90V)
unstable discharge and even discharge failure happened because of uneven
coating of barium oxide and the
damaged
surface.
The
results presented
above were
obtained with a second cathode obtained near the end of the research period.
noted,
it
was
As
also difficult to operate, though the experience gained in the initial
experiments helped.
42
Fig. 4.6 Picture of
Damaged Cathode
43
Surface
Fig. 4.7
Broken Tip of Ceramic Insulator
44
&
rim
II
tpim
I
<
Ijirm
III
cnrm
I
ym
I
I
I
wOO
WB
^
^
Fig. 4.8 Detailed
Diagram
of Disassembled
45
Spectra—Mat Hollow Cathode
CONCLUSION
V.
Extensive
investigations
the
of
and
behaviors
starting
pcircimetric
chaxact eristics of two (standard and Spectra—Mat) hollow cathode designs have
shown how the
ignition
cind
operation are dependent
on propellant flow rate,
cathode tip temperature (for standard hollow cathode), biasing potential, geometry,
and the availability of a low work function material
temperature or flow rate are increased, the
voltage
range both decrease until,
potentials often below 40 volts.
high
at
maximum
values,
like
As
barium oxide.
value and width of this
starting
is
reproducible
at
This behavior appears to be strongly influenced by
the site and rate of dispensation of the low work function material.
For the Spectra—Mat hollow cathode,
after (within
idle
mode disch^ge occurred
10 seconds) applying a voltage of ^315 volts.
—
extraction, an external anode should be
mounted
biasing potential should be applied.
Discharge extraction
cind
50
For the discharge
100 volts anode to keeper
is
possible,
condition was so erratic and the values of data were unpredictable.
reasons,
Spectra—Mat hollow cathode
is
not highly
emission purposes.
46
right
recommended
but,
the
For these
for cJl electron
REFERENCES
1.
Kim
Guntber, "Hollow Cathode Plasma Source" ( Spectra—Mat Hollow
), Spectra—Mat Inc., Watsonville, California
Cathode Manual
Aston,Graeme, "Summary Abstract: A Hollow Cathode for Ion Beam
Processing Plasma Sources", Jet Propulsion Laboratory, CcJifornia Institute of
Technology, Pasadena, California 91109
2.
3.
Daniel E. Siegfried and Panl J. Wilbnr, "An Investigation of Mercury Hollow
Colorado State University, Fort Collins, Colorado
Cathode Phenomena
4.
',
Daniel E. Siegfried and Paul
J.
Wilbur, "Studies on an Experimental Quartz
University, Fort Collins, Colorado
Tube Hollow Cathode", Colorado State
CM. Philips and D.G. Feam, "Recent Hollow Cathode Investigations at the
Royal Aircraft Establishment", Royal Aircraft Establishment, Farnborough,
Hampshire, Englaind
5.
6.
D.G. Fearn, Angela S. Cox and D.R. Mof&tt, "An Investigation
Hollow Cathode Discharges"
of the
Initiation of
7.
Daniel E. Siegfried,
"A Phenomenological Model Describing
Orificed,
Hollow
Cathode Operation"
Daniel E. Siegfried and Paul J. Wilbur, "A Model for Mercury Orificed
Cathodes: Theory ajid Experiment", Colorado State University, Fort
Collins, Colorado
8.
Hollow
Daniel E.Siegfried "Xenon and Argon Hollow Cathode Research", Colorado
State University, Fort Collins, Colorado
9.
10.
M.F. Shatz, "Heaterless Ignition of Inert Gas Ion Thruster Hollow
Cathodes", National Aeronautics and Space Administration, Lewis Research Center,
Cleveland, Ohio
William D. Deiniger, Graeme Aston, and Lewis C. Pless, "Hollow Cathode
for Active Spacecraft Chaige Control", Jet Propulsion Laboratory,
California Institute of Technology, Pasadena, California
11.
Plasma Source
12.
Graeme Aston, "Hollow Cathode Startup Using a Microplasma Discharge",
Sr. Engineer Electric Propulsion and Advenced Concepts Group, Jet Propulsion
Laboratory, Pasadena, Ccdifornia
13.
D.G. Pearn and CM. Philip, "An Investigation of Physical Processes in a
Hollow Cathode Discharge", Space Department, Royal Aircraft Establishment,
Farnborough, Hampshire, England
47
14.
CM.
15.
Alfred von Engel, "Ionized Gases", Keble College, Oxford, England
Philip, "A Study of Hollow Cathode Discharge
Aircraft Establishment, Faxnborough, Hampshire, England
Chaj act eristics", Royal
16.
Cobine,J.D.,
"Gaseous
Conductors—Theory
Engineering
and
Apphcations",lst ed., Dover Publications, Inc., New York, 1958, chps.5,8,9.
NASA
V.K. Rawlin and E.V. Pawlik, "A Mercury Plasma-Bridge Neutralizer",
Lewis Hesesirch Center, Cleveland, Ohio
18.
Aston, G., "Holow Cathode Apparatus"
17.
NASA
Pasadena
Office,
CA. 1981
Maslyany N.V., Pridantzev V.F., Prisnyakov V.F.,Khitko A.V., "Parametric
19.
Investigations of the Hollow Cathodes for Ion Thrusters", Dniepropetrovsk State
University, U.S. S.R.
CM. Phlip and M.O. Woods, "Some Investigations of Hollow Cathodes with
Porous
Dispensers",
Aircraft
Space
Department,
Royal
Establishment,
Faxnborough, Hampshire, England
20.
K. Groh and S. Walther, "Neutralizer Investigations: Performance Mapping,
Position Optimization, Durability Tests". 1st Institute of Physics, University of
Giessen, FRG
21.
G.L. Davis (CML, Mullard, Mitcham, Surrey), D.G. Feam (RAE,
Fcirnborough, Hants), "The Detailed Examination of Hollow Cathodes Following
Extensive Life Testing"
22.
D. Newson, M.G. Charlton, G.L. Davis, "Starting Behaviours
Cathodes including Multiple Starts", CML Mullard, Mitcham, England
23.
M.G. Chilton, G.L. Davis, D. Newson,
24.
for Ion Thrusters", Mullaxd,
Mitcham, Central
of
Hollow
"Investigations on Hollow Cathodes
Matericils Laboratory
D.G. Fearn, "The Operation of Hollow Cathodes under Conditions Suitable
Beam Neutralization", Space Department, Royal Aircraft Establishment,
Farnborough, Hampshire, England
25.
for Ion
D.G. Fearn and T.N. Williams, "The Behavior of Hollow Cathodes During
LongJ-Term Testing in a 10cm Ion Thruster and in a Diode Discharge System'
Space Department, Royal Aircraft Establishment, Fcu-nborough, Hampshire,
England
26.
,
S.E. Walther, K.H. Groh and H.W. Loeb, "Experimental and Theoretical
Investigations of the Gissen Neutralizer System", Gissen University, Gissen, FRG
27.
28.
Chaj
Chang Shi—liang, Hu Yong—nian. Hong Huai—yen, Sun Yu—zhu, Ren Ni,
Xiao—ming, "Hollow Cathode for Electron Bombardment Mercury Ion
Thruster Pulse Ignition Characteristics", Lanzhiu Institute of Physics, China
48
Y. Nakamura and K. Miyazaki (National Aerospace Laboratory, Chofu,
Toko, Japan), S. Kitamura and K. Nitta (National Space Developement Agency,
Tokyo, Japan), "Long Time Operation Test of Engineering model of ETS—III Ion
Thruster"
29.
Isao Kudo, Kazuo Machida and Yoshitsugu Toda, "10,000-Hours Neutralize!
Hollow Cathode Endurcince Test", Electrotechnical Laboratory, Ibaraii, Japan
30.
31.
NASA
32.
John D. Williams, Paul J. Wilbur, "Space Plasma Contactor Research",
Plasma Contactor Research 1988 pl-27
John D. Williams "Investigation of a Hollow Cathode Failure",
Advenced Electric Propulsion Research 1988, p39—46
NASA
John D. Williams, Paul J. Wilbur, "Plasma Contacting;
Technology", Colorado State University, Fort Collins, Colorado
33.
An Enabling
John D. Williams, Paul J. Wilbur, "Ground-Based Tests of Hollow Cathode
Plasma Contactors", Colorcuio State University, Fort Collins, CO 80523
34.
35.
J.W.
Neutrcilizers",
Plasma
Ward, H.J. King, "Mercury Hollow Cathode
Hughes Research Laboratories, Malibu, California
Bridge
David F. Hall, Robert F. Kemp, Haywood Shelton, "Mercury Discharge
Devices and Technology",
Systems Group, Redondo Beach, California
36.
TRW
37.
J.
Thewlis, "Encyclopaedic Dictionary of Physics", Vol.
6,
Park, Young—Cheol, "Hollow Cathode Plasma Source
Master's Thesis, Naval Postgraduate School, Monterey, CA
38.
49
P
414.
Characteristics",
INITIAL DISTRIBUTION LIST
No. Copies
Defense Technical Information Center
2
Cameron
Station
22304-6146
Alexandria,
VA
Library,
Code 0142
Naval Postgraduate School
Monterey, CA 93943--5002
2
Physics Library Code 61
Department of Physics
Naval Postgraduate School
Monterey, CA 93943-^5000
1
Department Chairman, Code 61
Department of Physics
Naval Postgraduate School
Monterey,
CA
CA
1
93943-^5000
Professor R.C. Olsen Code 61
Department of Physics
NavaJ Postgraduate School
Monterey,
WH
OS
10
93943-5000
S. Gnanalingum Code 61
Department of Physics
Naval Postgraduate School
Monterey, CA 93943-5000
Professor
Maj. Han, Hwang—Jin
Postal Code 134-00
Song-Pa Gu, Ga-Rak Dong,
Seoul, Republic of Korea
GM
1
5
Ga-Rak APT
59 Dong 401
Ho
Maj. Kim, Jong—Ryul
Postal Code 500-00
Book-Gu, Du-Am Dong, 874-14
Kwang— Ju, Republic of Korea
1
Maj. Yoon, Sang—II
Postal Code 138-150
1
Gane-Dong Gu, Bang-I Dong, Sam-Ik APT 201-1103
Seoul, Republic of
Korea
50
10.
Yong—Seok
Maj. Seo,
SMC
1448,
Monterey,
11.
Maj.
SMC
Jin,
Naval Postgraduate School
93943
CA
Won-Tae
1737,
Monterey,
Naval Postgraduate School
93943
CA
12.
Maj. Ryu, Joong—Keun
SMC 1499, Naval Postgraduate School
Monterey, CA 93943
13.
Cpt. Song, Tae-Ik
SMC 2686, Naval Postgradute School
Monterey,
14.
CA
93943
Cpt. Pcu:k, Jeong—Hyun
1818, Naval Postgraduate School
SMC
Monterey,
15.
93943
Cpt. Ryu, Jong—Soo
SMC 2039, Naval Postgraduate School
Monterey,
16.
CA
CA
93943
Library, P.O. Box 2
Korea Military Accidemy
Do-Bong Gu, Gong—Neung Dong 556—21
Seoul, Republic of
Korea
51
^/7-5S^
:,002
Thesis
H1795
c.l
Han
Physical processes In
hollow cathode discharge
sources.