General Disclaimer
One or more of the Following Statements may affect this Document
This document has been reproduced from the best copy furnished by the
organizational source. It is being released in the interest of making available as
much information as possible.
This document may contain data, which exceeds the sheet parameters. It was
furnished in this condition by the organizational source and is the best copy
available.
This document may contain tone-on-tone or color graphs, charts and/or pictures,
which have been reproduced in black and white.
This document is paginated as submitted by the original source.
Portions of this document are not fully legible due to the historical nature of some
of the material. However, it is the best reproduction available from the original
submission.
Produced by the NASA Center for Aerospace Information (CASI)
-A
r
NASA Technical Memorandum 87087
f'
r
Electrode Erosion in Arc Discharges
at Atmospheric Pressure
^y
k
y
(NASA--M-87J87) ELECTRODE EFOSIUN IN XEC
N85-34176
DISCHIRGLS AT ATMOSPHERIC PFRssurF ;NASF)
21 P HC
A02
/MF
1.1
C3CL 21;.
Unclas
,13/20 22224
Terry L. Hardy
Lewis Research Cenier
*.
Cleveland, Ohio
z
Prepared for the
18th International Electric Propulsion Conference
cosponsored by the AIAA, BASS, and DGLR
Alexandria, Virginia, September 30-October 2, 1985
7
J
4
ELECTRODE EROSION IN ARC DISCHARGES AT ATMOSPHERIC PRESSURE
Terry L. Hardy
National A.ronautics and Space Administration
Lewis Research Center
Cleveland, Ohio 44135
Abstract
An experimental investigation was performed
in an effort to measure and increase lifetime of
electrodes in an arcjet thruster. The electrode
erosion of various anode and cathode materials was
measured after tests in an atmospheric pressure
nitrogen arc discharge at powers less than 1 kW.
A free-burning arc configuration and a constricted
arc configuration were used to test the materials.
Lanthanum i.s^;aburide and thoriated tungsten
had low cathode erosion rates while thoriated
tungsten and pure tungsten had the lowest anode
erosion rates of the materials tested. Anode
coolinq, reverse gas flow, and external magnetic
fields were all found to reduce electrode mass
loss.
Introduction
The direct current arcjet thruster is pre-
sently being considered as an auxiliary propulsion
device for space applications because of its wide
thrust range, high specific impulse capability
(400 to 1000 sec), and inherent simplicity. Due
to available power and spacecraft interaction
considerations near-term applications appear to
be at power levels less than 5 kW with ammonia or
hydrazine as the propellant. Although arcjets had
been investigated in the 1960's. the research was
primarily concerned witn power levels onthe order
of 30 kW with hydrogen as the propellant.
)
At
30 kW, two tests were run with hydrogen for over
500 hr at approximately 1000 sec specific impulse
with thrust efficiencies greater than 40
percent.+
3
At 2 kW, a test was run in hydrogen
for 150 hr;
4
however, tests performed at power
levels of 1 kW experienced severe electrode ero-
sion, especially with gases other than hydrogen.5
Therefore, a study of electrode erosion at low
power levels was required.
Because increased energy efficiencies can be
obtained at higher pressure, it is desired to
operate
arc,,
p
ts
at 1 atm or higher.
Much of
the research on electrode erosion in arc di-
charges has been under vacuum conditions.]]
Because the electron emission and electrode ero-
sion mechanisms are different in vacuum and
atmospheric pressure arcs, the erp$ion rates are
probauly also different. Kimblin
10
has done a
study on cathode erosion in the transition from
vacuum to atmospheric pressure, but copper and
graphite were the only material$ considered in
this study. Cobine and Burger
present data
on anode erosion for various materials in high
pressure arcs, but the current level was greater
than 11 000 A, which is at least two orders of
magnitude greater than the current levels being
considered in the low puwer arcjets.
Therefore, as a f?ntinuation of research by
Hardy and Nakanishi,
the present study
addresses the problem of achieving long electrode
life (>1000 hr) in low powe
r
arcjets by measuring
the electrode erosion in a free-burning arc dis-
charge, as well as the erosion in a constricted
arc similar to previous arcjet designs. Tests
have been conducted to characterize the electrode
erosion of various materials in nitrogen at 1 atm
at operating powers of 1 kW or less. Also, in an
effort to reduce erosion the qas flow direction,
electrode cooling, and the number of starts have
been examined. Preliminary results of the mass
loss of the electrodes in an external magnetic
field, used to reduce erosion, are also presented
Apparatus
The experimental apparatus for the evaluation
of electrode erosion is shown in Fig. 1. The test
chamber was constructed of pyrex and a roughing
pump was used to obtain pressures as low as
0.3 torr. Gas could be fed into the chamber
through an external port to increase the pressure.
A current-controlled, voltage-regulated do power
supply with a 600 V, 25 A capacity was used with
a 0.5 mH, 150 A inductance coil and a variable
resistor, normally set at 5 ohm, completing the
circuit. Direct current meters were used to read
voltages and currents. Due to a power supply
malfunction, the 600 V. 25 A power supply was
replaced with a 100 A, 100 V power supply for some
tests. This supply used an additional 500 V power
supply to ignite the arc, and a series-pass linear
transistorized current regulator was used in place
of the resistors and inductance coil. All gases
used were 99.995 percent pure.
Free-Burning Arc
Figure 2 is a schematic diagram of the free-
burning or unconstricted arc configuration. This
configuration had no confinement of the gas flow.
The electrodes were held in place wi
t
h brass sup-
port pieces fabricated to screw into 1.21 cm o.d.,
19.1 cm long copper tubes. A ceramic tube
1.91 cm o.d., 19 cm long was placed over the
cathode-side copper tube to prevent arc formation
on the copper. Thr anode-side copper tube was
fabricated to alto+ .ter cooling if desired, and
gas could either be fed through the cathode (con-
ventional flow) or over the anode (reverse flow).
Boron nitride covers were fabricated for both the
anode and cathode to prevent arc formation on the
brass holder. A 5.08 cm diameter quartz tube was
placed over the electrodes to prevent eroded
material from coating the inside of the pyrex test
chamber.
Cathodes
Five different cathode materials were inves-
tigated using this free-burning arc apparatus:
1.
Lanthanum hexaboride tube: Becausp4of
proniisiig results rom previous tests,llss,
a
lanthanum hexaboride (LaB6) tube, 0.53 cm o.d.,
0.30 cm i.d. was used. Lanthanum haxaboride has
a melting point of 2210
*
C. The tube was fabri-
cated by a hot-pressing sintering technique and
had a typical density of 80 percent. Because
lanthanum hexaboride was found to develop thermal
stress cracks upon heating, a supporting tube of
molybdenum was placed over the LaB6 tube with
tantalum foil placed between the molybdenum and
the LaB6 to insure a tight fit and good
electrical contact.
2.
Hafnium Carbide tube: Due to its high
melting point 3890 C , hafnium carbide (HfC) has
been sugggstgd for use as cathode material in
arcjets.
Il
b
A 0.53 cm o.d., 0.30 cm i.d.
tube, hot-pressed sintered from powder, had a
typical density of 75 percent. Initial tests
showed that this material also developed thermal
stress cracks, requiring a molybdenum outer tube
for support similar to the LaB6 configuration.
3.
Two Percent Thoriated Tun sten tube: Pre-
vious experiments indicated tat thoriate tung-
sten (Th02-W), when operated in an oxygen-free
environment, may have a low cathode erosion
rate.
?
Thoriated tungsten has a melting point
of 3410
+
C and the tube used was 0.64 cm o.d.,
0.53 cm i.d., with a density of 97 percent.
4.
Graphite Rod: Graphite with a density of
87 percent was used in a rod configuration,
0.32 cm diameter, ground flat on three sides to
allow gas flow. Graphite sublimes at 3652 'C.
5.
Copper Rod: To test the reliability of the
cathode erosion measurements, a copper rod, melt-
ing point 1083
*
C, was used allowing a comparison
with data found in the literature. This copper
rod was 0.64 cm diameter, ground flat on three
sides to allow gas flow.
Anodes
Six different anode configurations were used
in this free-burning arc configuration:
Water-cooled copper rod, 0.64 cm
diameter.
2. Two percent Thoriated tungsten rod, 0.64
cm diameter. This configuration was used both
with and without water cooling.
Pure tungsten rod, 0.64 cm diameter
4.
Hafnium carbide tube, 0.53 cm o.d.,
0.30 cm id.
5.
Barium Oxide Impregnated Porous Tungsten
tube, 0.53 cm o.d., 0.30 cm i.d.
6.
Two percent thoriated tungsten insert,
O.b4 cm diameter, 0.10 cm orifice. This insert
was fabricated in the shape of a converging-
diverging nozzle.
Magnetic Field
Figure 3 is a schematic diagram of the fixed
magnet configuration. A 0.32 cm diameter thori-
ated tungsten rod was used as the cathode and a
0.64 cm diameter thoriated tungsten insert was
used as the anode. This an;)de typically had a
0.10 cm orifice and a 45' nozzle angle. Four
rectangular 35 percent Samarium, 65 percent Cobalt
magnets, 3500 G at the surface, 1.75 by 1.25 by
0.50 cm, were placed on a steel support to provide
the magnetic field. The magnets provided an axial
magnetic field of 100 G at the electrodes. A
gaussmeter was used to measure the magnetic field
strength both before and after the tests. Two
boron nitride cylinders were used to isolate the
steel support from the copper tubes supporting the
electrodes. This configuration was also used
without the magnets for comparison purposes.
Constricted Arc
In addition to the free-burning arc configu-
ration, a constricted arc configuration with con-
fined gas flow was constructed to measure erosion
in configurations similar to previous arcje,
designs. Figures 4 and 5 show the configurati ns
used. Figure 4 is the conventional configuration
where the water-cooled copper constrictor acts as
the anode with gas flow from cathode to anode.
Two anodes were used in this configuration. One
anode was a 1 cm didmeter copper disk, 0.96 cm
thick, with a 0.10 cm orifice to allow flow
through the constrictor. The other anode was a
0.64 cm diameter thoriated tungsten insert fabri-
cated in a nozzle configuration. This insert had
an orifice 0.10 cm diameter. The thoriated tung-
sten was pressed into a copper uisk 1.00 cm o.d.,
0.64 cm i.d., 0.96 cm thick. Both copper disks
were fabricaued to screw into a brass fitting,
which was soldered to the inner copper tube. This
copper tube was 1.58 cm i.d. and 0.96 cm long, and
was soldered to an outer copper tube 2.70 cm i.d.,
7.62 cm long. Copper disks were soldered to the
copper tubes to contain the cooling water. The
cathodes used in this configuration were a lan-
thanum hexaboride tube, a thoriated tungsten tube,
and a 0.32 cm diameter thoriated tungsten rod,
ground flat on three sides to allow gas flow over
the cathode. A 2.54 cm o.d. quartz tube was
placed inside the outer copper tube to restrict
the gas flow. This quartz tube extended from the
copper tube to the flange on the pyrex test
chamber.
Figure 5 shows the modification to the appa-
ratus shown in Fig. 4. This modification allowed
the cathode to be placed in the position of the
anode and the gas flow to be from anode to cathode
in a reverse flow pattern. In this configuration,
a 0.76 cm long LaB6 tube was placed in a 0.96 cm
long molybdenum tube. The LaB^ tube acted as
the cathode and was press fit into a copper disk,
which was fabricated to f
i
t into the brass fitting
as described previously. Similarly, a ThO2-W
tube was used as a cathode. On the downstream end
of the cathode, a thoriated tungsten nozzle with
a 0.10 cm orifice was pressed into copper disk and
fit inside the brass support piece to constrict
the flow. The anode was a 0.64 cm diameter thor-
iated tungsten rod ground flat just enough
(0.02 cm) on three sides to allow gas flow. One
boron nitride cover fit against the quartz tube
and under the outside copper tubing on the con-
strictor to prevent gas leakage. Another boron
nitride cover was placed over the anode to prevent
possible arc formation on the brass supporting the
anode, and a third boron nitride disk placed over
the cathode tube prevented arcing to the copper
constrictor.
Procedure
Arc Characteristics
The arc V-I characteristics were obtained with
both argon and nitrogen for various gaps, or dis-
tance between the electrodes, for a current range
of 5 to 15 A. The arc was started with a voltage
of 400 to 500 V with argon at 0.5 to 10 torr
background pressure at flow rates 100 to 140 SCCM.
After initiation of the arc, the pressure was
raised to 760 torr in argon and the arc was
allowed to run for 5 min before readings were
taken. The first voltage measurement was then
taken at 5 A for a gap width of 0.3 cm. Then the
current was raised by 1 A, and after 2 or 3 min
another measurement ,,as taken. This procedure was
repeated for the full range of currents, after
which the gap was increased to 1 cm and the volt-
age measurements were taken as described for a gap
of 0.3 cm. After the measurements were taken in
argon, the chamber was pumped down to approxi-
mately 50 torr, nitrogen was allowed to flow into
the chamber from the cathode, and then nitrogen
was bled into the chamber from an external port
to raise the chamber pressure to 760 torr. The
pressure was maintained by drawing a vacuum to
offset the flow. The procedure was then repeated
as described or argon.
Mass In
—
, TPSts
Measurements were made of the electrode mass
loss of various electrode materials tested in high
purity nitrogen. Tests were also performed which
compared the effects of anode cooling, gas flow
direction, and the number of starts on electrode
erosion. The mass loss rates in free-burning arc,
constricted arc, and magnetic field configurations
were compared.
Due to difficulties in starting an arc in
nitrogen, ar
g
on was used in all tests to initiate
the arc. As in the arc characteristic tests, the
arc was started with a 400 to 500 V input at 0.5
to 10 torr in argon at 10 A and 0.3 cm gap with
flows of 140 SCCM. Once the arc was stable at
these pressures, the pressure was raised to
760 torr by back-filling argon into the chamber.
After 5 min the argon was shut off, the chamber
was pumped down to 50 torr, nitrogen was allowed
to flow through the cathode tube at 100 SCCM, and
nitrogen was bled into the tank to raise the
chamber pressure to 760 torr. This was repeated
to insure that the argon was pumped out of the
test chamber. In the constricted arc configura-
tion tests, some experiments were performed with
a gas flow of 4000 SCCM. In these high flow tests
a constrictor chambe pressure of 600 torr was
achieved with a background presssure of 10 to
15 torr.
After arc initiation, most tests were run for
2.5 hr, after which the test was stopped and the
chamber was allowed to cool 30 to 45 min. The
starting and running procedures were then repeated
to ensure constant arc conditions. The total run
time per experiment was approximately 5 hr. In
the tests comparing the number of starts, the arc
was initiated 4 and 8 times in a 5 hr period, as
opposed to two starts in most other experiments.
After the run was completed, the chamber was
allowed to cool for at least 1 hr after which the
electrodes were removed and weighed on a scale
with an accuracy of 0.001 g. The entire test
procedure was then repeated with new electrodes
and an average mast loss for the two tests was
reported. The results are presented as a mass
loss per unit time, or a mass loss rate, with
error bars representing the span of the electr-ide
erosion.
Results and Discussion
Free-burning Arc
Arc characteristics. Figure 6(a) shows the
voltage-currEn c arac eristi:s of a hafnium
carbide cathode at different gap widths while
Fig. 6(b) show, the V-I characteristics for a
lanthanum hexaboride cathode. As was shown in
previous tests of a thoriated tungsten cathode,12
the arc voltage in nitrogen here was 20 to 45 V
higher than the arc voltage in argon, a difference
attributed to a difference in the plasma proper-
ties of the two gases. Also, the voltage
decreased as the current was incrr^sed. As dis-
cussed by Finkelnburg and Maecker , as the
current increases the conductivity of the gas
increases, thus decreasing the electrical impe-
dance and hence the voltage. Finally, as the
distance between the electrodes increased, the arc
voltage increased. This increase in arc voltage
may have been due to an increase in the potential
fall across the column, or possibly due to an
increase in the anode voltage drop. This effect
is important as a change in arc voltage may
reflect electrode ablation. It is to be noted
that, due to buoyancy forces at gaps larger than
0.5 cm, the actual arc length may have been longer
than the measureu gap between the e)^ctrodes.
This effect was reported by Cobine.11
A separate arc characteristic test was per-
formed using graphite electrodes in N2. The
results obtained were compared with values found
in the literature, es shown in Fig. 7. Because
the V-I characteristics agree within 3 V of those
presented by Cobine,
19
it is assumed that the
arc characteristic results obtained for other
materials reported herein are valid.
Cathode mass loss. The cathode mass loss for
various materials in a nitrogen free-burning arc
is given in Fie. 8. Water-cooled copper was
chosen as the anode to prevent eroded anode
material from being transferred to *he cathode,
thus affecting the results. The mac
s
icss
rate
of copper was 3.4x10
-3
g/min. This re^,-,!t agreed
closely Wjth the results of Semiciiova and
Petrova,
ZO
who found the mass loss of copper in
nitrogen at 9 A to be 3.6xl0
-3
9/min. Graphite
had a mass loss of 8.7x10
-
g/min. '^fnium car-
bide had a mass loss rate of 6.6x10 g/min; how-
ever, after the 5 hr run, the cathode was examined