-
AECD-2682
UNITED STATES ATOMIC ENERGY COMMISSION
LIFE TESTS OF TUNGSTEN AND TANTALUM FILAMENTS IN ION GAUGES
by
W. E. Bush
University of CaliforniaRadiation Laboratory
Date of Manuscript:Date Declassified:
January 8, 1944August 26, 1949
Issuance of this document does not constituteauthority for
declassification of classifiedcopies of the same or similar content
and title
and by the same author.
This copy is reproduced direct from copyas submitted to this
office.
Technical Information Branch, ORE, Oak Ridge, TennesseeAEC, Oak
Ridge, Tenn., 2-14-50--850-A14019
PRINTED IN USAPRICE 10 CENTS
metadc100793
-
LIFE TESTS OF TUNGSTEN AND TANTALUM FILAMENTS IN ION GAUGES
By W. E. Bush
Purpose
This paper presents the results of an investigation carried on
for the following purposes:1. To determine the relative life of
filaments when operated at low emission (low temperatures)
as compared to those operated at high emission values.2. To
determine the nature of the dependence of life upon the initial
diameter of the filament.3. To determine the effect of varying
concentrations of water vapor and air upon the life of the
filaments.4. To determine the relative life characteristics of
tantalum as compared to tungsten.
Method
A large tank (1180 liters) equipped with a suitable vacuum
system was arranged to mount sevenion gauges. A leak was prepared
to independently admit air or water vapor or both into the
tank.
The gauges used for pressure determinations were calibrated for
air by the rate of rise method.Calibration of the gauges for water
vapor was accomplished by opening a water leak of considerable
size(after a good vacuum had been obtained) and maintaining a trap
between the gauge and the tank at thetemperature of solid CO2 (dry
ice). This gives a water pressure in the gauge which is that of
water atthe solid CO2 temperature, namely 0.50 microns. (This value
varies slightly with the barometric pres-sure, for the CO 2 point
is -78.51*C at 760 mm Hg. Tables may be consulted to find the vapor
pressureof water at temperatures corresponding to different
pressures on the Co2 ')
Gauges to be tested were put on the tank and pumped out for some
time before being operated. Then,under high vacuum, they were
turned on and allowed to out-gas for an hour or so. (This latter
procedurewas not followed in the case of tantalum, since outgassing
took but a very few minutes.)
The time during the rate of rise experiment was then counted
from the instant the leaks were opened.An occasional check of the
relative air to water vapor ratio was made by putting liquid air on
the
traps to eliminate the water. In the case of the tungsten
filaments, a coiled filament 10 cm long wasused (except in tests of
tetrode-type gauges, in which case it was 6 cm long and coiled),
while in thecases of the tantalum a hairpin filament 3.8 cm long
was used in a Distillation Products Type VG-1Agauge.
The grid voltage was approximately 130 volts and the plate
voltage was -22 volts in all cases.A total of 18 filaments was
tested, 7 of these being tungsten and the rest tantalum.
Discussion
Figure 1 presents the life history of four of the tungsten
filaments, started all at the same time andunder identical
circumstance except as noted. Numbers 1, 2, and 3 were 10 mil
diameter coiled fila-ments, while No. 4 was a 15 mil coil. Until
Nos. 1, 2 and 3 failed, and for some time afterward, theatmosphere
was held constant at a pressure equal to 1 micron of water plus 1
micron of air. At first,some difficulty was experienced in doing
this and a scattering of points resulted, as is shown in Fig.
1.However, this was repeated with gauges No. 5 and No. 6, this time
with well regulated control through-out the experiment. The results
are nearly identical to those presented on Figure 1.
AECD-2682 1
-
As time went on, it was apparent that the low emission (0.010
ma) filament was becoming consider-ably tapered in dimension, so
when the tests were finally finished the remains were carefully
savedand measured with a micrometer. The results of these
measurements are shown in Figure 2. It willbe noted that only in
the very low emission tube did tapering seriously enter into the
destruction of thefilament. This tube, operating at an emission
1/500 that of the hottest one, started at a lower wattageinput and
ran to very nearly zero, the last readings being well below one
watt.
It is well to note that during the tapering of filament No. 1,
the calibration did not materially change.From tube No. 4 on Figure
1 (which has not yet failed after some 90 hours and retains about
a
2 mil filament), it can be seen that the decay rate responds
immediately to changes in concentration ofair and water vapor.
Tubes Nos. 5 and 6 show this same phenomena, as do Figures 4, 5,
and 6. Theyalso indicate that there is nearly a threshold
concentration value (around 2 x 10-4 mm Hg) for waterbelow which
decay is very slow. This may also be true in the case of air.
Further, it seems true that this threshold decreases as the
operating temperature is lowered. Thisis again shown in all of the
decay curves for tantalum and tungsten, for in all cases excepting
one ofeighteen, the life is longer as the emission is increased.
This may be due to the fact that though thegauge pressure is more
or less the same in each case, the higher the temperature, the
lower the gaseousconcentration.
It seemed surprising that the power required for a given
emission is a linear function of the timeas long as the conditions
are constant, but all cases indicate this is true for tungsten. No
linear re-lationships were found for tantalum.
The surmise was made that in the case of tungsten (if we assume
no poisoning or activating) thepower for a given emission is
constant for a given area and a decrease in power means a decrease
inradius. To check this, the filaments of gauges Nos. 1, 2, and 3
were measured and the results com-pared with those computed for the
ratio of final to initial power. These results are here shown:
Tube Measured Diameter Calculated
1 0.0017 (from R) 0.0025 " (from R)2 0.0021 0.00228 "3 0.0047
0.00508"
This indicates that if tapering is not serious, the relation is
essentially true.The case of tantalum presented unexpected results
inasmuch as it was reported to be superior in
performance to tungsten. For this reason observations on the
first pair of tubes was inadequate, as isshown in Figure 8. Also in
this case, coiled filaments were used similar to the tungsten ones.
Failureseemed to be due to sagging and consequently hairpin
filaments were adopted for later tests. Subsequentresults, however,
indicate that the sagging was merely incidental with burnout.
The results of the several tests made may be summarized by the
following table:
Tube Atmosphere Emission Life
No. 7 (coiled fil10mil) 1air + 1 H2 0 0.5 ma 1hr. 50 min.
No. 8 (coiled fil10 mil) 1 air + 1 H2 0 5.0 ma 2 hrs. 30
min.
TetrodetubeNo. 9A (coiled fil
20mil) 1air + 1 H 2 0 0.5 ma 2 hrs. 44 min.No. 10 hairpin 10 mil
1 air + 1 H 20 0.5 ma 1 hr. 50 min.No. 11 hairpin 10 mil 1 air + 1
H 2 0 5.0 ma 2 hrs. 0.4 spin.No. 12 hairpin 10 mil 1 air + 1 H2 0
0.5 ma 2 hrs. 39 min.No. 13 hairpin 10 mil 1 H 2 0 5.0 ma 2 hrs. 55
min.No. 15 hairpin 10 mil 1 H 2 0 0.5 ma 6 hrs. 48 min.No. 16
hairpin 10 mil 1 air 5.0 ma 7 hrs. 21 min.No. 17 hairpin 10 mil 1
H2 0 0.5 ma 3 hrs. 18 min.No. 18 hairpin 10 mil 1 H2 0 5.0 ma 3
hrs. 26 min.
2 A ECD -2682
-
AECD-2682
The most noticeable characteristic of the tantalum filaments was
their irregular behavior as com-pared to that of the tungsten. At
burnout the sketches (Figure 7) illustrate typical appearance.
Thewhole filament mass seems to become sponge-like and quite
irregular in size. This has been observedin other cases of the use
of tantalum for filaments. This swelling, resulting in increased
surface, isreflected in the sometimes rather continuous increase
required in wattage. Different individual fila-ments showed
different properties so far as details were concerned, but the life
seemed to be moreor less specifically determined by the emission
and the atmosphere.
With tantalum, as in the case of tungsten, somewhat better life
results at high rather than lowtemperatures. Gauge No. 18 disproved
this rule.
During the experiments, reports from F. R. Elder of the General
Electric Company indicated thatvery long life was being obtained
from tantalum with an "atmosphere of water at 1 micron". Since
thisin no way checked the results herein obtained, an analysis was
requested of the wire being used here.Also, the G. E. Research Lab.
sent a sample of the 10 mil Ta that they were using. Gauges #17
and#18 were constructed from this wire (which also is undergoing
analysis); so far, no report has beenmade concerning these
determinations. However, since the results obtained with the G. E.
sample are
very similar to those of the previously used filaments under
similar conditions (#12 and #13), the as-sumption seems good that
both samples actually are tantalum. The wattage behavior of the two
samplesis different, but the life is essentially the same.
Conclusions
From the foregoing it would seem that the following several
conclusions are justified.1. For the concentrations of air and
water vapor tried, there is no advantage to be gained in
operating
tungsten filaments at excessively low emissions in order to
reduce the temperature. This seems likelyfor tantalum also.
2. In the case of tungsten, the wattage required for a given
emission is a linear function of the time
of operation, and the ratio of the final to the initial power is
the same as the ratio of the final to theinitial diameters of the
filaments.
3. Tapering of tungsten filaments at usual emission values is
negligible.4. The life of a tungsten filament is proportional to
the initial diameter.
5. No great changes (more than 10%) of calibration were noted
during the life of any filament ex-cepting in tantalum a few
minutes before burnout.
6. As already known, filament disintegration is in a way
proportional to the concentration of eitherair or water, the water
being more effective. This seems to be true of tantalum as well as
tungsten.
7. Tantalum filaments of size equal to tungsten filaments can be
expected to have from 5 to 10%
the life of tungsten in the same atmosphere of air and
water.
Further Remarks
Gauges Nos. 9, 9A, and 14 were tetrodes. These showed no
particular variation in characteristicsover the triode design.
However, the tests made were all at constant emission, the emission
being con-trolled by the filament current rather than the control
grid. This nullifies any possible conclusion thatmight be drawn
about them.
In the case of large filaments, a large amount of deposited
material must be allowed for. This meansadequate shielding but such
is easily furnished, the VG-1A and others being satisfactory in
this respect.
It might also be noted that low temperatures increase the
sensitivity of the gauges. Thus, for
example, the gauges when run at 0.010 ma were about 25% more
sensitive (positive ions per electron)than the same tube run at 5
ma. This is most likely due to alteration of inter-electrode gas
concentra-
tion as a result of difference in temperature.
Further work should be done to obtain more quantitative measures
of probable life in the atmospheres
described and also in chlorine and combinations of chlorine,
water, and air.
3
-
I I F I______
4 4I4I4I4 1I-r
UCRL VAG RESEARCH DEPT4-7-44 )
f0.0OH MS
7o
GO
POWER-GAUGE NO.4 RESISTANCE-GAUGE NO. 3 RESISTANCE-GAUGE
NO.4SLOPE=-4.05 WATT/ HR.
50
GAS PRESSURE
PWER-GUEN.34.0 AIR + 4.0 H2OLsOPE=-0.98 WAT T/ HR, o
40
--- POWE R-GAUGE NO. 2
SLOPE=- 4.43 WAT T/HR30 0
BURN-OUT
(DUE TO VOLTAGE SURGE)o POWER-GAUGE NO. 4
SLOPE=-0.65 WATT/ HR.20
(+0.040 AIR)
40>BURN-OUT -4.0 AIR -
(+0.046 p H20) 0.46 p .
BURN -OU T_________- -------___________ _____________
-------________ --- --_______ ____________
__________________________40
OPERATING TIME IN HOURS
Figure 1. Tungsten Filaments in air and water vapor. Power
consumption-resistance as functions of (1) emission (temperature);
(2) gas concentration; (3) origin al filament size.
Filament Specifications
1. Length: 10 cm.2. Form: 8 turn coil3. Size: Gauges 1, 2, and
3: 10 mil;
gauge 4: 15 mil
Operation
Gauge 1Gauge 2Gauge 3Gauge 4
Ep Eg
Emission 0.010 maEmission 0.50 ma
Emission 5.00 maEmission 0.50 ma
- 22V-22V-22V-22V
Concentration of Gases
Time
From To128V128V128V128V
0 hrs 00 min41 hr 40 min51 hr 00 min58 hr 20 min
41 hr 40 min51 hr 00 min58 hr 20 min81 hr 55 min
Concentration
Air Water
1.0 u0.035 yu1.0 .L0.04 j
1.0 A1.0 AL40.016 u0.16 AL
80
z
Li
0a-
44.0
10.0
-1 9.0
~-0---
H 2O0+O.02 *AIR0. tH2
I-h
(U 80 90
8.0
7.0
6.0 0
.0
z
0z
cr
4.0
3.0
2.0
4.0
0o
I I
-. I
20 30 40 50 60 70 80 90
-
AECD-2682
12.5 7 7 7 7 7 7 7 7 7
1 V.V 5
E7.5l0z
GAUGE NO. 1
55.0 GAUGE NO. 3 BURN-OUTc0
0
2.5 __ __- _0_
00
0 1 2 3 4 5 6 7CM FROM UPPER END OF 10-CM FILAMENT
8 9 10
Figure 2. Tapering of tungsten filaments. Coiled (8 turn) 10 mil
filaments at various emissions useduntil failure in an atmosphere
of 1 air + 1 y H 2 0. No. 1 = emission 0.010 ma; No. 2 =
emission0.50 ma; No. 3 = emission 5.0 ma. Note: #3 failed as result
of voltage surge.
5
-
AECD-2682
50
40
30
UCRL VAC RESEARCH DEPT12/29/43 A
20
ATMOSPHERE1. AIR + 1 H20
-A C(OO
x
5 A
x___ __V.
A
00,x
0 5 10 15 20 25 30 5TIME IN HOURS
5
Figure 3. Resistance of tungsten filaments 10 mil, 10 cm long.
G)= No. 1 emission 0.010 ma; A = No. 2emission 0.50 ma; x = No. 3
emission 5.0 ma.
0
zw0ZHCo
w
6
-
oC __ X WATER (2 /Lm) (TAKEN FROM AVERAGE AFTER 23 HR
PERIOD)
AIR (,O a)x
0 43WATE(55 67a)8
._.WATER + AIR
PRESSURES Oo 40 /L, H2O + ,0/La. AIR WATER (40 a,)0 40 pm H2O t
0.25#a. AIRX 40,u. AIR + 0.20,uaH2O0 p ,hH2 + 0.4 paAIR
TRIODE UCRL 4/5/44
TIME IN HOURS
Figure 4. Ion gauge No. 5. Tungsten filament 10 mil, 10 cm long.
Emission 0.5 ma; Eg = 130 volts;Ep = -23 volts.
W
U)'-
3
4
5
6
C)
-
AVERAGE VALUEWATER (2pa) (FOR 23 HOURS)
xlx
x x AIR
-WATER (5 a )
2WATER (10 a)
WATER + AIR
3
PRESSURES
4 0 aO H2O + O1a AIR
A 1Opa H2 0 + 0.25pa AIRX 10 a AIR + 0.20 a H20
TRIODE UCRL 1/1/44 0 1.6pa H2O + 0.4MAIR
5
60 12 3 4 5 6 07 8 9
TIME IN HOURS
Figure 5. Ion gauge No. 6. Tungsten filament 10 mil, 10 cm long.
Emission 5.0 ma; Eg = 130 volts;Ep = -23 volts.
WU),QWacw
I-4
-
SWATER 2o (TAKEN FROM 23 HR AVE)
x X xWATER (5 a)
\Q x
2
AIR (10p )
3
WATER (10p~a)
TETRODE UCRL 1/5/444
5A
2 3 4 5TIME IN HOURS
6 7 8 9
Figure 6. Ion gauge No. 14. Tungsten filament 10 mil, 6 cm long.
Emission 0.5 ma; Eg = 130 volts.
WNIQW
UH
F-Q
06
d
uI
i
-
AECD-2682
F
GE
DH
C
1 iB
A J
E
D F
C
G
B
H
IA
Figure 7. Tantalum filaments.
A - 0.0098" Dia.B - 0.0143C - 0.0072D - 0.0119E - 0.0151F -
0.0111G - 0.0135H - 0.0072I - 0.0135 nI - 0.0103
A - 0.0103" Dia.B - 0.0123 "C - 0.0052D - 0.0123E - 0.0142F -
0.0088G - 0.0097H - 0.0124I - 0.0108
Filament No.10 Filament No. 11
10
-
35
BU0RN-OUT 1 HR 50M
25BURN-OUT
2 HR 30 M
20
45
TUBE- GLENTA NO. 2 (COIL FILAMENT)UCRL VAC RESEARCH
DEPT1/T/44
10"0 2
TIME IN HOURS
Figure 8. Tantalum filaments. 10 mil, 10 cm long. E Gauge No.
7;0.5 ma; gauge No. 8 at 5.0 ma; pressure 10 ya H2 0 + 10 a
air.
A Gauge No. 8. Gauge No. 7 at1-+
N)9-3-2
0CL
az
)
OD
I 3
-
25
20
zza00 15ZZ
UzW
0W1
0
2TIME IN HOURS
3
Figure 9. Tantalum filaments. Tube VG-1A (Nc. 10 and No. 11
only) 10 mil, 3.8 cm long. o Gauge No.10, 0.5 ma emission; A Gauge
No. 11, 5.0 ma emission; x Gauge 9A, 0.5 ma emission. Note: No. 9
is atetrode designed by Frank Kirby. Filament 6 cm, 20 mil Ta.
UCRL VAC RESEARCH DEPT12/30/43
NO. 10
BURN-OUT 41HR 50 M
NO.1
NO. 9 (CURRENT) \,-BURN-OUT 2 HR 04M
BURN-OUT 2 HR 44 M
NOTE: ATMOSPHERE OF 1.O AIR + 1.O H2 0
0
0
-
21%
2
U- 1
Q-3
0.4
2TIME IN HOURS
3
Figure 10. Life tests of four tantalum filaments. Tube used:
DPI-VG-1A. Filament length = 3.8 cm;Ep = -23 volts; Eg = 130
volts.
5
0UCRL VAC RESEARCH DEPT
N0.14 12/30/43
NO.10NO. 1
NO.12
NO. 43
GAUGE SYMB. FIL. EMISSION PRESSURE LIFEAIR WATER
5 NO. 10 -O- 10 MIL Ta 0.5 10~3 MM 10 3 MM 1 HR 50M11 -A- 5.0
10~3 MM 10-3 MM 2 HR 04 M
12 -- X- 0.5 2X10 5 MM 10-3MM 2HR 39M43 -D- 5.0 2X105 MM 10 3 MM
2HR 55M
0
a~
0
i
I
-
0 NO.15 RUN AT 0.5 mae NO. 16 RUN AT 5.0 ma
UCRL VAC RESEARCH DEPT
GAUGE NO.15
GAUGE NO. 16
0 p-BURN-OUT 6 HR 48M
______________ _______________ ______________ _____________
______________ ______________ ________ ______________I
___________________ ____ ____ ____ _____ NA__0____ ____
2 4 5TIME IN HOURS
BURN-OUT
6
7 HR 21M-\
7 8 9
Figure 11. Tantalum filaments. Tube VG-1A: 10 mil Ta, 3.8 cm
long in atmosphere of 10 sa air.
25
20
j~15
z
0o10
5
O0
i i
-
-NO.18 BURN-OUT (3 HR 18 M)
25
A
20NO.18
BURN-OUT (3 HR 26 M)
P NO. 17
15
O GAUGE NO.17
A GAUGE NO.18
TANTALUM FURNISHED BY GENERAL10 ELECTRIC CO.
UCRL VAC RESEARCH DEPT1/7/44
O 1 2 3 4 5 6TIME IN HOURS
Figure 12. Tantalum filaments. Tube VG-1A: 10 mil, 3.8 cm long.
Gauge No. 17 at 0.5 ma. GaugeNo. 18 at 5.0 ma. Pressure: 10 ga H 2
0 + 0.4 a air.
N-
H~
z~
a
CD
C,'