TungstenRhenium Thermocouples for Use at High Temperatures
R. R. Asamoto and P. E. Novak
Citation: Rev. Sci. Instrum. 38, 1047 (1967); doi: 10.1063/1.1720964
View online: http://dx.doi.org/10.1063/1.1720964
View Table of Contents: http://rsi.aip.org/resource/1/RSINAK/v38/i8
Published by the American Institute of Physics.
Additional information on Rev. Sci. Instrum.
Journal Homepage: http://rsi.aip.org
Journal Information: http://rsi.aip.org/about/about_the_journal
Top downloads: http://rsi.aip.org/features/most_downloaded
Information for Authors: http://rsi.aip.org/authors
Downloaded 07 Dec 2011 to 192.107.52.30. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
THE REVIEW OF SCIE"-'TIFIC INSTRUMENTS
VOLUME 38, NUMBER 8
ACGlJST 1%7
Tungsten-Rhenium Thermocouples for Use at High Temperatures*
R, R.
ASAMOTO AND
P. E.
NOVAK
Advanced Products Operation, General Electric Company, SunnY'l'aie, California 94086
(Received 13 February 1967; and in final form, 17 April 1967)
Tungsten-rhenium thermocouple systems were evaluated for use in measuring temperatures between 1600 and
3000°C. Temperature-millivolt relationships were extended from 2300 to 3000°C for bare-wire \V3Re/W25Re and
W5Re/W26Re thermocouples in vacuum. The performance of W26Re sheathed, high-fired beryllia-insulated
thermocouples was limited only by the melting point of the beryllia insulation C-v2550°C). The thermoelectric
output of high-fired thoria-insulated thermocouples sheathed in W26Re was determined and found to be reliable
up to at least 2850°C, when corrected for electrical shunting through the insulation. Stability of high temperature
thermocouples using thoria insulation was within ±0.3 mV (±40°C) when held at 242SoC for 148 h.
INTRODUCTION
PRIOR WORK
increasing need for reliable high temperature
sensors is being realized in the research and development programs of both the aerospace and nuclear
industries. Thermocouples are one of the most widely
used temperature sensors because of their relative simplicity of construction (not to be confused with insensitivity to manufacturing variables), capability for remote
operation, and general availability.
It is desirable to be able to measure ceramic fuel temperatures under both steady state and transient operating
conditions in the Southwest Experimental Fast Oxide
Reactor (SEFOR). The data from these measurements
will aid in the correlation and evaluation of reactor
physics parameters. Target requirements of the fuel, and
hence the temperature sensors, include a 1500 h lifetime
under cyclic operating conditions up to 26oo°C in contact
with (U,PU)02 fuel, and a maximum neutron dose of
l.4X 102° nvt (fast). Based upon a survey of available
sensing devices, it was concluded that tungsten-rhenium
alloy thermocouples insulated with beryllia or thoria have
the best prospect of meeting the requirements for SEFOR
fuel temperature measurement. 1 To aid in the determination of adequate thermocouple systems for use in SEFOR,
an out-of-pile and in-pile development and testing program was undertaken. The results obtained from the
out-of-pile phase of the program are presented here
and include: (1) extension of the calibration curve of
bare wire W3Re/W25Re and W5Re/W26Re thermocouples to 3000°C in vacuum, (2) evaluation of berylliaand thoria-insulated thermocouple systems up to 2550
and 2850°C, respectively, and (3) investigation of the
thermal stability of a thoria-insulated thermocouple system for 148 h at 2425°C.
Tungsten-rhenium alloy thermocouples are the most
widely used thermocouples for temperature measurements
above 20OO°C, because of their excellent thermoelectric
characteristics and high melting point (>3OO0°C). The
temperature-millivolt relationships for these thermocouples have been established by the manufacturers to
2320°C for tungsten vs tungsten-26% rhenium (W/W26Re)
and tungsten-5% rhenium vs tungsten-26% rhenium
(W5Re/W26Re),2 and to 24oo°C for the tungsten-3%
rhenium (alkaline-earth doped) vs tungsten-25% rhenium
(W3Re/W25Re).3,4 Some data have been reported in the
literature to over 30OO°C for W/W26Re 5 ,6 and to 2800°C
for W3Re/W25Re.1 Tungsten-rhenium thermocouple stability has been demonstrated in the temperature range of
1200 to 1450°C for 1000 h,8 1400 to 1900°C for several
hundred hours,9 and 2320°C for 240 h.lD Only the W3Re/
W25Re and W5Re/W26Re thermocouples were tested in
the present investigation because of the improved initial
low temperature ductility of the low-rhenium content
legs compared to a pure tungsten thermoelement. Use of
the W5Re alloy, however, has been reported to offer no
ductility advantage when utilized above 2200°C. 5
AN
2 Hoskins' calibration curves for W /W26Re and W5Re/W26Re,
adopted Nov. 23, 1962.
3 Englehard's
calibration curve for W3Re/W2SRe, adopted
Nov. 1, 1965.
4 Hoskins' calibration curve for W3Re/W25Re, adopted Nov. 23,
1965.
6 W. C. Kuhlman, "Research and Evaluation of Materials for
Thermocouple Application Suitable for Temperature Measurements
up to 4500°C on the Surface of Glide Re-Entry Vehicles", ADDTDR-63-233, (May 1963).
6 B. F. Hall, Jr., and N. F. Spooner, "Study of High Temperature
Thermocouples", AFCRL-6S-251 (23 March 1965).
7 "Reactor Development Program Progress Report", Al'."L-7046
(May 1965).
8 J. W. Hendricks and D. L. McElroy, "High Temperature HighVacuum Thermocouple Drift Tests", ORNL-TM-883 (August 1964).
9 S. K. Danishevskii, S. I. Ipatove, E. r. Pavlova, and N. 1.
Smirnova, "Tungsten-Rhenium Alloy Thermocouples for Temperature Measurements up to 2500°C", Dounreay trans!. 109, from
Zavod. Lab. 29, 1139 (1963).
10 R. B. Clark, "Calibration and Stability of W /WRe Thermocouples to 2760°C (5000°F)", Preprint, 19th Annual ISA ConI. and
Exhibit, New York, 12-15 October 1964.
* Work performed by General Electric for the Southwest Atomic
Energy Associates, under contract AT(04-3)-540 with the U. S.
Atomic Energy Commission.
1 R. R. Asamoto and P. E. Novak, "A Survey for a.High Temperature Sensor for SEFOR", GEAP-4903 (July 1965).
1047
Downloaded 07 Dec 2011 to 192.107.52.30. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
1048
R.
R.
ASAMOTO
The prime purpose of a ceramic insulator in a thermocouple system is to maintain separation between the
thermoelements. Reliable thermocouple performance requires that consideration be given to the properties of the
insulator in the temperature range of use and to the
environment in which it is to be exposed.
Beryllia is the most widely used thermocouple insulator
for high-temperature work at present, but its melting
point (2SS0°C±2S0C) limits its usefulness at higher temperatures. Although molten beryllia no longer has the
capability of keeping the thermoelements separated, a
beryllia-insulated thermocouple has been observed to
operate satisfactorily for a short period in this condition.ll Observations have been reported concerning the
compatibility between beryllia and tungsten-rhenium
alloys. One observer reports reaction between molten
beryllia and W26Re,12 while another states that W IW26Re
is compatible up to at least 2400°C. 13
The use of thoria as an insulator appears attractive
because of its high melting point (3300°C± 100°C) and
its relatively good compatibility with most of the refractory metals.u Thoria, however, has been largely
ignored as an insulator material because of reported
discrepancies in its high temperature electrical resistance
properties. 5 ,1O,15,16 There is also concern of self-heating and
degradation of thoria in a nuclear environment. 17 ,18
A sheath or protection tubing is generally incorporated
around the insulator and thermoelements to protect these
components from the environment. Tungsten-26% rhenium tubing has been used primarily in this investigation
because of its high melting point (3120°C) and improved
ductility properties over pure tungsten.
EXPERIMENTAL PROCEDURE
Thermocouple calibrations were made in an inductivelyheated vacuum furnace, Fig. l,l9 An axial hole was bored
11 Private communication with R, A. Pustell of the G. E. Co.,
West Lynn, Mass.
12 E. J. Brooks, W. C. Kramer, and R. D. McGowan, "High Temperature Sensors for Borax-V Boiling Fuel Rods", ANL-6636
(October 1963).
13 J. A. McGurty and W. C. Kuhlman, "Tungsten/TungstenRhenium Thermocouple Research and Development", presented at
SAE National Aeronautic Meeting and Production Engineering
Forum, New York (6 April 1962).
14 R. J. Runck, "Oxides," in High Temperature Technology, 1. E.
Campbell, Ed., (Electrochemical Society, New York, 1956), p.
29.
16 W. R. Prince and W. L, Sibbit, "High-Temperature Reactor
Core Thermocouple Experiments," Los Alamos Scientific Laboratory
of the Univ, of Calif., LA-3336-MS (July 1965).
16 E. J. Brooks and W. C. Kramer, "Tungsten-Rhenium Alloy
Thermocouples and Their Use in a U0 2-Fueled Reactor," ANL-6981
(November 1965).
i7 L. Fisher, J. Pendleton, J. O. Pounder, and A. B. Washington,
Xucl. Eng. 11,600 (1966).
18 L. N. Grossman, "Electrical and Spacers for Nuclear Thermionic
Devices," GEST-2022 (February 1964).
19 P. E. Novak and R. R. Asamoto, "An Out-of-Pile Evaluation
of W-Re Thermocouple Systems for Use to ·1700°F in PU02-U0 2,"
GEAP-5166 (June 1966).
AND
P.
E.
NOVAK
PRIS'.'
QUARTZ
WINDOW
lilDUCTOR COIL
5~ACK
BODY HOLE
CONCENTRATOR
ALUMINA
INSULATED
THERMOCOUPLE
COPPER
COOurjG BLOCK
FIG. 1. High-temperature thermocouple calibration furnace.
through the approximately 2S~mm long tungsten susceptor
to accommodate the thermocouples. The hole size was
reduced from 3.2 to 0.64 mm approximately 90% of the
way through the susceptor, to serve as a black body
sighting hole for optical temperature measurements. The
LID ratio of this cavity was greater than S. The test
thermocouple was inserted into the susceptor so that the
thermocouple junction became the bottom of the black
body hole. Corrections for the optical pyrometer, the
quartz sighting window, and the prism were made by
calibrating against a National Bureau of Standards calibration tungsten ribbon lamp. To prevent tungsten
vaporization on the quartz window, a swinging metal
shield, magnetically operated, was used to cover the
window during nonmeasurement periods. The corrected
temperature values observed from the optical pyrometer
were correlated with the corresponding emf output generated from the thermocouples. A vacuum of approximately 10- 5 Torr was maintained during all tests.
During the calibration of bare-wire thermocouples,
great care was taken to insure that the self-supporting
thermoelements were not touching each other or the sides
of the susceptor. High purity alumina insulation provided
support for the wires below the tungsten susceptor. The
thermocouple wires were extended directly from the
furnace to a potentiometer, without the use of compensating lead wire.
Thoria insulator shunting tests were performed in vac-
Downloaded 07 Dec 2011 to 192.107.52.30. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
TUNGSTE~-RHENIUM
uum in a tungsten resistance furnace. 2o The temperature
characteristics of a 76 mm long hot zone were predetermined with calibrated, bare-wire W5Re/W26Re thermocouples. The thermoelectric outputs of these thermocouples were related to optical pyrometer readings on a
tungsten plate at the bottom of the furnace. The thermoelectric outputs of the test thermocouples were then compared to true temperatures as determined by the optical
pyrometer.
450
400
.
.
--(
../.
RESULTS AND DISCUSSION
1. Bare Wire Calibration
350
The successful utilization of a thermocouple system
operating with electrical shunting at high temperatures
requires an initial knowledge of the thermoelectric characteristics of the nonshunting thermocouple. Because of
the limited data presently available for tungsten-rhenium
thermocouples at temperatures above 2300--2400°C, additional values were measured to extend the calibration
curves for W3Re/W25Re and W5Re/W26Re thermocouples to 3000°C. The data obtained showed good agreement within the range of the manufacturers' published
1049
THERMOCOUPLES
,
I
I','
~.l
.1'
LEGEND
RUN:11
• RUN '2
• RUN '3
x RUN '4
- - MANUFACTURER'S DATA
THIS 1!ORK
o
300
I,"
I
250
1500
1000
2500
3000
TEMPERATURE (oC)
FIG.
values, Figs. 2 and 3. In addition, the values from ANV
for W3Re/W25Re thermocouples are in good agreement
with the results obtained in this investigation.
With reliable calibration data available, the calibration
curves for the tungsten-rhenium thermocouples were defined mathematically for practical utilization. To fully
describe the emf-temperature relationship for both the
W3Re/W25Re and W5Re/W26Re thermocouples, the
manufacturers' data from 0 to 1600°C were integrated
with the observed test data from 1600 to 3000°C. The
resulting sixth-order polynomial equations, derived by
the least squares technique are
400
LEGEND
o RUN "I
• RUN '1
... RUN ~3
x RUN 04
o ANLDATA
- - MANUFACTUREn DATA
THIS WORK
350
>--'
'"".
::;
-'
~
~
~
3. Calihration of W5Re/W26Re thermocouples in vacuum.
W3Re/W25Re
300
/
/
F= 0.176+ 7. 773X 1O-3T+2.565X 10- 5 ]'2- 2.433X lO-sP
I
+ 1.162 X lo-lITc 2.88X lo-15 P+2.837X lo- l9 m,
W5Re/W26Re
15.0
1500
1000
TEMPERATURE
1500
3000
1°c,
FIG. 2. Calibration of W3Re/W25Re thermocouples in vacuum.
20
Brew furnace model no. 424C.
E=0.293+9.015X lCr3T+2. 701X lo-5]'2-3.158X lo- BT3
+ 1. 788X lo- lI P- 5.049X lo-15 P+ 2.837X 1O-19T6,
where
T=Temperature (OC),
and
E=emf (mV).
Downloaded 07 Dec 2011 to 192.107.52.30. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
1050
R.
R.
ASAMOTO AND
TABLE I. Accuracy of calibration equations.
W3Re/W25Re
W5Re/W26Re
±1%
±1%
±2%
±2%
Table I lists the accuracy of these equations, relative to
the experimental data in various temperature ranges.
Values derived from these equations between 1600-3000°C
are tabulated in Table II and are compared with the
manufacturers' data within their respective limits.
TABLE II. Measured emf-temperature relationship W3Re/W25Re
and W5Re/W26Re therniocouples-Ref. O°C.
Temperature
1600 0 C
1700
1800
1900
2000
2100
2200
2300
2400
2500
2600
2700
2800
2900
3000
This work
Mfg. values'
This work
Mfg. values'
W3Re/W25Re W3Re/W25Re W5Re/W26Re W5Re/W26Re
29.277 mV
30.921
32.504
34.022
35.468
36.835
38.116
39.303
40.391
41.376
42.259
43.044
43.746
44.386
44.991
29.448 mV
31.130
32.745
34.287
35.751
37.108
38.338
39.395
40.253
28.052 mV
29.551
30.995
32.371
33.669
34.875
35.977
36.969
37.847
38.619
39.303
39.936
40.570
41.286
42.181
28.078 mV
29.528
30.922
32.298
33.632
34.915
36.089
36.929
2. Beryllia-Insulated Thermocouples
Four commercially-obtained, beryllia-insulated (highfired), W5Re/W26Re thermocouples, sheathed in 3.2 mm
o.d. W26Re tubing, were immersed 25 mm in the inductively heated vacuum furnace, and cycled twice to 1900°C.
The outputs agreed to within ± 1% of the manufacturer's
calibration data. The stability of beryllia-insulated thermocouples below the melting point of beryllia has also been
demonstrated in other investigations. 5 •1o
A W3Re/W25Re (0.25 mm wire diam) thermocouple
with a 3.2 mm o.d. W26Re sheath, and insulated with
99.9% pure high-fired beryllia was tested to above the
melting point of beryllia. Because of the possible detrimental effect impurities could cause at high temperatures,
great care was taken in the fabrication of the thermocouple. The W26Re sheath was electropolished in a 2%
~aOH solution, and together with the thermoelements
ultrasonically cleaned in acetone. Complete assembly
of the thermocouple system was performed in a glove
box under an argon atmosphere C",v100 ppm O2). A
tungsten end plug was welded onto one end of the sheath.
The insulator and junctioned thermoelements were placed
into the sheath and the system was closed by a vacuum
sealant at the low temperature end.
The thermocouple was immersed 76 mm in the hot zone
P.
E.
NOVAK
of the furnace and showed good stability and agreement
with the bare-wire calibration data up to 2550°C. A
sudden decrease in emf at this point was observed. On
disassembly of the system, the beryllia was observed to
have melted. The thermoelements had maintained their
integrity, but showed a change in appearance, i.e., one
leg had a grainy structure while the other appeared
polished. Identity of the thermoelements was not determined. The beryllia appeared to have melted and adhered
to the thermocouple wires.
It is concluded that the most serious problem encountered after the beryllia melts is that the thermoelements are no longer supported and can create secondary
junctions between themselves and/or the sheath. This, in
turn, causes erroneous thermoelectric outputs that are
not necessarily corrected upon resolidification of the
beryllia (the thermoelements may remain in the displaced position). Pustell has observed that a W /W26Re
thermocouple system exposed for 20 min in molten beryllia
showed no deviation from the standard calibration curve.l1
This indicates that molten beryllia maintains its insulating
properties and that the thermocouple wires, in his test,
did not reposition. This also suggests either compatibility
between the molten beryllia and the W-Re alloys, or that
reactions in the above time period do not cause error in
the thermocouple output. Deleterious effects due to the
hexagonal-to-cubic phase inversion of beryllia at 2050°C21,22
have not been detected. Although some cracking and
spalling may occur, the beryllia retains integrity for wireto-wire separation. It is thus apparent that beryllia-insu. lated, tungsten-rhenium theromcouples operate reliably
(no shunting effects) up to the melting point of the beryllia
insulation.
3. Thoria-Insulated Thermocouples
The high melting point of thoria makes it an attractive
insulator material. However, uncertainties in its high
temperature electrical properties necessitated further evaluation. Two 3.2 mm o.d. W26Re sheathed, W3Re/W25Re
(0.25 mm diam wire) thermocouples were fabricated,
utilizing 99 and 99.9% pure vitrified thoria insulation.
The bore diameter of the two bore insulator beads was
0.64 mm. Fabrication procedures were similar to that
described for the beryllia-insulated thermocouple. In
addition, the thoria insulation was outgassed for 2 h at
1700°C in a hydrogen furnace prior to its use. The thermocouples were tested to 2850°C in the tungsten resistance
furnace at immersion depths of 25 and 76 mm. Results
of the tests are illustrated in Fig. 4. The data show that,
for the thermocouple system tested, performance was
D. K. Smith and C. F. Cline, J. Nucl. Mater. 6, (3), 265 (1962).
S. B. Austerman, "Decrepitation of Beryllium Oxide At High
Temperature," NAA-SR-6428 (September 1961).
21
22
Downloaded 07 Dec 2011 to 192.107.52.30. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
TU;-;GSTEN-RHENIUM
reproducible and insensitive to the thoria purities and
immersion depths tested. Shunting effects appeared to
start at 2300°C and continued up to 8% at 2850°C. The
thermocouples were cycled from 1300 to 2850°C two to
three times and exhibited good stability in all cases. A
plot of emf vs temperature for the thoria-insulated thermocouples, Fig. 5, illustrates the decreased but reproducible
thermoelectric output relative to an uninsulated thermocouple. As indicated from this plot, a calibration curve
may be generated for this type of thermocouple system
up to at least 3000°C when operated in a 25 to 76 mm hot
zone with a steep thermal gradient at the lead out.
To further evaluate the thoria shunting effects, thoriainsulated thermocouples of various designs were obtained
from commercial vendors and calibrated. Two 1.0 mm
o.d. tantalum sheathed, swaged thoria-insulated, WjW26Re
thermocouples were calibrated to 2000°C. One of the
thermocouples showed no signs of shunting at 2000°C,
while the other indicated up to 3% shunting at the same
temperature. This difference was probably due to differing
as-fabricated swage densities in the hot zone. A 1.6 mm
o.d. W26Re sheathed, thoria-insulated (high-fired), W3Rej
W25Re thermocouple agreed within ±2% of the calibration data up to 2200°C.
The discrepancies reported on the shunting effects of
tho ria-insulated thermocouples appears to be very dependent on the systems design. Other investigators have
also observed large shunting effects in swaged thoriainsulated thermocouples. 1O ,15 Hall and Spooner,6 on the
other hand, have reported no shunting effects up to 2100°C
for a W jW26Re thermocouple insulated with high-fired
thoria, also in agreement with the data of this work.
The available experimental data indicate that the type
and degree of contact between the thermoelements and
the insulator greatly affect the thermocouple output.
Swaged thermocouple designs with complete surface contact between the insulator and the wire results in larger
shunting. Vitrified insulator beads permit the use of large-
~:
!'
x •
;,,'
Ix x" I
".
"
,
•
'"
~.
., .~
~
'r
Oz
2J 1I1r11 IMM ER::' IU'~
PUHE n 02
2~ 111111 I~ME'ISiUN
99 9 ", PU III T
•
99 '.
x
999 '.PIJRETIDZ 16 111111 I'.1M[RSICH,
Pllfi~
450
113Re Vl25R.
[BARE WIRE)
400
W3R. W25Re
iTlt°2INSULATED)
350
25 - 76 Illrn IMMERSION DEPTHS
1.2 mm 0.0. W26R. SHEATH
0.25 mm W1R./W15R. THERMO·
ELEMENTS
l~o-u~li96'6u~tl ~6~~FJ~~ ~~01
BORE OIA.l
..
100
250
1500
1000
2500
1000
TEMPERATURE (oC)
FIG. 5. Calibration curve for thoria-insulated thermocouples.
diameter holes in the insulator relative to the diameter
of the thermocouple wire. This results in reducing the
contact area to a point or line contact between the insulator and thermo element. The larger diameter insulator
holes also allows a sine-shaped wiggle to be placed along
the wire, to reduce differential expansion problems.
Theoretical analysis of insulator shunting has been
reported by Tallman23 and Popper,24 but the results are
somewhat limited at present because of the lack of adequate materials properties and contact resistance data.
These analyses, however, do indicate a large effect of
wire-insulator contact, immersion depth, and temperature
gradient on the shunting characteristics of a thermocouple
system.
~!
,,.
•
1051
THER:\IOCOUPLES
4. Thermocouple Drift Test
Reliable use of thermocouple systems at high temperatures also requires a knowledge of the time-temperature
T J2 :611111 IMMFRSil)',
:J
~------~l~dll,--------~L---------~--------~3~~C~
FIG. 4. Shunting characteristic of ThOe-insulated thermocouples.
23 C. R. Tallman, "Analytical Model for Study of Thermocouple
Error Attributed to Electrical Conduction in Insulation," LA-DC7055, presented at Twentieth Annual Conference and Exhibit Instrument Society of America (4-7 Oct. 1965).
'
24 G. F. Popper, and T. Z. Zeren, "Refractory Oxide Insulated
Thermocouple Analysis and Design," WASH-1067, presented at the
High Temperature Thermometry Seminar, Washington, D. C. (24-26
February 1965).
Downloaded 07 Dec 2011 to 192.107.52.30. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
1052
R.
R.
ASAMOTO
stability of the thermocouple output. A WSRe/W26Re
thermocouple insulated with BeO remained stable for
240 h at 2300°C. 1O To evaluate higher temperatures, a
1.6 mm o.d. W26Re sheathed, thoria-insulated (high-fired),
W5Re/W26Re thermocouple was tested at 2425°C for
148 h, in an induction heated vacuum furnace. The output
of the thermocouple was monitored continuously by a
THE REVIEW OF SCIENTIFIC INSTRUMENTS
AND
P.
E.
~OVAK
multipoint recorder and the furnace was thermal cycled
several times during the test. The results of the test
showed that the thermoelectric output of this thermocouple
was stable within ±2%, with no observed negative or
positive drift. Additional tests indicated similar stability
of bare wire W3Re/W25Re and W5RejW26Re thermocouples (0.51 mm wire diam) for 15 h at 2600°C±50°C.
VOLUME 38, NUMBER 8
AUGUST 1967
Techniques for Conducting Shock Tube Experiments with Mixtures of
Ultrafine Solid Particles and Gases*
R. WATSON, A. L. MORSELL, AND W. J. HOOKER
Heliodyne Corporation, Van Nuys, California 91405
(Received 5 April 1967 ; and in final form, 1 May 1967)
Techniques have been developed for conducting shock tube experiments on mixtures of fine solid particles and
gases. Particles with diameters of 1 IJ. and smaller are prepared by grinding in a dry atmosphere. The powdered material is injected into the shock tube test section by opening a valve connecting the evacuated shock tube with a
tank containing gas into which the powder has been previously suspended by a pressurized injection. The axial
distribution of powder in the shock tube has been measured both by determining the weight of powder which has
settled out on small stainless steel slides placed in the shock tube and by observing the optical absorption from a
long-life reaction product of the shock-heated powder-gas mixture as the mixture sweeps by an observation port.
Typical operating results are presented, including particle size distribution of a solid material after grinding,
powder distribution along the shock tube axis after injection, and oscilloscope records of absorption at 2536 A by
the CF 2 formed from Teflon decomposition behind the incident shock wave.
THE
INTRODUCTION
techniques used in the shock tube testing of
gases and gas mixtures are reasonably well understood and widely used. However, only a limited number of
shock tube laboratories have been involved in tests requiring the introduction of solid particles into the test
section. It is the purpose here to describe a technique
which is being used successfully in this Laboratory in tests
designed to measure the optical transmission properties of
air-powder mixtures at equilibrium after the powders have
vaporized and reacted with other gases behind the
shock front.
The present article reviews the main aspects of the
technology involved in the preparation of powders and
the injection of mixtures of powders and gases into the
test section. Only a brief discussion of the theoretical aspects of the particle burnup behind the shock front is
given here since this is treated in detail elsewhere.l Except for the injection and presence of the dispersed solid
* This work was supported in part by the Advanced Research
Projects Agency, Department of Defense, under Contract No. D A04495-AMC-458(Z), and by the Air Force Ballistic Systems Division
under Contract No. AF 04(694)-777.
1 W. J. Hooker, "A Summary of Calculations of Chemical and
Fluid Dynamic Parameters in the SAPAG Facility," Heliodyne Corp.
Research Note 17, Los Angeles, Calif. (June 1965).
particles in the test section, the operation of the powder
injection shock tube is the same as the usual shock tube
such as those used for equilibrium thermal radiation
measurements, etc.
PARTICLE PREPARATION
The problems associated with satisfactory preparation
of the desired particle species are mainly those of attaining
sufficiently small sizes to achieve (a) low settling rates,
i.e., minutes, and (b) rapid vaporization and burnup rates.
Associated with these are the obvious requirements involved with handling (and purity) of the powder and
adequate dispersion throughout the test gas in the
shock tube.
The combined requirements of low settling rate and
rapid burnup dictate micron and submicron size particles.
The settling velocity is governed by the classical Stokes
formula2 for the velocity of fall of a sphere of diameter d
in a viscous gas, namely
Id2
(A)
V=--(Ps-pg) 1+2- ,
18 JL
d
(1)
2 Fundamentals of Gas Dynamics, H. W. Emmons, Ed., Princeton
Series of High Speed Aerodynamics and Jet Propulsion (Princeton
University Press, Princeton, N. T•• 1958), Vol. 3, pp. 722-724.
Downloaded 07 Dec 2011 to 192.107.52.30. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions