U
-
1
...SiT
3
-
UNCLASSIFIED
ORNL
2.
LA1
- .
'.', Arun
misma
WO
relea
beton at tim
149
TOFI
LI
CU
.
1.
1.
.
.
22
1
11
.1
PEŁ
internetesimen
LEGAL NOTICE
This report was prepared as an account of
Government sponsored work. Neither the
United States, nor the Commission, nor any
person acting on behalf of the Commission:
A. Makes any warranty or representa-
* tion, expressed or implied, with respect to
the accuracy, completeness, or usefulness
of the information contained in this report,
or that the use of any information, appa-
ratus, method, or process disclosed in this
report may not infringe privately owned
rights; or
B. Assumes any liabilities with respect
to the use of, or for damages resulting from
the use of any information, apparatus,
method, or procese disclosed in this report.
As used in the above, “person acting on
behalf of the Commission” includes any em-
ployee or contractor of the Commission, or
employee of such contractor, to the extent
that such employee or contractor of the
Commission, or employee of such contractor
prepares, disseminates, or provides access
to, any information pursuant to his employ-
ment or contract with the Commission, or
his employment with such contractor.
2
-
1
.
.
LE: CITY
TE
.
XW
.
Y
.
m
.
2
.
.
.
14
.
.
I
01
.
2
CL
OUT
TRI
,
4-
.
THE
.
7.
..
.
-
1
4
.
.
.
10
ü
Wern
Wien
N
N
SFC
.
....
CA/
WY
*
XXX
.
.
V
..
"
T
.
2.
LS
1
AUT
".
21
.
..
..
'.'.
'.'.
in.l. -.,.".
.
1
DITE
MICROCARD
ISSUANCE
DATE
-TE
.
-

.
*
.
i.
9/10


..
1
.1
..
:.
SY
.
:
-
G
.
.
*
T-
.
in-
.
1964
.
F
J.
pl
.
:5**
......
-
I

ORMP.149 DTTE
CONF-638-1
Drite: Add to

"
-
but
564
This note is an adáendum to ORNL-3556 and was prepared to allow presenta-
tion of the most recent racial heat flow apparatus measurements at the
National Physical Laboratory Conference on Thermal Conductivity, Teddington,
England, July 15–17, 1964.
+
SISTER
AUG 1 1 1984
.
.
-
:
.
.
:
arvan ...
COMPARISON OF THE THEMAL CONDUCTIVITY, ELECTRICAL RESISTIVITY,
AND SEIBECK COEFFICIENT OF A HIGH-PURITY TRON
AND ARMCO IRON TO 1000°C*
J. P. Moore, W. Fulkerson, and D. L. McElroy
Metals and Ceramics Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee, USA
--.--.-bio.
-LEGAL NOTICE -
The man wear wu
Oren
t


Der
siia
r
them,
A. When my warranty
o
mey,
s
o wel
c
the
omed
Month
the report or the
.
Focsimile Price $_Lulee
Microfilm Price $ _igo
. .
I. A My Wobmane me portw a , or for a rool rehe
aw
rm
orou Cartoonport.
Ao w
d
own to why arent "
w we
mogus or more d
t, hogy me contract, we
а олмаси
с не горю,
to, o
m
e ln, any Information
a t
wiu We Camom, *
w nowote.
. .
.
Available from the
Office of Technical Services
Deportment of Commerce
Washington 25, D. C.
.
. .
.
. .
.
.
.
Research sponsored by the U. 8. Atomic Energy Commission under
contract with the Union Carbide Corporation.
· COMPARISON OF THE THERMAL CONDUCTIVITY, ELECTRICAL RESISTIVMY,
AND SEOBECK COEFFICIENT OF A HIGH-PURITY IRON
AND ARM.CO IRON TO 1000°C
en omst'. .
J. P. Moore, W. Fulkerson, and D. In McEnroy
merkunden
S1
INTRODUCTION
The thermophysical properties of Arnico iron such as thermal conduc-
tivity,l electrical resistivity, and Seebeck coefficient have been
extensively investigated and reviewed up to 1000°C. Few investigations
of such properties have heen made on higher purity iron. If such a study
is made using the same apparatus to determine the properties of two purity
levels of iron, then several significant intercomparisons can be made
which add meaning to data on a single material. The systemic errors for a
single apparatus are the same, therefore comparison of a property of two
similar materials 18 more significant. A comparison of the property
changes with temperature and purity can show the effects of impurities on
the mechanisms contributing to a property and allows prediction of the
properties of iron as a function of purity. For these l'easons a study was
i initiated on a high-purity iron for comparison to Armco iron.
SPECIMEN CHARACTERIZATION
unitat
Sahib
The high-purity iron specimens were obtained by first arc melting
Armco iron stock in a pure inert atmosphere to produce pancake shap 'd
billets. These pancakes were then rolled into sheets and cut to make feed
stock for electron-beam melting. The pracluct of the electron-beam casting
was a 4-10. -diam by 6-in.-long billet. The outside edges of the billet
were trimmed off and two radial heat flow disks were machined from the
had 5:
h e
taviti
="
S
id
.
s
.
,,
R. W. Powell, "Armco fron as a Thermal Conductivity Standard.
Part I, Review of Published Data," p. 454 in Progress in International
Research on Thermodynamic and Transport Properties (ed. by Joseph F. Mesi
and Donald i. Toal), The American Society of Mechanical Engineers and
Academic Press, New York, 1962.

m
davriniam
center portion of the billet. One of the disks was 1.130 in. thick and
the other was 1.450 in. thick. X-ray radiographs showed that these pieces
had several small voids in them so the thermocouple wells were carefully
placed so that these voids would not interfere with radial heat flow
between the inside and outside thermocouples. A rod shaped specimen
0.15 in. in diameter and 3 in. long was cut from the center portion of
the billet for electrical resistivity and Seebeck coefficient measurements,
X-ray radiography showed that this rod was free of voids.
Table 1 is a comparison of impurity contents of the high-purity iron
and the Armco iron used in this experiment. It is apparent that the high-
purity iron has an impurity concentration smaller than the Armco iron by
a factor of 10. The Amco iron had a substantial amount of oxygen present
as a second phase. This second phase is very apparent in the microstruc-
ture of Armco iron and amounted to about 1 vol 4.2 The microstructure of
the high-purity iron was very clean showing no second phase. The high-
purity iron exhibited large grains on the order of 1/8-in. diam.
DESCRIPTION OF THE THERMAL CONDUCTIVITY APPARATUS
The ORNL radial beat flow apparatus2 was used for the thermal
conductivity measurements on the high-purity iron with the following minor
modifications.
Core Heater
The core heater was made by winding a double strand of 0.020-in. -diam
94% Pt-6% Rh wire on a 3/8-10. -OD by 15-1n.-long Coors AD-99 alumina tube
grooved with 12 turns per inch.
21. G. Godfrey, et al., Thermal Conductivity of Uranium Dioxdde and
Armco Iron by an Improved Radial Heet Flow Technique, ORNL-3556 Tune 1964).
Table 1. Chemical Analysise of the High-Purity Iron and Armco Iron
High-Purity Iron
Concentration
Equivalent
(at. %)
Carbonb
Arnco Iron
Concentration Equivalent
(at. $)
Carbon
Element
0.00021-0.0021
0.00014-0.0014
< 0.105
0.000045-0.00045
0.000030-0.00030
< 0.022
0.00009 0.0009
0.000019-0.00019
0.00095-0.0095
0.002-0.02
2888 LEGE DUADOR
0.00020-0.0020
0.00043-0.0043
< 0.054
·0.09
0.051
< 0.029
0.095
< 0.0%
< 0.012
< 0.022
0.061
0.01
0.040
< 0.0056
0.304
0.020
< 0.0116
0.0192
0.0110
< 0.0062
0.0203
< 0.0086
< 0.0026
< 0.0047
0.0124
0.0024
0.0086
< 0.0012
0.0651
0.0043
:).014
0.0020
0.0052
< 0.0056
0.0088
0.0020
0.00300
0.00043
0.00110
< 0.00120
0.00188
0.00042
Totals:
Minimum including oxygen
Minimum without oxygen
Maximum including oxygen
Maximum without oxygen
For elements considered
by Powell
0.04099
0.03219
0.0589
0.0501
0.00875
0.00687
0.01527
0.01339
0.00518 min
0.01102 max
0.672
0.368
0.960
0.636
0.1432
0.0782
0.2002
0.1351
0.0739 min
0.0825 max
a Al, Ca, Cr, Cu, Mn, Mo, Ni, si, Ti, v analyzed for by Emission Spectroscopy (semi quantitative); C, P,
S, H, O, N analyzed for by Quantitative analysis.
Diwelve times the sum of the ratio of the weight percent of each constituent to the atomic weight of
the element.
Thermocouple System and Specimen Configuration
The high-purity Iron specimen consisted of a 9-in.-high by 3 1/4-in.-
diam cylindrical column made up of six disks. Both of the measuring disks
of high-purity iron contained six 0.0625-in.-OD by 3/4-in.-deep thermocoupie
wells, three at an inside radius of 0.375 in, and three at an outside radius
of 1.437 in. These two measuring disks formed the central section of the
9-in.-high specimen. Since the supply of high-purity iron was limited,
Armco iron was used to construct four backup disks for extending the
column to 9 in. The two Armco iron disks at the ends of the specimen
stack had thermocouple wells at radii of 0.375 and 1.437 in, for deter-
mining the temperature profile in the specimen.
The thermocouples for the measuring planes were positioned as in the
Armco iron specimen, except they were insulated by two-hole 0.062-10. - OD by
0.015-in.-ID Degussit A123 alumina tubes. The 0.010-in. -diam thermocouple
Wiro extended from the measuring plane to lavito tie-dowa blocks on the
baseplate. This allowed unbroken thermoelerents in regions where
temperature gradients existed.
DESCRIPTION OF THE ELECTRICAL RESISTIVITY AND
SHIBECK COEFFICIENT APPARATUS
Auxiliary measurements or the electrical resistivity and Seebeck
coefficient were made on rod specimens in a vacuum of 100% to 10-7 torr.
Thermocouples were attached near each end of the specimen by welding
annealed reference grade Sigmund Cohn 0.010-in.-diam Pt and 90% Pt-10% Rh
wire onto the iron specimens with the hot junction made through the
specimen. The Pt and 90% Pt-10% Rh thermoelements were used to determine
the potential drop along the specimen during electrical resistivity meas-
urements. Current leads of 0.020-in.-diam PA wire were welded onto the
extreme ends of the rod specimen. The exact distance between the two
thermocouples was measured electrically by comparison to knite edges of
known spacing.
Low-temperature electrical resistivity data were obtained by immersing
the lastrumented specimen in three baths: 1ce and water (0°C); dry ice
and acetone (-86°C); and liquid nitrogen (-200°c). Eectrical resistivity
and Seebeck coefficient data for temperaturer, above 23°C were obtained by
positioning the rods in a 0.750-in. -OD by 30-in.-long alumina tube which
had been ground to fit standard tapered glass joints. The alumina tube
was attached to a vacuum system through a glass manifold. All electrical
leads were vacuum sealed with Apiezon black wax. The alumina tube was
then placed in a 15-in.-long clamshell furnace. An awd llary Nichrome
heater was wound around the outside of the alumina tube so that a tempera-
ture gradient could be established along the rod specimen for Seebeck
coefficient measurements.
A Rubicon six-dial potentiometer with a certifieå accuracy of
+0.01% +0.01 uv was used to measure the potential drop along the specimen,
the thermocouple emf's, and the potential drop across a standard 0.01-oh
resistor for current determinations. The accuracy of the electrical
resistivity measurements over the temperature range from 23 to 1000°C
was better than 10.2% and the precision was within 10.1%. The accuracy
of the Seebeck coefficient measurements is estimated to be approximately
+1%.
EXPERIMENTAL RESULTS
Thermal conductivity values for Armco iron in the ORNL radial heat
flow apparatus2 will be used here for comparison to high-purity iron. The
thermal conductivity data obtained on Armco iron in the temperature range
100 to 1000°C could be described to within 11.5% by four linear equations:
(2)
(2)
K = 0.7273 - 6.260 x 10-4t (100 st s 436°C),
k = 0.6554 - 4.609 10-4t (436. st 5786°C),
k = 0.2703 + 2.854 x 10-5t (786 st s 910°c), and
k = 0.1736 + 1.211 x 10-4t (910 st = 1000°c).
These four equations are plotted in Figs. 1 and 2 as kuste for Armco iron.
Thermal conductivity data for high-purity iron were obtained during
three runs. The results from these runs are tabulated in Appendix A.
Run I began with a data point at 59°C and extended to 900°C. Run 2 con.
818ted of three data points in the goiron temperature range from 920 to
UNCLASSIFIED
ORNL-DWG 64-5874
--HIGH PURITY IRON
---ARMCO IRON
....
...
.
.....-..-
K, THERMAL CONDUCTIVITY (w °C-4 cm-4)

K, , LATTICE THERMAL CONDUCTIVITY
(w ºc-' cm-4)
.....
.........
.
.
.
.
.
.
... *****
O
L
........
..
..
.........
-..do ..
............
...........
0
200
800
1000
400 600
TEMPERATURE (°C)
Fig. 1. Thoral Conductivity Conparisons Between Arrieo Iron una
lligh-Purity Iron Using the Wiecicmann-Ivan%-jorena Relation.
UNCLASSIFIED
ORNL-DWG 64-5875

+ 14 =
.
.
.
......
..
..
-
...-
..
.HIGH PURITY IRON...
---ARMCO IRON
...
T
I
.
5 PS
K, THERMAL CONDUCTIVITY (W °C -4 cm-1)
O
--------
K; , LATTICE THERMAL CONDUCTIVITY
(w °C -4 cm-)
.
-+
O
O
....
.....
...
..
....
200
800
1000
400 600
TEMPERATURE (°C)
Fig. 2. Thorld Conductivity Compuisons Retwocu Aimco Iron and
High-Purity Iron Usine the Backlund-Lindo Correlation.

1000°C. During Run 2, considerable thermocouple instability was
encountered which made the taking of data rather tedious. The temperature
was then raised to 1065ºC and left overnight to stabilize the thermocouples.
The temperature was then lowered and Run 3 was made between 893 to 999°C.
During Run 3, thermocouple instability was not a problem and it is
interesting to note that the data from Runs 2 and 3 agree with each other
within 12%. The data from these three runs on high-purity iron are
plotted in Figs. 1 and 2, and smooth curves are drawn through the data.
The basic characteristics of the thermal conductivity curves are the
same for the Armco iron and the high-purity iron. For example, there is
a slope change in both curves at about 436 and 786°C and a discontinuous
drop in both curves at the awy transition temperature. The thermal con-
ductivity of the high-purity iron 18 greater than that of the Armco iron
over the entire a-iron range, but the difference decreases greatly with
increasing temperature.
'Electrical resistivity and Seebeck coefficient results on ooth Armco
iron and the high-purity iron are tabulated in Appendixes A and B. The
Seebeck coef:ficients were assigned positive values since the hot junction
was positive with respect to the cold junction. The Seebeck coefficients
for the two irons are plotted 2,5 a function of temperature in Fig. 3.
Both curves show a sharp minimum at about 440°C and a large discontinuous
drop at the c-y transition temperature. The electrical resistivity data
were corrected for thermal expansion using the equation,
o corrected = p uncorrected (1 + at),
(5)
where t is in °C. A mean value for a of 14.8 x 10-6(°c)-d was used from
23 to 912°C and a value of 23.0 x 10-6 between 912 and 1000°C. ? A linear
specimen contraction of 0.33% was assumed to occur at the any transition. 3
The electrical resistivity of high-purity iron 18 lower than that of Armco
irca from -200 to 1000°C. The electrical resistivity of both irons shows
.
.
.
g
3Taylor Lyman (ed.), Metals Handbook, p. 2207, vol. 1, 8th ed., The
American Society for Metals, Novelty, Ohio, 1961.
.
UNCLASSIFIED
ORNL: DWG 64-5876

• ARMCO IRON
• HIGH PURITY IRON
..
SEEBECK COEFFICIENT (HV 90-1)
i
200
400 600
TEMPERATURE (°C)
800
1000
Fig. 3.
Seebeck Coefficient for Armco Iron and lich-Purity Iron.
a discontinuous drop at the dog transition, but this drop is less than 1%
in both cases. Values of the thermal conductivity, electrical resistivity,
and Seebeck coeflicient were taken from smooth curves and are given for
both irons in Table 2.
DISCUSSION OF RESULTS
Thermal Conductivity
In comparing Armco 1ron to the high-purity iron, one would expect
that the increase in k due to purification would be indicative of an
increase in both ke, the electronic portion, and ky, the lattice portion
of the thermal conductivity. The electronic portion may be calculated
from the electrical resistivity data assuming the validity of the Wiedemann-
Franz-Lorenz (W-F-L) relation. The results are plotted in Fig. I and show
the expected increase in ke with purity. However, the magnitude of the
increase is greater than that of the increase in the total thermal conduc-
tivity, which leads to the result that the lattice portion actually de-
creases with increasing purity as shown by the bottom two curves of Fig. 1.
This seemingly unreasonable conclusion is probably reconcilable by experi-
mental error. There is a probable determinate error band of about 1.5%
· around the Armco ironk data and a similar band of 1% around the high-purity
Iron k data. (The improved accuracy for the high-purity iron was due to
a larger ratio of outside-to-inside thermocouple radii.) If one adds 2.5%
of k to the measured difference in the total thermal conductivities, the
result is always larger than the corresponding difference in ke. Therefore,
the change in k 18 best explained as being due primarily to a change in ke
and any change in ki cannot be detected because the meusurements are not
sufficiently accurate.
The electronic portion of k was also calculated by a modified W-F-L
relation applied to iron by Backlund and initially proposed by Linde. S
"N. G. Backlund, "An Experimental Investigation of the sectrical and
Thermal Conductivity of Iron and some Dilute Iron Alloys at Temperatures
Above 100°K," J. Phys. Chem. Solids 21 (1/2), 1–16 (1961).
5J0. Linde, "An Investigation of the Validity of the Wiedemann-
Franz-Lorenz Law, " Arkiv. Syaik (38), 341 (1951).
11
Table 2. Snoothed Values of the Physical Properties
of Armco Iron and a High-Purity Irona
.
S
t(°C)
(w cm2 •c-2)
Armcob High-Purity
(uv/°K)
Armco High-Purity
-200
-100
(uohm-cm)
Arnco High-Purity
1.92 1.037
5.55 4.59
7.69 6.63
10.19 9.04
12.95 11.75 :
16.08 14.78
-150
50
19.59
100
18.26
14.85
200
17.17
16.50
13.76
10.65
8.95
21.84
300
11.50
30.64
41.29
53.90
9.02
400
500
600
700
750
· 800
850
900
0.696
0.665
0.602
0.539
0.477
0.425
0.379
0.333
0.310
0.294
0.295
0.296
0.296
0.285
0.289
0.295
0.748
0.698
0.621
0.555
0.489
0.436
0.386
0.338
0.313
0.297
0.296
0.296
0.297
0.280
0.287
0.298
32.35
43.27
56.41
71.72
90.35
101.55
108.98
112.70
114.72
114.97
114.94
116.15
117.95
11.30
15.20
17.45
19.03
19.85
68.64
85.85
96.13
105.32
109.45
112.33
112.78
112.33
113.45
115.34
9.44
9.39
11.47
15.27
17.49
19.39
20.33
21.05
910
13.64
920
950
1000
13.75
14.27
15.40
& See Appendixes A and B for actual measurtd values.
Calculated from the four equations applicable to Armco iron.
12
This relation may be written
(6)
ke = LT/ (0 + pod,
where p is the electrical resistivity, I is the theoretical W-F-L constant,
T is the absolute temperature, and po is a constant determined by Backlund
for iron. to be 2.66 uohm-cm. Figure 2 shows ke as calculated from Eq. (5) ·
for Armco iron and the high-purit.r iron as well as the corresponding curves
for ktQualitatively, the results are similar to those obtained using
the W-F-I relation although the difference in ke is more nearly equal to
the difference in k.
A further use of the data generated in the studies on Armco iron and
tre high-purity iron is the calculation of the thermal conductivity of any
reasonably pure iron for use as a thermal conductivity standard. Since
there is no preference between the W-F-L and Backlund treatments, an arbi.
trary choice of either can be made. If the S-F-I relation is chosen, the
effective lattice component of the thermal conductivity can be read as the
average of the ki results for the Armco and the high-purity iron. The elec-
tronic component can then be calculated from prior electrical resistivity
measurements as a function of temperature for the particular iron under
consideration. The sum of these two gives a reasonable (+1.5%) assessment
of the total thermal conductivity of the particular iron without the
necessity of performing the difficult-to-make total tiermal conductivity
measurements.
Powell. has plotted the thermal conductivity of iron at 200°C as a
function of impurity content. The ORNL values for high-purity iron and
Arrco iron correlate well with Powell's analysis and indicate an extrapo-
lated value of 0.70] w ºkº?cm-for iron with no impurities, assuming a
linear relation. Powell would predict a nominal value of 0.694 w °K-2 cm-1
for a pure iron. Powell considers the impurity elements, Cu, Mn, Ni, Si,
C, P, and s, as seen in Table 1 to be the major impurities in the iron.
The only other impurity of any consequence is oxygen which appears as a
second phase, probably iron oxide, and should not have as important an
effect on thermal conductivity as the dissolved impurities.
6R. W. Porell, "Armco Iron as a Thermal Conductivity Standard, Part I,
Review of Published Data," p. 461 in Progress in International Research on
Thermodynamic and Transport Properties (ed. by J. F. Mesi and D. H. Tsai),
The American Society of Mechanical Engineers and Academic Press, New York,
1962.
Electrical Resistivity
The electrical resistivity of the high-qurity iron as a function of
temperature lies below that of Armco iron over the entire temperature range
of measurement. The dirrerence in p increases slowly from a value of
0.88 wohm-cm at 200°C to 5.42 pohm-cm at 750°C. Above 750°C the differ-
ence decreases to 2. 19 uohm-cm at 910°C and then increases again to a
constant value of 2.6 Holm-em in the y region. The behavior of this dif-
ference is not as expected from elementary theory which would predict that
the impurity contribution to p is independent of temperature, at least at
high temperatures. It may be that the impurity concentration changes at
the higher temperatures because of increased solubility of second phase
materials; however, this is hardly a suitable explanation at low tempera-
tures and, furthermore, no change of electrical resistivity with time was
detected as might be expected for a dissolution phenomenon. The increase
in the difference to a maximum below the Curie point may be a reflection
of a magnetic interaction with the impurities.
Several measurements of p were made on the high-purity iron while
passing through the a-y transformation. The drop in p always occurred at
about 915 to 916°C, a few degrees above the accepted value of 910 to
912°c.? This higher value may be because of the purity of the specimen
since impurities decrease the c-y transformation temperature.? However,
the higher value may also be because the data were taken while heating, and
an appreciable superheating may have occurred because the pure iron does
not have sufficient nucleation centers to initiate the reaction. In fact,
during one run the electrical resistivity decreased slowly at 916°C over
a period of more than 1 hr to a final constant value.
At a given temperature, the electrical resistivity of high-purity
iron and Armco iron can be plotted as a function of impurity concentration
to obtain a reasonable extrapolation of p to zero impurity content at that
tenperature. At 100°C, a value of 14.62 uohm-cm was obtained assuming
7Taylor Lyman (ed.), Metals Handbook, p. 1207, vol. 1, 8th ed., The
American Society for Metals, Novelty, Ohio, 1961.
that p in a linear function of equivalent carbon concentration which is
the same assumption that was made above to extrapolate the k data. Only
the impurity elements considered by Powell were used to obtain the extrap-
olation. The asnmption of linearity of p vs equivalent carbon concentra-
tion is the same as assuming that all impurities are equally important
electron scatterers.
Seebeck Coefficient
In discussing the Seebeck coefficient, one should amit the effect of
the platinum reference and refer to the absolute Seebeck coefficient of
iron. This was done by Rudnitski8 and the general shape of the absolute
curve is the same as that of Fig. 3 for the iron-platinum couple. There-
fore, in the following qualitative discussion we will refer to Fig. 3 as
if it were an absolute plot. The saddle shaped curves of Fig. 3 are
characteristic for a ferromagnetic material. Rudnitski shows a sudden
change in the slope of the Seabeck coefficient vs temperature at
the Curie temperature. Such a sudden change was not apparent for either
the Armco or high-purity Iron in these studies although there is an
inflection near the Curie temperature.
. The most striking phenomenon in the Seebeck coefficient behavior is
the stepwise drop at the A- transformation which for the high-purity iron
was a -43% drop. This is all the more startling if it 18 compared to the
very small drop in the electrical resistivity. The general expression for
the Seebeck coefficient on the basis of the one electron moiel may be
written. 9
ya
k*2
s = get them la neces
S-
(7)
JES
8A. A. Rudnitski, Thermoelectric Properties of the Noble Metals and
Their Alloys, AEC-tr-3724, Publishing House of the Academy of Sciences, a
USSR, Moscow, 1956.
9J. M. Ziman, Electrons and Phonons, The Theory of Transport
Phenomena in Solids, pp. 397401, Oxford at the Clarendon Press, 1963.
.
...
...
.
.
.
.
.
;
..
... .
where
k* = Boltzmann's constant,
e = electronic charge,
€ - energy of electron, and
ple) = electrical resistivity calculated as a function of e.
...
....
..
-
-
-
-
-
-
The partial derivative is evaluated at the Fermi level, s. Although pls)
does not change appreciably at the c-y point, 'its derivative undergoes a
large change. Perlaps this is not mysierious since the change in lattice
structure also changes the band structure.
Comparison of the curves in Fig. 3 for Armco iron and high-purity iron
shows that the effect of impurities is to decrease the Seebeck coefficient,
S, of iron. The difference in s between the two irons decreases with
increasing temperature and becomes small between 550 and 750°C as can be
again rises above that for Armco tron, but the coefficients agree within
experimental error in the y region. If one writes the electrical resis-
tivity as
DE PL +40 ,
; where 40 is due to impurities and pr, is the remainder and then substituting
this expression into Eq. (7), we have
OL A + JOB
where A is a en pI/de)ers and B 18 (a en so/de)e=s. For iron 40
above room temperature'60 we can write
52-K (A + Beeld
S = - LeT (A + B
(10)
where L is the theoretical W-F-L constant. This expression shows that at
a given temperature, s 18 a linear function of 40. Since 10 is a more or
lees linear function of the impurity concentration, a similar relation 18
expected for S, so one can extrapolate to zero impurity concentration.
Such an extrapolation at 100°C yields a value for S for pure iron of
18.46 uv/°K which is only 1.1% above the value measured for high-purity
iron. Equation 10 also indicates why the difference in S between pure
iron and Armco iron decreases with increasing temperature since, in
general, 40/or decreases. This will be true providing A and B are rela-
tively insensitive functions of temperature.
CONCLUSIONS
1. The results of this study show Armco iron to be an excellent
thermal conductivity standard and the temperature dependence of the thermal
conductivity can be calculated to 11.5% using electrical resistivity
measurements.
2. The shape and magnitude of the temperature dependence of the
total thermal conductivity of iron se controlled by the electronic contri-
bution; this can be calculated from either the Wiedemann-Franz-Lorenz
relation or the Backlund relation.
3. The total thermal conductivity of iron has (a) a change of slope
at 436°C which possibly is related to a minimum in the Seebeck coefficient
at this temperature; (b) a minimum at 780°C which is associated with the
change in magnetic properties; and (c) a 4 to 7% drop at the a-y transform -
ation which is not predicted by the electrical resistivity measurements
but which is due to a 30 to 50% decrease in the lattice portion of the
thermal conductivity.
4. Current theoretical treatments of the total themal conductivity
of iron are not consistent in yielding a linear temperature dependence of
the thermal resistance due to the lattice portion. Furthermore, the effect
of impurities on the electrical resistivity is not a simple additive amount
independent of teraperature.
5. The Seebeck coefficient of iron 18 sensitive to impurity content
below 550°C and increases with purity. The Seebeck coefficient exhibits
a pronounced minimum near 430°C and undergoes a 43% decrease at the a-y
transformation.
-
---
-
-
6. The thermophysical property measurements on the two irons at
100°C can be used to predict the following property values for an iron free
of impurities: k, 0.701 w Ka?cm-2; 2, 14.62 uobm-cm; and s, 18.46 UV/°K.
U11
ACKNOWLEDGMENTS
The authors would like to thank C. F. Leitten, Jr., and R. B. McDonald
of ORNL for preparing the electron-bean-melted high-purity iron.
..
..
.
.
.
.
...
•
.
is
18
APPENDIX A
Physical Properties of High-Purity Iron
ردم. م ) و
ثم
مير)
(
(x) و
59
13.29
37
Run No. 1
0.7367
lll
0.6998
197
0.6223
290
0.5633
370
0.5072
450
0.4603
531
0.4201
614
0.3782
700
0.3382
754
0.3121
796
0.2973
848
0.2958
0.2964
100
23
149
200
251
300
352
398
243
299
349
397.5
447
501
550
601
647
702
898
847.5
799
750
923
950
996
11.51
10..29
9.45
9.16
9.39
10.11
11.47
13.06
15.27
21.04
20.28
19.38
17.49
13.77
14.27
15.30
450
90.).
Run No. 2
0.2828
958
0.2871
1000 0.3017
10.35
11.06
14.74
10.29
18.05
21.92
26.14
30.67
35.94
41.07
47.38
53.98
61.18
68.33
77.02
85.85
96.16
105.48
109.45
112.35
102.06
90.25
96.18
112.55
112,93
112.21
112. 20
212.47
113.92
115.30
9.048
5.170
1.0378
920
893
· 501
551
598
651
700
750
80.2
850
900
774
722
750
904
913
917
918
925
964
1000
0.0
Run No. 3
0.2966
920
0.2837
937
0.2822
959
0.2866
999
0.3000
200
& Corrected for core heater expansion.
Corrected for thermal expansion of specimen.
cwith respect to platinum.
duncorrected for expansicn.
19
APPENDIX B
Physical Properties of Armco Iron
is
.
. .
sb
.
t
(uchen.com
todo
lebens (wohacm) Fotos
(wv>°K)
.
.
34
.
.
73
68
102
152
201
250
.
..
.
..
350
12.04
12.99
14.39
17.47
20.30
23.25
24.72
16.20
23.20
28.05
34.06
34.1l
41.54
12.1
161
199
217
103
199
255
317
318
385
480
534
584
647
670
452
544
..
637
701
747
799
847
873
625
705
804
894
952
1002
915
664
531
384
78.30
90.40
100.97
108.90
112.51
113.84
76.15
91.91
109.35
114.31
116.18
117.98
114.71
82.84
60.78
41.35
13.05
398
449
500
550
598
651
699
746
774
801
819
864
16.97
16.51
15.21
13.83
12.11
9.57
8.97
8.76
9.05
9.87
11.19
13.04
15.08
..17.17
18.38
53.57
18.99
61.24
68.96
79.89
84.23
49.77
62.80
78.63
0
10.190
11.400
19.34
20.04
19.73
13.63
893
924
6.430
1.92€
630
Corrected for thermal expansion of specimen.
"With respect to platinum.
"Uncorrected for expansion.
END