MATERIALS

SCIENCE &
ENGINEERING

ELSEVIER Materials Science and Engineering A202 (1995) 134 141

A

Investigation of thermal and hydrogen effects on emissivity of

refractory metals and carbides
Yo Ozaki, Ralph H. Zee
Materials Engineering Program, Auburn University, Auburn, AL 36849, USA
Received 6 June 1994; in revised form 27 January 1995

Abstract

Direct thermal-to-electric power-conversion systems operate at high temperatures to ensure high power density and proper heat
transfer in such systems, critical to their stability. Radiation in this temperature regime is a main means of heat transfer and the
control of the surface characteristics of radiative components is thus an important issue. High-temperature normal spectral
emissivity of various high-temperature components was investigated. Materials examined include tungsten, a dispersion strength-
ened tungsten (W-HfC), a chemical vapor deposition (CVD) coating (CVD-W) and five types of carbides (WC, TaC, NbC, ZrC
and HfC). Spectral emissivities were measured using single- and dual-wavelength radiation thermometries. Emissivity of tungsten
was found to be 0.5 and remained constant even after vacuum annealing up to 2723 K. The five carbides were exposed to high
temperat.ures, both in vacuum and in hydrogen. Results show the process to be thermally activated and hydrogen to have little
effect on the ceramics examined. Tungsten carbide was the least stable and it transformed into W2C upon annealing whereas NbC
decomposed after similar exposure. Hafnium carbide and ZrC were found to be the most stable and possessed the highest
emissivity (0.9) among the carbides investigated. The experimental observation was examined in terms of thermodynamics of the
materials. The implications of emissivity on the thermal characteristics of a typical space nuclear power system were examined.

Keywords: Thermal effects; Hydrogen; Emissivity; Refractory metals; Carbides

1. Introduction ity of a material [11,12]. One method of determining the

emissivity is to measure the reflectivity, and emissivity is
One o f the key issues in future space exploration' is complementary to reflectivity. Another way is to mea-
the development of advanced power and propulsion sure the absolute radiant power and use this value to
systems [1,2]. These systems will require materials to calculate the emissivity. The most direct way is to
withstand high temperatures in combination with cor- compare the radiation from the material with that from
rosive hydrogen gas and large thermal gradients [3-8]. a perfect emitter (black body), the ratio of which yields
The effectiveness of the propulsion system is a strong the emissivity value. The difficulty with many of these
function of temperature. Radiative coupling between methods is the precise measurement of the surface
components is crucial, especially in systems which rely temperature of the sample.
on such coupling as the main source of heat transfer Components in nuclear propulsion systems may be
from the nuclear fuel [9,10]. Surface coatings with exposed to hot hydrogen. Hydrogen is known to perme-
materials o f high emissivity are frequently employed to ate increasingly through metals with increasing tempera-
enhance the thermal characteristics o f the system. In the ture. It reacts with graphite to form methane at elevated
case of the nuclear propulsion system, the hydrogen temperatures and it interacts with carbide precipitates to
fuel enters the nuclear reactor as a cryogenic fluid and enhance cracking [13-15]. These reports also suggest
exits the reactor at a temperature in excess of 2300 K. that hydrogen would alter the surface of the material,
Thermal radiation depends directly on emissivity. There thereby changing the emissivity. Since the refractory
are numerous methods to measure the surface emissiv- metal carbides are the candidate materials for heat-

0921-5093/95/$09.50 © 1995 -- Elsevier Science S.A. All rights reserved

SSDI 0921-5093(95)09793-7
 Y. Ozaki, R.H. Zee / Materials Science and Engineering A202 (1995) 134-141 135

transfer components, the effects of hot hydrogen on the nation of the radiative power as a function of wave-
emissivity of carbides need to be investigated. length [17] reveals that under this high-temperature
condition ( ~ 2200 K), 95% of the radiative power
occurs at wavelengths below 6 p m (mid-infrared range).
2. Experimental procedures Owing to the conductive nature of the materials investi-
gated, it is anticipated that free electrons are responsi-
Dual-wavelength and single-wavelength radiation ble for the emissivity observed in this wavelength range.
thermometers were used to measure high-temperature The wavelength employed in this study was 0.81 pm,
emissivity. A dual-wavelength thermometer involves near the peak of the radiation curve. These conditions
measuring the spectral radiance at two different wave- should result in emissivity values relatively independent
lengths and taking the ratio of two radiant values [16]. of wavelength. To examine the effect of a different
If the material of interest has the same emissivity at the emissivity value for the far-infrared regime (above 6
two wavelengths, the emissivity contribution to the /tm), an error analysis was conducted which divided the
radiant power cancels out and the surface temperature integral over wavelength into two parts: one up to 6
of the sample can be calculated from the ratio of the ktm, and the other above. The emissivity values ob-
spectral radiance measured at the two wavelengths. tained in this study (wavelength of 0.81 pm) were used
Therefore, the absolute temperature of the sample can for the first integral whereas the two extreme values
be determined without knowing the emissivity. In con- of zero and one were used for the far-infrared
trast, the temperature measured by a single-wavelength range. Results show that even such extreme emissivity
thermometer requires a correction based on the emissiv- values in the far-infrared range, only lead to uncertain-
ity of the sample. The emissivity of the sample can be ties of the order of _+ 3%. This is owing to the low
determined by comparing the signals obtained from the radiative power in the large wavelength region at high
single- and the dual-wavelength thermometers in the temperature.
following manner. The general equation relating the Wavelengths of 0.71 p m and 0.81 p m were chosen
black-body function and emissivity is given by for the dual-wavelength thermometer, and a wavelength
of 0.81 p m was chosen for the single-wavelength instru-
e~M;~(T) = M~(Th) (1)
ment. Both signals were simultaneously measured by an
where ex is the spectral emissivity at a wavelength 2, optical pyrometer (Model Tempmatic 8120G-C-T,
M~(T) is the black-body function at this wavelength manufactured by the Williamson Corporation). The
and at temperature T, and Tb is the effective black-body samples were heated by electron-beam bombardment in
temperature. Assuming that the spectral emissivity is an ultrahigh vacuum furnace with a base vacuum of
independent of wavelength ( = e) and taking the ratio 10-~l Torr. The temperature was measured through a
Mx(TD/Mx(T) and subsequently integrating over all glass viewport with a transmittance of 90% for the two
wavelengths, yields the following simple relationship wavelengths of interest.
between emissivity and measured temperatures: Tungsten with 0.4 tool.% hafnium carbide (disper-
sion-strengthened tungsten denoted as W - H f C ) , which
e = (Tb / T) 4 (2)
is close to the optimum composition for mechanical
where e is the effective emissivity, Tb is the effective properties for this class of materials according to Klopp
black-body temperature (from a single-wavelength ther- et al. [18], was chosen as one of the materials to be
mometer) and T is the true temperature (from a dual- investigated. This alloy is a candidate material for fuel
wavelength thermometer). The assumption of a cladding for space nuclear systems. In order to enhance
constant emissivity (independent of wavelength) can be the surface characteristic o f this refractory alloy, a
justified as follows. All the materials examined in this chemical vapor deposition (CVD) coating of pure tung-
study (tungsten and carbides) are conductive materials sten is commonly applied to the substrate. In this paper
with no band gap and therefore should not possess this coating is denoted as CVD-W. The W HfC alloy
electronic transitions which have highly energy-depen- was obtained from Westinghouse Corporation, Pitts-
dent emissivity. This implies that only two mechanisms burgh, Pa, and CVD-W was placed on the W - H f C by
are responsible for the radiative behavior: free electrons General Atomics Corporation, San Diego, CA. The
and phonons. In most metallic materials, the former emissivity of pure polycrystalline tungsten was also
phenomenon dominates at the near-infrared and mid- included in this study as a standard.
infrared regimes (up to about 6/~m) whereas the latter Refractory metal carbides have high emissivity values
is important at the far-infrared range (up to 15 /zm). and are therefore suitable for use as thermal compo-
The emissivity in this study was measured at high nents in nuclear thermal propulsion systems. These
temperatures (up to 2500 K) where the peaks of the materials are also being considered as coatings to en-
radiation curves occur at wavelengths of approximately hance the heat transfer characteristics of radiative com-
1.2 g m which is in the near-infrared range. An exami- ponents. Therefore, the intrinsic emissivity of carbides
136 Y. Ozaki, R.H. Zee /Materials Science and Engineering A202 (1995) 134 141

is an important parameter. Emissivity and the effects of Table 1

thermal and hydrogen degradation in refractory metal Emissivity results for the tungsten materials

carbides of WC, TaC, NbC, ZrC and HfC were exam- W-HfC CVD-W Pure W
ined in this study. These carbides were obtained from
Oak Ridge National Laboratory. Temperature 1850-2250 1700-2400 1800 2400
range (K)
As-received 0.52 0.51 0.49
3. Results (polished surface)
Annealed at 2723 K 0.48 0.49 --
for l h
The normal spectral emissivities of the samples at a
(polished surface)
wavelength of 0.81 Hm were measured using graphite as
As-received -- -- 0.51
a reference. Graphite was used as a reference because
(rough surface)
its emissivity has been well studied and it is relatively
independent of wavelength in the micrometer range. A
graphite emissivity value of 0.9 was used to determine
the absolute emissivity values of the samples in this The emissivity values of these materials are summarized
study. In this research a rough surface refers to a in Table 1.
surface finished with a # 240 abrasive paper and a The normal spectral emissivities of WC, TaC, NbC,
polished surface implies a surface finished with a ZrC and HfC at a wavelength of 0.81 Hm were also
# 4000 abrasive paper. measured using graphite as a reference. Results of the
Fig. 1 shows the single-wavelength intensity signal emissivity measurements are shown in Table 2. Fig. 2
versus temperature plot of three materials: W - H f C shows the single-wavelength signal versus temperature
(tungsten with 0.4% HfC), CVD-W (tungsten with coat- plot of the materials investigated in this group. These
ing) and the graphite reference. The experiments for the emissivity values are higher than previously reported in
W - H f C and the CVD-W samples were conducted in the literature [19]. The main difference between this
the as-received condition as well as after vacuum an- study and the earlier investigation is the methods used
nealing at 2723 K for 1 h. The single-wavelength inten- to measure the absolute temperature of the materials.
sity signal is a direct function of emissivity and absolute In this research the surface temperature was directly
emissivity values were obtained by comparing this in- measured using the dual-wavelength radiation ther-
tensity with that of graphite (with an emissivity of 0.9). mometer whereas bulk temperatures were measured
using thermocouples in the previous study. In the case
,~ 100
of WC and TaC, the emissivity was found to increase
> when the surface roughness increased whereas the emis-
V
sivity of NbC, HfC and ZrC remained constant for the
._1 two surface conditions. One question that must be
Z addressed in the present study is the assumption that
(.9 the emissivity values measured are independent of tem-
t/) perature within the range examined. Figs. 1 and 2 show
-I- the measured intensities as a function of temperature
I.--
0 for the various materials investigated. These radiation
z 10 -1 D
intensities were examined according to the expected
I.iJ
_1 [] values from'a constant emissivity value and the results
LIJ
> were found to agree very well. If the emissivities were a
3: • C function of temperature, a significant deviation of the
I zx C V D - W as-recelved data from the predicted radiation intensities should
I.iJ have occurred. Such a deviation was not observed in
-J • CVD-W heat treated
£.0 any of the measurements. The temperature dependence
z zx n W-HfC as-received of the emissivity in conductors is expected to arise from
(D • W-HfC heat treated an electrical conductivity variation. At the high temper-
10 - 2 ' ' ' ' ' ' ' ' ' I ' ' ' ' ' ' ' ' ' I ' ' ' , , , , , , I . . . . . . ,

1700 1900 21 O0 2300 ature of interest, the electrical conductivity of the car-
bides exhibits only slight temperature dependence.
TEMPERATURE (K) Results from a previous study of tungsten [20] show
that its emissivity varies by only _+ 5% between 1600 K
Fig. 1. Plot of single-wavelength signal versus temperature for W
HfC, CVD-W in the as-received condition and after annealing at
and 2800 K in agreement with our analysis.
2723 K for 1 h. Also included are the signal intensity obtained for the The emissivity of NbC was found to decrease after
graphite standard. each measurement, i.e. after high-temperature exposure
 Y. Ozaki, R.H. Zee /Materials Science and Engineering A202 (1995) 134 141 137

Table 2
Emissivity results for carbides

WC TaC NbC ZrC HfC

Temperature range (K) 1700-2500 1800 2400 1700 2200 1700-2200 1700 2200
Polished surface 0.82 0.43 0.79 0.9 0.9
Rough surface 0.9 0.61 0.79 0.9 0.9
After thermal treatment 0.71 0.43 0.71 0.9 0.9

in vacuum. The emissivity gradually decreased from verted from WC to W2C owing to thermal exposure.
0.79 with thermal exposure and approached a constant Fig. 4 shows the X R D patterns obtained from the
value of 0.71. According to X-ray diffraction ( X R D ) sample before and after exposure.
analysis, there was no apparent difference in phase The emissivities of the carbides that were exposed to
structure before and after the measurement (both spec- hydrogen at high temperatures were also measured.
tra indicate N b C only). F r o m scanning electron mi- Table 3 shows the conditions for hydrogen exposure
croscopy (SEM) observations, black spots were found and the results of the emissivity measurements. Details
on the surface of the sample in the as-received condi- of this hydrogen exposure are given in a previous paper
tion. The concentration of these black spots increased [3]. The exposure has no observable effect on either
after each measurement. The energy-dispersive X-ray HfC or ZrC. The emissivity of N b C was reduced from
spectroscopy (EDS) result indicates that the constituent 0.79 to 0.72 as a result of exposure to hydrogen at 1536
of the black particles is carbon (graphite) which is likely K for 2 h. The SEM micrograph of the N b C sample
to be a residual from the fabrication of the NbC. The after hydrogen exposure shows that the amount of
reduction of emissivity with thermal vacuum exposure graphite particles on the surface of this sample is more
is likely to be a consequence of the decomposition of than that in the same material in the as-received condi-
NbC, tion. The surface of the T a C was polished prior to
In the case of WC, the emissivity was also found to hydrogen exposure with an emissivity of 0.43. However,
decrease after each measurement, especially after the the emissivity of the material after exposure to hot
WC sample was exposed to temperatures above 2330 hydrogen at 1390 K for 1 h was found to have in-
K, as shown in Fig. 3. The emissivity was found to creased to a value similar to that of T a C with a rough
decrease from 0.9 to 0.71. Analysis of the X R D spectra surface (emissivity of 0.63).
indicates that the surface of the sample has been con-

10 0
10 0 i

-J

_. //
[] D []
0 v 0 -r
I,,-
r~ 10_1
Z 1 0 -~
,.,J"' ///iI WC
IaJ
> 0 WC > I st M e s u r e m e n t
<( o /
v NbC / 2nd
I [] ,., 3rd
I.d [] T a C /
..J
(.9 (.~
• C, H f C a n d ZrC data z
Z i

(/1 10 -2 . . . . . . . . , . . (/) 10-2 ............................................

1700 1900 21 O0 2300 1700 1900 21 O0 2300 2500
TEMPERATURE(K) TEMPERATURE(K)
Fig. 2. Plot of single-wavelength signal versus temperature for WC, Fig. 3. Plot of single-wavelength signal versus temperature for WC
NbC, TaC, HfC and ZrC in the as-received condition. The graphite showing a decrease in emissivity owing to vacuum annealing towards
standard is included for comparison. progressively higher temperatures.
138 Y. Ozaki, R.H. Zee / Materials Science and Engineering A202 (1995) 134-141

5000
(o)

4000

Z
0
5000

z 2000
0

1000

30 50 70 90 110 130 150

2 THETA

2000
(b) Fig. 5. Microstructure of the CVD coating deposited on a W - H f C
substrate. The sample has been annealed at 2673 K for 1 h and the
substrate has undergone grain growth whereas the C V D coating
1500 remained unaffected.
Z
0

~ 1000
most likely to be a direct consequence of the change in
the grain structure. An increase in the surface rough-
Z
0
ness results in an increase in the emissivity (unalloyed
tO
5OO tungsten with a polished surface was found to have a
lower emissivity than that with a rough surface, as
shown in Table 1) and small grains could have the same
0 [ effect as a rough surface. A larger grain size is expected
30 50 70 90 110 130 150 to result in a more polished surface. On the contrary,
2 THETA the thermal effect on emissivity in the CVD-W material
Fig. 4. X-ray diffraction patterns obtained from W C in (a) the
is smaller (decreased from 0.51 to 0.49). The CVD
as-received condition showing W C peaks, and (b) after vacuum heat coating appears to be more resistant to grain growth
treatment at 2500 K showing W2C peaks. than the W - H f C substrate as illustrated in the mi-
crostructure shown in Fig. 5. After annealing for 1 h at
4. Discussion 2673 K, the W - H f C substrate had undergone signifi-
cant grain growth whereas the CVD-W coating was still
All the tungsten materials investigated have similar unaffected.
emissivity values of 0.5. High-temperature annealing The emissivity of WC and TaC was found to increase
resulted in a small decrease in emissivity as shown in with surface roughness whereas the emissivity of NbC,
Table 1. The CVD-W and W - H f C samples that were HfC and ZrC remained constant with surface morphol-
annealed at 2723 K had a slightly lower emissivity (0.49 ogy. Increasing the surface roughness usually leads to
and 0.48 respectively) than the same samples but in the an increase in emissivity. However, the high porosity of
as-received condition (0.51 and 0.52 respectively). Re- N b C , HfC and ZrC (their densities were 68%, 62% and
sults from the microstructure analysis show that the 66%, respectively, of their respective theoretical densi-
grain size of the annealed W - H f C samples is of the ties) makes it impossible to obtain a polished surface (in
order of 5 0 0 / l m which is approximately ten times that contrast, the densities of WC and TaC were 86% and
found in the as-received materials. The decrease in 94% respectively). Therefore, the emissivity was the
emissivity from 0.52 to 0.48 as a result of annealing is same for both polished and rough surfaces.
After each set of emissivity measurement (i.e. high-
temperature exposure in a vacuum), the emissivity of
Table 3
Hydrogen exposure conditions and emissivity results for carbides NbC was found to decrease. Niobium carbide is pro-
duced by combining Nb205 and carbon. However, at
TaC NbC ZrC HfC high temperatures (above 1800 K), the Nb metal was
reported to be formed during the process [21-23]. This
Temperature (K) 1390 1536 1180 1180
indicates that NbC may decompose into Nb metal and
Time (h) 1 2 1 1
Emissivity 0.63 0,72 0.9 0.9 carbon at high temperatures. Therefore, the decrease in
the emissivity may be owing to a partial decomposition
 Y. Ozaki, R.H. Zee / Materials Science and Engineering A202 (1995) 134 141 139

of NbC into Nb which forms a thin layer on the surface The effect of surface emissivity on the thermal char-
of NbC. Surface analysis using SEM/EDS showed an acteristics of a fuel rod in a direct thermal-to-electric
increase in the graphite particles owing to high-temper- nuclear power conversion system was calculated. The
ature exposure. However the X R D result does not differential equation governing the temperature distri-
conclusively indicate the presence of metallic niobium bution in a cylindrical fuel geometry at steady state is
on the surface. The Nb layer may be sufficiently thin given by
that it escaped detection by XRD, but still affects the
measured emissivity value. Niobium possesses an emis-
l d (rd )
rdr\ -~r ÷K-=0 (5)
sivity of about 0.35 which is lower than that of NbC.
The similarity of the data obtained from the samples with the boundary conditions:
exposed to hot hydrogen and to high temperature in a
dT
vacuum indicates that the degradation process is domi- K--~r + EaT4 = 0 at the surface of the fuel rod (6a)
nated by a thermal reaction.
In the case of WC, a decrease in emissivity was and
detected during the measurement. A survey of the
dT
possible high-temperature reactions leads to the follow- - - = 0 at the center of the fuel rod (6b)
ing decomposition process, dr
where Ao is the power density of the fuel, e is the
2WC ~ W2C + C (3) emissivity, ~r is the Stefan-Boltzmann constant
(5.67 x 10 -8 W m -2 K -4) and K is the thermal con-
The free energy of the reaction [24] is governed by
ductivity of the fuel. The thermal conductivity of UO2
AG(cal mole ~) = 6440 - 2.80T (4) fuel is a relatively constant value of 2.3 W m ~ K ~ at
the high temperature (above 1500 K) of interest [25].
This implies that the equilibrium of the reaction The analytical solution to the above equations is [26]:
(AGr = 0) occurs at 2300 K. Above this temperature,
the decomposition of WC into W2C and C is favored T(r)= ~Ao( a 2 --r 2)+ \( A°a ) (7)
energetically. This agrees with the result of this study.
The emissivity of WC is 0.9 and that of WzC is 0.71. where a is the radius of the fuel element. The emissivity
After hydrogen exposure at 1390 K for 1 h, the employed in this analysis corresponds to the spectrally
emissivity of TaC with a polished surface increased averaged hemispherical emissivity. To use the emissivity
from 0.43 to 0.63. The emissivity value is about the values measured in this heat-transfer analysis requires
same as TaC with a rough surface. In contrast, high- the condition that the emissivity of the materials be
temperature exposure without hydrogen did not alter insensitive to wavelength, temperature and direction.
the emissivity of TaC. Analysis of the surface of the The first two assumptions were discussed and justified
sample exposed to hot hydrogen reveals a texture simi- earlier in this paper. The last criterion, directional
lar to that from the sample with a rough surface after dependence, is also expected to be small at the high
measurement. Therefore the emissivity enhancement temperature examined in this study where energy emis-
owing to hydrogen exposure is caused by surface sion is diffused. Results from earlier studies [27-29]
roughening due to hot hydrogen. show that the emissivity of metallic surfaces is relatively
The ability of thermal components to dissipate their independent of the angle. Deviation from constant
heat is proportional to their emissivity. Results from emissivity occurs only at high angles (above 75°). Since
this study show that the surface condition of WC, TaC the radiative power is weighed by the cosine of this
and NbC could be affected by high-temperature expo- angle, the effect of such deviation on radiative heat
sure in hydrogen or in vacuum resulting in changes in transfer is secondary.
the emissivity. However, no effect was observed in HfC According to Eq. (7), the maximum temperature
and ZrC. This can be attributed to the high stability of occurs at the fuel centerline position (r = 0) and the
HfC and ZrC. The free energies of formation for HfC, minimum at the outside surface (r = a). The form of the
ZrC, WC, TaC and NbC at 2000 K are --210 kJ solution indicates that the profile of the temperature
mole l, _ 1 8 4 k J m o l e l, _ 5 8 k J m o l e i, _ 1 4 2 k J distribution (the first term in Eq. (7)) is independent of
mole J and - 1 3 4 kJ mole-1 respectively [24]. Fur- surface emissivity. Based on the design proposed for a
thermore, the surface emissivity of ZrC and HfC is the 570 kW thermal system [30] with a fuel pellet diameter
highest (0.9) among all the materials investigated. of 15.7 mm (7.9 mm for the parameter a) and a
Therefore, from both the long-term stability and radia- uranium enrichment of 93%, the power density (Ao) of
tion efficiency points of view, HfC and ZrC are the the heat source is 0.117 W mm 3. The temperature
ideal materials for advanced nuclear energy conversion difference (AT) between the center and the surface of
applications. the fuel is simply given by Aoa2/4K. For the conditions
140 Y. Ozaki, R.H. Zee / Materials Science and Engineering A202 (1995) 134-141

3000 .... , . . . . . . . . ,... (HfC and ZrC). In a nuclear power system it is antici-
pated that the enriched UO2 fuel will be clad inside a
refractory material and possibly with an emissivity en-
v hancement coating. It is important to maintain the fuel
v
in its solid state and the cladding material at sufficiently
LM 2500
g low temperature in order to retain its mechanical stabil-
ity. The melting temperature of UO2 is 3100 K and the
.5 m a x i m u m operation temperature of W - H f C is approx-
n-
LM imately 2000 K before a significant loss of creep
El. strength occurs. This implies that the centerline temper-
2000
LM .9 ature and the surface temperature must be maintained
I-- below 3100 K and 2000 K respectively. According to
Fig. 7, a minimum emissivity of 0.25 is required to
Emissivit satisfy the former condition but a higher emissivity
1500 I I I I [ I I . . . . . . I . . . . value of 0.5 is necessary to prevent degradation of the
0 -5 0 5 cladding. Experimental results obtained in this investi-
gation show that the emissivities of tungsten (pure,
DISTANCE FROM CENTER (mm)
W - H f C and CVD-W) are all approximately 0.5 and
Fig. 6. Radial temperature distributions in a UO2 nuclear fuel rod for that of TaC is also in the same range. On the contrary,
two emissivity values. a higher emissivity of 0.9 was found in HfC and ZrC
and both carbides were very stable. According to the
selected, this A T corresponds to a constant value of 783 above analysis, the use of tungsten without appropriate
K. A lower emissivity value of the radiative surface emissivity coatings will lead to a high cladding tempera-
results in a higher surface temperature and a corre- ture of 2000 K. Under a prolonged operation, degrada-
spondingly higher centerline temperature while main- tion of the mechanical properties of the cladding
taining a constant AT. Fig. 6 illustrates the temperature material may occur. Coating of the cladding with either
distribution in claddings with two different emissivity HfC or ZrC reduces the cladding temperature to 1700
values of 0.5 and 0.9. Fig. 7 summarizes the fuel K and this reduction will significantly lengthen the life
centerline temperature and the surface temperature as a expectancy of the cladding material. The fuel centerline
function of emissivity calculated for the power and temperature also decreases from 2735 K to 2463 K with
geometrical conditions assumed. The temperature of the application of a coating which leads to a more
the fuel increases dramatically when the emissivity stable fuel element.
drops below a value of 0.5. The emissivity of the
materials experimentally investigated in this study 5. Conclusions
ranges from a low of 0.5 (tungsten) to a high of 0.9
High-temperature normal spectral emissivity at a
3500 . . . . wavelength of 0.81 a m of tungsten, W - ' H f C , C V D - W
and five types of carbides (WC, TaC, NbC, ZrC and
HfC) were measured using single- and dual-wavelength
radiation thermometers. The emissivity of the three
3000
types of tungsten examined was found to be similar
with a value of 0.5 and thermal treatment had little
effect on their surface emissivity. The effects of high-
2500 temperature exposure, with and without hydrogen, on
the emissivities of the carbides were also investigated.
Tungsten carbide was the least stable and it trans-
formed into W2C upon annealing whereas N b C decom-
2000 posed. Hafnium carbide and ZrC were found to be the
most stable among the carbides investigated in this
research. The experimental results were in accordance
with the thermodynamics of the materials. The effect of
1500 i , , i
0.0 0.2 0.4 0.6 0.8 1.0 emissivity on temperature distribution in a nuclear fuel
cladding of a typical space reactor was evaluated. The
EMISSIVITY use of an emissivity coating (such as HfC and ZrC)
Fig. 7. Dependence of peak centerline and surface temperature on would significantly enhance the thermal stability of the
emissivity for a 570 kW thermal nuclear system. system.
 Y. Ozaki, R.H. Zee / Materials" Science and Engineering A202 (1995) 134-141 141

Acknowledgements [11] R.N. Wall, D.R. Basch and D.L. Jacobson, J. Mater. Eng.
Perform., 1(5) (1992) 679.
[12] J.T. Gier, R.V. Dunkle and J.T. Bevans, J. Opt. Soe. Amer., 44
This work was supported by the U.S. Air Force (1954) 558.
Phillips Laboratory at Kirtland Air Force Base, New [13] C.A. Hippsley and C.E. Lane, Mater. Sci. Technol., 6 (1990)
Mexico and the Department of Energy Office of De- 735.
fense Energy Projects through the Auburn University's [14] H. Huschka, G. Luthardt and V.E. Portscher, Nuclear Technol.,
Space Power Institute which the authors greatly appre- 66 (1984) 562.
[15] N. Kishimoto, T. Tanabe, H. Araki, H. Yoshida and R. Watan-
ciate. abe, Nuclear Technol., 66 (1984) 578.
[16] W.R. Barron, Sensors, 17(1) (1990) 15.
[17] R. Siegel and J.R. Howell, Thermal Radiation Heat TransJer,
References McGraw-Hill, New York, 1972, p. 29.
[18] W.D. Klopp, P.L. Raffo and W.R. Witzke, J. Metals, (1971)
[1] R. Ewell, Proc. 4th Int. ConJi Thermoelectric Energy Conversion, 27.
Institute of Electrical and Electronics Engineers, 1982, p. 29. [19] A. Sala, Radiant Properties of Materials', PWN-Polish Scientific
[2] J.S. Clark, S.K. Borowski, M.C. Mcllwain and D.G. Pellaccio, Publishers, Warsaw, 1986.
NASA TM-105870, 1992. [20] P. Hrycak, AIAA J., 1 (1963) 2619.
[3] J. Cohron, Y. Ozaki, R. Zee and B. Chin, in M.S. El-Genk and [21] E.K. Storms, The Refractor), Carbides, Academic Press, New
M.D. Hoover (eds.), Proc. llth Symp. on Space Nuclear Power York, 1967.
and Propulsion, American Institute of Physics, Part 2, 1994, p. [22] P. Schwarzkopf and R. Kieffer, Refractor), Hard Metals,
939. Macmillan, New York, 1953.
[4] J. Cohron, R. Zee and B. Chin, J. Nuclear Mater., 212 215 [23] G.P. Shveikin and P.V. Gel'd, Soy. J. Non-Ferrous Metals, 34
(1994) 1488. (1961) 41.
[5] Y. Ozaki and R.H. Zee, in M.S. E1-Genk (ed.), Proe. 12th Symp. [24] H.L. Schick, Thermodynamics of Certain ReJractory Compounds,
on Space Nuclear Power and Propulsion, American Institute of Vol. II, Academic Press, New York, 1966.
Physics, Part 1, 1995, p. 371. [25] D.R. Olander, Fundamental Aspeets of Nuclear Reactor Fuel
[6] J. Liu and R.H. Zee, in M.S. EI-Genk (ed.), Proc. 12th Symp. on Elements, Technical Information Center, Office of Public Affairs,
Space Nuclear Power and Propulsion, American Institute of Energy Research and Development Administration, Springfield,
Physics, Part 1, 1995, p. 383. VA, 1976.
[7] H.P. Gao and R.H. Zee, in M.S. EI-Genk (ed.), Proe. 12th [26] H.S. Carslaw and J.C. Jaeger, Conduction of Heat in Solids',
Symp. on Space Nuclear Power and Propulsion, American Insti- Oxford University Press, London, 1959.
tute of Physics, Part 1, 1995, p. 283. [27] M.F. Modest, Radiative Heat Transfer, McGraw Hill, New
[8] Y. Ozaki and R.H. Zee, Scripta Metall. Mater., 30(10) (1994) York, 1993, p. 98.
1263. [28] N.O. Stockman and J.L. Kramer, NASA TN D-1878, 1963.
[9] W.A. Ranken and J.A. Turner, Proc. 26th Intersoc. Energy [29] B.V. Karlekar and B.T. Chao, Int. J. Heat Mass TransJer, 6
Conversion Engineering Conj'., American Nuclear Society, Vol. 1, (1963) 33.
1991, p. 425. [30] J.C. Mills, W.R. Determan and T.H. Van Hagan, in M.S.
[10] R.D. Rovang, T.R. Palamides and M.E. Hunt, Proc. 27th Inter- El-Genk and M.D. Hoover (eds.), Proc. llth Symp. on Spaee
soc. Energy Conversion Engineering Conf, American Nuclear Nuclear Power and Propulsion, American Institute of Physics,
Society, Vol. 2, 1992, p. 319. Part 2, 1994, p. 695.