Processing of solid solution, mixed uranium/refractory metal carbides for advanced space nuclear power and propulsion systems

Material Information

Processing of solid solution, mixed uranium/refractory metal carbides for advanced space nuclear power and propulsion systems
Knight, Travis Warren ( Dissertant )
Anghaie, Samim ( Thesis advisor )
Dugan, Edward T. ( Reviewer )
Hanrahan, Robert J. ( Reviewer )
Kaufman, Michael J. ( Reviewer )
Vernetson, William G. ( Reviewer )
Place of Publication:
State University System of Florida
Publication Date:
Copyright Date:


Subjects / Keywords:
Carbides ( jstor )
Carbon ( jstor )
Graphite ( jstor )
High temperature ( jstor )
Hydrogen ( jstor )
Melting points ( jstor )
Propulsion ( jstor )
Sintering ( jstor )
Solid solutions ( jstor )
Uranium ( jstor )
Dissertations, Academic -- Nuclear and Radiological Engineering -- UF ( lcsh )
Nuclear and Radiological Engineering thesis, Ph. D ( lcsh )
Space vehicles -- Propulsion systems ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
non-fiction ( marcgt )


Nuclear thermal propulsion (NTP) and space nuclear power are two enabling technologies for the manned exploration of space and the development of research outposts in space and on other planets such as Mars. Advanced carbide nuclear fuels have been proposed for application in space nuclear power and propulsion systems. This study examined the processing technologies and optimal parameters necessary to fabricate samples of single phase, solid solution, mixed uranium/refractory metal carbides. In particular, the pseudo-ternary carbide, UC-ZrC-NbC, system was examined with uranium metal mole fractions of 5% and 10% and corresponding uranium densities of 0.8 to 1.8 gU/cc. Efforts were directed to those methods that could produce simple geometry fuel elements or wafers such as those used to fabricate a Square Lattice Honeycomb (SLHC) fuel element and reactor core. Methods of cold uniaxial pressing, sintering by induction heating, and hot pressing by self-resistance heating were investigated. Solid solution, high density (low porosity) samples greater than 95% TD were processed by cold pressing at 150 MPa and sintering above 2600 K for times longer than 90 min. Some impurity oxide phases were noted in some samples attributed to residual gases in the furnace during processing. Also, some samples noted secondary phases of carbon and UC2 due to some hyperstoichiometric powder mixtures having carbon-to-metal ratios greater than one. In all, 33 mixed carbide samples were processed and analyzed with half bearing uranium as ternary carbides of UC-ZrC-NbC. Scanning electron microscopy, x-ray diffraction, and density measurements were used to characterize samples. Samples were processed from powders of the refractory mono-carbides and UC/UC2 or from powders of uranium hydride (UH3), graphite, and refractory metal carbides to produce hypostoichiometric mixed carbides. Samples processed from the constituent carbide powders and sintered at temperatures above the melting point of UC showed signs of liquid phase sintering and were shown to be largely solid solutions. Pre-compaction of mixed carbide powders prior to sintering was shown to be necessary to achieve high densities. Hypostoichiometric, samples processed at 2500 K exhibited only the initial stage of sintering and solid solution formation. Based on these findings, a suggested processing methodology is proposed for producing high density, solid solution, mixed carbide fuels. Pseudo-binary, refractory carbide samples hot pressed at 3100 K and 6 MPa showed comparable densities (approximately 85% of the theoretical value) to samples processed by cold pressing and sintering at temperatures of 2800 K. ( ,, )
KEYWORDS: space, nuclear, fuel, mixed carbides, uranium carbide
Thesis (Ph. D.)--University of Florida, 2000.
Includes bibliographical references (p. 162-166).
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
General Note:
Title from first page of PDF file.
General Note:
Document formatted into pages; contains vii, 167 p.; also contains graphics.
General Note:
Statement of Responsibility:
by Travis Warren Knight.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
45825450 ( OCLC )
002566163 ( AlephBibNum )
AMT2444 ( NOTIS )


This item has the following downloads:

Full Text







This work is dedicated to my parents, Dannis and Frances Knight,
for their love, support, and encouragement through the years.


This work was performed in the Ultra-high Temperature Materials Laboratory of the

Innovative Nuclear Space Power and Propulsion Institute (INSPI) at the University of Florida.

Support for this research came from the NASA Marshall Space Fight Center under grant

NAG8-1251 and from the Department of Defense, Ballistic Missile Defense Organization

(formerly SDIO), Innovative Science and Technology Office under contract NAS-26314,

managed by NASA Glenn Research Center through INSPI.

The author wishes to thank Dr. Robert J. Hanrahan Jr. of Los Alamos National

Laboratory for his many valuable insights into material science and laboratory research and

many other contributions and assistance in this study.

Many heartfelt thanks belong to Dr. Samim Anghaie for his kind mentorship and sage

advice. His wisdom and encouragement in matters both academic and professional have been a

true source of inspiration.

A special thanks is owed to the other members of my doctoral committee, Dr. Edward

T. Dugan, Dr. Robert J. Hanrahan, Dr. Michael J. Kaufman, and Dr. William G. Vemetson, for

their advice and comments throughout this research and the preparation of this manuscript.


A C K N O W L E D G M E N T S ................................................................................................... iii

A B ST R A C T ....................................................................................................... . ........ .. vi

IN TR O D U C TIO N ........................................................................ ......... ... .... ............... 1

M motivation and O objective ........................................... .............................................1.........
A application Fundam entals............................................ .............................................3........
N nuclear Therm al Propulsion .................................................................. ...............3......
A advanced T errestrial R actors .................................................................... ...............6...
Historical Background on Carbide Nuclear Fuel Development ..................... ..................... 8
Space Pow er and Propulsion Studies........................................................... ...............8...
Advanced Terrestrial Reactor Fuel Studies ................................................................ 12
Technical Background on Carbide Fuel Development ...................................................... 15
Melting Point and Carbon-to-Metal Ratio................................................................. 16
Processing and Fabrication ..................................................................... ............... 19
Processing by extrusion ........................................... .. ....................... ... .......... .. 19
Processing by sintering .................................................................... ................ 2 1
Processing by hot pressing......................................... ....................... ................ 25
O their processing m ethods ..................................... ........................ ................ 26
Fuel E lem ent F fracture .. ...................................................................... ..... ............... 27
Hot Hydrogen Corrosion and Mass Loss...................................................................28

M E TH O D ................................................................................. ...... ..... ..................... 34

Preparation and H handling of Pow ders.............................................................. ................ 34
C om position .................................................................................. ....................... 34
U ranium Hydride Processing .................. ............................................................. 38
M ixing and H handling of Pow ders...................................... ...................... ................ 39
P ro c e ssin g ...........................................................................................................................4 1
C old U niaxial Pressing .............................................................................. .. .............. ....... 4 1
Uniaxial pressing in a graphite die/susceptor.........................................................42
U niaxial pressing in stainless steel dies.................................................. ................ 47
Sintering B y Induction H eating......................................... ....................... ................ 48
Equipm ent design and perform ance...................................................... ................ 48


Tem perature m easurem ent and control................................................. ................ 53
Testing schedule ....................... .. ........... .....................................53
D ynam ic M agnetic Com paction....................................... ....................... ................ 54
H o t P re ssin g ................................................................................................................ .. 5 4
Equipm ent design and perform ance...................................................... ................ 54
Equipm ent redesign .... .. ................................. ............................. ............ 56
Testing schedule ....................... .. ........... .....................................60
D density M easurem ents ................................................... .............................................. 60
M elting Point D eterm nation ............................................................................... ................ 61

R E SU L T S .......................................................................................................... . ....... .. 64

B in ary C arb id e s ................................................................................................................. .. 6 4
T ernary C arbides ................................................................ ............. ............... 68
Density Measurements and Microscopy Results.........................................................68
X -ray D iffraction R results ....................................................................... ................ 72
H ot P dressing ...................................................................................................... . ........ .. 7 6

D ISC U S SIO N ............................................................................... .. . . ................... 137

B in ary C arb id es ................................................................................................................ 13 7
T ern ary C arb id es ............................................................................................................... 14 0
Microscopy Results ...................... .. .......... ............................. 140
Tim e and tem perature. ................................................................. ............... 140
Pre-compaction .............................................................. ......... ............................... 143
Solid Solution Formation and X-ray Diffraction Results...................... .................. 144
P re-C o m p action n ............................................................................................................... 15 5
Hot Pressing ........................................................................................ 155
Suggested Processing M ethodology................................... ...................... ............... 157

CONCLUSIONS AND RECOMMENDATIONS...... .... ....................................159

L IST O F R E FE R E N C E S .................................................................................................. 162

BIOGRAPH ICAL SKETCH ................................................................. ................ 167

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Travis Warren Knight

May 2000

Chairman: Samim Anghaie
Major Department: Department of Nuclear and Radiological Engineering

Nuclear thermal propulsion (NTP) and space nuclear power are two enabling

technologies for the manned exploration of space and the development of research outposts in

space and on other planets such as Mars. Advanced carbide nuclear fuels have been proposed

for application in space nuclear power and propulsion systems. This study examined the

processing technologies and optimal parameters necessary to fabricate samples of single phase,

solid solution, mixed uranium/refractory metal carbides. In particular, the pseudo-ternary

carbide, UC-ZrC-NbC, system was examined with uranium metal mole fractions of 5% and

10% and corresponding uranium densities of 0.8 to 1.8 gU/cc. Efforts were directed to those

methods that could produce simple geometry fuel elements or wafers such as those used to

fabricate a Square Lattice Honeycomb (SLHC) fuel element and reactor core.

Methods of cold uniaxial pressing, sintering by induction heating, and hot pressing by

self-resistance heating were investigated. Solid solution, high density (low porosity) samples

greater than 95% TD were processed by cold pressing at 150 MPa and sintering above 2600

K for times longer than 90 min. Some impurity oxide phases were noted in some samples

attributed to residual gases in the furnace during processing. Also, some samples noted

secondary phases of carbon and UC2 due to some hyperstoichiometric powder mixtures having

carbon-to-metal ratios greater than one.

In all, 33 mixed carbide samples were processed and analyzed with half bearing

uranium as ternary carbides of UC-ZrC-NbC. Scanning electron microscopy, x-ray diffraction,

and density measurements were used to characterize samples. Samples were processed from

powders of the refractory mono-carbides and UC/UC2 or from powders of uranium hydride

(UH3), graphite, and refractory metal carbides to produce hypostoichiometric mixed carbides.

Samples processed from the constituent carbide powders and sintered at temperatures above

the melting point of UC showed signs of liquid phase sintering and were shown to be largely

solid solutions. Pre-compaction of mixed carbide powders prior to sintering was shown to be

necessary to achieve high densities. Hypostoichiometric, samples processed at 2500 K

exhibited only the initial stage of sintering and solid solution formation. Based on these findings,

a suggested processing methodology is proposed for producing high density, solid solution,

mixed carbide fuels.

Pseudo-binary, refractory carbide samples hot pressed at 3100 K and 6 MPa showed

comparable densities (approximately 85% of the theoretical value) to samples processed by

cold pressing and sintering at temperatures of 2800 K.


Motivation and Objective

Because of its high performance potential, nuclear thermal propulsion (NTP) could be

utilized for manned missions and cargo transport to the moon or Mars, unmanned explorations

of the outer planets, and earth orbit transfers of satellites or other space-based assets. The

Rover/NERVA programs, a joint effort between NASA and the Atomic Energy Commission,

ran from 1955 to 1973 during which four major fuel types were tested (Bennett et al., 1994;

Davidson, 1991; Taub, 1975). The last advanced nuclear fuel considered before the program

was cancelled in 1973 was a solid solution, mixed carbide, (U, Zr)C (Lyon, 1973). Other

advanced fuels for terrestrial reactors have been tested including pyrolytic carbon coated

microspheres of UC2 embedded in a graphite matrix, graphite coated spheres, and mixed

carbides such as (U, Pu)C and (U, Th)C2. These advanced fuels have been studied with

particular interest for fast breeder reactor programs because of the higher heavy-metal atom

density in carbides over oxide fuels leading to shorter doubling times. Observed high melting

point, thermochemical stability, and high thermal conductivity of single phase, solid-solution

mixed uranium/refractory metal carbides such as the pseudo-ternary carbide, (U, Zr, Nb)C,

portend their usefulness as an advanced fuel for next generation terrestrial as well as space

reactor design applications. This study was undertaken to develop and optimize processing

techniques for producing high density, solid solutions of the mixed carbide, (U, Zr, Nb)C,

containing five to ten mole percent uranium carbide (UC).

Mixed carbide fuels have several advantages over the more widely studied oxide fuels

used in virtually all nuclear reactors today. Perhaps chief among these advantages is their higher

thermal conductivity, which can approach that of metallic uranium. This higher thermal

conductivity lowers the fuel centerline temperature permitting a higher linear heat generation rate

and larger diameter fuel rod which decreases fabrication costs. Savings are also obtained

through reduced emergency cooling requirements due to less thermal energy being stored in the

core at any given time. Mixed carbides have the advantage of increasing the melting point of

UC with melting points for typical compositions greater than 3600 K making them even more

desirable for high temperature applications.

Furthermore, the higher uranium density of carbide fuels permits the design of more

compact reactor cores. While high temperature, compact cores are the domain of space-based

reactors, design studies utilizing mixed carbide fuel for terrestrial reactors have revealed

additional cost savings possible through smaller reactor vessels and containment buildings.

Mixed carbides of uranium and either thorium or plutonium have been investigated as a fuel for

fast breeder reactors enabling shorter doubling times (Matske, 1986). These and additional

issues have been discussed in several monographs on the subject (Matzke, 1986; Holden,


Despite these many benefits, some concerns regarding carbide fuels include

compatibility issues with coolant and/or cladding materials. Uranium carbide is compatible with

sodium up to 1143 K and with helium at all temperatures. By alloying with refractory metal

carbides, the resistance of UC to attack by water can be increased (Holden, 1966). Because

of their improved thermochemical stability in a hot-hydrogen environment over graphite matrix

fuels, mixed carbide fuels have been more widely investigated for potential space nuclear power

and propulsion applications. Their projected endurance at very high temperatures far exceeds

that of fuels previously tested and signifies their potential as a fuel for increased performance

characteristics (i.e. higher specific impulse and/or longer lifetime, etc.). However, insufficient

test data exist to fully evaluate their performance under conditions required for NTP such as

temperature and hot hydrogen environment. Further, previous studies of the (U, Zr, Nb)C

system did not clearly define carbon-to-metal (C/M) ratios. Studies by Carmack (1991) and

Czechowicz et al. (1991) have shown that the C/M ratio greatly affects the melting point and

performance of carbides. Any study attempting to characterize mixed carbide fuels for high

temperature applications must include this determination.

Application Fundamentals

Nuclear Thermal Propulsion

Because of their high temperature, high radiation, and hot hydrogen environment, high

performance space nuclear reactors for power and/or propulsion present a unique and

challenging set of materials engineering requirements. To understand the origin of these

requirements and the motivation for studying mixed carbide fuels, it is instructive to examine the

factors that contribute to nuclear rocket performance. Specific impulse (Isp), also called

specific thrust, is used to measure performance and is defined as thrust divided by propellant

mass flow rate (see Eq. 1).

ISP thrust / T
Isp = ------oc -
( propellant A MW
w mass flow rate)
where (1)
T =temperature of reactor core
MW = molecular weight of propellant

A nuclear thermal rocket operates by the same basic principles as chemical rockets--

namely the expansion of hot gas (propellant) through a rocket nozzle to provide thrust. As

shown in Figure 1-1, the propellant flows through coolant channels of the solid-fuel reactor core

where it is heated to very high temperatures (>3000 K proposed for pseudo-ternary carbides).

To achieve high performance (as measured by Eq. 1), the fuel is required to operate at very high

temperatures. Hydrogen has been used as a propellant during all rocket reactor tests and is

preferred because it has the lowest molecular weight. However, hot hydrogen can react with

the fuel resulting in corrosion and mass loss. Furthermore, mission cost constraints require a

compact, lightweight reactor necessitating high power densities (high neutron flux) with

associated radiation damage and increased susceptibility of the fuel elements to fracture.

High operating temperatures are also desirable for waste heat rejection by space

nuclear power reactors. The energy rejected per unit area by a radiator is proportional to the

fourth power of its temperature as shown in equation 2 (Angelo and Buden, 1985). Further,

because it is desirable to minimize the mass of a space power system due to high launch costs,

the mass and therefore the area of any radiator is at a premium. Higher operating temperatures

possible with mixed uranium/refractory metal carbides allow for greater waste heat rejection per

unit radiator area helping to minimize payload. Of course, to reject heat at these temperatures,

the radiator material must have good performance characteristics at high temperatures.


a) Nozze


Figure 1-1. Nuclear thermal rocket engine a) drawing showing propellant flow (after
Koenig, 1986) b) photograph of nuclear rocket engines tested in the Rover/NERVA

However, their study is outside the scope of this work, which was focused on the development

of the high temperature nuclear fuels alone.

E =aT4
e =emissivity of the radiator (2)
a = Stefan Boltzman constant, 5.67E 8 W m-2 K-4
T = temperature of reactor core (K)

Therefore, mixed carbides make possible smaller payloads reducing launch costs in

several ways. Higher melting points lead to smaller radiators and associated hardware

necessary for space power generation. Compact reactor cores for power and/or propulsion

are made possible through higher thermal conductivity and higher heavy-metal atom density. To

achieve these goals, their performance must be qualified against all the deleterious effects

concomitant with high performance--namely high temperature, high neutron flux, and hot

hydrogen environment.

Advanced Terrestrial Reactors

Carbide fuels have a number of advantages over oxide-based fuels for fast breeder

reactor applications by increasing breeding ratios and shortening doubling times. Higher

breeding ratios, the amount of fissile fuel produced over the amount destroyed, leads to higher

fissile material gains (Fgain) per cycle. Equation 3 shows the relationship of reactor doubling time

(RDT), the time it takes to produce enough fissile material to fuel an additional reactor, to Fgain

and FBOC, the amount of fuel at the beginning of a cycle or fissile inventory. The higher thermal

conductivity of carbides allows for higher linear heat generation rates (LHGR) for more

compact cores and decreased fissile inventory (FBOC) leading to shorter RDT (see Equation 3).

RT gon fuelcvcle
vear 1 (3)

From a safety perspective, high thermal conductivity also decreases the thermal energy

stored in the fuel that must be contended with in an accident/transient scenario. This reduces the

emergency core cooling requirements and therefore the cost of such systems. Also, the higher

thermal conductivity permits larger diameter fuel pins which are not only less expensive to

fabricate but result in fewer number of pins and therefore less associated cladding and

hardware. This leads to a decrease in parasitic captures (increase in Fgain). Similarly, higher

heavy-metal atom densities and higher metal to non-metal atom ratios also serve to increase

Fgain. Having proportionately more fuel in the core also serves to harden the spectrum

increasing the number of neutrons released in fission per neutron absorbed in fissile material (r1)

and therefore Fgain (Harry, 1983; Matzke, 1986).

Advantages of carbide fuels for other advanced reactor types derive mainly from the

benefits of higher thermal conductivity and higher heavy-metal atom densities through higher

bumup, higher LHGR, lower fissile inventories (smaller cores) and therefore decreased costs.

Historical Background on Carbide Nuclear Fuel Development

Space Power and Propulsion Studies

The processing and testing of solid solution, mixed uranium/refractory metal carbide

fuels have been conducted in both the U.S. and former Soviet Union (also later in Russia) with

particular interest for their application to space nuclear power and propulsion. Solid solution,

pseudo-binary carbides--namely (U, Zr)C--were the last in a series of fuel designs investigated

during the Rover/NERVA (nuclear engine for rocket vehicle applications) programs. It was

demonstrated that the solid solution uranium-zirconium carbide fuel elements have the potential

for higher service temperatures than any other candidate fuel types.

The Rover/NERVA programs, a joint effort between NASA and the Atomic Energy

Commission, ran from 1955 to 1973 during which four major fuel types were tested as depicted

in Figure 1-2. These fuel types were: 1) UC2 particles dispersed in graphite 2) UO2 and

eventually UC2 particles with a pyrolytic-carbon (PyC) coating dispersed in a graphite matrix,

3) A composite of graphite/(U, Zr) C with the carbide forming a continuous webbed structure,

and 4) Solid-solution (U, Zr) C. All fuel elements except the solid solution carbide type had a

protective zirconium carbide coating. The latter two fuel types were tested in the Nuclear

Furnace 1 (NF-1) Test Reactor at Jackass Flats, Nevada in 1972 (Lyon, 1973). Further large

scale, in-core testing of fuel elements was cancelled in 1973 with the termination of the

Rover/NERVA program. Despite being judged a technical success, the program was cancelled

due to changing national priorities (Angelo and Buden, 1985).

CVD Coaling M 6 -
0 C UC2 Particle 0 Substrate 0

.-- PyC Coaling +
(~) Grapfirp i

-ILC2 Sphee F l %
---o___ c p ___ _-- _

aj UC2 POrticles Dispersed b) PyC Coated UC2 Spheres

Figure 1-2. Microstructure of various NERVA fuels (after Matthews et al., 1991).
inBecause of the limited number of solid-solution carbide fuel elements tesGraphted (28 fuel

elements in two reactor cells), its usefulness as a fuel for nuclear thermal propulsion (NTP)
Hy/drogen [,, Z otig Hdrogen

%L1 Xr XCoating


z[U,Zr]C .

c) Composile Fuel d) Solid Solution (U,Zr)C

Figure 1-2. Microstructure of various NERVA fuels (after Matthews et al., 1991).

Because of the limited number of solid-solution carbide fuel elements tested (28 fuel
elements in two reactor cells), its usefulness as a fuel for nuclear thermal propulsion (N'TP)

could not be fully evaluated. However, depending on the required operational lifetime, the solid
solution carbides are expected to operate for short periods with propellant exit temperatures as


RV p -W

high as 3200 K and many hours at lower temperatures (2600 3000 K) as shown in Figure 1-

3 (Lyon, 1973; Koenig, 1986). More recently, studies and modeling by Storms (1992) have

shown that mass loss due to vaporization will dominate. This was shown to be the life limiting

phenomenon near the reactor core's exit where surface temperatures exceed 2900 K. Based

on the thermochemical analysis of vaporization behavior of the Zr-U-C and Nb-U-C systems,

Butt et al. (1993) fit the predicted mass loss rates with the following two equations as a function

of fuel temperature (see Equations 4a and 4b).

(Nbo85,Uo15)C, log io(Ri) =-18.313+4.295x10-3T (Butt etal., 1993) (4a)

(Zro9,Uo1)C, log 1o0(R2)

Carbide fuel e;
mass loss ratf

0S Cor
0 4
i li

16.882+3.97x10-3T (Butt etal., 1993)


0 I I I I1
2200 2400 2600 2800 3000 3200
temperature (K)

Figure 1-3. Performance of NERVA tested nuclear fuels and expected performance of carbide
fuels at the conclusion of the Rover/NERVA programs (after Koenig, 1986). Curves A and B
are also included as more recent estimates of pseudo-binary carbide fuel performance
calculated based on modeling from Butt et al. (1993).

From equations 4a and 4b, the projected endurance of these pseudo-binary carbide

fuels can be calculated. Curves A and B in Figure 1-3 were calculated based on the fuel

geometry of the solid solution, all-carbide fuel elements used in the NF-1 experiments (see

Figure 1-4) and assuming an arbitrary limit of 10% mass loss for the core. It should be pointed

out that the estimates calculated in this manner are conservative since they assume a uniform

maximum operating temperature and hence a uniform mass loss rate for the entire core. In

reality, the peak temperature will occur near the exit of the core where the mass loss rates by

vaporization will be greatest.

With the interest generated as a result of the Space Exploration Initiative (SEI), other

studies of mixed carbides were reported in the early 1990s. Experimental work reported by

Czechowicz et al. (1991) examined the pseudo-binary, UC-ZrC, system. Storms (1992)

contributed a great amount of work on the properties of refractory carbides and more recently

reported on some thermochemical modeling of mass loss from (U, Zr)C fuels in flowing hot

hydrogen. Carmack (1991) investigated the processing of refractory monocarbides and mixed

carbides of Ta, Hf, and U and also reported on the melting points of some monocarbides.

Wang et al. (1994) produced samples of (U, Zr)C using a self-propagating high-temperature

synthesis method.

From 1993 to 1997, a collaborative effort was established between the Innovative

Nuclear Space Power and Propulsion Institute (INSPI) at the University of Florida and the

Russian Scientific Research Institute, LUTCH. The goal of this collaboration was to verify

Russian data and research on pseudo-ternary carbide nuclear fuels for NTP, which were

carried out in the former Soviet Union between 1978 and 1988. Testing in that period reported

hydrogen exit-gas temperatures of 2800-3300 K with power densities as high as 20 MW/liter

and uranium mass loss estimates as low as 0.5 to 1.0 % (based on reactivity loss

measurements). However, vital information on the post-test condition of fuel elements was

lacking since no post-test analysis was conducted. Collaborative efforts between INSPI and

LUTCH were aimed at verifying these results and conducting post-test analysis of the fuel

(D'yakov and Tishchenko, 1994; Diaz, 1994; Knight, 1999).

Advanced Terrestrial Reactor Fuel Studies

A number of advanced nuclear fuel studies have been conducted using carbide-based

fuels with most efforts directed toward fast breeder reactor programs (see Table 1-1). Mixed

carbide nuclear fuels of (Uo.3,Puo.7)C have been used in the Indian Fast Breeder Test Reactor

(FBTR) operated from 1985 to the present. The higher plutonium content was chosen because

natural uranium could be used. For these higher plutonium contents, the carbide fuel is not only

advantageous for the above reasons but is necessary because early investigations showed that

(U0.24, Pu0.76)02 is not compatible with the sodium coolant (Ganguly et al., 1986). Experiments

conducted by Los Alamos National Laboratory in the Experimental Breeder Reactor II (EBR

II) used test assemblies with (U0.76, Pu0.24)C. Helium-bonded test assemblies achieved peak

bumups as high as 20.7 at% (192 MWd/kg) without failure. Sodium-bonded fuel pins achieved

peak bumups of 15.8 at% (146 MWd/kg) before failure. Both sodium and helium-bonded,

peak bumup fuel pins had 316 stainless steel cladding (Harry, 1983; Herbst and Matthews,


Other potential applications of carbide fuels that have been investigated include

applications in high temperature gas-cooled reactors (HTGR) and in pebble-bed reactors

Table 1-1. Experience with advanced carbide nuclear fuels for terrestrial reactors.

Reactor/ Country Oper- Fuel Comments Reference
Study ation

WR-1 Manitoba, 1965 to Cast enriched Organically cooled, (Matzke, 1986)
(Whiteshell Canada 1985 UC slugs, Zr heavy-water
Reactor) alloy sheath moderated; 60 MWt;
10 MWd/kg; few
defects noted in fuel

Dragon UK 1966 to (U, Th)C2 HTGR; 100 MWd/kg (Matzke, 1986;
1973 Lung, 1996)

Peach US 1967 to (U, Th)C2 HTGR; 40 MWe (El-Wakil, 1982;
Bottom 1974 Agnew, 1981;
Atomic Matzke, 1986)

AVR JYilich, 1967 to 20% enriched Pebble-bed reactor; 13 (El-Wakil, 1982;
Germany 1980s UC2 graphite MWe; He cooled, Agnew, 1981;
coated; ThC2 T=1223 K; 100 Matzke, 1986;
(fertile) MWd/kg Lung, 1996)

BR-10 USSR 1973 carbide FBR; 5 at% bumup; (Matzke, 1986)
switched to non-
carbide fuel

Fort St. US 1979 to UC2, Th2, coated HTGR; 330 MWe; (El-Wakil, 1982;
Vrain 1989 microspheres predicted 100 MWd/kg Agnew, 1981;
in graphite before shutdown in Matzke, 1986)
matrix 1989

FBTR Kalpakkam, 1985 to (U03, Pu07)C FBR; 42.5 MWt, 12.5 (Matzke, 1986)
India present MWe

EBRII test US 1974 to (UO 24, Pu0 76)C Advanced LMFBR (Herbst and
fuel 1980s study; 192 MWd/kg Matthews,
assemblies achieved; 120 1982; Harry,
MWd/kg, 100 kW/m, 1983)

a) (U, ZrK)/Graphile Composite b}) (U, r

005k0 0

(U, Zr)C/Graphite Composite Matrix
Coolant Channaels
19 rmm

c) Russian (U, Zr, Nb)C or (U, Zr, Ta)C Twisted Ribbon

2.2 mm

d) Square-Lartiec Honeycomb


I- IFa1 r

Figure 1-4: Various designs of nuclear fuel elements, a) Rover/NERVA NF-1
carbide/graphite composite fuel element (Lyon, 1973) b) Rover/NERVA, NF-1 carbide
fuel element (Lyon, 1973) c) Russian twisted ribbon carbide fuel element (D'yakov and
Tishchenko, 1994) d) Square-lattice Honeycomb fuel wafers, grid assembly, shroud, and
reactor core (Furman, 1999; Widargo, 1999)

(PBR). An example of the latter reactor type was the German Arbeitsgemeinschaft

Versuchsreaktor (AVR) that operated from 1967 to the 1980s. This reactor operated using

spherical graphite fuel pellets 2.36 in. in diameter with a center containing 3.5 g of 20% enriched

UC and graphite (El-Wakil, 1982). Burnups as high as 100 MWd/kg were achieved. Fuel

rods made of 600 pm diameter microspheres embedded in graphite were used in HTGRs such

as Peach Bottom and Fort St. Vrain. These microspheres had a UC2 or (U, Th)C2 nuclear fuel

center ~200um in diameter. Surrounding the fuel is a buffer layer of carbon to limit swelling by

accommodating fission gasses (El-Wakil, 1982; Kneif, 1992). A layer of pyrolytic graphite was

used to help contain the migration of fission products while a silicon carbide layer provides

strength. A final PyC layer protects the more brittle SiC coating (Agnew, 1981; Kneif, 1992).

The "amoeba effect", a problem noted in coated fuel particles, occurs when the centerline

temperature of the particle exceeds 1873 K. The result is a migration of the central carbide fuel

particle in the direction of the temperature gradient breaching the coating that provides the

fission product barrier (Matzke, 1985). An average fuel burnup of 100 MWd/kg was expected

for the Fort St. Vrain reactor before it was shutdown in 1989 due to economic factors (El-

Wakil, 1982; PSCC, 1995).

Technical Background on Carbide Fuel Development

Several studies are discussed in the following sections that illustrate four major factors to

be considered in the development of carbide nuclear fuels. These factors include: 1) controlling

microstructure and carbon-to-metal (C/M) ratio to prevent the formation of a second phase, 2)

difficulties associated with fabricating carbide fuel elements, 3) fracture problems during

operation, and 4) corrosion by the hot hydrogen propellant. Each of these factors is discussed

below as they relate to the development of pseudo-ternary carbide nuclear fuels.

Melting Point and Carbon-to-Metal Ratio

The highest melting point for the monocarbides of U, Zr, Nb, Ta, and W occurs for

congruent melting in the single phase, solid solution region of these nonstoichiometric

intermediate phases with C/M typically between 0.75 and 1.0. The congruent melting points of

several refractory monocarbides are listed below in Table 1-2 along with their corresponding

carbon-to-metal ratio (C/M). Similarly, the single phase, solid solution regions for pseudo-

binary and pseudo-ternary carbides lie within a narrow range of C/M values less than one. The

C/M ratio had to be carefully adjusted in the Rover/NERVA, NF-1 test program to prevent the

formation of a second phase, carbon, which drastically lowers the melting point. A C/M ratio of

0.88 to 0.95 was targeted for NF-1, (U, Zr)C fuel elements for a proposed maximum operating

temperature of 3200 K (Lyon, 1973). Outside the range of C/M for single phase, solid

solutions, these carbides experience eutectic melting at far lower temperatures (Butt et al.,

1993). However, it is desirable to produce fuel in the upper range of the C/M ratio due to the

high initial carbon mass losses during operation (Butt et al., 1993). More regarding carbon

mass loss will be discussed in later sections.

Along with other mixed refractory carbides, Tosdale (1967) also investigated the

pseudo-binary (U, Zr)C and the pseudo-ternary, (U, Zr, Nb)C. This study reported improved

oxidation resistance and higher melting points for ternary carbide mixtures with UC than for

binary carbide mixtures of NbC or ZrC with an equal amount of UC. For low uranium content

fuels (0.05 to 0.1 U/M), a maximum in the solidus temperature was observed to fall between a

zirconium to refractory metal ratio of 0.65 to 0.85. Unfortunately, no C/M determination was

reported for samples in this study. However, studies of stoichiometric (U, Zr)C by Czechowicz

et al. (1991) revealed the development of a second phase, carbon, in equilibrium with the solid-

solution (U, Zr)C. The melting temperature of the supposed single-phase, solid solution (U,

Zr)C from Tosdale (1967) is 100 K to 700 K higher (depending on the uranium content) than

for eutectic compositions, (U, Zr)Cx + C (Czechowicz et al., 1991). Figure 1-5 compares the

solidus temperatures for solid solution (U, Zr)C and eutectic (U, Zr)Cx+C. Butt et al. (1993)

noted that the addition of 10 at% uranium can be expected to lower the melting point by 200 to

500 K depending on C/M (Butt et al., 1993).

Table 1-2. Important Data on Some Refractory Carbides.

Binary alloy Melting point, (C/M Ratio)t Lattice Parameter (nm)

NbC 3873 K + 25 K, (0.79) 0.4469

TaC 4258 K, (0.89) 0.4454

UC 2803 K, (1.00) 0.49605*

WC 3058 K (0.61) 0.422**

ZrC 3813 K, (0.87) 0.4697

t(Massalski, 1986), (Weimer, 1997), *(Matthews et al., 1994), **(Storms, 1967)

Studies by Accary et al. produced uranium monocarbide by the decomposition of UH3

mixed with graphite. This study noted the very small size and flake-like shape ofUH3 particles.

Samples produced from these powders achieved high densities following sintering but showed

more dependence on cold pressing pressure than spherical uranium metal particles produced

from the calcium reduction ofUO2 (Accary and Caillat, 1961). Uranium metal hydride can be

produced relatively easily from uranium metal using a glove box. The annealed metal with a

clean surface is heated to between 423 to 473 K in a glove box under an atmosphere of Ar-

7%H. The hydride (UH3) will form on the surface producing small flakes that fall off exposing

more uranium metal to continue the reaction. The resultant UH3 powder can then be used as a

feed material for processing mixed carbides and controlling the C/M ratio to produce

hypostoichiometric mixed carbides.



Solid Solution (U, Zr)C
S ~ (Tosdale)


2800 ., (U, Zr)Cx+ C


0 20 40 60 80 100

UCI.00 Composition ZrC (mole %) ZrC .00

Figure 1-5. Comparison of solidus curves for solid solution (U, Zr)C and eutectic (U, Zr)Cx+C
(after Czechowicz et al., 1991).

Processing and Fabrication

Processing by extrusion

In reporting on the production ofUC, Accary and Caillat (1961) briefly touch on some

of the benefits of the extrusion process when using a mixture of uranium and graphite powders.

In general, the reaction is fairly complete owing likely to the high pressure and resultant

deformation of particles during extrusion. This provides an intimate contact between the various

particles and a disruption of the oxide layer on uranium particles, which promotes the

carburization process. However, complete densification was not achieved via extrusion.

Fabrication of solid solution (U, Zr)C fuel elements for the Rover/NERVA program

was accomplished in several steps beginning with the extrusion of a mixture of ZrC, U02, ZrO2,

graphite flour, and a binder, Varcum 8251. A long heat treatment process followed as shown

in Table 1-3. The free carbon remaining was removed by leaching with flowing hot hydrogen at

2200 to 2300 K for 40 to 60 hours. Fuel elements were then impregnated with zirconium to

varying degrees with overall mass gains of (0%, 3%, or 8%) using a chemical vapor deposition

process at -1900 K to produce carbide fuel elements that were hypostoichiometric in carbon.

A final heat treatment of two hours at 2800 K was applied to hypostoichiometric fuel elements.

A lack of graphite in the mixture made it difficult to extrude these elements and

produced severe wear on the dies (Lyon, 1973). The 19 mm wide, hexagonal Rover/NERVA

fuel element with its 19 coolant channels (2.3 mm diameter) could not be fabricated due to these

difficulties. Instead, the carbide elements had to be manufactured in a cylindrical form with a

single 3.2 mm diameter coolant channel and machined to a hexagonal geometry 5.5 mm wide

and 0.64 m long (see Figure 1-4). These combined processing steps produced fuel elements

with porosities of 18 to 23% and C/M ratios of 0.95 to 0.98.

Table 1-3. Heat treatment procedure of Rover/NERVA fuel elements.

Length of Time Temperature Range Comment

60 hours 325-385 K Electrically heated circulated-air ovens

10 hours 385 405 K

15-20 hours 405 525 K Decomposition of binder

54 hours 1125 K -10 torr argon flush

Moved to vertical induction furnaces, argon
2.5 hours Heat up to 1875 K a

3 hours 1875 2625 K Carbothermic reduction of U02 and ZrO2

0.5 hour Hold at 2625 K

3 hours Heat up and hold 2625 K

3.5 hours Hold at 2625 K Complete sold solution formation below UC2+C
3.5 hours Hold at 2625 K e 2 K
eutectic -2725 K

2 hours Increase and hold
between 2775 2875 K

Fabrication details of the Russian fuel elements are not reported, although it is generally

accepted that this is accomplished via an extrusion process. The geometry of the fuel elements

is different than that used in the US taking on the form of a "twisted ribbon" with individual pins

2 mm thick, 4 mm wide, and 35-100 mm long. Twisted ribbons are bundled together to form

fuel elements and stacked vertically to form a fuel assembly (D'yakov and Tishchenko, 1994).

Because of their segmented lengths, axial variation in enrichment of the fuel assembly is possible.

Russian cores also incorporate an axial reflector or "after-burner" region, which allows for the

use of reduced uranium content of the pseudo-temary carbides (uranium density of -0.8 g/cc).

A low uranium content is desired not only for higher melting temperatures as shown above but

also for improved thermochemical stability in the hot hydrogen environment encountered in NTP

systems (D'yakov and Tishchenko, 1994).

Processing by sintering

Processing of mixed carbides by sintering can follow many different routes with options

that include sintering with a chemical reaction, pre-sinter compaction, etc. For example, mixed

carbide powders can be compressed prior to sintering to increase the initial density by cold

pressing with a punch and die, by pulsed magnetic compaction, or some other method. Also,

the starting material can be varied from the constituent carbide powders sinteringg without a

chemical reaction) to powder mixtures selected for a particular chemical reaction (reaction

sintering) to produce mixed carbides. Some examples include the carbothermic reduction of the

mixed metal oxides with graphite or carburization of metal powders either directly or metal

powders produced from the decomposition of their hydride form. For sufficiently exothermic

reactions, a self-propagating, high-temperature synthesis method is possible that very rapidly

consumes the reactive starting material. In some cases, the resultant material may not be

acceptable due to porosity or impurities and a comminution step is required grinding the initial

carbide clinker forming the carbide powders for further processing and fabrication steps.

Storms (1967) noted that the production of hypostoichiometric carbides from the carbothermic

reduction of the oxide with graphite leads to unacceptable oxygen content. A high vacuum

anneal should be applied to carbides produced in this method.

The procedure used by Carmack (1991) for producing mixed carbide samples

consisted of cold pressing at 55 MPa the refractory metal carbide and hydride powders with

U02 powder and a binder. Metal hydride powders were added to produce samples that were

hypostoichiometric in carbon. Samples were then sintered in a vacuum resistance furnace with

tungsten elements following a program that first took the samples to 573 K to allow the binder

to vaporize. Next, the furnace was incrementally raised to 1773 K for one hour to allow the

hydrogen gas evolved from the metal hydrides to escape. Finally, the samples were raised to

2573 K at the rate of 100 K/hr and held for one to two hours for sintering. As a binder to

provide green strength for the pressed samples prior to sintering, Carmack used 3.0 wt%

polyethylene oxide and 1.5 wt% oleic acid dissolved in alcohol. Earlier attempts to use 1.0

wt% stearic acid did not provide adequate green strength for the pressed samples.

A comparison of the various processing parameters associated with sintering was made

by Accary and Caillat (1961) for the reaction sintering of uranium and graphite powders to

produce UC. An evaluation of cold pressing pressure revealed a critical value between 200

and 350 MPa above which there was no benefit or an actual lowering of the compact's final

density following sintering. Near full density could be achieved without any cold pressing by

sintering in a vacuum furnace above 1473 K which is greater than the melting point of uranium.

The heatup rate was found to be important with high heating rates causing the reaction to

proceed too rapidly producing a more porous material. Variation of both sintering time and

temperature saw little difference in final density. However, longer sintering times and higher

temperatures resulted in a more complete reaction of the starting materials as measured by the

percentage of carbon reacted.

Mixed uranium/plutonium carbide fuel elements were produced for the Advanced Fuels

Program for fast breeder reactors (Gutierrez and Herbst, 1980). These fuel types included

single phase (Uo.8, Puo.2)C fabricated at 87% TD and two-phase (U, Pu)C+10 vol% (U,

Pu)2C3 fabricated at 98%, 87%, and 81% of TD. A vacuum carbothermic reduction process

with U02, PuO2, and C (graphite) followed by grinding, cold pressing at approximately 70 to

200 MPa and sintering between 1723 and 2073 K for times up to nine hours depending on the

desired density. For the high density (98% TD), two-phase carbide, a sintering aid, nickel, was

added to shorten the sintering period. To process the single phase (U, Pu)C, an additional step

had to be added to remove excess carbon. Following grinding, the carbide powders were

subjected to a hydrogen treatment at 1173 K reacting with excess carbon to form methane,

CH4, gas. The subsequent carbide powders are very reactive and had to be guarded against

exposure to oxygen or moisture in concentrations greater than 10 ppm.

Dynamic magnetic compaction (DMC) was developed as an alternative to conventional

powder metallurgy (PM) for producing full-density, net-shape parts (Chelluri and Barber,

1999). The DMC process involves filling an electrically conductive container referred to as the

armature with the powders to be consolidated. A central spindle (die) can be placed in the

center of the armature prior to filling to give internal features to the part as the powders are

pressed against the die. The armature containing the powders is placed in a high-field

electromagnetic coil to which a high current pulse is applied. A magnetic field is produced in the

coil which induces currents in the armature. These opposite currents create magnetic forces that

repel each other and press the armature into the powder with a large force providing the

compaction (see Figure 1-6). The entire process occurs in less than a millisecond. Full-density,

net-shape powder consolidation has been reported using this method with various metal

powders. High green strengths are achieved and can eliminate or reduce the sintering time

required for material processing (Chelluri and Barber, 1999).

2. Pulse Coil
(high current)
S3. Remove
1. Insert Sample

Figure 1-6. Illustration of the dynamic magnetic compaction process.

Wang et al. (1994) produced samples of (U, Zr)C using a self-propagating, high-

temperature synthesis (SHS) method. A form of reaction sintering, the SHS method was used

to produce solid solution (U, Zr)C from powders of U, UC2, Zr, and graphite. The exothermic

nature of the reaction leads to a combustion wave consuming the reactants. Wang et al.

observed the onset of the reaction was observed to occur near the melting point of uranium with

the liquid providing for higher diffusion rates thus fueling higher reaction rates.

Wang et al. (1994) also found the initial heating rate of the reactants to be important.

Lower heat up rates provide greater time for solid state reactions forming intermediate products

that act as a diffusion barrier for the reactants. This reduces the driving force and delays or

prevents the onset of the combustion reaction. This method has the advantage of a high reaction

rate and a lower energy requirement for sintering since the heat of the reaction is used to drive

the further combustion of the reactants. A possible drawback includes the relatively porous

nature of the final product especially for very high reaction rates (Haggerty, 1991). This porous

nature was shown in the study of (U, Zr)C by Wang et al.(1994). Further, incomplete reactions

were noted with samples formed from initial reactants involving UC2 instead of uranium metal

because of its higher melting point.

Processing by hot pressing

Hot pressing has some advantages over sintering alone without applied pressure. Hot

pressing is typically conducted at temperatures of approximately half the melting point of the

material, which is less than for sintering operations which are usually performed at three-fourths

the melting point. Also reduced is the time at which this temperature is maintained because of

the accelerated nature of sintering due to the simultaneous application of pressure and high

temperature. The application of pressure allows for more contact between particles through

rearrangement and through increased stress or on particles at their contact area, which is where

sintering occurs and increases the energy available for sintering. These factors of lower

processing temperature and shorter times serves to reduce grain growth and usually leading to

greater compact strength (Richerson, 1992).

Fischer (1964) investigated hot pressing of mixed carbides of (Tao.8, Hf0.2)C and (Tao.8,

Zro.2)C. Hot pressing pressures between 7 and 50 MPa and temperatures of 2200 to 3000 K

were investigated. Heating of the graphite die and sample was accomplished using a 50 kW,

9.6 kHz electromagnetic induction furnace with a coil diameter of 19 cm (7.5 in.). Full density

was achieved for (Tao.8, Hf0.2)C after 15 minutes at 48 MPa and 2800 K. Similarly, 96% TD

for (Tao.8, Zro.2)C was achieved after 15 minutes at 41 MPa and 2800 K. The inability to

achieve 100% TD with (Tao.8, Zro.2)C was attributed to the starting material's larger particle size

of 2.8 rim, nearly twice the size of (Tao.8, Hf.2)C particles.

Accary and Caillat (1961) produced UC at nearly 98% TD by hot pressing uranium

and graphite powders. Pressures between 25-30 MPa were used with a double punch floating

graphite die, while a molybdenum cylinder placed over the graphite die was used to contain the

lateral pressure on the die. A minimum sintering temperature of 1123 K was noted to produce

samples that had some mechanical strength. These lower processing temperatures would be

expected for UC than for the mixed refractory carbides studied by Fischer (1964) due to the

much lower melting point of UC (-2800 K) vs. the greater than 4100 K melting point of the

aforementioned mixed refractory carbides.

Other processing methods

Other techniques have been applied to the processing and fabrication of carbide fuels.

For example, various methods involving freeze drying, sol-gel techniques, or uranium loaded

resins in a fluidized-bed furnace have been used to produce carbide microspheres (Matthews et

al., 1994; Stinton et al., 1979; Zaitzev, 1994). Such microspheres are or were used in particle

bed or dispersed fuel designs such as the graphite matrix fuel rods for HTGRs and early fuel

elements for the Rover/NERVA programs or refractory-metal matrix fuels such as the cermet

(U, Zr)CN-W. However, these particle fuel techniques were not pursued since they do not

conform to the solid solution, all-carbide fuel forms previously investigated for NTP reactors

and proposed by this work.

Fuel Element Fracture

The problem of low thermal-stress resistance among all-carbide fuel elements was

anticipated in the Rover/NERVA programs. It was in part due to this fact that the geometry of

the fuel element was changed to be thinner with a single coolant channel to minimize thermal

gradients. Other likely contributors to fuel element fracture include surface flaws introduced

during the extensive machining operations performed on solid solution carbide fuel elements

because they could not be extruded in the desired geometry. From the post-test analysis of

NF-1 fuel elements, the chief problem exhibited by the solid solution carbides was fuel element

fracture. The number of fractures and the fracture pattern could be correlated to some degree

with the amount of zirconium that was added to control the C/M ratio. Fuel elements with the

highest amount of zirconium added (lowest C/M ratio) showed fewer cracks both in the

transverse and longitudinal direction. For (U, Zr)C fuel elements, the greatest degree of fracture

occurred at the axial region corresponding to the highest power densities (highest neutron flux).

However, it was noted that no sign of millimeter sized fragments were seen (Lyon, 1973).

Component studies performed by LUTCH measured the strength of fuel elements

before and after exposure to flowing hot hydrogen. These results showed an increase in

torsional strength for (U, Zr, Nb)C following exposure while those made of (U, Zr, Ta)C

remained approximately constant (D'yakov and Tishchenko, 1994).

Butt et al. (1993) in their survey discussed the thermal shock resistance parameter as it

relates to mixed carbides. This factor is related to the thermal stress that would be developed in

a material due to temperature gradients and is proportional to the fracture strength and thermal

conductivity of the fuel and inversely proportional to its coefficient of thermal expansion as

shown in Equation 5. While noting that insufficient data exists for the mixed carbide fuels to

allow for direct comparisons, Butt et al. (1993) did note that low porosity and low uranium

content can be expected to increase thermal shock resistance. From this it would seem

favorable to have high density, high thermal conductivity, low uranium content fuels.

R'= fI(1 -/), (Butt et al., 1993; Richerson, 1992)
f = fracture strength (Pa)
c = thermal conductivi ty (W/mK) (5)
E = Young' s modulus of elasticity (Pa)
a = thermal expansion coefficien t (1/K)
y = poison ratio

Hot Hydrogen Corrosion and Mass Loss

Besides their high service temperature, single-phase, solid solution carbides have a good

resistance to corrosion by hot hydrogen compared with earlier Rover/NERVA fuel designs.

The study of carbide fuel elements at the end of the Rover/NERVA program was motivated by

carbon loss rates in earlier fuel types sufficient to adversely impact neutronic considerations.

These losses were due to the high chemical reactivity of hot hydrogen with free carbon. The

near absence of free carbon in the advanced solid solution carbide fuel elements reduces carbon

loss rates due to interaction of hot hydrogen with only chemically combined carbon (Lyon,



$ 60
E 50
E 40



0 200 400 600 800 1000 1200
Station (mm)

Figure 1-7. Mass loss rates for three different Rover/NERVA fuel elements showing the
characteristic midband corrosion pattern. Curve A is for NbC coated graphite matrix fuel
elements (Pewee- 1 tests). Similarly, curve B is for ZrC coated fuel elements (Pewee- 1 tests),
while curve C is for ZrC coated composite carbide/graphite fuel elements from NF-1 tests
(after Lyon, 1973).

Regardless of the fuel design, all fuel elements tested during the Rover/NERVA

program experienced some degree of corrosion and mass loss during testing as expected.

Insufficient test data exist for the solid solution carbide fuel elements but mass loss rates for

three other fuel designs are shown in Figure 1-4. The characteristic peak in mass loss rates

about the midsection of the core was termed "mid-band" corrosion. Mass loss during operation

can be attributed to several interrelated phenomena including radiation exposure, chemical

reaction with flowing hot hydrogen, vaporization, and creep among others. Isolated, single-

effect studies on graphite-matrix, composite, and carbide fuel materials did not indicate the

complex corrosion pattern exhibited by the Rover/NERVA tests. Instead, most single-effect



studies designed to test temperature dependence indicated a single thermally activated rate

limiting step (Barletta et al., 93). These and other findings point to competing processes that

give rise to the varied "mid-band" corrosion pattern based on the distinct local physical

conditions that are likely to exist at different stages along the fuel element length (see Figure 1-

7). Such gradients that occur along the propellant stream include temperature, pressure,

neutron fluence, and hydrocarbon concentration.

Suggested explanations of this characteristic corrosion pattern point to the high reactivity

of carbon with flowing hot hydrogen and the observed cracking in the fuel's zirconium carbide

coating. Cracking exposed the fuel to attack by flowing hot hydrogen and cracks were most

numerous around the mid-band. However, this understanding of corrosion patterns would not

apply to the carbide fuel elements due to the absence of both free carbon and a protective


However, some common phenomenon can help explain mass loss characteristics for

graphite matrix and composite design fuel and also predict losses for solid solution, carbide fuel

designs. The presence of hydrocarbons in the propellant stream serves to reduce carbon

losses. In the upper part of the reactor core where temperatures are less than 1500 K,

reactions of hydrogen with carbon will be negligible. Hydrocarbons can be added to the stream

but will be present regardless due to reactions of hot hydrogen with carbon. At high

temperatures these become unstable and their effect is negligible above 2900 K. Near the exit,

where temperatures exceed 2900 K, losses are largely due to vaporization from the exposed

carbide surfaces. Here also hydrogen corrosion is negligible and hydrogen gas actually serves

to reduce losses by vaporization by reflecting some vapor back to the surface. In between,

hydrogen will react with carbon forming hydrocarbons that build up in the propellant stream

serving to suppress losses downstream but ahead of the higher temperature region where

vaporization dominates.

Thermochemical modeling of the pseudo-binary (U, Zr)C based on available data

indicates that the initial loss of uranium and carbon will be large as they form concentration

gradients at exposed surfaces of the fuel element. A low uranium diffusion rate in the solid

solution carbide causes this gradient to be steep and limited to about the outer 40 im (Storms,

1992). Subsequent losses of uranium are predicted to be smaller and only in relative proportion

to zirconium losses. Storms (1992) concluded that this loss of uranium due to vaporization

would be the life-limiting phenomenon for solid solution carbide NTP systems. Porosity in the

carbide fuel either from fabrication or created by irradiation, serves to increase uranium diffusion

to the surface leading to higher mass loss rates. Also, increasing the surface area exposed to the

propellant gas stream, open pores further increase uranium mass losses.

Pseudo-ternary carbides of (U, Zr, Nb)C and (U, Zr, Ta)C were investigated by

LUTCH with average uranium content of ~1.1 g/cc and 0.85 g/cc respectively. The uncoated

fuel pins were fabricated in the "twisted-ribbon" design. Tests were conducted in flowing hot

hydrogen at 3300 K for one hour for (U, Zr, Nb)C and two hours for samples of (U, Zr, Ta)C.

Post-test analysis has shown uranium mass losses of -5% for (U, Zr, Nb)C samples. There

was a fair amount of uncertainty in tests for (U, Zr, Ta)C as different methods of analysis gave

conflicting results. Uranium mass losses for these samples could be 7% or as little as 1-2%

(D'yakov and Tishchenko, 1994).

In order to gauge the service life for these carbides, tests in flowing hot hydrogen were

performed at a lower temperature, 2800 K, for 10 hours. Fuel pins of (U, Zr, Nb)C

experienced average mass losses of 2%, while those of (U, Zr, Ta)C had average losses of less

than 1%. Analysis of the uranium distribution in cross-sections of fuel pins revealed that the loss

of uranium was largely from the surface of the fuel pins as shown in Figure 1-8. Solid solutions

with NbC showed losses mainly in the outer 200-300 [m while solid-solutions with TaC

revealed losses only from the outer 50-100 [m (D'yakov and Tishchenko, 1994). These

uranium mass loss profiles agree at least qualitatively with the predictions based on

thermochemical modeling of the pseudo-binary (U, Zr)C (Storms, 1992).






Distance From Surface of Fuel Pin (mm)

Figure 1-8. Changes in uranium distribution for Russian "twisted ribbon" fuel elements before
and after hot hydrogen testing (measured across fuel pin cross section). Lines 1, 2, and 5 are
pre-test measurements while 3, 4, 6, and 7 are post-test. (a) (U, Zr, Nb)C (b) (U, Zr, Ta)C
(after D'yakov and Tishchenko, 1994).

0 0.4 0.8 1.2








Preparation and Handling of Powders


Based on preliminary results from previous work with mixed uranium/refractory metal

carbides, studies of pseudo-ternary carbides such as (U, Zr, Nb)C should examine high density,

low uranium content fuels with C/M ratios of approximately 0.95. These characteristics are

most likely to provide the highest melting point, lowest mass loss rates, best thermal shock

resistance, and lowest theoretical density (through low U/M) for an advanced, high performance

nuclear fuel. Based upon previous Rover/NERVA designs utilizing highly enriched uranium

(93% U-235) with a density of approximately 0.3 gU/cc, such low uranium content fuels are

possible (Lyon, 1973). Other possible variations include Russian core designs using 0.8 to 1.1

gU/cc fuel with a reflector or "after burner" region. Therefore, this study examined the pseudo-

ternary (U, Zr, Nb)C, with a C/M of 0.95, U/M equal to 0.5 and 1.0, and zirconium to

refractory metal ratios (Zr/Mref) of 0.65, 0.75, and 0.85 yielding estimated uranium densities

ranging from 0.8 to 1.8 gU/cc (see Table 2-1).

Initial processing attempts were accomplished using mixtures of carbide powders with

uranium, C-T1, and without uranium, C-B1, C-B2, C-B3, and C-B4. These stoichiometric or

near stoichiometric compositions were easier to prepare since the carbide powders of UC,

c-i4 c4 c4 c,4 c,4 c,4

0 0


Ic o I

0 O


NbC, and ZrC were readily available. This initial phase was used to test equipment and

processing methods before attempting to process hypostoichiometric samples in the target

composition range. In all, four pseudo-binary, (Zr, Nb)C, and one pseudo-ternary, (U, Zr,

Nb)C, compositions were processed from the initial carbide powders. Table 2-1 lists these

various compositions along with the other compositions processed in this study.

As explained in the previous chapter, the low U/M ratios were chosen to maintain the

highest melting points of the refractory carbides while containing enough uranium for criticality

and desired excess reactivity. Initial studies of the UC-ZrC-NbC system indicate that the

highest melting point for compositions containing this fraction of UC should occur for Zr/Mref of

0.65 to 0.85 (Tosdale, 1967). Therefore, the pseudo-binary compositions C-B2, C-B3, and

C-B4 nominally correspond to a Zr/Mref of 0.65, 0.75, and 0.85 respectively. Similarly, ternary

carbide compositions with U/M equal to 0.1, C-T2, C-T3, and C-T4, have nominal Zr/Mrefof

0.65, 0.75, and 0.85 respectively. The same is also true for C-T5, C-T6, and C-T7 but with

U/M equal to 0.05.

Also shown in Table 2-1 are the corresponding estimates of theoretical density (TD) for

each composition. Theoretical density was taken as the crystallographic density for a solid

solution of the mixed carbides, which it was the goal of this work to produce. Crystallographic

density was calculated by dividing the mass of a unit cell weighted in proportion to the various

carbides present by the volume of the unit cell. The dimensions of the unit cell were obtained

from lattice parameters for the various carbides as listed in Table 1-2.

Table 2-2. Material data on powders used in sample compositions

Material Material Supplier Lot No. Particle Size Purity

1 UC CERAC 60 Mesh 99.5%
(-250 [m)

2 UH3 Alfa Lot No. 062174
Stock: 89000

3 ZrC (1) Alfa Lot No. F10E09 -325 Mesh 98%
(-44 [m)

4 ZrC (2) LANL Lot #5A APS 3.5 [im

5 NbC LANL Lot #46-C1 APS 3.5 pm

6 Graphite Johnson Lot #I24C08 300 Mesh 99.5%
Matthey (-48 [m)

7 Stearic Acid Aldrich Lot #07112AF 95%

8 WC Alfa Lot #B02A45 99.5%

All of the data for the mono-carbides listed in Table 1-2 are for the cubic, rock salt

[NaCl] structure. However, stoichiometric WC has an HCP structure and the cubic form

WCo.61 is only stable above 2800 K (Storms, 1967). Since the goal of this work was to

produce solid solution mixed carbides, the lattice parameter for the cubic form of WC was

chosen for this calculation and it was assumed that a limited amount of WC could be substituted

in a solid solution of largely ZrC and NbC in a cubic form. Scanning electron microscopy and

x-ray diffraction analysis would be used to determine if indeed solid solutions of all the mixed

carbide samples was achieved and whether this assumption was valid.

Vegard's Law, which assumes a linear relationship in lattice parameter with

composition, was invoked to estimate the lattice parameter of the mixed carbides by weighting

the lattice parameter of the individual carbides by their corresponding proportions (Cullity,

1978). No variation in lattice parameter was accounted for with regard to hypostoichiometric

compositions because this effect was expected to be small since no more than 5% of the carbon

atoms would be absent from their interstitial sites for the target C/M ratio of 0.95. However,

their mass was deducted from the unit cell in proportion to their deviation from stoichiometry.

This effect is likewise small since it amounts to no more than 5% of the total weight percent of

carbon, which itself is less than approximately 10 wt% of the mixed carbides and when these

factors are combined amount to less than 1% of the overall unit cell mass.

Uranium Hydride Processing

In order to produce samples with varying carbon-to-metal-ratios, powders of uranium

hydride (UH3) and graphite were mixed with carbide powders of zirconium and niobium.

Based on calculations of Gibb's free energies for the decomposition of UH3 (eq. 2-1) using the

FACT computer code (Bale and Pelton, 1996), the hydrogen is evolved at temperatures above

676 K (see Table 2-3). During sintering at temperatures of 2500 K or above as called for in

this study, all the hydrogen is predicted to be evolved from mixtures containing UH3, graphite,

and refractory metal carbides leaving behind uranium metal to form mixed uranium/refractory

metal carbides.

Uranium hydride for these samples was produced from uranium metal rod of 4.5 mm

diameter heated to 473 K in an atmosphere of flowing Ar-7%H. As the hydriding reaction

takes place, the rod appears to swell and crack and UH3 particles flake off exposing more

uranium metal (see Figure 2-1). These flakes ofUH3 were then mixed with the desired

compositions of graphite and carbides of zirconium and niobium to produce samples that were


2 UH3 2 U + 3 H2 (2-1)

Table 2-3. Gibb's free energies for decomposition of uranium hydride.

T (K) AG (J)*

300 144856.5

400 107651

500 69308.9

600 30077.6

675 176.2

676 -225

700 -9871.3

*Calculated from: FACT (Bale and Pelton, 1996)

Mixing and Handling of Powders

The material powders used in sample compositions are listed in Table 2-2 above. The

carbide powders and 3 wt% stearic acid were weighed on a Sartorius model R180D balance

and added to a 125 ml Nalgene HDPE bottle. This handling took place inside a fume hood.

Approximately 80 chrome steel balls (diameter 0.635 cm) were added to the bottle for mixing

in a ball mill. Mixing was done overnight for at least 18 hours. If only carbide powders were to

be used in a particular mixture, the bottle was closed and was mixed on a Lortone model 1.5E

rotary tumbler (see Figure 2-2). Two Buna-N o-rings were placed around the outside of the



Figure 2-1. Production of uranium hydride powder from uranium metal rod. a)
comparison of an unexposed uranium metal rod (d=4.5 mm) with a uranium metal rod
(d=3.7mm) exposed for approximately 36 hours. b) exposed uranium metal rod and
resulting UH3 particles.

bottle at the top and bottom to aid the bottle in making contact with the rollers of the rotary

tumbler to prevent slippage between the bottle and rollers preventing the bottle from rotating.

Because uranium hydride is pyrophoric in air, if the hydride powders were to be used, the

handling was done inside a glove box with an inert environment of ultra-high purity (UHP) argon

or Ar-7%H. The lids and open bottles were placed in the antechamber that was evacuated so

that no residual air would be left in the bottles to contaminate the glove box. The bottles and

lids were then transferred to the main chamber and the appropriate amount of UH3 was

weighed on an AND model HL-200 (0.1 g accuracy) scale and added to the bottle which was

then closed. As with the mixtures containing only carbide powders, the bottles were mixed on

the rotary tumbler overnight for at least 18 hours.

Prior to cold pressing, the stainless steel dies were transferred to the glove box. A 1 cm

tall graphite plug was placed in the bottom of each die and approximately 2.5 g of a mixture was

added to each die. A temporary graphite punch was placed on top of the powders to minimize

contact with the open air before transfer to the cold press where a stainless steel punch was

used during pressing. After pressing, the bottom die holder of the press was removed, and the

sample, along with the graphite plug, is pushed out the bottom of the die using the press. The

sample was then transferred to the graphite die for sintering.


Cold Uniaxial Pressing

The first attempts at processing mixed uranium/refractory metal carbides did not involve

cold pressing or any other method of compaction prior to sintering. These samples exhibited a

Figure 2-2. Mixing of powders was done on a rotary tumbler, Lortone model 1.5E.

large amount of porosity (low density) and had little mechanical strength. The samples were of

such low quality that they would be ground down during polishing. In order to produce high

quality samples of low porosity, it was decided to cold press samples prior to sintering.

Uniaxial pressing in a graphite die/susceptor

The first attempt at cold unidirectional pressing utilized a 20 ton hydraulic press to

compact powders. A two piece holder of brass and aluminum was machined to mount a 0.635

cm (0.25 in.) punch made of tool steel to the hydraulic press as shown in Figure 2-3. The

mixed powders were pressed in the same graphite die/susceptor (2.54 cm OD by 6.5 cm

length) that was used for sintering. A plexiglass shield was used during all cold pressing

20 ton
" jack



Figure 2-3. Design of the first uniaxial cold press used for pressing mixed carbide
powders in graphite dies before sintering. a) an illustration of the press. b) a photograph
of some of the hardware used. From left to right, the 0.635 cm punch and holder for
attaching to the 20 ton hydraulic press, steel sheath for containing the lateral pressure on
the die of 2.54 cm OD, and the copper mold for later dies of 1.9 cm OD.

operations to provide protection in case the die was to shatter. Almost invariably, the dies

would fail due to the high lateral pressure on the die and low strength of the graphite.

To contain the lateral pressure, a steel cylinder (3.5 cm OD) was used to sheath the

graphite dies. This setup is similar to the graphite die and molybdenum cylinder used by Accary

and Caillat (1961). To ensure a tight fit, heavy weight paper (punch cards) were wound around

the graphite die, which would be twisted into the steel sheath for a tight fit. This method

improved the success rate of cold pressing but the graphite dies would fail on approximately

every other pressing due to the inability to get a good fit. Also, it was difficult to prepare the die

and steel sheath for pressing. Therefore, the steel sheath was discarded and a new method was


Next a copper mold that was cut in half (4.2 cm on each side and 7.5 cm in length) was

used to brace the graphite die and contain the lateral pressure. The copper mold was machined

to exactly fit the graphite die by first drilling two holes along both edges of the block. It was

then cut in half perpendicular to the holes, which were used to bolt the block together again. A

central hole along the axis was then bored out to match the graphite die forming a mold to brace

it. Because this method was expected to better contain the lateral pressure, a smaller graphite

die, 1.9 cm in diameter and 7 cm tall, was used for pressing. A smaller diameter die/susceptor

was used for the added benefit of allowing for higher temperature without exceeding the limits of

the furnace's heat exchanger (see section titled "Sintering" below). The thin layer of copper

about 0.5 mm thick was filed away from the inside of each piece where the two halves contact

each other. This was done to ensure a tight fit around the graphite die when the mold and die

were bolted together. Ultimately however, this method proved no better than the steel sheath.

The failures of both these methods seemed to result from imperfect matching of the

holder and graphite die, which allowed the die to fracture at those areas of poor fit.

Recognizing the failure of the two previous methods, an improved method was sought where the

holder and die could achieve a better, if not perfect, fit. For this reason, a tapered graphite die

that could be buttressed by a much larger size matching graphite holder was developed. The

graphite die was machined to the specifications shown in Figure 2-4 to match with a tapered

hole bored into a large graphite block 9 cm by 9 cm in cross section and 12.5 cm tall using a

#10 reamer. The two complementary components could be wrung together to achieve a near

perfect fit. Pressing was achieved by applying a bi-directional force on the die/holder apparatus

(see Figure 2-4). This was accomplished using two separate hydraulic presses, one to press the

powders inside the die and the other to press the die into the graphite holder to apply an even

lateral force to the outside of the die for balancing the internal pressure form the other press.

Using this method, cold pressing was successful at pressures as high as 250 MPa.

Also, a tapered die was successfully used for pressing a second time following sintering of an

initial sample in the furnace. However, this seemed to be the limit of use obtainable from the

tapered die since most failed during a third attempt. Failure seemed to result again from the

inability to contain the lateral pressure because of poor fit of the graphite die and holder. The

outer surface of the tapered die would become uneven and its outer diameter would be reduced

in the middle where it would experience the highest temperatures during sintering. This is likely

the result of vaporization of carbon from the die surface at the high temperature encountered

during sintering or oxidation due to residual oxygen retained in the furnace during sintering. This

7.7 cm

1.84 cm

< -0.635 cm


1.68 cm

Figure 2-4. Re-design of the initial cold press using a tapered die. a) an illustration of
the press and tapered die showing dimensions. b) a photograph of the die with its larger
graphite holder for reinforcement during pressing.

large-scale unevenness could not be successfully eliminated by wringing together the tapered die

and matching holder and ultimately resulted in failure of the tapered graphite die.

Uniaxial pressing in stainless steel dies

Finally, a fourth cold pressing apparatus was designed. Because of graphite's low

strength, it was decided to press the samples in stainless steel dies and transfer the cold pressed

sample to the graphite die for sintering. Initial attempts failed because the sample lacked

cohesion after pressing and would crumble and breakup upon removal from the steel die. To

overcome this problem, a binder, stearic acid (C 18H3602), was added to the powder mixture.

The addition of a binder permitted the cold pressed sample to have a measure of green strength

to allow its transfer to the graphite dies for sintering and also provided some lubrication between

the powder compacts and die wall to allow their removal. Three weight percent stearic acid

was added to each mixture. With less than this amount, the cold pressed sample would crumble

and not hold its form for transfer to the graphite die.

An entirely new cold press was constructed to handle pressing operations using the new

stainless steel punch and dies. Instead of mounting the punch and die in a hydraulic press, it was

mounted between two steel blocks that can move up and down on two 2.54 cm (1 in.)

threaded steel rods. The bottom of each steel rod is mounted to a steel cart and pressure is

applied to the punch and die by tightening a nut on each threaded steel rod. Being able to

control each side of the press with a separate nut allowed better control over the pressing and

prevented the misalignment of the punch and die. This had become a challenge in the old

hydraulic press where there was not much precision and resulted in the bending the punch

because of misalignment. A trailer tongue scale was used during pressing for measuring up to

454 kg (1000 lbs.) for a possible equivalent pressure of 141 MPa for a 0.635 cm (0.25 in.)

diameter sample.

This press was later modified to shorten the time required to press a sample. Instead of

applying pressure by tightening the nuts on the upper die holder, a Buehler Simplimet-2

hydraulic mounting press was placed underneath the bottom die holder and replaced the trailer

tongue scale. Pressure was applied using the hydraulic jack contained in the mounting press,

which also contains a gauge for measuring the pressure applied to the sample up to 4536 kg

(10,000 lbs.) for a possible equivalent pressure of 1405 MPa for a 0.635 cm (0.25 in.)

diameter sample (see Figure 2-5).

Sintering By Induction Heating

Equipment design and performance

Carbide samples were sintered in an induction furnace following cold pressing. A

Taylor Winfield, model CE2000, 20kW, 450 kHz power supply was used with a four liter,

water-cooled test chamber (see Figure 2-6). Two water-cooled electrical power feedthroughs

from Insulator Seal Inc., model 9511020, rated for 10 kV/35 kW were welded into an 8 inch

diameter stainless steel rotatable, Conflat flange. This was bolted to one of the four arms of

the stainless steel, four-way cross test chamber. Different size coils were used along with

different diameter graphite dies. Initially, a four turn 3.3 cm ID coil made of ordinary copper

refrigeration tubing was used with a 2.54 cm (1 in.) diameter, 6.5 cm tall graphite die. A 0.635

cm (0.25 in.) diameter hole was drilled through the center of the graphite rod to form the die.

All samples sintered by induction heating were made to 0.635 cm (0.25 in.) in diameter. This

Figure 2-5. Photograph of the final cold press.

large die permitted cold pressing at approximately 600 lbs. with variable success but prevented

obtaining temperatures above 2400 K for any extended period of time. The large die has a

large surface area to radiate energy/heat to the coil causing the cooling water to boil or overheat

and trip the high temperature sensor on the power supply.

Depth Gauge

Fuel Material
Powders \

Figure 2-6. An illustration of the induction furnace chamber showing major components.

In order to achieve the higher temperatures targeted for processing samples, the

diameter of the graphite die was reduced to 1.9 cm in conjunction with the development of the

second cold pressing design. This medium sized graphite die which permitted cold pressing at

1000 lbs. with variable success, could operate at temperatures as high as 2600 K for more than

an hour. A slightly smaller, tapered diameter graphite die was developed for the third cold

pressing design. This design performed similarly to the medium graphite die but was abandoned

because of problems with cold pressing and problems removing the sample after sintering.


Tapered graphite dies as well as those of the medium size were often cracked while trying to

remove the sample. The problem resulted from the sample diffusion bonding to the graphite die

and punch during sintering (see Figure 2-7). Since the powders had been previously pressed

into the die at high pressure (as much as 250 MPa for the tapered design), the powder was

already in intimate contact with the punch and die surfaces allowing them to easily become

bonded with the sample during sintering. The fourth cold pressing design using a stainless steel

die to press the samples and then transfer to the graphite die virtually eliminated the difficulty of

removing the samples after sintering. Almost all samples could be removed by tapping the

sample from the bottom using a stainless steel rod.

Figure 2-7. A photograph showing a mixed carbide sample diffusion bonded to the graphite
punch and die following a sintering for 40 minutes at 2600 K.
I n. .: .". : :.:
-A-7 : : '

Figure ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~.. 2-7 A htgahsoigamie abd apedfuso oddt h rp it

Since the final cold press design no longer required a tapered graphite die, the design

was changed to be a uniform 1.6 cm in diameter and 8 cm long. Along with a smaller five turn,

2.2 cm ID coil, temperatures as high as 2800 K were achieved. Using this small graphite die

and small coil, samples were sintered as long as 90 minutes at or near this high temperature.

Slight decreases in power were necessary during some experiments due to inadequate cooling

from the heat exchanger that provides cooling for the coil and power supply.

Another problem encountered during sintering was arcing between the coil and graphite

die. Early coil designs were made from ordinary refrigeration copper tubing with a wall

thickness of 0.076 cm (0.030 in.). Arcing would melt the coil allowing the cooling water to

pour into the chamber terminating the experiment and requiring an extensive cleanup effort.

Later coil designs were made from heavy-duty copper tubing with wall thickness of 0.124 cm

(0.049 in.). The thicker wall tubing prevented most arcing events from destroying the coil and

would only occasionally produce a pinhole but would still require replacement and cleanup of

the chamber.

Pressure was applied to the graphite punch by means of a mechanical feedthrough with

model number, VF- 100-2, by Huntington Inc. Weights sufficient to produce a pressure of 3

MPa were added to the feedthrough for use during the sintering of mixed carbide samples. A

dial gauge graduated in thousandths of an inch was fitted on top of these weights and measured

the movement of the punch during sintering. Using this gauge, it was possible to note the onset

and cessation of compaction as well as estimate the degree of compaction based on gross

movement of the gauge.

Temperature measurement and control

Measurement of sintering temperature was done by using a MaxlineTM temperature

acquisition and control system. The control unit was a MaxlineTM model MX-MR04 with

infrared thermometers, which span a temperature range from 977 to 3866 K. These so-called

"two color" sensors operate by measuring the ratio of energy emitted by the target at two

infrared wavelengths of 0.7 and 1.07 microns. Because they operate on a dual wavelength

ratio, they are not susceptible to errors in temperature measurement owing to the emissivity of

non-blackbodies which is cancelled by taking the ratio of intensity at two different wavelengths

(Felice 1995). Instead, possible errors result from any wavelength dependence of emissivity for

the target in question. Any difference in the target's emissivity at the two different wavelengths

monitored by the sensors must be compensated for through an emissivity slope (e slope) factor.

Greybodies are a group of materials whose emissivities are the same at both wavelengths

measured by the sensors and therefore do not require a compensating factor or e slope factor

equal to 1.00. For applications involving the Maxline TM system, graphite can be considered a

greybody (Maxline, 1988). Measurements by Neuer (1995) have shown that the emissivity of

graphite is approximately linear at these wavelengths.

Testing schedule

Various heating schedules were tried to determine the processes occurring during

sintering and to find the optimum processing parameters such as sintering temperature and time

interval. Samples were processed at temperatures between 2500 and 2800 K for times ranging

from five minutes to two hours.

Dynamic Magnetic Compaction

Dynamic magnetic compaction (DMC) was investigated as an alternate method to cold

isostatic pressing of carbide powders. Carbide powders from batches C-T1 and C-T8 were

compacted using the DMC method. A copper tube 1.58 cm OD and 1.28 cm tall with a wall

thickness of 0.07 cm was used as the armature. It was filled with the powder mixture and

shaken using a vibrator to get a uniform filling so that the finished sample would be uniform in the

axial direction. Figure 2-8 shows the prepared sample just prior to insertion into the coil for

compaction. Following compaction, the samples were sintered in the induction furnace at

temperatures between 2700 and 2950 K for periods as long as 90 minutes. Sintering was

accomplished by placing the sample with its copper armature atop a 1.9 cm diameter graphite

pedestal. The pedestal was positioned so that the sample was at the center both vertically and

diametrically of a four turn, 2.2 cm ID induction coil. The copper armature was removed from

the sample by melting as the power was increased to the sintering temperature.

Hot Pressing

Equipment design and performance

Hot pressing of mixed carbides by direct resistance heating of the samples was

investigated. The first attempt at hot pressing involved the use of an arbor press to which a 1.27

cm (0.5 in.) graphite punch and die were added between the press ram and base plate. A

water-cooled, copper electrode was bolted around the graphite punch at the top and a copper

plate formed the bottom electrode attached to the graphite die underneath. As much as 650 A

at 5.9 volts of direct current was applied to the hot press from a Miller power supply, model

Figure 2-8. Picture showing a pseudo-ternary carbide sample just prior to insertion into
the electromagnetic coil for compaction using the dynamic magnetic compaction method.

number SR-1000-Cl, for a peak input power of 3.8 kW. The peak power output from the

power supply was 7.3 kW with significant losses in the cables and connections. Pressure was

applied to the punch by lead bricks attached to the arm of the arbor press. Loose graphite

powder was packed around the punch and die as insulation. This graphite insulation was

contained by an inconel foil shield, 5 cm (2 in.) in diameter, that together with the packed

graphite powder enclosed all of the punch and die except for the upper 2 cm (0.75 in.) that was

exposed. A depth gauge mounted on top of the ram of the arbor press that contacted the

graphite punch/electrode was used to monitor compaction. The dial face of the gauge was

marked in thousandths of an inch.

During the experiment the inconel shield would glow incandescent and the upper part of

the punch that was exposed glowed a bright orange. Problems associated with this setup

included burning of the electrode that was exposed and to a lesser extent that part which was

enclosed by the graphite insulation and inconel shield. Experiments were halted before its failure

and its final condition can be seen in Figure 2-9. In early experiments, the copper electrode lost

its contact with the graphite punch as the experiment heated up. This forced the addition of

water cooling for the copper electrode. This set of experiments was ultimately halted because

the insulation on the cables melted at 7.3 kW supplied by the power supply. Compaction was

noted to be continuing before the experiment was halted indicating that more compaction was

still possible.

Equipment redesign

The failures noted in the original hot press forced the design and construction of a

different press to prevent overheating of the cables and better protect the graphite

I _

Figure 2-9. Picture of damaged hot press components.

punch/electrode from oxidation. The second hot press used the larger framework of 20 ton

hydraulic press with the hydraulic hand jack replaced by lead bricks placed on top of the punch

(see Figure 2-10). The upper and lower parts of the graphite electrode that are exposed were

enlarged to 9 cm by 9 cm (see Figure 2-11). Water-cooled copper electrodes bolt directly into

these graphite electrodes. A 1.27 cm (0.5 in.) graphite punch and die were inserted between

the graphite electrodes. The outer diameter of the graphite die was 2.54 cm (1 in.). A larger,

28 cm (11 in.) diameter by 38 cm (15 in.) tall, steel drum was used to contain the loose graphite

insulation that was packed around the graphite electrodes and the punch and die. This was to

provide better insulation and more protection to reduce the oxidation of the punch and die. A

0.635 cm (0.25 in.) quartz tube was added penetrating the graphite powder and abutting the

graphite die. UHP argon was allowed to flow through this tube, which was used to make a





Water Cooling Lines

I 4- Top Electrode


4- Bottom
Water Cooling Lines

Figure 2-10. Schematic of the second hot press.

visual inspection of the color radiating from the die. At the highest power levels attempted

during testing, the die glowed a bright orange color and was estimated to be between 3000 K

and 3200 K.

; ,| ,

Figure 2-11. Picture of the second hot press electrodes.

Some operational difficulties occurred during the first experiment. The original graphite

punch was a uniform 1.27 cm (0.5 in.) in diameter and a significant part of the upper punch was

oxidized or was vaporized. The reduced diameter caused the punch to collapse under the

weight of the lead bricks. It was prevented from completely collapsing by steel blocks that

limited how far the punch could travel. The graphite punch was redesigned to be 2.54 cm (1

in.) in diameter in the upper part and taper down to a uniform 1.27 cm (0.5 in.) where it makes

contact with the die. The graphite insulation was repacked and the UHP argon was allowed to

flow into and permeate the insulation overnight. Compaction was noted in all experiments but in

each case the experiment had to be halted because of overheating of the cables.

Testing schedule

The hot press was operated in this final configuration for four experiments between 60

and 112 minutes. Hot pressing was continued until compaction ceased after reaching the

maximum output of the power supply or the cable overheated.

Density Measurements

Bulk density was measured to estimate the amount porosity contained in the samples.

Sample porosity is important since it affects material performance and high density (low

porosity) fuels are required to achieve the desired performance under NTP conditions. Also,

the relative amount of porosity provides an indication of the degree of sintering and whether

additional processing is needed. Also, it provides a means to compare the effectiveness of

different processing parameters such as temperature, compaction pressure, sintering time, etc.

Because the sample shapes were not entirely regular, their volume had to be estimated

by measuring the displacement of water from a known volume. This was accomplished by first

weighing the dry samples before any other steps were taken and recording their masses. Next,

the weight of a small container of water (d=lcm, h=lcm) was measured and recorded. Each

sample was then individually placed in the container, which was then filled with water and

weighed. The mass of the container filled with water and the dry sample mass minus the mass

of the sample and water together provides the mass of the water displaced by the sample. Then

assuming a density of 1 g/cc for water, the volume of the sample was obtained. The assumption

of Ig/cc for water at room temperature is valid since for water at 293 K, the density is 0.998

g/cc and does not vary more than 0.003 g/cc for +10 K.

Melting Point Determination

Because of the importance of high melting temperature to the usefulness of these

carbides, it was decided to examine the melting temperature of these mixed carbides. A test

stand for melting point determination was constructed in a 36 cm (14 in.) bell jar (see Figure 2-

12). Water-cooled, power feedthroughs made of copper and rated for 1000 amps were used.

These power feedthroughs (part number 9462017) were manufactured by Insulator Seal

Incorporated. A Miller power supply, model number SR-1000-C1, was used to supply power

to the test stand. The apparatus was designed to use direct resistance heating to melt the small,

0.635 cm (0.25 in.) diameter, samples processed in this study. Graphite electrodes were used

with a top electrode diameter of 1.27 cm (0.5 in.) and the bottom electrode 2.54 cm (1 in.) in


The top electrode was bolted to two pieces of copper while the bottom electrode was

drilled and bolted to a copper plate. Both ends were water-cooled. The bottom plate was

insulated from the other parts of the test stand using a mica sheet of thickness 0.5 mm. A small

shallow, 0.635 cm (0.25 in.) diameter depression was machined in the ends of the graphite

electrodes to hold the sample in place during heating. The top electrode was designed to move

freely vertically between two 1.27 cm (0.5 in.) threaded rods covered and electrically insulated

by quartz tubes. A dial gauge measuring movement in one thousandth of an inch monitored its

vertical movement. The electrodes and sample would be expected to expand upon heating. A

subsequent contraction at temperatures near the melting point would be indicative of the onset

of structural changes in the sample. Therefore, this is the temperature of interest because

structural changes in the proposed fuel could block coolant channels and the formation of any

liquid must be strictly avoided to maintain mass loss rates at acceptable levels.

Figure 2-12. Apparatus for melting point determination.

Temperature measurements were made using a MaxlineTM, model MX-MRO4,

temperature acquisition and control system with infrared thermometers. As discussed above,

these so-called "two color" sensors operate by measuring the ratio of two infrared wavelengths,


0.7 and 1.07 microns, and are not susceptible to errors in temperature measurement owing to

the absolute emissivity of non-blackbodies (see "Sintering" above). However, ZrC and NbC

are not greybodies as graphite and require use of an e-slope factor of 1.1 to correct for non-

linearity in their emissivity with wavelength based on data for emissivity of ZrC, the dominant

component, from Touloukian and DeWitt (1972).


A total of 33 mixed carbide samples were processed using the various methods and

parameters discussed in the previous chapter. Table 3-1 and Table 3-2 list the samples, their

nominal composition, powder mixture batch number, measured bulk density, and key

parameters related to their processing. Of the 33 total samples, 20 were ternary carbides and

of those 20, 17 contained depleted uranium in amounts of either 5% or 10% metal mole

fraction. Nineteen of the ternary carbides and seven of the binary carbides of ZrC and NbC

were pre-compacted either by cold pressing or the DMC method and sintered. Additionally,

three samples were successfully processed by hot pressing for comparison with pre-compacted

and sintered samples. Results of these experiments are summarized below.

Binary Carbides

The binary carbide samples B4 through B 10 were processed by cold pressing at 120

MPa and sintering at temperatures as high as 2800 K for as long as one hour (see Table 3-1).

For samples B through B3, no pre-compaction was done prior to sintering. A portfolio of

scanning electron micrograph (SEM) images is shown in Figures 3-1 through 3-10. These

SEM images reveal a largely open, porous microstructure. Additionally, many of the binary

samples were not particularly consolidated and some would be ground down noticeably during

polishing. This pattern is true despite sintering times longer than 30 min. at 2800 K and pre-

7 7 t \C q1
c C \Cq [~ C

z C

2 0 0 0 0 0
si^a a a o

o a o o o a

Z Z~ l Z1 Z Z C> Z
d c
o ^ ~ ~c ~c o c- oc c- 0c )-

- 00 00 C> C> (^ Cl C> ( C00


o u - u o u u o u u
NU N N/- N/i < D m ti < r r r

- ^ ^ ^ N N N N N N

-^ i- Cl Cl rn c ^ r ^ r
0<*0 0 0 0
-^ - ^ ^ ^

6 _
S ^ ^ [^Cl>^ / >/ >/C C>/


o a
o 0 0 00 00
o o o l oD en o Cl

e en en ND \ 0 Cl N >

e 4N t e en 2 C> C>

O D C> Cl
o N (N ( t 00 m N

0 0 0 0 0 o 0 0 0 0 0 U
0 0 0 e Cl 0 0 00 N N N tN eN C -

\ A A m N O O .i 6 en ^ O Cl Cl
o N0 C> 00 C 00 C> (N 0> 00 00) N 00 00 00

-r 0 en 'C 0 ^1- 1l- 0> '- 'C N C> N^ ( 0 00 Cl
Cl ~ 'C 'C N C enCl ~ M

C)I 'C NN '- ND ND N- (^ N 'C 'C trN 'C 'C 'C


Cr j ^! ^! J ^ .JtJ!J
S -r -r -r -r -r -r -r ^r ^- ^- -r ^ i

o N N N N N N N N N N N

u u u u u u 'C 'C 'C 'C u 0 0

-r Cl en 1- tn 'C N, 00 C> C l en 1 tfl r

o en en ~t



o o ~ 00 N
Cl Cl o


o ~ C)
Cl Cl Cl ~ C
C Cl

o ~> o -
H 00 a> od ~- ~
o 00 00 00 N 00

m0 0> 00 'C 'C
'C 'C 'C tfl 'C

Cl Cl cn ~
o Cl -
? ? ? ?
0t Cl -
S 00


-~ Cl en ~t ~ N

C 'C N 00 0> C

compaction by cold pressing at 120 MPa. Further, a pressure of 3 MPa was applied to all

binary samples during sintering in the induction furnace. These characteristics lie in stark

contrast to most of the ternary samples, which after sintering were quite hard and appeared

fairly consolidated from their SEM images (see "Ternary Carbides" below).

Ternary Carbides

Density Measurements and Microscopy Results

The ternary carbides were processed by sintering for various times and temperatures as

shown in Table 3-2. All samples were pre-compacted prior to sintering except for sample TI

and T19. Samples T13, T14, and T15 were compacted using the DMC method. The

measured density and percentage of theoretical density for each sample are also shown in Table


Many of the various problems encountered during processing have been outlined in the

previous chapter describing the development of the different processing methods. In some

cases, these problems affected the processing of one or more samples and changed the

prescribed processing parameters. These variations were documented and are reflected in the

data contained herein. The fact that problems occurred does not invalidate the experiments as

those processing parameters were to be intentionally varied. This was done in order to gain

perspective on the various phenomena occurring during processing and to identify the critical

parameters for processing mixed carbides.

A portfolio of scanning electron micrograph (SEM) images are shown in Figures 3-11

through 3-29 at the end of this chapter. For most samples, a standard SEM image generated

from secondary electrons emitted from the sample reveals the topographic features of the

sample. Also, where possible a second SEM image showing compositional contrast is shown

by imaging using backscattered electrons. For these images, lighter areas indicate regions of the

mixed carbide sample with a largely higher atomic number than the surrounding material.

Therefore, since Nb (Z=41) and Zr (Z=40) have similar atomic numbers, uranium (Z=92)

provides a marked contrast when concentrated in regions of the sample providing indication of

inhomogeneity in the mixed carbide. All images were taken on a JEOL 35C with a beam

voltage of 25 kV.

The SEM image of sample TI is shown in Figure 3-11. A fair degree of residual

porosity is evident and seems to agree with the estimate of 76.8% TD based on density

measurements. The numerous rounded pores are indicative of the intermediate stage of

sintering, while the large, light shaded areas of Figure 3-1 lb indicates a fair degree of

inhomogeneity comprised mostly of UC.

The microstructure of sample T2 is shown in Figure 3-12. Its higher density can be

attributed in part to pre-compaction by cold pressing at 140 MPa. Rounded and elongated

pores are indicative of the intermediate stage of sintering. Grain growth is evident with grain

sizes of approximately 30 to 40 pm indicating the latter stages of sintering. The large amount of

uranium contained in the grain boundaries points to large scale inhomogeneity. This

concentration in the grain boundaries indicates a process of liquid phase sintering where a liquid

(UC) fills the open pores and grain boundaries aiding in the sintering process.

Figure 3-13 shows a larger amount of porosity for sample T3 than would be expected

based on the longer sintering time of 68 min. at 2600 K. This is in part due to the fact that the

die cracked during pressing and only achieved approximately 80 MPa pressure instead of the

intended 140 MPa applied to most samples. Also, the graphite punch was believed to have

been stuck initially during sintering since the onset of compaction was not observed as expected.

More weight was added to the mechanical feedthrough and indication of compaction was noted

thereafter. The large areas concentrated in UC (see Figure 3-13b) note a large degree of


Similar to sample T2, Figure 3-14 shows spherical and elongated pores for sample T4.

Its higher density could be expected from the longer sintering time it received. Accordingly, the

sample also appears to be more homogenous. The heating schedule for sample T5 was very

abbreviated and as expected the porosity is much higher and the large areas of concentrated

UC indicate a lack of homogeneity (see Figure 3-15). In contrast, sample T6 appears much

less porous (96.0% TD) with a much more homogenous microstructure owing to the extended

sintering time of 142 min. at 2500 K. The fraction of UC that is still visible in the grain

boundaries is small and points to liquid phase sintering (see Figure 3-16). The large grains on

the order of 100 pm indicate a significant degree of grain growth which would be expected for

such long times at high temperatures.

The largest amount of porosity (72.8% TD) for any sample was exhibited by sample T7

with a much abbreviated heating schedule of 4 min. at 2400 K and 8 min. at or above 2300 K.

Figure 3-17 shows a large amount of porosity and inhomogeneity. Sample T8 points to the

effect of temperature on sintering with a short heating schedule of 5 min. at a higher temperature

of 2800 K and 10 min. at or above 2700 K. A moderate amount of porosity (90.7% TD) is

shown in Figure 3-T8 with smooth, mostly spherical pores and some inhomogeneity remaining.

Samples T9 through T12 mark a shift in the processing methodology for ternary

samples. These samples were produced by mixing the refractory carbide powders with uranium

hydride and stearic acid as a binder/lubricant. These samples were pressed in stainless steel

dies prior to transfer to the graphite susceptor for sintering. A persistent arcing problem with

the induction furnace prevented high temperature sintering and resulted in abbreviated heating

cycles for most samples. As a result, lack of consolidation prevented any SEM image of

sample T9 from being obtained. Instead, the sample was virtually ground away to a powder

during the polishing step after processing.

Figure 3-19 indicates a large amount of porosity for sample T10 with 85.9% TD.

Because of the lack of consolidation, a compositional contrast image was not possible. Instead

of revealing contrast in atomic number of the sample, the only contrast visible was between the

peaks and the deep holes present in the open microstructure. Sample T 11 (see Figure 3-20)

and T12 (see Figure 3-21) appear much as sample T10 with a large amount of open porosity.

These samples exhibit only the initial stages of sintering.

Samples T13 through T15 were pre-compacted using the DMC method without a

binder. Figure 3-22 shows the large spherical pores of sample T13 with 81.2% TD. A virtually

uniform composition is shown in Figure 3-22b, which can be attributed to a long sintering time

of 90 min. at 2700 K. Samples T14 (see Figure 3-23) and T15 (see Figure 3-24) were

processed from a mixture of WC-ZrC-NbC. Both samples have similar microstructures and

densities of 82.1% TD and 82.6% TD respectively even though sample T15 was sintered at a

temperature of 2900 K for 97 min. compared with 2700 K for 94 min. for sample T14.

The SEM images of samples T16 through T20 are shown in Figures 3-25 through 3-29.

These samples were processed from hypostoichiometric mixtures similar to samples T9 through

T12. Likewise, these samples except for T19 were also pre-compacted by cold pressing at

120 MPa prior to sintering. Each sample was sintered for different time intervals for 40 min. to

128 min. above 2500 K. Peak temperatures above 2800 K were achieved for samples T16

through TI 8 for three to four minutes. This temperature could not be maintained because the

energy radiated by the susceptor began to boil the water cooling the induction coil.

From their SEM images (Figures 3-25 through 3-27), samples T16 through T18 appear

to be well compacted in agreement with their relative densities of approximately 89% TD.

These same samples also show evidence of liquid phase sintering. Sample T19, which was not

pre-compacted, appears far less consolidated at only 74.5% TD (see Figure 3-28). Because

of their large amount of porosity, meaningful compositional contrast images could not be

obtained for samples T19 and T20.

X-ray Diffraction Results

The diffraction patterns of the original powders, all the ternary samples, and samples B7

and HP2 are shown in Figure 3-32 through 3-59 at the end of this chapter. Results of this

analysis are summarized in Table 3-3. Also shown is the diffraction pattern of the UH3

processed from uranium metal rod (see Figure 3-37).

Samples Tl and T5 appear to be largely solid solutions with the peaks corresponding to

d-values for a cubic phase between those of pure ZrC and pure NbC. For example, the d-

value of the { 111} plane falls at 2.66 which is the weighted average between ZrC, NbC,

and UC. The slightly broader peaks indicate there is some variation in the d-values and

Table 3-3. X-ray diffraction results.

No. Solid Solution Determination Additional Phases/Impurities

Tl s.s. with possible concentration Significant amount U02

T2 s.s. with possible concentration Very small amount graphite, UC2, and oxide
gradients of Zr/Nb

T3 s.s. with possible concentration Very small amount graphite and UC2

T4 s.s. with possible concentration Small amount graphite and UC2 to a lesser
gradients extent

T5 s.s. with possible concentration Very small amount U02

T6 s.s. with possible concentration Very small amount UC2

T7 s.s. with possible concentration Small amount UC2 very small fraction oxide
gradients of Zr or Nb/U

T8 distinct peaks; some indication significant amount U02
of s.s. regions

T9 Not solid solution; ZrC/NbC

T10 Not solid solution

T11 Not solid solution Large amount of graphite

T12 Not solid solution

T13 Solid solution

T14 Solid solution

T15 Solid solution

T16 Solid solution

T17 Solid solution

Table 3-3. continued.

T18 Solid solution Small amount of graphite

T19 Not solid solution

T20 Solid solution

B7 Solid solution

HP2 Solid solution Significant amount of graphite

therefore variation in composition and that it may not be entirely homogenous. Also, a small

ZrC peak is discernable for the { 111 } plane indicating a small amount of ZrC is still present. In

addition, some impurity U307 is also present based on the additional peaks identified in Figures

3-39 and 3-43. This is likely the result of residual oxygen in the chamber during processing

since this phase does not appear in the starting materials.

Similar results were obtained for sample T2 except that no residual ZrC appears to be

present leaving a solid solution with some slight concentration gradients possible as noted from

the small degree of broadening exhibited by the solid solution peaks. A very small fraction of

other phases are noted also including graphite and UC2. An oxide is also present in a very small

amount of either ZrNb207 or Nb14ZrO37.

Samples T3 and T4 are a solid solution with some small composition gradients as noted

by the slightly broadened peaks. Also, graphite and UC2 are noted to be present in small

amounts. Samples T6 and T7 show only a small fraction of UC2 and no graphite while T7

shows an additional oxide of either ZrO2 or Nb3UO10.

Beginning with sample T8, a marked difference in the sample diffraction patterns

appears as noted by the virtual absence of a solid solution. Sample T8 shows distinct peaks for

NbC and ZrC as well as the presence of solid solution regions. Also, present is a significant

amount of U02 which similar to the oxides found in other samples are likely the result of residual

oxygen in the chamber during processing.

Samples T9 through T12 and T19 have distinct ZrC peaks with tails stretching to higher

angles of 20 or lower d-values. These tails stretch as far as the d-values for NbC and,

combined with the lack of a distinct NbC peak in most cases, point to the early stages of solid

solution formation. On the other hand, samples T16, T17, T18, and T20 are shown to be

largely solid solutions by the sharp well defined peaks for the mixed carbide at the expected d-

values (angles of 20) as shown in Figures 3-54, 3-55, 3-56, and 3-58 respectively. The only

second phase evident is graphite in samples T 11 and T18.

Sample T13 appears to be a solid solution with sharp peaks corresponding to expected

d-values of the mixed carbide (see Figure 3-51). Similarly, T14 and T15 are solid solutions of

NbC and ZrC with WC instead of UC as in the previous examples (see Figures 3-52 and 3-

53). The peaks are slightly broader for T14 and T15 indicating a range in d-values and hence

the presence of possible concentration gradients. No lines corresponding to d-values for the

hexagonal mono-carbide, WC, are present indicating that it forms a solid solution with NbC and

ZrC at least in the concentration range up to 10% WC.

Hot Pressing

Three samples were hot pressed using the apparatus described in the previous chapter

with the results of these experiments listed in Table 3-4. The highest temperatures achieved

were estimated between 3000 K and 3200 K. Compaction was noted in all hot pressing

attempts and was noted to be continuing although extremely slowly before each experiment was

terminated due to overheating of the electrical cables. SEM images of samples HP1 and HP2

are shown in Figures 3-30 and 3-31 respectively. These samples have a small amount of open

porosity and are fairly consolidated as shown in the SEM images. This was further made

evident while polishing the samples prior to imaging as they were hard and required several

grinding disks to polish the samples. The last sample, HP3, was not fully consolidated since the

hot pressing operation had to be stopped due to melting of the electrical cable insulation.

The binary carbide samples B7 and HP2 both indicate the formation of solid solutions.

However, only sample HP2 shows a significant amount of graphite also present in the sample.

Table 3-4. Measured data and processing parameters for hot pressed samples.
Processing Time
No. Composition ensity %TD Above -3100 K
(g/cc) (min.)
HP1 WC 13.9 88.8 55
HP2 (Zro.7, Nbo.3)C 5.91 85.1 30
HP3 WC 13.18 84.3 0

Figure 3-1. Binary sample Bl, (Zro.7, Nbo.3)C, with p= 5.6g/cc (80.6% TD) processed for
42 min. at = 2800 K after cold pressing at 120 MPa. Scale indicator is 10 rim. a)
SEM; b) higher magnification SEM.

Figure 3-2. Binary sample B2, (Zro.7, Nbo.3)C, with p= 5.79 g/cc (83.3% TD) processed
for 62 min. at = 2800 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

Figure 3-3. Binary sample B3, (Zro.7, Nbo.3)C, with p= 6.19 g/cc (89.2% TD) processed
for 47 min. at = 2800 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

Figure 3-4. Binary sample B4, (Zro.7, Nbo.3)C, with p= 6.07 g/cc (87.4% TD) processed
for 30 min. at = 2800 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

Figure 3-5. Binary sample B5, (Zro.75, Nbo.25)C, with p= 5.48 g/cc (79.5% TD) processed
for 18 min. at = 2700 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

Figure 3-6. Binary sample B6, (Zro.75, Nbo.25)C, with p= 5.31 g/cc (77.0% TD) processed
for 5 min. at = 2400 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

Figure 3-7. Binary sample B7, (Zro.65, Nbo.35)C, with p= 6.02 g/cc (85.9% TD) processed
for 37 min. at = 2750 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

-p. --

Figure 3-8. Binary sample B8, (Zro.65, Nbo.35)C, with p= 5.49 g/cc (78.3% TD) processed
for 30 min. at = 2750 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

1, ISO#

Figure 3-9. Binary sample B9, (Zro.75, Nbo.25)C, with p= 5.63 g/cc (81.6% TD) processed
for 17 min. at = 2700 K after cold pressing at 120 MPa. Scale indicator is 10 rim.
a) SEM; b) higher magnification SEM.

Figure 3-10. Binary sample B10, (Zro.75, Nbo.25)C, with p= 5.48 g/cc (79.5% TD)
processed for 19 min. at = 2800 K after cold pressing at 120 MPa. Scale indicator
is 10 rm. a) SEM; b) higher magnification SEM.


', .o t t ,.1 F : ''" :"* '^ ^ *. *ity. .

A J 1

1 , ..*

. - *. . ..' ..a u d.' .



Figure 3-11. Ternary sample TI, (Uo.I, Zr0.45, Nbo.45)C, with p=6.23 g/cc (76.8% TD)
processed for 19 min. at = 2600 K without pre-compaction. Scale indicator is 10
rm. a) SEM; b) SEM with compositional contrast.

.'"i .

I ,' A?



* --Am
frS .fc '







a -


Figure 3-12. Ternary sample T2, (Uo.i, Zro.45, Nbo.45)C with p=7.46 g/cc (91.9% TD),
cold pressed at 140 MPa and sintered for 20 min. at = 2600 K Scale indicator is
10 gm. a) SEM; b) SEM with compositional contrast.



a. 100 rm

b. 10 pn

Figure 3-13. Ternary sample T3, (Uo.1, Zro.45, Nbo.45)C with p=6.60 g/cc (81.3% TD)
processed for 68 min. at = 2600 K after cold pressing at -80 MPa. a) SEM; b)
SEM with compositional contrast.

a. 10 pm




I, *tr4.


*h. ,g~

* '4


"ii. V
-~ a


100 pm

Figure 3-14. Ternary sample T4, (Uo.i, Zr0.45, Nbo.45)C, with p=7.64 g/cc (94.2% TD)
processed for 38 min. at = 2600 K after cold pressing at 140 MPa. a) SEM; b)
SEM with compositional contrast.



1.' ^
^ ^te* ^ ^

' t* 1 ~4ir

4- 4; ,
M L ..J ~~fr.. *

- ,t ,

Figure 3-15. Ternary sample T5, (Uo.i, Zro.45, Nbo.45)C, with p=7.14 g/cc (88.0% TD)
processed for 6 min. at = 2500 K after cold pressing at 250 MPa. Scale indicator
is 10 gm. a) SEM; b) SEM with compositional contrast.


:' r .: ',
.'. .,


0-0. "


Si _

* I,

Is. -


'It *1






t. ~,1



"r V

t: A, 1


;t t-~-.-


A .




Figure 3-16. Ternary sample T6, (Uo.i, Zr0.45, Nbo.45)C, with p=7.79 g/cc (96.0% TD)
processed for 20 min. = 2600 K and 142 min. = 2500 K after cold pressing at 150
MPa. Scale indicator is 10 gm. a) SEM; b) SEM with compositional contrast.