FFR_chap20.txt

CHAPTER 20
COMPOSITION AND PROPERTIES OF LIQUID-METAL FUELS*

20-1. Core Furn ComrosiTioN

In Chapter 18, the advantages and disadvantages of liquid metal fuels
were discussed in a general way. The point was made that a liquid-metal
fuel has no theoretical limitation of burnup, suffers no radiation damage,
and is easily handled for fission-product poison removal. In this chapter,
the results of research and development on various liquid-metal fuels are
presented. This work has been largely concentrated on uranium dissolved
in bismuth.

At the contemplated operating temperatures of approximately 500°C,
it was found that uranium has adequate solubility in bismuth when present
by itself. However, as the work progressed, it soon became evident that
other materials would have to be added to the solution in order to obtain a
usable fuel. The present fuel system contains uranium as the fuel, zir-
conium as a corrosion inhibitor, and magnesium as an oxygen getter.

An LMI'R operating on the contemplated Th**? to U?* breeding cycle
an be designed with an initial U3 concentration of 700 to 1000 ppm in
bismuth. The actual figure, of course, 18 dependent upon the specifie de-
sign and materials used. In Chapter 24, in the design studies, such figures
are given. The concentrations of zirconium and magnesium are each ap-
proximately 300 ppm. It is contemplated that these concentrations will
have to be varied depending upon desired operating conditions.  In their
use as corrosion inhibitor and antioxidant there is enough leeway for this
purpose.,

The fuel deseribed in the previous paragraph is the clean fuel which would
be charged initially. During reactor operation, however, fission products
will build up in the fuel and would be maintained at a level dictated by the
economics of the chemical reprocessing system used. It has been found
that the fission products and other additives to the bismuth have an im-
portant effect on the solubility of uranium in bismuth. These have been
arefully investigated in order to permit selection of reactor temperatures
that will ensure that all the uranium remains in solution during reactor
operation. Likewise, the solubility of steel corrosion products has been in-
vestigated to determine their effect on uranium solubility in bismuth.

*Based on contributions by D. H. Gurinsky, D. G. Schweitzer, J. R. Weeks,
J. 8. Bryner, M., B. Brodsky, C. J. Klamut, J. G. Y. Chow, . A. Meyer, R. Bour-
deau, and O. F. Kammerer, Brookhaven National Laboratory.

722
20-2] SOLUBILITIES IN BISMUTH 723

 

 
     

 

 

 

635° C 560 496 441 394 352 315
5.0
' I [ | I I [
K -
1.0 =
B2 -
§ 5 }I:
>
10 &
05 L | I | | | |
1.0 i 1.2 1.3 1.4 1.5 1.6 1.7

1000/7 K

Fra. 20-1. Solubility of uranium in bismuth.

It 1= important to note that although the basie fuel is a simple one, the
uranium used for liquid metal fuel reactors using the Th—U?233 cycle must
be almost completely enriched 233 or 235 in the initial charge. Ifurther,
since the concentrations are measured in parts per million by weight, it is
not an easy matter to maintain a striet accounting of all fuel. When deal-
g with such small amounts, losses due to reaction of uranium with
carbon and adsorption of uranium on steel and graphite walls can be sig-
mificant,

The fuel for the LMFR is still under extensive study. At present, most
of the major information for the design of an LMFR experiment is at hand.
Thix information is primarily solubility data and other fuel information,
presented in the following pages.

20-2. SOLUBILITIES IN BIisMUTH

20-2.1 Uranium. The experimental techniques used to measure solu-
bilities in liquid bismuth have been described previously [1,2]. Several
workers [3-7] have investigated the solubility of uranium and bismuth.
Recently, with improvements in analytical techniques, redetermination of
the solubility curve has been undertaken. The latest results are at variance
with the older work of Bareis [5], as shown In I'ig. 20-1. It can be seen
that the recent data obtained at Brookhaven National Laboratory are, at
some temperatures, as much as 20 to 259, lower than those obtained some
vears ago.

This variance in solubility determinations may be due to several factors,
but 1t 1= believed that the improved techniques are more reliable, and that
the newer values are consequently more precise. The presence of such
724 PROPERTIES OF LIQUID-METAL FUELS [cHAP. 20

Temperature, °C
1200 1000 8GO 700 600 500 400 300
IR '

 

10.0

o
S

Solubility, Wt %

o
o

010

 

 

 

 

/T K x 104

Fig. 20-2. Solubility of uranium and thorium in bismuth.

727 636 560 494 441 394 352 315 283
10%

 

   
 

 

 

 

| ] [ [
B
— ].05/0 [—— —
=
2
=
; Los Alamos
S01% -
>
a .
01% | 1 | | 1 | { 1 1
0% 110 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

1000/T °K
Fra. 20-3. Solubility of plutonium and uranium in bismuth.
20-2] SOLUBILITIES IN BISMUTH 725

other materiils as nickel, copper, manganese, ete., in the bismuth in quan-
tities large enough to affect the uranium solubility still remains to be in-
vestigated.  For example, nickel has been shown to markedly reduce the
uranium =olubility in hismuth [1].

It 1= obvious that even slight variations of the solubility of uranium in
bismuth nmight be of considerable importance in LMI'R reactor design.
The solubility of uranium, according to the preferred data (the solid curve
i g, 20-1), allows a rather small leeway in uranium concentration in the
reactor cycle when the lowest temperature of 400°C in the heat exchangers
1s taken into account.

20-2.2 Thorium and plutonium. The solubility of thorium in bismuth,
as determined by Bryner, is compared with the solubility of uranium
in Iig. 20-2. In the temperature range 100 to 500°C, the solubility of
thorium 1= markedly lower than that of uranium. In fact, it is =0 low that
a breeding cvele using only thorium in solution with bismuth cannot be
arried out,

To fill out the mformation on fissionable fuel solubility in bismuth,
Fig. 203 shows the solubility of plutonium in bismuth, as determined at
the Los Akunes National Laboratory. In comparing plutonium with
uranium, it 1= =een that plutonium is significantly more soluble.

20-2.3 Fission-product solubility. The solubilities of most of the impor-
tant fis<ion products have been determined, and are shown in g, 20-1.
In general, all the fission produets are soluble enough so that they will stay
in solution throughout the reactor eyele. This is not true of molybdenum
however,  Attempts at determining the solubility of Mo huve indicated
that it i less than 1 ppm (the Limit of detection) at temperatures below
800°C". =inee o tuir amount of the Mo is produced by fission, this meuns
that o =ludee might form during reactor operation. (Beryllium presents
similar ditticulties, since at temperatures below 800°C the solubility of
Be has been shown to be less than 10 ppm.)

20-2.4 Magnesium and zirconium. The solubility of magnesium in
bismuth 1 the temperature range 100 to 500°C is approximately 5 wt.0g,
which 1= coustderably higher than the amounts of magnesium being con-
sidered 1n this work {300 ppm). Little work has been done on this partie-
ular determination at Brookhaven.

The =olubility of zirconium in bismuth has been determined and is shown
in Fig. 20-5. This information is important in showing that the saturation
solubility of zirconium is very close to the amounts desired for corrosion
inhibition in the temperature range 400 to 500°C.
726 PROPERTIES OF LIQUID-METAL FUELS [crHAP. 20

 

T, °C
560 495 440 394 352 315 283
100
60 ]
40 - ]

     
   

oLt Hl

Atomic %

T \IHH‘

o

o

D
T

 

 

 

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
1000/T, °K

Fic. 20-4. Solubility of fission products in bismuth.

20-2.5 Solubility of corrosion products in bismuth. An alloy steel is
contemplated as the tube material for containing the circulating fuel in
the LMTFR. Hence it has been pertinent to determine the solubility of alloy
steel constituents in bismuth. Figure 206 shows the solubilities of 1romn,
chromium, nickel, and manganese, all of whose solubilities are fiirly high
from a corrosion point of view. Nickel and manganese are particularly
high.

The solubility of titanium is shown in Fig. 20-7. It has been shown [8]
that titanium will reduce the mass-transfer corrosion of steels by liquid
bismuth.

20-2.6 Solubilities of combination of elements in bismuth. The effect
of Zr on the U solubility. The mutual solubilities of uranium and zirconium
in bismuth have been measured over the temperature range 325 to 700°C.
The data are plotted in Fig. 20-5. When bismuth is saturated with zir-
conium. the uranium solubility is appreciably decreased. On the other
20-2] SOLUBILITIES IN BISMUTH 727

 

    
  

636 560 496 441 394 352°C
o———(Barels) Uin Bi n

DGS &JRW}  Uin8i 4

U in Bi saturated with Zr

Uor Zrin Bi, Wt %,

 

 

 

1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7
1600/T, °K

Fre. 20-3. Mutual solubility of uranium and zirconium in bismuth.

e, only o slight decrease is noted in the Zr solubility. The addition of
1000 ppm magnesium had no effeet on either the uranium or zirconium
sulntality. This, of course, is in considerable excess of the quantity of mag-
nestin contemplated for use in the fuel.

The puual solubility effects were further studied by determining the
teroy system U-Zr-Bi at three temperatures, 375, 400, and 425°C.
These ure shown in Fig. 20-8.

The v ieet of fission products on the solubility of U-Bi. Considerable work
hus been done on determining the mutual solubility effect of fission prod-
urts on uranin and bismuth. A good typical example is shown in Fig.
20-9. which <shows that the selubility of uranium and bismuth is affected
b 230 pp Zr, 350 Mg, 60 Nd, 15 Sm, 15 Sr, 10 Cs, and 8 Ru. There is
el doabt that this small amount of fission products, 120 ppm, has a
<tnull bt defimite effect on uranium solubility.

Eifects o adiditives on solubility of corrosion products in liquid bismuth.
The ordinary concentrations of zirconium (250 to 300 ppm) do not affect
the equilibrium iron solubility at temperatures from 500 to 700°C. TFor
728 PROPERTIES OF LIQUID-METAL FUELS [cHAP. 20

Temperature ,°C

 

 
 
  

 

 

 

 

 

 

 

 

 

727 636 560 4% 441 394 352 315
100 T~ T 1 | I
. .-""--.._ Mn Bi stable i
- Ni in Bi 1
- B Ni - Bi) -
=
w O —j 100 <
S —s0 °
@ i i &
£ g
£ B — 30 £
= | . £
2 2
Z
01 I~ —410 £
- 7 o
[ ] a
o l | l j I l | I 1
09 1.0 1 12 13 1.4 15 16 17 18
1000/T °K
Fig. 20-6. Solubility of Fe, Cr, Ni, and Mn in bismuth.
T, °C
727 637 561 497 442 395 353 317
05
| | | l l 1
03 |—
0.10 |—
5 |
z 007 —
= -
005 }—
0.03
0 11 12 13 14 15 16 17

1000 /T, °K
F1g. 20-7. Solubility of titanium in bismuth.
20-2] SOLUBILITIES IN BISMUTH 729

 

  

 

 

 

340 i T T [ T T
Saturation Zr in Bi (425°C)

S00 e —

450 — \\ —

420 \ —
Saturation Zr in Bi (400°C) ‘

80 = CoE —

— \

340 [ --_'-—- \ -

E 300 [~ \ —

e 260 Saturation Zr in Bi (375°C)

N __ \c425°c1 N
220 | \ —
180 = |\(400°c1 \ ]
140 |— \\ —]
100 S— \\ —

-.__
- =" ‘."-..__
60 — - 1 ——
2 | 1 1 L | L
700 900 1100 1300 1500 1700 1900 2100 2300 2500

U, ppm

Fic. 20-8. The U-Zr-Bi ternary system: liquidus curves at 375, 400, and 425°C.

 

 

 

 

636°C 560 496 441 394 352

3.0 l I I I

\\

10 — Normal U Solubility ]
== ]
$ 06— _ ]
5 | UinBi 4+ 250 Zr ppm __|

03 |— UinBi -+ 250 Zr ppm ]

-+ 350 Mg ppm -+
| 120 ppm Mixed ]
Fission Products N
X
N
1.0 1.2 1.4 1.6

1000/T °K

Frg. 20-9. Solubility of U in Bi+ 250 ppm Zr, and in Bi+ 250 ppm Zr 4+ 350
ppm Mg + 120 ppm mixed fission products. Original alloys 3.99, U and 3%, U
respectively,

»
730 PROPERTIES OF LIQUID-METAL FUELS [cHAP. 20

 

      

 

 

 

T, °C
5727 637 561 497 442 395 353 317
10
1 | | ! | |
A\ -
\ ]
i Product of Normal Fe and i
104 k \‘\/'/Normol Cr Solubility 4
- \ =
a Apparent Solubility Product
£ A of Fe and Crin Bi ]
& 13
20 —
s F ]
E L"\ ]
a2 N 4 300
¢ - ~ b
€ INO N
I 10¢ E=™\ — 100
5 BN 317
¥ N ~ ] £
- 430 2
[~ i S
10! {10 s
- 3 @
- ~N q7 <
- N
= ~N Normal 3
= Normal Fe Solubility N\Cr Solubility 4
: L1 N,
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7

1000/T. °K

Fic. 20-10. Effect of Cr on the solubility of Fe in Bi.

higher concentrations (above 700 ppm zirconium), the iron solubility is
increased in this same temperature range.

Zirconium in all concentrations up to saturation does not affect the sol-
ubility of chromium in bismuth.

Uranium, with magnesium additions up to 2000 ppm, does not affect
the solubility of iron in bismuth. The possible effects on chromium solu-
bility are not known at this time.

Chromium has a marked effect on the solubility of iron, whereas the
chromium solubility itself is not affected. An apparent solubility product
is observed as is shown in I'ig. 20-10 by the line titled “Apparent solu-
bility product. Below 450°C, the iron solubility appears to be increased by
saturating the solution with chromium. Above that temperature, the iron
solubility is markedly reduced by chromium.

Titanium, at concentrations greater than 100 ppm, has been found to
reduce the iron solubility in the temperature range 475 to 685°C [9].

20-2.7 Salts. In some of the contemplated chemical fuel processing
methods the liquid bismuth fuel will be brought in contact with chloride
20--5] FUEL STABILITY 731

and fluoride salts. A typical chloride salt is the eutectic mixture of
NaCl-KCl-MgCls. It is important that none of the salts dissolve in the
hismuth and get carried over into the core, since chlorine is a neutron
poison. Preliminary investigations at BNL indicate that the solubility of
these chloride salts is less than the detectable amount, 1 ppm.

20-3. PuysicaL PROPERTIES OF SOLUTIONS

20-3.1 Bismuth properties. The physical properties of bismuth are
listed in Table 23-1.

20-3.2 Solution properties. lLittle work has been done on determining
physical properties of the solutions. The available results indicate that
the small amount of dissolved material does not appreciably affect the
phy=ical properties of density, viscosity, heat capacity, and vapor pressure.
I'or design purposes, the properties of pure bismuth can probably be used
with safety,

20-3.3 Gas solubilities in bismuth. The question of the solubility of the
fission-product gases xenon and krypton in bismuth is of extreme im-
portance. In particular, Xe'3% o strong neutron poison, must be removed
from the system as fast as it is formed in order to have a good neutron
eCOnomy.

Attempts at measuring and calculating the solubility of these gases in
bi=muth have proved extremely difficult, because of the extremely small
<olubilities, Mitra and Bonilla [10] have measured the solubility of xenon
i bismuth at 492°C as 8 X 1077 atom fraction at atmospheric pressure.
On the other hand, MeMillan [11] has calculated the solubility as 10712
atony fraction at 300°C. It 1s probable that the amount of gases produced
i the reactor lies between these two determinations. At present, the ques-
tion of xenon and krypton solubility in bismuth is open to more intensive
reseurch,

20-4. FugeL PREPARATION

I'uel has been prepared at BNL by simply dissolving the solid uranium,
mugnesium, and zireonium in molten bismuth. The solids are usually in
the form of small chips and are placed in a small metal basket which is
then ~uspended in the bismuth.

20-5. TFureL StaBiLiTy

It is essential to maintain a homogeneous fuel and to prevent the uranium
from concentrating in any particular region of the reactor. Stability tests
have been conducted to determine conditions necessary for keeping the
732 PROPERTIES OF LIQUID-METAL FUELS [cuap. 20

uranium in solution by preventing its reaction with the steel and graphite
of the system. Measurements have also been made of the rate of oxidation
of uranium in the liquid fuel stream. This study indicates the effect of an
accidental air leak during the reactor operation.

20-5.1 Losses of uranium from bismuth by reaction with container ma-
terials. FEarly attempts to make up uranium-bismuth solutions resulted
in about a 50% loss of uranium even though very high-purity bismuth
(99.999%) was used. Apparently the uranium reacted with the few im-
purities in bismuth or adsorbed on the walls of steel containers. Sand-
blasting and acid-pickling of the container walls, deoxidizing the bismuth
by hydrogen firing, and adding 250 ppm Zr and 350 ppm Mg before intro-
duction of U reduced this loss to about 5%. 1t is possible that even this
5% loss may not be real, but is attributable to analytical and sampling
techniques.

Only small decreases in the zirconium and magnesium concentrations
have been observed, and in tests where titanium was used as an oxygen
scavenger, no loss of U was ohserved.

When the fuel solution is brought in contact with graphite, usually 10 to
15% of the uranium is lost from the solution. Apparently it reacts with the
graphite or impurities present in the graphite. Research on this is under
way at present. However, it is proving to be extremely difficult since the
amounts of materials involved are so small.

Since zirconium reacts with graphite to form zirconium ecarbide in
preference to uranium forming uranium carbide, addition of zirconium to
the solution should help prevent loss of uranium. This effect has been
observed.

Generally it has been found that zirconium concentration will initially
drop and then maintain a constant level throughout the exposure of the
fuel solution to graphite.

20-5.2 Reaction of fuel solution with air. Should an air leak oceur in
the LMFR, the uranium, magnesium, and zirconium will all tend to oxi-
dize in preference to the bismuth. Figure 20-11 shows the results of an
experiment in which air was bubbled through fuel kept at a temperature
of 405°C. These results indicate that the preference of oxidation is in the
order magnesium, uranium, zirconium,

The reaction data indicate that the uranium oxidation rate is one-half
order dependent on the UOz present. The magnesium oxidation rate, in
general, is first order with respect to magnesium concentration. Other
experiments show that if additional amounts of magnesium are added to
the solution after the oxidation, most of the UOy can be reduced back to
uranium. These data are given in Table 20-1.
20-5] FUEL STABILITY 733

 

100

IZirconium l

   
   
  

20

80

70

60

50

Magnesium

Percent of Original U, Mg, or Zr Remaining in Solution

30 —

 

 

2 [ | | 1
0 4 8 12 16 20 24
Time in Minutes

 

F1c. 20-11. Concurrent oxidation of U, Zr, and Mg from Bi containing 750 ppm
U, 284 ppm Mg, and 280 ppm Zr.

TaBLE 20-1

RepuctioN or UQOs, BY Mg 1nv Bl

 

U (ppm) U (ppm)
present in . U (ppm)
as UO2 . Time after | . )
o solution | Mg (ppm) . | in solution | 9 UO2*
T°C | before ad- Mg addi-
dit; before ad- added . after Mg | reduced
ition of . tion
; Mg dition of added
1 . Mg
405 960 10 6600 25 min 710 75-100
400 150 530 5000 48 hr 660 90
360 510 310 2500 10 hr 460 30
360 550 10 1000 48 hr 290 50

 

 

 

 

 

 

 

 

 

 

*The values listed as 9, UQ2 reduced are probably lower than equilibrium
values, since the samples were taken at arbitrary times after the Mg was added.
734 PROPERTIES OF LIQUID-METAI, FULLS [caap. 20

Work on fuel stability ix obviously of great importance, and ix being
continued.  Very little has been done so far on observation of stability
under neutron bombardment. A program is getting underway for the study
of radiation effects on the fuel concurrently with a study of corrosion
effects. For this purpose the Brookhaven Pile will be used together with
Radiation Effects Loop No. 1.

20-6. Tioritm BisMUTHIDE BLANKET SLURRY

20-6.1 Status of development. In developing a blanket system for the
LMIER, 1t has seemed logical to select one which is as similar as possible to
the core fuel.  After considerable evaluation the prineipal emphasis has
been placed upon a bismuth fluid containing thorium bismuthide in the
form of very small particles, This is commonly called the thorium bis-
muthide slurry system.

Since this fluid has practically the same physical properties as that of
the core, it would be possible to balance pressures across the graphite wall
separating the blanket from the core and, in the event of mixing the core
and blanket fluids, no violent reactions would ensue. Furthermore, from a
chemical processing point of view, an all-metallic blanket system offers
considerable advantage when pyrometallurgical processing techniques
are used.

This does not mean that other types of blankets are not being studied.
Work is coneurrently under way on thorium oxide-bismuth slurries. Also,
thorium carbide, thorium fluoride, and thorium sulphide slurries are under
consideration.

At the Ames Laboratory (Iowa State College) the solution of thorium
I magnesium has received considerable attention in the past few years.
This is a true solution, and certainly offers another possibility for a blanket
fluid. However, unless an absolute method for keeping the magnesium
solution separate from the core bismuth solution is found, this svstem
would be hazardous when used with the contemplated uranium-bismuth
core fluid, since magnesium and bismuth will react violently and cause
a marked temperature rise.

20-6.2 Chemical composition of thorium bismuthide. "The thorium
bismuthide intermetallic compound discussed in this section has the chemi-
:al formula ThBio. This compound is 35.7 w/o thorium. A sccond com-
pound, ThsBiy, also can exist and has been observed in alloys containing
greater than 50 to 55 w/o thorium.

20-6.3 Crystal chemistry of thorium bismuthide. ThBi; has a tetragonal
crystal structure (with ap = 4.942 A, and ¢y = 9.559 A) containing two tho-
rium atoms and two bismuth atoms per unit cell. The density as determined
20-6] THORIUM BISMUTHIDE BLANKET SLURRY 735

by x-ray measurement, is 11.50 g/cc at 25°C. It is estimated to be approxi-
mately 11.4 g/ce at 550°C.

Th3Bis has a body-centered cubic structure (ap = 9.559 A) containing
12 thorium atoms and 16 bismuth atoms in the unit cell. The density is
11.65 g/ce.

Ordinarily, when thorium bismuthide is prepared at 500°C, very small
equiaxed particles (less than 0.5 micron) are formed. These equiaxed
particles grow until they reach the average size of 50 to 60 microns, and
under certain conditions they can grow to considerably larger dimensions.

When a 5 to 10 w/o thorium bismuthide slurry is cooled from a tempera-
ture of complete solution (above 1000°C), ThBiz precipitates in the
form of platelets having diameter-to-thickness ratios greater than 50:1.
The plane of the platelet is parallel to the 001 plane of the crystal. Platelet
diameters up to 1 em have been observed in alloys cooled at moderate
rates. The diameters can be decreased by increasing the cooling rate.

Whereas equiaxed particles tend to grow equally in all three dimensions,
it has been found that the platelets, when heated isothermally at tempera-
tures above 300°C, tend to grow faster in thickness than in diameter. The
solid particles thus tend to approach an equiaxed shape. The rate of ap-
proach to equiaxiality inecreases as the temperature of isothermic treat-
ment 1< increased.

("onsiderable work has been carried out on control of crystal structure
and =ize. The addition of tolerable amounts of Li, Be, Mg, Al, 51, Ca, Tj,
Cr. Mn, Fe, Co, Ni, Cu, Zn, Zr, Mo, Pd, Ag, Sn, Sh, Te, Pa, La, Ce, Tr,
Nd, Ta, W, Pt, Pb, and U has little effect on the mode of thorium bis-
muthide when it 1s precipitated. It has been found, however, that tellurium
inhibits the thorium bismuthide particle growth, agglomeration, and de-
position during thermal cycling. The platelet mode of bismuthide precipi-
tation 1< not modified by addition of tellurium. The amount of tellurium
used 1n these experiments has been 0.10 w/o.

The mechanisms by which tellurium additions inhibit ThBis particle
arowth, agelomeration, and deposition are as yet uncertain. Although
additions of tellurium in larger concentrations decrease the solubility of
thorium in bismuth markedly, the coneentration of tellurium required for
inhibition decreases the solubility only slightly. These small amounts of
tellurium appear to be associated with the solid phase rather than the
liquid phase. They do not appear to alter the erystal structure.

It has been observed that under certain conditions ThBig particles
suspended in liquid bismuth can be pressure-welded to one another and to
container materials by the forces of impact. This pressure-welding phe-
nomenon has been observed at 525°C and higher temperatures. Since this
phenomenon might cause plugging by agglomeration at points of high
impact, it will be necessary to take this factor into account in the design
of slurry circulation systems.
736 PROPERTIES OF LIQUID-METAL FUELS [crar. 20

20-6.4 Thorium-bismuth slurry preparation. Dispersions of small equi-
axed particles of ThBiz in bismuth ecan be prepared by heating finely
divided thorium, in the form of powder or chips, in contact with liquid bis-
muth at 500 to 600°C under an inert atmosphere. The intermetallic com-
pound is formed by an exothermic reaction at the thorium-bismuth
interface, when the convex radius of curvature of the thorium surface is
suitably small. The compound exfoliates into the liquid as agglomerates of
very small particles (less than 0.5 micron). These small particles grow
very rapidly, the larger at the expense of the smaller, as equiaxed single
crystals of ThBis. Rapid growth ceases when the maximum crystal di-
mensions approach approximately 50 to 60 microns. The time necessary
for complete reaction varies with the dimensions of the thorium. For
example, 325-mesh thorium powder reacts completely in 5 min at 500°C,
thorium chips 1/2” X 1/8”7 % 0.010" require 2 hr at 500°C, and thorium
chips 3/4"" X 3/16” X 0.020" require 13 hr at 500°C. The thorium dimen-
sions have only a slight effect upon the ultimate particle size. The reaction
can be accelerated by raising the temperature. Higher temperatures,
however, increase both the particle size and the tendency to form sintered
agglomerates rather than single crystals.

If thorium powder is added to the liquid bismuth surface at the reaction
temperature, it is necessary to stir the thorium into the liquid. Otherwise
a crust of intermetallic compound forms on the surface which is rigid enough
to support subsequent additions, thus preventing contact between the
thorium and the bismuth.

During the reaction, evolution of an unidentified gas (possibly hydrogen
from thorium hydride) has been observed. It is necessary to stir the slurry
under vacuum to remove the undesirable trapped bubbles of this gas.

A photomicrograph of a typical slurry produced by the exfoliation
method is shown in Fig. 20-12. The dark ThBiz particles appear in a
white matrix of solidified bismuth. The method has been used to prepare
90-1b batches of slurry and may readily be adapted to tonnage-scale prepa-
ration. The method is suitable for preparation of the initial blanket charge,
but would probably not be used for slurry reconstitution during subsequent
blanket processing.

A modification of this method has been studied in which finely divided
thorium {from a supernatant mixture of fused chlorides is electrolytically
deposited on a molten bismuth cathode at the desired temperature [13].
The thorium must be stirred through the interface. Slurries that are satis-
factory with respect to thorium content and particle size and shape have
been produced by the electrolytic method in batches of up to 10 1b. No
evolution of gas has been detected during the thorium-bismuth reaction.
Unfortunately, the necessary stirring introduces chloride inclusions which
are difficult to remove completely. Since these inclusions would decrease
20-6] THORIUM BISMUTHIDE BLANKET SLURRY 737

 

Fie. 20-12. 5 w/0 Th-95 w/0 Bi. Dispersion of equiaxed ThBi» particles in Bi
Produced by heating Th chips in Bi at 500°C for 2 hr. (150x)

the efficiency of neutron utilization in a breeder blanket because of the
high cross section of chlorine, the electrolytic method of slurry preparation
must, at present, be considered unsatisfactory.

Another preparation method for thorium-bismuth slurry is by quenching
and heat treatment. In this method a solution of thorium, for example
5 w o, 18 very rapidly cooled from about 1000°C down to about 600°C.
This can be accomplished by pouring a hot solution into a container having
a sufficiently high heat capacity or by pouring the hot solution into an
equal volume of liquid bismuth heated just above the melting point. When
this 1s done tiny platelets are formed.

As will be discussed in the following section, the platelet form of crystal
is unsatisfactory from a fluidity point of view. When these fine platelets
are heat-treated for 20 min at 800°C, or for 5 min at 900°C, dispersions of
ThBi» particles having maximum dimensions less than 100 microns and
diameter-to-thickness ratios equal to or less than 5 to 1 are produced.
Platelet formation during cooling is avoided by agitating the slurry to
suspend the particles.

Figure 20-13 shows the fine platelets produced by the quenching and
Fig. 20-14 shows the larger particles produced from these fine platelets
by the heat treatment at 800°C for 20 min. Such a slurry exhibits high
fluidity after concentration to 10 w/o thorium by removal of excess liquid
phase, and is suitable for use in the reactor blanket.

Other possible ways for reconstituting a satisfactory slurry after heating
to complete solution involve the use of ultrasonic energy [14]. It has been
demonstrated that application of ultrasonic energy to a thorium-bismuth
solution during cooling causes the formation of essentially equiaxed particles
rather than platelets. It has also been demonstrated that application of
738 PROPERTIES OF LIQUID-METAL FUELS [cHAP. 20

T T i . S
i .- i o e YA
. ,W # »f* - ks SN
Nfi“" ‘Y, L,J‘ » & § . Ve .
. w0
« ,;fiéf“ " ot e, *y!‘i L
L. S ' - v s
v‘ - ,(é.,»" a : ‘;#& g i # 3
L 1 w Fon 7 : g"; v
P 5 - :\ L tx
o BT e : o e . S
vy &r‘i E,’ « - . .E: 5 e
N R - - - v : pn - * 4 .s/
\ :;,/ ‘¢ & ST \ ! J éj,‘
= . N H g
. PP AN A
. L <
ot 3 S0 ; ~ ¥ "‘.:“"\‘1
. L - - . T 4
% {s ~\ ¥ T i g ‘:'E Tyt
T . e gl f‘:?;/?_‘ i s
x .7 - * % .
v wle oL T e s .
3y ool g BN
L, Coe L - ¥ Gt ey
T ¥ s - » -*fi» RS
¥ W __»t iy L i \% * s o
I o . L F
T LT ‘,;’f‘ > 2Nt T o
B Aee Cae e
- "R i * Lo Tl e Tl
T N R N

Fra. 20-13. 5 w/o Th-95 w/o Bi. Dispersion of ThBis platelets in Bi. Alloy
heated to 1000°C and quenched by pouring into graphite crucible at 25°C. (150x)

 

- L.y S

Fic. 20-14. 10 w/0 Th-90 w/o Bi. Dispersion of reconstituted ThBi2 particles
in Bi. Produced by heating fine-platelet dispersion to 800°C for 20 min. (150%)

ultrasonic energy to platelet dispersions causes the platelets to break up
into essentially equiaxed fragments.

20-6.5 Engineering studies of slurries. The intermetallic compound
ThBig is quite soft, having a Rockwell 15-T hardness of approximately
60 at room temperature. It is brittle at room temperature but appears to
exhibit some ductility at 400°C. The compound is pyrophoric and must be
protected against oxidation.

When slurries of equiaxed bismuthide in bismuth are prepared, they are
fluid at temperatures above the melting point of bismuth, 271°C. In
these slurries the solid phase is in thermodynamic equilibrium with the
20-6] THORIUM BISMUTHIDE BLANKET SLURRY 739

liquid phase and is perfectly wetted by it. At the proposed reactor tem-
peratures (350 to 550°C) practically all the thorium in the slurry appears
in the solid phase, since the solubility in the liquid is very low.

The ideal slurry composition represents a balance between a desire for a
high thermal neutron utilization factor (i.e., a high thorium content) and
the necessity for high fluidity. IFluidity studies have shown that the upper
limit of thorium concentration for high fluidity at reactor temperatures is
approximately 10 w/o of thorium. This corresponds to 24.9% by volume
of ThBis2, and a thermal neutron utilization factor of 0.957. Although the
viscosity of Th-Bi slurries has not been measured, ealculations based on
the viscosity of liquid bismuth and the behavior of similar systems indicate
that at 530°C the viscosity of a 10 w/o0 Th suspension of 50-micron, equi-
axed ThBi. particles should be approximately 2.5 centipoises. It has been
observed that increasing the thorium content beyond 10 w/o Th causes a
disproportionately large increase in the viscosity, so that the consistency
approaches that of a mud or paste. The maximum thorium concentration
for high fluidity decreases when the ThBi» particle shape departs signifi-
cantly from an equiaxed shape.

The density of liquid bismuth varies from 9.97 at 350°C to 9.72 at 550°C,
and should not be changed appreciably by the small amount of thorium
dix=olved at these temperatures. Therefore the solid particles should sink
in the liquid. Although settling rates have not been measured, the mag-
nitude of expected settling rates can be caleulated. The setiling rate for
100-micron spheres at 330°C, as caleulated by Stokes’” Law, is 0.030 fps.
The settling rate in a 10 w/o Th-B1 dispersion of 100-micron spheres at
550°C, us ealeulated by the hindered settling equation, 1s 0.0026 fps.

It has been observed in small systems that equiaxed ThBia particles
<ettle to n relatively stable layer in which the thorium concentration is
approximately 15 w/0 Th. Such layers can be redispersed by mild agita-
tion of the supernatant liquid. When the thorium concentration in the
<ettled laver is Increased to 18 to 20 w/o Th (by centrifugation, for ex-
ample), the viscosity of the layer is so high that mechanical agitation of
the layver itself is necessary to redisperse the particles.

Fxperiments have shown that the viscosity of a 10 w/o Th slurry, using
platelets of 50- to 100-micron size, is so high as to make the slurry com-
pletely unsuitable for use.

Sherry behavior under conditions of reactor blanket operation. It 1s antici-
pated that the slurry would be contained in a low-permeability graphite
within the reactor blanket. IFor heat removal and processing, the slurry
would be circulated externally through pipes and heat exchangers fabri-
cated of low-chromium steel or comparable material. During circulation
for heat removal, the slurry would be subjected to thermal eycling between
a probable maximum temperature of 550°C in the blanket and a possible
740 PROPERTIES OF LIQUID-METAL FUELS [cHAP. 20

minimum of 350°C in the heat exchangers. Capsule and pumped-loop ex-
periments have been carried out to study the behavior of the slurry under
conditions of thermal eyeling and flow.

In the capsule experiments, small specimens of slurry are caused to flow
back and forth at 6 cyecles/min in periodically tilted tubes fabricated of
the container material under test. The tubes, which are sealed under
vacuum, are heated to a higher temperature at one end than at the other.
When specimens of slurry containing 10 w/o Th, with and without addi-
tions of 0.025 w/o zirconium, were cycled between 350 and 350°C in 2197,
Cr-19, Mo steel tubes, nearly all the ThRBi» was deposited-in the cooler
ends of the tubes in less than 500 hr. Fxamination of the deposits disclosed
that a deposit due to mass transfer of the steel had formed on the tube
walls prior to deposition of the ThBi>. This suggested that mass transfer
of the steel may have been instrumental in starting the ThBis deposition,
perhaps by roughening the walls or perhaps by altering the composition
of the tube surface.

Specimens of 5 w/o Th slurries have been cycled for 300 hr between 350
and 580°C in graphite tubes with no evidence of plug formation. In these
experiments, a relatively rapid increase in ThBis particle size (from 50 to
225 microns in 500 hr) was observed. This increase was due to particle
agglomeration rather than growth of single erystals. No evidence of graph-
1te erosion was observed.

Specimens of slurries containing 10 w/o thorium and 0.10 w/o tellurium
have been cyceled between 350 and 580°C in graphite, and between 350
and 550°C in 239, Cr-19, Mo steel for 500 hr with no evidence of ThBiy
plug formation or mass transfer of the steel. The speeimens showed no
inerease in the maximum particle dimension and no particle agglomeration.
When a specimen of slurry containing 10 w/o Th, 0.10 w/o Te was cycled
at higher temperatures in a 239 Cr-19, Mo stecl tube, mass transfer of
steel and deposition of ThBis in the cooler end were observed after less
than 100 hr.

Slurries containing up to 7 w/o Th and minor additions of zirconium
have been circulated through small 239, Cr-19,, Mo steel loops by means
of a propeller pump. Isothermal circulation at 450°C has been carried
out for more than 450 hr at velocities between 0.3 and 1.5 fps, with no
difficulty in circulation or maintaining suspension. Attempts to circulate
these slurries through a temperature differential, however, have resulted
in the formation of ThBiz> deposits in the coldest section of the loop. In
a modified loop containing a graphite liner in the finned-cooler section,
1sothermal cireulation was maintained without difficulty. ThBiz, however,
again deposited in the finned-cooler section when a temperature differential
was applied.

When a slurry containing 7 w/o Th. 0.025 w/0 Zr, and 0.10 w/0 Te was
20-7] THORIUM COMPOUND SLURRIES 741

circulated in a 219, Cr-19,, Mo steel loop through a temperature differ-
ential, ThI3iz deposited in the finned-cooler section. The rate of buildup
of the deposit was markedly less than in the case of slurries containing no
tellurtum,

The problem of ThBis deposition during circulation through a tempera-
ture differential is one which must be solved before the Th-Bi slurry is
aceeptable as a {luid breeder-blanket material. The favorable results ob-
tained by tellurium additions in the capsule experiments offer hope that
the problem can be solved.

20~7. Tuoritm CoMPOUND SLURRIES

20-7.1 Thorium oxide. Probably the best blanket material, next to the
thorium bismuthide slurry, 1s the suspension of thorium oxide in bismuth.
The thorium-oxide shurry should be compatible with the graphite and steel
in the reactor structure. Iixperiments have shown that ThOs is wetted
by the liquid bismuth if some zirconium or thorium iz dissolved i the bis-
muth. =lurries of 10 w/o thorium oxide have been prepared.

The separation of thorium oxide from the liquid bismuth for processing
could be achieved by mechanical means, and the oxide could then be
proceszed by the existing Thorex process.

The thorium-oxide hlanket slurry is gaining increased attention. A
loop of several pounds per minute capacity has been completed for forced
cireulation of the oxide slurries at BNL and an 800 Ib/min loop is ready
at Babeock' & Wilcox.

20-7.2 Other thorium compounds. A small amount of attention has
been directed toward ThCs, ThS, and Thl'y slurries in bismuth, How-
ever, the major effort is on the thorium bismuthide and thorium-oxide
<lurries.
742 PROPERTIES OF LIQUID-METAL FUELS [cHAP. 20

REFERENCES

L. J. R, Werks et al., Corrosion Problems with Bismuth-Uranium Fuels, in
Proceedings of the Firslt International Conference on the Peaceful Uses of Atomic
Energy, Vol. 9. New York: United Nations, 1956. (P/118, pp. 341-355); D. H.
Gurinsky and G. J. Diens, Nuelear Fuels. Princeton, N. J.: D. Van Nostrand
Co., Inc., 1956. (Chap. XTIT); J. R. Werks, Metallurgical Studies on Liquid
Bismuth and Bismuth Alloys for Reactor Fuels or Coolants, in Progress in Nuclear
Energy, Series IV, Technology and Engineering, Vol. 1. New York: Pergamon
Press, 1956. (pp. 378-408)

2. J. E. Arnerron et al, Studies in the Uranium-Bismuth Fuel System, in
Chemrcal Engincering Progress Symposium Series, Vol. 50, No. 12. New York:
American Institute of Chemiecal Engincers, 1954, (pp. 23-37); Nucleonics 4(7),
40-42 (1954).

3. D. H. Aumany and R. R. BawpwiN, The Urantum-Bismuth System,
USAEC Report CT-2961, Iowa State College, 1945.

4. Massacnuserrs INSTITUTE oF TrcHNOLOGY, Progress Report for the Month
of October 1946, USAEC Report CT-3718.

5. D. W. Barurs, Liquid Reactor Fuels: Bismuth-Uranium System, USAEC
Report BNL-75, Brookhaven National Laboratory, 1950,

6. R.J. Trrrer, Uranium-Bismuth System, J. Mefals 9, 131-136 (1957).

7. G. W. GREENWOOD, personal communication to J. R, Weeks, Aug. 29, 1957.

8. 0. 1. Evgerr and C. J. Ecan, Dynamic Corrosion of Steel by Liquid Bismuth,
USALEC Report MTA-12, California Rescarch and Development Co., 1953.

9. J. R. Weeks and D. H. Gurinsky, Solid Metal-Liquid Metal Reactions
in Bismuth and Sodium, in ASM Symposium on Liquid Metals and Solidification,
ed. by B. Chalmers. Cleveland, Ohio: The American Socicty for Metals, 1958.

10. C. R. Mitra and C. F. Bontura, Solubility and Stripping of Rare Gases
in Molten Metals, USAEC Report BNT.-3337, Columbia University Department
of Chemical Engineering, June 30, 1955.

11. W. G. Mcwmiuoan, Estimates of the Solubility and Diffusion Constant of
Xenon tn Liguid Bismuth, USAEC Report BNIL-353, Brookhaven National
Laboratory, June 1955.

12. M. E. Seserr, Investigation of Methods for Preparation of Thorium
Bismuthide Dispersions tn Liquid Bismuth, Final Progress Report, Horizons,
Inc., Oct. 31, 1956.

13. ArrorrosecTs, Inc., Applications of Ulirasonie Energy, Progress Report
No. 4, USAEC Report NYO-7918, 1957.