188 Journal of Nuclear Materials 155-157 (1988) 18X-201

North-Holland. Amsterdam

CERAMIC BREEDER MATERIALS

C.E. JOHNSON
Argonne National Laboratory3 Argonne, IL 60439 USA

K.R. KUMMERER
Kernforschungszentrum Karlsruhe, Po~tfac~ 3640, O-7500 Kar~s~he~ Fed. Rep. Germany

E. ROTH
Commissariat a IEnergie Atotnique, Saclay, France

Ceramic materials are under investigation as potential breeder material in fusion reactors. This paper will review candidate
materials with respect to (a) fabrication routes and characterization, (b) properties in as-fabricated and irradiated condition,
and (c) experimental results from laboratory and in-pile investigations on tritium transport and release. Also discussed are the
resources of beryllium, which is being considered as a neutron multiplier. The comparison of ceramic properties that is
attempted here aims at the identification of the most-promising material for use in a tritium breeding blanket.

1. Introduction ture limit is based upon thermal sintering or restructur-

ing effects. A high tritium solubility can be detrimental
Current fusion reactor designs include a tritium because it indicates potential for high blanket inven-
breeding blanket for production of tritium. To be con- tory. A high lithium atom density and thermal conduc-
sidered, a candidate material must (a) release tritium, tivity, wide operating temperature range, and low tri-
(b) exhibit thermophysical, chemical, and mechanical tium solubility are all favorable quantities. Accordingly,
stability at high temperature, (c) be compatible with L&O appears favorable because of its high lithium atom
other blanket and structural materials, and (d) possess density and thermal conductivity and its low tritium
desirable irradiation behavior. Lithium, in some form, is solubility; LiAlO, appears almost as favorable because
the only material suited for this task. Present candidates of its broad operating temperature range. Except for
for the breeding material are liquid lithium and its Li,O, all other candidates are certain to require a
alloys and certain lithium-containing solids, principally neutron multiplier due to their low lithium atom den-
the lithium-containing ceramics. This paper will focus sity.
only on the latter materials, namely, the oxide, y- Comparisons of the candidates can also be made by
aluminate, silicate, and zirconate of lit~um. examining the rate of change with temperature in the
In materials evaluation, tabulations of properties value of a given property. Fig. 1 displays the rates of
data are critically examined. Table 1 lists the melting change of several properties as a function of tempera-
point (T,), lithium atom density (pLi), bulk density ture. The figure is intended only to show trends to
( prr,), thermal conductivity (K), operating temperature achieve acceptable performance; no attempt has been
range, and tritium solubihty of the candidate materials. made to quantify the axes. The steepness of the slopes
The lower operating temperature limit is based upon for a given property is intuitive and will only be firmly
estimates of tritium diffusion, and the upper tempera- established once data become available. Performance

Table 1
Property values and temperatures for candidate solid breeder materials

Breeder Property Operating Tritium

temperature solubility )
PTD
range (K)
2) $ycrn ) (g/cm3 ) &/mK)
Li,O 1706 0.93 2.01 3.4 673-1073 L
LiAlO, 1883 0.27 2.61 2.2 723-1473 L
Li 2SiO, 1473 0.36 2.53 1.5 683-1073 H
Li 4SiOe 1523 0.54 2.35 1.5 583-1223 M
Li,Zr03 1888 0.33 4.15 1.3 673-1673 M
Li,ZrOh 1568 0.68 2.99 1.5 623-1253 M

) L = low, M = moderate, H = high.

0022-3115/88/$03.50 0 Elsevier Science Publishers B.V.

(North-~oiland Physics Publis~ng Division)
 C. E. Johnson et al. / Ceramic breeder materials 189

MASS TRANSPORT _

SoLualLlTY
AND TRAPPING

GRAIN GROWTH
AND SINTERING

TEMPERATURE

Fig. 1. Breeder material properties as a function of temperature.

must be considered to represent a measure of the overall upon possible fabrication routes, their properties in the
operating performance of a given breeder material. A as-fabricated and irradiated condition, and experimen-
most interesting aspect of this figure is that tritium tal results from laboratory and in-pile investigations on
diffusion is the only property exhibiting a positive re- tritium transport and release.
sponse to temperature.
Transformation of materials during neutron irradia- 2. Fabrication of breeder material compounds
tion must also be considered because the neutron cap-
ture process eliminates one chemical component (Li) The geometrical shape of the ceramic breeder body
and produces two others (T, He). These compositional depends on the design concept chosen for the breeder
changes can lead to the formation of second phases blanket module. There are several configurations for
either as internal precipitates or as solid solutions. The solid breeders: monolithic block, pressed and sintered
impact of a change in composition for a ternary oxide is pellets or plates, and spherical (or sphere-pat) particles.
likely to be reflected in changes in its properties, e.g., Thermal stress-induced cracking and fragment reloca-
melting point, tritium diffusivity, thermomechanical tion impose difficulties in precise temperature control
performance. Further, in any detailed evaluation of of monolithic blankets. To overcome the cracking prob-
candidate blanket materials, strong consideration must lem and its associated gap conductance uncertainties,
be given to their tritium breeding potential. Currently, blanket designs can employ either pressed and sintered
blanket designs use beryllium for neutron multiplication pellets and plates or the sphere-pat configuration. With
to ensure an adequate tritium breeding ratio (tritium the BIT (Breeder In Tube) concept, the lithium ceramic
atoms produced per fusion neutron). is embedded within a cladding, e.g., as cylindrical pins,
The performance parameters considered are not of being cooled from outside. With the BOT (Breeder Out
equal importance. The tritium breeding ratio is a key of Tube) concept, the situation is the reverse: the coolant
parameter since tritium self-sufficiency is a necessary is in tubes that penetrate the breeder. Hence, the geom-
requirement. A safety concern is the tritium inventory etry of the breeder ceramic is the pellet shape for the
within the blanket. The large uncertainty in tritium-re- BIT concept and a sphere-pat for the BOT concept. The
lated properties makes data comparisons difficult. In fabrication of ceramic breeders into the desired form
this paper, we assess the candidate materials based needs to be demonstrated. Attention also needs to be
190 C. E. Johnson et al. / Ceramrc breeder materials

+ Ca(Oi-02 LiOH-Solution
-.i

I irj

Calc at 900 YI

Mdlrng

Sintergrade Ceramic Powder

Fig. 2. The preparation of sinter-grade lithium ceramic powders,

placed on microstructural tailoring (e.g., controlled den- methods were used to prepare MgO-doped lithium
sity, grain size, pore morphology), on impurities, and on aluminate. Also, the preparation of y-LiAlO, from
stability to ensure the development of solid breeders liquid mixtures of aluminum and lithium salts, as well
with superior properties. as from pasty mixtures of Al,O, and lithium salts. was
For material fabrication, two steps are considered: examined [2,3]. In a wet process, aluminum metal chips
(a) the chemical conversion of the commercially availa- are dissolved in aqueous LiOH solution, spray-dried,
ble source material (Li,CO, or LiOH) into the desired and calcined at 900 OC to a sinterable y-LiAlO, powder
oxide (fig. 2), and (b) the compaction of the powder to [41.
the desired shape (fig. 3). y-LiAlO, may also be prepared (1) by spray-drying
The preparation of sinterable powders from Li,CO, of lithium aluminum sols [5-71, (2) by microwave
and LiOH source material is schematically presented in denitration of the lithium aluminum nitrate solution as
fig. 2. The wet processes (using spray-ding) avoid successfully demonstrated for fission reactor fuels {S,9],
introduction of impurities in the course of the dry or (3) by agglomeration of oxide powders. The spray-
milling step. The source materials normally contain drying of lithium-aluminum sols was found to give
lithium with natural isotopic composition, i.e., 7.5% Li-6 good results; however, the Li/Al stoichiometry fre-
and 92.5% Li-7. For experiments that require other quently varied. The flow-sheet for large scale (kg) pro-
isotopic compositions either depleted or enriched in duction by this method is still under examination.
Li-6 content, Li,CO, or LiOH source material is availa- There are several silicates of lithium to be consid-
ble from special isotope suppliers. ered. The metasilicate (LiaSiO,) and the orthosilicate
The fabrication of LiAlO, aims at producing the (Li4Si04) are well known and thoroughly investigated.
high-temperature y-phase. In a dry process, Li,CO, As recently demonstrated, Li,SiO, exists under certain
and Al,O, powders are milled together and heated to conditions [lO,ll] but is difficult to prepare and has a
700C. At this temperature, a slow reaction leads to very limited stability range.
formation of the denser a-phase material, which is A standard method for the synthesis of Li,SiO, is
completely transformed into the y-phase at tempera- the co-milling of Li,CO, with highly reactive SiOZ
tures above 900C. Control of grain size and porosity powder in an aqueous mixture followed by heat treat-
is obtained by adjustments in the mixing ratio of start- ment between 500 and 750C for many hours [12-141.
ing materials and the sintering temperature [l]. Similar Different intermediate drying techniques, e.g., spray-
 C.E. Johnson ef al. / Ceramic breeder materials 191

drying or slowly drying in a kneader, lead to powders of

Sintergrade Ceramic Powder
quite different sinterability. Final reaction products
1. __
often require repeated milling and calcining. The lengthy Extrusion /
milling may also cause an increased impu~ty level. All ._. .~
Melting
these drawbacks may be avoided by a recently devel- :::__c. ..
oped wet chemical process that starts with LiOH [4,15]. Granulation

In this process, a LiOH solution is slowly fed into a I ~~1 ---~-- .J

highly dispersed aqueous suspension of SiO,. A final Sintermg __ Drop
spray-drying and calcining at 600 C lead to a sinter- TV _-Solidifmtion
able Li&O, powder.
The lithium orthosilicate can be synthesized in a
manner similar to the wet process for Li,SiO,. The
I
starting mixture, of course, contains correspondingly 1 I I

more LiOH according to the required stoichiometric Fig. 3. Fabrication of solids from ceramic powder
relationship. In the initial precipitation step, Li,SiO, is
formed with excess LiOH; the remaining hydroxide is
either oxidized to Li,O, by introduction of H,O, or is commonly used in mass production of ceramic
transformed to dispersed carbonate by CO,. The con- materials. The sintering temperature varies between 800
cluding calcining step at 6OOC results in sinterable and 1100 o C and is an important factor for controlling
Li,SiO, powder. Without this step, the final product the pellet structure. The sintering time is of the order of
would be quite coarse and have very limited sinterabil- 5 h, depending on preconditioning of the powder.
ity. In addition to this wet process, dry mixing and Hot pressing in a temperature range between 500
calcining in vacuum at very high temperatures have and 900C may lead to compacts of very high density.
been successfully applied. Other wet processes using In this more expensive process, the original grain size of
alcohol instead of water [lo] also seem attractive. Spe- the powder will not increase due to the lower temper-
cially doped orthosilicates (with Al- and P-ions) have ature and the short exposure time [14].
been prepared by suspension in alcohol [16]. Also an The fabrication of spherical particles can be carried
organometallic process 1171 resulted in a very fine out by the granulation of extruded material. Ideal
grained Li,SiO, powder. spherical shapes could be obtained by melting and
For the fabrication of Li,O several methods are freezing. However, free flowing particle beds that are
applicable. In a customary process, the source material normally desired can be formed satisfactorily using
(Li,CO,) is calcined in vacuum and ball-milled to a pebbles of irregular size.
sinterable powder [2]. A more favorable process, as far In qualification of the fabricated ceramic breeder,
as the impurity content is concerned, starts with an the following factors are important: shape (dimensions),
aqueous LiOH solution followed by precipitation of the density, chemical composition. isotope content, impur-
peroxide (Li,@) with H,Oz, which is converted to ity content, crystallographic phases, and microstructure
sinterable Li,O powder [18] by vacuum-calcining at of the solid. The geometry and density of pellets, to-
300-400 o c. gether with the theoretical density of the compound,
A similar procedure is used for the preparation of can be used to derive a direct measure for the average
Li,ZrO,. The intermediate compound Li,O, is mixed porosity. The open porosity is measured by mercury
with ZrO, powder and calcined at 900C [IS]. A porosimetry and can be correlated to the effective intru-
preparation by the organometalhc process mentioned sion channel diameter [19]. The latter can be evaluated
above [17] has also been reported. from the cumulative open porosity by a modeling calcu-
The quality of a sinterable powder is primarily de- lation. The chemical and isotopic composition of the
fined by its specific surface area. For example, Li,SiO,, ceramic is determined by routine analytical methods.
produced by spray-drying is easily processed to yield The crystallographic structure is determined by X-ray
homogen~us ceramic compacts [El. To obtain open diffractomet~. The microst~cture is primarily char-
pore structures, further preparational steps like granu- acterized by the internal geometry of the ceramic
lation and subsequent calcination are necessary. solid, which is highly important for tritium transport
For compaction of the ceramic powder to shaped and also for mechanical strength. The microscopic grain
solids, different procedures are applicable, especially (a) structure is visualized by ceramographic investigations.
cold-pressing and sintering, (b) precompacting and hot In this context the grain size and shape are important
isostatic pressing, (c) extrusion, granulation, and sinter- because they relate to transport and release of tritium.
ing, and (d) melting, liquid drop formation, and solidifi- The adsorption of tritium on internal surfaces of the
cation. These procedures, as shown in the scheme of fig. ceramic can be responsible for retention of tritium.
3, result either in pellets or in spherical particles, Ceramographic structural data can be converted to the
The simple procedure of cold-pressing and sintering relevant specific surface area within the ceramic solid
192 C. E. Johnson et al. / Ceramic breeder materials

using simple modeling methods. Also special char- in turn, is better than that of Li,SiO,. At high temper-
acterisation methods like photoacoustic analysis are re- atures and very low oxygen activities, Li,O would be
ported [20]. The thermomechanical behavior of sphero- better than LiAlO, with respect to tritoxide retention in
dized particle beds requires additional attention [21]. the condensed phase [30].
Partial pressures of the gaseous tritium-containing
3. Materials properties species T2 and T,O were calculated over Li 20, Li,SiO,.
and LiAlO, at 900 K and 1300 K for the oxygen
Properties data presented in this chapter are based activity range between 1Om25 and 1O-5. Results show
on comparative measurements from various studies that both T2 and T,O partial pressures are greater over
[11,22-251; a rigorous comparison and assessment of LiAlO, than over Li,O and Li,SiO, at both temper-
candidate material can only be achieved using values atures in the oxygen activity range between 10m2a and
measured under comparable conditions and on com- lop5 [30].
parable specimens. The impact of microstructure and Vaporization data for several candidate breeder
impurities on properties, on material stability, and on materials have been reported [22,32,33]. Partial pres-
performance of ceramics has long been recognized [26]. sures of the different species over the ceramics have
For example, porosity, pore morphology, and grain size been expressed as a function of temperature and differ-
have profound effects on thermal conductivity, mecha- ent structural metals showing that higher lithium vapor
nical properties, stability, and tritium inventory. Impur- pressures exist for materials with higher lithium content.
ities can affect thermodynamic and chemical properties Measurements made using a carrier gas stream of vary-
as well as transport behavior. The thermodynamic, ther- ing H,O partial pressure showed that moisture en-
mophysical, and mechanical properties for the as-fabri- hances the volatility of Li,O as LiOH. Results are
cated ceramic materials, as well as their irradiation derived from [32] in terms of the equilibrium constants
behavior, largely determine design guidelines. for the reaction Li,O(s) + H,O(g) = 2 LiOH(g).
In the FUBR 1A irradiation experiment carried out
3.1. Thermodynamic properties in the US fast reactor EBR II, lithium transport from
the ceramic pellets to the capsule walls was investigated
Measurements have been made of the solubility of [34]. Four materials, namely, Li,O, y-LiAlO,, Li ,SiO,,
hydroxide in Li,O and of the vaporization of hydroxide and Li,ZrO,, were irradiated at 500-900 o C. Tempera-
from Li,O into a helium gas stream containing varying ture differences in the helium-filled pellet cladding gap
amounts of H,O vapor [27-291. The solubility data were - 300Cat 900C. - 155 o C at 700 o C. and less
were used to enable corrected thermodynamic calcu- than 40C at 500 o C. The fractional lithium transfer
lations (corrected for nonideality of solutions of hydrox- was found to be temperature dependent. It was signifi-
ide in Li,O) to be made on the Li,O system [30]. A cant for Li,O at all temperatures, reaching 1 at.% at
general conclusion from the calculations is that the 900 o C, the largest observed for any material. No lithium
behavior of breeders will be strongly dependent on the transport was detected from the LiAlO, and Li,ZrO,
oxygen activity prevailing in the system. Oxygen activity pellets but a slight amount of transport was detected
controls the form of the released tritium, i.e., either the from the Li,SiO, pellets.
oxidized or reduced form, the amount of tritium re-
tained in the condensed phase, and the amount of
3.2. Thermophysical properties
blanket transported as LiOH/LiOT vapor. The sensitiv-
ity to oxygen activity in determining either oxidized or
reduced forms of released tritium is such that nominally The most complete study of thermal conductivity for
pure sweep gas containing only parts per million Li,O and Li,ZrO, as a function of porosity and tem-
levels of oxygen and/or H,O will be strongly on the perature is reported in [35]. A comprehensive compari-
oxidizing side. If oxygen activity is reduced sufficiently, son of this property for materials of 80% th.d. is re-
it is possible for hydride (tritide) to become an im- ported in [25] and presented in fig. 4. These results show
portant species in the condensed phase [31]. decreasing conductivity with increasing temperature;
Thermodynamic calculations have been carried out the order among solids is Li,O > LiAlO, > Li,SiO, >
in detail for the condensed and gaseous phases for the Li,ZrO,. Estimates have been made for thermal con-
breeder materials Li,O, LiAlO,, and Li,SiO,. From ductivity of sintered product and sphere-pat Li,O and
this work, a comparison of predicted thermochemical y-LiAlO, ) esp. in the TRIO experiment in the Oak
performance emerged. This comparison was based on Ridge reactor ORR [36-381. The TRIO estimates
the amounts of tritium retained in the condensed phase, showed little change from unirradiated values, while the
the amount of tritium in the gas phase, the amount of sintered product and sphere-pat estimates showed sig-
elemental lithium vaporized, and the amount of LiOH nificant change after irradiation. However, recent mea-
vapor transport. The calculations suggest that the per- surements on materials from FUBR-1A [39] showed
formance of LiAlO, is better than that of Li,O, which, essentially no change. Irradiation of LiAlO, to fluences
 C. E. Johnson et al. / Ceramic breeder materials 193

o LigZrOS 05% T.O.

A Lip0 65% T.D.

b I.6
<
1.6 + Li4Si046S%T.D.
d

A
0.6

0.4

0.2

0
11111111111
-0.2
0 200 400 600 600 1000
TEMPERATURE, *C

Fig. 5. Thermal expansion of solid breeder and structural

materials.

were measured [40]. For this submicron material lithium

loss appears at about 85OC and shrinkage sets in at
200 400 600 800 1000
9OOC. Under identical conditions the behavior of
Temperature, Y
Li,SiO, and Li,SiO, was poorer. No change of micro-
Fig. 4. Thermal conductivity of ceramic breeder materials (of
structure was observed for aluminates in the ALICE 1
80% th.d.).
experiment up to 600 C (OSIRIS Reactor, 102 n
cme2 fluence) but sintering was detected in ALICE 2 at
of 1020 n cmm2 at temperatures to 750C exhibited irradiation temperatures of about 750 o C for a 0.4 pm
little change in thermal conductivity [40]. grain size material [42]. In a low bumup irradiation
The differences between the thermal expansion of experiment, (OG5 experiment in Oak Ridge Reactor,
four solid breeders and two structural materials as a fluence < 5 x 102 n cmm2) 70% th.d. Li,O showed
function of temperature are shown in fig. 5 [41]. Li,O significant pore closure at temperatures above 850 C,
exhibited the largest expansion, while y-LiAlO, and whereas most pores remained open below 750C. No
Li,ZrO, showed the least expansion. pore closure was observed for 80% th.d. LiAlO, and for
Thermal fracture studies for LiAlO, of several differ- 90% th.d. Li,SiO, below 950 o C [43].
ent geometries and at different temperature gradients
showed better behavior for cylinders than that for disks. 3.3. Mechanical properties
In the latter case the critical temperature difference was
400-5OOC at about 600 C [40]. Dienst et al. [25] Elastic moduli reflect the atomic bond strength of
showed a critical temperature difference of 450-500 o C materials and provide information on lattice defects.
for LiAlO, and I 240 o C for Li,SiO,; Li ,SiO, was
found more resistant than LiAlO,.
Sintering of the ceramic material is strongly temper- Table 2
ature dependent, increasing with time and decreasing Thermal stability test after 8 h at 9ooC
particle size. Stability tests (8 h at 900 QC) are reported
Material Weight Dimensional Grain growth (pm)
in [41] (see table 2) for hot-pressed materials. The
loss (W) change ( W) before/after
greatest changes were observed for Li,O, somewhat
smaller changes for Li,Si04, and insignificant changes y-LiAlO, 0.04 0.00 <l/Cl
for Li,ZrO, and y-LiAlO,. Annealing experiments were Li,ZrO, 0.08 - 0.03 2/2
Li,SiO, 0.3 - 0.74 2/20
carried out on lithium aluminate for up to 40 d in air at
Li,O 1.34 - 2.4 6/80
400-1100 o C; dimensional, porosity, and weight changes
194 C. E. Johnson er a(. / Ceramic breeder wtatrnafs

160 500

LiA 10, Il.

ILi,SiO,
ALi,SiO,. KFK
I-
9 400

i
cl
0)
: 300
L
;I

9
.?I 200
II)
2
k
E 100
u
40

0
20 0 5 10 15 20 25 30 35
. .- Porosity, %
0 c Fig. 7. Compressive strength of ceramic breeder materials.
0 5 10 15 20 25 .30

Porosity. %
was measured for LiAIO: and found to he about 0.1
that of compressive strength.
Fig. 6.Young's modulus of ceramic breeder materials. Creep rate at 10 MPa has been measured as a
function of temperature and porosity for several materi-
als 125,401 (see fig. 8). The creep rate of Li,SiO, was
Youngs modulus was measured as a function of poros- found to be about two orders of magnitude higher than
ity at room temperature on lithium oxide [44], on lithium that of Li,SiO,. The steady-state thermal creep rate for
aluminates [40], and on lithium metasilicate [25]; Liz0 was investigated as a function of stress (15-45
Youngs modulus and shear modulus were measured on MPa) and temperature (700-800C) [45]. From creep
92% th.d. lithium oxide [45]. A comparison of Youngs
modulus of several breeders is displayed in fig. 6 [25]. In
the ALICE 1 irradiation experiment on LiAlO, (fluence
lo** n cm -2), no change in Youngs modulus was
detected at 400C, but a 10% reduction was observed
at 600C [40]. No change was observed on lithium
oxide specimens irradiated at 600 o C [44].
Compressive strength was studied as a function of
porosity and grain size and data are displayed in fig. 7.
Materials of low porosity and small grain size exhibit
the greatest strength. No changes in compressive
strength were noted for LiAlO, at 600 C for 1 at.%
burnup or up to 800 C for 2 at.96 burnup. Compressive
strength for LiilO, with 75% th.d. and 0.4 pm grain
size decreased from 275 MPa at 2OC, to 153 MPa at
500 o C, and to 100 MPa at 900 o C [40]. The compres-
sive strength and bending strength of LiAlO, differ by a
factor of 10. Bending strength of 80% th.d. aluminate,
metasilicate, and metazirconate (50-60 MPa) did not
show significant change up to 600 C, but that of Li 2O 4
begin to decrease at 400 o C 14.41.For brittle materials, 0 5 10 15 20 25 30 35
compressive strength is substantialIy larger than tensile
Porosity, %
strength, generally by a factor of eight. No direct mea-
surements of tensile strength for the candidate materials Fig. 8. Mean compressive creep rate of ceramic breeder
are available. However, diametral compressive strength materials in 100 h at 10 MPa.
 C.E. Johnson ei al. / Ceramic breeder materials 195

results, estimation was made of the dimensional and

microstructure stability of breeder materials. Consid-
ered as rigid materials at creep rates of 10m6/h and 10
MPa, temperature limits for 10 to 20% porosity material
are: = 950 C for LiAlO,, = 825 C for Li,SiO,, =
750C for Li,SiO,, and = 625 o C for Li,O. These
limits were confirmed by experimental measurements
for LiAlO, and Li ,SiO,.
An attempt [25] has been made to assess the resis-
tance of different breeder materials to thermal crack
A
initiation. This assessment used a reduced parameter LI ,SIO,

k/Ecx, instead of the usual thermal shock parameter

o,k (1 - v)/Ecx (q = tensile strength, k = thermal con-
0.1

0
t 1
ductivity, v = Poisson number, E = Youngs modulus,
I I I I I I I 1
a = thermal expansion). The highest resistance to ther- 0.20 0.30 0.40
0.00 0.10
mal crack initiation was found for LiAlO, and the CALCULATED THERMAL STRAINS (%I
lowest for Li,SiO, [25]. For Li,O, elastic and creep
Fig. 10. Thermal strains and pellet integrity for lithium ceramics
behavior indicate that it may be easily restrained by
under irradiation.
structural design [45].

3.4. Irradiation behavior rounded by stainless steel rather than an aluminum-

magnesium alloy.
Recent FUBR 1A irradiation tests tend to support The materials from FUBR-IA showed increase in
the selection of upper temperature limits in table 1. fragmentation in the order Li,ZrO,, LiAlO,, Li,O, and
Data indicate that tritium retention and pellet integrity Li,SiO, (see fig. 10). LiAlO, at 95% th.d. was less
are reasonably good [46-481. Analysis of helium and fragmented than this same material at 85% th.d. Frag-
tritium retention showed that Li,O exhibits the greatest mentation was not burnup dependent and was thought
retention for helium, especially at the intermediate tem- to result from thermal stresses arising from thermal
perature (700C), while tritium retention was lowest gradients and thermal expansion. In the ALICE experi-
for Li,ZrO,. Fig. 9 shows tritium retention data at ments, no significant fragmentation of LiAlO, was
700 o C as a function of burnup for the four candidate observed up to 600 o C and 1 at.% burnup; severe frag-
materials. All four materials show an approximately mentation was observed at about 750 o C with thermal
linear proportionality for tritium and helium retention gradients of about 80 K/mm [40]. In the EXOTIC
with burnup. In the ALICE experiments [49] tritium experiments carried out in the reactor HFR in Petten,
retention was strongly dependent on the nature of the The Netherlands, the integrity of specimens was gener-
structural material surrounding the sample during ally good up to 600 C.
irradiation. Retention was much lower when sur- In the FUBR 1A experiment, swelling at 700 and

/ LirSiOa I

4.0% - :;
0 :
E
L 3.0% - :::

s
5
Li ZrOo _
0.3% P
I A A
A
bf o- I I I I
I- 0.0 2 4 6 8 10 12
0.0 2 4 6 6 10 12

TRITIUM GENERATED (lOzo atOmlCm3) BURNUP (10 2o atom/cc)

Fig. 9. Tritium retention at 700C in Li,ZrO,, Li,O, Li,SiO,, Fig. 11. Volumetric swelling at 700C in Li,ZrO,, Li,O,
and LiAIO, Li,SiO,, and LiAlO,.
196 C.E. Johnson et al. / Ceramic breeder materials

900 C was observed to be very low for 85% th.d. tests carried out on Li,SiO, showed comparable weight
LiAlO, and Liz ZrO,, slightly higher for Li,SiO,, and gain per unit surface area.
rather high for Liz0 (see fig. 11). Swelling was thought
to arise from the retained helium in the material. A
4. Tritium transport and release
small shrinkage was observed (I 0.03%) for 75% th.d.
LiAlO, in the ALICE 1 experiment [40]. Significant
In an operating fusion reactor, the tritium breeding
changes were noticed in grain growth for Li,O. a slight
blanket will reach a steady state in which the tritium
growth for Li,SiO,, and no change for LiAlO, and
release rate equals the production rate. The tritium
Liz ZrO, 1481.
retention must be low to ensure that the tritium inven-
tory in the blanket does not become excessive. The
3.5. Material compatibility kinetics that govern the transport and release of tritium
from the breeder material must be fully understood if
we are to establish optimum operating temperature
Compatibility studies have been carried out at 873 K
limits for the blanket. Tritium generated within a
between couples comprised of the breeder materials
candidate material likely follows the mechanism (a)
(Li,O, LiilO,, and Li,SiO,) and the structural materi-
diffusion to the grain surface, (b) surface diffusion, (c)
als (Type 316 stainless steel, Incoloy 800, Inconel 625,
surface reaction, (d) desorption/evolution from the
HT-9, and Ti-6242) [50-533. Li,FeO, and Li,CrO,
grain surface, and (e) percolation through the pore
were common corrosion products. Similar studies [54,55]
structure. Thus, parameters that must be investigated
on the same ceramic materials and alloys at g73-973 K
include tritium diffusion within the solid and desorp-
identified LizNisO,, as a corrosion product with In-
tion/evolution from the grain surface.
cone1 625. Of these breeder materials, Li,O was most
reactive and LiAlO, was least reactive. The corrosion
chemistry of Li,O has been established in a basic 4. I. Tritium release experiments
thermodynamic study [56,57]. L&O was found to be
inert to the metals, provided it was rigorously freed of In-situ tritium recovery experiments provide a
moisture impurity. Corrosion observed previously was method of evaluating the feasibility of the lithium
attributed to LiOH conta~nation of Li,O. Porter et al. ceramics as tritium breeders. Many experiments
[SS] reported similar findings on long term studies of [36,61-771 have been performed to determine the tri-
doped Liz0 (with H,O) in contact with austemtic and tium release rate. For some materials the rates have
ferritic alloys. Materials doped to 6% H,O exhibited shown significant variance. Further, differences in in-
shallow attack (3-4 pm) after 700 h at 837 K. Experi- terpretatioia concerning the controlling mechanism for
ments on similar materials run for more than 2000 h tritium release appear throughout the literature gener-
exhibited identical behavior. The reaction rate between ally because of poor characterization of system varia-
oxide breeders and stainless steei cladding can be de- bles e.g., grain size. extraction temperature, material
rived from work of Hofmann and Dienst [59]. Temper- impurity content, carrier gas composition during extrac-
ature limits for a 20 pm thick reaction zone after 100 h tion. These and other tritium control parameters were
are 700 C for Li,O, 900 C for L,SiO,, and 1000 o C identified at a recent workshop [78] on tritium behavior.
for Li,SiO,, LiAlO,, and Li,ZrO,. Laboratory and in-situ tritium recovery experiments
Compatibility studies between beryllium and Li,SiO, have focused on the release process with the objective of
at 600 and 750C show that reaction effects seem characte~~ng the rate-limiting phenomena. In-situ tri-
acceptable at 600 o C but not at 750 o C 1591. Studies of tium recovery experiments have been conducted on
the compatibility between beryllium and LiAlO, and Li,O [61-671, on LiAlO, [36,65-721 and on
600 and 700 o C for 1500 h show that beryllium under- LizSiO,/Li,SiOd [67,70-741. A striking feature of these
goes no measurable oxidation under these conditions in-situ release experiments is the ease with which tri-
1601. tium is released from the solid. Characteristically, tri-
No reaction is expected between solid breeders and tium is observed in the purge gas stream within hours
pure helium, only between solid breeders and coolant after irradiation was initiated. Further, residual tritium
impurities, e.g., H,O. Li,O readily reacts with moisture content in the solid has been found to be very low,
to form LiOH. Sensitivity of LiAlO, to moisture has indicative of low tritium solubility, and therefore, a
been investigated by Rasneur et al. [40]. Changes in potentially low inventory. Studies continue in order to
weight and in phase for specimens exposed to a range of rigorously define release parameters.
humidity levels were studied as a function of time; the For the TRIO experiment [36], tritium recovery was
phases LiOH .2 Al(OH), f 2 H,O and Liz0 . H 2O were readily achieved to temperatures as low as 770 K. In
identified. In the presence of CO,, the latter phase is this experiment, it was found that the addition of Hz to
transformed into Li,CO,. No alteration of aluminate is the purge gas stream significantly enhanced tritium
detected for relative humidity less than 25%. Similar release. This implies that surface processes play a sig-
 C.E. Johnson et al. / Ceramic breeder materials 197

nificant role in the overall release process. Throughout the solid and desorption from the surface. The program
the course of the TRIO experiment, the predominant is based on the solution of the differential equation
species evolved from the test capsule was the noncon-
1ac la ,ac G
densible tritium form (HT). In sealed capsule tests,
D at g ar k-1 ar -
+D
where tritium is evolved by postirradiation annealing,
the HT fraction is generally < 10% of the total. Recent where D is the diffusivity, C is the tritium concentra-
in-situ tritium recovery tests on Li,O (VOM [72], tion, r is the grain radius, G is the tritium generation
CRITIC [75]) also showed a very high fraction (= 90%) rate. The equation for tritium release R, is given by
of the condensible tritium species (HTO, T,O). The
LILA-l experiment [68] in which LiAlO, was placed in

1+K&t
quartz and stainless steel capsules showed that material

n=,
in the quartz capsule released tritium predominantly in
the condensible form, while material in the steel capsule exp( - Dcxit)
released it in the noncondensible form. These data +*Cy,2+.h(ah-l)]
suggest the potential for immediate reduction of the
released material on structural members. where a = grain radius, K, = effective first-order de-
Tritium diffusion data on various lithium ceramic sorption rate constant, C, = tritium concentration
breeder candidates have been tabulated from both evaluated at r = a prior to a temperature change, h =
laboratory and in-reactor release-type experiments [84]. K,/D, t = time. In this equation, (Y, represents the
Much of the release data pertain to Li,O [79-831. roots of aa cot U(Y= 1 - ah. The release rate was
Below about 400 C the activation energy varies be- calculated using the diffusion-desorption model, and
tween about 96 and 143 kJ/mol for four separate the calculated results were compared to the observed
groups of data [79-821. However, only one data set [82] tritium release rate for a sample of Li,SiO, from the
exhibits the relatively large activation energy of about LISA-l experiment [70]. A typical plot is shown in fig.
143 kJ/mol. The rest of the data [79-811 show activa- 12. The model predicts the initial rise in release rate and
tion energies clustering around 96 kJ/mol. There are the release rate after 100000 s quite well but does a
several data sets for tritium diffusion at SOO-950 C poorer job of predicting the release rate between 10000
[83-901. The activation energy in these data varies con- and 75000. The calculated.tritium release profile is also
siderably from a low value of about 29 to 77 kJ/mol to
a high value of about 116 kJ/mol. The last two values
are reasonably close to the low temperature ( < 400 o C)
results [79-811. However, the low value of 29 kJ/mol is
consistent with the activation energies of deuterium and
hydrogen diffusion in Li,O [84,90].
There is considerable spread in the apparent activa-
tion energy for tritium diffusion in lithium ceramics.
The reason for this is not understood at present. Possi-
ble causes may include (a) different experimental condi-
tions (e.g., wet or dry), (b) different sample characteris-
tics (e.g., single crystal versus polycrystal, pellets, powder
forms), and (c) effects of other mechanisms besides bulk
diffusion contributing to the tritium release process.
There is also a wide spread in the values of the derived
diffusion coefficient D,. For example, for studies on
Li z0, Dr (at 600 o C) can range between a low value of
5 X 1OK cm*/s [87] to a high value of 3.5 x lo-*
cm/s [83]; but, as discussed before, this could be
attributed partly to uncertainty in the sample character-
istics.

4.2. Model development t 1 1 I I

0 200000 400000 600000 800000
TIME, set
A computer program is being developed at Argonne Fig. 12. Calculated and observed tritium release from meta-
National Laboratory, USA, to describe tritium behavior silicate sample Pl of LISA-1 experiment; D = 0.75 X10-
in ceramic breeder materials [94]. As presently con- exp( -35.0 kJ mol-/RT) cm/s, K, = 0.9970 exp( -30 kJ
structed, the program focuses on modeling diffusion in mol-/RT); - - - observed, - calculated.
198 C.E. Johnson et al. / Ceramic breeder materials

in good agreement with the observed data for time ture water shows a wide variation in behavior. However,
periods following a large temperature decrease (- high purity beryllium in deionized water at 590 K
250000, 450000 and 675000 s) or increase (= 350000 s). exhibited corrosion rates of < 1 mg/cm2.d. Beryllium
One shortcoming of the current model is that it uses a can also interact with steel to form a hard surface
characteristic grain radius of the sample to calculate the coating. In particular, beryllium has a high affinity for
release rate. The samples used in the LISA-1 experi- nickel, and the compound NiBe predominates on the
ments have a large distribution of grain sizes, and it is surface of high nickel steel. Other compounds which
difficult to assign a characteristic grain size to the can form include FeBe, and Be,C. The extent of the
sample. Since the tritium release is dependent on the interaction with stainless steel should be small, however,
grain size, this could have a substantial effect on the at blanket operating temperatures.
release profile. To correct this shortcoming the model is
being altered to allow for a distribution of grain sizes as 5.3. Betyllium resources
input.
The availability of beryllium is summarized in table
5. Neutron multipliers 3, where US Bureau of Mines and US Geological Survey
estimates are shown [97,98]. The Bureau of Mines data
Aside from its neutronic properties, the neutron mul- focused on the US reserve and resources. and the Geo-
tiplier should be resistant to change during irradiation logical Survey data focused on the world-wide reserve
and be compatible with the surrounding materials. Neu- and resources. The Bureau of Mines defines known
tronically, the elements Be, Bi, Pb, and Zr appear resources of beryllium to consist primarily of bertran-
acceptable. However, beryllium is favored as the refer- dite (4BeO. 2Si0,. H,O) located at Spor Mountain and
ence neutron multiplier because it combines excellent Gold Hill, Utah, and the Seward Peninsula, Alaska, and
multiplication with low density, high specific heat, and of beryl (3BeO. Al,O,. 6Si0,) located in the coarse-
high thermal conductivity. The main concerns for beryl- zoned pegmatites of New England, South Dakota, and
lium are irradiation swelling caused by helium genera- Colorado.
tion, corrosion, and its limited resources. (The STAR- To address the resource issue, it is necessary to
FIRE design required 5 X lo4 kg of beryllium.) consider beryllium recycle [99]. Then the losses can be
limited to the sum or recycle process losses and beryl-
S. 1. Irradiation behavior lium burn-up losses. It is possible that, with sufficient
incentive, the process loss rate could be reduced to
Helium gas is generated in beryllium from the (n, 2n) about 1%. The burnup losses are likely to be small, but
and (n, cx) reactions. Small grain size and high tempera- will compete with this figure. For extended service,
ture will generally enhance helium release, thereby re- efficient recycle is required and the recycle process
ducing the driving force for swelling. Postirradiation losses must decrease as the multiplier thickness in-
annealing studies [92-951 of beryllium indicate that (1) creases.
there is a threshold temperature below which swelling is
insignificant, and (2) there is a limit to volume increase
( = 30%) beyond which no further swelling occurs. The Table 3
swelling threshold temperature depends on the helium World beryllium resources (in lo3 tons of contained beryllium)
gas content, decreasing with increasing neutron ex-
Country Reserves Other Total
posure. The maximum volume increase of 30% appears
to be a geometrical limit at which the majority of the North America
internal bubbles touch and release their gases. Such United States 25 48 73
Canada _ 23 23
volume increases were only observed when the anneal-
Mexico _ 2 2
ing temperatures are above 1300 K.
South America
Argentina 25 41 12
5.1. Corrosion behavior
Brazil 140 256 396
Asia: India 64 119 1x3
The corrosion behavior of beryllium in oxygen,
Africa
nitrogen, and water environments has been examined
Mozambique 6 11 17
[96]. The reaction of beryllium with both oxygen and 11 20 31
Rwanda
nitrogen follows the parabolic rate. Significant oxida- Rep. South Africa 15 29 44
tion appears between 870 and 970 K and increases Uganda 15 26 41
gradually up to 1100 K, after which the temperature of Zaire 7 14 21
oxidation increases rapidly until actual burning com- Europe: USSR 61 111 172
mences at 1470 K. The reaction of beryllium with Oceania: Australia 11 19 30
nitrogen follows the same pattern, but the reaction rates World total 380 125 1105
are lower. The corrosion of beryllium in high tempera-
 C. E. Johnson et al. / Ceramic breeder materials 199

Based upon these observations, it appears reasonable Li,O, LiAlO,, Li,Si04. and Li,ZrO,; however, one
to consider beryllium multipliers for the first and sec- must recognize the potential for reordering as the prop-
ond generations of fusion reactor service. Since the erties data base develops. Critical to a rigorous assess-
impurities in beryllium will activate in the fusion en- ment are the tritium-related properties, e.g., solubility,
vironment, a remote refabrication technology will be release, diffusivity, and bumup. Advanced methods of
required. Without very efficient recycle, thick beryllium preparations of the four candidates may provide materi-
multipliers should be considered for use only in the first als of superior performance and may affect their rank-
generation of fusion reactors. ing. Further studies on the above materials appear
needed.
6. Summ~
We wish to acknowledge the extensive and very
In this assessment. one finds that the fusion com- beneficial help and assistance of Dr. N. Roux (CEA) in
munity has given considerable attention to Li,O, less the preparation of the manuscript for this paper.
attention to LiAlO, and Li,SiOa, and little attention to
LizZrO,. The result is that the data base for these References
materials is very uneven, and rigorous comparisons are
hard to make. Currently, effort is being directed to 111E. Roth, J. Abassin, F. Botter, M. Briec, M. Masson. G.
evening out the data base, especially with respect to Micaud. B. Rasneur, N. Roux, J. Sannier and M. Trota-
Li,ZrO,. This material has shown excellent irradiation bas, Transactions ENC86, Vol. 3 (1986) p. 145.
performance and is now receiving attention in labora- PI R.M. Arons, R.B. Poeppei, M. Tetenbaum and C.E. John-
son, J. Nuci. Mater. 103 & 104 (1981) 555.
tory studies and purge flow irradiation experiments.
Among the candidates, lithium oxide is highly re- 131 B. Rasneur, Proc. 13th Symp. Fusion Techn. (13th SOFT)
Varese, Italy, Sept. 1984, Vol. 2, p. 1017.
garded because of its high lithium atom content, low
141 D. Voiiath. H. Wedemeyer and E. Gunther, Proc. 13th
tritium solubility. high thermal conductivity, reasonably Symp. on Fusion Technology (13th SOFT) Varese, Italy,
rapid tritium release, and high tritium recovery feasibil- Sept. 1984, Vol. 2, 1967.
ity. It is probably the only material that could satisfy [51 C. Aivani, S. Casadio, L. Lorenzini. and G. Brambeiia,
the tritium breeding requirements without a neutron Fusion Technoi. 10 (1986) 106.
multiplier. Its principal disadvantage results from its 161 L. Yang, R. Medico and W. Baugh, US Patent 4405595
high hydroxide/trito~de and atomic lithium vapor (Sept. 20, 1983).
pressure, which could result in transport of blanket [71 J.F. Palmer and C.W. Turner, Proc. 89th Ann. Meet. Am.
Ceramic Sot., Pittsburgh. PA, USA, April 1987. Adv. in
material to cooler zones and closure of porosity. Lithium
Ceram. 1988, in print.
aluminate might be considered superior to Liz0 be-
PI M. Korzumi, K. Ohtsuka, H. Isagawa, H. Akiyama and A.
cause of its better swelling resistance, higher melting
Todokoro, Nucl. Technoi. 61 (1983) 55.
point, broader operating temperature window, slow re-
[91 B.J.F. Palmer, L.E. Bahen and A. Ceiii. Ceram. Bull. 63
action with water, and its excellent irradiation behavior, (1984) 1030.
especially for high density material (95% th.d.). How- [lOi D. Voiiath and H. Wedemeyer, J. Nucl. Mater. 141-143
ever, its tritium diffusivity characteristics are poorer; (1986) 334.
therefore, this material requires a higher minimum oper- Ull A. Skokan. H. Wedemeyer, D. Voiiath and E. Gunther,
ating temperature to maintain low tritium inventory. Proc. 14th Symp. Fusion Technology (14th SOFT) Avig-
Li,SiO, behaves much like Li,O in many respects. It non, France, Sept. 1986, Vol. 2, p. 1255.
exhibits good tritium diffusivity and chemical stability. 1121 A.J. Fiipot, P. Diets and R. Lecocq, J. Nucl. Mater. 133 &
134 (1985) 226.
However, recent closed capsule irradiation test showed
(131 A.J. Fiipot and P. Dieis. J. Nuci. Mater. 141-143 (1986)
this material to fragment severely, probably due to
339.
thermal expansion. Li,ZrO, has good stability and low F. Rigby and P. Kennedy, Proc. 13th Symp. Fusion
u41
tritium retention, and it has exhibited excellent irradia- Technology (13th SOFT) Varese, Italy, Sept. 1984, Vol. 2,
tion characteristics. Its principal disadvantage is activa- p. 953.
tion of zirconium under irradiation, which could cause 1151 D. Voiiath, H. Wedemeyer and E. Gunther, J. Nuci.
maintenance problems. Fuller evaluation is hampered Mater. 133 & 134 (1985) 221.
by a lack of data. [I61 D. Voiiath. H. Wedemeyer, Proc. 89th Ann. Meet. Amer.
The above brief assessment has focused on material Ceramic Sot., Pittsburgh, PA, USA, April 1987, Advances
properties and material behavior. Another factor that in Ceramics 1988, in print.
1171 D.J. Suiter, J.W. Davis and B.A. Kirkpatrick, J. Nuci.
could strongly influence the selection process is that of
Mater. 103 & 104 (1981) 579.
blanket design. It is conceivable that each ceramic pos-
WI P. Kennedy. K.E. Gilchrist, D.E. Walker and S. Brough-
sesses characteristics that could make it the preferred
ton. Proc. 14th Symp. Fusion Technology (14th SOFT)
candidate for a given blanket design. Maturing of the Avignon, France, Sept. 1986, Vol. 2, p. 1013.
blanket design activity should clarify this possibility. [I91 H. Eibei, in: these Proc. (ICFRM-3). J. Nucl. Mater.
Currently, the materials may be ordered as follows: 155-157 (1988).
200 C. E. Johnson et al. / Ceramic hreeder materrals

[20] B.K. Bein, S. Krueger and J. Pelzl, J. Nucl. Mater. 141-143 [50] P.A. Finn, S.R. Breon and N.R. Chellew, J. Nucl. Mater,
(1986) 119. 103 & 104 (1981) 561.
[21] G. Schumacher, M. Dalle Donne. S. Dorner, in: these [51] O.K. Chopra and D.L. Smith. J. Nucl. Mater. 103 & 104
proc. (ICFRM-3) J. Nucl. Mater. 155-157 (1988). (1981) 555.
[22] D. Suiter, Report MDC E 2677 (June 1983). [52] K. Chopra, T. Kurasawa and D.L. Smith, Proc. Topical
[23] C. Denuziere and N. Roux, CEA/NET Contract No. Conf. Ferritic Alloys for Use in Nuclear Energy Technol-
5/FU-F/NET (November 1984). ogies, Snowbird, Utah, USA. June 1983, p. 209.
[24] K.R. Kummerer and W. Dienst. J. Nucl. Mater. 133 & [53] P. Hofmann and W. Dienst. J. Nucl. Mater. 141-143
134 (1985) 176. (1986) 289.
[25] W. Dienst and H. Zimmermann, in: these Proc. (ICFRM- [54] H. Takeshita. T. Kurasawa, S. Muraoka, S. Nasu. M.
3). J. Nucl. Mater. 155-157 (1988). Miyake and T. Sano. J. Nucl. Mater. 80 (1979) 249.
[26] W.D. Kingery, K. Bowen and R. Uhlmann, Introduction [55] T. Kurasawa, H. Takeshita and S. Nasu. J. Nucl. Mater.
to Ceramics, 2nd Ed. (Wiley, New York, 1976). 92 (1980) 67.
[27] J.H. Norman and G.R. Hightower, J. Nucl. Mater. 122 & [56] R.J. Pulham, W.R. Watson and J.S. Collinson. J. Nucl.
123 (1984) 913. Mater. 122 & 123 (1984) 1243.
[28] M. Tetenbaum and C.E. Johnson. J. Nucl. Mater. 126 [57] R.J. Pulham, W.R. Watson and N.P. Young, Proc. 14th
(1984) 25. Symp. on Fusion Technology (14th SOFT) Avignon.
[29] M. Tetenbaum, A.K. Fischer and C.E. Johnson. Fusion France, Sept. 1986, p. 963.
Techn. 7 (1985) 53. [58] D.L. Porter. J.R. Krsul, M.T. Lang. L.C. Walters and M.
[30] C.E. Johnson and A.K. Fischer, J. Nucl. Mater. 130 Tetenbaum. J. Nucl. Mater. 122 & 123 (1984) 929.
(1985) 445. [59] P. Hofmann and W. Dienst, in: these Proc. (ICFRM-3) J.
[31] A.K. Fischer and C.E. Johnson, Proc. 89th Ann. Meet. Nucl. Mater. 1555157 (1988).
Am. Ceramic Sot., Pittsburgh, PA, USA, April 1987. [60] T. Flament, P. Fauvet and J. Sannier, in: these Proc., J.
Advances in Ceramics 1988, in print. Nucl. Mater. 155-157 (1988).
[32] M. Tetenbaum and C.E. Johnson, J. Nucl. Mater. 120 [61] T. Kurasawa, H. Takeshita, W. Watanabe, H. Yoshida. Y.
(1984) 213. Naruse. T. Miyauchi, K. Mimura and H. Umei. J. Nucl.
[33] H.R. Ihle, R.D. Penzhorn and P. Schuster in: these Proc. Mater. 122 & 123 (1984) 902.
(ICFKM-3) J. Nucl. Mater. 155-157 (1988). [62] T. Kurasawa, H. Yoshida. H. Takeshita, W. Watanabe. T.
[34] G.W. Hollenberg, D.L. Baldwin, R. Wisner and B. Carl- Miyauchi. T. Matui, H. Umei and Y. Nat-use. J. Nucl.
son, Proc. Int. Symp. on Fusion Reactor Blanket and Fuel Mater. 133 & 134 (1985) 196.
Cycle Technology, Tokai-Mura, Japan. Oct. 1986, p. 90. [63] R.A. Verrall. J.M. Miller, and S.R. Bokwa. Proc. 89th
[35] T. Takahashi and T. Kikuchi. J. Nucl. Mater. 91 (1980) Annual Meeting American Ceramic Society, Pittsburgh,
93. PA, USA, April 1987, Adv. Ceram. (1988). in print.
[36] R.G. Clemmer, P.A. Finn, B. Misra, M.C. Billone, A.K. [64] S. Tanaka. T. Terai, H. Mohri and Y. Takahashi, in: these
Fischer, SW. Tam, C.E. Johnson and A.E. Scandora, J. Proc. (ICFRM-3) J. Nucl. Mater. 155-157 (1988).
Nucl. Mater. 133 & 134 (1985) 171. [65] H. Kwast, R. Conrad, A.J. Flipot. P. Kennedy and J.D.
[37] SW. Tam, L. Leibowitz, Y.Y. Liu, R. Blomquist and C.E. Elen. in: these Proc. (ICFRM-3) J. Nucl. Mater. 155-157
Johnson, J. Nucl. Mater. 133 & 134 (1985) 234. (1988).
[38] Y.Y. Liu and S.W. Tam, Fusion Techn. 7 (1985) 30. [66] T. Kurasawa. H. Watanabe, G.W. Hollenberg, Y. Ishii, A.
[39] J.L. Ethridge and D.E. Baker. Proc. 89th Ann. Meet. Am. Nishimura, H. Yoshida. Y. Naruse. M. Aizawa, H. Ohno
Ceramic Sot., Pittsburgh, PA, USA. April 1987. Advances and S. Konishi, J. Nucl. Mater. 141-143 (1986) 265.
in Ceramics 1988, in print. [67] H. Kwast, R. Conrad, P. Kennedy, A.J. Flipot and J.D.
[40] B. Rasneur, F. Botter and M. Briec, CEA Note 2536 Elen. J. Nucl. Mater. 141-143 (1986) 300.
(1987). [68] E. Roth, J.J. Abassin, F. Botter, M. Briec, P. Chenebault.
[41] D.L. Smith et al., Blanket Comparison and Selection M. Masson, B. Rasneur and N. Roux, J. Nucl. Mater. 133
Study, ANL/FPP84-1, Vol. 2 (1984), Chapter 6.3. & 134 (1985) 238.
[42] F. Botter, B. Rasneur and E. Roth, in: these Proc. [69] M. Briec, F. Botter, J.J. Abassin, R. Benoit. P. Chenebault.
CICFRM-3). J. Nucl. Mater. 1555157 (1988). M. Masson. B. Rasneur. P. Sciers. H. Werle and E. Roth,
[43] L. Yang, R. Medico, W. Baugh and K. Schultz. J. Nucl. J. Nucl. Mater. 141-143 (1986) 357.
Mater. 103 & 104 (1981) 585. [70] H. Werle. J.J. Abassin, M. Briec, R.G. Clemmer, H. Elbel.
[44] P. Kennedy, Proc. 14th Symp. on Fusion Technology H.E. Hafner, M. Masson. P. Sciers and H. Wedemeyer. J.
(14th SOFT) Avignon, France, Sept. 1986, p. 1013. Nucl. Mater. 141-143 (1986) 321.
[45] M.C. Billone, Y.Y. Liu, R.B. Poeppel, J.L. Routbort, K.C. [71] M. Briec, J.J. Abassin. M. Masson, E. Roth. P. Sciers and
Gorreta and D.S. Kupperman, J. Nucl. Mater. 141-143 H. Werle. in: these Proc. (ICFRM-3). J. Nucl. Mater.
(1986) 282. 155-157 (1988).
[46] G.W. Hollenberg, HEDL 7643 (1987). [72] T. Kurasawa, H. Watanabe, E. Roth and D. Vollath. m:
[47] G.W. Hollenberg, Proc. 89th Ann. Meet. Am. Ceramic these Proc. (ICFRM-3) J. Nucl. Mater. 155-157 (1988).
Sot., Pittsburgh, PA, USA, April 1987, Advances in [73] R.G. Clemmer. H. Werle and M. Briec, Proc. Int. Symp.
Ceramics 1988, in print. on Fusion Reactor Blanket and Fuel Cycle Technology,
[48] G.W. Hollenberg and D.L. Baldwin, J. Nucl. Mater. 133 Tokai-Mura, Japan, Oct. 1986, p. 51.
& 134 (1985) 242. [74] H. Werle. M. Briec, R.G. Clemmer, H. Elbel, H.E. Hafner.
(491 J.P. Bonal, Memoire du Conservatoire National des Arts M. Masson. G. Schumacher and H. Wedemeyer, in: these
et Metiers (1987). Proc. (ICFRM-3) J. Nucl. Mater. 155-157 (1988).
 C.E. Johnson et a(. / Ceramic breeder materials 201

[75] R.A. Verrall, J.M. Miller, S.R. Bokwa and C.E. Johnson, Fusion Technology, Garmisch-Partenkirchen, FRG, June
J. Nucl. Mater. 155-157 (1988). 1976, p. 337.
[76] T. Kurasawa, G.W. Hollenberg, and H. Watanabe, Proc. [87] D. Guggi, H.R. Ihle, D. Bruning, U. Kurz, S. Nasu, K.
Int. Symp. on Fusion Reactor Blanket and Fuel Cycle Noda and T. Tanifuji, J. Nucl. Mater. 18 (1983) 100.
Technology, Tokai-Mura, Japan, Oct. 1986, p. 43. [88] T. Tanifuji, K. Noda, T. Takahashi and H. Watanabe, J.
[77] H. Watanabe, T. Kurasawa, E. Roth and D. Vollath, ibid. Nucl. Mater. 149 (1987) 227.
ref. [76], p. 33. [89] R. Wiswall and E. Wirsing, BNL-50748 (1977).
[78] 1. Hastings, Ed., Modeling Workshop - Chalk River, [90] H. Katsuta, S. Konishi and H. Yoshida, Proc. Fall Meet-
Canada, April 1987. ing of Atomic Energy Sot. of Japan, F-16, Kyushu, Japan,
[79] H. Kudo and K. Okuno, J. Nucl. Mater. 101 (1981) 38. Oct. 1981.
[80] K. Okuno and H. Kudo, Proc. Fall Meeting of Atomic [91] J.P. Kopasz, S.W. Tam and C.E. Johnson, in: these Proc.
Energy Society of Japan, F-17, Kyushu, Japan, Oct. 1981. (ICFRM-3). J. Nucl. Mater. 155-157 (1988).
[81] S. Nasu, H. Kudo, K. Shiozawa, T. Takahashi, T. [92] J.B. Rich, G.B. Redding and R.S. Barnes. J. Nucl. Mater.
Kurasawa, M. Tachiki and K. Tanaka, J. Nucl. Mater. 68 1 (1959) 96.
(1977) 261. [93] C.E. Ells and E.C.W. Perryman, ibid. ref. [92], p. 73.
[82] V.G. Vasilev, S.R. Borisov, N.N. Riazantseva and A.A. (941 J.B. Rich and G.P. Walters, Proc. Int. Conf. on the
Vashman, US/USSR Workshop on Engineering and Eco- Metallurgy of Beryllium, London, UK, Oct. 1961, p. 362.
nomic Problems of ETF, Moscow and Leningrad, USSR, [95] J.B. Rich, G.P. Walters and R.S. Barnes, J. Nucl. Mater. 4
September 1979. (1961) 287.
[83] K. Okula and D.K. Sze, Proc. ANS Meeting on Tritium [96] W.W. Beaver and D.W. Lillie. in: Reactor Handbook,
Technology in Fission, Fusion and Isotopic Applications, Second Ed., Vol 1 - Materials, Ed. CR. Tipton (Intersci-
April 1980, Dayton, OH, USA, p. 286. ence Publishers, New York, 1960) p. 897.
[84] J. Shearer, SW. Tam and C.E. Johnson, Trans. Am. Nucl. [97] D. Petkof, Beryllium, Mineral Facts and Problems, US
Sot. 44 (1983) 139. Bureau of Mines Bulletin 650 (1980).
[85] H. Ohno, S. Konishi, T. Nagasaki, T. Kurasawa, H. 1981 W.R. Griffitts, Beryllium, US Geological Survey Prof.
Katsuta and H. Watanabe, J. Nucl. Mater. 141-143 (1986) Paper 820 (1973).
181. [99] D.H. Berwald and S. Zenczak, Fusion Technol. 8 (1985)
[86] D. Guggi, H.R. Ihle and U. Kurz, Proc. 9th Symp. on 1143.