Journal of Nuclear Materials 318 (2003) 325–332

www.elsevier.com/locate/jnucmat

Corrosion behaviour of steels and refractory metals

and tensile features of steels exposed to flowing
PbBi in the LECOR loop
a,*
C. Fazio , I. Ricapito b, G. Scaddozzo c, G. Benamati a

a
ENEA – CR Brasimone, I – 40032 Camugnano Bologna, Italy
b
ESA s.a.s, I – 40033 Casalecchio di Reno, Bologna, Italy
c
Politecnico di Torino, Corso Duca degli Abruzzi 24, I – 10129 Torino, Italy

Abstract

An experimental activity has been started using the LECOR loop at the ENEA Brasimone centre to investigate the
corrosion behaviour of steels and refractory metals as well as the tensile properties of steels exposed to flowing liquid
lead bismuth with low oxygen activity. The oxygen content in the liquid metal was controlled and monitored by a
dedicated system. The compatibility test was performed at 673 K and the corrosion and tensile results herein reported
concern the first 1500-h run of the loop operation. All the materials tested suffered from liquid metal attack exhibiting a
weight loss. The consequent evaluation of the corrosion rate showed that, under the given test conditions, the refractory
metals are more resistant than the steels. The tensile properties of austenitic steel are not affected by the liquid metal
corrosion, while the martensitic steel exhibited a mixed brittle–ductile fracture surface.
Ó 2003 Elsevier Science B.V. All rights reserved.

1. Introduction the MYRRHA project, which consists in the develop-

ment of a Multipurpose Neutron Source for R & D
Liquid lead or eutectic lead–bismuth have been pro- Applications on the basis of an ADS, in Belgium [4]
posed both as spallation target and as coolant in ac- have the aim of demonstrating the operability of such
celerator driven systems (ADS), which aim to burn liquid metal targets. Both of these initiatives have
radioactive waste from conventional nuclear power planned to use liquid eutectic lead–bismuth, which sat-
plants. As has been well described in the past [1,2] the isfies the requirements for neutron production and also
ADS consists of proton beam impinging a liquid metal because it possesses suitable physical properties, i.e. low
target and a subcritical core. Neutrons are produced by melting point and low vapour pressure. However, it is
the proton beam striking the liquid metal target, thereby well known that materials compatibility issues should be
inducing a spallation reaction. A design of a Demon- considered when using liquid metal. Indeed, it has al-
strator (DEMO) reactor has been produced by the ready been comprehensively reported that the mecha-
Italian company named ANSALDO [2] and a large R & nism and rate of the corrosion of structural materials
D programme is ongoing within the European Union to when exposed to liquid lead or eutectic lead–bismuth, at
demonstrate the feasibility of ADS. Within the R & D the operation temperature of the target, are strongly
programme the international project MEGAwatt Pilot dependent upon the oxygen activity in the liquid metal
target Experiment (MEGAPIE) in Switzerland [3] and [5,6]. For instance, it has been demonstrated that for a
sufficiently high oxygen activity in heavy liquid metals,
oxide layers are developed both on steels [7] and on re-
*
Corresponding author. Tel.: +39-0534 801463; fax: +39- fractory metals [8,9]. Although during the 1960s, some
0534 801225. corrosion tests were performed on low-alloy steel ex-
E-mail address: concetta.fazio@psf.fzk.de (C. Fazio). posed to deoxidised heavy liquid metal [10], a complete

0022-3115/03/$ - see front matter Ó 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-3115(03)00009-6
326 C. Fazio et al. / Journal of Nuclear Materials 318 (2003) 325–332

perspective on the corrosion behaviour under the pro- submerged impeller, sized to provide the loop with a
posed operation conditions cannot be drawn from the maximum liquid metal flow rate of 4 m3 /h correspond-
existing literature. The aim of the experimental activities ing to a velocity in the test sections of 3 m/s.
here presented is to describe the corrosion mechanism The three test sections (TS1/2/3) containing the
and to estimate the corrosion rate of selected materials specimens to be investigated are placed downstream of
exposed to flowing Pb–Bi with a low oxygen activity. the electrical heater (H). The samples are piled up in the
The experiments were performed in the LECOR (LEad test section and separated from each other with spacers.
CORrosion) loop designed and constructed at the In the steady state, the thermal power provided by the
ENEA site of Brasimone. The loop testing parameters heater is equal to that extracted by the forced air heat
were chosen to be as similar as possible to the experi- exchanger (C). The electrical heater (H) can supply a
mental conditions of the MEGAPIE target and also, to maximum power of 82 kW, and is operated by a PID
some extent, to the MYRRHA target. A detailed de- thermal regulator connected to a socket thermocouple
scription of the loop, as well as the oxygen control sys- located at the exit, with H and C responsible for the DT
tem employed will be presented. The materials analysed between the hot and cold branches. This configuration
were austenitic steel, martensitic steel and refractory enables a continuous transport of corrosion products
metals. In addition to the identification of the corrosion from the hot region, where the test sections are located,
process, tensile properties of the steels after exposure to to the cold region where they are partially released,
the liquid metal were also tested. The experimental simulating the actual behaviour of coolant fluid in a
programme of LECOR is planned to extract samples thermal production plant. The total liquid metal flow
from the test section every 1500 h, up to a maximum of rate available to the test sections is adjusted through the
4500 h. The results here presented refer to analyses by-pass line, whilst the liquid metal flow rate in each test
performed on the materials corroded for 1500 h. section is automatically controlled by electro-pneumatic
regulation valves driven by feedback signals from elec-
tromagnetic flowmeters (EF).
2. Experimental To carry out the experimental campaign with well-
defined oxygen content in the liquid metal, an oxygen
2.1. LECOR plant electrochemical sensor (Ox) is placed in the hot region,
close to one of the test sections. Moreover, the vessel S1,
According to Fig. 1, the LECOR plant has a Ôfigure- which acts as drain and storage tank, is connected to an
of-eightÕ configuration, typical for a corrosion loop, with external gas system that allows the oxygen content in the
a high temperature branch running from the economiser liquid metal to be reduced by injection of Ar/H2 mixture.
(EX) to the test sections (TS) and a low temperature The structural materials of the loop are as follows:
branch including the delivery and return pipes between the cold part is fabricated in austenitic steel (AISI 316L),
the S2 tank and the economiser, including the by-pass while the hot region is in ferritic steel (Cr–Mo Steel)
line. The S2 vessel contains a mechanical pump, with representing a good compromise between the require-

EF
gas out

P
HG
Ox
EX

TS1 TS2 TS3

S2

EF
EF EF EF
C H
Ar + H2 MFC

TS test section
gas out S2 recirculation vessel
MFC S1 drain / storage vessel
EX economiser
HG
P mechanical pump
C air cooler
S1 EF electro-magnetic flow-meter
H heater
Ox oxygen sensor
MFC mass flow controller
HG hygrometer

Fig. 1. Scheme of the LECOR loop.

 C. Fazio et al. / Journal of Nuclear Materials 318 (2003) 325–332 327

Table 1 water extraction rate from the drain tank. During cir-
Main characteristics of the LECOR loop culation of the liquid metal, the oxygen activity was
Pb–Bi inventory (m3 ) 0.6 measured with an oxygen sensor placed in the hot part
Maximum Pb–Bi flow rate in the 4 of the loop. This oxygen sensor, constructed at the IPPE
pump delivery line (m3 /h) of Obninsk, has the red-ox pair Bi/Bi2 O3 as reference
Maximum temperature in the test 550 electrode and Ytrria stabilised zirconia acting as a solid
section (°C) electrolyte. The probe had been previously calibrated in
Total installed electrical power 150 the Russian laboratories and was delivered to ENEA
(kW)
with the calibration equation [22]. During the 1500-h
Maximum heating power of the 82
heater H (kW)
run of the LECOR loop, the electromotive force read-
Maximum heat exchange power 240 ings of the probe were recorded by a data acquisition
in EX (kW) system, and those data and the corresponding oxygen
Design pressure (barg) 4 activity are reported in Fig. 2, where oxygen activity was
calculated using the calibration equation. As can be seen
in that diagram, the measured oxygen activity was be-
ments of acceptable ductility and corrosion resistance in tween 2:3  10 6 and 5:5  10 4 , corresponding to
a liquid metal environment. 3:1  10 10 and 7:3  10 8 wt% of oxygen dissolved in
Loop operation is automatically controlled by a the liquid metal.
PLC, connected to a supervisor PC that also runs the
data acquisition software. Characteristics of the main 2.3. Materials, test parameters and post test analysis
LECOR components are given in Table 1.
Cylindrical machined corrosion specimens 10 mm in
diameter and 50 mm in length, as well as cylindrical
2.2. Oxygen control and monitoring system tensile specimens 3 mm in diameter and gauge length
10 mm were machined for the testing programme. The
To perform the corrosion tests in the LECOR loop tested materials were AISI 316L and T91 steels and the
using liquid metal with a low oxygen activity, 80 wppm refractory metals tungsten and molybdenum. The two
of pure Mg were added to the eutectic lead–bismuth. In steels were tested for both corrosion and tensile perfor-
addition, pre-treatment with hydrogen gas was per- mance, whilst only the corrosion behaviour of the two
formed in the storage tank for a few days at 603 K. The refractory metals was tested. The W and Mo, supplied
aim of adding Mg was to deoxidise the liquid metal, by Elettroleghe s.r.l of Italy, were fabricated by sintering
since Mg forms much more stable oxides than Pb and Bi and had purities of 99.9 and 99.95 wt% respectively,
[11]. The hydrogen pre-treatment was performed taking while the composition of the two steels is reported in
advantage of the oxide reduction kinetic studies previ- Ref. [7]. The lead–bismuth used in the loop, was sup-
ously performed [12], and monitored by evaluating the plied by Stachow of Germany, and was composed 44.8

-4
5.7x10
-4
5.0x10
4.3x10 -4
-4
3.6x10
-4
2.9x10
-4
2.1x10
1.4x10 -4
7.1x10 -5
0.0
0.5

0.4

0.3

0.2
10 20 30 40 50 60 70

Fig. 2. Electromotive force and oxygen activity in the liquid metal recorded during the 1500 h run.
328 C. Fazio et al. / Journal of Nuclear Materials 318 (2003) 325–332

by wt% of Pb and 55.2 wt% of Bi with some minor

impurities. The temperature in the test sections of the
loop was 673 K and the cold part of the loop was at 573
K. The flow of liquid metal around the samples was 120
l/h corresponding to a velocity of 1 m/s and the Rey-
nolds number was about 1  104 , indicating a turbulent
flow. Every 1500 h of liquid metal circulation in the
loop, a set of samples was extracted from the test sec-
tions, where two corrosion samples for each type of
material and four tensile samples for each type of steel
were taken. In order to extract the samples from the test
sections the loop is drained and cooled down, the flan-
ged test sections are opened and the specimens were
extracted. After extracting the samples from the test
section, the corrosion rates of the four materials were Fig. 3. SEM micrograph of the AISI 316L cross section etched
with oxalic acid.
evaluated by means of the weight change measure-
ment technique. For each material, one of the corro-
sion samples was cleaned with a solution made by
CH3 COOH, H2 O2 , C2 H5 OH in a ratio of 1:1:1, with the As it can be observed in Fig. 3 the surface of the steel
cleaning procedure performed until the weight of the underwent a liquid metal attack, developing a corroded
specimen remained constant. Each second corrosion porous layer with laminar morphology. The EDX ana-
specimen was cut, mounted, lapped and etched in order lyses performed on this layer, but on a not etched part,
to perform metallurgical analyses on its cross section showed a relevant Ni depletion and also the Cr content
with the aid of an optical microscope and scanning was less than in the bulk material. The electron probe
electron microscope (SEM). The chemical compositions micro analysis (EPMA) indicates that ferritisation had
of the corroded layers were evaluated by means of en- occurred, and it was also observed that the thickness of
ergy dispersive X-ray spectroscopy (EDX) using ZAF the ferrite layer was not uniform, ranging from few
correction method. The ZAF correction method con- micrometer to about 14 lm. The observed ferritisation
verts the X-ray intensities to concentration. It is based indicates that leaching of the steel elements in the liquid
on measurement of the pure element intensities and metal is a relevant parameter in the AISI 316L corrosion
composition of corrected factors for the Ôatomic number mechanism under the given test conditions. In fact the
effectÕ, Z, the Ôabsorption correctionÕ, A and the Ôfluo- solubility of Ni, Cr and Fe in the eutectic Pb–Bi at 673 K
rescence correctionÕ, F. is about 16000 wppm, 3.90 wppm and less than 1 wppm
The sets of four tensile specimens were tested with the [6] respectively, considerably supporting the experi-
SYNTECH test machine at 673 K and under Ar at- mental results.
mosphere without removing the traces of Pb–Bi alloy The appearance of the cross sections of the corroded
from their surfaces. The tensile tests were conducted sample of martensitic steel T91 is shown in the SEM
under strain-controlled conditions with an extension micrographs of Figs. 4 and 5. In those figures, liquid
rate of 2 mm/min which corresponds to an initial strain metal penetration was visible as well as the presence of
rate of 3  10 3 s 1 . cavities at the surface of the steel. Considering the
morphology of the steel–liquid metal interface, it ap-
pears that the corrosion proceeds via transgranular and
3. Results and discussion intergranular dissolution of the steel elements. Further,
a preferential dissolution of steel elements could not be
3.1. Steels observed, as was for instance reported for martensitic
steel exposed to Pb–17Li [13,14] where Cr depleted
The weight change measurements performed on the corrosion layers were detected. Nevertheless, it should
samples after 1500 h of exposure to the liquid metal be taken into account that in these works the flow rate of
showed a weight loss for both steels. The correspond- the Pb–17Li was about 100 times lower than the flow
ing evaluated corrosion rate is 1:52  10 5 mg/mm2 h rate of the Pb–Bi in LECOR. Thus, in assuming the
(1:9  10 3 lm/h) for the austenitic steel AISI 316L, and same corrosion mechanism in Pb–17Li and in Pb–Bi, a
2:47  10 5 mg/mm2 h (2:9  10 3 lm/h) for the mar- flaking of the Cr depleted corrosion layer might oc-
tensitic steel T91. Fig. 3 shows the SEM micrograph of curred with the higher flow rate.
the AISI 316L cross section. In order to reveal the mi- Literature data about the corrosion rate of steels in in
crostructure of the corroded layer and the bulk, the Pb–Bi with low oxygen content is sparse [15], but liter-
cross section was etched with a solution of oxalic acid. ature data on the corrosion of steels in Pb–17Li is
 C. Fazio et al. / Journal of Nuclear Materials 318 (2003) 325–332 329

Table 2
Tensile test results (mean values) obtained at 673 K on non-
corroded and corroded AISI 316L and T91 steels
UTS R (0.2%) A% Z%
(MPa) (MPa)
AISI 316 L
As received 457  40 165  35 69  3 73  3
1500 h corr. 477  9 183  26 68  3 72  5
T91 steel
As received 644  26 532  42 22  1 72  2
1500 h corr. 639  14 522  10 18  1 42  8

Fig. 4. SEM micrograph of the unetched T91 cross section.

500

400

300

200

AISI 316L as received

100 AISI 316L exposed to PbBi for 1500 h

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Fig. 6. Engineering tensile curves of the non-corroded (as re-

ceived) and corroded AISI 316L samples.
Fig. 5. SEM micrograph of the T91 cross section etched with
oxalic acid.

700
plentiful [16]. Since the oxygen solubility at 673 K in
Pb–17Li is about 10 8 wt% [24], the Pb–17Li environ- 600
ment is arguably similar to Pb–Bi from the low oxygen
activity point of view, thus the Pb–17Li data can be 500

considered useful. Both of these references report that

400
austenitic steels exhibit a higher corrosion rate than
martensitic steels. Thus, it appears that the comparable 300

corrosion rates of the austenitic and martensitc steels T91 steel as received
200
here observed do not agree with the cited literature, al- T91 steel exposed
though it could be assumed that the steel element dis- 100
to PbBi for 1500 h
solution was itself preceded by decomposition of the
natural oxide layer present on the steel surfaces. A 0

similar process has been described in the past to explain 0.00 0.05 0.10 0.15 0.20 0.25

the observed corrosion rate data obtained in Pb–17Li,

which exhibited a discontinuity with increasing exposure Fig. 7. Engineering tensile curves obtained on the non-cor-
time [17,23]. Further measurements on specimens ex- roded (as received) and corroded T91 samples.
tracted from the LECOR loop after 3000 and 4500 h of
exposition could help to confirm this assumption.
The results obtained from the tensile tests performed
on the two types of steels before and after corrosion are elongation (A %) and area reduction (Z %) factors are
reported in Table 2. Here the average values of the yield given. In Fig. 6 a pair of stress–strain curves of AISI
(R (0.2%)) and ultimate tensile strength (UTS) as well as 316L steel corroded and non-corroded and in Fig. 7 a
330 C. Fazio et al. / Journal of Nuclear Materials 318 (2003) 325–332

pair of stress–strain curves of T91 specimens corroded

and non-corroded are reported.
Comparing the tensile behaviour of the non-corroded
and corroded austenitic steel it can be seen that the Pb–
Bi corrosion and the consequent ferritisation did not
affect its tensile behaviour (see Table 2 and the curves
reported in Fig. 6). Conversely, the corroded martensitic
steel produced a much lower area reduction factor (42%,
see Table 2) with respect to the non-corroded specimens
(72%, see Table 2). The metallographic analysis of the
AISI 316L and T91 fracture surfaces are reported in the
SEM micrographs of Figs. 8 and 9 respectively. As is
shown in these figures, the bulk surface area of both
steels shows a dimpled fracture, which is the typical
morphology of ductile behaviour. Pb–Bi could be de-
tected in the external part of both types of steel, indi-
cating an intimate contact between the liquid metal and
the steel. The main differences that were observed be-
tween the two steels was that on the periphery of the T91
fracture surface a flat morphology could be seen with the
presence of microcracks, whilst on the AISI 316L steel
the same morphology was not present. This behaviour
of the corroded T91 steel is quite different to similar past
work performed on martensitic steels exposed to Pb–
17Li [13,18,19], where the liquid metal did not affect the
tensile properties. Moreover, liquid metal embrittlement

Fig. 9. SEM micrograph, fracture surface of the T91 sample.

tests performed in Pb [20] and Pb–17Li [21] have shown

no influence on the mechanical behaviour of normal
heat-treated martensitic steels.
It could be that the more aggressive Bi in the Pb–Bi
alloy together with long exposure of the T91 steel to the
liquid metal might play some role in the decrease of
the area reduction coefficient observed. Tensile tests of
the LECOR corroded samples after 3000 and 4500 h are
expected to give a clearer idea of this phenomenon. It
could be assumed that also low cycle fatigue tests might
give more indications about the liquid metal effect on the
mechanical performances of the steels.

3.2. Refractory metals

The measured weight loss of both the tungsten and

molybdenum samples was 7:26  10 7 mg/mm2 h (7:1 
10 5 lm/h) and 9:1  10 7 mg/mm2 h (4:7  10 5 lm/h)
respectively. From a metallographic point of view, as is
shown in the SEM micrograph of Figs. 10 and 11 both
metals exhibited an almost smooth interface with the
lead–bismuth alloy. No evidence of liquid metal attack
on the surface of the two materials, or of the growth of
Fig. 8. SEM micrograph, fracture surface of the AISI 316L an oxide layer, could be detected. The weight loss
sample. measured could be associated with both the uniform
 C. Fazio et al. / Journal of Nuclear Materials 318 (2003) 325–332 331

the test with an oxygen probe installed on the hot part of

the loop. All the materials tested in the LECOR loop
exhibited a weight loss after 1500 h of exposure. The
measured corrosion rate of W and Mo, in the range of
10 5 lm/h, was two order of magnitude lower than that
of the AISI 316L and T91 steels. As regards the corro-
sion mechanism, the AISI 316L steel exhibited mainly a
Ni depletion and consequent ferritisation to a maximum
depth of 14 lm from the surface. The T91 steel under-
went steel element dissolution with the formation of
cavities and no depleted layer could be observed.
Moreover, the estimated corrosion rates of the austentic
Fig. 10. SEM micrograph of the W cross section. steel and the martensitic steel were of the same order of
magnitude, which can be explained assuming a first step
of decomposition of the natural oxide layer usually
present on the steels. Further ongoing experiments will
confirm this assumption. The tensile behaviour of the
AISI 316L seems to be unaffected by corrosion in
flowing Pb–Bi, while the T91 steel exhibited a decrease
of the area reduction factor with a mixed ductile–brittle
fracture morphology. It was assumed that a relationship
between the brittle fracture mode occurring mainly on
the peripheral part of the T91 tensile samples, and the
presence of Bi plus the long exposure time to the liquid
metal could give rise to this phenomenon. More exper-
iments will be needed to confirm this assumption.

Fig. 11. SEM – micrograph of the Mo cross section.

Acknowledgements
dissolution and the low solubility of W and Mo in Pb–Bi
This work has been supported by the EU project
liquid at 673 K. In comparing the estimated corrosion
TECLA. The authors are grateful to Mr L. Rapezzi and
rate obtained on the refractory metals with that ob-
his collaborators for operation of the LECOR loop, to
tained on steels, it is evident that the low oxygen content
Mr A. Agostini and his colleagues of the ENEA-Brasi-
in the liquid metal, the refractory metals exhibited two
mone Laboratory of Metallurgy for the specimen prep-
orders of magnitude greater resistance to corrosion than
aration assistance and to Mr S. Storai for performing
the steels. On the other hand, previous studies [8,9]
the tensile tests.
performed in static and oxygen saturated lead at 793 K
have shown that W and Mo exhibit friable and not
protective oxides. Even though the test temperature in
References
these studies is higher than the LECOR temperature,
calculations of the free energy of oxides formation sug- [1] C. Rubbia, J.A. Rubio, S. Buono, F. Carmianti, N. F
ıetier,
gest that these would be likely even at 673 K, given J. Galvez, C. Gel
es, Y. Kadi, R. Klapish, P. Mandrillioni,
sufficient oxygen activity in the liquid metal. J.P. Revol, C. Roche, European Organisation for Nuclear
Research, CERN report AT/95-44 (ET), 1995.
[2] L. Cinotti, M. Bruzzone, S. Cardini, G. Corsini, G.
4. Conclusions Saccardi, XADS Pb–Bi Cooled Experimental Accelerator
Driven System Reference Configuration, Doc. ANSALDO
The aim of the LECOR loop is to perform materials ADS1 SIFX 0500, June 2001.
compatibility tests in Pb–Bi eutectic having a low oxy- [3] G.S. Bauer, M. Salvatores, G. Heusener, J. Nucl. Mater.
296 (2001) 17.
gen activity, thereby simulating the conditions of
[4] H. A€ıt Abderrahim, P. Kupschus, E. Malambu, Ph. Benoit,
the structural materials of both the MEGAPIE and K. Van Tichelen, B. Arien, F. Vermeersch, P. DÕhondt, Y.
the MYRRHA targets. The low oxygen activity was Jongen, S. Ternier, D. Vandeplassche, Nucl. Instrum. and
achieved by adding Mg, which acts as oxygen getter, and Meth. A 463 (2001) 487.
by pre-conditioning the liquid metal with hydrogen gas. [5] G. M€ uller, G. Schumacher, F. Zimmermann, J. Nucl.
The oxygen activity was successfully monitored during Mater. 278 (2000) 85.
332 C. Fazio et al. / Journal of Nuclear Materials 318 (2003) 325–332

[6] B. Xiaoyi He, N. Li, M. Mineev, J. Nucl. Mater. 297 (2001) [16] G. Benamati, in: Proceeding of the Fifth Workshop on
214. Liquid Metal Blanket Experimental Activities, 16–18
[7] F. Barbier, G. Benamati, C. Fazio, A. Rusanov, J. Nucl. September 1997, Paris, France, CEA report DMT 97/442,
Mater. 295 (2001) 149. SERMA/LCA 2113.
[8] G. Benamati, P. Buttol, V. Imbeni, C. Martini, G. [17] H.U. Borgstedt, H.D. Rohrig, J. Nucl. Mater. 179–181
Palombarini, J. Nucl. Mater. 279 (2000) 308. (1991) 596.
[9] G. Benamati, P. Buttol, V. Imbeni, C. Martini, G. Poli, [18] O.K. Chopra, D.L. Smith, J. Nucl. Mater. 141–143 (1986)
Workshop on Heavy Liquid Metal Technology, Karlsruhe, 566.
16–17 September 1999, FZK Report: FZKA 6389, De- [19] H.U. Borgstedt, M. Grundmann, Fusion Eng. Des. 6
cember, 1999. (1988) 155.
[10] J.R. Weeks, A.J. Romano, Corrosion – Nace 25 (3) (1969) [20] G. Nicaise, A. Legris, J.B. Vogt, J. Foct, J. Nucl. Mater.
131. 296 (2001) 256.
[11] O. Kubaschewski, C.B. Alcock, P.J. Spencer, Materials [21] T. Sample, H. Kolbe, J. Nucl. Mater. 283–287 (2000)
Thermochemistry, 6th Ed., Pergamon, Oxford, 1993. 1336.
[12] I. Ricapito, C. Fazio, G. Benamati, J. Nucl. Mater. 301 [22] V. Troyanov, A. Rusanov, IPPE doc. UDC 621.029.553,
(2002) 60. Obninsk, 2001.
[13] G. Benamati, C. Fazio, I. Ricapito, Mechanical and [23] M. Broc, P. Fauvet, T. Flament, A. Terlain, J. Sannier,
corrosion behaviour of EUROFER 97 steel exposed to Proceeding on the Fourth International Conference on
Pb–17Li, 10th International Conference on Fusion Reactor Liquid Metal Engineering and Technology, 17–21 October
Materials, 14–19 October 2001, Baden-Baden, Germany. 1988, Avignon, France.
[14] N. Simon, A. Terlain, T. Flament, J. Nucl. Mater. 254 [24] H.U. Borgstedt, Chemische Eigenschaftern des fl€ ussigen
(1998) 185. Blanketstoffs Pb17Li, FZK report, KfK 4620, August
[15] G. Ilinecev, REZ doc.: Nr. DITI 11 641 M, July 2001. 1989.