Beruflich Dokumente
Kultur Dokumente
The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
a r t i c l e i n f o a b s t r a c t
Article history: Self-passivating tungsten-based alloys may provide a major safety advantage in comparison with pure
Available online 30 April 2011 tungsten, which is presently the main candidate material for the plasma-facing protection of future
fusion power reactors. WCrSi alloys were manufactured by mechanical alloying (MA) and HIP at 1300 ◦ C
Keywords: and 200 MPa for 1 h. Different MA conditions were investigated to obtain powders with lowest possible
Tungsten alloys amount of contaminants and small and homogeneous particle and crystallite size. Milling in WC vials
Mechanical alloying
under Ar without process control agent provided best results. After HIP densities close to 100% were
HIP
obtained. First oxidation tests on preliminary alloys showed self-passivating behavior with rates com-
Oxidation resistance
Armor material
parable to WCrSi thin films at 800 ◦ C but worse performance at 1000 ◦ C. In all cases a Cr2 WO6 protective
layer is formed at the surface.
© 2011 Elsevier B.V. All rights reserved.
1. Introduction (MA) + hot isostatic pressing (HIP)) are presented. First results of
oxidation tests up to 1000 ◦ C are shown.
Tungsten is presently the main candidate material for the
plasma exposed areas of future fusion reactors (DEMO) mainly 2. Experimental
due to its low erosion yield by sputtering with plasma particles
[1]. However, the use of W as first wall (FW) material implies a Elemental powders of pure W (99.9%, 10 m), Cr (99%, 40 m)
safety concern in case of a loss-of-coolant accident with air ingress and Si (99.5%, 40 m) were mechanically alloyed in a SPEX 8000D
into the reactor vessel. In this situation, temperatures up to 1100 ◦ C mill, using vials and grinding media of both hardened steel (HS)
would be achieved in the in-vessel components of a He-cooled reac- and WC with ball-to-powder weight ratios (BPR) of 4:1 and 10:1,
tor within 10 to 30 days due to the decay heat [2], leading to the respectively and balls of two different diameters: 12 and 5 mm. Two
formation of WO3 and subsequent evaporation of (WO3 )x -clusters different compositions were studied: WCr10Si10 and WCr5Si5 (wt.
[3,4]. A possible way for avoiding this would be the addition to W %). Initially, milling was performed in air (only HS vials), and later
of stable oxides-former alloying elements, so that at high temper- on, in Ar to minimize powder contamination by O. Grinding was
ature in the presence of oxygen a self-passivating layer is formed performed both using hexane as process control agent (PCA) and
protecting the material from further oxidation. In previous works without PCA (only WC vials). Different milling times (15 min to 6 h)
[5,6] different binary and ternary W alloys have been manufac- were investigated. The powders were uniaxially pressed at 100 MPa
tured via magnetron sputtering, demonstrating that W thin films to cylinders of ∅ 16 mm and ∼4 mm height. HIP of glass encapsu-
containing about 10 wt.% Si and 10 wt.% Cr exhibited excellent self- lated compacts was carried out at 1300 ◦ C for 1 h and 200 MPa.
passivating behavior when exposed to air at temperatures up to The O and C content of the powders was measured using a
1000 ◦ C. However, the PVD route is not applicable to DEMO because LECO analyzer. Powders and sintered samples were characterized
for the blanket FW, coatings or tiles with a thickness of several mm by FEG-SEM, EDX mapping, cross sectioning FIB and XRD. The open
are required for economic reactor operation [2]. porosity of the samples was measured by He pycnometry, and the
Powder metallurgy is a suitable and relatively low cost route density was determined from the geometrical and the He densities
for the production of alloys of complex composition with the assuming no residual porosity. Vickers microhardness measure-
desired microstructure. In this paper, first results on the manufac- ments were performed applying a load of 9.8 N for 15 s. Finally, first
turing of W-Cr-Si alloys by powder metallurgy (mechanical alloying oxidations experiments were performed at 600, 800 and 1000 ◦ C
with synthetic air at atmospheric pressure. The mass gain of the
samples due to oxide formation was measured by a thermobalance
∗ Corresponding author. Tel.: +34 943 212 800; fax: +34 943 213 076. Netzsch STA 449 F1 with a sensitivity of 0.025 g. The procedure is
E-mail address: cgrosales@ceit.es (C. García-Rosales). described in detail in [5].
0920-3796/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.fusengdes.2011.03.107
Author's personal copy
Table 1
MA conditions of manufactured alloys, contaminants after MA and some properties after HIP.
Sample code Composition MA conditions Impurities after MA (wt.%) Properties after HIP
3. Results and discussion reason for a reduction of the milling time. These observations were
the same for the other composition.
First trials of MA were carried out in HS vials (BPR 4:1) under In Fig. 2 powders 9W-WC-ih-2 and 9W-WC-i-1 at different
air with milling time ranging from 15 min to 6 h, using hexane as magnifications are shown. After 2 h milling in hexane, a very effec-
PCA to achieve effective particle fracture and avoid excessive cold- tive reduction in particle size is observed, more pronounced than
welding [7]. With increasing milling time Fe contamination from in powders milled in HS vials because of the larger density of
the vial and balls (measured by EDX) rose steadily, reaching val- WC and the higher BPR. In addition, the initial Cr and Si parti-
ues as high as 6.1 wt.% after 6 h milling. Also C contamination from cles cannot be distinguished anymore with BSE. Most W particles
hexane increased steadily up to 0.31 wt.% after 6 h MA. On the other are flake-shaped. After 1 h milling without PCA, a larger and less
hand, the O content of the powders (initially 0.16% for WCr10Si10, homogeneous particle size is observed, most probably because the
0.11% for WCr5Si5) increased up to 0.76% after 2 h and then sta- absence of PCA promotes cold welding; this can be verified at larger
bilized at 0.50% after 6 h MA. Under all conditions, contamination magnification (Fig. 2 (d)), where cold-welded flakes forming lamel-
of WCr5Si5 was slightly lower than that of WCr10Si10. XRD spec- lar structures are visible (yellow arrow). Again, no Cr and Si can be
tra of the powders showed the peaks of the elemental powders detected by BSE.
without formation of new phases, indicating that no alloying but According to these results, future work in a planetary ball mill
energetic milling took place. Estimation of the crystallite size from with larger amounts of powders will be performed in WC vials
the W peaks width using the Williamson-Hall method [8] indi- under Ar without PCA.
cated a strong decrease of crystallite size with milling time, which All powders milled in Ar were compacted, encapsulated and
stabilized between 2 and 3 h. This result is in line with SEM obser- HIPed, achieving in most cases a densification above 93%, and in
vation of the powders. Therefore, 2 h milling time was selected for some cases close to 100%. The exceptions are samples 8W-HS-
future work at the SPEX as a good compromise between an effective ih-2 (87.5%) and especially 8W-WC-ih-2 (75.9%), where possibly
milling and tolerable amount of contamination. some encapsulation problems occurred. No clear correlation of the
In Table 1 the milling conditions and contaminants after MA of densification after HIP with the manufacturing conditions can be
the materials prepared during this work are listed. First samples deduced from these results. Samples with composition WCr5Si5,
for HIPing were prepared with WCr10Si10 and WCr5Si5 powders i.e. with larger amount of W, seem to densify better. Fig. 3 shows the
milled in HS under air with hexane as PCA for 2 h (8W-HS-ah-2 and microstructure of samples 9W-WC-ih-2 and 9W-WC-i-1 (milled in
9W-HS-ah-2). Densities above 93% were achieved, proving that it WC vials with and without hexane for 2 and 1 h, respectively) after
is feasible to obtain dense alloys by this method. HIP. As expected from the powder morphology, the microstruc-
To reduce the O content, powders of both compositions were ture of sample 9W-WC-ih-2 is finer and more homogeneous, while
milled under the same conditions than before but in inert atmo- sample 9W-WC-i-1 exhibits larger W flake-shaped particles. A
sphere (Ar) (8W-HS-ih-2 and 9W-HS-ih-2). As shown in Table 1, larger amount of dark phase is present in sample 9W-WC-ih-2,
the O content could be reduced by a factor of 2-3. Thus, all fur- which had a significantly higher amount of C (s. Table 1). Accord-
ther powders were milled in Ar. Surprisingly the C content of these ing to EDX mapping, this dark phase corresponds to carbides and
powders was also reduced by a factor of ∼1.5. oxides in sample 9W-WC-ih-2, and to oxides in sample 9W-WC-
To eliminate Fe contamination, subsequent millings were per- i-1. Cross-sectioning by FIB demonstrated the practical absence of
formed using WC vials and balls (BPR 10:1), and maintaining porosity. At higher magnification, ultrafine intergranular oxides can
hexane as PCA (8W-WC-ih-2 and 9W-WC-ih-2). However, while be detected, which may have a beneficial effect by preventing grain
the HS vials are tight, the WC vials do not completely ensure leak
tightness. Nevertheless, after several trials it was possible to obtain
powders with an O content comparable to that of previous powders
milled in HS and Ar (s. Table 1). Additionally, a significant increase
in the C content, coming now not only from hexane but also from
the WC vials, was observed. Hence, new powder batches were pre-
pared under the same conditions without hexane as PCA and for a
milling time of 1 h (8W-WC-i-1 and 9W-WC-i-1). It was managed
to produce powders of both compositions with relatively low O
content (< 0.3 wt.%) and with a C content (< 0.06 wt%) 6-10 times
lower than in the previous case. Nevertheless, it should be taken
into account that the milling time was lower. Besides, whereas WC
was clearly detected by XRD in 8W-WC-ih-2, no WC peaks were
visible in the XRD spectra of the 8W-WC-i-1 powder (Fig. 1). More-
over, the Si peaks almost disappeared in the 8W-WC-i-1 powders, Fig. 1. XRD spectra of the 8W-WC-ih-2 and 8W-WC-i-1 powders milled in WC vials
which may indicate a more effective milling; this was actually the under Ar with and without hexane as PCA and for 2 and 1 h, respectively.
Author's personal copy
Fig. 2. BSE-SEM images of WCr5Si5 powders after MA in WC vials under Ar: (a)-(b) with hexane as PCA, 2 h milling (9W-WC-ih-2); (c) -(d) without PCA, 1 h milling
(9W-WC-i-1).
growth and improving ductility. Further investigation with TEM is sponds mainly to (WCr)5 Si3 and to a small amount of pure Cr. This
required to clarify the type of oxides. EDX and XRD analysis revealed is in agreement with [9], whereby complete solid-solution of Cr5 Si3
that the brightest phase corresponds to pure W, possibly with a and W5 Si3 exists. According to [9], only the silicide and almost no
small amount of Cr in solid solution, while the grey phase corre- pure W should be found in samples of composition WCr10Si10;
Fig. 3. BSE-SEM images of HIPed WCr5Si5 samples milled in WC vials: (a)-(b) 9W-WC-ih-2; (c)-(d) 9W-WC-i-1.
Author's personal copy
Fig. 5. Arrhenius plot of parabolic oxidation rates of pure W and WSi8Cr12 thin
films, and of sample 8W-HS-ah-2.
Fig. 4. SEM cross sectional view of sample 8W-HS-ah-2 after oxidation at 1000 ◦ C
for 3 h. A Cu protection layer was deposited prior to FIB examination.
Acknowledgments [3] C.J. Smithells (Ed.), Metal Reference Book, 8th ed., Butterworth, London,
2004.
[4] Y.A. Yang, Y. Ma, J.N. Yao, B.H. Loo, J. Non-Cryst. Solids 272 (2000) 71.
This work has been funded by the European Community [5] F. Koch, H. Bolt, Pys. Scr T128 (2007) 100.
within the framework of EFDA WP08-09 under the contract of [6] F. Koch, S. Köppl, H. Bolt, J. Nucl. Mater 386–388 (2009) 572.
EURATOM/CIEMAT, and by the Spanish Ministry for Science and [7] C. Suryanarayana, Prog. Mater Sci 46 (2001) 1.
[8] G.K. Williamson, W.H. Hall, Acta Metal 1 (1953) 22.
Innovation (ENE2009-09511, subprogram FTN). [9] R. Kieffer, O. Schob, H. Nowotny, F. Benesovsky, Monatsh. Chem 93 (1962) 517.
[10] S.W.H. Yih and C.T. Wang, Tungsten, Plenum Press, New York and London
References (1979).
[11] R. Meyer, H. Pastor, B. Soc. Fr. Ceram (1969) 59.
[12] W. Kunnmann, S. La Placa, L.M. Corliss, J.M. Hastings, E. Banks, J. Phys. Chem.
[1] H. Bolt, V. Barabash, W. Krauss, J. Linke, R. Neu, S. Suzuki, et al., J. Nucl. Mater.
Solids 29 (1968) 1359.
329–333 (2004) 66.
[2] D. Maisonnier, I. Cook, P. Sardain, R. Andreani, L. Di Pace, R. Forrest et al., A
Conceptual Study of Commercial Fusion Power Plants, Final Report, EFDA-RP-
RE-5.0, 13 April 2005.