Simultaneous effects of neutron irradiation and long–term sputtering on the surface relief of ITER–grade tungsten were studied. The effects of neutron–induced displacement damage have been simulated by irradiation of tungsten target with W6 + ions of 20 MeV energy. Ar+ ions with energy 600 eV were used as imitation of charge exchange atoms in ITER. The surface relief was studied after each sputtering act. The singularity in the WJ–IG surface relief was ascertained experimentally at the first time, which determines the law of roughness extension under sputtering. As follows from the experimental data, the neutron irradiation has not to make a decisive additional contribution in the processes developing under impact of charge exchange atoms only.

Due to its good thermal properties (high melting point and good thermal conductivity) and low sputtering yield tungsten (W) is foreseen as one of plasma–facing materials in fusion reactors such as International Thermonuclear Experimental Reactor (ITER).1 Tungsten is chosen as the material of the baffle and dome of the divertor in ITER, both will be plasma facing components.2 Recently, it was decided that divertor target plates will also be made of tungsten. Following ITER, an electricity producing fusion reactor DEMO is under development, in which tungsten is a leading candidate as plasma–facing material.3 It is clear, that tungsten should be highly resistive to radiation impact of various types.1,4–6

Tungsten polycrystalline material with enhanced thermal conductivity along a certain direction is called ITER–grade tungsten. During manufacturing process W material is rolled either in the form of a rod or a plane. Rolling results in grain elongation anisotropy since the grains are elongated along the rolling direction. The elongated grain orientation of ITER–grade W should be parallel to the heat transfer direction. The thermal conductivity is supposed to be higher for the material with perpendicular grain elongation with respect to the surface than for the material with parallel grain elongation. That is why W material with perpendicular grains is favorable as a plasma facing material in the divertor region.2 

According to Ref. 7, W components with the total area about one hundred  m2 will receive comparatively low heat flux but high charge exchange neutral fluxes of a wide energy distribution. The main criterion to the ITER–grade tungsten should be the high resistance to radiation damages due to impact of neutrons and charge exchange atoms (CXA), i.e. D and T atoms or ions (D+, T+ and impurity ions). Simultaneous impact of these factors can result in faster modification of the surface of ITER–grade tungsten in comparison with the case of CXA only, due to the defects created by neutrons in the near–surface layer. The clearing–up this question is important for ITER with the present in–vessel construction conception.

At present a few Companies are trying to produce tungsten with properties satisfying those requirements. The A.L.M.T. Corp. in Japan is one of such Companies, and below we present and discuss the results obtained with ITER–grade W specimens, produced by this Company, when they were subjected to long–term sputtering with Ar+ ions after preliminary bombardment with 20 MeV W+6 ions (self–damaged ITER–grade tungsten).

By this time several papers were devoted to simulation of neutron irradiation effects on character of interaction of ITER–grade tungsten with deuterium, when neutron impact process was simulated by exposure of W specimens to bombardment with 20 MeV W+6 ions.8,9 However, the important question relating the effects of neutron irradiation on formation of the surface structure under CXA was only slightly touched in Refs. 10,11. In the present paper the results of detailed investigations of surface structure of self–damaging ITER–grade tungsten specimens under simultaneous impact of neutron irradiation and long–term sputtering are presented.

The experiments were provided with ITER–reference tungsten grade. The tungsten plates with purity of 99.99wt% and with 99.7% of theoretical density were prepared (A.L.M.T. Corp., Japan) by a quite complicated technology.8 Polycrystalline W workpieces were deformed (rolled, swaged and/or forged) with following appropriate heat–treatments to obtain better mechanical properties (strength and toughness). Then plates were cut into samples of 10 × 10 × 2 mm3, and double-side mechanically and electrochemically polished to a mirror–like surface. Later in the text this type of tungsten is marked as WJ–IG.

The microstructure of the ITER–grade tungsten consists of anisotropic elongated grains along the deformation axis and the grain size is about 1−3 μm in section and up to 5 μm in length.8 The elongated grain orientation is defined to be parallel to the heat transfer direction, i.e., mainly perpendicular to the surface.

To model neutron irradiation in ITER, two WJ–IG specimens were bombarded with W+6 ions of 20 MeV energy. The irradiated area was a circle with 8 mm in diameter. The calculated dose depth profile has a maximum 0.3 or 3.0 dpa (the details of calculations can be found in Refs. 8,9). The unirradiated area of specimens was not exposed to W+6 ions and used as the control surface in following ion sputtering acts, simulating CXA sputtering in ITER.

The initial reflectance of specimens was found to be noticeably below the nominal values for W,12 therefore a “soft” cleaning of their surfaces was provided13,14 before sputtering acts were started. The unirradiated with W+6 ions sides were cleaned by exposing to low energy (∼60 eV) ions of deuterium plasma, and the irradiated sides — by short time exposing to ions of Ar plasma with 200 eV energy. In the both cases the electron cyclotron resonance (ECR) discharge (frequency f = 2.35 GHz) was used for plasma production in deuterium and argon, respectively. Similar cleaning procedure is routinely used for removing organic contaminants from the surface of mirror specimens prepared for the sputtering test.14 After cleaning, the reflectance of specimens agreed closely with the nominal values for W.12 

Following the cleaning process, a number of identical sputtering acts were provided and after every one the specimen surface was analyzed ex–situ via optical microscopy and interferometry, reflectometry (measurement of reflectance at normal incidence R(λ) within 220–650 nm) and ellipsometry (ellipsometric parameters Ψ and Δ within 450–760 nm spectral range) in dependence on Ar ions fluence. The results of optical and ellipsometry measurements have been described and discussed in detail in our previous paper11 and therefore will be mentioned here only in the Discussion section. The sputtering on both sides was performed in 7 steps, with the accumulation of fluences (Table I).

Table I.

Numbers of sputtering acts and the corresponding fluences.

Act numberFluence,  ion/m2
F2.3 × 1022 
F5.7 × 1022 
F1.2 × 1023 
F2.9 × 1023 
F4.4 × 1023 
F5.2 × 1023 
F6.5 × 1023 
Act numberFluence,  ion/m2
F2.3 × 1022 
F5.7 × 1022 
F1.2 × 1023 
F2.9 × 1023 
F4.4 × 1023 
F5.2 × 1023 
F6.5 × 1023 

After exposure to plasma ions was finished the surface of specimens was analyzed with confocal laser scanning microscopy (CLSM) and electron backscatter diffraction (EBSD) methods.

To shorten the experiment time, the Ar+ ions with 600 eV energy were used instead of hydrogen isotope ions. Both specimens were exposed simultaneously, first one side, then the other.

Microimages of the specimen surfaces were obtained using optical microscopes. Microinterferometric setup15 was used to investigate the surface structure.

After cleaning, the surfaces of both sides were smooth as follows from the straight interference lines seen in an interferometer microscope. Fig. 1 displays interference patterns for an unirradiated (0 dpa) and irradiated (3 dpa) sides of a specimen after several successive sputtering acts. Each sputtering has led to substantial changes in the sample surface, which are related to the appearance of specific roughness, first, in separate regions of the surface (shown by white arrows in Figs. 1(a)–1(c)), then on almost the entire surface (Figs. 1(e) and 1(f)).

FIG. 1.

Interference patterns of the surface of samples of WJ–IG: (a, c, e) the unirradiated side (0 dpa); (b, d, f) the irradiated side (3 dpa); (a, b) after cleaning and sputtering by Ar ion fluence in the regime F1 = 2.3 × 1022 ion/m2; (c, d) F2 = 5.7 × 1022 ion/m2; and (e, f) F7 = 6.5 × 1023 ion/m2. White arrows indicate shifts of interference fringes on the developing roughness. Black arrows show the preferred direction of the roughness.

FIG. 1.

Interference patterns of the surface of samples of WJ–IG: (a, c, e) the unirradiated side (0 dpa); (b, d, f) the irradiated side (3 dpa); (a, b) after cleaning and sputtering by Ar ion fluence in the regime F1 = 2.3 × 1022 ion/m2; (c, d) F2 = 5.7 × 1022 ion/m2; and (e, f) F7 = 6.5 × 1023 ion/m2. White arrows indicate shifts of interference fringes on the developing roughness. Black arrows show the preferred direction of the roughness.

Close modal

Thus, with every next sputtering act the interference patterns become more and more distorted and are practically indistinguishable after the last exposure to Ar+ ions bombardment (Figs. 1(e) and 1(f)), which indicates the occurrence of fundamental changes in the surface. The scale of the longitudinal wavelength of inhomogeneity is different in the different direction and consists on an average ∼30 μm and ∼100 μm along the horizontal (⊥ the arrow), and vertical (∥ the arrow), respectively. It is hard to say anything about the inhomogeneities height since the interference image is very eroded.

Preferred direction of roughness elements are evident (shown with black arrows on Figs. 1(e) and 1(f)). This preferred direction is related to the technique of the sample preparation. Anisotropy of polycrystale properties is connected with the preferential orientation of grains, which causes the anisotropy of the roughness on the surface. The sintering process of such kind of texture polycrystales is known now. This axes of deformation (texture) appears as a result of the preliminary thermohardening treatments, including some operations: different kinds of hot rolling, annealing et al. This preferential directional inhomogeneity of the relief is clearly seen also from CLSM and EBSD data (Figs. 2 and 3). The grains perpendicular to the surface are shown using definite grades of grey color on Figs. 2 and 3.

FIG. 2.

2D–results of CLSM measurements for ITER–grade W after sputtering with ions of argon plasma (F7 = 6.5 × 1023 ion/m2). Size of the image is 1100 × 1100 μm2.

FIG. 2.

2D–results of CLSM measurements for ITER–grade W after sputtering with ions of argon plasma (F7 = 6.5 × 1023 ion/m2). Size of the image is 1100 × 1100 μm2.

Close modal
FIG. 3.

(a) — EBSD data for ITER–grade W specimen (0.3 dpa), size is 90 × 90 μm2; (b) — 3D–results of CLSM and EBSD measurements for smaller fragment (10 × 10 μm2) of the surface after sputtering with ions of argon plasma F7 = 6.5 × 1023 ion/m2, indicating the orientation of grains perpendicular to the surface using grades of grey color.

FIG. 3.

(a) — EBSD data for ITER–grade W specimen (0.3 dpa), size is 90 × 90 μm2; (b) — 3D–results of CLSM and EBSD measurements for smaller fragment (10 × 10 μm2) of the surface after sputtering with ions of argon plasma F7 = 6.5 × 1023 ion/m2, indicating the orientation of grains perpendicular to the surface using grades of grey color.

Close modal

From EBSD data (Fig. 3) one can see that some areas with equal colors are tens of microns in size and have a tendency to be oriented from top–left to down–right. Colors equality means that orientations of grains inside such area (we can call them “conglomerates”) are close to each other. Analysis of the Fig. 3(b) data is in a good agreement with the CLSM and EBSD results for polycrystalline tungsten:10 if the inclination of grain orientation from a given crystallographic axis is not more then ±15°, the sputtering rate will be similar to grains oriented exactly along the given axis. Each group of grains (conglomerate) has the definite color.

Correspondingly, the sputtering rate of areas with closely oriented grains are different, namely, according to the results of Ref. 10 the areas with domination of grains with orientation close to (110) have highest sputtering rate (dark grey), forming “valleys”, and the areas with domination of grains with orientation close to (111) have much lower sputtering rate (light grey) forming the main parts of “ridges” with “peaks” composed of grains close to (111) orientation having lowest sputtering yield. Both, “ridges” and “valleys”, are clearly seen in Figs. 1(e) and 1(f). These conglomerates form an anisotropic sub–structure of the material, with groups of varying sizes, thus contributing to long wavelength surface roughness. The mean distance between such groups might be a factor in determining the scale that characterizes the longitudinal wavelength of inhomogeneity.

Such surface relief differs very much from the relief of conventional polycrystalline specimens being sputtered in similar conditions10,16 and can probably be the consequence of a quite complicated technology procedures used for preparation of this particular kind of tungsten.8 Note that such surface peculiarities are qualitatively similar for sputtered unirradiated and for preliminary irradiated specimens, independently on the dose (0.3 or 3.0 dpa).

As can be seen on the Fig. 2 the scale of the longitudinal wavelength of inhomogeneity is different in the different direction and consists ∼30 μm and ∼130 μm along the horizontal and vertical, respectively. These sizes are in very good accordance with the data of the interference images (Figs. 1(e) and 1(f)).

It should be mentioned that characteristic size of the relief features along the surface (tens of microns) is much longer then the size along the surface of individual grains (1–3 μm, as was mentioned above), which are practically not distinguishable in Figs. 1–3.

Thus, the structural data obtained in these experiments demonstrate that, similar to the cases of polycrystalline tungsten,10 the real relief developing on the ITER–grade W surface under sputtering has much more large–scale characteristics than the micro–structural constituent elements of this kind tungsten, i.e., the size of the crystallines of ∼5 μm in depth and ∼1–3 μm in width are not determinative for characteristics of microrelief that develops due to sputter erosion.

Analysis of the experimental data shows a strong directional inhomogeneity of all measured parameters which appears due to sputtering, independently if the specimen was self–damaged or not. This is a reflection of the inhomogeneity of surface relief which starts to develop soon after beginning of sputtering acts (Fig. 1). There is no doubt that the reason of relief inhomogeneity lies in the technology of preparation of these particular W specimens. According to Ref. 8, the technology included the processes, which are known to result in appearance of some preferred orientation of grains. The details of preferred grain orientation depend on the processing history and on the material itself.

The structure of this particular tungsten is characterized by formation of “conglomerates” with closely–oriented grains, i.e., with different “average” sputtering yield, separated from each other. The size of the conglomerates that much exceeds (by more than one order of magnetude) the size of individual grains does define the appearance of long wavelength surface roughness. The conglomerates form an anisotropic substructure of the W. They have an evident elongation in some preferred direction (in Figs. 1(e) and 1(f) this direction is indicated with arrows) resulting in a strong spatial inhomogeneity of the relief that developed under sputtering. As a result, the surface of WJ–IG specimens starts to reasonable mountain “ridges” and “valleys”, which correspond to grains with lesser and grater sputtering rates in the main , respectively. In accordance to the EBSD results on the surface of recrystallized tungsten specimens,10 in “valleys” the grains with (110) orientation are predominant, and in “ridges” predominate (100) grains with “peaks” corresponding to (111) orientation.

The independence of normal incidence reflectance degradation under long–term sputtering on whether W mirror specimens were preliminary irradiated with 20 MeV W ions or not,11 is in agreement with previously published results on sputtering of copper and stainless steel mirror specimens preliminary irradiated to varied doses with Cu and Cr ions of 3 MeV energy, correspondingly,16 as well as with recently published data for recrystallized tungsten.10 

The singularity in the WJ–IG surface relief was ascertained experimentally at the first time, which connected with fact that the crystallites are combined in local groups (conglomerates) of closely orientated grains with approximately the same sputtering yield. These conglomerates form an anisotropic sub–structure of the material, with groups of varying sizes, thus contributing to long wavelength surface roughness. The mean distance between such groups might be a factor in determining the scale that characterizes the longitudinal wavelength of inhomogeneity and determines the law of roughness extension under sputtering. Thus, the roughness in WJ–IG surface is based on the two factors: 1) the dependence of sputtering rate on crystalline orientation and 2) technology procedures used for preparation of this particular kind tungsten. It is clear why the roughness develops unevenly over the WJ–IG surface under sputtering (Fig. 1): in the first place the roughness emerges on such kind of the surface, which is least resistant to sputtering.

The results of ellipsometry are rather intriguing, since a quite large difference was found of parameters Ψ and Δ for specimens irradiated (3 dpa) and unirradiated with 20 MeV W ions. The attempt to explain the difference of ellipsometry data with data of other methods was made in Ref. 11, where special investigation and modeling was provided. A comparative analysis of experimental data obtained by reflectometry and ellipsometry made it possible to suggest a model of the process of surface modification for samples of ITER–grade tungsten that were preliminarily irradiated with high energy tungsten ions. The results of the present paper are in a good agreement with the realistic model of process of surface modification for ITER–grade tungsten.11 Two different scales in the surface roughness exist: coarse–scaled and fine–scaled. The factor that is responsible for coarse–scaled roughness is sputtering, which is equal for both irradiated and unirradiated sides of the samples. The fine–scaled roughness forms due to the preliminary irradiation by tungsten ions only. As follows from results of Ref. 11, the neutron irradiation is connected with the defects created by neutrons in the near–surface layer. So, at the damage rate that would be achieved in ITER, it has already to make some additional contribution in the processes developing under impact of charge exchange atoms only, but this contribution is not strong and decisive in the roughness formation.

As far as the fine–scaled roughness is concerns, the height is less than 10 nm; the lateral size is less than optical microscope resolution.

The coarse–scaled roughness of the surface relief should strongly influence on the reflection from tungsten surfaces of atoms and electromagnetic radiation emanating by burning plasma, and thus can change the heat balance of the baffle and dome elements of the divertor in ITER.

Summarizing our results, we may come to conclusion that for exploitation of ITER there should be no a synergistic action of two key factors (neutrons and charge exchange atoms) on the sputtering rate of the material of baffle and dome of the divertor, made of ITER–grade tungsten.

The authors would like to thank Dr. V.S. Voitsenya and Dr. M. Balden for helpful discussions. We are also grateful to Dr. M. Balden for providing data for Figs. 2 and 3 for the analysis.

A.A. Galuza and I.V. Kolenov wish to express their gratitude for support from the National Academy of Science of Ukraine funded grant for young scientists under contract no. 0111U008185.

1.
O. V.
Ogorodnikova
,
T.
Schwarz-Selinger
,
K.
Sugiyama
, and
V. K.
Alimov
,
Journal of Applied Physics
109
,
013309
(
2011
).
2.
A.
Rusinov
,
M.
Sakamoto
,
H.
Zushi
,
R.
Ohyama
,
K.
Honda
,
I.
Takagi
,
T.
Tanabe
, and
N.
Yoshida
,
Plasma and Fusion Research
7
,
1405105
(
2012
).
3.
R.
Toschi
,
P.
Barabaschi
,
D.
Campbell
,
F.
Elio
,
D.
Maisonnier
, and
D.
Ward
,
Fusion Engineering and Design
56–57
,
163
(
2001
).
4.
I.
Bizyukov
,
K.
Krieger
,
N.
Azarenkov
, and
U.
Toussaint
,
Journal of Applied Physics
100
,
113302
(
2006
).
5.
I.
Bizyukov
and
K.
Krieger
,
Journal of Applied Physics
102
,
074923
(
2007
).
6.
O. V.
Ogorodnikova
,
J.
Roth
, and
M.
Mayer
,
Journal of Applied Physics
103
,
034902
(
2008
).
7.
G.
Federici
,
C. H.
Skinner
,
J. N.
Brooks
,
J. P.
Coad
,
C.
Grisolia
,
A. A.
Haasz
,
A.
Hassanein
,
V.
Philipps
,
C. S.
Pitcher
,
J.
Roth
,
W. R.
Wampler
, and
D. G.
Whyte
,
Nuclear Fusion
41
,
1967
(
2001
).
8.
V. K.
Alimov
,
B.
Tyburska-Püschel
,
S.
Lindig
,
Y.
Hatano
,
M.
Balden
,
J.
Roth
,
K.
Isobe
,
M.
Matsuyama
, and
T.
Yamanishi
,
Journal of Nuclear Materials
420
,
519
(
2012
).
9.
B.
Tyburska-Püschel
,
V. K.
Alimov
,
O. V.
Ogorodnikova
,
K.
Schmid
, and
K.
Ertl
,
Journal of Nuclear Materials
395
,
150
(
2009
).
10.
V. S.
Voitsenya
,
M.
Balden
,
A. I.
Belyaeva
,
V. K.
Alimov
,
B.
Tyburska-Püschel
,
A. A.
Galuza
,
A. A.
Kasilov
,
I. V.
Kolenov
,
V. G.
Konovalov
,
O. A.
Skoryk
, and
S. I.
Solodovchenko
,
Journal of Nuclear Materials
434
,
375
(
2013
).
11.
A. I.
Belyaeva
,
A. A.
Galuza
,
I. V.
Kolenov
,
V. G.
Konovalov
,
A. A.
Savchenko
, and
O. A.
Skoryk
,
The Physics of Metals and Metallography
114
,
703
(
2013
).
12.
E. D.
Palik
,
Handbook of Optical Constants of Solids
(
Academic Press, Inc.
,
New York
,
1985
).
13.
A. I.
Belyaeva
,
A. A.
Galuza
,
A.
Savchenko
, and
K. A.
Slatin
,
Bulletin of the Russian Academy of Science. Physics.
75
,
721
(
2011
).
14.
A. I.
Belyaeva
,
A. A.
Galuza
,
I. V.
Kolenov
,
A. A.
Savchenko
,
S. N.
Faizova
,
G. N.
Raab
, and
D. A.
Aksenov
,
Bulletin of the Russian Academy of Science. Physics.
76
,
854
(
2012
).
15.
A. I.
Belyaeva
,
A. A.
Galuza
, and
A. D.
Kudlenko
,
Pribory i Technika Experimenta
6
,
135
(
2008
), (in Russian).
16.
M.
Balden
,
A. F.
Bardamid
,
A. I.
Belyaeva
,
K. A.
Slatin
,
J. W.
Davis
,
A. A.
Haasz
,
M.
Poon
,
V. G.
Konovalov
,
I. V.
Ryzhkov
, and
A. N.
Shapoval
,
Journal of Nuclear Materials
329–333
,
1515
(
2004
).