- Email: [email protected]
Deuterium-induced nanostructure formation on tungsten exposed to high-flux plasma
Journal of Nuclear Materials xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Deuterium-induced nanostructure formation on tungsten exposed to high-flux plasma H.Y. Xu a,b,⇑, G. De Temmerman c,d, G.-N. Luo e, Y.Z. Jia a, Y. Yuan a, B.Q. Fu a, A. Godfrey a, W. Liu a,⇑,1 a
Key Laboratory of Advanced Materials, Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, China Science and Technology on Surface Physics and Chemistry Laboratory, Mianyang, Sichuan 621907, China c FOM Institute DIFFER, Dutch Institute For Fundamental Energy Research, Ass. EURATOM-FOM, Trilateral Euregio Cluster, Postbus 1207, 3430BE Nieuwegein, The Netherlands d ITER Organization, Route de Vinon-sur-Verdon CS 90046-13067, St Paul Lez Durance Cedex, France e Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, Anhui 230031, China b
a r t i c l e
i n f o
Article history: Available online xxxx
a b s t r a c t Surface topography of polycrystalline tungsten (W) have been examined after exposure to a low-energy (38 eV/D), high-flux (1.1–1.5 1024 m2 s1) deuterium plasma in the Pilot-PSI linear plasma device. The methods used were scanning electron microscopy (SEM), transmission electron microscopy (TEM), positron annihilation Doppler broadening (PADB) and grazing incident X-ray diffraction (GI-XRD). After exposure to high flux D plasma, blisters and nanostructures are formed on the W surface. Generation of defects was evidenced by PADB, while high stress and mixture of phases were detected in depth of 50 nm by GI-XRD. TEM observation revealed fluctuations and disordered microstructure on the outmost surface layer. Based on these results, surface reconstruction is considered as a possible mechanism for the formation of defects and nanostructures. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Tungsten (W) is currently the plasma-facing material for the ITER divertor and a candidate for future fusion devices owing to its good thermal properties and low erosion by incoming plasma particles. Blister formation has been widely observed on W surfaces for bombardment by deuterium (D) plasma with flux lower than 1023 m2 s1, and depends on ion fluence, energy and surface temperature [1–6]. Recently, the surface modification of W exposed to high flux D plasma (1024 m2 s1) similar to that expected in the ITER divertor region during operations was investigated by the authors [7]. Much smaller sized blisters are observed on tungsten surface with insisting appearance at elevated temperature (up to 873 K). A new type of surface nanostructure was also identified on the surface under those conditions suggesting that the ion flux plays an important role on the enhancement of the surface modification. The importance of ion flux was already identified in case of helium ion irradiation, where the development of ‘‘nano-fuzz’’ depends on the incoming He ion flux, and that for fluxes higher than 7 1021 m2 s1, it only depends on the expo⇑ Corresponding authors at: Science and Technology on Surface Physics and Chemistry Laboratory, Mianyang, Sichuan 621907, China (H.Y. Xu). E-mail addresses: [email protected] (H.Y. Xu), [email protected] (W. Liu). 1 Presenting author.
sure time [8]. With that in mind, the significance of such morphology changes for the behavior of tungsten under fusion-relevant conditions is far from clear, in particular in the divertor region where the particle flux is the highest (>1024 m2 s1). In this paper, efforts are made to explore the mechanism of surface modification by high flux D plasma exposure. The surface morphology is observed by high resolution scanning microscope (SEM). Positron Annihilation Doppler broadening (PADB) is used to monitor the defect evolution by plasma exposure. Grazing incident Xray diffraction (GI-XRD) and transmission electron microscopy (TEM) are applied to analyze the microstructure of exposed surface.
2. Experimental procedures The preparation of the tungsten samples used in this study has been described in more details in [9]. Briefly, polycrystalline tungsten samples with diameter of 30 mm and thickness of 3 mm were cut from a rolled sheet and electro-polished, and then outgassed at 1273 K for 1 h at a background pressure of 5 104 Pa before implantation. The purity of the samples is 99.95 wt% with main impurities Mo, Fe, C around 50 ppm, O about 2500 ppm. The plasma exposure is performed in the Pilot-PSI linear plasma device, which is uniquely capable of producing low energy plasmas with a high flux ranging from 1023–1025 D m2 s1 and with energy
http://dx.doi.org/10.1016/j.jnucmat.2014.11.039 0022-3115/Ó 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039
2
H.Y. Xu et al. / Journal of Nuclear Materials xxx (2014) xxx–xxx
fluxes up to 50 MW m2 [10]. The ion energy is fixed at 38 eV by negatively biasing the target. The electron density and temperature are measured by Thomson scattering, and are used to calculate the ion flux. The plasma beam with a Gaussian profile is obtained with the peak flux of about 1–2 1024 D m2 s1 in the center and a beam diameter of 1 cm. Several consecutive shots are carried out to reach the required fluence, with effective heating-up and cooling down of 1–2 s. The surface temperature is determined by the heat flux of the plasma and active cooling of the target holder, and is monitored by a fast infrared camera (FLIR SC7500-MB) during experiment. Surface morphology is observed by secondary electron (SE) mode and inlens mode [11] in a high resolution SEM. PADB was used to investigate the formation of defects in the crystal lattice caused by plasma exposure, the details of the set-up of experiments can be found in [7]. GI-XRD and FIB–TEM are carried out to characterize the surface structure. The samples are analyzed with 50 KV Cu Ka XRD analysis, the step width is 0.04°, and time per step is 7.0 s. 3. Results and discussions Fig. 1 shows a typical view of the sample surface after exposure in Pilot-PSI to a plasma fluence of 7 1026 D m2 at a surface temperature of 423 K. In Fig. 1(a), blisters with irregular shapes and relatively small size are observed widespread on the exposed surfaces. The most popular blister size is measured to be 0.3–0.4 lm, while the maximum diameter dmax is found to be lower than 2 lm. Fig. 1(b) shows a high magnification image of the highlighted rectangle in Fig. 1(a). Some nano-structures are present homogeneously on the surface, even on the top of blisters, but appear in different shapes within different cells/grains. In total, three types of nanostructures are found on the surface, which can be described as triangles, ripples and sponge-like structures. Detailed information of the evolution of nanostructures with fluence and surface
temperature and its dependence on grain orientations has been reported in previous investigations, which can be found in [7,12]. To obtain insights on the effect of high flux deuterium plasma exposure, the sample was then analyzed by PADB with the Variable Energy Positron (VEP) beam at Delft Technology University, as has been reported in [7]. Fig. 2 shows the measured S parameter as a function of positron annihilation depth in Fig. 2(a) and S vs. W data in Fig. 2(b). The Doppler broadening spectrum of a virgin rolled-W sample is also included for comparison (Fig. 2, ). Taking the S parameter of 0.46 from annealed ‘defect-free’ tungsten as the reference bulk S parameter, the S parameter of the rolled-W sample is increased by 4.3% (Fig. 2(a)), suggesting that the non-exposed material readily contains defects such as dislocations and small vacancies, which is typical for a rolled polycrystalline material. After analysis by VEPFIT program [13], a positron diffusion length of 20 nm is obtained, which is much smaller than the value of 100 nm for the defect-free tungsten. This also evidences the presence of defects in rolled-W before plasma exposure. After plasma exposure, a significant increase of S profile in tungsten (Fig. 3, ) is observed. The increase of the S parameters is found to be 10% above the defect free reference value (indicated by the red dash line in Fig. 2(a)). The affected area by D plasma is extended to a depth of 250 nm, and the significantly affected depth is 100 nm, as depicted in Fig. 2(b). In general, for a positron trapped in an open volume defect (such as a dislocation, a vacancy or vacancy cluster or voids) the probability of annihilations with core electrons is reduced compared with that for valence electrons resulting in a higher S parameter [14]. Therefore, the generation of defects, mainly dislocations, vacancy/vacancy clusters by high flux D plasma is suggested by the PADB measurements. And the narrow distribution of created defects can be related to the low mobility of defects and D-defect complexes at 423 K. The question rises as how could defects be created by D ions with energy well below the displacement threshold. Since several hypotheses have been proposed based on the high stress caused
Fig. 1. Surface morphology of tungsten after exposure to high flux D plasma with fluence of 7 1026 D m2 at 423 K, (a) blisters shown by SE image and (b) nano-structures shown by inlens image.
Fig. 2. PADB results of W before and after exposure to high flux D plasma with fluence of 7 1026 D m2 at 423 K, with S parameter versus depth in (a), and SW values in (b).
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039
H.Y. Xu et al. / Journal of Nuclear Materials xxx (2014) xxx–xxx
Fig. 3. GI-XRD results of W after exposure to high flux D plasma with fluence of 7 1026 D m2 at 473 K, parts of the continuous scan from 20° to 140° with a grazing incident angle of 2°, with blue lines denote characteristic spectrum of alpha tungsten, and red lines denote those of beta tungsten. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
by supersaturated D in the surface layer, the GI-XRD measurement is thus applied to detect the surface structure. Results are shown in Fig. 3. Three locations on the exposed surface are continuously scanned from 20° to 140°, one spot at the center of the plasma beam, and two spots 8 mm away from the center. The grazing angle is 2°, corresponding to an information depth of 50 nm. Reference spectrum is collected with incident X-ray beam. A broadening of the diffraction peaks is observed for the center spot compared to the reference spectrum, while no changes are found for the two spots 8 mm away from the center. This is in line with the Gaussian profile of plasma flux. The broadening of peaks is generally related to an increased stress level in the lattice, in case of no changes in crystal size, which is apparently caused by the high flux D plasma exposure. Since the plasma flux in this study is several orders of magnitude higher than in previous studies [1–4,8,15], it is reasonable to expect that much higher stresses are established by higher transient D concentration during plasma exposure. In addition, two different peaks appear in the diffraction pattern at the center spots. The two peaks are found at around 36.15° and 63.79°, corresponding to peaks from (2 0 0) and (2 2 2) of b-W, respectively. This leads to the issue of a partial phase transition on the surface layer caused by high stress in the surface layer, namely the surface reconstruction. Detailed discussions will be given as following.
3
To answer the question on the relationship of the defects and nanostructures, cross-section observation of the exposed sample was carried out by focus ion beam assisted TEM. A W foil with thickness less than 100 nm is cut by FIB with a Pt–C protection layer on the surface, and the microstructures are shown in Fig. 4. In Fig. 4(a), typical deformed microstructures are observed in the bulk, corresponding to the initial rolling process. The inset of Fig. 4(a) shows a selected area diffraction (SAD) image corresponding to the bcc lattice of tungsten, and no measurable impurities are found in that surface layer. A layer with fine fluctuations appears in the outmost surface, with an apparent thickness at around 30 nm. Different contrast can be distinguished from the fluctuations, indicating that the fluctuations may be stacked by several structures. High resolution observation of the surface layer shows the structures in the fluctuations are in a disordered state in Fig. 4(b). The inserted picture presenting the fast Fourier transform (FFT) pattern of the highlighted area in Fig. 4(b), reveals haloes and a faint diffraction patters, with no diffraction points from single crystal. The haloes may be caused by diffraction of the X-rays from the defects and the presence of mixture phases, in line with the results from PADB and GI-XRD measurements. It should be noted that besides the fluctuations, no observable nano-voids are found in the outmost surface and bulk material, while vacancy/vacancies are beyond the detection limit of TEM. This is inconsistent with our previous observations of small bumps and voids (<30 nm in diameter) in the top surface on cross-sections (by SEM) of tungsten samples exposed under similar conditions [7]. The first possible reason is that most of the nanostructures are solid mounds, while only very few will evolve into nano-voids, as that shown in Fig. 2 in Ref. [7]. Interestingly, no bubbles are reported close to the surface in tungsten exposed to low energy hydrogen isotope plasma, while they are extensively found on tungsten surface exposed to low energy helium plasma [16]. Therefore, no bubbles are expected on tungsten surface by deuterium plasma exposure. Secondly, the thin tungsten foil prepared by FIB might tend to recover to release the high stress on the surface layer during the cutting, also leading to no observable nano-voids. For the latter possibility, observations on the surface plane by FIB–TEM is planned to obtain more details of the nanostructures. Nevertheless, the TEM observations reveal the existence of an abnormal surface layer with defects and mixture phases, and fluctuations with size at 10 nm, which are considered corresponding to the nanostructures. Based on the above results, a possible effect of surface interaction is taken into account, which is proposed for tungsten surface covered by adsorbed gases or ultrathin metal films to relieve the
Fig. 4. TEM results of W after exposure to high flux D plasma with fluence of 7 1026 D m2 at 423 K, (a) is bright field TEM image and (b) is high resolution photograph of cross section at the outmost surface, with inserted pictures of SAD pattern and FFT pattern of the highlighted area, respectively. Pt–C layer was deposited on the surface for protection.
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039
4
H.Y. Xu et al. / Journal of Nuclear Materials xxx (2014) xxx–xxx
high stresses, named as ‘‘strain-relief mechanism’’ [17] or metal atoms displacements [18]. Coincidently, by such mechanism(s), three-fold symmetry triangular terraces were formed on W (1 1 1) crystal [17,18]. Such features are very similar to the nanostructures of triangular shape found in grains with surface normal direction near h1 1 1i in the present conditions. More importantly, the proposed model includes the formation of atomic-sized vacancies in each reconstructed unit cell in the first surface layers. Such vacancies play a significant role in surface reconstruction and a dominant configuration of surface structure with more defects is considered to be energetically favourable. This leads to the possibility that the defects, together with the nanostructures, in the outmost surface are generated by mechanism of surface reconstruction. As expected during high flux D plasma exposure, a very high D concentration can be built up, which produces high stresses in the surface layer. The displacement of metal atoms then occurs on the top surface to relieve the high stresses, which is referred to as surface reconstruction. 4. Summary Pronounced surface modification including blister and nanostructure occurs on the top surface of tungsten samples exposed to ITER-relevant D plasma fluxes. Investigations to explore the mechanisms were carried out by PADB, GI-XRD and FIB–TEM. Defects are created by high flux D plasma, as found by PADB. The generation of defects is significantly in the surface layer, while high stress are detected in the depth less than 50 nm by GI-XRD. An abnormal surface layer with fluctuations and disordered microstructure is evidenced by TEM observation, which consists with defects and mixture phases and is suggested to be corresponding to the surface nanostructures. Supersaturated deuterium induced surface reconstruction is considered as a possible mechanism for the generation of defects and formation of nanostructures, which needs to be further investigated.
Acknowledgements The authors would like to acknowledge the Pilot-PSI group for their technical support and great help. The authors further wish to thank Prof. X. Huang’s group of Risø DTU for their help in observation by HRSEM and their beneficial discussion. This work was supported by National Magnetic Confinement Fusion Science Program of China under Grant 2013GB109004 and the National Nature Science Foundation of China under contract No. 51071095. This work was partially financially supported by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO).
References [1] V.Kh. Alimov, B. Tyburska-Püschel, S. Lindig, et al., J. Nucl. Mater. 420 (2012) 519–524. [2] G.-N. Luo, W.M. Shu, M. Nishi, J. Nucl. Mater. 347 (2005) 111–117. [3] T. Funabiki, T. Shimada, Y. Ueda, et al., J. Nucl. Mater. 329–333 (2004) 780–784. [4] V.Kh. Alimov, W.M. Shu, J. Roth, et al., J. Nucl. Mater. 417 (2011) 572. [5] Y. Zayachuk, M.H.J. ’t Hoen, P.A. Zeijlmans van Emmichoven, et al., Nucl. Fusion 53 (2013) 013013. [6] L. Buzi, G. De Temmerman, B. Unterberg, et al., J. Nucl. Mater. 455 (1–3) (2014) 316–319. [7] H.Y. Xu, G.N. Luo, H. Schut, et al., J. Nucl. Mater. 447 (2014) 22. [8] M.J. Baldwin et al., J. Nucl. Mater. 390–391 (2009). [9] H.Y. Xu, W. Liu, G.N. Luo, et al., submitted for publication. [10] G. De Temmerman, J.J. Zielinski, S. van Diepen, et al., Nucl. Fusion 51 (2011) 073008. [11] K. Kumagai, T. Sekiguchi, Ultramicroscopy 109 (2009) 368. [12] H.Y. Xu, Y.B. Zhang, Y. Yuan, et al., J. Nucl. Mater. 443 (2013) 452. [13] A. Van Veen, H. Schut, J. De Vries, et al., in: AIP Conference Proceedings of 4th International workshop on: slow-positron beam techniques for solids and surfaces, vol. 218, 1991, pp. 171–198. [14] P.J. Schultz, K.G. Lynn, Rev. Mod. Phys. 60 (1988) 701. [15] O.V. Ogorodnikova, J. Roth, M. Mayer, J. Appl. Phys. 103 (2008) 034902. [16] K.O.E. Henriksson, K. Nordlund, A. Krasheninnikov, et al., Appl. Phys. Lett. 87 (2005) 163113. [17] C.H. Nien, T.E. Madey, Surf. Sci. 433–435 (1999) 254–260. [18] T.E. Felter, R.A. Barker, P.J. Estrup, Phys. Rev. Lett. 38 (20) (1977) 1138–1141.
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Deuterium-induced nanostructure formation on tungsten exposed to high-flux plasma H.Y. Xu a,b,⇑, G. De Temmerman c,d, G.-N. Luo e, Y.Z. Jia a, Y. Yuan a, B.Q. Fu a, A. Godfrey a, W. Liu a,⇑,1 a
Key Laboratory of Advanced Materials, Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, China Science and Technology on Surface Physics and Chemistry Laboratory, Mianyang, Sichuan 621907, China c FOM Institute DIFFER, Dutch Institute For Fundamental Energy Research, Ass. EURATOM-FOM, Trilateral Euregio Cluster, Postbus 1207, 3430BE Nieuwegein, The Netherlands d ITER Organization, Route de Vinon-sur-Verdon CS 90046-13067, St Paul Lez Durance Cedex, France e Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, Anhui 230031, China b
a r t i c l e
i n f o
Article history: Available online xxxx
a b s t r a c t Surface topography of polycrystalline tungsten (W) have been examined after exposure to a low-energy (38 eV/D), high-flux (1.1–1.5 1024 m2 s1) deuterium plasma in the Pilot-PSI linear plasma device. The methods used were scanning electron microscopy (SEM), transmission electron microscopy (TEM), positron annihilation Doppler broadening (PADB) and grazing incident X-ray diffraction (GI-XRD). After exposure to high flux D plasma, blisters and nanostructures are formed on the W surface. Generation of defects was evidenced by PADB, while high stress and mixture of phases were detected in depth of 50 nm by GI-XRD. TEM observation revealed fluctuations and disordered microstructure on the outmost surface layer. Based on these results, surface reconstruction is considered as a possible mechanism for the formation of defects and nanostructures. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Tungsten (W) is currently the plasma-facing material for the ITER divertor and a candidate for future fusion devices owing to its good thermal properties and low erosion by incoming plasma particles. Blister formation has been widely observed on W surfaces for bombardment by deuterium (D) plasma with flux lower than 1023 m2 s1, and depends on ion fluence, energy and surface temperature [1–6]. Recently, the surface modification of W exposed to high flux D plasma (1024 m2 s1) similar to that expected in the ITER divertor region during operations was investigated by the authors [7]. Much smaller sized blisters are observed on tungsten surface with insisting appearance at elevated temperature (up to 873 K). A new type of surface nanostructure was also identified on the surface under those conditions suggesting that the ion flux plays an important role on the enhancement of the surface modification. The importance of ion flux was already identified in case of helium ion irradiation, where the development of ‘‘nano-fuzz’’ depends on the incoming He ion flux, and that for fluxes higher than 7 1021 m2 s1, it only depends on the expo⇑ Corresponding authors at: Science and Technology on Surface Physics and Chemistry Laboratory, Mianyang, Sichuan 621907, China (H.Y. Xu). E-mail addresses: [email protected] (H.Y. Xu), [email protected] (W. Liu). 1 Presenting author.
sure time [8]. With that in mind, the significance of such morphology changes for the behavior of tungsten under fusion-relevant conditions is far from clear, in particular in the divertor region where the particle flux is the highest (>1024 m2 s1). In this paper, efforts are made to explore the mechanism of surface modification by high flux D plasma exposure. The surface morphology is observed by high resolution scanning microscope (SEM). Positron Annihilation Doppler broadening (PADB) is used to monitor the defect evolution by plasma exposure. Grazing incident Xray diffraction (GI-XRD) and transmission electron microscopy (TEM) are applied to analyze the microstructure of exposed surface.
2. Experimental procedures The preparation of the tungsten samples used in this study has been described in more details in [9]. Briefly, polycrystalline tungsten samples with diameter of 30 mm and thickness of 3 mm were cut from a rolled sheet and electro-polished, and then outgassed at 1273 K for 1 h at a background pressure of 5 104 Pa before implantation. The purity of the samples is 99.95 wt% with main impurities Mo, Fe, C around 50 ppm, O about 2500 ppm. The plasma exposure is performed in the Pilot-PSI linear plasma device, which is uniquely capable of producing low energy plasmas with a high flux ranging from 1023–1025 D m2 s1 and with energy
http://dx.doi.org/10.1016/j.jnucmat.2014.11.039 0022-3115/Ó 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039
2
H.Y. Xu et al. / Journal of Nuclear Materials xxx (2014) xxx–xxx
fluxes up to 50 MW m2 [10]. The ion energy is fixed at 38 eV by negatively biasing the target. The electron density and temperature are measured by Thomson scattering, and are used to calculate the ion flux. The plasma beam with a Gaussian profile is obtained with the peak flux of about 1–2 1024 D m2 s1 in the center and a beam diameter of 1 cm. Several consecutive shots are carried out to reach the required fluence, with effective heating-up and cooling down of 1–2 s. The surface temperature is determined by the heat flux of the plasma and active cooling of the target holder, and is monitored by a fast infrared camera (FLIR SC7500-MB) during experiment. Surface morphology is observed by secondary electron (SE) mode and inlens mode [11] in a high resolution SEM. PADB was used to investigate the formation of defects in the crystal lattice caused by plasma exposure, the details of the set-up of experiments can be found in [7]. GI-XRD and FIB–TEM are carried out to characterize the surface structure. The samples are analyzed with 50 KV Cu Ka XRD analysis, the step width is 0.04°, and time per step is 7.0 s. 3. Results and discussions Fig. 1 shows a typical view of the sample surface after exposure in Pilot-PSI to a plasma fluence of 7 1026 D m2 at a surface temperature of 423 K. In Fig. 1(a), blisters with irregular shapes and relatively small size are observed widespread on the exposed surfaces. The most popular blister size is measured to be 0.3–0.4 lm, while the maximum diameter dmax is found to be lower than 2 lm. Fig. 1(b) shows a high magnification image of the highlighted rectangle in Fig. 1(a). Some nano-structures are present homogeneously on the surface, even on the top of blisters, but appear in different shapes within different cells/grains. In total, three types of nanostructures are found on the surface, which can be described as triangles, ripples and sponge-like structures. Detailed information of the evolution of nanostructures with fluence and surface
temperature and its dependence on grain orientations has been reported in previous investigations, which can be found in [7,12]. To obtain insights on the effect of high flux deuterium plasma exposure, the sample was then analyzed by PADB with the Variable Energy Positron (VEP) beam at Delft Technology University, as has been reported in [7]. Fig. 2 shows the measured S parameter as a function of positron annihilation depth in Fig. 2(a) and S vs. W data in Fig. 2(b). The Doppler broadening spectrum of a virgin rolled-W sample is also included for comparison (Fig. 2, ). Taking the S parameter of 0.46 from annealed ‘defect-free’ tungsten as the reference bulk S parameter, the S parameter of the rolled-W sample is increased by 4.3% (Fig. 2(a)), suggesting that the non-exposed material readily contains defects such as dislocations and small vacancies, which is typical for a rolled polycrystalline material. After analysis by VEPFIT program [13], a positron diffusion length of 20 nm is obtained, which is much smaller than the value of 100 nm for the defect-free tungsten. This also evidences the presence of defects in rolled-W before plasma exposure. After plasma exposure, a significant increase of S profile in tungsten (Fig. 3, ) is observed. The increase of the S parameters is found to be 10% above the defect free reference value (indicated by the red dash line in Fig. 2(a)). The affected area by D plasma is extended to a depth of 250 nm, and the significantly affected depth is 100 nm, as depicted in Fig. 2(b). In general, for a positron trapped in an open volume defect (such as a dislocation, a vacancy or vacancy cluster or voids) the probability of annihilations with core electrons is reduced compared with that for valence electrons resulting in a higher S parameter [14]. Therefore, the generation of defects, mainly dislocations, vacancy/vacancy clusters by high flux D plasma is suggested by the PADB measurements. And the narrow distribution of created defects can be related to the low mobility of defects and D-defect complexes at 423 K. The question rises as how could defects be created by D ions with energy well below the displacement threshold. Since several hypotheses have been proposed based on the high stress caused
Fig. 1. Surface morphology of tungsten after exposure to high flux D plasma with fluence of 7 1026 D m2 at 423 K, (a) blisters shown by SE image and (b) nano-structures shown by inlens image.
Fig. 2. PADB results of W before and after exposure to high flux D plasma with fluence of 7 1026 D m2 at 423 K, with S parameter versus depth in (a), and SW values in (b).
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039
H.Y. Xu et al. / Journal of Nuclear Materials xxx (2014) xxx–xxx
Fig. 3. GI-XRD results of W after exposure to high flux D plasma with fluence of 7 1026 D m2 at 473 K, parts of the continuous scan from 20° to 140° with a grazing incident angle of 2°, with blue lines denote characteristic spectrum of alpha tungsten, and red lines denote those of beta tungsten. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
by supersaturated D in the surface layer, the GI-XRD measurement is thus applied to detect the surface structure. Results are shown in Fig. 3. Three locations on the exposed surface are continuously scanned from 20° to 140°, one spot at the center of the plasma beam, and two spots 8 mm away from the center. The grazing angle is 2°, corresponding to an information depth of 50 nm. Reference spectrum is collected with incident X-ray beam. A broadening of the diffraction peaks is observed for the center spot compared to the reference spectrum, while no changes are found for the two spots 8 mm away from the center. This is in line with the Gaussian profile of plasma flux. The broadening of peaks is generally related to an increased stress level in the lattice, in case of no changes in crystal size, which is apparently caused by the high flux D plasma exposure. Since the plasma flux in this study is several orders of magnitude higher than in previous studies [1–4,8,15], it is reasonable to expect that much higher stresses are established by higher transient D concentration during plasma exposure. In addition, two different peaks appear in the diffraction pattern at the center spots. The two peaks are found at around 36.15° and 63.79°, corresponding to peaks from (2 0 0) and (2 2 2) of b-W, respectively. This leads to the issue of a partial phase transition on the surface layer caused by high stress in the surface layer, namely the surface reconstruction. Detailed discussions will be given as following.
3
To answer the question on the relationship of the defects and nanostructures, cross-section observation of the exposed sample was carried out by focus ion beam assisted TEM. A W foil with thickness less than 100 nm is cut by FIB with a Pt–C protection layer on the surface, and the microstructures are shown in Fig. 4. In Fig. 4(a), typical deformed microstructures are observed in the bulk, corresponding to the initial rolling process. The inset of Fig. 4(a) shows a selected area diffraction (SAD) image corresponding to the bcc lattice of tungsten, and no measurable impurities are found in that surface layer. A layer with fine fluctuations appears in the outmost surface, with an apparent thickness at around 30 nm. Different contrast can be distinguished from the fluctuations, indicating that the fluctuations may be stacked by several structures. High resolution observation of the surface layer shows the structures in the fluctuations are in a disordered state in Fig. 4(b). The inserted picture presenting the fast Fourier transform (FFT) pattern of the highlighted area in Fig. 4(b), reveals haloes and a faint diffraction patters, with no diffraction points from single crystal. The haloes may be caused by diffraction of the X-rays from the defects and the presence of mixture phases, in line with the results from PADB and GI-XRD measurements. It should be noted that besides the fluctuations, no observable nano-voids are found in the outmost surface and bulk material, while vacancy/vacancies are beyond the detection limit of TEM. This is inconsistent with our previous observations of small bumps and voids (<30 nm in diameter) in the top surface on cross-sections (by SEM) of tungsten samples exposed under similar conditions [7]. The first possible reason is that most of the nanostructures are solid mounds, while only very few will evolve into nano-voids, as that shown in Fig. 2 in Ref. [7]. Interestingly, no bubbles are reported close to the surface in tungsten exposed to low energy hydrogen isotope plasma, while they are extensively found on tungsten surface exposed to low energy helium plasma [16]. Therefore, no bubbles are expected on tungsten surface by deuterium plasma exposure. Secondly, the thin tungsten foil prepared by FIB might tend to recover to release the high stress on the surface layer during the cutting, also leading to no observable nano-voids. For the latter possibility, observations on the surface plane by FIB–TEM is planned to obtain more details of the nanostructures. Nevertheless, the TEM observations reveal the existence of an abnormal surface layer with defects and mixture phases, and fluctuations with size at 10 nm, which are considered corresponding to the nanostructures. Based on the above results, a possible effect of surface interaction is taken into account, which is proposed for tungsten surface covered by adsorbed gases or ultrathin metal films to relieve the
Fig. 4. TEM results of W after exposure to high flux D plasma with fluence of 7 1026 D m2 at 423 K, (a) is bright field TEM image and (b) is high resolution photograph of cross section at the outmost surface, with inserted pictures of SAD pattern and FFT pattern of the highlighted area, respectively. Pt–C layer was deposited on the surface for protection.
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039
4
H.Y. Xu et al. / Journal of Nuclear Materials xxx (2014) xxx–xxx
high stresses, named as ‘‘strain-relief mechanism’’ [17] or metal atoms displacements [18]. Coincidently, by such mechanism(s), three-fold symmetry triangular terraces were formed on W (1 1 1) crystal [17,18]. Such features are very similar to the nanostructures of triangular shape found in grains with surface normal direction near h1 1 1i in the present conditions. More importantly, the proposed model includes the formation of atomic-sized vacancies in each reconstructed unit cell in the first surface layers. Such vacancies play a significant role in surface reconstruction and a dominant configuration of surface structure with more defects is considered to be energetically favourable. This leads to the possibility that the defects, together with the nanostructures, in the outmost surface are generated by mechanism of surface reconstruction. As expected during high flux D plasma exposure, a very high D concentration can be built up, which produces high stresses in the surface layer. The displacement of metal atoms then occurs on the top surface to relieve the high stresses, which is referred to as surface reconstruction. 4. Summary Pronounced surface modification including blister and nanostructure occurs on the top surface of tungsten samples exposed to ITER-relevant D plasma fluxes. Investigations to explore the mechanisms were carried out by PADB, GI-XRD and FIB–TEM. Defects are created by high flux D plasma, as found by PADB. The generation of defects is significantly in the surface layer, while high stress are detected in the depth less than 50 nm by GI-XRD. An abnormal surface layer with fluctuations and disordered microstructure is evidenced by TEM observation, which consists with defects and mixture phases and is suggested to be corresponding to the surface nanostructures. Supersaturated deuterium induced surface reconstruction is considered as a possible mechanism for the generation of defects and formation of nanostructures, which needs to be further investigated.
Acknowledgements The authors would like to acknowledge the Pilot-PSI group for their technical support and great help. The authors further wish to thank Prof. X. Huang’s group of Risø DTU for their help in observation by HRSEM and their beneficial discussion. This work was supported by National Magnetic Confinement Fusion Science Program of China under Grant 2013GB109004 and the National Nature Science Foundation of China under contract No. 51071095. This work was partially financially supported by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO).
References [1] V.Kh. Alimov, B. Tyburska-Püschel, S. Lindig, et al., J. Nucl. Mater. 420 (2012) 519–524. [2] G.-N. Luo, W.M. Shu, M. Nishi, J. Nucl. Mater. 347 (2005) 111–117. [3] T. Funabiki, T. Shimada, Y. Ueda, et al., J. Nucl. Mater. 329–333 (2004) 780–784. [4] V.Kh. Alimov, W.M. Shu, J. Roth, et al., J. Nucl. Mater. 417 (2011) 572. [5] Y. Zayachuk, M.H.J. ’t Hoen, P.A. Zeijlmans van Emmichoven, et al., Nucl. Fusion 53 (2013) 013013. [6] L. Buzi, G. De Temmerman, B. Unterberg, et al., J. Nucl. Mater. 455 (1–3) (2014) 316–319. [7] H.Y. Xu, G.N. Luo, H. Schut, et al., J. Nucl. Mater. 447 (2014) 22. [8] M.J. Baldwin et al., J. Nucl. Mater. 390–391 (2009). [9] H.Y. Xu, W. Liu, G.N. Luo, et al., submitted for publication. [10] G. De Temmerman, J.J. Zielinski, S. van Diepen, et al., Nucl. Fusion 51 (2011) 073008. [11] K. Kumagai, T. Sekiguchi, Ultramicroscopy 109 (2009) 368. [12] H.Y. Xu, Y.B. Zhang, Y. Yuan, et al., J. Nucl. Mater. 443 (2013) 452. [13] A. Van Veen, H. Schut, J. De Vries, et al., in: AIP Conference Proceedings of 4th International workshop on: slow-positron beam techniques for solids and surfaces, vol. 218, 1991, pp. 171–198. [14] P.J. Schultz, K.G. Lynn, Rev. Mod. Phys. 60 (1988) 701. [15] O.V. Ogorodnikova, J. Roth, M. Mayer, J. Appl. Phys. 103 (2008) 034902. [16] K.O.E. Henriksson, K. Nordlund, A. Krasheninnikov, et al., Appl. Phys. Lett. 87 (2005) 163113. [17] C.H. Nien, T.E. Madey, Surf. Sci. 433–435 (1999) 254–260. [18] T.E. Felter, R.A. Barker, P.J. Estrup, Phys. Rev. Lett. 38 (20) (1977) 1138–1141.
Please cite this article in press as: H.Y. Xu et al., J. Nucl. Mater. (2014), http://dx.doi.org/10.1016/j.jnucmat.2014.11.039