1 Introduction

Tungsten has emerged to be one of the most important materials for fusion reactors. Its high melting point, high thermal conductivity and low tritium affinity being its most attractive features. The International Thermonuclear Experimental Reactor (ITER) tokamak, where sustained high heat flux is received at the divertor, will use tungsten as the plasma-facing or armour material. Recently it has also been proposed to replace the beryllium tiles of the First Wall (FW) by tungsten[1,2] as disruption events can damage Be-FW very significantly as shown by JET.[3] In spite of a long history of work on tungsten, its behaviour when exposed to reactor conditions remains ill understood. A part of the problem lies in the fact that there would not be any facilities, other than the reactor itself, where the problem can be studied in-depth. The second part has to do with the complexity of the formation, evolution and impact of the defects on the very material properties that drove the choice of tungsten. The neutron exposure creates defect in tungsten which can affect its thermal and mechanical properties. Furthermore, the armour material is also subjected to bombardment of helium ions and hydrogen isotopes as well. Since the defects can act as trapping centers for hydrogen isotopes and helium[4,5,6,7,8] there is a need to understand the structure of defects. Since there exist no fusion relevant 14 MeV neutron sources with the desired flux at present,[9] one needs to adopt ion-irradiation experiments to create the dpa similar to that generated by the neutron-induced damage. The primary knock-on atom (PKA) energy for high energy neutron produces displacement damage in tungsten lattice. The heavy ion irradiation techniques are being adopted to introduce defects by displacement damage in tungsten with faster rate than that of by neutrons.[10,11,12] Ion irradiation techniques can be used to create defects by cascade displacement without any transmutation process unlike neutron irradiation.[13,14,15,16] A systematic study of the dynamics of defect production due to high energy collision cascade in tungsten has been carried out by simulations and experiments over a decade.[12,13,17,18,19,20,21] Different kinds of defect structures such as vacancy and interstitial clusters, dislocation loops and voids were noticed in the cascade region.[11,22,23,24,25,26,27] These defects are responsible for the degradation of their mechanical properties as well as enhancing the H-isotope trapping.

In the previous works we have shown that dislocation lines, loops and defect clusters are formed in both heavy and light ion-irradiated tungsten foils.[18,28,29] In heavy-ion irradiated tungsten, the positron spectroscopy analysis showed the formation of significant vacancy clusters, in fact, the saturation of defects in positron analysis.[29,30,31] This indicates that the defects created after irradiation were far beyond the saturation limit of the positron technique precluding further analysis of the defect clusters. In this paper, we show the direct evidence of the defect saturation after irradiation using bright-field (BF) transmission electron microscopy (TEM) analysis. The weak-beam dark-field (WBDF) analysis[32,33,34] confirms that the defect clusters are predominantly of vacancy type. The rest of the paper is organized as follows: the experimental details are given in Section II, the results are given in Section III, the analysis and discussions are given in Section IV and the conclusion are given in Section V.

2 Experimental Details

Cold rolled tungsten foils of 0.01 mm thickness with 99.96 wt pct purity procured from Princeton Sci. Corp. were used in the experiments. The samples of 8 \(\times \)8 mm size were fine polished with high napped soft cloth impregnated with diamond paste suspension of 0.05 \(\mu \)m diamond particles. The initial sample consists of elongated grains with a high density of dislocations (\(\ge 2.3 \times 10^{10}/\text{cm}^2\)) along the grain boundaries and within the grains and an average grain size of about 2 \(\mu \)m.[28] This is attributed to the plastic flow due to heavy deformation during the rolling. The samples were annealed at 1838 K for about 1 hour in a vacuum furnace at a base pressure of 0.1 Pa. The annealing was carried out in an inert environment of 10\(^4\) Pa Ar with 8 pct H\(_2\) gas. The hydrogen was added to prevent the surface oxidation during high temperature annealing.[28] The annealing temperature is well above the recrystallization threshold of tungsten (\(\sim \)1473 K). The annealing resulted in grain growth with grain size varied from 25 to 40 \(\mu \)m with a preferred orientation along (200) direction.[18] The dislocation density also reduced to 5.7 \(\times \) 10\(^8\)/cm\(^2\).

The recrystallized samples were irradiated with 80 MeV gold ions (Au\(^{7+}\)) for a fluence of 1.3 \(\times \) 10\(^{14}\) ions/cm\(^2\) using a 15 MV van de Graff generator (Pelletron) at Inter University Accelerator Center (IUAC) Delhi.[18,28] The beam current was 1.1.5 pnA (particle nano ampere = current/charge state) with a flux of 1\(\times \)10\(^{10}\) ions/cm\(^2\) for a total irradiation time of 3 hours. The beam was rastered uniformly over the sample surface throughout the irradiation. The sample temperature was about 300 K during the entire irradiation process. The total displacement damage estimated is about 0.22 by using a Kinchin-Pease damage model in Stopping Range of Ions in Matter (SRIM) software for a displacement energy of 90 eV, surface energy of 2 eV and a lattice energy of 3 eV.[18] The dpa profile is shown in Figure 1.

Fig. 1
figure 1

The dpa profile in tungsten due to 80 MeV Au ion-irradiation for a fluence of 1.3 \(\times \) 10\(^{14}\) ions/cm\(^2\). The locations where the TEM analysis was carried out is marked by vertical dashed lines

Full size image

Irradiated, and annealed samples were analyzed under transmission electron microscope for defects produced due to the ion-irradiation. The TEM images were taken at two different depths: (1) near-surface region within 200 nm from the irradiated side of the sample which is referred as ‘near-surface’ and (2) at 2 \(\mu \)m depth from the irradiated side which is referred as ‘bulk’ hereafter. This is to analyze two different dpa (vacancy) concentrations in the sample (see Figure 1).

2.1 TEM Characterization

For TEM analysis it is necessary to have 3 mm diameter and thickness typically about 200 nm or less. The sample preparation process involved various sequential thinning and polishing steps. The main process steps followed to get an electron transparent specimen for TEM investigation were: (1) Preparation of 3mm diameter discs from 100 \(\mu \)m thick foil by disc punching method, (2) grinding of the disc to achieve \(\sim \) 80 \(\mu \)m thickness and (3) low-angle ion-beam milling with dual Ar-ion gun mounted at 5 deg angle in vacuum (\(5.33 \times 10^{-3}\) Pa) for further thinning of the specimen to achieve the electron transparency.

A disc of 3 mm diameter was punched out from a 100 \(\mu \)m thick foil of the dimension of 8 \(\times \) 8 mm using a Gatan make disc punch (model 659). This 3 mm disc then, was ground, using a Gatan make disc grinder (model 623), on a fine lapping sheet of 5 \(\mu \)m sized abrasive particles and thickness of the disc was brought down to 80 \(\mu \)m. This ground disc was subjected to pre-thinning and polishing using a Gatan make dimple grinder (model 656). In this process a phosphor bronze wheel along with diamond paste and alumina suspension were used for achieving the necessary pre-thinning and polishing of the disc sample. After this process, the thickness of the sample was reduced to 20 \(\mu \)m at its central area while at the rim the thickness was maintained as 80 \(\mu \)m which is required to facilitate the handling of the sample. After the dimple grinding, the sample was subjected to ion milling process using Gatan make Precision Ion Polishing System (PIPS-II). In this process two ion-guns were used for polishing the sample by etching the material from the surface with argon ions of 5 keV energy at 4 deg incident angle. After the completion of the milling, the thickness of a very small region of the sample at its middle was reduced to less than 200 nm which was electron transparent. The ion milling was done at a beam energy of 5 keV to achieve thinning up to 60 to 200 nm in 3.5 to 5 hours.

3 Results

Figure 2 shows the bright-field TEM image of the recrystallized tungsten foil annealed at 1873 K. The analysis was carried out at the near-surface region. The microstructure shows a few dislocations on a 6 \(\times \) 6\(\,\mu \)m\(^{2}\) area with line length varied from 200 to 900 nm. The dislocation analyses of the samples were carried out using ImageJ software[35] and the dislocation density was estimated to be about 3.6 \(\times \) 10\(^{8}\) cm\(^{-2}\). The dislocation density was estimated using the line-intercept method where three lines of length (L) 10 \(\mu \)m were drawn along random directions on the image. The dislocation density (\(\rho \)) is calculated by taking the ratio of the number of intersection points of these lines with the dislocations, N, to the product of line length (L) and sample thickness (t), \(\rho = (N/Lt)\). From the requirements of electron transparency, an average sample thickness of 200 nm was considered for the calculations.

Fig. 2
figure 2

TEM image of the tungsten foil recrystallized at 1873 K

Full size image

The TEM image of the near-surface region of the tungsten foil irradiated with 80 MeV Au ions for a fluence of 1.3 \(\times \) 10\(^{14}\) /cm\(^2\) is shown in Figure 3. Dislocation lines are observed on the surface with the line length varied between 200 and 1100 nm which is larger compared to the annealed sample. The dislocation analysis shows a high amount of mixed-screw dislocations (60.3 pct) in comparison with edge dislocations (22.4 pct). The ion irradiation resulted in an increase in the mixed-screw dislocations (60.3 pct from 51.5 pct) and a reduction in the edge dislocation (22.4 pct from 27.3 pct). This might indicate the inter-conversion of edge type into mixed-screw types. The presence of long segments of dislocation lines could be due to the extension of mixed-screw type dislocation by cross slip to reduce its energy (since edge dislocation has higher energy) after the ion irradiation. The dislocation line density was about 6.1 \(\times \) 10\(^{8}\) cm\(^{-2}\) which is higher than that of annealed foil (3.6 \(\times \) 10\(^{8}\) cm\(^{-2}\)).

Fig. 3
figure 3

TEM images of tungsten foil after 80 MeV Au-irradiation for a fluence of 1.3 \(\times \) 10\(^{14}\) ions/cm\(^{2}\)

Full size image

The high resolution image of the near surface region is shown in Figure 5. The defect clusters are seen throughout along with long dislocation lines. The average defect cluster size is found to be about 8.3 nm (Fig. 4).

Fig. 4
figure 4

Defect clusters along with the dislocation lines observed in the near-surface region of the Au-irradiated tungsten

Full size image

The high resolution image of 2 \(\mu \)m bulk sample is shown in Figure 5. The defect clusters are shown in the black contrast. The clusters are uniformly distributed over the surface with a higher cluster density compared to the surface. The average cluster size is found to be about 5.9 nm. A summary of different defect types, their size and density is given in Table I.

Fig. 5
figure 5

Defect clusters observed at 2 \(\mu \)m depth of Au-irradiated W-foil for a fluence of 1.3 \(\times \) 10\(^{14}\) ions/cm\(^{2}\)

Full size image
Table I Different Defect Types Observed in TEM in the Annealed and 80 MeV Au-Irradiated Sample
Full size table

The cluster-size distribution observed at the near-surface region and at 2 \(\mu \)m depth in the irradiated tungsten is shown in Figure 6.

Fig. 6
figure 6

Cluster-size distribution observed at the near surface and at 2 \(\mu \)m bulk tungsten

Full size image

4 Discussion

In order to understand the nature of dislocation lines and defect clusters, detailed analysis has been carried out by superposing the selective area diffraction (SAED) pattern along with the stereographic projections. The SAED pattern of the recrystallized foil along with the crystallographic orientations is shown in Figure 7. The dislocations with Burgers vector \(\textbf{b} = 1/2~[-111]\) and 1/2 [111] are identified with the diffraction vector g[011] near zone axis [001] and is shown by the white arrows. The yellow arrows represent the stereographic projection of the foil and different planes are marked by circles around the bright spots. The SAED patterns of the recrystallized foil, irradiated foil both at near-surface region and at the bulk along with the two-beam diffraction condition is shown in Figure 8. Using this, the bright field images shown in Figures 2, 3, 4 and 5 were analyzed to understand the dislocation characteristics using the diffraction contrast conditions. The dislocations are observed when the reciprocal lattice vector \(\textbf{g}\) has an angle with the burgers vector \(\textbf{b}\), i.e., \(\textbf{g}.\textbf{b} \ne 0\). The dislocations were analyzed using the diffraction contrast condition for \(\textbf{g}\) [011] in the bright field imaging. The diffraction condition \(\textbf{g}.\textbf{b} \ne 0\) is satisfied for pure screw dislocation and \( \textbf{g}.(\textbf{b} x \textbf{u}) \ne 0\) is satisfied for pure edge dislocation where, \(\textbf{u}\) is unit vector of dislocation. The summary of the analysis is given in Table II. The samples show a significant fraction of mixed dislocations. The angle between the Burgers vector of the dislocation and the orientation of the dislocation line with respect to the stereographic projection is used to measure the dislocation angle. For example, screw 45 indicates 45 deg angle between the screw-type dislocation line oriented along (\(\bar{1}\)10) direction with the stereographic projection. The Burgers vector of these lines could be different, for example for screw 45, it could be either (0\(\bar{1}\)0) or (100). In the analysis we adopt a counting scheme where dislocations having angles between 40 and 50 deg are labelled 45 and those between 55 and 65 deg as 60. The total screw dislocation fraction is found to be much smaller in the near-surface region after irradiation when compared to the recrystallized case. This could be due to the radiation induced annealing, perhaps originating from the huge electronic loss due to high-energy high-mass ions.

Fig. 7
figure 7

Electron diffraction pattern of bcc tungsten superimposed with crystallographic poles showing crystallographic orientation of the foil (Color figure online)

Full size image
Fig. 8
figure 8

Selected area electron diffraction patterns of the annealed foil (a), near-surface region of the irradiated foil (b), 2 \(\mu \)m region of the irradiated foil (c), and the two-beam diffraction condition of the irradiated foil (d) (Color figure online)

Full size image
Table II Dislocation Fractions Observed in the Recrystallized and Au-Irradiated Foil
Full size table

Weak-beam images were analysed to understand the nature of the defect clusters observed in Figures 4 and 5. Figure 9 shows the weak-beam diffraction contrast in dark field mode which was reconstructed by (\(\bar{2}\)11) beam at 2 \(\mu \)m depth. The white contrast in the image shows the local strain field arising from the defect clusters which is typically measured in the weak-beam imaging.

Fig. 9
figure 9

Weak beam (\(\bar{2}\)11) diffraction contrast dark field image showing local strain field due to the defect clusters

Full size image

Figure 10 shows the weak-beam diffraction contrast in dark field mode which was reconstructed by a low-index plane (0\(\bar{1}\bar{1}\)) beam at 2 \(\mu \)m depth. The high-index plane (\(\bar{2}\)11) shows higher defect concentration in comparison to the low-index plane (0\(\bar{1}\bar{1}\)).

Fig. 10
figure 10

Weak beam (0-1-1) diffraction contrast dark field image showing local strain field due to the defect clusters

Full size image
Table III The Strain Obtained from the Weak-Beam Image Analysis
Full size table

The lattice strain observed at 2 \(\mu \)m depth is given in Table III. The strain is calculated from the d-spacing value of diffraction beam (\(\bar{2}\)11) using the ration of camera constant of the TEM (1.1 nm-mm) to distance between (\(\bar{-2}\)11) plane and centre (000). The data shows compressive strain which indicates the presence of vacancies. The analysis confirms the findings of positron spectroscopy measurements,[29] where the formation of vacancy clusters was observed throughout the sample in the heavy-ion irradiated tungsten.

5 Conclusions

The characterization the type of defects created in recrystallized tungsten foil due to 80 MeV gold ions using TEM under both bright and weak field imaging has been presented. The defects were characterized at two different depths, near-surface and 2 \(\mu \)m deep, to understand the defect profile as a function of the irradiation depth. The analysis shows that the heavy ion irradiation led to the formation of both, long dislocation lines up to 3 \(\mu \)m length and defect clusters of about 5 to 6 nm size. The dislocations lines are of edge, screw and mixed types. The mixed dislocations were found to increase during the irradiation at the expense of pure edge and screw dislocations. The dislocation line density did not change significantly with the depth but the size of the segments is found to increase. The defect clusters however, showed a significant increase with the depth. At the near-surface it was about 1.9 \(\times \) 10\(^{10}\) cm\(^{-2}\), increasing up to 6.9 \(\times \) 10\(^{12}\) cm\(^{-2}\) in the bulk with a reduction in the maximum cluster size.

The weak-beam imaging conclusively shows that the defect clusters are of mainly vacancy type both at the near-surface region and 2 \(\mu \)m depth. The local strain analysis shows a compressive strain which confirms the formation of vacancy clusters in heavy ion irradiation. The analysis further confirms the observations of positron spectroscopy measurements of vacancy clusters in heavy ion irradiation throughout the sample.