2026 Volume 21 Article ID: 1405047
Tungsten (W) is a key plasma facing material (PFM) for future fusion devices, requiring improved ductile-to-brittle transition temperature (DBTT) and recrystallization resistance for long-term operation. Various modification methods of W—including second-phase dispersion (potassium doping (K-doping) and lanthanum oxide (La2O3) particle dispersion) and solid-solution alloying (alloying with rhenium (Re) and tantalum (Ta))—were evaluated in the present study for expanding the operation temperature range of PFMs from DBTT to recrystallization temperature. Both second-phase dispersion and solid-solution alloying were found to expand the operation temperature range, offering greater flexibility in designing plasma facing components. Among the materials evaluated, K-doped W-1%Ta and W-1%Ta appeared particularly promising. Nevertheless, alloying with Ta may have concerns, including higher post-irradiation surface dose rate and poorer machinability. Thermal properties must also be considered for PFM, however solid-solution alloying reduces thermal diffusivity and conductivity, and large dispersed particles such as La2O3 can introduce anisotropy and may reduce thermal diffusivity and conductivity. Therefore, selecting suitable W-based materials requires carefully balancing mechanical property improvements with thermal property limitations to ensure dependable operation in future fusion devices.
Tungsten (W) is used as a plasma facing material (PFM) for the divertor of ITER [1]. In addition, W has been selected for the blanket first wall of ITER instead of beryllium [2]. Based on the intrinsic material properties and knowledge gained from ITER, it is highly likely that W-based materials will be applied to the divertor and blanket first wall of future fusion devices. Since these fusion devices need long-term operation, most components and materials are required to have a lifetime much longer than that of ITER. In addition, to broaden the design window—the range of operating conditions (e.g., temperature, heat loads, neutron irradiation and other service environments) under which materials maintain their integrity—it is necessary to expand the range of conditions (including temperature) in which microstructural changes are suppressed and key properties, such as mechanical properties, remain stable. Consequently, extensive efforts have been made to improve material properties of W.
Mechanical properties—particularly the ductile-to-brittle transition temperature (DBTT)—and high-temperature microstructural stability, especially recrystallization resistance, are among the most important factors, which may determine the lifetime and operation temperature range of W-based materials. Previous studies have shown that the DBTT and recrystallization temperature of W can be improved by various methods. Our series of studies demonstrated that potassium doping (K-doping) [3–5] and alloying with rhenium (Re) [3, 4, 6] and tantalum (Ta) [7, 8] were effective in simultaneously improving these properties.
K-doping is a common dispersion-strengthening method for W [9]. K-doped W contains nano-bubbles that include K. Because K-bubbles can hinder the motion of grain boundaries and dislocations, they lead to strengthening especially at high temperature and suppression of recrystallization [10, 11]. In addition, K-doping can produce finer grains compared to pure W because K-bubbles inhibit grain boundary migration. This grain refinement also leads to improvement of strength and toughness [11].
Solid-solution alloying is well known as another common method to improve mechanical properties and resistance to recrystallization of W [12, 13]. Representative elements that form solid-solutions with W include Re, molybdenum (Mo), vanadium (V), and Ta. When solid-solute elements substitute into the crystal lattice, lattice distortion occurs, inhibiting dislocation motion and grain boundary movement. This enhances mechanical properties and recrystallization resistance of W.
In our series of studies, the DBTT and recrystallization temperature were evaluated for the following seven types of materials to clarify the effects of each modification approach [3–8].
All materials were produced by A.L.M.T. Corp. using powder metallurgy, hot rolling, and subsequent stress-relief heat treatment. All manufacturing conditions (such as reduction ratio and heat treatment temperature) were identical, and the final plate thickness was also fixed at around 7 mm, except for the differences introduced by K-doping and/or alloying. In our series of studies, the concentration of K and Re was approximately 30 wt. ppm and 3 wt.%, respectively. The K content was determined based on manufacturing constraints (details cannot be disclosed). The Re content (3%) was selected based on the known effects of Re addition on thermo-mechanical properties, thermal conductivity, recrystallization resistance, and resistance to neutron irradiation [15]. DBTT was evaluated by Charpy impact test. The recrystallization temperature was evaluated through 1 h isochronal heat treatment followed by hardness measurement and grain structure observation. The extent of the modification effects varied, but all materials had a positive impact on DBTT and/or recrystallization temperature (Note that the Charpy impact test for W-5%Ta was not performed, and therefore its effect on DBTT remains unknown.) [3–8].
In the present study, DBTT and recrystallization temperature were evaluated using the same method for the following two types of materials.
The evaluation of lanthanum (La) oxides dispersed W (W-1%La2O3) aimed to clarify the effect of solid particle dispersion on DBTT and recrystallization temperature. Previous studies have published numerous similar approaches using oxides (e.g., La oxides [16] and yttrium (Y) oxides [17]) and carbides (e.g., titanium (Ti) carbides [18, 19] and Ta carbides [19]) to improve mechanical properties and microstructural stability. The present study specifically clarified the effect of La2O3 particle dispersion. The evaluation of K-doped W-1%Ta aimed to clarify the combined effect of K-doping and alloying with Ta. K-doping and alloying with Ta can influence the characteristics of K-bubbles, such as their size, number density, and spatial distribution. Consequently, differences may occur between K-doped W and K-doped W-1%Ta. In contrast, because Ta is fully solid-solved in W, the state of Ta should be identical in both W-1%Ta and K-doped W-1%Ta.
Based on these evaluation results, this paper clarifies mechanical-property-based modification methods for expanding the operation temperature range. This is expected to contribute to expanding the design window and available operating conditions for future fusion devices. Furthermore, thermal properties—particularly thermal conductivity—are the most critical factors for PFM. Therefore, this paper presents the thermal diffusivity of the aforementioned materials as a parameter used to estimate thermal conductivity and concludes with a discussion of the advantages and limitations from the perspective of thermal properties.
W-1%La2O3 and K-doped W-1%Ta plates were produced by A.L.M.T. Corp. Powder metallurgy (CIP and sintering) and hot rolling followed by stress-relief heat treatment were used to produce these plates, whose fundamental manufacturing conditions and final plate thickness (around 7 mm) were the same as those used for the aforementioned materials in our previous studies. The La2O3 content (1 wt.%) was determined based on manufacturing constraints (details cannot be disclosed). The concentration of K was approximately 30 wt. ppm, which was the same as those used for the aforementioned materials in our previous studies. In the evaluation of the W-Ta binary alloy system in our previous studies, an increase in Ta addition tended to raise the recrystallization temperature, but the DBTT was significantly lower with 1 wt.% addition than with 3 wt.%. Therefore, the Ta addition was set to 1 wt.%. Concentration of major elements and interstitial impurity atoms (carbon (C), nitrogen (N), and oxygen (O)) in those plates are summarized in Table 1.
| Material | Re [wt.%] | Ta [wt.%] | K [wt. ppm] | La [wt.%] | C [wt. ppm] | O [wt. ppm] | N [wt. ppm] |
|---|---|---|---|---|---|---|---|
| Pure W | — | — | — | — | < 10 | < 10 | < 10 |
| K-doped W | — | — | 30 | — | < 10 | < 10 | < 10 |
| W-3%Re | 3 | — | — | — | < 10 | < 10 | < 10 |
| K-doped W-3%Re | 3 | — | 28 | — | < 10 | < 10 | < 10 |
| W-1%Ta | — | 0.96 | — | — | < 10 | 20 | < 10 |
| W-3%Ta | — | 3.0 | — | — | < 10 | 10 | < 10 |
| W-5%Ta | — | 5.0 | — | — | < 10 | 20 | < 10 |
| W-1%La2O3 | — | — | — | 0.84* | < 10 | 1,400* | < 10 |
| K-doped W-1%Ta | — | 1.02 | 28 | — | < 10 | 10 | < 10 |
*Theoretically, 1,473 wt. ppm of O and 0.85 wt.% of La result in 1 wt.% of La2O3.
Charpy impact tests were performed at temperatures below 1,000°C in vacuum. A KLST Charpy V-notched specimen aligned in the L–S direction was used. Nomenclature of L-, T-, and S-directions of materials and specimens is shown in Fig. 1. The details of the test conditions are described in our previous papers [4–7, 14].
Vickers hardness measurements and grain structure observations on L × S surface were conducted for the specimens in the as-received condition and after the isochronal heat treatment for 1 h to evaluate the recrystallization behavior. The heat treatment temperature was up to 2,300°C. The load and dwell time of the hardness measurements were 1.96 N and 15 s, respectively. Electron backscatter diffraction (EBSD) images and optical micrographs were used for grain structure observation.
2.4 Thermal diffusivity measurementThermal diffusivity measurements were performed in the S-direction at temperatures up to 500°C in argon using a xenon-flash apparatus. Disk-shaped specimens with a diameter of 10 mm and thickness of 2 mm were prepared for these measurements. The details of the test conditions are described in our previous papers [7, 20]. Thermal diffusivity measurement was performed for W-1%Re as well.
For W-1%La2O3, measurements were performed in the L-, T-, and S-directions to evaluate the anisotropy of the thermal diffusivity (see Fig. 1). In our previous study, K-doping and alloying with Re resulted in almost no anisotropy in the thermal diffusivity of hot-rolled W plates [20]. La2O3 has been reported to have low thermal diffusivity [21]. Furthermore, it is assumed that the La2O3 particles in a rolled plate do not have an equiaxed shape due to the effects of rolling, which could cause anisotropy in the thermal diffusivity of the plate. Therefore, the thermal diffusivity was measured in three directions. Since the rolled plate has a thickness of 7 mm, disc-shaped specimens with a diameter of 6 mm and thickness of 2 mm were used.
Test temperature dependence of absorbed energy obtained from Charpy impact tests of materials in the as-received condition is shown in Fig. 2. Although the variability in absorbed energy has not been evaluated through a large number of tests, repeated tests on several materials at the same temperature in our previous studies indicated that this variability was not significant [3, 5, 14]. In the present study, DBTT determined by Charpy impact test (DBTTCharpy) was defined as the intermediate temperature within the transition region where the absorbed energy decreases from a finite value to nearly zero. Based on this definition, the DBTTCharpy values for each material (evaluated in 50°C increments) were: 550°C for pure W, 350°C for K-doped W, 450°C for W-3%Re, 250°C for K-doped W-3%Re, 250°C for W-1%Ta, 550°C for W-3%Ta, 250°C for K-doped W-1%Ta, and 350°C for W-1%La2O3. Although slightly lower DBTTCharpy values may be estimated for Pure W, W-3%Ta, and K-doped W-1%Ta, these differences do not significantly affect the relative comparison; therefore, the values listed above were used.
K-doping and La2O3 particle dispersion were introduced as second-phase dispersions in the present study, and their effects on DBTTCharpy were comparable. Because the amounts of K and La2O3, as well as the sizes and number densities of K-bubbles and La2O3 particles, are expected to differ, a more detailed evaluation will become clearer once these factors are elucidated.
The effect of alloying with Ta was significant at a Ta content of 1%, whereas almost no effect was observed at 3%. When comparing Ta and Re as solid-solute elements, Re exhibited a greater effect on DBTTCharpy at the 3% addition level. However, this effect remained smaller than that produced by the second-phase dispersions described above. Although a Charpy impact test has not yet been conducted for W-5%Ta, its DBTTCharpy is expected to be similar to that of W-3%Ta, considering its Vickers hardness in the as-received condition and the temperature dependence of its tensile properties, particularly elongation [7]. The behavior at Ta contents below 1% is of particular interest. Because the surface dose rate of Ta after neutron irradiation is higher than that of W, there remain concerns regarding its use as a PFM subjected to neutron irradiation [22]. In addition, Ta causes relatively significant solid-solution hardening in W, which is unfavorable for component fabrication due to reduced machinability. Therefore, determining the extent to which mechanical properties can be improved with a small amount of Ta is crucial for assessing the potential of W-Ta alloys.
K-doped W-3%Re exhibited a superior DBTTCharpy compared with both K-doped W and W-3%Re, demonstrating the combined effect of K-doping and alloying with Re. In contrast, for K-doped W-1%Ta, the combined effect of K-doping and alloying with Ta on DBTTCharpy was not clearly observed, and the DBTTCharpy was essentially the same as that of W-1%Ta. Alloying with Re and Ta can influence the characteristics of K-bubbles, including their size, number density, and spatial distribution. Therefore, in K-doped W-1%Ta, the state of the K-bubbles may not have been sufficient to cause a further decrease in DBTTCharpy compared to W-1%Ta. However, closer examination of the absorbed energy values at 200°C and 300°C reveals that K-doping slightly decreases the DBTTCharpy of W-1%Ta, indicating that the effect is not negligible.
3.2 Recrystallization and grain growthThe effects of various modification methods on recrystallization resistance were evaluated based on changes in Vickers hardness and grain structure after isochronal heat treatment for 1 h. In general, heat treatment leads to microstructural recovery, followed by primary recrystallization and subsequently secondary recrystallization, with grain growth occurring during these processes. Hardness is a suitable parameter for assessing recovery and primary recrystallization. In contrast, hardness remains largely unchanged during secondary recrystallization and the associated grain growth because primary recrystallization eliminates most dislocations in the matrix. Therefore, observation of the grain structure is more appropriate for evaluating these stages.
Figure 3 shows heat treatment temperature dependence of Vickers hardness. Except for W-1%Ta, all plates exhibited a gradual decrease in hardness from the as-received state, followed by a rapid decrease with increasing heat treatment temperature. After this rapid decrease, only negligible decreases were observed up to 2,300°C. These gradual, rapid, and negligible decreases correspond to microstructural recovery, primary recrystallization, and secondary recrystallization, respectively. In contrast to the other materials, W-1%Ta exhibited a temperature range in which hardness decreased gradually before the rapid decrease. This gradual decrease will be attributed to partial recrystallization occurring even after the onset of primary recrystallization. Partial recrystallization was confirmed by the EBSD image shown in Fig. 4, obtained after heat treatment at 1,500°C. In the present study, the recrystallization temperature (TReX_50%/1h) was defined as the temperature corresponding to the average of the hardness at the end of the first gradual decrease and the hardness at the end of the rapid decrease. Based on this definition, the TReX_50%/1h values (evaluated in 25°C increments) were 1,225°C for pure W, 1,325°C for K-doped W, 1,475°C for W-3%Re, 1,425°C for K-doped W-3%Re, 1,525°C for W-1%Ta, 1,625°C for W-3%Ta, 1,625°C for W-5%Ta, 1,575°C for K-doped W-1%Ta, and 1,375°C for W-1%La2O3.
When comparing second-phase dispersion and solid-solution alloying, the latter had a more pronounced effect on increasing the recrystallization temperature (i.e., improving recrystallization resistance). Suppression of recrystallization by second-phase dispersion and solid-solution alloying is generally attributed to the Zener pinning and the solute drag effect, respectively. The mechanism responsible for the quantitative differences observed in W will be clarified in future studies. When comparing Ta and Re as solid-solute elements, Ta produced a greater improvement in recrystallization resistance at the 3% addition level. According to the periodic table, Ta has a larger atomic radius than W, while Re has an atomic radius very close to that of W. Therefore, when dissolved in W, W alloys containing Re are expected to exhibit smaller lattice distortion and weaker solute drag, resulting in lower recrystallization resistance than alloys containing Ta [23, 24]. The combined effects of K-doping and alloying with 3% Re, as well as those of K-doping and alloying with 1% Ta, were not significant. Although K-doping clearly improves recrystallization resistance when comparing pure W and K-doped W, the mechanism by which this effect is not observed in W-3%Re and W-1%Ta requires further investigation.
EBSD images and optical micrographs of materials after isochronal heat treatment for 1 h are summarized in Fig. 4 [3, 5, 6]. Using these images, the grain size along the S-direction (dS) was measured, and its heat treatment temperature dependence is shown in Fig. 5 [3, 4]. Although grain sizes differed among materials, all plates exhibited elongated grains in the rolling direction prior to primary recrystallization, whereas equiaxed grains were observed after primary recrystallization. These observations are consistent with the standard recrystallization mechanism. In the temperature range corresponding to primary recrystallization, which is accompanied by a rapid decrease in hardness, pure W exhibited a substantial increase in grain size, whereas grain growth was suppressed in the other materials. This trend was also observed during secondary recrystallization. In this temperature range, pure W exhibited a marked increase in grain size (approximately a tenfold increase), whereas the increase in grain size for the other materials was much smaller. The significant grain growth in pure W and its suppression in other materials indicate that grain boundary motion is hindered by second-phase dispersion and solid-solution alloying. Among the modified materials other than pure W, no significant differences were observed in grain growth during primary recrystallization. However, during secondary recrystallization, the two Re-containing materials (W-3%Re and K-doped W-3%Re) exhibited a substantially greater increase in grain size than the other materials. Because grain growth during secondary recrystallization is smaller in W-3%Re than in pure W, but larger in K-doped W-3%Re than in K-doped W, this phenomenon cannot be explained solely by the effect of Re solid-solution. The mechanism underlying this behavior will be clarified in future studies.
Figure 6 summarizes the temperature range from DBTTCharpy to TReX_50%/1h. Although the extent varies among materials, the second-phase dispersion and solid-solution alloying methods employed in the present study, as well as their combined application, expanded this temperature range (DBTTCharpy remained unchanged for W-3%Ta and is unknown for W-5%Ta.). When these results are considered as a provisional operation temperature range of PFM, K-doped W-1%Ta and W-1%Ta appear particularly promising. However, given the previously mentioned concerns regarding the surface dose rate and machinability of W-Ta alloys, the appropriateness of using Ta as a solid-solute element must be evaluated from multiple perspectives, including machinability, thermal properties, and features observed during and after neutron irradiation (e.g., surface dose rate).
The discussion thus far has focused on material modifications aimed at improving mechanical properties and thereby expanding the operation temperature range. However, as noted earlier, thermal properties are equally important for PFM. Therefore, changes in thermal diffusivity resulting from the material modification were also evaluated. Figure 7 shows the test temperature dependence of thermal diffusivity [7, 20]. When the laser incident direction was the S-direction, the thermal diffusivity of K-doped W was equivalent to that of pure W, whereas all other materials exhibited lower thermal diffusivity, with the difference becoming more pronounced at lower temperatures. Furthermore, when the thermal diffusivity of W-1%La2O3 was measured along three directions at room temperature, anisotropy was observed; in the T- and L-directions, thermal diffusivity was equivalent to that of pure W.
In general, solid-solution alloying introduces atoms of different sizes and masses into the crystal lattice, creating structural disorder. This disorder increases the scattering of electrons and phonons, the primary heat carriers in metals, thereby reducing both thermal conductivity and thermal diffusivity compared with pure metals. The lower thermal diffusivity observed in the alloyed materials in the present study is consistent with this general mechanism. Considering that solid-solute microstructures contain no precipitates, anisotropy in thermal diffusivity is not expected. Indeed, Fukuda et al. reported no significant anisotropy in the thermal diffusivity of W-1%Re [20].
For materials dispersed with second-phases, it is generally understood that the small volume fraction and small size of the dispersed particles or bubbles do not significantly disturb the crystal lattice. Because electrons and phonons are only weakly scattered by these features, the heat-carrying mechanisms remain largely unchanged, and thermal conductivity and thermal diffusivity remain close to those of the pure metal. K-doped W is consistent with this general understanding; Fukuda et al. reported no significant anisotropy in its thermal diffusivity [20].
In contrast, the thermal diffusivity of W-1%La2O3 in the S-direction was lower than that of pure W, and clear anisotropy was observed among the L-, T-, and S-directions. The La2O3 particles are believed to have a pancake-like morphology, elongated in the L-direction and slightly elongated in the T-direction due to hot-rolling. As shown in the EPMA image in Fig. 8, the La2O3 particle sizes in the L- and S-directions were approximately 10 μm and a few micrometers, respectively. Therefore, particularly when the laser incident direction was the S-direction, the presence of these relatively large La2O3 particles may have been sufficient to influence heat conduction. According to Fornarini et al. [21], the thermal diffusivity of La2O3 itself is very low compared with that of W (approximately 2.7 mm2/s at room temperature). Thus, the reduced thermal diffusivity of W-1%La2O3 in the S-direction will be attributed to the influence of these relatively large La2O3 particles.
Based on the above discussion, solid-solution alloying reduces the thermal conductivity of W regardless of the alloying element, and the magnitude of the reduction increases with alloying content. This must be considered when using alloyed W as a PFM. Furthermore, for W dispersed with second-phases, there is a concern that thermal conductivity may decrease when the dispersed particles or bubbles become sufficiently large, as observed for W-1%La2O3. Such effects, including shape-dependent anisotropy, must therefore be taken into account.
The present study investigated how to expand the operational temperature range of W for fusion reactor PFM by improving the DBTT and recrystallization resistance. Nine materials were examined: pure W, K-doped W, W-3%Re, K-doped W-3%Re, W-1%Ta, W-3%Ta, W-5%Ta, W-1%La2O3, and K-doped W-1%Ta. All were fabricated by powder metallurgy and hot rolling. Their DBTT, recrystallization behavior and grain growth, and thermal diffusivity were evaluated using Charpy impact tests, hardness measurements and grain structure observations after isochronal heat treatments, and thermal diffusivity measurements.
1) K-doping and La2O3 particle dispersion both lowered DBTT. Re alloying (3 wt.%) also reduced DBTT, although its effect was smaller than that of second-phase dispersions. Ta alloying improved DBTT only at 1 wt.% and showed little benefit at 3 wt.%.
2) Solid-solution alloying significantly increased recrystallization temperature. Materials alloyed with Ta showed the highest recrystallization temperatures (1,525–1,625°C), outperforming both materials alloyed with Re and those dispersed with second-phases.
3) Pure W exhibited rapid grain growth during both primary and secondary recrystallization. Modified materials suppressed grain growth, whereas materials alloyed with Re showed comparatively larger grain growth during secondary recrystallization.
4) K-doped W-3%Re achieved the lowest DBTT among all evaluated materials. This demonstrated a combined effect of K-doping and alloying with Re. Recrystallization temperature of this material was higher than pure W and K-doped W, but slightly lower than W-3%Re. Thus, the combined effect was unclear.
5) DBTT of K-doped W-1%Ta was essentially the same as W-1%Ta. Thus, no strong combined effect of K-doping and alloying with Ta was observed. Recrystallization temperature of this material was slightly higher than W-1%Ta. The improvement was modest, and combined effect was limited.
6) All alloyed materials showed reduced thermal diffusivity. W-1%La2O3 exhibited pronounced anisotropy due to large, elongated La2O3 particles.
The results show that both second-phase dispersion and solid-solution alloying can expand the temperature range from DBTT to recrystallization temperature, potentially broadening the design window and operational conditions available for plasma facing components. However, alloying with Ta raises concerns on higher surface dose rate after neutron irradiation compared with W, as well as reduced machinability. Furthermore, although thermal properties must also be considered for PFM, solid-solution alloying lowers thermal diffusivity, and large second-phase particles—such as La2O3—can introduce anisotropy and may reduce thermal diffusivity. Thus, selecting W-based materials requires balancing mechanical benefits with thermal constraints to ensure reliable performance in future fusion devices.
The present study was conducted through a collaboration between A.L.M.T. Corp. and Tohoku University. We are grateful to the many staff members and students of Prof. Hasegawa’s and Prof. Nogami’s laboratories at Tohoku University for their contributions to this work. We also thank Dr. Michael Rieth, Dr. Philipp Lied, Mr. Sven Sickinger, and Mr. Siegfried Baumgärtner of the Institute for Applied Materials (IAM), Karlsruhe Institute of Technology (KIT), for their support with the Charpy impact tests conducted at KIT. We further express our gratitude to Prof. Ryuta Kasada and Dr. Hao Yu of the Institute for Materials Research, Tohoku University, for their support with the thermal diffusivity measurements.
