At the National Institute for Fusion Science (NIFS) researchers have introduced a new metallurgical method (brazing technique) for bonding tungsten and copper alloys. They achieved a superior bonding layer that is tough and difficult to fracture by using a technique that does not utilize intermediate material except for the filler material. They have succeeded in fabricating a small-scale divertor mock-up which has an excellent heat removal capability.
The divertor is the device that continuously receives the extremely high heat and particle loads from the nuclear fusion plasma. Research and development on the divertor, which is highly reliable for heat removal, is being conducted around the world. Tungsten (W) block is being considered as the divertor armor material. Tungsten has great advantages, such as low hydrogen isotope retention and low sputtering yield. On the back side of the tungsten armor, the water-cooled heat sink made of copper alloys which excel at heat transfer will be bonded. They are considering adoption of the same structure in the helical reactor (FFHR-d1), for which NIFS is advancing with design research. For this it is necessary to bond tungsten and copper alloys. However, because these two materials do not make an alloy, a bonding material called filler material is inserted into the space between the tungsten and the copper alloy, and is heated up to the high temperature of more than 900˚ C. Further, because the thermal expansion coefficients of tungsten and copper alloy are largely different, in methods used to date intermediate materials that absorb thermal stress must be inserted simultaneously with the filler material. Until now, the method without using an intermediate material has been considered to be technically difficult. However, by inserting the intermediate material, the number of bonded interfaces and the bonded area increases, and, moreover, the strength weakens, and there emerge the problems of falling heat removal performance and of rising production costs.
Research Results and Their Significance
Professor Masayuki Tokitani and his research group at the National Institute for Fusion Science have developed a new technique for direct tough bonding between tungsten and copper alloys by making the bonding layer as a cushion even without using an intermediate material. Using this bonding technique, they succeeded in the fabrication of a small-scale divertor mock-up with excellent heat removal capability even under the reactor relevant condition (~15 MW/m2).
The divertor components must endure the extreme high-heat flux. Further, during the heat treatment phase for brazing, since the entire component is heated up to approximately 900˚C, it is then cooled to the room temperature. Therefore, the thermal stress is induced in the bonding interface of the armor and heat sink material. Such thermal stress should be reduced as much as possible. This time, in order to meet these requirements simultaneously, the research group used the filler material BNi-6 (Ni-11%P) and the oxide dispersion strengthened copper alloy (ODS-Cu), GlidCop® (Cu-0.3wt%Al2O3) and fulfilled the optimal bonding condition.
More specifically, Professor Tokitani’s group set the thickness of the brazing material at 38 mm, and the heat treatment temperature and the duration at 960˚C and 10 minutes, respectively, for that time when the brazing is undertaken. Then, in cooling from 960˚C to 100˚C they used the extremely slow natural cooling. In cooling from 100˚C to room temperature they used nitrogen gas cooling. After the brazing, the three-point bending test was carried out for evaluating the bonding strength. Surprisingly, the bonding layer has a ductile property as shown in Fig. 1. Fig. 1(a) shows the schematic drawing of the ductile feature of the brazing layer. Fig. 2(b) shows the stress-strain curve of the bending test. The line in red shows the result. The yield strength reached to around 200 MPa. Since yield strengths of both tungsten and GlidCop® are over 300 MPa even after the brazing heat treatment, the deformation region must be focused on the brazing layer itself. When the strain is 0.2%, at a glance it may not be thought to be a particularly significant plastic deformation. However, since the actual plastic deformable region is very thin, e.g., a few tens of micrometers, the absolute local strain should be significantly greater than 0.2%. This is a surprising result.
This means that the bonding layer obtains toughness, and that the induced thermal stress during brazing heat treatment can be absorbed by the brazing layer. Further, such a relaxation capability of the applied stress has great merit from the viewpoint of the reliability of the divertor components even when they accept the unexpected thermal stress during the reactor operation. Incidentally, the dashed line in green is an example of a failed brazing in which different copper alloy and brazing materials are used. In the case of the failed brazing, the bonding layer was fractured with a brittle feature at 1/4 of the stress as compared with the advanced bonding technique of this study. Further, the small-scale divertor mock-up of the W/BNi-6/GlidCop® was then successfully fabricated by the advanced brazing technique as shown in Fig. 2(a). The heat loading test under the reactor relevant condition to the mock-up was carried out by using the electron beam device ACT2 at NIFS. Fig. 2(b) and (c) show the schematic view of the cross-section of the mock-up and the temperature of the mock-up during a steady state heat loading up to ~15 MW/m2, respectively. The temperature of 650˚C was sufficiently lower than that of the melting point of the BNi-6 (875˚C) and recrystallization temperature of tungsten (~1500˚C). The reason why such an excellent heat removal capability was obtained is that since the direct bonding without any intermediate material was adopted, a minimum heat transfer resistance from the armor to the heat sink could be maintained.
The advanced brazing technique of this study will contribute not only to constructing the superior divertor but also to greatly reducing the construction cost of the entire divertor structure in the future fusion reactor. In the future work, using this method, we will produce a large-scale divertor component whose structure will be similar to the divertor that will be utilized in the nuclear fusion reactor. We will aim at a divertor design and construction that will make long-range operation and safe usage possible.
This research result was presented at the 26th IAEA Fusion Energy Conference held in Kyoto, Japan, October 17-22, 2016.
(a) Schematic view of the deformed BNi-6 bonding layer during the three-point bending test. Bonding layer is locally deformed.
(b) Representative stress-strain curve (red lines) of the toughened bonding layer. Obvious yield point can be confirmed in point (1). The dashed line in green is an example of a failed brazing in which different copper alloy and brazing material are used.
(a) Small-scale divertor mock-up fabricated by W/BNi-6/GlidCop® brazing.
(b) Schematic view of the cross-section of the small-scale divertor mock-up.
(c) Temperature of the small-scale divertor mock-up during a steady state heat loading. Point A and B correspond to the position A and B in (b), respectively. The highest temperature under 15 MW/m2 is ~650˚C in the W plate. This temperature is quite acceptable for this mock-up from the viewpoint of structural reliability. After the 15 MW/m2 heat loading, the mock-up did not show any sign of a visible destruction.
In the nuclear fusion reactor, the divertor has the function of maintaining the purity of a nuclear fusion plasma by neutralizing the charged particles of impurities including fuel gas particles flowing along the magnetic field lines and by exhausting them. Conversely, because the kinetic energy of the charged particles becomes heat on the divertor surface, this place continuously receives the highest heat load.
Recrystallization occurs when a metal material is heated to a temperature of from 1/3 to 1/2 of the melting point. When tungsten recrystallization occurs, brittleness is accelerated because the grain boundaries are easily fractured. This is called recrystallization embrittlement.
(3) ODS-Cu: Oxide Dispersion Strengthened Copper alloy
By mixing the nano-size oxide particles (typically less than 1%) in the copper matrix, the strength is increased.
National Institute for Fusion Science grant administrative budget: NIFS11UFFF032
Dr. Masayuki Tokitani
National Institute for Fusion Science