Tritium inventory and recovery in next-step fusion devices

Future fusion devices will use tritium and deuterium fuel. Because tritium is both radioactive and expensive, it is absolutely necessary that there be an understanding of the tritium retention characteristics of the materials used in these devices as well as how to recover the tritium. There are three materials that are strong candidates for plasma-facing-material use in next-step fusion devices. These are beryllium, tungsten, and carbon. While beryllium has the disadvantage of high sputtering and low melting point (which limits its power handling capabilities in divertor areas), it has the advantages of being a low-Z material with a good thermal conductivity and the ability to get oxygen from the plasma. Due to beryllium's very low solubility for hydrogen, implantation of beryllium with deuterium and tritium results in a saturated layer in the very near-surface with limited inventory (J. Nucl. Mater. 273 (1999) 1). Unfortunately, there are nuclear reactions generated by neutrons that will breed tritium and helium in the material bulk (J. Nucl. Mater. 179 (1991) 329). This process will lead to a substantial tritium inventory in the bulk of the beryllium after long-term neutron exposure (i.e. well beyond the operation life time of a next-step reactor like ITER). Tungsten is a high-Z material that will be used in the divertor region of next-step devices (e.g. ITER) and possibly as a first wall material in later devices. The divertor is the preferred location for tungsten use because net erosion is very low there due to low sputtering and high redeposition. While experiments are still continuing on tritium retention in tungsten, present data suggest that relatively low tritium inventories will result with this material (J. Nucl. Mater. 290–293 (2001) 505). For tritium inventories, carbon is the problem material. Neutron damage to the graphite can result in substantial bulk tritium retention (J. Nucl. Mater. 191–194 (1992) 368), and codeposition of the sputtered carbon with the tritium from the plasma will produce a layer of carbonaceous material potentially containing kilograms of tritium in the cooler areas of the tokamak (J. Vac. Sci. Technol. A5 (1987) 2286). This paper reviews the tritium retention mechanisms for the three materials discussed above. Tritium removal techniques, including those used in situ to minimize in-vessel inventories as well as those used to reduce contamination prior to waste disposal, are discussed.