Workshop on runaway electrons in ITER

During plasma disruptions in tokamaks, the fast current quench generates substantial electromagnetic forces on the vacuum vessel and can also produce a significant current of runaway electrons (REs). The REs are accelerated in the high electric field associated with the current quench and resemble an electron beam with velocities close to that of light. Model predictions indicate that in ITER a runaway electron current of up 70% of the initial plasma current could be generated due to a "secondary avalanche" process. REs are eventually lost to the first wall and this can result in significant local energy deposition, which can potentially cause damage such as localized melting of the first wall surface. Assessment of the severity of such events requires an accurate quantitative specification of the characteristics of the energy deposition of REs, e.g., probable locations of deposition, time duration of loss process and incident angle of electrons, is of primary importance.

To eliminate the potential for such damage, development of techniques for mitigation of REs (i.e. avoidance of their generation and/or suppression of generated REs) is a key issue for ITER. To date, two potential methods, massive gas injection (MGI) and stochastic magnetic field application, have been investigated in current tokamaks. Both methods would require design changes or some upgrading of foreseen systems to be implemented. Therefore, their suitability for ITER must be assessed and the relevant physics guidelines must be specified in the near future.

To this end, a three-day workshop was held in Cadarache during 15-17 July under the leadership of Masayoshi Sugihara from the ITER Fusion Science & Technology Division, to review the latest information on the experimental characteristics of REs and the results of mitigation experiments, to update the ITER specifications in this area and to promote R&D studies in support of the design of the ITER RE mitigation system. 25 scientists and engineers from laboratories around the world and about 20 IO members participated in the workshop, listening to a series of presentations which revealed considerable progress in this area in recent years and contributing to several lively and productive discussion sessions on the implications of the results for ITER and for future R&D activities in response to ITER's needs.

Sugihara summarizes the main outcome of the workshop as follows:

Many devices observe a threshold in toroidal magnetic field (remarkably uniform at ~2T) and/or safety factor (qeff ~2) for the generation of REs at disruptions. Larger RE currents are generated for higher current quench rates. These data suggest that the present specifications of the frequency with which REs occur in ITER (i.e. it is assumed that REs are generated at every major disruption) and of anticipated RE current fractions may be pessimistic.

The possibility of controlling the vertical position of the RE beam using the proposed in-vessel coils was considered in relation to the fact that the spatial profile of the RE current will probably be more peaked than that of typical ITER plasmas. Nevertheless, if the position of the RE-dominated plasma could be controlled for sufficiently long, several possible RE suppression schemes could be applied, e.g., impurity gas injection, application of reverse one-turn voltage, interception by a dedicated sacrificial wall element, or application of a stochastic magnetic field.

Understanding of the mechanism by which massive gas injection mitigates RE generation has steadily progressed. The figure shows the penetration of injected gas into the TEXTOR plasma. All tokamaks demonstrate that REs are suppressed, even when density is below the predicted threshold density for suppression (commonly referred to as the "Rosenbluth" density). According to numerical studies, this can be explained by the induced stochastic magnetic perturbation in the plasma which occurs at the disruption.

The suppression of REs by an external magnetic perturbation has been demonstrated. The required magnetic perturbation is estimated as ~0.1% of the toroidal field in the spatial region in which REs are most likely to be generated (usually close to the pre-disruptive plasma centre).

Helium has been recommended as the optimum gas species for MGI. Pumping down the torus down to the required operating pressure after the injection of ~5 x 10^5 Pam3 of gas appears possible in a time of 3-4 hours with some additional investment, allowing an acceptable recovery time for plasma operation.

Building on the results of the workshop, the specification of possible mitigation schemes (including combinations of several techniques) will be updated and a programme of studies will be developed between the IO and Members' fusion communities to support further improvements in the understanding of RE characteristics and the processes contributing to their mitigation. This will provide important input for the design of the ITER disruption mitigation system.