The effect of magnetic balance and particle drifts on radiating divertor behavior in DIII-D

The success of the 'puff-and-pump' radiating divertor approach to heat flux reduction at the divertor targets while maintaining high performance H-mode tokamak plasmas depends sensitively on the divertor magnetic geometry and the ion ∇B drift. In the puff-and-pump scenari used here, Argon (Ar) was injected near the outer divertor target, while plasma flows into the inner and outer divertors were enhanced by a combination of particle pumping near divertor targets and D₂ gas puffing upstream of the divertor targets. The key to the success of this approach is in preventing the injected Ar from escaping the divertor and contaminating the main plasma. Under similar conditions, Ar accumulated at a 2 - 3 times faster rate in the main plasma of magnetically balanced double-null (DN) configurations than in magnetically unbalanced DN shapes (where |dRsep| was typically 1.2 cm). For balanced DNs, stronger increases in Ar concentration were observed in the divertor opposite to the ion ∇B drift direction, regardless of the divertor into which the Ar was initially injected. For unbalanced DNs, there was significantly higher divertor enrichment and lower Ar leakage out of the divertor when the ion ∇B drift was directed away from the dominant X-point. That configuration provides the best prospect for coupling a radiating divertor approach to a high performance H-mode plasma. Exhaust-to-core argon concentration ratios ∼ 35 were achieved under this scenario, along with good H-mode conditions: τ_E/ /τ_89P = 2, β_N = 2, n_e /n_eG = 0.6, P_RAD /P_INJ ∼ 0.6, and Z_eff ∼ 2. Related experiments using ELMing H-mode discharges compared the plasma behavior in the open lower divertor to that of the more closed upper divertor and demonstrated that control of D₂ and Ar inventory was more sensitive to the ion ∇B drift direction than to the relative closure of the divertor. The roles of particle drifts in the divertor(s) and scrape-off layer are discussed in the context of UEDGE analysis. (author)