Enhancement of Mid-/High-Z Impurity Transport by Continuous Li-granule Dropping in a Stellarator Plasma
An enhancement of core impurity transport is observed in high-density plasmas of the stellarator LHD heated by neutral beam injection when continuous lithium (Li) granule dropping is performed. In the experiments reported here, in which the TESPEL is employed to inject trace amounts of titanium (Ti) and molybdenum (Mo) into the plasma core, confinement times for these impurities are seen to reduce significantly when Li dropping is applied, this reduction being more notable for Mo. In order to gain some initial insight into these observations, simulations are performed using the drift-kinetic transport code SFINCS for the Mo case. These simulations indicate that, while neoclassical transport prevails for the main plasma components (electrons, majority ions and low Z impurities), the classical contribution seems to be dominant for transporting Mo impurities. In summary, this work reports the first experimental observation of the degradation of mid-Z and high-Z impurity confinement induced by the continuous dropping of Li granules into a high-density stellarator plasma. In the case of the Mo impurity, simulations suggest that classical transport is the key mechanism underlying the enhanced impurity transport.
💡 Research Summary
This paper reports the first experimental observation that continuous dropping of lithium (Li) granules into the edge of high‑density Large Helical Device (LHD) plasmas significantly degrades the confinement of mid‑Z (titanium, Z = 22) and high‑Z (molybdenum, Z = 42) impurity ions. The experiments were performed in an inward‑shifted magnetic configuration (Rₐₓ = 3.6 m, Bₐₓ = 2.75 T) with hydrogen as the working gas. The plasma was heated by 3 MW of electron cyclotron resonance heating (ECRH) and up to 7 MW of neutral beam injection (NBI). Typical parameters were line‑averaged electron density nₑ,avg ≈ 5.3 × 10¹⁹ m⁻³, central electron temperature Tₑ₀ ≈ 2.2 keV, ion temperature Tᵢ₀ ≈ 1 keV, and stored energy ≈ 880 kJ.
Lithium granules were released continuously from ≈ 4.8 s until the end of the discharge by gravity using the Impurity Powder Dropper (IPD). The Li injection caused modest changes in bulk plasma: edge electron density increased by < 5 %, central Tₑ rose by ~12 %, stored energy grew by ~10 %, and effective charge Z_eff increased from ~1.8 to ~2.3. Simultaneously, Hα emission dropped by ~25 %, indicating a reduction in particle fueling or altered wall recycling.
Trace impurity experiments employed the TESPEL (Tracer‑Encapsulated Solid Pellet) system to inject known quantities of Ti and Mo at a radial location r_eff/a₉₉ ≈ 0.75 (inside the density shoulder). Approximately 5 × 10¹⁷ atoms of each species were delivered. The injected cloud was assumed to remain at the deposition radius because outward drift of solid pellets is negligible.
Impurity confinement times were extracted from the decay of characteristic EUV/VUV spectral lines: Ti XX (25.93 nm) and Mo XXXII (12.79 nm). Without Li, Ti confinement time τ_Ti = 1.38 ± 0.18 s, while with Li it reduced to 1.15 ± 0.4 s (≈ 17 % decrease). For Mo, the effect was far more dramatic: τ_Mo without Li was 6.34 ± 2.63 s, but with Li it fell to 1.43 ± 0.8 s, a reduction of about 78 %. The Z‑dependence suggests that Li granule dropping preferentially enhances transport of higher‑Z species.
To interpret the measurements, the 1‑D impurity transport code STRAHL was used. Reasonable agreement with the experimental decay curves was obtained using spatially uniform diffusion coefficients D ≈ 0.08–0.11 m² s⁻¹ and convective velocities V with minima of –0.9 m s⁻¹ (Ti without Li) to –2.0 m s⁻¹ (Mo without Li). The presence of Li required only modest adjustments to V, indicating that Li mainly modifies the convective component that drives impurities outward.
The radial electric field E_r, inferred from 2D‑PCI phase‑velocity measurements, remained negative and essentially unchanged by Li injection, ruling out E_r‑driven changes as the primary cause. However, the amplitude of the measured turbulence (identified as resistive‑interchange, RI, modes) was significantly reduced when Li granules were present. While reduced RI turbulence can explain the observed improvement in energy confinement, it would normally be expected to lessen impurity transport, contrary to the experimental finding.
To resolve this paradox, drift‑kinetic simulations with the SFINCS code were performed for the Mo case. The calculations separated neoclassical (NC) and classical (C) contributions to particle fluxes. For bulk electrons and H⁺ ions, NC fluxes dominate and increase slightly with Li. For the trace Mo⁺³¹ ions, the total flux is overwhelmingly classical, and the classical component grows markedly when Li is present. This suggests that, under the high‑density, ion‑root (negative E_r) conditions of the experiment, the transport of high‑Z impurities is governed by collisional classical mechanisms rather than neoclassical ones, and that Li granule deposition enhances this classical channel.
In summary, continuous Li granule dropping in high‑density LHD plasmas produces a subtle modification of edge plasma conditions that improves overall particle and energy confinement but simultaneously accelerates the outward transport of mid‑ and high‑Z impurities. The effect is especially strong for Mo, where classical transport dominates. These results highlight a previously unappreciated route to impurity control (or degradation) in stellarator‑type fusion devices and imply that operational scenarios for future high‑performance reactors must carefully consider the impact of low‑Z material injection on high‑Z impurity behavior.
Comments & Academic Discussion
Loading comments...
Leave a Comment