Search

Large edge-localized modes (ELMs) were mitigated by gravitational injection of lithium granules into the upper X-point region of the experimental advanced superconducting tokamak (EAST) device with tungsten plasma-facing components. The maximum ELM size was reduced by ∼70% in high β N H-mode plasmas. Large ELM stabilization was sustained for up to about 40 energy confinement times, with constant core radiated power and no evidence of high-Z or low-Z impurity accumulation. The lithium granules injection reduced the edge plasma pedestal density and temperature and their gradients, due to increased edge radiation and reduced recycling from the plasma-facing components. Ideal stability calculations using the ELITE code indicate that the stabilization of large ELMs correlates with improved stability of intermediate-n peeling-ballooning modes, due to reduced edge current resulting from the profile changes. The pedestal pressure reduction was partially offset by a core density increase, which resulted in a modest ∼7% drop in core stored energy and normalized energy confinement time. We surmise that the remnant small ELMs are triggered by the penetration of multiple Li granules just past the separatrix, similar to small ELMs triggered by deuterium pellet Futatani et al (2014 Nucl. Fusion 54 073008). This study extends previous ELM elimination with Li powder injection Maingi et al (2018 Nucl. Fusion 58 024003) in EAST because (1) use of small, dust-like powder and the related potential health hazards were eliminated, and (2) use of macroscopic granules should be more applicable to future devices, due to deeper penetration than dust particles, e.g. inside the separatrix with velocities ∼10 m s−1 in EAST.

We report an observation of robust suppression of edge-localized modes (ELMs) in the Experimental Advanced Superconducting Tokamak (EAST), enabled by continuous boron (B) powder injection. Edge harmonic oscillations appear during B powder injection, providing sufficient particle transport to maintain constant density and avoid impurity accumulation in ELM-stable plasmas. Quasi-steady ELM suppression discharges are demonstrated with modest energy confinement improvement and over a wide range of conditions: heating power and technique variation, electron density range over a factor ∼3.5, deuterium or helium ion species, and with either direction of the toroidal magnetic field. ELM suppression is observed above a threshold edge B intensity and ceases within 0.5 s of termination of the B injection. In contrast to ELM suppression accompanied by recycling reduction during Li powder injection in NSTX and EAST (Maingi et al 2018 Nucl. Fusion 58 024003), reduced recycling due to hydrogenic species retention is unnecessary for the ELM suppression with B powder injection, paving the way for its consideration as an ELM control tool for future fusion devices.

A decrease in the L-H transition power threshold P L-H under the lithium wall conditioning has been observed since 2010 in EAST. The physical mechanism responsible for the reduced P L-H is found to be related to three phases of wall recycling as determined by the accumulation of lithium in the vessel. In phase I (a few rounds of wall conditioning) the wall recycling is low, while no clear variation of L-H transition power threshold is observed. In phase II (after another few rounds of wall conditioning) a clear increase in the wall recycling is observed, leading to decreased L-H transition power threshold. In phase III the wall recycling is higher and the hydrogen concentration n H / ( n H + n D ) is lower compared to phases I and II. In this third phase the L-H transition power threshold is mainly reduced by the decreased n H / ( n H + n D ) . During phase III of the wall recycling, the quantitative relationship between the hydrogen content and normalized L-H transition power threshold P L-H / P s c a l e is studied for campaigns in 2010, 2012, 2015 and 2016. Data scaling reveals that P L − H / P s c a l e = ( − 1.513 ± 0.135 ) × e ( − 6.989 ± 0.362 ) × n H / ( n H + n D ) + ( 1.698 ± 0.050 ) in the hydrogen concentration range of 3–45%, i.e. P L-H / P s c a l e increases significantly for 3 % < n H / ( n H + n D ) < 20 % , while it increases mildly for 20 % < n H / ( n H + n D ) < 45 % .

Significant engineering and physic progress was made by the upgrade of a flowing liquid lithium limiter (FLiLi) inserted into EAST H-mode discharges in 2016. The progress includes an improvement of liquid Li coverage uniformity, no macroscopic surface erosion by plasma materials interaction, effective removal of 55% of the discharge heat flux using high pressure He cooling during ohmic discharges, and the demonstrated ability to restart Li flow for a second experiment performed a week after the first. With increasing Li flow rate, fuel particle recycling and Fe impurity radiation gradually decreased, and plasma performance slightly improved, but there was no impact on the edge tungsten influx. No obvious Li bursts appeared during operation with up to about 4.5 MW auxiliary heating power with peak heat flux reaching ~4 MW m−2 on FLiLi. In addition, a gradual mitigation of edge localized mode (ELM) activity was observed with FLiLi operation. Finally, short-lived ELM-free phases were observed for the first time in EAST during FLiLi insertion, with increasing τ E and transient H98(y, 2) < 2.

The ability of an injected lithium granule to promptly trigger an edge localized mode (ELM) has been established in multiple experiments. By horizontally injecting granules ranging in diameter from 200 microns to 1 mm in diameter into the low field side of EAST H-mode discharges we have determined that granules with diameter >600 microns are successful in triggering ELMs more than 95% of the time. It was also demonstrated that below 600 microns the triggering efficiency decreased roughly with granule size. Granules were radially injected from the outer midplane with velocities ~80 m s−1 into EAST upper single null discharges with an ITER like tungsten monoblock divertor. These granules were individually tracked throughout their injection cycle in order to determine their efficacy at triggering an ELM. For those granules of sufficient size, ELM triggering was a prompt response to granule injection. By simulating the granule injection with an experimentally benchmarked neutral gas shielding (NGS) model, the ablatant mass deposition required to promptly trigger an ELM is calculated and the fractional mass deposition is determined.

We report the first successful use of lithium (Li) to eliminate edge-localized modes (ELMs) with tungsten divertor plasma-facing components in the EAST device. Li powder injected into the scrape-off layer of the tungsten upper divertor successfully eliminated ELMs for 3–5 s in EAST. The ELM elimination became progressively more effective in consecutive discharges at constant lithium delivery rates, and the divertor D α baseline emission was reduced, both signatures of improved wall conditioning. A modest decrease in stored energy and normalized energy confinement was also observed, but the confinement relative to H98 remained well above 1, extending the previous ELM elimination results via Li injection into the lower carbon divertor in EAST (Hu et al 2015 Phys. Rev. Lett. 114 055001). These results can be compared with recent observations with lithium pellets in ASDEX-Upgrade that failed to mitigate ELMs (Lang et al 2017 Nucl. Fusion 57 016030), highlighting one comparative advantage of continuous powder injection for real-time ELM elimination.

A new flowing liquid Li limiter (FLiLi) based on the concept of a thin flowing film has been successfully designed and tested in the EAST device in 2014. A bright Li radiative mantle at the plasma edge was observed during discharges using FLiLi, resulting from passive Li injection and transport in the scrape-off layer (SOL) plasma. Li particle efflux from the FLiLi surface into the plasma was estimated at >5 × 1020 atom s−1, due to surface evaporation and sputtering, and accompanied with a few small Li droplets ~1 mm diameter that were ejected from FLiLi. The Li efflux from FLiLi was ionized by the SOL plasma and formed a Li radiation band that originated from the FLiLi surface, and then spread toroidally by SOL plasma flow. The Li radiative mantle appeared to partly isolate the plasma from the wall, reducing impurity release from the wall materials, and possibly leading to a modest improvement in confinement. In addition, strong Li radiation reduced the particle and heat fluxes impacting onto the divertor plate, with certain similarities to heat flux reduction and detachment onset via low-Z impurity injection.

As an alternative choice of solid plasma facing components (PFCs), flowing liquid lithium can serve as a limiter or divertor PFC and offers a self-healing surface with acceptable heat removal and good impurity control. Such a system could improve plasma performance, and therefore be attractive for future fusion devices. Recently, a continuously flowing liquid lithium (FLiLi) limiter has been successfully designed and tested in the EAST superconducting tokamak. A circulating lithium layer with a thickness of 3 s−1 was achieved. A novel in-vessel electro-magnetic pump, working with the toroidal magnetic field of the EAST device, was reliable to control the lithium flow speed. The flowing liquid limiter was found to be fully compatible with various plasma scenarios, including high confinement mode plasmas heated by lower hybrid waves or by neutral beam injection. It was also found that the controllable lithium emission from the limiter was beneficial for the reduction of recycling and impurities, for the reduction of divertor heat flux, and in certain cases, for the improvement of plasma stored energy, which bodes well application for the use of flowing liquid lithium PFCs in future fusion devices.

The first results of edge-localized mode (ELM) pacing using small spherical lithium granules injected mechanically into H-mode discharges are reported. Triggering of ELMs was accomplished using a simple rotating impeller to inject sub-millimetre size granules at speeds of a few tens of meters per second into the outer midplane of the EAST fusion device. During the injection phase, ELMs were triggered with near 100% efficiency and the amplitude of the induced ELMs as measured by Dα was clearly reduced compared to contemporaneous naturally occurring ELMs. In addition, a wide range of granule penetration depths was observed. Moreover, a substantial fraction of the injected granules appeared to penetrate up to 50% deeper than the 3 cm nominal EAST H-mode pedestal width. The observed granule penetration was, however, less deep than suggested by ablation modelling carried out after the experiment. The observation that ELMs can be triggered using the injection of something other than frozen hydrogenic pellets allows for the contemplation of lithium or beryllium-based ELM pace-making on future fusion devices. This change in triggering paradigm would allow for the decoupling of the ELM-triggering process from the plasma-fuelling process which is currently a limitation on the performance of hydrogen-based ELM mitigation by injected pellets.