Experimental characteristics of lost fast negative ions on EAST tokamak

PAPER • OPEN ACCESS Experimental characteristics of lost fast negative ions on EAST tokamak To cite this article: Z.X. Zhang et al 2025 Nucl. Fusion 65 046001 View the article online for updates and enhancements. You may also like EDA H-mode in ASDEX Upgrade: scans of heating power, fueling, and plasma current L. Gil, T. Pütterich, C. Silva et al. - Low frequency m = 1 modes during standard and improved confinement scenarios in W7-X Dario Cipciar, Carsten Killer, Christian Brandt et al. - Simulation and analysis of a high-k electron-scale turbulence diagnostic for MAST-U D.C. Speirs, J. Ruiz Ruiz, M. Giacomin et al. - This content was downloaded from IP address 193.147.173.203 on 24/04/2025 at 10:24 International Atomic Energy Agency Nuclear Fusion Nucl. Fusion 65 (2025) 046001 (12pp) https://doi.org/10.1088/1741-4326/adb7ee Experimental characteristics of lost fast negative ions on EAST tokamak Z.X. Zhang1,2, J. Huang1,∗, J.F. Chang1,∗, C.R. Wu3, J. Galdon-Quiroga4, A. Snicker5,6, M. Isobe7,8, K. Ogawa7,8, K.Y. He1, S.S. Wang9, Y.X. Sun1,2, X.H. Wang1,2 and the EAST Team1,a 1Institute of Plasma Physics, Hefei Institutes of Physical Science Chinese Academy of Sciences, Hefei 230031, China 2University of Science and Technology of China, Hefei 230026, China 3Shiyan Key Laboratory of Quantum Information and Precision Optics, School of Mathematics, Physics and Optoelectronic Engineering, Hubei University of Automotive Technology, Shiyan 442002, China 4Department of Atomic, Molecular and Nuclear Physics, University of Seville, E-41012 Seville, Spain 5Department of Applied Physics, Aalto University, PO Box 11100, Helsinki FI-00076 AALTO, Finland 6VTT Technical Research Centre of Finland Ltd, Kivimiehentie 3, Espoo, Finland 7National Institute for Fusion Science, National Institutes of Natural Sciences, Toki 509-5292, Japan 8SOKENDAI (The Graduate University for Advanced Studies), Toki 509-5292, Japan 9Qingdao Sifang SRI Intellectual Technology Co., Ltd, Qingdao 266031, China E-mail: [email protected] and [email protected] Received 12 March 2024, revised 29 November 2024 Accepted for publication 19 February 2025 Published 3 March 2025 Abstract It is the first time that fast negative-ion loss signals are observed by Fast-ion Loss Detector (FILD) and analyzed in EAST neutral beam injection experiment. Velocity-space distributions of lost fast negative and positive ions reconstructed by the FILDSIM code are basically consistent. Orbit analysis backtracked by the ASCOT code shows that the fast negative ions are generated at the low-field side (LFS) scrape-off layer, whereas the fast positive ions are generated near the last closed flux surface (LCFS) at both the highand LFSs. The existence of fast negative-ion loss indirectly suggests a very high neutral particle density near both the main limiter and antenna limiter. The differing birth locations of fast positive and negative ions result in distinct loss behaviors during ELM H-mode and n=3 rotating RMP experiments. The asymmetry of neutral particle density in tokamak suggest that previous studies may have underestimated the fraction of charge exchange (CX) losses due to the assumption of a uniform neutral particle density outside the LCFS. Therefore, simulations evaluating the CX loss fraction in the future should account for a three-dimensional distribution of neutral particle density, including local gas puffing and the neutral particle density around various limiters. Keywords: fast negative-ion loss, EAST tokamak, FILD, velocity-space distribution (Some figures may appear in colour only in the online journal) aSee Gong et al 2024 (https://doi.org/10.1088/1741-4326/ad4270) for the EAST Team. ∗Authors to whom any correspondence should be addressed. Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 1741-4326/25/046001+12$33.00 Printed in the UK 1 © 2025 The Author(s). Published by IOP Publishing Ltd on behalf of the IAEA Nucl. Fusion 65 (2025) 046001 Z.X. Zhang et al 1. Introduction Good confinement of fast ions plays an important role in fusion power generation for future fusion reactors, such as ITER and CFETR. However, fast-ion losses can occur due to various issues, such as MHD instabilities [1], resonant magnetic perturbations [2,3], toroidal field ripple [4], and edgelocalized modes (ELMs) [5], etc. Fast-ion losses not only degrade the core performance, but also could cause damage to plasma-facing components (PFCs). Scintillator-based Fast Ion Loss Detector (FILD) is a crucial diagnostic and widely used to investigate fast-ion losses in many magnetic confinement devices, including JET [6], DIII-D [7], ASDEX Upgrade [8], KSTAR [9], NSTX [10], MAST-U [11], HL-2A [12], LHD [13] and W7-X [14]. It can provide both velocity-space distribution (pitch angle and gyroradius) and absolute fluxes of fastion losses [15]. Including the measurement results of the FILD detector, the full-orbit codes such as ASCOT [16] and SPIRAL [17] have been used to track the lost fast ions to understand the loss mechanism [2,18]. FILD has also been applied on EAST to investigate the fast-ion loss behavior in neutral beam injection (NBI) plasma [19], with the correctness of the velocityspace distribution of fast-ion losses analyzed by FILDSIM and ASCOT [20]. In this paper, using FILD the lost fast positive and negative ions are simultaneously detected and distinguished for the first time with the detailed analysis shown in the content. Here, the generation of fast negative ions may be attributed to charge exchange (CX) reactions. Fast positive ions can undergo a CX reaction with background neutrals, thereby generating fast neutrals (D+ f+M→D0 f+M+, where D+ frepresents fast positive ions, Mdenotes background neutrals, and D0 frefers to fast neutrals) [21]. The fast neutrals undergo second-step CX reactions either with background charged particles to become fast positive ions (D0 f+D+ s→D+ f+D0 s, where D+ srepresents background charged particles) [22–26], or with background neutral particles to generate fast negative ions (D0 f+M→ D− f+M+, where D− frepresents fast negative ions). The multistep CX process responsible for generating the fast negative ions may be analogous to the multi-step interactions between solar wind and background neutrals from planets and comets in space physics [27,28]. The paper is organized as follows: section 2describes the experimental setup and the principle of observing the loss of fast negative ions by FILD. Section 3presents the velocityspace distribution of the lost fast negative ions observed in the NBI experiment. Section 4presents the birth location and origin of fast positive and negative ions determined through orbit analysis. Section 5discusses the influence of the birth location of fast positive and negative ions on their loss behavior in the experiments. Finally, in section 6we present conclusions and future work. 2. Experimental setup and methods EAST is equipped with four neutral beam injectors (NBIs) which use deuterium as the working gas for auxiliary heating Figure 1. Layout of the auxiliary heating system on EAST, showing the relative position of the FILD diagnostic (top view). [29]. The NBI system consists of beamlines NBI1 located at the A port, and NBI2 located at the F port, as shown in figure 1. Beamlines NBI1R and NBI2L provide perpendicular injections in the co-Ipand ctr-Ipdirections, respectively, while NBI1L and NBI2R are tangential injections in the coIpand ctr-Ipdirections, respectively. The main limiter (ML) is installed between the G and H ports to prevent contact between the core plasma and the vacuum wall. Additionally, limiters are installed on the horizontal sides of the LHW antennas (E and N ports) and the ICRF antennas (B and I ports) to shield the antennas from excessive thermal loads. The scintillator-based FILD is used to detect lost fast ions and is installed on the J port of the EAST tokamak [19]. In the poloidal cross-section, the FILD is located near the EAST midplane at z∼0.224 m. The FILD consists of six main components: the exchange box, long shaft, probe head, relay optical system (ROS), CCD system, and PMT system, as illustrated in figure 2(a). A vacuum isolation valve is positioned between the exchange box and the port flange, allowing the scintillators in the probe to be replaced without compromising the integrity of the primary vacuum. The long shaft connects the probe to the drive motors, which include both linear and rotary types that control the radial movement (>2 m) and rotational motion around its axis, respectively. The probe is primarily composed of a stainless-steel shield, a scintillator (4 cm ×4 cm, ZnS:Ag, decay time: 70 ns, emission peak: 450 nm), and collimators. The scintillator is mounted on the rear side of the stainlesssteel shield, and the probe is blackened to minimize the impact of internal surface reflections on the signal. In order to enhance the collection efficiency of the light emission of scintillator, a ROS is employed. The optical system includes seven lenses, three of which are field lens groups. The effective focal length of the ROS is 358.4 mm, with a distance of 680 mm between the scintillator plane and the first lens. The light emission pattern of scintillator is transmitted through the optical path in the long shaft to the optical splitter, where half of the light 2 Nucl. Fusion 65 (2025) 046001 Z.X. Zhang et al Figure 2. (a) Schematic of the FILD diagnostic on EAST, showing the relay optical system, which includes optical 1, optical 2, optical 3, and the optical splitter. (b) Correspondence between the CCD image of the light emission plane of scintillator and the 25 optical fiber arrays. Figure 3. Fast positive ions (in green lines) and negative ions (in red lines) in the rev-Btoperating regime. (a) Schematic of the entry of fast ions into the FILD probe head through the collimator (in blue). (b) Illustration of fast ions strike the scintillator (in white). is reflected to the CCD system and the other half is directed to the PMT system. The CCD system uses a high-speed CMOS type Phantom V2010-96GB camera (1280 ×800 pixels) from Vision Research, Inc. The PMT system employs a multichannel PMT array (Hamamatsu Photonics, No. H10492003, sampling rate: 2 MHz) to synchronously collect signals, with each data acquisition system operating independently. The light emission pattern of scintillator is collected by a uniformly distributed array of 25 optical fibers (5 ×5), which are then transmitted to each PMT subsystem [30]. The correspondence between the light emission pattern of scintillator and the 25 optical fiber array is illustrated in figure 2(b). The FILD has a probe head geometry with top and bottom collimators, allowing it to detect both lost fast positive and negative ions. As shown in figures 3(a) and (b), in the rev-Bt(with Ipand Btdirected counterclockwise from the top view) operating regime [31], lost fast negative ions are expected to strike the scintillator in the FILD probe head through the top collimator, while fast positive ions are expected to strike through the bottom collimator. In contrast, lost fast negative ions are expected to strike the scintillator through the bottom collimator, while fast positive ions are expected to strike through the top collimator in the nor-Btoperating regime (with Ipdirected counterclockwise and Btdirected clockwise from the top view), respectively. 3. Characteristics of lost fast negative ions Figures 4(a)–(c) show the time evolution of the line-averaged density, NBI power, and gap-out (radial width of scrape-off layer (SOL)) for discharges #93723 and #94994 on EAST. The photomultipliers (PMTs) signal of the FILD shown in figures 4(d) and (e), which can be used to detect fast ion loss signals at different positions on the scintillator. It can be observed that fast ion loss signals are generated only during the NBI injection. Fast negative-ion losses have been observed in a variety of plasmas, covering a wide range of parameters, including plasma density (from 2.5 ×1019 m−3 to 5 ×1019 m−3), plasma current (from 0.4 MA to 0.5 MA), and gap-out (from 4.5 cm to 7 cm). Notably, the intensity of fast positive and negative ion loss signals shows a weak positive correlation with electron density in discharge #93723. Furthermore, the fast positive and negative ion loss signals exhibit negative and positive correlations, respectively, with the radial width of the SOL, particularly at the power step (t∼3.51 s). A detailed discussion is provided in section 5. Scintillator signals from the FILD were extracted at two time slices from two different shots for the analysis of fast ion loss characteristics. The velocity-space distribution of fast ion loss was obtained using FILDSIM code [32] which incorporates orbit trajectories within the geometry of the FILD probe head. The FILDSIM code maps the velocityspace grid onto the FILD signals. FILD observed two main populations of beam ion losses during NBI2L injection, as shown in figure 5(a). The results indicate that the pitch (arccos(v///v)∼68◦) and gyroradius (RL∼3 cm) are nearly identical for both populations of beam-ion losses. The beamion population in the upper half of the scintillator corresponds to fast negative-ion losses, while the population in the lower half corresponds to fast positive-ion losses. FILD observed three distinct beam-ion loss populations during NBI1R injection, as shown in figure 5(b). One of the fast negative-ion populations is consistent with the velocity-space distribution of fast positive-ion losses, as determined by pitch (arccos(- v///v)∼70◦) and gyroradius (RL∼2.5 cm). Another fast negative-ion population with a pitch of ∼45◦is expected to correspond to the loss distribution of fast positive ions with the same velocity-space distribution. However, fast positive ions with pitch of ∼45◦cannot be detected, as they are outside the range of the FILD. 3 Nucl. Fusion 65 (2025) 046001 Z.X. Zhang et al Figure 4. Time evolution for discharges #94994 and #93723 of (a) the line-averaged density, (b) NBI power, (c) gap-out (radial width of SOL), (d) fast positive-ion loss and (e) fast negative-ion loss. Figures 5(c)–(f) show the velocity-space distribution of fast negative and positive ions, obtained by performing triangular interpolation on the strike map. The vertical axis is converted from gyroradius RLto beam energy Eusing E=(e·Z·RL·BFILD)2/(2·MD), where Zis the ion charge number, BFILD is the magnetic field at the FILD probe head, and MDis the mass of the Datom. As shown in figures 5(c) and (e), the energy of the lost fast negative and positive ions is approximately 50 keV, which is consistent with the main energy of the NBI2L beamline. Figures 5(d) and (f) show that the energy of the lost fast negative ions with a pitch angle of 45◦matches the main energy of the NBI1R beamline, while the energy of the lost fast positive and negative ions with a pitch angle of 70◦is approximately 65 keV. Additionally, the fast ion energy in figures 5(d) and (f) extends up to nearly 100 keV, due to the limited size of the FILD collimator, which restricts the resolution of the detector in the Larmor radius [20,32]. The distribution measured on the scintillator can be considered a distortion of the velocity space, resulting from the finite resolution of the detector aperture. Fast ion loss signals from the NBI1R beamline may include energy components of E/2 and E/3, but these cannot be distinguished by the FILD probe head. 4. Origin and birth positions of fast negative ions 4.1. Initial deposition locations of lost beam ions The origin of the lost beam ions can be determined by analyzing the characteristics of the fast ions. The characteristic time for the first-orbit fast-ion losses is on the order of ∼10−5s (by recording the time taken for fast ion losses to the first wall using orbit-following code) in EAST. In this timescale, collisions between background and fast ions can be neglected (τD∼10−3s,ne∼1019m−3,Te∼1keV, τDis deuterium ion collision time [33]). Therefore, energy can be treated as a conserved quantity [34]. Moreover, the gyromotion of fast ions follows the conservation of magnetic moment due to the slow variation of magnetic fields both in time and space. Since both energy and magnetic moment are conserved during the gyromotion of the fast ions, the relationship between the pitch θof the beam-ions losses and the radial initial deposition position Rcan be derived as follows: θ(R) = arctan√1 E µ(R)BFILD −1.(1) Eis the neutral beam energy (eV). µ(R) is the distribution of the magnetic moment along the radial initial deposition position of the beam lines, which can be calculated by considering the geometry and magnetic field distribution of the beamline. BFILD is the magnetic field strength (T) at the location of the FILD probe head. Based on equation (1), the initial deposition position of the neutral beam can be determined by observing the velocityspace distribution using the FILD. Figure 6(a) shows the relationship between the initial deposition position of the neutral beam from different injectors on EAST and the pitch angle of beam-ion losses. It can be seen that beam-ion losses originate from the initial deposition of the beamlines near the last closed flux surface (LCFS). Notably, beam ions with pitch angles of ∼45◦and ∼70◦, generated by NBI1R, originate from the initial deposition of neutral beam on the high-field side (HFS) and low-field side (LFS) of the SOL, respectively. However, the calculation in equation (1) does not account for the sign of the of beam ions charge, and thus it can only indicate that the origin of fast negative ions is the neutral beam deposited near the LCFS. Further analysis using full-orbit following is required to determine the specific locations where the fast negative ions are generated. Based on the geometry and beam energy of different beamlines, the initial pitch angle of beam ions can be determined as a function of the initial deposition location of the neutral beams. The initial pitch of the lost beam ion can be obtained by calculating the radial deposition locations of different loss beam-ions using equation (1), as shown in figure 6(b). The initial pitch of the lost beam ions generated by NBI2L is 73.98 ◦, and the initial pitch of the lost beam ions generated by NBI1R are 61.48 ◦(HFS) and 70.79 ◦(LFS), respectively. 4 Nucl. Fusion 65 (2025) 046001 Z.X. Zhang et al Figure 5. Strike map of beam-ion losses measured by FILD probe head with (a) #[email protected] s (b) #[email protected] s. (c) and (e) Correspond to the velocity-space distribution of lost fast positive and negative ions from NBI2L. (d) and (f) Represent the velocity-space distribution of lost fast positive and negative ions from NBI1R. 4.2. Comparison of trajectories of fast positive and negative ions In order to investigate the birth location of fast negative ions, their trajectories are tracked backwards based on their velocity-space distribution using ASCOT [16], which is a fullorbit following code using a Monte Carlo approach. In the equilibrium magnetic field, fast negative and positive ions can be followed backwards in time from the FILD probe position, then the trajectories and initial birth positions of fast positive and negative ions can be attained. Markers originating from the respective beam lines were fixed with a single pitch angle and main energy, as shown in figure 5(c), where the orbits are followed backwards in time from the FILD position until they overlap with the corresponding beamlines. Figures 7(a) and (c) show that the fast positive ions deposited by NBI2L at R∼2.2 m are rapidly lost to the FILD probe head. However, for fast negative ions, reverse-following from the FILD probe position only backtraces it to the first wall of the device, as shown in figures 7(b) and (d). The significant difference in the loss trajectories between fast positive and negative ions is due to the fact that the direction of the magnetic field gradient drift and magnetic field curvature drift velocity depends on the charge sign of the particles. It is clear that fast negative ions cannot originate from the first wall, meaning that they must be produced at a location between the first wall strike point and the FILD probe position. Figure 7(b) illustrates a typical trajectory of fast positive ions transitioning into fast negative ions. The fast positive ions (in green line, R=2.2 m, pitch =73.36◦) motion toward the SOL, where they undergo CX reactions with background neutral particles at a certain position (in red circle), acquiring two electrons and transforming into fast negative ions (in blue line). Ultimately, the fast negative ions are lost to the FILD probe. By varying the CX location for fast positive ions and subsequently following their trajectory until overlapping with the beamline, the approximate toroidal birth locations of fast negative ion were determined to be between φ∼309◦and φ∼21◦, as shown in figure 7(d). 4.3. Comparison of fast negative ion trajectories originating from HFS and LFS To illustrate the process by which neutral beam, initially deposited on the HFS, is converted into fast negative ions lost 5 Nucl. Fusion 65 (2025) 046001 Z.X. Zhang et al Figure 6. (a) The pitch of the beam ion losses and (b) beam-ion initial pitch, originated by different neutral beam injectors as a function of the initial deposition radial position of the neutral beam. to the FILD probe head, an orbit analysis of fast negative ion loss originating by NBI1R was conducted in the nor-Btoperating regime. The population of fast negative ion loss observed by the FILD with a pitch angle of 45◦and an energy of 65 keV originates from the initial deposition of neutral beam inside the HFS LCFS, with the birth position located outside the LFS LCFS, as shown in figure 8(a). Figure 8(c) illustrates the toroidal birth positions of fast negative ions with a pitch angle of 45◦, which are approximately between φ∼297◦and φ∼19◦. Figure 8(b) shows the orbits of fast negative ion loss with a pitch angle of 70◦and energy of 65 keV, indicating that they originate from the initial deposition of neutral beam outside the LFS LCFS. The approximate toroidal birth locations of fast negative ions with a pitch angle of 70◦were determined to be between φ∼297◦and φ∼18◦, as shown in figure 8(d). The birth positions of the fast negative ions are also located outside the LFS LCFS, but the toroidal birth positions are closer to the FILD toroidal position. Thus, it is imperative to consider the initial deposition of neutral beam in the SOL in order to accurately validate the velocity-space distribution of both fast positive and negative ions measured by the FILD. It should be noted that the potential birth locations of fast negative ions generated by NBI1R are confined to a specific region between the FILD probe position and the first wall strike point (such as the division of the fast negative ion trajectory into blue and gray lines in figure 8(a)). This limitation arises from the fact that when the CX position of the fast positive ions falls within the gray line region, they are unable to follow to the position overlapping with the beamline. 5. Discussion When backtracking to the position where the fast ions overlap with the beamline, the determined initial deposition positions and pitch angles are generally consistent with those calculated based on the conservation of energy and magnetic moment, further validating the accuracy of the orbit analysis. This also implies that regardless of the nor-Btor rev-Btoperating regimes, or whether the fast negative ions originate from the HFS or the LFS, the FILD can only observe fast negative ions birth in the SOL near the FILD probe head (i.e. ML (φ ∼325◦) and ICRF limiter (φ∼350◦and φ∼10◦) in EAST). The possible fact that the limiters are located closer to the LCFS result in higher neutral particle density near them. As described in section 3, figure 4demonstrates that the intensity of fast positive-ion loss signals shows a positive correlation with electron density and a negative correlation with the radial width of the SOL in discharge #93723, which can be attributed to two main factors. Firstly, lower electron density leads to a higher fraction of neutral beam deposition in the core, thereby reducing the fast positive-ion loss fraction [29]. Secondly, the increasing of fast positive-ion losses may be attributed to the reduction of SOL radial width and the simultaneous NBIs (NBI1R +NBI1L) at t=3.51 s, with the former bringing fast positive ions closer to the PFCs near the LFS LCFS and the later increasing the density of fast ions directly. On the contrary, the intensity of fast negative-ion loss signals is positively correlated with both electron density and the radial width of the SOL. Figures 9(a) and (b) show the cross sections for the reaction processes that generate fast negative ions (the ‘source’ processes) and those that annihilate fast negative ions (the ‘sink’ processes), respectively, as listed in table 1. The reaction rates (R) of the ‘source’ and ‘sink’ processes are positively correlated with the densities (n) of the reacting species. A decrease in plasma density weakens the ‘source’ process, as the density of fast positive ions in the SOL decreases. The reduction in the radial width of SOL leads to the motion of fast negative ions into the LCFS, thereby enhancing the ‘sink’ process due to increased background ion density. Furthermore, the reduction in the radial width of SOL results in more fast positive ions striking the limiters, which strengthens local particle recycling and enhances simultaneously both the ‘source’ and ‘sink’ processes due to increased local neutral particle density. As shown in figure 9, the cross section for the ‘sink’ process (reaction 8) is significantly higher than those of other ‘sink’ processes and the ‘source’ processes. The results suggest that the ‘sink’ process (reaction 8) is the primary cause of the observed reduction in fast negative-ion loss intensity, particularly at power step (t=3.51 s). In conclusion, the intensity of fast negative-ion loss is governed by the competition between ‘source’ and ‘sink’ processes, and the ‘sink’ processes seem dominant at power step. The effects of SOL radial width and plasma density on fast negative-ion losses will be further systematically investigated in future experiments. 6 Nucl. Fusion 65 (2025) 046001 Z.X. Zhang et al Figure 7. Typical trajectories of fast positive and negative ions originating from NBI2L. (a) Schematic of the trajectory of fast positive ions observed by FILD. (b) Schematic of the trajectory where a fast positive ion in the SOL converts into a fast negative ion observed by FILD. The green and blue lines indicate the trajectories of the fast positive and negative ions, respectively. The red circle indicates the position where fast positive ions become fast negative ions. (c) Is top-view projection of the trajectories shown in (a). (d) Is top-view projection of the trajectories shown in (b). Gray areas indicate possible birth regions of fast negative ions. The lost fast positive ions observed by the FILD originate near the high and low field side of the LCFS, while the lost fast negative ions are generated exclusively in the low field side SOL through interactions with the lost fast positive ions and neutral particles. The different birth locations result in distinct loss behaviors of fast positive and negative ions in ELMy H-mode experiments. Figures 10(a)–(c) illustrate the time trace of Dαemission, line-averaged density and fast ion losses for discharges #93723 at 4.6–4.7 s. In this period, q95 =6.35, fELM ∼160 Hz and βN∼1.0. In figure 10(a), line-averaged density is observed to drop at each ELM crash. It was observed that each ELM crash leads to an enhancement in fast positive ion losses, as shown in figures 10(b) and (d). This phenomenon can be explained as the ELM crash causing edge-pedestal collapse, further resulting in the transport of fast positive ions from the edge-pedestal region (near the FILD probe head) into the SOL [36–38]. Consequently, the FILD probe in the SOL detects an enhanced loss signal of fast positive ions. However, ELM crash leads to a reduction in fast negative ion losses, as shown in figures 10(c) and (f). Orbit analysis indicates that the observed fast negative ions are generated near the limiters on the LFS, outside the LCFS. During an ELM crash, fast positive ions and other charged particles in the edge-pedestal region (near the limiters) are directly lost to the limiters. Consequently, the number of fast positive ions available to generate fast negative ions near the limiters sharply decreases (sudden weakening of the ‘source’ processes). The reduction results in a momentary reduction of the fast negative ion loss signal detected by the FILD probe. Before the next ELM crash, sputtering of the first wall by fast 7 Nucl. Fusion 65 (2025) 046001 Z.X. Zhang et al Figure 8. Typical trajectories of fast negative ions originating from NBI1R. (a) Schematic of the trajectory where fast positive ions deposited on the HFS convert into fast negative ions in the SOL. Gray line indicates the trajectories in which the fast negative ions will not be birth, and the same as below. (b) Schematic of the trajectory where fast positive ions deposited on the LFS convert into fast negative ions in the SOL. (c) Is top-view projection of the trajectories shown in (a). (d) Top-view projection of the trajectories shown in (b). Figure 9. (a) Cross-section for charge exchange and electron capture (‘sources’) in table 1. (b) Cross-section for electron stripping (‘sinks’) in table 1. Cross section as a function of the relative particle energy per nucleon is identical for the reaction processes of hydrogen isotopes. A typical 60 keV EAST deuterium beam with its full, half, and third energy components are indicated. 8