- Email: [email protected]
Wall conditioning towards the utilization in ITER
Journal of Nuclear Materials 415 (2011) S35–S41
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Wall conditioning towards the utilization in ITER J. Li a,⇑, M. Shimada b, Y. Zhao a, J. Hu a, X. Gong a, Y.W. Yu a, G.Z. Zhuo a a b
Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, China ITER Organization, St Paul Lez Durance, France
a r t i c l e
i n f o
Article history: Available online 2 November 2010
a b s t r a c t Wall conditioning provides an effective means for reducing both impurities and recycling from the plasma surrounding surface. A wide variety of techniques have been developed during the last few decades for conditioning the plasma facing surface. With the presence of magnetic fields, electron cyclotron resonant (ECR) and ion cyclotron resonant (ICR) discharge cleaning techniques have been explored, which could be used for next generation superconducting magnetic confined devices, such as ITER. Efforts have been made on the application of ICR conditioning on many devices and significant progress has been made. A new and simple method for future wall conditioning, high frequency glow discharge cleaning (HF-GDC), has been developed. HF-GDC operates in the presence of strong magnetic field (0.5–2 T) at frequencies of 20–100 kHz stably for a wide range of gas pressures. In this paper, all these techniques are reviewed and their proper application in ITER are discussed. Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction Wall conditioning provides an effective means for reducing both impurities, and recycling from the plasma surrounding surface. The experience in present short-pulse tokamaks already indicates that wall conditioning plays a crucial role in achievement of reproducible and clean plasma conditions and improvement on overall plasma performance [1–5]. Some long pulse experiments suggest that the impact of wall condition on the plasma performance could be more serious in ITER with plasma discharge time from 400 to 3000 s. Furthermore, ITER, as a nuclear device, must control the in-vessel inventory of tritium and dust for safety reason. (1) The objectives of wall conditioning are: reduction of gaseous impurities influx from in-vessel components (the PlasmaFacing Component surfaces and surfaces inaccessible from the plasma) to maintain/improve plasma purity. (2) Reduction of hydrogenic species influx from in-vessel components to facilitate density control. (3) Removal of tritium from in-vessel components. (4) Removal of dust from Plasma-Facing Components. A wide variety of techniques have been developed during last few decades. Baking of the internal components inside vacuum vessel, glow discharge cleaning (GDC) [6,7], Taylor discharge clean⇑ Corresponding author. Address: Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, Anhui, 230031, China. Tel.: +86 551 5591371; fax: +86 551 5591310. E-mail address: [email protected] (J. Li). 0022-3115/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jnucmat.2010.10.048
ing (TDC) [8], which were developed in the early days of tokamak experiments, have been demonstrated to be effective, and are still in use today. With the presence of magnetic fields, electron cyclotron resonant (ECR) and ion cyclotron resonant (ICR) discharge cleaning techniques have been explored, which could be used for next generation tokomaks, such as ITER. Wall conditioning for above mentioned issues has been studied extensively for past few decades and many papers, especially many excellent review papers have been published [3,4,9–14]. The principle of all the cleaning procedures using plasma discharges is to reduce metal or carbon oxides by hydrogenic plasma bombardment to form volatile species, thus depleting the contamination layers on the wall surfaces. These volatile substances, such as water vapour, methane and other hydrocarbons, are then desorbed from the surface, both thermally and by particle impact, and ultimately evacuated from the device via the torus pumping system. Two key factors for achieving a large cleaning effect are the maximum volatile substance formation by plasma discharge, and the effective exhaust of the formed impurities before they are re-ionized, so that these impurity products are not re-ionized, dissociated and redeposited on the surfaces in the cleaning plasma itself. This requires plasmas to have low Te and a low ionization fraction. In addition, high wall temperatures are useful not only for increasing the rates of the reactions leading to the formation of the volatile species on the surface, but also for promoting their desorption from the surface. Furthermore, a large pumping speed is desired for efficient removal of the volatile species from the torus. The essential difference between various conditioning methods lies in the way the atomic hydrogen species are created and interact with the surface. Other effects such as charge exchange
S36
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
neutrals, and sputtering due to high-energy ion bombardment may also play a role. After introduction, GDC will be described in Section 2. ECR and ICR wall conditioning will be given in Sections 3 and 4. A new technique, high frequency glow discharge cleaning will be described in Section 5. Other techniques, such as strike point sweeping, flash lamp and oxidation for effective moving T retention will be given in Section 6. Discussion and summary for future needs and application to ITER will be discussed in last section of this paper.
2. Glow discharge cleaning The most effective method actually used for surface impurity cleaning is hydrogen or deuterium GDC. One or more anodes are either located in portholes, or moved by a bellows assembly into the main volume of the vacuum vessel. The wall surfaces and limiters are at ground potential. Typical hydrogen pressures, range from 0.1 Pa to a few Pa, the voltage between anode and cathode is of the order of a few hundred volts. Superimposed RF power (100 W) is often used to ease the breakdown of the glow, and to stabilize the plasma at low pressure. During D2-GDC, both ions and charge exchange neutrals interact with the first wall surface, and produce by chemical sputtering molecular impurities such as H2O (D2O), CxHy (CxDy) and CxOy, which are continuously pumped out during the cleaning process. Simultaneous wall baking is very helpful during discharge cleaning, especially for removing oxygen from the wall. Results from TEXTOR showed that a high wall temperature (over 350 °C) is favourable for quick conditioning of first wall during the GDC process [15]. Oxygen and water contamination can be very effectively removed from the vacuum wall by long glow discharges (100 h). He GDC is usually used for devices with a large graphite coverage. The main reason is that large quantities of hydrogen are stored within the penetration range of hydrogen ions. They can be released upon particle impact during a tokamak discharge and make density control difficult, in particular, when the walls cannot be baked to temperatures above 300 °C. For carbon wall, fuel retention is dominated by codeposition. Another reason for fuel retention is the porous structure of graphite that leads to storage of significant amounts of water vapour and hydrogen, rendering recycling control difficult during tokamak discharges. Helium glow discharge conditioning for hydrogen removal has been optimized in DIII-D [16], and routinely used between each tokamak discharge. It has been instrumental in achieving high-performance H-mode plasmas, and a wide operating space in DIII-D where major parts of the surface were covered by graphite. Tritium removal efficiency with H2 and He DC glow has been measured in JT-60U by Nakamura et al. [17]. The hydrogen isotope (tritium and deuterium) release rate by H2 GDC is much larger (by a factor of 2–3) than by He GDC due to chemical processes induced by the hydrogen discharge shown in Fig. 1. The dominant removal processes for the He GDC and H2 GDC after a few hours can be attributed to physical sputtering and isotope exchange reaction assisted by chemical sputtering induced by the H2 discharge, respectively. This graph suggests that the removal rate, or the tritium near the surface, could be reduced by three orders of magnitude with a half day of H2 GDC even at a modest baking temperature of 100 °C. Its drawback for a machine with superconducting magnets like ITER is that it can be used only occasionally e.g. just during the recovery from a vent or major air/water leak or when the magnets are not turned on (the frequency depends on the build-up rate of tritium inventory and the efficiency of other wall cleaning methods and can range from once a year to once a
Fig. 1. Tritium release rates from JT-60U vacuum vessel during: (a) H2 GDC and (b) He GDC at various wall temperature [17].
month). Thus it is excluded from the routine clean-up schemes during operation for a machine like ITER. Another possible drawback is that because the ion energy exceeds the sputtering threshold, erosion of PFCs and coating on the diagnostics mirrors and windows are concern. For ITER wall conditioning, GDC is still required when vent to air or during machine non-operation maintenance when there is no magnetic field. For this reason, GDC is still baseline requirement for ITER [18]. The detail of ITER GDC is shown in Table 1. However, due to the presence of high permanent magnetic fields, conditioning techniques based solely on glow discharges cannot be used. In future superconducting devices, such as ITER, which will be operated over long discharge durations and use deuterium/tritium mixtures, new techniques for wall conditioning are needed, firstly to obtain a reproducible and controlled plasma discharges, and secondly to reduce the tritium wall inventory after each discharge. ECR and ICR conditioning techniques have been developed during the past 20 years, and very good results have been obtained.
Table 1 Glow discharge requirements. Parameters First wall current density Bias voltage Pumping speed (He, molecular flow) Conditioning gas Number of electrodes Operating pressure Coil currents (TF, PF, CS and CC)
Unit 2
A/m V m3/s
Pa A
Value 0.1–0.4 300–600 30–40 H2, D2, He, O2 6 0.01–0.5 0
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
3. ECR conditioning Electron cyclotron resonance (ECR) discharge cleaning was first carried out on JFT-2[19] in 1980s, by using 2.45 GHz LH system under the magnetic field of 0.0875 T, and many experiments were done on JIPPT-II [20], the Gamma-10 tandem mirror [21], Alcator C-Mod [22], LHD [23], TRIAM-1 M [24,25], TOMAS [26], and recently HT-7 [27,28], except for JT-60U by using 1.7 GHz LH system under 0.06 T [29]. ECRF system with the frequency of 110 GHz on Textor-94 was used for ECR discharge cleanings under 1.65–2.28 T [30]. He, H2, D2, Ar and CH4 or CD4 were used as working gas during ECR discharges in the pressure between 1 10 4 and 6 10 2 Pa, for the experimental studies on inactive or non-reactive cleaning, thin film coating and co-deposited layer removal. Fig. 2 shows power densities defined as the injected ECR power normalized to the particle content [28], in different devices. It ranged from 2 to 4 103 kW Pa 1 m 3. Typical electron temperature of ECR plasma was 1–10 eV, and plasma density (1–5) 1016 m 3. ECR plasmas were nearly toroidally homogeneous, but poloidally inhomogeneous. Due to the ExB drift induced by the absence of plasma current and therefore poloidal magnetic field component, ECR plasmas were mainly concentrated on the resonance layer and low field side. A typical picture of ECR plasmas on JIPP T-II is shown in Fig. 3 with the electron density peaking at the resonance layer decreasing towards low field side [20]. Improved homogeneity of ECR plasmas were obtained on JT-60U with 110 GHz ECRH system by applying a horizontal field Bh (0.3–0.5% Bt), and Bh was crucially important to expand ECR plasma toward the high field side [31]. Compared with other wall conditioning methods, ECR discharge cleaning had similar or better effects than TDC on JFT-2 with either H2 or He working gas, and it was effective for performance improvement of both OH and LHCD plasmas on TRIAM-1M. But the removal rate of ECR discharge cleaning was 7–20 times lower than ICRF discharge cleaning on Textor-94 and HT-7. Recent ECR discharge cleaning on JT-60U using 110 GHz ECRH system showed that the H2 outgas efficiency was equivalent to 79% of TDC with the power of 1.27 MW. Pulsed mode with 1 s on and 59 s off was used. It was considered that higher outgassing rate could be attributed due to this duty ratio. One ITER relevant application on JT-60U is that after a disruption of Ip = 1 MA, Bt = 3.5 T, PNBI = 19 MW and the stored energy (W) = 3.3 MJ, the plasma was soon recovered by only one shot ECR cleaning with 2.5 MW of 110 GHz for 1.5 s without change of Bt. Removal of deposit or co-deposited hydrogen isotope by ECR discharge was also studied. Due to the localized ECR plasmas, maximum removal rate was at a few centimetres outside the resonance location on Alcator C-Mod [32]. For the removal of tritium from
Fig. 2. Power densities of ECR discharge cleaning on different devices.
S37
co-deposited layers, laboratory experiments showed that the best conditions for a fast and effective method is oxygen ECR plasma at a low pressure to achieve high ion fluxes and at relatively high wall temperature [26]. ECR plasma position should be scanned by changing magnetic field in order to obtain homogeneous deposition or removal. However, for the future use in ITER with either full field (5.3 T) or half field (2.65 T), the frequency of ECR must be either 170 GHz or 85 GHz with a power of 0.5–1.0 MW. More uniform plasma both in toroidal and poloidal directions should be produced. Significant efforts are still needed towards this goal. The cleaning efficiency of ECR is much lower than that of GDC. This is caused by the extremely localized power deposition area which makes it very difficult to illuminate all surfaces homogeneously.
4. ICR wall conditioning The theory of plasma production in tokamaks using RF in the ICRF range has been investigated by some authors [33,34]. To initiate the discharge, an antenna has to generate an RF electric field E| parallel to the magnetic field lines. The E|| at the antenna is responsible for neutral gas ionization by highly accelerated electrons during gas breakdown. The primary plasma generation is not directly related to IC-resonant layers but caused by RF-breakdown in the strong E.M. field near the antenna. This explains why the experimentally observed dependence on toroidal field at constant frequency is weak. Poloidal ICRF antennas can generate Ez-field (evanescent in vacuum) electrostatically and inductively due to: (i) RF voltage difference between the strap and the sidewalls (RF limiters). (ii) RF voltage induced between the tilted FS bars by the time-varying magnetic flux. To get a higher ionization ratio, the harmonic resonance absorption layers should be inside the plasma. The ion cyclotron frequency x is qB/m, where B is the magnetic field, q the ionic charge and m the mass. Plasma waves (slow wave, ion Bernstein wave, fast wave) excite and propagate over the torus in a relay-race regime governed by the antenna kZ-spectrum. The mechanism of cleaning wall by ICR WC is mainly due to [10]: (a) A massive flux of low energy charge exchange neutrals with fluxes of 1–10 1020 p/m2 s. (b) A low flux of high energy charge exchange neutral hydrogenic atoms fluxes of 1–10 1018 p/m2 s. (c) Flux of ions where the resident flux-tubes cross the wall with fluxes of a few times 1021 p/m2 s, however they are very localized. By using almost the same RF system for heating with an RF power from few tens to few hundreds kW and toroidal magnetic field from 0.2 to 3 T, RF plasma could be easily initiated with a wide range of pressure from 1 10 3 Pa to 0.1 Pa. ICR-WC experiments were performed in Tore Supra [11], TEXTOR [35], HT-7 [36,37], W7-AS [38], ASDEX-U [39], JET [39], KSTAR, LHD and EAST and significant progress has been achieved since 1996. Extensive ICR conditioning has been routinely used in HT-7 and EAST superconducting tokamaks for wall cleaning, recycling and isotopic control and thin film coating (B, Si, Li) by using different working gases. In both HT-7 and EAST, ICR conditioning showed a higher particle removal rates than GDC. The basic RF plasma parameters produced by different antenna configurations were measured in different machines. Quadruple mass spectrometer (QMS) analysis during and after the RF conditioning was routinely carried out. Electron temperature is in the range 3–50 eV. The electron temperature for helium RF plasmas
S38
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
Fig. 3. He ECR plasmas (a) and electron density distribution (b) of ECR plasmas on JIPP T-II tokamak.
is higher than that in deuterium plasmas by a factor of two with the same discharge condition. Electron density is in the range of 0.5–8 1017 m 3. Ion temperature is as high as 0.5–2 keV with high-energy tail of a few tens of keV. The two effects which govern the conditioning efficiency are outgassing rate of particles from the wall and the ionization rate of desorbed molecules which induces a redeposition [3]. For getting higher outgassing rate, a higher particle flux to the wall with suitable energy is favourable. For reducing ionization rate, the pulsed-RF-mode was used. Helium RF discharge is used mainly for hydrogen removal. This makes plasma start up very easy much and allows the control of particle recycling. He-RF discharges are also very effective for the replacement of deuterium by isotopic exchange. The pulsed mode with higher filling pressure (0.05–0.1 Pa) increases the conditioning efficiency by controlling electron density and temperature separately to minimize the hydrogen redeposition probability. The partial pressure of hydrogen was increased by two orders of magnitude after RF plasma was initiated, as a result of the desorption from the wall. High hydrogen removal rates have been obtained. Deuterium ICRF discharge is used for surface cleaning and wall isotope control in HT-7 and EAST tokamaks. The pulsed mode with optimized pressure of 1 10 1 Pa is used. The electron temperature is lower (2–3 eV) compared with helium discharge. Electron temperature changes a little within the injection RF power of 30 kW. The low temperature minimizes the strong plasma reionization and wall redeposition of neutral species. High temperature deuterium ion and its energetic tail are observed. Both energetic and reactive particles bombard the first wall and induce the desorption of molecular impurities. These are main reasons for the good conditioning efficiency of the deuterium RF discharge. It is clearly shown that for future application in ITER, ICR has more favourable features than ECR-WC since it can be easily initiated for relatively uniform plasma without the need for extra hardware, and with higher particle removing efficiency compared with ECR-WC at high density. Even thin film coating is not foreseen in ITER due to possible co-deposited layers which is more difficult to remove, but it is still very useful for getting a quick plasma start-up after minor leakage or disruption. Two possible ways could be RF boronization and RF siliconization. ICR boronization using less hazardous carboranes C2B10H12, was fully developed in HT-7 and EAST, and became a routine coating process [37]. Pulsed RF plasmas with a magnetic field of 2 T and 20 kW of RF power were produced by using a mixture of carborane
(C2B10H12) 50% in helium (50%). The boronization process lasted a few hours and was followed by a He ICR discharge for removing the high hydrogen content in the film. A fine homogeneous and hard a-B/C:H film was produced by pulsed ICRF plasma. The film shows stronger adhesion, greater thickness, and longer lifetime compared with those by GDC method. Good uniformity of the film in both toroidal and poloidal directions has been obtained by using a long antenna on the high field-side of the vessel. The re-emission of hydrogen was easily controlled by using helium ICR discharge after boronization, leading to very strong wall pumping. The plasma performance was significantly improved after boronization with a higher density limit, and wider operation space was obtained. In addition, the hard X-rays accompanied by high power lower hybrid current drive were dramatically suppressed. This gives direct evidence that the thin boron film serves as a protecting layer against the energetic particles, which is very important for future long-pulse operation. RF siliconization was also performed in HT-7 and EAST by using a silane (SiH4) gas mixture (10%) with helium (90%) to facilitate gas handling. The RF discharge was operated in pulsed mode with a 1 s on, 1 s off cycle, and by using the long ICRF antenna side of the vessel. This results in a homogeneous coating in the toroidal and poloidal directions with a film thickness of 50–80 nm. The promising results obtained from ICR plasma-assisted deposition have shown that such techniques are applicable to future superconducting fusion devices with fully W wall after disruption or quick start-up of plasma discharge even the coating techniques are not foreseen in ITER.
5. High frequency glow discharge cleaning Recently, a new method, called High Frequency glow discharge cleaning (HFGDC) with magnetic field, has been invented and worked in both EAST and HT-7 superconducting Tokamaks. Normal electrodes for GDC have been used for this new method with a high frequency power supply. The parameters of HFGDC power supply are: U = 2.0 kV, f25–100 kHz, I10.0 A. Plasma performance for High frequency glow discharges was investigated by reciprocating Langmir probe under the conditions with and without toroidal magnetic field. It shows that the electron temperature of HF_GD plasma with B-field is similar with those of normal DC_GDC without B-field. The cleaning effects are also similar compared with GDC without magnetic field from QMA measurements.
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
Preliminary results showed that plasma can be easily initiated with different gas filling pressure. With higher filling pressure (from 10 10 2 Pa to a few Pa), plasma is much localized around electrodes. When pressure gets lower (from 1 10 3 Pa to 1 10 4 Pa), plasma becomes more uniform not only in toroidal direction but also in poloidal direction. Fig. 4 shows HF-GDC plasma in HT-7 with different filling pressure. It was clearly shown that a uniform plasma was formed with filling pressure between 10 3 and 10 4 Pa. Both helium and deuterium working gases have been used and results are similar. Siliconization technique by using HF_GD was also studied in HT-7. Similar results have been obtained compared with RF-siliconization. Strong impurities radiation caused by a 20 Pa leakage in HT-7 could be suppressed by only 1 h HFGDC with BT = 1.8 T and SiH4 filling gas. Fig. 5 shows two shots before and after HFGDC siliconization.
6. Other methods for tritium removal 6.1. Oxidation The long-term retention of tritium fuel in codeposition layer, plasma facing surface and bulk of materials in fusion devices is one of the major problems. Removal of tritium from amorphous carbon layers will have an important impact on machine operation and safety in the next generation tokamaks, such as ITER. Oxidation treatment is one of promising methods for amorphous tritiated carbon layers removal. Oxidation wall conditioning has been carried out in TFTR [40], TEXTOR [41,42], AUG [43]. Very recently, a systematic experiments of oxidation wall conditioning, including thermo-oxidation, glow discharge (O-GDC) and radio frequency wave associated oxidation (O-ICR), were carried out in a limiter machine – HT-7 [44–46], and in an ITER-relevant full superconducting tokamak – EAST with both metal and full carbon wall conditions. In the O-GDC experiment on HT-7, the removal rates of H-atoms and D-atoms in form of H2O, HDO and D2O were higher than that of H2 and D2 by factors of about 20 and 50, respectively, as shown in Fig. 6. Compared to He-GDC cleaning, O-GDC has higher removal rates for H-atoms and C-atoms by a factor of about 2–4 and about 25, respectively. Average removal rates of 5.2 1022 H-atoms/h, 5.6 1021 D-atoms/h and 5.5 1022 C-atoms/h, respectively, were obtained during 145 min O-GDC cleaning in a pressure range of
S39
0.5–1.5 Pa. However, this procedure led to a significant O contamination. About 5.4 1022 O-atoms were adsorbed on the walls. During a long O-GDC operation in 2005, the removal rates for H, D and C remained very high within 9.5 h. The removal rates for H and C decreased faster than that for D, possibly due to the fact that H and C were desorbed from Stainless Steel and carbon walls whereas D mostly originated from deposits. Continuous or pulsed (0.3 s on and 1.2 s off) O-ICR was performed in the presence of a magnetic field of 1.5–2.0 T on HT-7. The highest removal rates of H, D and C atoms reached to 2.6 1022, 7.8 1021 and 1.5 1022 atoms/h, respectively, during 40 kW 9.0 10 2 Pa pure pulsed O-ICR cleaning shown in Fig. 7. Compared to He-ICRF cleaning, O-ICR has a higher removal rate for H-atoms by a factor of about 6 and for C-atoms by a factor of 20. The reduced amount of oxygen under the 4:1 He/O-ICR cleaning leads to a reduced cleaning efficiency but a reduced O contamination on the wall. Pulsed He/O-ICR were also carried out on EAST with full metal walls at 420 K in 2007 and on full carbon walls at a room temperature in 2009 for a divertor configuration. With the same parameters (20 kW 1.4 10 2 Pa), H and C removal rates during He/O-ICR (ratio of He to O2 is 1) cleaning in HT-7 were 2 1021 H-atoms/h and 1.9 1021 C-atoms/h, whereas they were 2.5 1022 H-atoms/h and 1.3 1022 C-atoms/h in EAST with full metal walls, and 4.1 1022 H-atoms/h and 1.5 1022 C-atoms/h in EAST with full carbon walls. To remove the retained oxygen from the walls, both He-ICR and He-GDC were found to be effective. The oxygen removal rate depends on how much oxygen was retained on the walls, the ICR power (or the current of GDC) and the gas pressure. The oxygen removal rates in D2-ICR cleanings were about 5–10 times higher than in He-ICR cleanings. After wall cleaning, plasma discharges could be recovered but only after a few tens of disruptive plasmas (or a few hours in total time). Plasma recovery depended on the amount of oxygen retention on the walls before plasma operation. After He/O-ICR in the 2007 campaign of HT-7, both D2-ICR and He-ICR were used for oxygen removal, boronization with 1 g C2B10H12 being afterwards carried out aiming for suppressing the remaining oxygen. During this boronization, H2O and CO was effectively suppressed. 90 min 15 kW He-ICRF was used for H removal following the boronization. After those methods, plasmas were easily recovered and typical Zeff in those recovered plasma were about 2.5, which was similar as that after normal boronization.
Fig. 4. HF_GDC plasma in HT-7with Bt = 1.0 T.
S40
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
Fig. 5. Comparison before and after HF_GD Siliconization in HT-7. Bt = 1.8 T, Ip = 125 kA.
Fig. 6. Dependence of particle removal rates on oxygen pressure during O-GDC.
Even either O-GDC or O-ICR is very effective to remove T retention, but unfortunately this method is strongly limited by amount DTO produced and will bring serious corrosion problem in ITER. 6.2. Surface heating Tritium can be desorbed from the hydrocarbon film or the deposit layer can be ablated by the external heating e.g., using separatrix scanning, controlled low Ip disruption, and a laser or flash lamp heating. For conditioning the PFC surfaces that are in contact with the plasma, separatrix scanning might be useful for removal of tritium-Separatrix sweeping with an L-mode hydrogen plasma with plasma current of 7.5 MA and input power of 60 MW [18] may be effective for removal of deposit layer, which might reduce
the risk of flaking of deposit layer during the burn and reduce the tritium retention. Photonic cleaning by means of a flash lamp and laser was demonstrated in JET [47] and CEA [48]. T removal efficiency was estimated for ITER: 0.075 g T/h/150 m2 for flash lamp and (0.03–0.3 g T/h/150 m2 for laser. Further improvement of efficiency and combination with manipulator could open the possibility of tritium cleaning at places where it is difficult or impossible to clean with other techniques. Disruption at low plasma current may be also effective for removal of deposit layer and tritium retention. Calculations in [49] show that for a radiactive termination of an ITER 9 MA discharge with a stored energy of 120 MJ (0.2 MJ/m2) triggered by injection of neon gas, the surface temperature of a beryllium substrate may reach 800 °C.
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
S41
should be carried out soon in near future. Routine clean-up of deposit layer, with e.g. RF discharges, separatrix sweeping, mitigated low plasma current disruption, may be required to avoid operational difficulty with tritium. Acknowledgements The Authors would like to thank the fruitful discussion with Drs. A. Lyssoivan, D. Douai, D.K. Mansfield, Ch. Schüller, V. Philipps, R.A. Pitts, N. Ashikawa Y. Sakamoto , C. Grisolia, members of ITPA Sol & Divertor Topical Group from With contribution from TEXTOR, TORE SUPRA, ASDEX Upgrade, ITER-IO, JET ,W7-AS T, LHD, C-mode, Triam-IM, KSTAR, HT-7/EAST Teams. This work is partially supported by Chinese national Science Foundation No. 10721505. References Fig. 7. Removal rate of H, D and C-atoms during O-ICR cleanings.
7. Discussion and summary ICR-WC appears to be the most suitable method from the available conditioning techniques at least under those circumstances that a full or semi-permanent toroidal magnetic field will be present. At the moment, it still does not meet the requirement for T removal within a relative short period. In order to meet future need of ITER, more efforts should be done to understand the mechanism of RF plasma. Detail measurements of RF plasma parameters should be carried our which needs more special diagnostics. More realistic numerical modeling also need to be developed by updating present RF codes, such as TOMCAT, METS, CYRANO. It would be desirable to develop 1-D self-consistent plasma code which combines particle balance, RF-physics and PWI codes into a single integrated one. Multi-machine experiments are also needed for optimizing the RF system towards better coupling and more uniform RF plasma production under different gas mixture, toroidal/poloidal magnetic field, RF frequency, RF antenna phasing and power level conditioning. Since HF-GDC is such an easy approach in presence of high magnetic field compared with other methods, it is worth while to do further research for understanding the mechanism of high frequency glow discharge. Comparison for removal efficiency of impurities and hydrogen isotopic with conventional GDC and ICRWC should be done. Optimizing the antenna configuration and power supply for better coupling for a more uniform plasma needs to be carried out as soon as possible. Detail measurements for the HF-GDC plasma parameters and more realistic numerical modeling should be done in parallel to meet ITER urgent requirement. As summary: Wall conditioning provides an effective means for reducing impurities, recycling, T retention from the plasma surrounding surface, to facilitate plasma operation and to enhance plasma performance. Baking and GDC are still necessary in ITER. ICR wall conditionings are the most promising method with toroidal field at moment. Further simulation and multi-machine experiments are needed for an optimized RF cleaning system. HF-GDC provides a very simple approach for future use in ITER. More efforts
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]
G. Saibene et al., J. Nucl. Mater. 220–222 (1995) 617. S. Higashijima et al., J. Plasma Fusion Res. 75 (1999) 1297 (in Japanese). J. Winter, Plasma Phys. Control. Fusion 38 (1996) 1503. G. Federici et al., Nucl. Fusion 41 (2001) 1967. V. Philipps et al., IAEA FEC (2008). FT/4-2Ra. H.F. Dylla et al., J. Vac. Sci. Technol. 17 (1980) 286. F.G. Waelbroeck et al., Vac. Sci. Technol. A2 (1984) 1521. L. Oren, R. Taylor, J. Nucl. Fusion 17 (1977). 1143. V. Philipps, Trans. Fusion Sci. Technol. 45 (2004) 237. E. de la Cal, E. Gauthier, Control Fusion 47 (2005) 197. G. Counsell et al., Plasma Phys. Control Fusion 48 (2006) B189. A. Loarte et al., Nucl. Fusion 47 (2007) S203. J. Roth et al., Plasma Phys. Control Fusion 50 (2008) 103001. C.H. Skinner et al., Fusion Sci. Technol. 54 (2008) 891. R.E. Clausing et al., J. Nucl. Mater. 122–123 (1984) 1145. G.L. Jackson, T.S. Taylor, P.L. Taylor, Nucl. Fusion 30 (1990) 2305. H. Nakamura et al., Fusion Eng. Des. 70 (2004) 163. M. Shimada, These Proceedings, P2-1. Y. Sakamoto et al., J. Nucl. Mater. 93–94 (1980) 333–337. Y. Sakamoto et al., J. Nucl. Mater. 111–112 (1982) 485–488. Y. Nakashima et al., J. Nucl. Mater. 162–164 (1989) 812–817. R.T. Nachtrieb et al., J. Nucl. Mater. 266–269 (1999) 896–900. A. Sagara et al., J. Plasma Fusion Res. 75 (1999) 263–267. E. Jotaki et al., Fusion Eng. Des. 36 (1997) 447–450. K. Hanada et al., Fusion Eng. Des. 54 (2001) 79–86. J. Ihde et al., in: 26th EPS Conf. on Controlled Fusion and Plasma Physics, vol. 23, J (ECA), Maastricht, The Netherlands, 1999, p. 709. Y.W. Yu et al., Phys. Scripta 79 (2009). 015502 (6pp). Y.W. Yu et al., J. Nucl. Mater. 390–391 (2009) 1051–1054. K. Ushigusa et al., Fusion Eng. Des. 45 (1999) 137–144. A. Lyssoivan et al., in: 14th Topical Conf. on Radio Frequency Power in Plasmas (Oxnard), AIP Conf. Proc, vol. 595, 2001, pp. 146–149. K. Itami et al., J. Nucl. Mater. 390–391 (2009) 983–987. R.T. Nachtrieb et al., J. Nucl. Mater. 266–269 (1999) 896. A.I. Lyssoivan et al., Nucl. Fusion 32 (1992) 1361. M.D. Carter et al., Nucl. Fusion, 30 (1990)723. H.G. Esser et al., Nucl. Mater 241–243 (1997) 861. Y.P. Zhao et al., Phys. Lett. 14 (1997) 916. J. Li et al., Fusion 39 (1999) 973. R. Brakel et al., Nucl. Mater. 290–293 (2001) 1160. D.Douai, These Proceedings, I-09. D. Mueller et al., J. Nucl. Mater. 241–243 (1997) 897. V. Philipps et al., J. Nucl. Mater. 266–169 (1999) 386. V. Philipps et al., J. Nucl. Mater. 363–365 (2007) 929. A. Lyssoivan et al., J. Nucl. Mater. 363–365 (2007) 135. J.S. Hu, J.G. Li, et al., Plasma Phys. Control. Fusion 47 (2005) 1271. J.S. Hu, J.G. Li, et al., Plasma Phys. Control. Fusion 48 (2006) 807. J.S. Hu, J.G. Li, et al., Plasma Phys. Control. Fusion 49 (2007) 421–434. A. Widdowson et al., J. Nucl. Mater. 363–365 (2007) 341. C. Grisolia et al., J. Nucl. Mater. 363–365 (2007) 1138. D. Whyte et al., J. Nucl. Mater. 337–339 (2005) 560.
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Wall conditioning towards the utilization in ITER J. Li a,⇑, M. Shimada b, Y. Zhao a, J. Hu a, X. Gong a, Y.W. Yu a, G.Z. Zhuo a a b
Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, China ITER Organization, St Paul Lez Durance, France
a r t i c l e
i n f o
Article history: Available online 2 November 2010
a b s t r a c t Wall conditioning provides an effective means for reducing both impurities and recycling from the plasma surrounding surface. A wide variety of techniques have been developed during the last few decades for conditioning the plasma facing surface. With the presence of magnetic fields, electron cyclotron resonant (ECR) and ion cyclotron resonant (ICR) discharge cleaning techniques have been explored, which could be used for next generation superconducting magnetic confined devices, such as ITER. Efforts have been made on the application of ICR conditioning on many devices and significant progress has been made. A new and simple method for future wall conditioning, high frequency glow discharge cleaning (HF-GDC), has been developed. HF-GDC operates in the presence of strong magnetic field (0.5–2 T) at frequencies of 20–100 kHz stably for a wide range of gas pressures. In this paper, all these techniques are reviewed and their proper application in ITER are discussed. Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction Wall conditioning provides an effective means for reducing both impurities, and recycling from the plasma surrounding surface. The experience in present short-pulse tokamaks already indicates that wall conditioning plays a crucial role in achievement of reproducible and clean plasma conditions and improvement on overall plasma performance [1–5]. Some long pulse experiments suggest that the impact of wall condition on the plasma performance could be more serious in ITER with plasma discharge time from 400 to 3000 s. Furthermore, ITER, as a nuclear device, must control the in-vessel inventory of tritium and dust for safety reason. (1) The objectives of wall conditioning are: reduction of gaseous impurities influx from in-vessel components (the PlasmaFacing Component surfaces and surfaces inaccessible from the plasma) to maintain/improve plasma purity. (2) Reduction of hydrogenic species influx from in-vessel components to facilitate density control. (3) Removal of tritium from in-vessel components. (4) Removal of dust from Plasma-Facing Components. A wide variety of techniques have been developed during last few decades. Baking of the internal components inside vacuum vessel, glow discharge cleaning (GDC) [6,7], Taylor discharge clean⇑ Corresponding author. Address: Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, Anhui, 230031, China. Tel.: +86 551 5591371; fax: +86 551 5591310. E-mail address: [email protected] (J. Li). 0022-3115/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jnucmat.2010.10.048
ing (TDC) [8], which were developed in the early days of tokamak experiments, have been demonstrated to be effective, and are still in use today. With the presence of magnetic fields, electron cyclotron resonant (ECR) and ion cyclotron resonant (ICR) discharge cleaning techniques have been explored, which could be used for next generation tokomaks, such as ITER. Wall conditioning for above mentioned issues has been studied extensively for past few decades and many papers, especially many excellent review papers have been published [3,4,9–14]. The principle of all the cleaning procedures using plasma discharges is to reduce metal or carbon oxides by hydrogenic plasma bombardment to form volatile species, thus depleting the contamination layers on the wall surfaces. These volatile substances, such as water vapour, methane and other hydrocarbons, are then desorbed from the surface, both thermally and by particle impact, and ultimately evacuated from the device via the torus pumping system. Two key factors for achieving a large cleaning effect are the maximum volatile substance formation by plasma discharge, and the effective exhaust of the formed impurities before they are re-ionized, so that these impurity products are not re-ionized, dissociated and redeposited on the surfaces in the cleaning plasma itself. This requires plasmas to have low Te and a low ionization fraction. In addition, high wall temperatures are useful not only for increasing the rates of the reactions leading to the formation of the volatile species on the surface, but also for promoting their desorption from the surface. Furthermore, a large pumping speed is desired for efficient removal of the volatile species from the torus. The essential difference between various conditioning methods lies in the way the atomic hydrogen species are created and interact with the surface. Other effects such as charge exchange
S36
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
neutrals, and sputtering due to high-energy ion bombardment may also play a role. After introduction, GDC will be described in Section 2. ECR and ICR wall conditioning will be given in Sections 3 and 4. A new technique, high frequency glow discharge cleaning will be described in Section 5. Other techniques, such as strike point sweeping, flash lamp and oxidation for effective moving T retention will be given in Section 6. Discussion and summary for future needs and application to ITER will be discussed in last section of this paper.
2. Glow discharge cleaning The most effective method actually used for surface impurity cleaning is hydrogen or deuterium GDC. One or more anodes are either located in portholes, or moved by a bellows assembly into the main volume of the vacuum vessel. The wall surfaces and limiters are at ground potential. Typical hydrogen pressures, range from 0.1 Pa to a few Pa, the voltage between anode and cathode is of the order of a few hundred volts. Superimposed RF power (100 W) is often used to ease the breakdown of the glow, and to stabilize the plasma at low pressure. During D2-GDC, both ions and charge exchange neutrals interact with the first wall surface, and produce by chemical sputtering molecular impurities such as H2O (D2O), CxHy (CxDy) and CxOy, which are continuously pumped out during the cleaning process. Simultaneous wall baking is very helpful during discharge cleaning, especially for removing oxygen from the wall. Results from TEXTOR showed that a high wall temperature (over 350 °C) is favourable for quick conditioning of first wall during the GDC process [15]. Oxygen and water contamination can be very effectively removed from the vacuum wall by long glow discharges (100 h). He GDC is usually used for devices with a large graphite coverage. The main reason is that large quantities of hydrogen are stored within the penetration range of hydrogen ions. They can be released upon particle impact during a tokamak discharge and make density control difficult, in particular, when the walls cannot be baked to temperatures above 300 °C. For carbon wall, fuel retention is dominated by codeposition. Another reason for fuel retention is the porous structure of graphite that leads to storage of significant amounts of water vapour and hydrogen, rendering recycling control difficult during tokamak discharges. Helium glow discharge conditioning for hydrogen removal has been optimized in DIII-D [16], and routinely used between each tokamak discharge. It has been instrumental in achieving high-performance H-mode plasmas, and a wide operating space in DIII-D where major parts of the surface were covered by graphite. Tritium removal efficiency with H2 and He DC glow has been measured in JT-60U by Nakamura et al. [17]. The hydrogen isotope (tritium and deuterium) release rate by H2 GDC is much larger (by a factor of 2–3) than by He GDC due to chemical processes induced by the hydrogen discharge shown in Fig. 1. The dominant removal processes for the He GDC and H2 GDC after a few hours can be attributed to physical sputtering and isotope exchange reaction assisted by chemical sputtering induced by the H2 discharge, respectively. This graph suggests that the removal rate, or the tritium near the surface, could be reduced by three orders of magnitude with a half day of H2 GDC even at a modest baking temperature of 100 °C. Its drawback for a machine with superconducting magnets like ITER is that it can be used only occasionally e.g. just during the recovery from a vent or major air/water leak or when the magnets are not turned on (the frequency depends on the build-up rate of tritium inventory and the efficiency of other wall cleaning methods and can range from once a year to once a
Fig. 1. Tritium release rates from JT-60U vacuum vessel during: (a) H2 GDC and (b) He GDC at various wall temperature [17].
month). Thus it is excluded from the routine clean-up schemes during operation for a machine like ITER. Another possible drawback is that because the ion energy exceeds the sputtering threshold, erosion of PFCs and coating on the diagnostics mirrors and windows are concern. For ITER wall conditioning, GDC is still required when vent to air or during machine non-operation maintenance when there is no magnetic field. For this reason, GDC is still baseline requirement for ITER [18]. The detail of ITER GDC is shown in Table 1. However, due to the presence of high permanent magnetic fields, conditioning techniques based solely on glow discharges cannot be used. In future superconducting devices, such as ITER, which will be operated over long discharge durations and use deuterium/tritium mixtures, new techniques for wall conditioning are needed, firstly to obtain a reproducible and controlled plasma discharges, and secondly to reduce the tritium wall inventory after each discharge. ECR and ICR conditioning techniques have been developed during the past 20 years, and very good results have been obtained.
Table 1 Glow discharge requirements. Parameters First wall current density Bias voltage Pumping speed (He, molecular flow) Conditioning gas Number of electrodes Operating pressure Coil currents (TF, PF, CS and CC)
Unit 2
A/m V m3/s
Pa A
Value 0.1–0.4 300–600 30–40 H2, D2, He, O2 6 0.01–0.5 0
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
3. ECR conditioning Electron cyclotron resonance (ECR) discharge cleaning was first carried out on JFT-2[19] in 1980s, by using 2.45 GHz LH system under the magnetic field of 0.0875 T, and many experiments were done on JIPPT-II [20], the Gamma-10 tandem mirror [21], Alcator C-Mod [22], LHD [23], TRIAM-1 M [24,25], TOMAS [26], and recently HT-7 [27,28], except for JT-60U by using 1.7 GHz LH system under 0.06 T [29]. ECRF system with the frequency of 110 GHz on Textor-94 was used for ECR discharge cleanings under 1.65–2.28 T [30]. He, H2, D2, Ar and CH4 or CD4 were used as working gas during ECR discharges in the pressure between 1 10 4 and 6 10 2 Pa, for the experimental studies on inactive or non-reactive cleaning, thin film coating and co-deposited layer removal. Fig. 2 shows power densities defined as the injected ECR power normalized to the particle content [28], in different devices. It ranged from 2 to 4 103 kW Pa 1 m 3. Typical electron temperature of ECR plasma was 1–10 eV, and plasma density (1–5) 1016 m 3. ECR plasmas were nearly toroidally homogeneous, but poloidally inhomogeneous. Due to the ExB drift induced by the absence of plasma current and therefore poloidal magnetic field component, ECR plasmas were mainly concentrated on the resonance layer and low field side. A typical picture of ECR plasmas on JIPP T-II is shown in Fig. 3 with the electron density peaking at the resonance layer decreasing towards low field side [20]. Improved homogeneity of ECR plasmas were obtained on JT-60U with 110 GHz ECRH system by applying a horizontal field Bh (0.3–0.5% Bt), and Bh was crucially important to expand ECR plasma toward the high field side [31]. Compared with other wall conditioning methods, ECR discharge cleaning had similar or better effects than TDC on JFT-2 with either H2 or He working gas, and it was effective for performance improvement of both OH and LHCD plasmas on TRIAM-1M. But the removal rate of ECR discharge cleaning was 7–20 times lower than ICRF discharge cleaning on Textor-94 and HT-7. Recent ECR discharge cleaning on JT-60U using 110 GHz ECRH system showed that the H2 outgas efficiency was equivalent to 79% of TDC with the power of 1.27 MW. Pulsed mode with 1 s on and 59 s off was used. It was considered that higher outgassing rate could be attributed due to this duty ratio. One ITER relevant application on JT-60U is that after a disruption of Ip = 1 MA, Bt = 3.5 T, PNBI = 19 MW and the stored energy (W) = 3.3 MJ, the plasma was soon recovered by only one shot ECR cleaning with 2.5 MW of 110 GHz for 1.5 s without change of Bt. Removal of deposit or co-deposited hydrogen isotope by ECR discharge was also studied. Due to the localized ECR plasmas, maximum removal rate was at a few centimetres outside the resonance location on Alcator C-Mod [32]. For the removal of tritium from
Fig. 2. Power densities of ECR discharge cleaning on different devices.
S37
co-deposited layers, laboratory experiments showed that the best conditions for a fast and effective method is oxygen ECR plasma at a low pressure to achieve high ion fluxes and at relatively high wall temperature [26]. ECR plasma position should be scanned by changing magnetic field in order to obtain homogeneous deposition or removal. However, for the future use in ITER with either full field (5.3 T) or half field (2.65 T), the frequency of ECR must be either 170 GHz or 85 GHz with a power of 0.5–1.0 MW. More uniform plasma both in toroidal and poloidal directions should be produced. Significant efforts are still needed towards this goal. The cleaning efficiency of ECR is much lower than that of GDC. This is caused by the extremely localized power deposition area which makes it very difficult to illuminate all surfaces homogeneously.
4. ICR wall conditioning The theory of plasma production in tokamaks using RF in the ICRF range has been investigated by some authors [33,34]. To initiate the discharge, an antenna has to generate an RF electric field E| parallel to the magnetic field lines. The E|| at the antenna is responsible for neutral gas ionization by highly accelerated electrons during gas breakdown. The primary plasma generation is not directly related to IC-resonant layers but caused by RF-breakdown in the strong E.M. field near the antenna. This explains why the experimentally observed dependence on toroidal field at constant frequency is weak. Poloidal ICRF antennas can generate Ez-field (evanescent in vacuum) electrostatically and inductively due to: (i) RF voltage difference between the strap and the sidewalls (RF limiters). (ii) RF voltage induced between the tilted FS bars by the time-varying magnetic flux. To get a higher ionization ratio, the harmonic resonance absorption layers should be inside the plasma. The ion cyclotron frequency x is qB/m, where B is the magnetic field, q the ionic charge and m the mass. Plasma waves (slow wave, ion Bernstein wave, fast wave) excite and propagate over the torus in a relay-race regime governed by the antenna kZ-spectrum. The mechanism of cleaning wall by ICR WC is mainly due to [10]: (a) A massive flux of low energy charge exchange neutrals with fluxes of 1–10 1020 p/m2 s. (b) A low flux of high energy charge exchange neutral hydrogenic atoms fluxes of 1–10 1018 p/m2 s. (c) Flux of ions where the resident flux-tubes cross the wall with fluxes of a few times 1021 p/m2 s, however they are very localized. By using almost the same RF system for heating with an RF power from few tens to few hundreds kW and toroidal magnetic field from 0.2 to 3 T, RF plasma could be easily initiated with a wide range of pressure from 1 10 3 Pa to 0.1 Pa. ICR-WC experiments were performed in Tore Supra [11], TEXTOR [35], HT-7 [36,37], W7-AS [38], ASDEX-U [39], JET [39], KSTAR, LHD and EAST and significant progress has been achieved since 1996. Extensive ICR conditioning has been routinely used in HT-7 and EAST superconducting tokamaks for wall cleaning, recycling and isotopic control and thin film coating (B, Si, Li) by using different working gases. In both HT-7 and EAST, ICR conditioning showed a higher particle removal rates than GDC. The basic RF plasma parameters produced by different antenna configurations were measured in different machines. Quadruple mass spectrometer (QMS) analysis during and after the RF conditioning was routinely carried out. Electron temperature is in the range 3–50 eV. The electron temperature for helium RF plasmas
S38
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
Fig. 3. He ECR plasmas (a) and electron density distribution (b) of ECR plasmas on JIPP T-II tokamak.
is higher than that in deuterium plasmas by a factor of two with the same discharge condition. Electron density is in the range of 0.5–8 1017 m 3. Ion temperature is as high as 0.5–2 keV with high-energy tail of a few tens of keV. The two effects which govern the conditioning efficiency are outgassing rate of particles from the wall and the ionization rate of desorbed molecules which induces a redeposition [3]. For getting higher outgassing rate, a higher particle flux to the wall with suitable energy is favourable. For reducing ionization rate, the pulsed-RF-mode was used. Helium RF discharge is used mainly for hydrogen removal. This makes plasma start up very easy much and allows the control of particle recycling. He-RF discharges are also very effective for the replacement of deuterium by isotopic exchange. The pulsed mode with higher filling pressure (0.05–0.1 Pa) increases the conditioning efficiency by controlling electron density and temperature separately to minimize the hydrogen redeposition probability. The partial pressure of hydrogen was increased by two orders of magnitude after RF plasma was initiated, as a result of the desorption from the wall. High hydrogen removal rates have been obtained. Deuterium ICRF discharge is used for surface cleaning and wall isotope control in HT-7 and EAST tokamaks. The pulsed mode with optimized pressure of 1 10 1 Pa is used. The electron temperature is lower (2–3 eV) compared with helium discharge. Electron temperature changes a little within the injection RF power of 30 kW. The low temperature minimizes the strong plasma reionization and wall redeposition of neutral species. High temperature deuterium ion and its energetic tail are observed. Both energetic and reactive particles bombard the first wall and induce the desorption of molecular impurities. These are main reasons for the good conditioning efficiency of the deuterium RF discharge. It is clearly shown that for future application in ITER, ICR has more favourable features than ECR-WC since it can be easily initiated for relatively uniform plasma without the need for extra hardware, and with higher particle removing efficiency compared with ECR-WC at high density. Even thin film coating is not foreseen in ITER due to possible co-deposited layers which is more difficult to remove, but it is still very useful for getting a quick plasma start-up after minor leakage or disruption. Two possible ways could be RF boronization and RF siliconization. ICR boronization using less hazardous carboranes C2B10H12, was fully developed in HT-7 and EAST, and became a routine coating process [37]. Pulsed RF plasmas with a magnetic field of 2 T and 20 kW of RF power were produced by using a mixture of carborane
(C2B10H12) 50% in helium (50%). The boronization process lasted a few hours and was followed by a He ICR discharge for removing the high hydrogen content in the film. A fine homogeneous and hard a-B/C:H film was produced by pulsed ICRF plasma. The film shows stronger adhesion, greater thickness, and longer lifetime compared with those by GDC method. Good uniformity of the film in both toroidal and poloidal directions has been obtained by using a long antenna on the high field-side of the vessel. The re-emission of hydrogen was easily controlled by using helium ICR discharge after boronization, leading to very strong wall pumping. The plasma performance was significantly improved after boronization with a higher density limit, and wider operation space was obtained. In addition, the hard X-rays accompanied by high power lower hybrid current drive were dramatically suppressed. This gives direct evidence that the thin boron film serves as a protecting layer against the energetic particles, which is very important for future long-pulse operation. RF siliconization was also performed in HT-7 and EAST by using a silane (SiH4) gas mixture (10%) with helium (90%) to facilitate gas handling. The RF discharge was operated in pulsed mode with a 1 s on, 1 s off cycle, and by using the long ICRF antenna side of the vessel. This results in a homogeneous coating in the toroidal and poloidal directions with a film thickness of 50–80 nm. The promising results obtained from ICR plasma-assisted deposition have shown that such techniques are applicable to future superconducting fusion devices with fully W wall after disruption or quick start-up of plasma discharge even the coating techniques are not foreseen in ITER.
5. High frequency glow discharge cleaning Recently, a new method, called High Frequency glow discharge cleaning (HFGDC) with magnetic field, has been invented and worked in both EAST and HT-7 superconducting Tokamaks. Normal electrodes for GDC have been used for this new method with a high frequency power supply. The parameters of HFGDC power supply are: U = 2.0 kV, f25–100 kHz, I10.0 A. Plasma performance for High frequency glow discharges was investigated by reciprocating Langmir probe under the conditions with and without toroidal magnetic field. It shows that the electron temperature of HF_GD plasma with B-field is similar with those of normal DC_GDC without B-field. The cleaning effects are also similar compared with GDC without magnetic field from QMA measurements.
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
Preliminary results showed that plasma can be easily initiated with different gas filling pressure. With higher filling pressure (from 10 10 2 Pa to a few Pa), plasma is much localized around electrodes. When pressure gets lower (from 1 10 3 Pa to 1 10 4 Pa), plasma becomes more uniform not only in toroidal direction but also in poloidal direction. Fig. 4 shows HF-GDC plasma in HT-7 with different filling pressure. It was clearly shown that a uniform plasma was formed with filling pressure between 10 3 and 10 4 Pa. Both helium and deuterium working gases have been used and results are similar. Siliconization technique by using HF_GD was also studied in HT-7. Similar results have been obtained compared with RF-siliconization. Strong impurities radiation caused by a 20 Pa leakage in HT-7 could be suppressed by only 1 h HFGDC with BT = 1.8 T and SiH4 filling gas. Fig. 5 shows two shots before and after HFGDC siliconization.
6. Other methods for tritium removal 6.1. Oxidation The long-term retention of tritium fuel in codeposition layer, plasma facing surface and bulk of materials in fusion devices is one of the major problems. Removal of tritium from amorphous carbon layers will have an important impact on machine operation and safety in the next generation tokamaks, such as ITER. Oxidation treatment is one of promising methods for amorphous tritiated carbon layers removal. Oxidation wall conditioning has been carried out in TFTR [40], TEXTOR [41,42], AUG [43]. Very recently, a systematic experiments of oxidation wall conditioning, including thermo-oxidation, glow discharge (O-GDC) and radio frequency wave associated oxidation (O-ICR), were carried out in a limiter machine – HT-7 [44–46], and in an ITER-relevant full superconducting tokamak – EAST with both metal and full carbon wall conditions. In the O-GDC experiment on HT-7, the removal rates of H-atoms and D-atoms in form of H2O, HDO and D2O were higher than that of H2 and D2 by factors of about 20 and 50, respectively, as shown in Fig. 6. Compared to He-GDC cleaning, O-GDC has higher removal rates for H-atoms and C-atoms by a factor of about 2–4 and about 25, respectively. Average removal rates of 5.2 1022 H-atoms/h, 5.6 1021 D-atoms/h and 5.5 1022 C-atoms/h, respectively, were obtained during 145 min O-GDC cleaning in a pressure range of
S39
0.5–1.5 Pa. However, this procedure led to a significant O contamination. About 5.4 1022 O-atoms were adsorbed on the walls. During a long O-GDC operation in 2005, the removal rates for H, D and C remained very high within 9.5 h. The removal rates for H and C decreased faster than that for D, possibly due to the fact that H and C were desorbed from Stainless Steel and carbon walls whereas D mostly originated from deposits. Continuous or pulsed (0.3 s on and 1.2 s off) O-ICR was performed in the presence of a magnetic field of 1.5–2.0 T on HT-7. The highest removal rates of H, D and C atoms reached to 2.6 1022, 7.8 1021 and 1.5 1022 atoms/h, respectively, during 40 kW 9.0 10 2 Pa pure pulsed O-ICR cleaning shown in Fig. 7. Compared to He-ICRF cleaning, O-ICR has a higher removal rate for H-atoms by a factor of about 6 and for C-atoms by a factor of 20. The reduced amount of oxygen under the 4:1 He/O-ICR cleaning leads to a reduced cleaning efficiency but a reduced O contamination on the wall. Pulsed He/O-ICR were also carried out on EAST with full metal walls at 420 K in 2007 and on full carbon walls at a room temperature in 2009 for a divertor configuration. With the same parameters (20 kW 1.4 10 2 Pa), H and C removal rates during He/O-ICR (ratio of He to O2 is 1) cleaning in HT-7 were 2 1021 H-atoms/h and 1.9 1021 C-atoms/h, whereas they were 2.5 1022 H-atoms/h and 1.3 1022 C-atoms/h in EAST with full metal walls, and 4.1 1022 H-atoms/h and 1.5 1022 C-atoms/h in EAST with full carbon walls. To remove the retained oxygen from the walls, both He-ICR and He-GDC were found to be effective. The oxygen removal rate depends on how much oxygen was retained on the walls, the ICR power (or the current of GDC) and the gas pressure. The oxygen removal rates in D2-ICR cleanings were about 5–10 times higher than in He-ICR cleanings. After wall cleaning, plasma discharges could be recovered but only after a few tens of disruptive plasmas (or a few hours in total time). Plasma recovery depended on the amount of oxygen retention on the walls before plasma operation. After He/O-ICR in the 2007 campaign of HT-7, both D2-ICR and He-ICR were used for oxygen removal, boronization with 1 g C2B10H12 being afterwards carried out aiming for suppressing the remaining oxygen. During this boronization, H2O and CO was effectively suppressed. 90 min 15 kW He-ICRF was used for H removal following the boronization. After those methods, plasmas were easily recovered and typical Zeff in those recovered plasma were about 2.5, which was similar as that after normal boronization.
Fig. 4. HF_GDC plasma in HT-7with Bt = 1.0 T.
S40
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
Fig. 5. Comparison before and after HF_GD Siliconization in HT-7. Bt = 1.8 T, Ip = 125 kA.
Fig. 6. Dependence of particle removal rates on oxygen pressure during O-GDC.
Even either O-GDC or O-ICR is very effective to remove T retention, but unfortunately this method is strongly limited by amount DTO produced and will bring serious corrosion problem in ITER. 6.2. Surface heating Tritium can be desorbed from the hydrocarbon film or the deposit layer can be ablated by the external heating e.g., using separatrix scanning, controlled low Ip disruption, and a laser or flash lamp heating. For conditioning the PFC surfaces that are in contact with the plasma, separatrix scanning might be useful for removal of tritium-Separatrix sweeping with an L-mode hydrogen plasma with plasma current of 7.5 MA and input power of 60 MW [18] may be effective for removal of deposit layer, which might reduce
the risk of flaking of deposit layer during the burn and reduce the tritium retention. Photonic cleaning by means of a flash lamp and laser was demonstrated in JET [47] and CEA [48]. T removal efficiency was estimated for ITER: 0.075 g T/h/150 m2 for flash lamp and (0.03–0.3 g T/h/150 m2 for laser. Further improvement of efficiency and combination with manipulator could open the possibility of tritium cleaning at places where it is difficult or impossible to clean with other techniques. Disruption at low plasma current may be also effective for removal of deposit layer and tritium retention. Calculations in [49] show that for a radiactive termination of an ITER 9 MA discharge with a stored energy of 120 MJ (0.2 MJ/m2) triggered by injection of neon gas, the surface temperature of a beryllium substrate may reach 800 °C.
J. Li et al. / Journal of Nuclear Materials 415 (2011) S35–S41
S41
should be carried out soon in near future. Routine clean-up of deposit layer, with e.g. RF discharges, separatrix sweeping, mitigated low plasma current disruption, may be required to avoid operational difficulty with tritium. Acknowledgements The Authors would like to thank the fruitful discussion with Drs. A. Lyssoivan, D. Douai, D.K. Mansfield, Ch. Schüller, V. Philipps, R.A. Pitts, N. Ashikawa Y. Sakamoto , C. Grisolia, members of ITPA Sol & Divertor Topical Group from With contribution from TEXTOR, TORE SUPRA, ASDEX Upgrade, ITER-IO, JET ,W7-AS T, LHD, C-mode, Triam-IM, KSTAR, HT-7/EAST Teams. This work is partially supported by Chinese national Science Foundation No. 10721505. References Fig. 7. Removal rate of H, D and C-atoms during O-ICR cleanings.
7. Discussion and summary ICR-WC appears to be the most suitable method from the available conditioning techniques at least under those circumstances that a full or semi-permanent toroidal magnetic field will be present. At the moment, it still does not meet the requirement for T removal within a relative short period. In order to meet future need of ITER, more efforts should be done to understand the mechanism of RF plasma. Detail measurements of RF plasma parameters should be carried our which needs more special diagnostics. More realistic numerical modeling also need to be developed by updating present RF codes, such as TOMCAT, METS, CYRANO. It would be desirable to develop 1-D self-consistent plasma code which combines particle balance, RF-physics and PWI codes into a single integrated one. Multi-machine experiments are also needed for optimizing the RF system towards better coupling and more uniform RF plasma production under different gas mixture, toroidal/poloidal magnetic field, RF frequency, RF antenna phasing and power level conditioning. Since HF-GDC is such an easy approach in presence of high magnetic field compared with other methods, it is worth while to do further research for understanding the mechanism of high frequency glow discharge. Comparison for removal efficiency of impurities and hydrogen isotopic with conventional GDC and ICRWC should be done. Optimizing the antenna configuration and power supply for better coupling for a more uniform plasma needs to be carried out as soon as possible. Detail measurements for the HF-GDC plasma parameters and more realistic numerical modeling should be done in parallel to meet ITER urgent requirement. As summary: Wall conditioning provides an effective means for reducing impurities, recycling, T retention from the plasma surrounding surface, to facilitate plasma operation and to enhance plasma performance. Baking and GDC are still necessary in ITER. ICR wall conditionings are the most promising method with toroidal field at moment. Further simulation and multi-machine experiments are needed for an optimized RF cleaning system. HF-GDC provides a very simple approach for future use in ITER. More efforts
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]
G. Saibene et al., J. Nucl. Mater. 220–222 (1995) 617. S. Higashijima et al., J. Plasma Fusion Res. 75 (1999) 1297 (in Japanese). J. Winter, Plasma Phys. Control. Fusion 38 (1996) 1503. G. Federici et al., Nucl. Fusion 41 (2001) 1967. V. Philipps et al., IAEA FEC (2008). FT/4-2Ra. H.F. Dylla et al., J. Vac. Sci. Technol. 17 (1980) 286. F.G. Waelbroeck et al., Vac. Sci. Technol. A2 (1984) 1521. L. Oren, R. Taylor, J. Nucl. Fusion 17 (1977). 1143. V. Philipps, Trans. Fusion Sci. Technol. 45 (2004) 237. E. de la Cal, E. Gauthier, Control Fusion 47 (2005) 197. G. Counsell et al., Plasma Phys. Control Fusion 48 (2006) B189. A. Loarte et al., Nucl. Fusion 47 (2007) S203. J. Roth et al., Plasma Phys. Control Fusion 50 (2008) 103001. C.H. Skinner et al., Fusion Sci. Technol. 54 (2008) 891. R.E. Clausing et al., J. Nucl. Mater. 122–123 (1984) 1145. G.L. Jackson, T.S. Taylor, P.L. Taylor, Nucl. Fusion 30 (1990) 2305. H. Nakamura et al., Fusion Eng. Des. 70 (2004) 163. M. Shimada, These Proceedings, P2-1. Y. Sakamoto et al., J. Nucl. Mater. 93–94 (1980) 333–337. Y. Sakamoto et al., J. Nucl. Mater. 111–112 (1982) 485–488. Y. Nakashima et al., J. Nucl. Mater. 162–164 (1989) 812–817. R.T. Nachtrieb et al., J. Nucl. Mater. 266–269 (1999) 896–900. A. Sagara et al., J. Plasma Fusion Res. 75 (1999) 263–267. E. Jotaki et al., Fusion Eng. Des. 36 (1997) 447–450. K. Hanada et al., Fusion Eng. Des. 54 (2001) 79–86. J. Ihde et al., in: 26th EPS Conf. on Controlled Fusion and Plasma Physics, vol. 23, J (ECA), Maastricht, The Netherlands, 1999, p. 709. Y.W. Yu et al., Phys. Scripta 79 (2009). 015502 (6pp). Y.W. Yu et al., J. Nucl. Mater. 390–391 (2009) 1051–1054. K. Ushigusa et al., Fusion Eng. Des. 45 (1999) 137–144. A. Lyssoivan et al., in: 14th Topical Conf. on Radio Frequency Power in Plasmas (Oxnard), AIP Conf. Proc, vol. 595, 2001, pp. 146–149. K. Itami et al., J. Nucl. Mater. 390–391 (2009) 983–987. R.T. Nachtrieb et al., J. Nucl. Mater. 266–269 (1999) 896. A.I. Lyssoivan et al., Nucl. Fusion 32 (1992) 1361. M.D. Carter et al., Nucl. Fusion, 30 (1990)723. H.G. Esser et al., Nucl. Mater 241–243 (1997) 861. Y.P. Zhao et al., Phys. Lett. 14 (1997) 916. J. Li et al., Fusion 39 (1999) 973. R. Brakel et al., Nucl. Mater. 290–293 (2001) 1160. D.Douai, These Proceedings, I-09. D. Mueller et al., J. Nucl. Mater. 241–243 (1997) 897. V. Philipps et al., J. Nucl. Mater. 266–169 (1999) 386. V. Philipps et al., J. Nucl. Mater. 363–365 (2007) 929. A. Lyssoivan et al., J. Nucl. Mater. 363–365 (2007) 135. J.S. Hu, J.G. Li, et al., Plasma Phys. Control. Fusion 47 (2005) 1271. J.S. Hu, J.G. Li, et al., Plasma Phys. Control. Fusion 48 (2006) 807. J.S. Hu, J.G. Li, et al., Plasma Phys. Control. Fusion 49 (2007) 421–434. A. Widdowson et al., J. Nucl. Mater. 363–365 (2007) 341. C. Grisolia et al., J. Nucl. Mater. 363–365 (2007) 1138. D. Whyte et al., J. Nucl. Mater. 337–339 (2005) 560.