The Next Step in a Development of Negative Ion Beam Plasma Neutraliser

The Next Step in a Development of Negative Ion Beam Plasma Neutraliserfor ITER NBI. V.M. KULYGIN 1), E.D.DLOUGACH 1), E.P.GORBUNOV 1), E.YU. KLIMENKO, 1) A. A. MEHED’KIN 2) I.V.MOSKALENKO 1), A.A.PANASENKOV 1), YU.M.PUSTOVOIT 1), A. A. SKOVORODA 1), V.A.SMIRNOV 1), V.A.ZHIL’TSOV 1), V.F.ZUBAREV 1)

1) Institute of Nuclear Fusion, RRC Kurchatov Institute, Moscow, Russia 2) Moscow Radio Technical Institute,Moscow, Russian Federation e-mail: [email protected]

Abstract. Injectors of deuterium atom beams developing for ITER plasma heating and current drive are based on the negative ion acceleration and further neutralisation with a gas target .The maximal efficiency of a gas stripping process, is 60%. The replacement of the gas neutraliser by a plasma one must increase the neutral yield to 80%. An overview оn experimental study of microwave discharge in a multicusp magnetic system chosen as a base device for Plasma Neutraliser (PN) realisation and design development of PN for ITER Neutral Beam Injectors (NBI) is presented. The experimental results achieved at a plasma neutraliser model PNX-U are discussed. Plasma confinement, gas flows, 18 -3 ionisation degree were investigated. The high density plasma with ne ≈ 10 m with low electron and ion temperatures (~5-6 eV) and high ionisation degree (not less than 40%) at its centre has been generated in operation with Argon.

1. Introduction

To provide an effective stripping of D beam PN plasma linear density have to be at a 15 -2 level of nel~ 2-3 10 cm . As the plasma neutralizer must be able to replace the gas neutralizer in the ITER NBI of the reference concept its length, l , is defined and equal to 3-4 12 -3 m. So the necessary plasma density in neutralizer should be ne~5-7 10 cm . The multicusp 3D magnetic wall configuration has been chosen for PN plasma confinement system [1]. An -4 electrodeless microwave discharge at a low gas pressure (10 Torr, H2 ,Ar) has been used to generate a plasma in the multicusp configuration. After some preliminary experiments the model PNX-U (plasma volume 0.5m3) was constructed for proof of principle of the proposed PN scheme. The first results of the experiments have been presented at IAEA Conference in Yokohama [2]. The experiments were continued to get more comprehensive database for PN ITER NBI design. Plasma confinement, gas flow and ionization degree were investigated. H- mode of operation with plasma density exceeded the critical one (for 7GHz) have been got. Some diagnostics were prepared and installed additionally. Namely: diagnostic H  beam injector (10 – 100mA, up to 350keV); local passive spectrometry for visible spectrum region; laser fluorescence diagnostic for local measurements of ionization degree in experiments with Argon plasma. A microwave interferometer, probes, calorimeters and ultraviolet/soft X- radiation detector were used as well as ion and electron multigrid energy analyzers.

2. PNX-U experiment brief description

To refresh in one’s mind the PNX-U magnetic field configuration and the coils arrangement the FIG.1 is presented. The magnetic system consists of cylindrical side part and two end parts: a front one and a back one. The side part is arranged with ring coils connected in pairs with interchanging current direction from pair to pair. The inner diameter of the side part coil is 60 cm. The ends closing of the formatted multicusp magnetic system is fulfilled with

1 transition to a ring coil of reduced diameter and finally to a coil of a racetrack form. The large part of experiments was fulfilled at a coil current in the range of 2.5 – 2.7 kA. The tangential microwave inputs are located at gaps between coils of a co-current pair to prevent large plasma flow to a launcher. The diagnostic laser radiation is injected through a magnetic slit.

FIG.1. PNX-U magnetic field configuration IB=2.5 kA

. FIG.2. The diagnostics displacement.

The schema of diagnostic system is shown on FIG.2. The PNX-U vacuum chamber is divided into three parts with two partitions. The central part and the right one have their own independent vacuum pumping with turbomolecular pumps. The left part is evacuated via edge side windows and axial hole. There are two collectors of axial holes plasma flows at the front and the back ends (1), two such a collectors at the end ring slits (3) (between the last side coil of a general size and the reduced diameter coil at the end part). These collectors were constructed to intercept fully the plasma flow going to end magnetic slits. They are water cooled (calorimetric measurements) and electrically insulated. Each of the ring collectors is divided into two independent parts to estimate a vertical uniformity of the correspondent plasma flows. Each of the collectors can be biased with different potentials. This allows to get their probe characteristics and in particular to measure a full ion current to each of the collectors. The axial collectors volumes are connected to pressure gauges with tubes (2) . This gives a possibility to measure gas pressure variation in the collectors during plasma flowing at different conditions.Pressure gauges displacement is figured as (4) throughout the FIG.2; optical windows for the main checking and spectrum measurements are marked as (5).

2 Plasma linear density is measured with 4 mm microwave interferometer, whose box horns are displaced at opposite end windows (5 at FIG.2). A retarding field method is used to measure plasma parameters with multigrid analysers. These analysers give a possibility to get an information about ion and electron mean energyand a plasma potential in different parts of the installation. Two of such analysers are displaced out of the main installation vacuum vessel (MA1 & MA2) and two (MA3 & MA4) – inside the chamber (FIG.2). Magnetic lines going from different parts of magnetic system lead plasma particles to the analysers. The MA1 position and the related magnetic lines structure lets this analyser “see” and analyse the particle flow coming from the “corner” of the magnetic system just behind the ring region with zero magnetic field. The MA2 analyser “sees” the central region. As for MA3 and MA4 the related magnetic field leads to MA3 particles from more deep region (up to 0.1m from the axis) than to MA4 (up to 0.2m from the axis) FIG.3. An example of an ion flux analysis is sited on the FIG.4. Electrons are blocked up by U-=–300 V at one of grids; the other one analyses ions with a sawtooth voltage was applied to it. In the case of MA1 the ion current wasn’t changed till U+= 53 V

FIG.3. A magnetic field value distribution. along gap and position of multigrid analysers

MA3 and MA4. IB=2.5 kA.

4,5 ions 4,0 82 V 3,5 MA1 3,0 53 V . u . a

, 2,5 e g a

t 2,0 l o

v 0 V 1,5 A M

t 1,0 n e r

r 0,5 Ti~6 eV a c

A 0,0 M -0,5

-1,0 0,0 0,1 0,2 0,3 t, s

FIG.4.Oscillogramms of MA1 ion current and positive retarding potential. U-= -300 V.

-0,2 . u . -0,3 a

n -0,4 e r r u

c -0,5

M -0,6

-0,8 MA-1 MA-2

-0,9 0 50 100 150 200 250 300 t, ms

FIG.5. Oscillogram of the MA ion current. U+= 0 V, U-= -300 V.

0,5 1 0,0

-0,5 . u . a

t -1,0 n e r r u

c -1,5

e b o r -2,0 p

r 2 i u

m -2,5 g n a

L -3,0

-3,5 0 200 t, ms

FIG.6. Oscillogramms of Langmuir probes. UL= -50 V. 1- peripheral region, 2- central plasma region. at the analyzing grid being decreasing to zero after that. Such an oscillogram can be simply explained: the beginning point of the ion current decreasing gives us the value of plasma potential and the ion current diminishing incline lets a possibility to find out the value of an ion temperature. FIG.5 demonstrates oscillogramms of a full ion current (U+=0 V) for MA1 and MA2. Let us remind that MA1 intercepts the magnetic lines coming from a plasma periphery while MA2 “sees” a central region of a plasma. One can see that during the pulse a transition to improved plasma confinement was occurred: the “central” signal is increasing. Plasma density and electron temperature were measured locally with movable Langmuir probes (12,13 on FIG.3). An arrange of immovable Langmuir probes was situated behind a side magnetic slit. Ten probes are displaced across the slit plasma flow to measure its plasma parameters. The probe ion current oscillogramms for “central” and “peripheral” interception are shown on FIG.6. The same situation as for multigrid analyzers is demonstrated : a central region plasma density is increasing. Fig.8. demonstrates the improvement of a plasma confinement near 125th ms by other way. Plasma loss along the axis is going down, than became approximately 30% less very sharply.

nl=1.8 1014cm-2 4

1 Phf Reflection

0 Ar8115

0,00 0,05 0,10 0,15 0,20 t,sec

FIG.7. The H-regime signal oscillogramms: nl-interferometer,Phf –the injected microwave power, Ar line intensity

There marked also on Fig.3 fiber light guide (6), monochromators for an emission spectroscopy and for laser-induced fluorescence (LIF) technique (7,8), laser irradiation system for LIF (9). Multigrid plasma flow analyzers, MA, provides particle energy spectrum measurements in different parts of the installation are marked by figure (10). 4-mm microwave interferometer shines through the plasma along the installation via the right window at the entrance end (the left one is used for diagnostic H- beam) to measure the line density nel (11).

1,5 . u .

a ion flux

, 1,0 e r gas pressure u s s e r p

, 0,5 x u l f

0,0 0 50 100 150 200 250 300 t, ms

FIG.8.Oscillogramms of ion flux and gas pressure in axial plasma collector.

There are also a movable Langmuir probes in a gap (12) and in a magnetic slit (13) and a system of such a probes is displaced across a slit to measure a correspondent plasma flow distribution (14) . An integral light intensity is measured by photomultiplier (15); mass

5 spectrum control provides with a mass-spectrometer (16); microwave directional couplers (17) lets a possibility to measure an incident and a reflected microwave power.

3. Experimental results overview and discussion

The main experimental results can be briefly listed as follows:  Use of piezo-valve for an additional gas puffing gave a possibility to get a steady-state H- 14 -2 12 -3 mode of operation with ne ≈ 1.5 ncr . An Argon plasma with nel≈2 10 см (ne≈10 см ) was generated at average injected microwave power density 0.08 W/cm3 .  Significant improvement of plasma confinement in compare with а simple magnetic case was achieved during microwave injection. It can be explained by ambipolar potential barrier at a plasma periphery and in cusp slits because of ECR heating of electrons at a periphery. Electron temperature at a periphery ( 20-30 eV) is much higher then in a main plasma body (5-8 eV). Plasma potential is equal correspondingly ~50V in slits and ~15- 20V in a plasma center. Periphery confinement time is ~0.4ms that is significantly less than in center (~0.4s). A full volume-averaged energy confinement time is ~1ms. When the klystrons are switched off the hot electrons leaves the trap during ~30 μs, while the main cold plasma decays during ~3 ms in accordance with a classical cusp confinement scaling. The oscillogramms presented on FIG.7 demonstrate relatively long time decay of nel after microwave power switching off while Argon light stimulated with the hot electrons disappears immediately.

 An average ionization degree ne/ng ≈ 0.25. It is very nonuniform in radial direction and achieves a value not less than ≈40% near plasma axis i.e. in the region which is important for beam neutralization.  In case of PNX-U operation with hydrogen the plasma density was 1.5-2 times less than with Argon. The Main Statements which are demonstrated with the experimental results: An improved confinement mode of operation (potential confinement) can be achieved in the low pressure Ar discharge at a multicusp system of PNX-U with a periphery ECR electron heating. At this High Mode (HM) regime  Plasma of cutoff density (about 6 – 7 1011 sm-3) was generated with 15 – 20 kW of injected microwave power in volume of 0.5 m3 ;  Low electron and ion temperatures were observed near plasma axis;  The potential confinement (HM-regime) was achieved quickly enough with plasma body formation and shown plasma density redoubling, axial plasma flows diminishing; side slit current depression; improvement of launch coupling with a plasma.  A parabolic radial profile of Ar1 spectrum line intensity was observed (not a flat one, as we might wait for at a large track length). This is the same shape of profile that was measured with probes in Argon plasma. This could be explained by potential distribution; we ought to remember that electrons (which could plane the potential distribution) are magnetized (in contrast to ions) practically throughout the whole volume of PNX-U. Because of these facts we can hope to get an Argon plasma density of 5 1012 cm in spite of usual Ar discharge, where radiation losses at Te ~ 10 eV would be too large. The explanation scheme of the phenomena could be as follows: the radiation cooling of Ar ions at near axis region increases with ion density; the ion temperature diminishing leads to the confinement time increasing (exponential !), i.e. to further increasing of a plasma density. The main energy losses in PNX-U experiments take place at a plasma periphery and can be done relatively small because of a thin periphery layer of relatively “warm” plasma and low ion confinement time at a periphery. Such a picture can be accepted for heavy ions (radiational discharge); its special investigation is necessary.

6 High degree of a plasma ionization was achieved. An intensive “plasma pumping” in PNX-U (gas pressure at a plasma boundary ~ 10-5 Torr ) was detected with a direct measurement of an Argon ion spectrum line intensity distribution. The optical measurement results have confirmed the results of model calculations about high ionization degree at a plasma axis. We can fearlessly say about more than 40% of the ionization in central plasma region Some apartness of the cusps which form the trap was detected. The new diagnostic possibilities have permitted us to detect spatial inhomogeneities of PNX-U plasma. It was found out that Te is higher in that cusps which are equipped with microwave inputs; plasma flow along the axis to North direction 1.5 time exceeds such a flow to South direction (both microwave inputs are displaced in North half of the installation); two points of gas puffing being displaced at opposite edges of PNX-U provide more uniform plasma flows to slits and edges. All that gives the foundation to recommend a “closed” configuration with near plasma copper liner and gas boxes behind slits. Such a configuration could provide microwave and gas input uniformity throughout the system.

4. PN-ITER preliminary design

The status of experimental achievements at PNX-U in compare with the designed and with the necessary for PN ITER parameter meanings is demonstrated the table below. PNX-SU is the proposed intermediate step which objectives are:  To achieve plasma parameters close to those necessary for PN-ITER.  To verify the analytical models used for the design of PN-ITER.  To test a PN device with parameters close to those of PN-ITER.  To gain technological and operational experience with the construction and running of the system similar to the full scale one. The design of PN-ITER is closely related to that of PNX-SU. Preliminary calculations have predicted the worst case power requirement of 700 kW for PN-ITER operated with deuterium. This value can be minimised if Ar will be acceptable. The future experiment have to specify in particular this number.

PNX-U PNX-SU PN ITER experiment design length, m 2.2 2.2 ~2 3 Plasma volume, m3 0.5 0.5 1 6 Linear density, m-2 ~2 х 1018 1.4 х 1018 1 х 1019 2.1 х 1019 Plasma density, m-3 1x1018 0.6x1018 7 x1018 7 x1018 Ionization degree ≥ 0.4 ≥0.2 0.3 ≥0.3 Microwave power, kW ~ 50 50 150 ≥500 Frequency, GHz 7 (klystrons) 24 (gyrotrons) Magnetic system copper Superconducting (NbTi) Magnetic field, T 0.36 0.5 1 >1

7 Fig.9. Plasma Neutralizer integrated Into ITER NBI

5. Conclusion We conclude that at any rate the cut off plasma density can be provided in the multicusp configuration with microwave near wall ECR discharge at low Argon pressure. The potential improvement of the cusp confinement allows to get High Moderegime of operation which gives a hope to get a plasma density more than the cut off one. An ionization degree more than 30% can be achieved at low plasma electron temperature and acceptable microwave power. PN could be integrated into existing ITER NBI design.

References

[1] ITER Design Description Document 1997, Joint Common Team Rep. N53DDD1557- 11-30W0.2 [2] Pamela J. 1995 Plasma Phys. Control. Fusion 37 A325 [3] Dimov G. I., Roslyakov G. V. 1975 Nuclear Fusion 15 551 [4] Dimov G. I., Ivanov A. A., Roslyakov G. V. 1980 Sov. Plasma Phys. 6 933 [5] Trow J. R., Moses K.J. 1986 Proc. 4th Int. Simp. Prod. and Neutral. Negative Ions, Brookhaven, p.651 [6] Hershcovitch A., Kovaric V., Prelec K. 1988 Proc. 3th Eur. Workshop on Production and Application of Light Negative Ions, Amsterdam, p.217 [7] Moses K.J., Trow J. R., Dooling J. C. 1990 Proc. 5th Int. Simp. Prod. and Neutral. Negative Ions, Brookhaven, p.717 [8] Yochikawa K., Yamamoto Y., Toku H., Komoda N. 1991 IAEA Committee Meeting on NBI, Japan [9] Skovoroda A. A. 1992 Plasma Neutralizer for ITER Injector, Rep. IAE-5544/6, RRC Kurchatov Inst., Moscow [10] Okumura Y. et al 2000 Proc. 18th IAEA Fusion Energy Conference, Sorrento, Italy, FTP1/18 at press [11] Leung K.N., Hershkowitz N., Mac Kenzie K.R. 1976 Phys.Fl. 19 1045 [12] Lagarde T., Pelletier J., Arnal Y. 1997 Plasma Sources Sci. Technol. 6 53 [13] Zhil’tsov V. A., Kosarev P. M., Skovoroda A. A. 1996 Plasma Phys. Rep. 22 246

8 [14] Lagarde T., Arnal Y., Pelletier J. 1997 Plasma Sources Sci. Technol. 6 386 [15] Gil’denburg V. B., Skovoroda A. A. 1998 Plasma Phys. Rep. 24 339 [16] Kuftin A. N., Flyagin V. A., Lygin V. K. et al 2000 Proc. Int. Workshop Strong Microwaves in Plasmas, Nizhny Novgorod, v.2 671 [17] Kulygin V. M., Skovoroda A. A., Zhil’tsov V. A. 1998 Plasma Devices and Operations 6 135 [18] Zhil’tsov V. A., Klimenko E. Yu., Kosarev P. M. et al 2000 Nuclear Fusion 40 No.3Y 509 [19] Kulygin V. M. et al 2000 Proc. 18th IAEA Fusion Energy Conference, Sorrento, Italy, FTP1/19 at press [20] Skovoroda A. A., Zhil’tsov V. A. 1997 Proc. Int. Workshop Strong Microwaves in Plasmas, Nizhny Novgorod, v.1 368 [21] Skovoroda A. A. and OGRA Team 2000 Proc. Int. Workshop Strong Microwaves in Plasmas, Nizhny Novgorod, v.1 371 [22] ITER Physics Basis 1999 Nucl. Fusion 39 2400 [23] Zubarev V. F., Panasenkov A. A. 2000 VANT, Ser. Thermonuclear synthesis, RRC Kurchatov Institute, Moscow N1 34 [24] Zhilt’sov V.A., Skovoroda A.A., Scherbakov A.G. 1991 Sov. Plasma Phys. 17 785 [25] Clark R., Abdalleh J., Post D.E. 1995 J. Nucl. Mater. 220-222 1016