P2-2

ICRF physics aspects of wall conditioning

with conventional antennas in large-size tokamaks

A. Lyssoivana*, R. Kocha, D. Douaib, J.-M. Noterdaemec,d, V. Philippse, V. Rohdec,

F.C. Schüllerf, G. Sergienkoe, D. Van Eestera, T. Wautersa,b, T. Blackmang, V. Bobkovc,
S. Brémondb, S. Brezinseke, E. de la Calh, R. Dumontb, M. Garcia-Munozc, E. Gauthierb,
M. Grahamg, S. Jachmicha, E. Joffrinb,i, A. Kretere, P.U. Lamallef, E. Lerchea, G. Lombardb,
F. Louchea, M. Maslovg, M.-L. Mayoralg,i, V.E. Moiseenkoj, P. Mollardb, I. Monakhovg,
J. Ongenaa, M.K. Paule, R.A. Pittsf, V. Plyusnink, W. Suttropc, E. Tsitroneb, M. Van Schoora,
G. Van Wassenhovea, M. Verviera, the TEXTOR Team, the TORE SUPRA Team,
the ASDEX Upgrade Team and JET EFDA Contributors♣

JET-EFDA, Culham Science Centre, Abingdon, OX14 3DB, UK

a
LPP-ERM/KMS, Association Euratom-Belgian State, 1000 Brussels, Belgium.
b
CEA, IRFM, Association Euratom-CEA, 13108 St Paul lez Durance, France.
c
Max-Planck Institut für Plasmaphysik, Euratom Association, 85748 Garching, Germany.
d
Gent University, EESA Department, B-9000 Gent, Belgium.
e
IEF-Plasmaphysik FZ Jülich, Euratom Association, 52425 Jülich, Germany.
f
ITER International Organization, F-13067 St Paul lez Durance, France.
g
CCFE/Euratom Fusion Association, Culham Science Centre, OX14 3DB, Abingdon, UK.
h
Laboratorio Nacional de Fusión, Association Euratom-CIEMAT, 28040 Madrid, Spain.
i
EFDA-CSU, Culham Science Centre, OX14 3DB, Abingdon, UK.
j
Institute of Plasma Physics, NSC KIPT, 61108 Kharkiv, Ukraine.
k
Instituto de Plasmas e Fusao Nuclear, Association EURATOM-IST, Lisboa, Portugal

Abstract
This paper focuses on a study of the principal operation aspects of standard ICRF heating

antennas in the Ion Cyclotron Wall Conditioning (ICWC) mode: (i) ability of the antenna to

ignite the cleaning discharge safely and reliably in different gases including those most likely

to be used in ITER – He, H2, D2 and their mixtures, (ii) the antenna capacity to couple a large

fraction of the RF generator power (>50%) to low density (~1016−1018 m-3) plasmas and (iii)

the RF power absorption schemes aimed at improved RF plasma homogeneity and enhanced

conditioning effect. The ICWC discharge optimization in terms of RF plasma wave

excitation/absorption resulted in successful simulation of the conditioning scenarios for ITER

operation at full field (JET) and half field (TEXTOR, TORE SUPRA, ASDEX Upgrade).
____________________________________________
PACS: 52.40.Hf, 52.50.-b, 52.55.Fa, 52.80.Pi
JNM keywords: Plasma-Materials Interaction, Gases, Helium, Plasma Properties
PSI-19 keywords: ITER, Wall conditioning, ICRF antenna, RF discharge
*Corresponding and presenting author address: Laboratory for Plasma Physics ERM/KMS,
30 Avenue de la Renaissance, 1000 Brussels, Belgium.
*Corresponding and presenting author E-mail: [email protected] (A. Lyssoivan).

-1-
 P2-2

♣
See the Appendix of F. Romanelli et al., Proceedings of the 22nd IAEA Fusion Energy
Conference 2008, Geneva, Switzerland
1. Introduction
Encouraging results obtained in present day tokamaks and stellarators (reviewed in Ref. [1])

have elevated ICWC to the status of one of the most promising techniques available to ITER

for routine wall conditioning in the presence of the permanent, high toroidal magnetic field

resulting from operation with superconducting magnets. The ability to operate in ICWC mode

has recently been confirmed as a functional requirement of the ITER main ICRF heating and

current drive system [2]. This paper focuses on multi-machine (TEXTOR, TORE SUPRA

(TS), ASDEX Upgrade (AUG) and JET) study of the principal RF physics aspects of

conditioning plasma production with standard ICRF heating antennas. It comprises (i) ability

of the antenna to generate Ez-field (along the toroidal magnetic field lines) and ignite the RF

cleaning discharge safely and reliably in different gases including those most likely to be used

in ITER – He, H2, D2 and their mixtures, (ii) antenna capacity to couple a large fraction of the

RF generator power (>50%) to low density (~1016−1018 m-3) plasmas and (iii) the RF power

absorption schemes (mainly collisional e-absorption) aimed at ICWC discharge performance

with improved plasma homogeneity and enhanced conditioning effect at ITER relevant

conditions.

2. Phases of RF discharge initiated with standard ICRF antenna

The initiation of an ICRF discharge in a toroidal magnetic field BT results from the absorption
B

~
of RF energy mainly by electrons. The RF E z -field (parallel to the BT-field) is considered to
B

be responsible for this process [3]. However, in the typical ICRF band (~10−100 MHz) in the

present-size fusion devices, for most of the antenna κ z -spectrum, the RF waves (cylindrical

modes) cannot propagate in the vacuum torus: κ ⊥2 = ω 2 c 2 − κ z2 <0, where κ ⊥ is the

perpendicular wave-vector, ω = 2πf , f is the RF generator frequency. Hence, the neutral gas
~
breakdown and initial ionization may only occur locally at the antenna-near E z -field

-2-
 P2-2

~
(evanescent in vacuum). Simplified analytical description of the antenna-near Ez -field

generation in vacuum [4] was found in a good agreement with numerical simulations done for

the real antenna configurations (JET A2 antenna [5]) using 3D electromagnetic codes (Fig.1).

Former analysis of the gas breakdown phase led to the following local ionization condition
~
[6]: (ω /e)(2meε i )1/2 ≤ E z (r ) ≤ meω 2 Lz /e . Here ε i is the ionization energy threshold for the
~ ~
molecules/atoms, Lz = E z (dE z dz ) is the parallel length scale of the RF ponderomotive
~
potential. Further analysis of the JET A2 antenna-near Ez -field amplitude in frame of this

model showed that the ionization condition in He (Ez~10 kV/m at ~1 cm outside the Faraday

screen) may be achieved with ~10 kV of applied RF voltage. This estimation was found in

agreement with the RF voltage operational range (≈12–14 kV) used for JET A2 antenna

operation in the RF plasma production mode [7]. After the first pre-wave local ionization

phase, as soon as plasma frequency ω pe becomes of the order of generator frequency,

ω pe ≥ ω , (it occurs at a very low density ~1012−1014 m-3 in the frequency range

10−100 MHz), the plasma waves (slow waves, SW) can start exciting/propagating in a relay-

race regime governed by the antenna κz-spectrum, causing further space ionization of the

neutral gas and plasma build-up over the torus (plasma wave phase). If plasma density

becomes high enough (ne>1018 m-3), the usual fast magneto-sonic wave (FW) can become

propagating for the typical antenna π-phasing operation. Because of the very low and frozen

plasma temperature during the ionization process (Te~2−5 eV [6]), the RF power is expected

to be dissipated mostly collisionally (random e-collisions with gas molecules, atoms or ions)

either directly or through conversion to SW and ion Bernstein waves (IBW) if ω > ωci or by

conversion at the Alfvén resonance if ω < ωci . Such a non-resonant coupling allows RF

plasma production at any BT [8].

3. Antenna-plasma coupling in ICWC mode of operation

The antenna-plasma coupling efficiency is the fraction of the generator power coupled to the

plasma, η=PRF-pl /PRF-G. The conventional ICRF antenna is designed for dense target plasma

-3-
 P2-2

(ne>1018 m-3) heating through excitation of FW with high coupling efficiency (η>0.9). Being

operated in the RF plasma production mode with the "plasma heating settings" (high kz-

spectrum of the radiated RF power), the conventional ICRF antenna gives evidence of poor

coupling (η0~0.2−0.3) to the low density RF plasmas ne~1016−1017 m-3, at which FW is

typically non-propagating. The present-day solutions for ICRF antenna enhanced coupling in

the ICWC mode are based on the development of scenarios with FW close to propagation or

propagating in low density plasmas: (i) antenna phasing to low kz-spectrum of the radiated RF

power, (ii) FW-SW-IBW Mode Conversion (MC) in RF plasmas with two ion species, (iii)

operation at High Cyclotron Harmonics (HCH), typically ω ≈ 10ωci . It should be noted that

the FW cutoff frequency strongly depends on the antenna kz-spectrum [9]. For the case of

TEXTOR ICRF antennas, it results in a dramatic reduction (about two orders) in the threshold

density for FW excitation results from changing the phase between RF current straps is

changed from π-phasing (ne-FW≈2.7×1018 m-3, kz≈6 m-1, f=32.5 MHz) to zero-phasing (ne-
16
FW≈7×10 m-3, kz<1 m-1, f=32.5 MHz), respectively. The recent ICWC experiments have

clearly demonstrated this effect as shown in Fig.2 for monopole-phasing (low kz, η η 0 ≈3−4)

and for higher kz. Another solution for rising the coupling is based on the effect of FW

conversion to SW and IBW in low density/temperature plasmas containing two ion species,

e.g. D++H+, 4He++D+ or 4He++H+ [8]. Generally, FW may be non-propagating over the

plasma cross-section except for the narrow conversion area located closer to the fundamental

ion cyclotron resonance (ICR) of the principal species, protons ( ω = ωcH + ) or deuterons

( ω = ωcD + ), depending on ICWC scenario. In this regime, antenna coupling becomes sensitive

to the radial location of MC layer: the closer MC layer to the antenna surface the higher

antenna-plasma coupling. This coupling effect is more pronounced for the standard π-phasing

antenna operation ( η η 0 >3) [8]. Antenna coupling to low density plasmas may also be

enhanced by increasing the ion cyclotron harmonic number by ~10 times either by decreasing

the BT-value or increasing the generator frequency. In terms of FW excitation, this effect is

-4-
 P2-2

attributed to strong shrinking of the wave evanescent layer at the plasma edge [4] and its

possible propagation even in the high kz-spectrum case, giving rise to the standard antenna

coupling, η η 0 ≈4 (Fig.3).
4. Impact of RF power absorption schemes on wall conditioning output

The progress achieved in the understanding of the antenna-plasma coupling characteristics in

the ICWC mode enabled to extend the ICWC operation window over a large range

(BT=0.2−3.3 T, BV=0−0.04 T, gas pressures ptot≈(0.2−8.0)×10-2 Pa, coupled power PRF-pl

B

≈10−230 kW and coupled power density PRF − pl V pl ≈0.3−3.0 kW/m3) and successfully

simulate ITER ICWC scenarios at half-field (TEXTOR, TS, AUG) and full field (JET) [10].

Simulation of the ITER ICWC scenarios in the present-day fusion machines means scaling of

the foreseen ITER BT-field (2.65T or 5.3T) and generator frequency band (40−55 MHz) to the

presently available values keeping the same ITER f/BT-ratio and on-axis location of the

fundamental ICR for protons/deuterons thus crossing over the divertor. The conditioning

output was studied by measuring the overall outgassing rate of several marker gases using

mass spectroscopy and/or optical penning gauges. Here we define the outgassing rate of a

given species as the quantity [11]: QRR (t ) ~ V (dp / dt ) + p ⋅ s + V (k d + k i ) pne , where V is the

volume, p and s are the partial pressure of the given mass and its pumping speed, respectively,

k i and k d are the ionization rate and dissociation rate and ne is the electron density.

In the case of H2 preloading at JET, analysis of the partial pressures for the mass 3 (HD) and

mass 2 revealed a noticeable increase in outgassing after the RF pulse termination for the

antenna low kz-spectrum case (0000-phasing) compared to 00ππ-phasing (JET D2-ICWC,

Fig.4). The net power coupled to the plasma was doubled (at the same generator power) and

(possibly partly due to this) higher and more homogeneous density was obtained. The H2-

injection into He plasmas made the conditioning more effective in the presence of the

fundamental ICR for protons (AUG (He+H2)-ICWC, Fig.5). The observed effect may be

attributed to the beneficial MC scenario [8]: (i) higher coupled power (~2 times) and (ii) the

-5-
 P2-2

improved plasma homogeneity due to radially extended (towards HFS) power deposition

profile for the electrons. The hydrogen induced chemical erosion/isotope exchange could also

contribute to the obtained result [12]. The conditioning effect at low BT≈0.2 T (HCH regime)

was comparable to that at high BT≈2.0−2.3 T (ICR presence). It may be related to the strongly

increased (~1.5−2 times) coupled power which is absorbed by the electrons through

collisional dissipation as TOMCAT code predicts [8]. The performed comparison of the

ICWC output indicates indirectly the minor role in conditioning process of the high energy

charge exchange (CX) atoms/ions usually generated in ICWC plasmas in the presence of ICR

and needs further investigation.

5. ICWC extrapolation to ITER

Modeling of the absorbed power scenarios for ICWC performance in ITER was performed

with the TOMCAT 1-D full-wave RF code [13] upgraded to low density/temperature non-

homogeneous plasmas and accounting for ion cyclotron, Landau damping and collisional

absorption mechanisms. The foreseen ITER magnetic field variation (BT=2.65–5.3 T) and

designed frequency band for the ICRF Heating and Current Drive (HCD) system (f=40–

55 MHz) gave a strong impact on selection of the ICWC operational parameters. The

TOMCAT code predicts that a more homogeneous power absorption by the electrons over the

ITER vessel may be achieved in the MC scenario at BT=3.6 T with two different frequencies

(f1 =40 MHz and f2 =48 MHz) and low-kz-spectrum (π/3-, π/6- or monopole-phasing) between

the RF currents in the toroidally adjacent antenna modules [8]. Performance of the MC

scenario at half-field (BT=2.65 T) or at full field (BT=5.3 T) may result in less homogeneous

ICWC discharge. However, plasma production with the antenna phased to low kz-spectrum of

the radiated RF power looks beneficial: (i) FW is already propagating in low density plasmas;

(ii) better antenna coupling is foreseen; (iii) larger fraction of the coupled RF power may be

transported to the antenna distant (>2 m) conversion layers. To improve the RF plasma

-6-
 P2-2

homogeneity, an application of the poloidal magnetic field (BV,R<<BT) and assistance with the
B B

on-axis ECRF ionization could be beneficial [8].

A 0-D plasma code [14] was used to simulate a scale of the RF power necessary to produce

and sustain ICWC hydrogen/deuterium plasmas in ITER-size machine ( a pl ≈2.4 m, R0=6.2 m)

in the presence of BT=5.3 T in the pressure range p≈(2−8)×10-2 Pa. The code predicts that RF
B

plasmas with density of ne≈(1−5)×1017 m-3, temperature Te≈1−2 eV and ionization degree

γi≈0.05−0.10 can be produced with the RF power coupled to the electrons in the range Ppl-

ITER≈0.3−1.5 MW depending on the gas pressure (Fig.6). Assuming an "optimistic" antenna

coupling efficiency η≥0.5 at the "monopole"-phasing, this corresponds to the generator power

range PG-ITER≈0.6−3.0 MW. The empirical direct extrapolation from the TEXTOR and JET

ICWC data (coupled power Ppl-TEXTOR≈12−30 kW, Ppl-JET≈230 kW, similar power density

scaling and antenna coupling) gives a power of Ppl-ITER≈1−2 MW and PG-ITER≈2−4 MW,

respectively.
6. Conclusions

The results from ion cyclotron wall conditioning experiments with conventional ICRF

antennas on several different tokamaks can be summarized as follows:

1. We have elaborated a general approach to the ICRF antenna functional

requirements for operation in the plasma production mode and established reliable working

parameters (antenna RF voltage and power, frequency, phasing and gas pressure) needed for

wall conditioning.

2. The found antenna operation solutions resulted in a successful simulation of ITER

ICWC scenarios at half-field (on the tokamaks TEXTOR, TORE SUPRA and ASDEX

Upgrade) and full field (on JET).

3. Extrapolation of the experimental data obtained on the various tokamaks,

complemented by simulations with 1-D RF and 0-D plasma codes indicate that the currently

planned ITER ICRF HCD system could be used for ICWC operations on ITER.

-7-
 P2-2

Acknowledgements

This work was supported by EURATOM and carried out within the framework of the

European Fusion Development Agreement. The views and opinions expressed herein do not

necessarily reflect those of the European Commission.

References
[1] E. de la Cal, E. Gauthier, PPCF 47 (2005) 197-218.
[2] ITER Team, ITER Design Change Request DCR-080 (2007).
[3] V.E. Golant, Usp. Fiz. Nauk 65 (1958) 39-86.
[4] A. Lyssoivan et al., Report LPP-ERM/KMS, vol. 114, 1998.
[5] A.Kaye, et al., Fusion Eng. Des. 24 (1994) 1.
[6] A. Lyssoivan, et al., Nucl. Fusion 32 (1992) 1361-1372.
[7] A. Lyssoivan, et al., J. Nucl. Mater. 337-339 (2005) 456-460.
[8] A. Lyssoivan, et al., J. Nucl. Mater. 390-391 (2009) 907-910.
[9] T. Stix, Waves in Plasmas, Springer-Verlag, N.Y. 1992.
[10] D. Douai, et al., this conference.
[11] T. Wauters, et al., this conference.
[12] G. Sergienko, et al., J. Nucl. Mater. 390-391 (2009) 979-982.
[13] D. Van Eester and R. Koch, PPCF 40 (1998) 1949-1975.
[14] J. Buermans and T. Matthys, Diploma Thesis, ERM/KMS, Brussels 2006.

-8-
 P2-2

Figure captions

Figure 1. Ez-field simulation for the JET A2 antenna with 3-D MWS code (f=30 MHz,
antenna straps in dipole-phasing, PRF-input=1 W).
Figure 2. Antenna coupling efficiency to ICWC plasmas (ne~3×1017 m-3) vs. kz-spectrum of
the radiated RF power for standard ICRF antennas.
Figure 3. Antenna coupling efficiency to low density (ne~3×1017 m-3) plasmas vs. ion
cyclotron harmonic number.
Figure 4. Effect of the antenna kz-spectrum on JET D2-ICWC output (H2 preloaded): 0000-
phasing (solid) vs. 00ππ-phasing (dashed), f=25 MHz, BT=3.3 T, p≈2×10-3 Pa.
Figure 5. AUG (He+H2)-ICWC vs. He-ICWC: f1=30 MHz, f2=36.5 MHz, BT=2.35T, p ≈
B

4×10-2 Pa, Ppl(He+H2)≈50kW, Ppl(He)≈30kW, AUG wall composition: 50%C+50%W.

Figure 6. 0-D simulation of the H2/D2-plasma production in ITER-size machine for averaged
power per neutral: P/N ≈ 70 kW/(Pa·m3), pH2=8×10-2 Pa, PRF-e=1.3 MW.

-9-
 P2-2

Figure 1

- 10 -
 P2-2

Figure 2

- 11 -
 P2-2

Figure 3

- 12 -
 P2-2

Figure 4

- 13 -
 P2-2

Figure 5

- 14 -
 P2-2

Figure 6

- 15 -