Shevelev DOI:10.1088/1741-4326/aa8cea EX/P7-42
Runaway Electrons Studies with Hard X-Ray and Microwave Diagnostics in the FT-2 Low-Hybrid Current Drive Discharges A. Shevelev1, E. Khilkevitch1, S. I. Lashkul1, V. Rozhdestvensky1, A. Altukhov1, D. Kuprienko1, I. Chugunov1, D. Doinikov1, L. Esipov1, D. Gin1, M. Iliasova1, V. Naidenov1, N. Nersesyan1, I. Polunovskiy1, A. Sidorov1, and V. Kiptily2 1Ioffe Institute, St. Petersburg, Russian Federation 2Culham Centre for Fusion Energy (CCFE), Culham Science Centre, Abingdon, UK Corresponding Author: A. Shevelev, [email protected]ffe.ru Analysis of the superthermal and runaway electrons behaviour in ohmic and low-hybrid current 19 3 drive FT-2 tokamak (R “ 0.55 m, a “ 0.08 m, Bt ď 3 T, Ip “ 32 kA, xny “ 1.9 ˆ 10 {m , f0 “ 920 MHz) plasmas has been carried out using information obtained from measurements of hard X-ray spectra and nonthermal microwave synchrotron radiation intensity in the frequency range 53–78 GHz r1s. A gamma-ray spectrometer developed for gamma-ray diagnostics of ITER (Nuclear Facility INB-174) and based on LaBr3(Ce) scintillator has been used in measurements of hard X-ray emission (E ą 0.1 MeV) generated by runaway electrons. An advanced digital processing algorithm of the detector signal recorded with high sampling rate has provided a pulse height analysis at rates exceeding 107{s. A spectrum deconvolution code DeGaSum has been used for reconstruction of the energy distribution of runaway electrons escaping from the plasma and interacting with materials of the FT-2 limiter in the vacuum vessel r2s. The developed digital signal processing technique for LaBr3(Ce) spectrometer has allowed studying the evolution of runaways energy distribution in the FT-2 plasma discharges with time resolution of 1 ms. Superthermal electrons accelerated up to 2 MeV by the L-H waves at the high-frequency pumping of the plasma 13 3 with low density xney „2 ˆ 10 {cm and then up to 6 MeV by vortex electric field have been found. A correlation between the hard X-ray and synchrotron radiations as well as a role of MHD activity is discussed. Analysis of the runaway electron beam generation and evolution of their energy distribution in FT-2 plasmas has been presented in the report. References r1s V. V. Rozhdestvensky, et al., Energy Environ. Eng. 3(3), 42-49, (2015). r2s A. E. Shevelev, et al., Nucl. Fusion 53, 123004 (2013). This work was supported in part by the RF State Contract No. N.4k.52.9B.14.1002 and the Russian Foundation for Basic Research projects Nos. 13-08-00411 and 14-08-00476.
Published as a journal article in Nuclear Fusion http://iopscience.iop.org/article/10.1088/1741-4326/aa8cea 1 EX/P7-42
Runaway Electron Studies with Hard X-Ray and Microwave Diagnostics in the FT-2 Low-Hybrid Current Drive Discharges
A.E. Sheveleva, E.M. Khilkevitcha, S.I. Lashkula, V.V. Rozhdestvenskya, A.B. Altukhova, D.V. Kouprienkoa, I.N. Chugunova, D.N. Doinikova, L.A. Esipova, D.B. Gina, M.V. Iliasovaa, V.O. Naidenova, N.S. Nersesyana, I.A. Polunovskiya, A.V. Sidorova and V.G. Kiptilyb aIoffe Institute, Politekhnicheskaya 26, St Petersburg 194021, Russian Federation bCCFE, Culham Science Centre, Abingdon, Oxon, X14 3DB, UK
E-mail contact of main author: [email protected]
Abstract Analysis of the super-thermal and runaway electrons behavior in ohmic and low-hybrid current drive FT-2 19 -3 tokamak (R0 = 0.55 m, a = 0.08 m, BT ≤ 3 T, Ipl = 32 kA,
1. Introduction
Studying occurrence and behavior of runaway electrons (RE) is one of the most important tasks ensuring safe operations of tokamak facilities. Avoiding a massive RE generation is especially important in large machines like ITER (Nuclear Facility INB-174), where the RE current can reach 10 MA and the energy transferred by fast electrons could exceed 200 MJ in a disruption. Very few diagnostics have capabilities of the RE energy estimation. One of them is hard X-ray (HXR) spectrometry, which allows inferring a maximum energy of accelerated electrons and, in some cases, RE energy distribution and a current transferred by the RE beam. The HXR spectrometry allows studying the response of RE on MHD instabilities, changes of the plasma shape etc. Analysis of the super-thermal and runaway electrons behavior in ohmic and low- hybrid current drive (LHCD) FT-2 tokamak (R = 0.55 m, a = 0.08 m, BT ≤ 3 T, Ipl =32 kA, 19 -3
2 EX/P7-42 in the frequency range (53 ÷ 78) GHz [1]. A gamma-ray spectrometer based on LaBr3(Ce) scintillator was used in measurements of hard X-ray emission (E > 0.1MeV) generated by runaway electrons [2]. LaBr3(Ce) provides very high light yield (63000 photons/MeV), perfect energy resolution (3 % on 662 keV line) and very fast decay time (~20 ns) [3]. In previous works [4, 5] it was demonstrated that the performance of LaBr3(Ce) detector can reach counting rates of a few MHz. LaBr3(Ce) spectrometers are proposed to be used in gamma-ray diagnostic systems at ITER [6, 7].
2. Experimental setup
Experimental setup for studies of accelerated electrons in the FT-2 tokamak was described in detail in [1, 2]. The detector with Ø25.4×76.2 mm LaBr3(Ce) crystal was placed in the experimental hall at a distance of about 4.5 meters from the vacuum vessel. The detector was surrounded by a lead shield to protect the spectrometer from scattered gamma-ray radiation and can observe the poloidal limiter in the vacuum vessel of the tokamak through the 5-mm collimator. A scheme of the hard X-ray detector arrangement on the FT-2 tokamak is shown in FIG.1.
FT-2 shot #060116_21
1000 1000 c) 29-31 ms a) 25-27 ms 1E13 1E13 e 100 100 Limiter 1E12 1E12 /dE /dE /dE (1/MeV) /dE HXR HXR /dE (1/MeV) /dE 10 1E11 RE 10 1E11 RE dN dN 1E10 1E10 dN (countsperchannel) dN (countsperchannel) 1 1 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 1000 1000 b) 27-29 ms e) 31-33 ms FT-2 1E13 1E13 30° 100 100 1E12 1E12
chamber /dE /dE HXR HXR /dE (1/MeV) /dE
1E11 (1/MeV) /dE 1E11
10 10 RE RE dN dN dN 1E10 (countsperchannel) 1E10 (countsperchannel) dN 1 1 Detector 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Energy (MeV) Energy (MeV)
FIG. 1. Scheme of the HXR detector FIG. 2. Measured HXR spectra (black dots) and arrangement on the FT-2 tokamak with two reconstructed RE distributions (red lines) for different poloidal limiters. time intervals of FT-2 shot #060116_21 with 96 kW LHCD.
A high light output of LaBr3(Ce) crystal provides a very good energy resolution of the detector of 3.5% for 662-keV gamma-ray line. Dimensions of LaBr3(Ce) detector have been optimized to reduce a scattered background gamma radiation. The similar detector was considered for use in vertical gamma-ray spectrometers of ITER [6]. An advanced digital processing algorithm of the detector signal recorded with high sampling rate has provided a pulse height analysis at rates exceeding 107 s-1. A spectrum deconvolution code DeGaSum [7, 8] was applied for reconstruction of the energy distribution of runaway electrons escaping from the plasma and interacting with stainless steel materials of the FT-2 limiter in the vacuum chamber. The code DeGaSum uses the maximum likelihood estimation using expectation maximization method (ML-EM) [9], known as Richardson-Lucy method [10, 11]. Hard X-ray spectrum y(ε) measured by the detector can be represented in the following convolution form
3 EX/P7-42
(1)
where ε, ε’, ε’’ – energies; n(∫ε) – statistical noise;∫ f - a runaway electron distribution function; hd - a gamma-ray detector response function; he is HXR generation function, i.e. an energy dependent density of probability for a bremsstrahlung emission produced by mono-energetic runaway electrons in the plasma volume of the detector field of view. This he function is calculated with MCNP (Monte Carlo N-Particle) code modelling the interaction of mono-energetic electrons with a solid target, bremsstrahlung emission production and transport of the emission in the detector. Super MC software also was used in the calculations [12]. More information about application of DeGaSum code for reconstruction of runway distributions can be found in [2, 7]. The applied technique allows analyzing the evolution of fast electron distributions in hot plasmas with time resolution of 1-5 ms. FIG. 2 shows examples of HXR spectra registered during LHCD run with 96 kW (black dots) and the reconstructions of escaped RE energy distributions (red lines). Microwave radiation in the FT-2 tokamak was investigated in the range of magnetic broadening of the first and second harmonics of thermal electron gyro-frequency and plasma frequency of electrons. A direct gain receiver (DR) with detector sections for wavelengths λ = 3, 1.5, and 0.8 –8 cm and sensitivity Pth ~ 10 W and a horn antenna located outside the discharge chamber from the low toroidal magnetic field side were used at the comparatively low-frequency microwave radiation recording. Simultaneously, heterodyne radiometers were used for registration of the 2 and 4 mm wavelength radiation. This allowed observing the appearance and correlation of MR flashes in the frequency range of f ~ 53 ÷ 78 GHz.
3. Runaway studies at FT-2 during LHCD
Waveforms of a typical LHCD discharge (red lines) in comparison with the waveforms of discharge with only ohmic plasma heating (OH) (blue lines) are shown in FIG.3. The LHCD pulse with 5 ms duration was applied on 30 ms of the shot. During the additional heating, the HXR emission (FIG.3c) increased due to super-thermal electrons generated by low-hybrid wave (LHW) rise and the plasma column slightly shift to the outer wall of the vacuum vessel. Usually, during LHCD, when Evortex drops, a re-appearance of the low-energy electrons (ERE < 2 MeV) is observed (FIG.3f). This appearance of the low-energy electrons correlated with the synchrotron radiation signal (FIG.3e). One can suppose that the appearance of low energy RE (ERE < 2 MeV) was caused by the RF wave during the LHCD, which was feeding the population of super-thermal electrons from the plasma. These electrons are accelerated in the vortex electric field and contribute to the HXR emission after RF pulse or at MHD instability during RF pulse. It seems, the enhanced output of RE during LHCD, as well as a noticeable increase of electron density (FIG.3 b), resulted in a reduction of the total number of RE, which escaped on the limiter right after RF pulse (FIGs.3f, 3g), while in the conventional ohmic discharge at that moment there was the birth of the new accelerated electrons at the plasma density decrease. Maximum energies of RE escaping onto the limiter (FIG.3h) have been derived in the course of analysis of electron energy distributions obtained during the deconvolution procedure.
4 EX/P7-42
OH #060215_10 (blue) and 30 a) Ip OH+LHCD #060215_15(red) FT-2 shots 20 kA 10 0 RF b)
-1 5 s
10
) d) dN /dt
6 HXR 1 4 0.1 (MeV) (*10
e) I max RE 3 synch 74GHz 70GHz E 2x103 2 a.u. 0 13 1x10 f) N (E <2 MeV) 1 RE RE 1x1012 RF 1x1011 0 1/(2 ms) 20 30 40 50 13 1x10 g) N (E >2 MeV) Time (ms) RE RE 1x1012 1x1011 1/(2 ms) 12 h) Emax 8 RE
MeV 4 0 10 20 30 40 50 60 Time (ms)
FIG. 3. Waveforms of OH FT-2 discharges FIG. 4. Measured maximum energies of electrons without (#060215_10; blue lines) and with escaped onto the limiter for the FT-2 discharges LHCD (#060215_15; red lines): a) Ipl - plasma with different input LH power. LHCD run is current; b)
RE energy distributions studies were carried out in shots with different power of LHCD varied from 66 to 160 kW. LHCD pulse was run at 25-33 ms of every shot. Evolution of maximum energy of runaways escaped onto limiter at shots with 66 (black line), 96 (red line) and 160 kW (blue line) input energy is shown in FIG.4. During the studies clear correlation between input LHCD power and Emax was observed only at low LH power input, when Zeff varied weakly at RF pulse. Deceleration of Emax ramp-up in this case mainly was caused by loop voltage decrease. At higher LH power the absence of this effect is probably connected with the fact that Upl during RF pulse is observed together with the rise of Zeff [13]. As it can be seen in FIG.4 in this case Emax continuously increased in the range of uncertainties up to 6-6.5 MeV at 48-50 ms thereafter it gradually decreased. As it is shown in FIG.5 during the RF pulse of discharge #071916_13 flashes of 10 GHz radiation, looking like sawtooth oscillations, correlated with spikes on Mirnov coil signal. The measurements were carried out with 1 μs time resolution. It was found these flashes also
5 EX/P7-42 correlated with bursts on the HXR monitor signal and with the LaBr3(Ce) spectrometer signal. Typical flashes reflect the same periodicity of collective processes as the data presented in FIG.5, that in the future will be used in the study of the instability development. Use of LaBr3(Ce) spectrometer allowed studying plasma processes with extremely short time resolution. Time trace of the detector count rate during LHDC pulse of shot #060116_21 with “sawtooth” oscillations is represented in FIG.6a with a time step of 10 μs. Preliminary analysis of energy distribution of RE escaped onto the limiter during the oscillations has shown that maximum energy of runaways in the bursts (FIG.6b, red line) exceeds RE maximum energy between the spikes approximately by 0.5 MeV (1.25 vs. 1.75 MeV). This could be explained by significant increase of electron diffusion during the “sawtooth” crashes, what leads to escaping RE from the plasma core, where electrons accelerated on the earlier stages of the discharge could be confined.
FT-2 shot #071916_13 FT-2 shot #060116_21 a) 4000 a) 10GHz signal s) 3000 20 /dt /dt
a.u. 2000 HXR 10
1000 dN
0 per10(counts 0 26.5 27.0 27.5 28.0 60 HXR 100 1E13 50 b) Time (ms) RF 40 b) 1E12
a.u. 30
/dE /dE 10 20 1E11 HXR /dE (1/MeV) /dE RE
10 dN 1E10
0 dN 1 (counts perchannel)(counts 1E9 0.0 0.5 1.0 1.5 2.0 2.5 3.0 400 c) MHD (Mirnov coil) 100 Energy (MeV) 1E13 200 0 c) 1E12 a.u. -200 /dE 10 1E11 HXR
-400 (1/MeV) /dE dN RE -600 1E10 dN 24 26 28 30 32 34 (counts perchannel)(counts 1 1E9 Time (ms) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Energy (MeV)
FIG. 5. Signals of FT-2 shot #071916_13: a) signal FIG. 6. HXR data obtained on 26.5-28.5 ms of of 10 GHz microwave radiometer; b) HXR monitor FT-2 shot #060116_21: a) Time trace of (black) and RF pulse of LHCD (red); c) Mirnov LaBr3(Ce) detector count rate; b) HXR spectrum coil signal recorded in the “sawtooth” bursts (black dots), reconstructed RE distribution (red line) and result of convolution of RE distribution with the detector response function (blue line); c) HXR spectrum recorded between the “sawtooth” bursts (black dots), reconstructed RE distribution (red line) and result of convolution of RE distribution with the detector response function (blue line)
6 EX/P7-42
Results of energetic and numerical analysis of RE beam generation and evolution in the FT-2 shot #060116_21 with input of 96 kW LHCD are represented in FIG.7. Increase of synchrotron radiation (blue line in FIG.7e) evidences about the rise of number of accelerated electrons during LHCD, which started on 25 ms of the shot. Deconvolution of HXR spectra demonstrated increase of low energy RE (ERE < 2 MeV) since 29 ms (FIG.7b). These electrons could be created by RF pulse (FIG.7e, red line) on plasma periphery and escaped from plasma due to development of MHD m=2/n=1 mode.
orbital shift limit FT-2 shot #060116_21 possible Emax a) measured Emax ripple-resonance n=1 12 8
(MeV) 4
RE 0 E E <2 MeV RE 1E14 b) E >2 MeV 1E13 RE 1E12 (a.u.)
RE 1E11
N 1E10 I p c) 1 30 Mirnov 20 0
(kA) 10 p MHD (a.u.) I bursts MHD 2/1 0 -1 )
U -3 pl 12 d) 4
U 4 (*10 e
0 0 n I e) synch. 3000 RF 100 2000
(a.u.) 50 1000 synch RF (a.u.)
I 0 0 10 20 30 40 50 60 70 Time (ms)
FIG. 7. Signals of FT-2 shot #060116_21: a) Measured RE Emax (black line) and Emax limited by orbital shift effect (red line), resonant energy for n=1 magnetic field ripple interaction (green line) and possible maximum energy of electrons accelerated by vortex electric field (blue); b) Evolution of numbers of electrons escaped onto limiter with 0.4
Possible causes limiting the maximum energy of RE during discharges with LHCD were investigated in the FT-2 experimental campaign. FIG.7a shows a graph of the measured RE maximum energy (black line). The blue line shows the maximum energy, which can be reached by runaways accelerated by vortex electric field (electron scattering and synchrotron radiation losses were not taken onto account). Red line represents the maximum energy that could be achieved by relativistic electrons at maximum orbital shift value (“the orbital shift limit”) [14].
7 EX/P7-42
The maximum energy limited by the orbital shift (assuming a flat current profile) is defined by the formula 1/2 2 r I p S 2R0 1 1 ,(2) rr17000 ll where R0 – major radius, Ip – plasma current, r – initial radius of accelerated electron measured from magnetic axis, rl – radius of limiter. The maximum possible electron kinetic energy ES=(γS - 1)mc2 in the orbital shift limit is achieved at the birth of the accelerated electrons on the magnetic axis r = 0. The RE maximum energy continuously increases in the limit of the maximum energy of electrons generated at the current ramp-up stage and accelerated by vortex electric field. When the plasma current ramped down, Emax decreased together with the energy of orbital shift limit. Another possible cause of RE maximum energy limitation could be a resonant interaction between their relativistic down-shifted cyclotron frequency ωce and the magnetic field ripple due to the finite number Nc of toroidal field coils. The resonant interaction between the electron gyromotion and the n-th toroidal harmonic will take place at an electron parallel momentum [15]
eB R p 00,(3) ||n nNc where B0 is the toroidal magnetic field. Resonant electron energy for n = 1 of FT-2 magnetic system shown in FIG.7a by green line could reach up to 13 MeV. As to measured maximal electron energy (black line), e. g., for 30 ms where Emax = 3 MeV, during LHCD run it can be described by resonances with n = 3 and 4. The effect of RE Emax limitation caused by interaction with the magnetic field ripple could be studied in future FT-2 campaigns with application of HXR spectrometry.
4. Summary
A gamma-ray spectrometer developed for gamma-ray diagnostics of ITER and based on LaBr3(Ce) scintillator has been used in measurements of hard X-ray emission generated by runaway electrons in the FT-2 tokamak discharges with LHCD. An advanced digital processing algorithm of the detector signal recorded with high sampling rate has provided a pulse height analysis at counting rates up to 107 s-1. A spectrum deconvolution code DeGaSum was used for reconstruction of the energy distribution of runaway electrons escaping from the plasma and interacting with materials of the FT-2 limiter in the vacuum vessel. The analysis of the LHCD discharges shows that the RE distribution function obtained with the HXR diagnostics is consistent with changes of the FT-2 plasma parameters. Evolution of runaway electron Emax in shots with LHCD was investigated with time resolution of 1-5 ms. During the studies clear correlation between input LHCD power and Emax was observed only at low LH power input, when Zeff varied weakly at RF pulse. Deceleration of Emax ramp-up in this case mainly was caused by loop voltage decrease. Bursts looking like sawtooth oscillations were observed on MHD and HXR signals during LHCD runs. Differences in RE energy distributions registered in bursts and between them were observed. The views and opinions expressed herein do not necessarily reflect those of the ITER Organization.
8 EX/P7-42
Acknowledgements
This work was fulfilled on faThis work was supported in part by the RF State Contract No. N.4k.52.9B.14.1002 and the Russian Foundation for Basic Research projects Nos. 13-08-00411 and 14-08-00476
References
[1] ROZHDESTVENSKY, V.V., et al., ”Nonthermal Microwave Emission Features under the Plasma Ohmic Heating and Low-hybrid Current Drive in the FT- 2 Tokamak”, Energy Environ. Eng. 3 (2015) (3) 42-49, [2] SHEVELEV, A.E., et al., “High performance gamma-ray spectrometer for runaway electron studies on the FT-2 tokamak”, Nucl. Instr. and Meth. Phys. Res. A 830 (2016) 102-108 [3] BrilLanCeTM Scintillators Performance Summary, http://www.crystals.saint-gobain.com/uploadedFiles/SG- crystals/Documents/Technical/SGC%20BrilLanCe%20Scintillators%20Performance%20 Summary.pdf [4] LOHER, B., et al., “High count rate gamma-ray spectroscopy with LaBr3:Ce scintillation detectors”, Nucl. Instr. and Meth. Phys. Res. A 686 (2012) 1 [5] NOCENTE, M., et al., “High Resolution Gamma Ray Spectroscopy at MHz Counting Rates With LaBr3 Scintillators for Fusion Plasma Applications”, IEEE Trans. Nucl. Sci. 60 (2013) 1408 [6] CHUGUNOV, I.N., et al., “Development of gamma-ray diagnostics for ITER”, Nucl. Fusion 51 (2011) 083010 [7] SHEVELEV, A.E., et al., “Reconstruction of distribution functions of fast ions and runaway electrons in fusion plasmas using gamma-ray spectrometry with applications to ITER” Nucl. Fusion 53 (2013) 123004 [8] KHILKEVITCH, E.M., et al., “Application of deconvolution methods to gamma-radiation spectra of thermonuclear plasma”, Tech. Phys. Lett. 39 (2013) (1) 63 [9] VARDI, Y., et al., “A statistical model for positron emission tomography”, J. Am. Stat. Assoc. 80 (1985) 8 [10] RICHARDSON, W.H., “Bayesian-Based Iterative Method of Image Restoration”, J. Opt. Soc. Am. 62 (1972) (1) 55-59 [11] LUCY, L.B., “An iterative technique for the rectification of observed distributions”, Astron. J. 79 (1974) 745 [12] WU, Y., et al., “CAD-based Monte Carlo program for integrated simulation of nuclear system SuperMC”, Annals of Nuclear Energy 82 (2015) 161-168. [13] LASHKUL, S.I., at al., “Impact of isotopic effect on density limit and LHCD efficiency in the FT-2 experiments”, Nucl. Fusion, 55 (2015) 7 [14] KNOEPFEL, H., and SPONG, D.A., “Runaway electrons in toroidal discharges”, Nucl. Fusion 19 (1979) 785 [15] MARTIN-SOLIS, J. R., et al., “Energy limits on runaway electrons in tokamak plasmas”, Phys. of Plasmas 6 (1999) (1) 238