Improvement of uniformity of the negative ion beams by tent-shaped magnetic field in the JT-60 negative ion sourcea) Masafumi Yoshida, Masaya Hanada, Atsushi Kojima, Mieko Kashiwagi, Larry R. Grisham, Noboru Akino, Yasuei Endo, Masao Komata, Kazuhiko Mogaki, Shuji Nemoto, Masahiro Ohzeki, Norikazu Seki, Shunichi Sasaki, Tatsuo Shimizu, and Yuto Terunuma Citation: Review of Scientific Instruments 85, 02B314 (2014); doi: 10.1063/1.4830365 View online: http://dx.doi.org/10.1063/1.4830365 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/85/2?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Long pulse production of high current D − ion beams in the JT-60 negative ion sourcea) Rev. Sci. Instrum. 79, 02A519 (2008); 10.1063/1.2821508 Uniform H − ion beam extraction in a large negative ion source with a tent-shaped magnetic filtera) Rev. Sci. Instrum. 79, 02C111 (2008); 10.1063/1.2816968 Improvement of beam uniformity by magnetic filter optimization in a Cs-seeded large negative-ion source Rev. Sci. Instrum. 77, 03A515 (2006); 10.1063/1.2165768 Improvement of a large negative ion source for the Large Helical Device neutral beam injector Rev. Sci. Instrum. 73, 1087 (2002); 10.1063/1.1430528 Negative ion beam production by a microwave ion source equipped with a magnetically separated double plasma cell system Rev. Sci. Instrum. 71, 1125 (2000); 10.1063/1.1150404
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 132.177.228.65 On: Wed, 18 Feb 2015 12:58:23
REVIEW OF SCIENTIFIC INSTRUMENTS 85, 02B314 (2014)
Improvement of uniformity of the negative ion beams by tent-shaped magnetic field in the JT-60 negative ion sourcea) Masafumi Yoshida,1,b) Masaya Hanada,1 Atsushi Kojima,1 Mieko Kashiwagi,1 Larry R. Grisham,2 Noboru Akino,1 Yasuei Endo,1 Masao Komata,1 Kazuhiko Mogaki,1 Shuji Nemoto,1 Masahiro Ohzeki,1 Norikazu Seki,1 Shunichi Sasaki,1 Tatsuo Shimizu,1 and Yuto Terunuma1 1 2
Japan Atomic Energy Agency, 801-1, Mukoyama, Naka 311-0193, Japan Princeton Plasma Physics Laboratory, Princeton, New Jersey 08543, USA
(Presented 10 September 2013; received 12 September 2013; accepted 12 October 2013; published online 9 December 2013) Non-uniformity of the negative ion beams in the JT-60 negative ion source with the world-largest ion extraction area was improved by modifying the magnetic filter in the source from the plasma grid (PG) filter to a tent-shaped filter. The magnetic design via electron trajectory calculation showed that the tent-shaped filter was expected to suppress the localization of the primary electrons emitted from the filaments and created uniform plasma with positive ions and atoms of the parent particles for the negative ions. By modifying the magnetic filter to the tent-shaped filter, the uniformity defined as the deviation from the averaged beam intensity was reduced from 14% of the PG filter to ∼10% without a reduction of the negative ion production. © 2013 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4830365] I. INTRODUCTION
In JT-60 Super Advanced for the fusion experiment, 22 A, 100 s D− ions are designed to be extracted from the largest ion extraction area in the world of 45 cm × 110 cm. One of the key issues for producing such as high current beams is a uniform production of the negative ions over the large ion extraction area.1, 2 Otherwise, the beam divergence might be locally degraded, which causes a significant direct interception of the beams and results in the limitation of the beam power and pulse duration time.3, 4 In the past studies,5–8 non-uniformity of the negative ions is experimentally found to be caused by a localization of the arc plasmas with H+ ions and H0 atoms which are converted to H− ions on the surface of plasma grid covered with cesium. It is also clarified from the electron trajectory calculation that this localization of the plasma is due to the magnetic drift of the primary electrons emitted from filaments. Based on these results, the magnetic filter in the negative ion source has been newly developed and tested.8 A tent-shaped filter is newly devised in order to improve the uniformity of the negative ions by suppressing the magnetic drift of primary electrons. In this paper, we report the design of the tent-shaped filter and the experimental results before and after the modification of the magnetic filter. II. MAGNETIC FIELD IN THE JT-60 NEGATIVE ION SOURCE A. The JT-60 negative ion source
Figure 1 shows schematic view of the arc chamber and the extractor of the JT-60 negative ion source. This is the a) Contributed paper, published as part of the Proceedings of the 15th
International Conference on Ion Sources, Chiba, Japan, September 2013. b) Author to whom correspondence should be addressed. Electronic mail:
[email protected] 0034-6748/2014/85(2)/02B314/4/$30.00
largest negative ion source in the world that has a semicylindrical arc chamber of 640 mm in diameter and 1220 mm in length, and ion extraction area of 450 mm in width × 1100 mm in length.9 Tungsten filaments of 48 are longitudinally placed at 50 mm from the inner of the side wall as shown in Fig. 1(a). Cesium is injected at the mid-way of the chamber to enhance the negative ion production. In the arc chamber, H+ /D+ ions and H0 /D0 atoms are produced and converted to H− /D− ions on the surface of plasma grid (PG) covered with cesium. The negative ions are extracted from an ion extraction area of (X, Y) = (±228 mm, ±580 mm), where there are 1080 apertures of 14 mm in diameter. Figure 1(b) shows diagnostic position of the Langmuir probes and line of the sight in emission spectroscopy on the plasma grids for the measurements of the plasmas. The plasma parameters, such as ion saturation current (Jis), electron temperature (Te), and density (Jes) were measured by 28 Langmuir probes, which were made of Mo with surface area of 0.07 cm2 (ϕ = 3 mm). The probes were located at 15 mm apart from the PG surfaces. The light intensity of the source plasmas in the longitudinal direction was measured through 5 optical fibers by spectroscopy. The measured spectra were ranged in 400–700 nm. The H− ions were extracted and dumped on the beam target. The beam target was made of copper with surface area of 1200 × 600 mm2 and thickness of 10 mm and located at 4.5 cm downstream from the extraction grids. Distribution of temperature rise on the beam target was measured by IR camera, from which the beam intensities of the beamlets were evaluated. The beam current density of the H− ion beams was defined by dividing a total H− ion beam current measured by a water calorimeter of the target by total areas of the ion extracted apertures. Figure 2(a) shows horizontally schematic view of the conventional magnetic configuration (PG filter) in JT-60 negative ion source. In the ion source, the arc chamber is fully
85, 02B314-1
© 2013 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 132.177.228.65 On: Wed, 18 Feb 2015 12:58:23
02B314-2
Yoshida et al.
Rev. Sci. Instrum. 85, 02B314 (2014)
(a)
0.04
Seg1
Seg2
Seg3
Seg4
Seg5
|B|
0.03
Tent shaped filter 0.02 0.01
Electron population [a.u.]
0 -600
(b)
PG filter
-400
-200
0
200
400
600
2.0
PG filter
Tent shaped filter
1.0
0 -600
-400
-200
0
200
400
600
Longitudinal direction [mm] FIG. 3. Comparison of the longitudinal profiles of (a) magnitude of magnet (|B|) and (b) electron population calculated by 3D finite-element magnetostatic calculation codes and Omni Trak codes (b) in the external and tentshaped filters.
B. Design of tent-shaped filter
FIG. 1. (a) Schematic view of longitudinal cross-section of JT-60 negative ion source. (b) The positions of the Langmuir probes and line of the sight in emission spectroscopy on the plasma grids.
surrounded by asymmetric 26 rows of the permanent magnets on the side wall. In addition, the PG filter is created by longitudinally flowing current into the plasma grids. Magnets are also installed in an extraction grid to suppress the extracted electrons with the negative ions. In the figure, vector diagram of the magnetic field of XZ-plane at Y = 0 mm with the PG current of 5 kA is also shown. The magnetic field is asymmetrically formed in the horizontal direction by the PG filter.
Figure 2(b) shows a new design of the tent-shaped magnetic field. This tent-shaped filter is formed by symmetrical arrangement of the permanent magnets surrounding the wall without the PG current. In the filter, horizontally symmetric magnetic fields and longitudinally asymmetric magnetic drifts are formed, and largely different from that of the PG filter, where the asymmetric magnetic fields and the symmetric magnetic drift are formed. Figures 3(a) and 3(b) show longitudinal profiles of the magnitude of the magnets (|B|) on the PG surfaces and populations of the primary electrons for the conventional PG filter and the tent-shaped filter, respectively. Magnum for 3D finite-element magnetostatic calculation codes and Omni Trak codes were used for these calculations.10 In both filters, the primary electrons are confined by strong magnetic fields around edges of the chamber. The filaments are immersed in relatively strong magnetic field formed by the magnets around the side wall, and the primary electrons emitted from the filaments drifts via the magnetic drifts. Since the source plasmas were mainly produced by the primary electrons, the trajectory of the primary electrons is calculated to investigate uniformity of the primary electrons as shown in Fig. 3(b). In the calculations, the primary electrons from the filaments were assumed to be emitted from the 48 filaments with arc voltage of −100 V. In the tent-shaped filter, the localization of the primary electrons is suppressed and the nonuniformity of the trajectory of the primary electrons is improved, compared with that in the PG filter where the primary electrons are localized in the 1st segment (Seg. 1) due to a one-way magnetic drift of the electrons.
III. EXPERIMENTAL RESULTS AND DISCUSSIONS A. Uniformity of source plasmas FIG. 2. Horizontally schematic view of two magnetic fields in JT-60 negative ion source; (a) a conventional PG filter and (b) a newly applied tent-shaped filter.
The plasma uniformities for the PG filter and the tentshaped filter were measured and compared. Figure 4 shows
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 132.177.228.65 On: Wed, 18 Feb 2015 12:58:23
Yoshida et al.
(a)
(b)
Jes [mA/cm2]
Te [eV]
1.5
Jis [A/m2]
(c)
Seg1
1
Rev. Sci. Instrum. 85, 02B314 (2014)
Seg2
Seg3
PG filter
Seg4
Seg5
80
Beamcurrent density extracted from each aperture [A/m2]
02B314-3
Tent-shaped filter
0.5 0 2.5 2 1.5 1 0.5 0
600 400
Seg4
Seg5
60 50 40
PG filter
30 20 10 0 -600
-400
-200
0
200
400
600
Longitudinal direction [mm]
1
FIG. 5. Comparison of the longitudinal profiles of the beam current density extracted from each aperture measured by the IR camera in the PG filter and in the tent-shaped filter with the Cs seeded conditions.
0.5
-400
-200
0
200
400
600
Longitudinal direction [mm]
FIG. 4. Longitudinal profiles of the typical plasma parameters; (a) Te, (b) Jes, (c) Jis (H+ ) ions at X = 0 mm in the new designed tent-shaped filter, compared with those in the conventional PG filter, and (d) H0 atoms determined by using measured Hα emission intensities and CR model.
longitudinal profiles of the typical plasma parameters; (a) Te, (b) Jes, and (c) Jis (H+ ions) on the PGs at X = 0 mm. From the measured intensity of the Hα line, the density of the H0 atoms was estimated by CR model.11, 12 In this calculation, the Hα line was simply assumed to be only emitted by excitation of the H0 atoms. The profile of the H0 atoms is shown in Fig. 4(d). In the PG filter, the Te is locally high at −500 < Y < −400 mm (at Seg. 1) and constant at Y > −400 mm. The Jes and Jis are also locally high at the same region of −500 < Y < −400 mm giving the high Te, and nearly constant at −400 < Y < 0 mm and quite low at Y > 0 mm. The H0 atom is also locally high at Seg. 1 and constant at the other segments. These localizations of the source plasma in Seg. 1 are similar to that of the calculated population of the primary electrons in the longitudinal direction as shown in Fig. 3(b). In the tent-shaped filter, Te and Jes in the regions of −500 < Y < −400 mm (Seg. 1) and 400 < Y < 500 mm (Seg. 5) are higher than those in the center regions of −400 < Y < 400 mm (Segs. 2–4), where the profiles of the Te, Jes, Jis, and H0 atoms are nearly constant. The tendency of these profiles were similar to those of the electron population in Fig. 3(b), showing a uniform population of the primary electrons. These results clearly show that suppression of the localization of the primary electrons by modifying the magnetic filter results in the improvement of the localization of the source plasmas of the parent particles for the negative ions. B. Uniformity of negative ion beams
Figure 5 shows the longitudinal profiles of the extracted H− ion beam measured by the IR camera with the Cs seeded
conditions and the arc power of 120 kW before and after the modification of the magnetic filter. Before the modification (PG filter), the intensity of the H− ion beam is the highest at the Seg. 2 and the lowest at the Seg. 5 in all segments. Lower beam intensity at the Seg. 1 might be caused by destruction of the H− ions due to collision with the localized fast electrons (>1 eV). In the tent-shaped filter, the negative beam intensities are relatively low in the Seg. 1 and high in the Segs. 4 and 5. Compared with the profile in the PG filter, the H− ion beam profile is clearly uniform in the region of Y = ±490 mm. These beam profiles in the both filters are qualitatively corresponded to the plasma distributions and electron populations as shown in Figs. 3 and 4. This indicates that suppression of the localization of the source plasmas allows to improve the non-uniformity of the negative ion beams. The averaged H− beam intensity in the tent-shaped filter was calculated to be 70 A/m2 by dividing the integrated profile of the intensity (total current) with total area. This value is almost the same as
20 15 PG filter Uniformity [%]
NH0 [x109/m3]
Seg3
Tent shaped filter
200
0 -600
Seg2
70
0
(d)
Seg1
10
5
0 50
Tent shaped filter
JT-60SA
100 150 Negative ion current [A/m2]
FIG. 6. The deviation as a function of the negative ion current density (Jacc) in the tent-shaped filter with the Cs seeded conditions.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 132.177.228.65 On: Wed, 18 Feb 2015 12:58:23
02B314-4
Yoshida et al.
that in the PG filter. This shows that the non-uniformity of the H− ion beams is improved without degradation of the H− ion beam production. Here, the beam uniformity is defined as the deviation from the averaged intensity in the region of Y = ±490 mm, which is an extraction region designed in JT-60SA. The beam uniformity was reduced from 14% of the PG filter to ∼10% of the tent-shaped filter at the current density of 70 A/m2 of 120 kW. In JT-60SA, the deviation is required to be lower than 10% at the current density of 130–170 A/m2 . The uniformity is plotted as a function of the negative ion current density in the tent-shaped filter with the Cs seeded conditions (Fig. 6). The uniformity was kept to be almost 10% for the current density. This indicates that uniform negative ion beam of