The new TORPEX in-vessel toroidal conductor for the generation of a poloidal magnetic field F. Avino, A. Fasoli, and I. Furno Citation: Review of Scientific Instruments 85, 033506 (2014); doi: 10.1063/1.4868588 View online: http://dx.doi.org/10.1063/1.4868588 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/85/3?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Edge electron density profiles and fluctuations measured by two-dimensional beam emission spectroscopy in the KSTARa) Rev. Sci. Instrum. 85, 11E434 (2014); 10.1063/1.4894839 Experimental investigation of density behaviors in front of the lower hybrid launcher in experimental advanced superconducting tokamak Phys. Plasmas 20, 062507 (2013); 10.1063/1.4812462 On the toroidal current density flowing across a poloidal-magnetic-field null in an axisymmetric plasma Phys. Plasmas 20, 040702 (2013); 10.1063/1.4801002 Two-dimensional imaging of edge-localized modes in KSTAR plasmas unperturbed and perturbed by n=1 external magnetic fields Phys. Plasmas 19, 056114 (2012); 10.1063/1.3694842 Rotation dependence of a phase delay between plasma edge electron density and temperature fields due to a fast rotating, resonant magnetic perturbation field Phys. Plasmas 17, 060702 (2010); 10.1063/1.3436614
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
REVIEW OF SCIENTIFIC INSTRUMENTS 85, 033506 (2014)
The new TORPEX in-vessel toroidal conductor for the generation of a poloidal magnetic field F. Avino,a) A. Fasoli, and I. Furno Ecole Polytechnique Fédérale de Lausanne (EPFL), Centre de Recherches en Physique des Plasmas (CRPP), CH-1015 Lausanne, Switzerland
(Received 28 October 2013; accepted 3 March 2014; published online 25 March 2014) TORoidal Plasma EXperiment (TORPEX) is a Simple Magnetized Torus featuring open helical magnetic field lines obtained from the superposition of a small vertical component on the main toroidal field. This work introduces the experimental setup developed to include a poloidal magnetic field. The toroidal and poloidal fields generate a rotational transform, making the magnetic geometry of TORPEX closer to that of a tokamak. This upgrade opens the possibility to deal with closed and open flux surfaces, as well as with the transition region across the last closed flux surface. The main technical solutions are discussed together with the physical considerations at the basis of the system design. Selected examples of the magnetic configurations accessible with the set of magnetic field coils available on TORPEX are discussed, ranging from single-null X-points to magnetic snowflakes. The simplest magnetic configuration of quasi-circular concentric flux surfaces is tested experimentally. Measurements of the two-dimensional electron plasma density profiles and the particle confinement time are presented, together with the first steps towards the understanding of plasma production mechanisms. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4868588] I. INTRODUCTION
The study of plasma heat, momentum, and particle transport is one of the key topics in the field of magnetically confined plasmas. In fusion oriented devices, this study is hampered by a limited diagnostic access. The research of solutions with higher experimental accessibility and flexibility has led to an increasing attention towards basic plasma physics devices.1–7 These low temperature and density experiments have proven to be very useful test beds to improve our understanding in plasma turbulence and instability driving mechanisms.8, 9 A rigorous validation procedure between experimental data and numerical simulations can be performed.10, 11 One of these machines is the TORoidal Plasma EXperiment (TORPEX),12 where important features of a tokamak Scrape-Off Layer (SOL), namely, density gradients in the presence of curvature and gradient of the magnetic field, are reproduced in a Simple Magnetized Torus (SMT). The SMT consists of a main toroidal magnetic field with a superimposed small vertical component to increase the connection length and improve the confinement time, forming open helical magnetic field lines.13 A comprehensive experimental study of plasma instabilities, including interchange modes generating radially outward propagating blobs, has been conducted in the last years5, 14–17 in parallel with numerical studies.18 Similarities with statistical properties of tokamak SOL turbulence have been found.19 The SMT lacks a poloidal magnetic field component, necessary to obtain, together with the toroidal magnetic field, a rotational transform. A possibility for generating it is the implementation of a central current-carrying conductor, initially explored in the early 1970s on several devices for the achievement of fusion conditions, in either levitating mode20
or suspended through magnetically screened pairs of leads.21 A layout with a central conductor supported by vertical holders has recently been implemented on the MISTOR device.7 An advantage of this solution with respect to an inductive current drive is the decoupling the plasma from the magnetic field source. Based on this concept, we present in this work the invessel toroidal conductor (TC) developed on TORPEX to generate a poloidal magnetic field.22 This allows investigating plasma fluctuations in tokamak-like magnetic geometries, opening the possibility to address SOL physics issues, to study the core region of closed flux surfaces and the topological transition between closed and open magnetic field lines. This paper is organized as follows. In Sec. II, we discuss the experimental setup and the technical constraints, defining the main accessible experimental conditions opened by the invessel TC. Magnetic field simulations and measurements are presented in Sec. III. In Secs. IV and V, we focus on the first measurements of plasma parameters and profiles in the presence of a poloidal magnetic field. Conclusions are presented in Sec. VI.
II. SYSTEM REQUIREMENTS AND DESIGN
The main technical features of the system design are set by the requirement of obtaining a safety factor within q = 5 on a sufficiently large portion of the plasma volume to perform tokamak-relevant studies of plasma instabilities. With an aspect-ratio A = R/a = 5, where R = 1 m is TORPEX major radius and a = 20 cm the minor radius, the safety factor radial profile can be computed as q(r)
a) [email protected]
0034-6748/2014/85(3)/033506/8/$30.00
85, 033506-1
2π r 2 Bφ,0 r Bφ (r) . R Bθ (r) (R + r) μ0 IT C
(1)
© 2014 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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-2
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
FIG. 1. Simulated safety factor radial profile in the large aspect-ratio approximation for a toroidal magnetic field of 79 mT and a TC current of 1 kA.
Here, r is the minor radius radial coordinate, Bφ,0 the toroidal field at the center of the vacuum vessel, Bθ (r) the poloidal field, and ITC the electrical current flowing inside the TC. The corresponding magnetic shear radial profile sˆ (r) is almost constant: sˆ (r) = r
dq(r) 1 2. dr q(r)
(2)
In the poloidal magnetic field expression Bθ (r) for the estimate of q(r), the vertical field contribution can be neglected. The toroidal magnetic field is constrained in the range Bφ,0 = 70–100 mT, such that the Electron Cyclotron (EC) resonance layer, responsible for the plasma breakdown, is inside the vacuum chamber. The EC position is given by rEC R(11.42 Bφ,0 − 1) , where rEC and R are expressed in centimeters and Bφ,0 is in Tesla. For Bφ,0 79 mT, corresponding to an EC layer at rEC −10 cm, ITC = 1 kA provides q 5 at r 12 cm (Fig. 1). Therefore, we choose 1 kA as the required current for our experiments. Active cooling along the toroidal length of the TC is avoided to keep the system as simple as possible. Assuming radiative emission as the only intrinsic cooling mechanism and a TC current with 1 s of flat-top at 1 kA, and 0.5 s for the ramp-up and for the ramp-down, we estimate the heating of the TC for an experimental day of 100 shots for different diameters of the TC poloidal cross section. A size of 2 cm diameter cross section allows keeping the system temperature under 30 ◦ C. The TC is composed of four similar plain copper sections covering toroidal angular extensions of 90 ◦ C, 102 ◦ C, 90 ◦ C, and 78 ◦ C. These values are chosen according to the TORPEX port availability to limit the stress on the junctions between the four sections. These are supported inside the TORPEX vacuum chamber at the junction points by four lateral and three vertical 1 mm stainless steel wires, as shown in Figs. 2(a) and 2(b). Figure 2(b) shows a picture of the TC installed in TORPEX. The system is designed such that the TC can be moved along the full vertical length, recovering the SMT if required. The possibility to set the TC at intermediate vertical positions constitutes a new feature with respect to other similar systems, such as the one reported in
FIG. 2. Schematic 3D view (a), and picture (b), of the in-vessel TC installed on TORPEX.
Ref. 7, and it allows us to explore more advanced magnetic configurations, as discussed in Sec. III. The seven supporting leads are kept electrically floating to reduce perturbations to the plasma. The fourth vertical holder is the electrical feed-through itself, consisting of a coaxial rod to minimize the magnetic field perturbations. The final design of the feed-through is sketched in Fig. 3, with a detailed view of the lower part, where the feed-through is connected to the TC, and the vacuum-tight sliding seal. The effective cross section of 62 mm2 for the inner conductor and 140 mm2 for the outer one of the coaxial feedthrough (respectively, red and orange in Fig. 3) requires active cooling along the whole vertical length to avoid overheating. The cooling system of the Toroidal Field Coils (TFC) is used for this purpose, forcing the circulation of demineralized water in the empty volume of the coaxial wire, as illustrated in Fig. 3. A nylon-based material, namely polyamide 6 (PA6), is chosen to isolate the coaxial conductors in the top and bottom parts (Fig. 3, in yellow). We currently use a power supply (PS) capable of reaching 1.1 kA. A 2 kA PS is available and could be implemented in future experiments. The PS is floating and is remotely controlled using an electronic module with its ground decoupled from the ground of the TC circuit. This allows changing the potential of the TC with respect the TORPEX
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-3
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
B1
E1
C1
D1 A1
Vertical IDEAL
100 A
100 A
A2 PS + PS −
D2 B2
1.7
1.8
1.9
C2
2
2.1
2.2 2.3 B [mT]
E2
2.4
2.5
2.6
2.7
FIG. 4. TORPEX cross section with the VFC and the corresponding current of 100 A in the coils used to obtain a vertical field component, namely, B1, B2, C1, and C2.
ther ON-OFF or multi-step mode, can be triggered at different times with respect the beginning of the plasma discharge, as will be discussed in Sec. V. III. MAGNETIC FIELD SIMULATIONS AND MEASUREMENTS
FIG. 3. Drawings of the current feed-through for the in-vessel TC. Demineralized water is injected from the top and flows between the inner (in orange) and outer (in red) conductors down to the bottom of the feed-through. Here, the water flows into the inner conductor through holes, as outlined in the zoomed inlet on the bottom right and is pumped out from the top. In yellow, the insulator in polyamide 6 (PA6) between the two conductors, used in a similar way on the top of the feed-through. On the top right, a zoomed view of the vacuum-tight sliding seal is provided. Vacuum sealing is insured by the component in peek shown in light gray, insulating the outer conductor from the vacuum vessel. The blue parts correspond to the holes for the O-ring, to keep the vacuum.
vacuum chamber. However, in the first experiments reported herein, a copper ground braiding is used to connect the feedthrough to the vacuum vessel to minimize the potential difference between the TC and the vacuum vessel, reducing the perturbations of the TC to the plasma. The control module is integrated in the TORPEX control software and it is at present programmed to work in ON-OFF mode, providing a single voltage step in the range 0–40 V. The total resistance of the circuit is Rtot 32 m. The electronic control module is designed to generate waveforms with multiple steps in the voltage signal, providing flexibility in the choice of the TC current time-trace. The voltage signal, in ei-
A set of 10 external poloidally distributed toroidal coils, herein referred to as VFC (Vertical Field Coils), is available on TORPEX,23 identified in Fig. 4 as A1, A2, B1, B2, C1, C2, D1, D2, E1, and E2. These coils can be combined to produce magnetic fields of different geometries, including a quasi homogeneous vertical field, as shown in Fig. 4. The combination of the magnetic fields generated by the VFC and the TC gives access to a wide range of magnetic configurations. These can be calculated using the measured currents in the TFC, TC, and VFC. The TC is modeled with several discrete sources sharing the total current, computing for each source the analytical solution of the generated poloidal field using the electromagnetic Green’s functions. Examples of configurations are wall-limited plasmas, in Figs. 8(e) and 8(f), single or double-null X-points, either vertical (Fig. 5) or horizontal, as well as magnetic snowflakes.24, 25 To obtain these advances diverted configurations, the TC could be placed in the upper part of the poloidal cross section (Fig. 6). The simulations are used to choose the experimental current values of the magnetic field coils (TFC, VFC) and TC to obtain the required magnetic geometries. In particular, the computed poloidal component is validated by measurements of the magnetic field with a single axis Hall transducer, Sentron Model ZS150 3-5-2I.26 The Hall probe is used to
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-4
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
20
10
8
6
4
P
0
B [mT]
z[cm]
10
−10 2
−20 −20
FIG. 7. (a) Radial magnetic field measurements and simulations for a current of 1025 A. (b) Relative error between simulations and experimental data.
−10
0 r[cm]
10
20
FIG. 5. Simulation of a double-null X-point on TORPEX using the VFC together with the poloidal field generated by the TC.
measure the vertical component of the poloidal magnetic field at z = 0 cm on the low field side (LFS), along the radial direction with a spatial step of 5 mm in the range r = 1.5–20 cm. The TC is positioned at (r = 0, z = 0) cm ± 3 mm. A current of about 1025 A is chosen to maximize the signal. At each radial position, the average over a time window of about 200 ms was performed on the acquired magnetic field signals. The experimental values are compared with the simulated magnetic field for the measured value of ITC = 1025 A. The results are shown in Fig. 7, with a good agreement within the error bars between computed and measured poloidal magnetic field. The relative error is below 3% on most of the LFS. IV. FIRST MEASUREMENTS OF TIME-AVERAGED PROFILES IN CLOSED FIELD LINE CONFIGURATIONS
Measurements of quasi circular-shaped plasmas in the presence of a poloidal field were performed. Clear differences with respect to the phase of no-TC current are observed. An example of hydrogen plasma with a neutral pressure pn 1.8 × 10−4 mbar, ITC 900 A, Bφ,0 79 mT (rEC −10 cm),
20 12 10
10
6
P
0
B [mT]
z[cm]
8
4 10
20 20
10
0 r[cm]
10
20
FIG. 6. Simulation of a magnetic second order null-point (magnetic snowflake) on TORPEX.
and constant power of approximately 350 W is provided in Fig. 8. A vertical field component of approximately 3 G is added to obtain concentric flux surfaces over most of the plasma volume, as illustrated by the simulated field lines in white in Figs. 8(e) and 8(f). From a top view, the toroidal magnetic field is counterclockwise, while the TC current is clockwise. Time averaged density measurements are performed using the HEXagonal Turbulence Imaging Probe (HEXTIP),27 a 2D array of 85 Langmuir probes (LPs) that can be used in ion saturation current regime (Isat ), floating or sweeping voltage, with an acquisition frequency of 250 kHz. Originally designed with 86 LPs, the HEXTIP central tip (r = 0, z = 0) cm was recently removed to make the diagnostic compatible with the presence of the TC in the middle of the vacuum vessel. The TC current time evolution and two temporal windows before [t = (10–60) ms] and during [t = (600–850) ms] the current flat-top, are shown in Fig. 8(a). The corresponding 2D profiles of the time averaged density signals ne and of their standard deviation σne are presented, respectively, in Figs. 8(c) and 8(e) and Figs. 8(d) and 8(f), together with the radial cuts at z = 0 cm in Fig. 8(b). In this work, a constant electron temperature of Te = 5 eV on the whole poloidal cross section is assumed √ to compute ne and σne from Isat = ne ecs /2, where cs = γ Zi kB Te /mi is the ion sound speed. An adiabatic index γ = 1 and a charge state Zi = 1 are considered. During the current flat-top the plasma shape changes, peaking in the region of higher poloidal field close to the TC, forming regions of almost constant density along the flux surfaces. Moreover, the intensity of plasma fluctuations and turbulence, quantified by the standard deviation of measured ne signals, is lower in the presence of the poloidal field.7 The plasma reproducibility is verified in quasi-concentric flux surfaces, for a set of typical experimental parameters (ITC 760 A, pn 1.8 × 10−4 mbar, rEC −10 cm, Pmag 300 W) with hydrogen gas, repeating the same shot three times. For each shot, a time window of about 600 ms during the current flat-top is used to calculate the time average and the standard deviation on the density signals of HEXTIP probes. The variation of the time averaged density between the analyzed shots is within 10% for all the probes and below 5% for most of them. The density standard deviation percentage variation is within 20% for all the probes and below 10% for most of them.
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-5
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
FIG. 8. (a) Toroidal current signal for the #55066 with the red and blue time traces corresponding to the time windows where the HEXTIP experimental data are analyzed. (b) Continuous lines: radial cuts at z 0 cm of the 2D time averaged density profiles displayed in (c) and (e). Dashed and dotted lines: cuts at z 0 cm of the 2D profiles of the density signal standard deviation displayed, respectively, in (d) and (f). In white on (c)–(f) the simulated magnetic field lines on the poloidal cross section. The black crosses correspond to the positions of the single HEXTIP probes, while the black central circle represents the TC.
V. PLASMA PRODUCTION AND CONFINEMENT
In the new closed field line configuration, we investigated whether the steady state plasma profiles are different if the plasma is generated before or during the current flat-top, when the required closed field lines geometry is obtained. A very high similarity of 2D steady-state plasma profiles in those two cases is verified for several plasma parameters: ITC (400, 620) A, rEC ( − 15, −12.5, −10, −7.5) cm, Pmag 300 W, and pn 1.8 × 10−4 mbar. In Fig. 9, we show the results for ITC 620 A and rEC −10 cm. Figure 9(a) shows the current time trace of a shot where the magnetron is triggered during the TC current flat-top at about 700 ms, as can be seen in Fig. 9(c). The time averaged density signals in the time window corresponding to the red time traces in Figs. 9(a) and 9(c) between 1100 and 1250 ms, are provided in Fig. 9(e). Figures 9(b) and 9(d) provide an example where the current is driven at about 80 ms, when the magnetron low power phase is already started, while the corresponding densities time averaged between 700 and 850 ms are visible in Fig. 9(f). The relative difference of the time-averaged signals of 83 out of 85 probes between the two shots is within 10%. Averaging over a sequence of short time windows starting from the magnetron trigger, during the current flat-top, the plasma density time evolution from the breakdown to the steady state condition can be obtained. Data shown in Fig. 10 correspond to ITC 620 A and an EC layer moved towards the high field side (HFS), at rEC −12.5 cm. The density layers corresponding to the EC and Upper Hybrid (UH) resonance are reported, respectively, in magenta and gray in
Fig. 10. The density nUe H corresponding to the UH resonance is given by28 nUe H =
(2π )2 me 0 eB 2 f , − rf e2 2π me
(3)
FIG. 9. (a) and (b) TC current time traces for shots #58397 and #58839, corresponding to the magnetron triggered, respectively, during and before the current flat-top. (c) and (d) Injected magnetron power signals; in red the time windows when the ne time average shown in (e) and (f) has been performed.
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
Avino, Fasoli, and Furno 15
(a)
10 z [cm]
2 0
x 1015 10
t = (663.5−664.5)ms
(b)
6 4
1 −10
−20 −20 20
-10
0
10
20
t = (668.5−669.5)ms
(c)
z [cm]
-10
0
10
20
t = (694.5−695.5)ms
(d)
10
0 16
x 10 4
10
3
0
2
1.5 0 1 −10 −20 −20
0.5
-10
0
r [cm]
10
20
16
10
2
0 −20 16 −20 x 10 2.5 20 2
(e)
8
10 0
−10
17
10
−10
0 −20 −20
n e [m -3 ]
x 10 3 20
t = (662.5−663.5)ms
n e [m -3 ]
20
Rev. Sci. Instrum. 85, 033506 (2014)
n e [m -3 ]
033506-6
15
10
14
10
1 13
-10
0 r [cm]
10
20
10 −20
−10
0
r [cm]
t = (662.5−663.5)ms t = (663.5−664.5)ms t = (668.5−669.5)ms t = (694.5−695.5)ms 10 20
FIG. 10. (a)–(d) Time averaged 2D plasma density at different times during the plasma formation; (e) Radial cuts at z = 0 cm for the four chosen times with the central conductor in black at z = 0 cm. The EC resonance layer and the upper hybrid resonance layer correspond, respectively, to the magenta and gray curves.
where me the electron mass, 0 the permittivity of free space, e the elementary charge, and frf = 2.45 GHz is the magnetron frequency. The particle confinement time τ can be evaluated by switching off the power source and fitting the decaying probe signals with an exponential Aexp (−t/τ ), for a variety of conditions. We fix Pmag 600 W, rEC −15 cm, and ITC 620 A, and vary the hydrogen neutral pressure in the range (0.4–7) × 10−4 mbar. As it can be seen in Fig. 11 for the case at 1.9 × 10−4 mbar, a few general features characterize the signal time-decay. Two characteristic times can be inferred, based on two different fits performed on two separate timewindows. The first one just after the switching off of the magnetron (zero of the time-axis of Fig. 11) provides a value in the range τ1 = 0.3–0.4 ms. The second one after about 0.50 ms gives τ2 = 1.2–1.6 ms. As we are acquiring Isat , both density
and temperature are responsible for the signal decay and they could vary on different time-scales. In particular, as reported for the SMT,28 a fast cooling of the plasma could be responsible for the fast decay of τ 1 . Increasing the toroidal current to ITC 920 A, with rEC −10 cm, Pmag 300 W, and pn 1.8 × 10−4 mbar has not given any significant improvement on τ 1 and τ 2 . In conclusion, the measured particle confinement time for the explored parameters results to be of the same order of magnitude as observed for the optimized SMT.28 Considering the Bohm time τ B for a cylindrical column of radius a, defined as29 τB =
a2 8a 2 eB = 5.1 ms, 2DB k B Te
(4)
where Te 5 eV and B 80 mT were used, we find that τ B is factor of five larger than the experimental τ . In a weakly ionized plasma, other processes such as recombination could take place and play a significant role in the losses. Others losses channels are the TC surface, which amounts to 1/20 the vacuum chamber surface, and the supports, even if electrically floating. We note that for the sake of the experiments to be conducted on TORPEX, a confinement time comparable with the growth rate of the expected plasma instabilities is sufficient. VI. CONCLUSIONS AND OUTLOOK
FIG. 11. Isat time signals for the shot #57838, taking the switching off of the magnetron as zero of the time axis. (a) and (b) Signals of probes located, respectively, at (z = 0, r = 3.5) cm and (z = 0, r = 7) cm. In red and green the data corresponding to the two time windows where different exponential fits are performed.
A new in-vessel copper TC has recently been installed on TORPEX. The poloidal magnetic field generated by the current driven in the TC combines with the toroidal component, closing the magnetic field lines and introducing a rotational transform. The main technical features of the system are reported, together with the first 2D measurements of plasma parameters in the new configuration. The plasma
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-7
Avino, Fasoli, and Furno
shape changes when the poloidal component is included in the magnetic geometry, resulting in density profiles peaked at the center of the vacuum chamber, close to the TC where the poloidal field is higher. However, although the presence of closed magnetic field lines eliminates the parallel losses, the plasma density is the same order of magnitude with respect to the SMT since different loss channels are still present. Similar to the SMT case,28 the maximum of the density is located between the EC and UH layers. As already assessed for the SMT configuration,30 the injected microwaves at frf = 2.45 GHz in O-mode provide the plasma breakdown at the EC layer, accelerating the free electrons, which ionize part of the neutrals. The partial change to X-mode of the wave polarization after multiple reflections on the chamber wall, allows the activation of the UH, which then absorbs most of the microwave power. Dedicated experiments will be conducted in the future to obtain a deeper insight of plasma production mechanisms in the presence of a poloidal field in different plasma conditions. In the range of the explored experimental parameters in the new closed flux surfaces configuration, the measured particle confinement time results to be comparable with the one of the optimized SMT. Experiments dedicated to the investigation of electrostatic fluctuations will be performed with closed flux surfaces.
FIG. 12. Example of a single-null X-point on the LFS at z = 0 cm. (a) ne time averaged profiles. (b) ne standard deviation.
Rev. Sci. Instrum. 85, 033506 (2014)
Configurations of increasing complexity will be explored in future experimental campaigns, focusing on several fusion relevant topics, from plasma turbulence studies to fast ions interaction with turbulence in closed field lines.31 An example of the first experimental measurements performed on TORPEX in presence of a single-null X-point on the LFS at z = 0 cm is reported in Fig. 12. The HEXTIP ne time-averaged signals are shown in Fig. 12(a), while the standard deviation of density signals is illustrated in Fig. 12(b). A current of ITC 880 A and a vertical field obtained with a current in the VFC of IBv −54 A is used, with Pmag 250 W and rEC 1 cm.
ACKNOWLEDGMENTS
This work was supported in part by the Swiss National Science Foundation. The authors wish to thank A. Bovet and P. Ricci for the very helpful discussions. We also acknowledge the fundamental support of the CRPP technical team, in particular J. Theile and R. Chavan. 1 K.
L. Wong, M. Ono, and G. A. Wurden, Rev. Sci. Instrum. 53, 409 (1982). 2 D. Bora, Phys. Lett. A 139, 308 (1989). 3 K. Rypdal, E. Grønvoll, Å. Fredriksen, R. J. Armstrong, J. Trulsen, and H. Pécseli, Plasma Phys. Controlled Fusion 36, 1099 (1994). 4 N. Lemoine and D. Grésillon, Phys. Plasmas 12, 092301 (2005). 5 A. Fasoli, B. Labit, M. McGrath, S. H. Müller, G. Plyushchev, M. Podestà, and F. M. Poli, Phys. Plasmas 13, 055902 (2006). 6 K. W. Gentle and H. Huang, Plasma Sci. Technol. 10, 284 (2008). 7 Th. Pierre, Rev. Sci. Instrum. 84, 013504 (2013). 8 C. Riccardi, D. Xuantong, M. Salierno, L. Gamberale, and M. Fontanesi, Phys. Plasmas 4, 3749 (1997). 9 A. Fasoli, M. Fontanesi, A. Galassi, C. Longari, and E. Sindoni, Plasma Phys. Controlled Fusion 31, 313 (1989). 10 P. Ricci, C. Theiler, A. Fasoli, I. Furno, B. Labit, S. H. Müller, M. Podestà, and F. M. Poli, Phys. Plasmas 16, 055703 (2009). 11 P. Ricci, C. Theiler, A. Fasoli, I. Furno, K. Gustafson, D. Iraji, and J. Loizu, Phys. Plasmas 18, 032109 (2011). 12 A. Fasoli, A. Burckel, L. Federspiel, I. Furno, K. Gustafson, D. Iraji, B. Labit, J. Loizu, G. Plyushchev, P. Ricci, C. Theiler, A. Diallo, S. H. Müller, M. Podestà, and F. M. Poli, Plasma Phys. Controlled Fusion 52, 124020 (2010). 13 S. H. Müller, A. Fasoli, B. Labit, M. McGrath, M. Podestà, and F. M. Poli, Phys. Rev. Lett. 93, 165003 (2004). 14 I. Furno, B. Labit, M. Podestà, A. Fasoli, S. H. Müller, F. M. Poli, P. Ricci, C. Theiler, S. Brunner, A. Diallo, and J. Graves, Phys. Rev. Lett. 100, 055004 (2008). 15 M. Podestà, A. Fasoli, B. Labit, P. Ricci, F. M. Poli, A. Diallo, S. H. Müller, and C. Theiler, Phys. Rev. Lett. 101, 045001 (2008). 16 C. Theiler, A. Diallo, A. Fasoli, I. Furno, B. Labit, M. Podestà, F. M. Poli, and P. Ricci, Phys. Plasmas 15, 042303 (2008). 17 C. Theiler, I. Furno, P. Ricci, A. Fasoli, B. Labit, S. H. Müller, and G. Plyushchev, Phys. Rev. Lett. 103, 065001 (2009). 18 P. Ricci and B. N. Rogers, Phys. Rev. Lett. 104, 145001 (2010). 19 B. Labit, I. Furno, A. Fasoli, A. Diallo, S. H. Müller, G. Plyushchev, M. Podestà, and F. M. Poli, Phys. Rev. Lett. 98, 255002 (2007). 20 S. Skellett, Cryogenics 15, 563 (1975). 21 B. Lehnert, Plasma Phy. 10, 281 (1968). 22 A. Fasoli, F. Avino, A. Bovet, I. Furno, K. Gustafson, S. Jolliet, J. Loizu, D. Malinverni, P. Ricci, F. Riva, C. Theiler, M. Spolaore, and N. Vianello, Nucl. Fusion 53, 063013 (2013). 23 S. H. Müller, Ph.D. thesis, EPFL (CRPP), 2006. 24 F. Piras, S. Coda, I. Furno, J.-M. Moret, R. Pitts, O. Sauter, B. Tal, G. Turri, A. Bencze, B. P. Duval, F. Felici, A. Pochelon, and C. Zucca, Plasma Phys. Controlled Fusion 51, 055009 (2009). 25 D. D. Ryutov, R. H. Cohen, T. D. Rognlien, and V. Umansky, Phys. Plasmas 15, 092501 (2008).
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-8 26 C.
Avino, Fasoli, and Furno
Schott, R. S. Popovic, S. Alberti, and M. Q. Tran, Rev. Sci. Instrum. 70, 2703 (1999). 27 S. H. Müller, A. Fasoli, B. Labit, M. McGrath, O. Pisaturo, G. Plyushchev, M. Podestà, and M. F. Poli, Phys. Plasmas 12, 090906 (2005). 28 M. Podestà, Ph.D. thesis, EPFL (CRPP), 2007.
Rev. Sci. Instrum. 85, 033506 (2014) 29 F.
F. Chen, Introduction to Plasma Physics and Controlled Fusion, 2nd ed. (Springer, New York, 2006), pp. 190–191. 30 M. Podestà, A. Fasoli, B. Labit, M. McGrath, S. H. Müller, and F. M. Poli, Plasma Phys. Controlled Fusion 48, 1053 (2006). 31 A. Bovet, I. Furno, A. Fasoli, K. Gustafson, and P. Ricci, Nucl. Fusion 52, 094017 (2012).
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
REVIEW OF SCIENTIFIC INSTRUMENTS 85, 033506 (2014)
The new TORPEX in-vessel toroidal conductor for the generation of a poloidal magnetic field F. Avino,a) A. Fasoli, and I. Furno Ecole Polytechnique Fédérale de Lausanne (EPFL), Centre de Recherches en Physique des Plasmas (CRPP), CH-1015 Lausanne, Switzerland
(Received 28 October 2013; accepted 3 March 2014; published online 25 March 2014) TORoidal Plasma EXperiment (TORPEX) is a Simple Magnetized Torus featuring open helical magnetic field lines obtained from the superposition of a small vertical component on the main toroidal field. This work introduces the experimental setup developed to include a poloidal magnetic field. The toroidal and poloidal fields generate a rotational transform, making the magnetic geometry of TORPEX closer to that of a tokamak. This upgrade opens the possibility to deal with closed and open flux surfaces, as well as with the transition region across the last closed flux surface. The main technical solutions are discussed together with the physical considerations at the basis of the system design. Selected examples of the magnetic configurations accessible with the set of magnetic field coils available on TORPEX are discussed, ranging from single-null X-points to magnetic snowflakes. The simplest magnetic configuration of quasi-circular concentric flux surfaces is tested experimentally. Measurements of the two-dimensional electron plasma density profiles and the particle confinement time are presented, together with the first steps towards the understanding of plasma production mechanisms. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4868588] I. INTRODUCTION
The study of plasma heat, momentum, and particle transport is one of the key topics in the field of magnetically confined plasmas. In fusion oriented devices, this study is hampered by a limited diagnostic access. The research of solutions with higher experimental accessibility and flexibility has led to an increasing attention towards basic plasma physics devices.1–7 These low temperature and density experiments have proven to be very useful test beds to improve our understanding in plasma turbulence and instability driving mechanisms.8, 9 A rigorous validation procedure between experimental data and numerical simulations can be performed.10, 11 One of these machines is the TORoidal Plasma EXperiment (TORPEX),12 where important features of a tokamak Scrape-Off Layer (SOL), namely, density gradients in the presence of curvature and gradient of the magnetic field, are reproduced in a Simple Magnetized Torus (SMT). The SMT consists of a main toroidal magnetic field with a superimposed small vertical component to increase the connection length and improve the confinement time, forming open helical magnetic field lines.13 A comprehensive experimental study of plasma instabilities, including interchange modes generating radially outward propagating blobs, has been conducted in the last years5, 14–17 in parallel with numerical studies.18 Similarities with statistical properties of tokamak SOL turbulence have been found.19 The SMT lacks a poloidal magnetic field component, necessary to obtain, together with the toroidal magnetic field, a rotational transform. A possibility for generating it is the implementation of a central current-carrying conductor, initially explored in the early 1970s on several devices for the achievement of fusion conditions, in either levitating mode20
or suspended through magnetically screened pairs of leads.21 A layout with a central conductor supported by vertical holders has recently been implemented on the MISTOR device.7 An advantage of this solution with respect to an inductive current drive is the decoupling the plasma from the magnetic field source. Based on this concept, we present in this work the invessel toroidal conductor (TC) developed on TORPEX to generate a poloidal magnetic field.22 This allows investigating plasma fluctuations in tokamak-like magnetic geometries, opening the possibility to address SOL physics issues, to study the core region of closed flux surfaces and the topological transition between closed and open magnetic field lines. This paper is organized as follows. In Sec. II, we discuss the experimental setup and the technical constraints, defining the main accessible experimental conditions opened by the invessel TC. Magnetic field simulations and measurements are presented in Sec. III. In Secs. IV and V, we focus on the first measurements of plasma parameters and profiles in the presence of a poloidal magnetic field. Conclusions are presented in Sec. VI.
II. SYSTEM REQUIREMENTS AND DESIGN
The main technical features of the system design are set by the requirement of obtaining a safety factor within q = 5 on a sufficiently large portion of the plasma volume to perform tokamak-relevant studies of plasma instabilities. With an aspect-ratio A = R/a = 5, where R = 1 m is TORPEX major radius and a = 20 cm the minor radius, the safety factor radial profile can be computed as q(r)
a) [email protected]
0034-6748/2014/85(3)/033506/8/$30.00
85, 033506-1
2π r 2 Bφ,0 r Bφ (r) . R Bθ (r) (R + r) μ0 IT C
(1)
© 2014 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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-2
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
FIG. 1. Simulated safety factor radial profile in the large aspect-ratio approximation for a toroidal magnetic field of 79 mT and a TC current of 1 kA.
Here, r is the minor radius radial coordinate, Bφ,0 the toroidal field at the center of the vacuum vessel, Bθ (r) the poloidal field, and ITC the electrical current flowing inside the TC. The corresponding magnetic shear radial profile sˆ (r) is almost constant: sˆ (r) = r
dq(r) 1 2. dr q(r)
(2)
In the poloidal magnetic field expression Bθ (r) for the estimate of q(r), the vertical field contribution can be neglected. The toroidal magnetic field is constrained in the range Bφ,0 = 70–100 mT, such that the Electron Cyclotron (EC) resonance layer, responsible for the plasma breakdown, is inside the vacuum chamber. The EC position is given by rEC R(11.42 Bφ,0 − 1) , where rEC and R are expressed in centimeters and Bφ,0 is in Tesla. For Bφ,0 79 mT, corresponding to an EC layer at rEC −10 cm, ITC = 1 kA provides q 5 at r 12 cm (Fig. 1). Therefore, we choose 1 kA as the required current for our experiments. Active cooling along the toroidal length of the TC is avoided to keep the system as simple as possible. Assuming radiative emission as the only intrinsic cooling mechanism and a TC current with 1 s of flat-top at 1 kA, and 0.5 s for the ramp-up and for the ramp-down, we estimate the heating of the TC for an experimental day of 100 shots for different diameters of the TC poloidal cross section. A size of 2 cm diameter cross section allows keeping the system temperature under 30 ◦ C. The TC is composed of four similar plain copper sections covering toroidal angular extensions of 90 ◦ C, 102 ◦ C, 90 ◦ C, and 78 ◦ C. These values are chosen according to the TORPEX port availability to limit the stress on the junctions between the four sections. These are supported inside the TORPEX vacuum chamber at the junction points by four lateral and three vertical 1 mm stainless steel wires, as shown in Figs. 2(a) and 2(b). Figure 2(b) shows a picture of the TC installed in TORPEX. The system is designed such that the TC can be moved along the full vertical length, recovering the SMT if required. The possibility to set the TC at intermediate vertical positions constitutes a new feature with respect to other similar systems, such as the one reported in
FIG. 2. Schematic 3D view (a), and picture (b), of the in-vessel TC installed on TORPEX.
Ref. 7, and it allows us to explore more advanced magnetic configurations, as discussed in Sec. III. The seven supporting leads are kept electrically floating to reduce perturbations to the plasma. The fourth vertical holder is the electrical feed-through itself, consisting of a coaxial rod to minimize the magnetic field perturbations. The final design of the feed-through is sketched in Fig. 3, with a detailed view of the lower part, where the feed-through is connected to the TC, and the vacuum-tight sliding seal. The effective cross section of 62 mm2 for the inner conductor and 140 mm2 for the outer one of the coaxial feedthrough (respectively, red and orange in Fig. 3) requires active cooling along the whole vertical length to avoid overheating. The cooling system of the Toroidal Field Coils (TFC) is used for this purpose, forcing the circulation of demineralized water in the empty volume of the coaxial wire, as illustrated in Fig. 3. A nylon-based material, namely polyamide 6 (PA6), is chosen to isolate the coaxial conductors in the top and bottom parts (Fig. 3, in yellow). We currently use a power supply (PS) capable of reaching 1.1 kA. A 2 kA PS is available and could be implemented in future experiments. The PS is floating and is remotely controlled using an electronic module with its ground decoupled from the ground of the TC circuit. This allows changing the potential of the TC with respect the TORPEX
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-3
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
B1
E1
C1
D1 A1
Vertical IDEAL
100 A
100 A
A2 PS + PS −
D2 B2
1.7
1.8
1.9
C2
2
2.1
2.2 2.3 B [mT]
E2
2.4
2.5
2.6
2.7
FIG. 4. TORPEX cross section with the VFC and the corresponding current of 100 A in the coils used to obtain a vertical field component, namely, B1, B2, C1, and C2.
ther ON-OFF or multi-step mode, can be triggered at different times with respect the beginning of the plasma discharge, as will be discussed in Sec. V. III. MAGNETIC FIELD SIMULATIONS AND MEASUREMENTS
FIG. 3. Drawings of the current feed-through for the in-vessel TC. Demineralized water is injected from the top and flows between the inner (in orange) and outer (in red) conductors down to the bottom of the feed-through. Here, the water flows into the inner conductor through holes, as outlined in the zoomed inlet on the bottom right and is pumped out from the top. In yellow, the insulator in polyamide 6 (PA6) between the two conductors, used in a similar way on the top of the feed-through. On the top right, a zoomed view of the vacuum-tight sliding seal is provided. Vacuum sealing is insured by the component in peek shown in light gray, insulating the outer conductor from the vacuum vessel. The blue parts correspond to the holes for the O-ring, to keep the vacuum.
vacuum chamber. However, in the first experiments reported herein, a copper ground braiding is used to connect the feedthrough to the vacuum vessel to minimize the potential difference between the TC and the vacuum vessel, reducing the perturbations of the TC to the plasma. The control module is integrated in the TORPEX control software and it is at present programmed to work in ON-OFF mode, providing a single voltage step in the range 0–40 V. The total resistance of the circuit is Rtot 32 m. The electronic control module is designed to generate waveforms with multiple steps in the voltage signal, providing flexibility in the choice of the TC current time-trace. The voltage signal, in ei-
A set of 10 external poloidally distributed toroidal coils, herein referred to as VFC (Vertical Field Coils), is available on TORPEX,23 identified in Fig. 4 as A1, A2, B1, B2, C1, C2, D1, D2, E1, and E2. These coils can be combined to produce magnetic fields of different geometries, including a quasi homogeneous vertical field, as shown in Fig. 4. The combination of the magnetic fields generated by the VFC and the TC gives access to a wide range of magnetic configurations. These can be calculated using the measured currents in the TFC, TC, and VFC. The TC is modeled with several discrete sources sharing the total current, computing for each source the analytical solution of the generated poloidal field using the electromagnetic Green’s functions. Examples of configurations are wall-limited plasmas, in Figs. 8(e) and 8(f), single or double-null X-points, either vertical (Fig. 5) or horizontal, as well as magnetic snowflakes.24, 25 To obtain these advances diverted configurations, the TC could be placed in the upper part of the poloidal cross section (Fig. 6). The simulations are used to choose the experimental current values of the magnetic field coils (TFC, VFC) and TC to obtain the required magnetic geometries. In particular, the computed poloidal component is validated by measurements of the magnetic field with a single axis Hall transducer, Sentron Model ZS150 3-5-2I.26 The Hall probe is used to
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-4
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
20
10
8
6
4
P
0
B [mT]
z[cm]
10
−10 2
−20 −20
FIG. 7. (a) Radial magnetic field measurements and simulations for a current of 1025 A. (b) Relative error between simulations and experimental data.
−10
0 r[cm]
10
20
FIG. 5. Simulation of a double-null X-point on TORPEX using the VFC together with the poloidal field generated by the TC.
measure the vertical component of the poloidal magnetic field at z = 0 cm on the low field side (LFS), along the radial direction with a spatial step of 5 mm in the range r = 1.5–20 cm. The TC is positioned at (r = 0, z = 0) cm ± 3 mm. A current of about 1025 A is chosen to maximize the signal. At each radial position, the average over a time window of about 200 ms was performed on the acquired magnetic field signals. The experimental values are compared with the simulated magnetic field for the measured value of ITC = 1025 A. The results are shown in Fig. 7, with a good agreement within the error bars between computed and measured poloidal magnetic field. The relative error is below 3% on most of the LFS. IV. FIRST MEASUREMENTS OF TIME-AVERAGED PROFILES IN CLOSED FIELD LINE CONFIGURATIONS
Measurements of quasi circular-shaped plasmas in the presence of a poloidal field were performed. Clear differences with respect to the phase of no-TC current are observed. An example of hydrogen plasma with a neutral pressure pn 1.8 × 10−4 mbar, ITC 900 A, Bφ,0 79 mT (rEC −10 cm),
20 12 10
10
6
P
0
B [mT]
z[cm]
8
4 10
20 20
10
0 r[cm]
10
20
FIG. 6. Simulation of a magnetic second order null-point (magnetic snowflake) on TORPEX.
and constant power of approximately 350 W is provided in Fig. 8. A vertical field component of approximately 3 G is added to obtain concentric flux surfaces over most of the plasma volume, as illustrated by the simulated field lines in white in Figs. 8(e) and 8(f). From a top view, the toroidal magnetic field is counterclockwise, while the TC current is clockwise. Time averaged density measurements are performed using the HEXagonal Turbulence Imaging Probe (HEXTIP),27 a 2D array of 85 Langmuir probes (LPs) that can be used in ion saturation current regime (Isat ), floating or sweeping voltage, with an acquisition frequency of 250 kHz. Originally designed with 86 LPs, the HEXTIP central tip (r = 0, z = 0) cm was recently removed to make the diagnostic compatible with the presence of the TC in the middle of the vacuum vessel. The TC current time evolution and two temporal windows before [t = (10–60) ms] and during [t = (600–850) ms] the current flat-top, are shown in Fig. 8(a). The corresponding 2D profiles of the time averaged density signals ne and of their standard deviation σne are presented, respectively, in Figs. 8(c) and 8(e) and Figs. 8(d) and 8(f), together with the radial cuts at z = 0 cm in Fig. 8(b). In this work, a constant electron temperature of Te = 5 eV on the whole poloidal cross section is assumed √ to compute ne and σne from Isat = ne ecs /2, where cs = γ Zi kB Te /mi is the ion sound speed. An adiabatic index γ = 1 and a charge state Zi = 1 are considered. During the current flat-top the plasma shape changes, peaking in the region of higher poloidal field close to the TC, forming regions of almost constant density along the flux surfaces. Moreover, the intensity of plasma fluctuations and turbulence, quantified by the standard deviation of measured ne signals, is lower in the presence of the poloidal field.7 The plasma reproducibility is verified in quasi-concentric flux surfaces, for a set of typical experimental parameters (ITC 760 A, pn 1.8 × 10−4 mbar, rEC −10 cm, Pmag 300 W) with hydrogen gas, repeating the same shot three times. For each shot, a time window of about 600 ms during the current flat-top is used to calculate the time average and the standard deviation on the density signals of HEXTIP probes. The variation of the time averaged density between the analyzed shots is within 10% for all the probes and below 5% for most of them. The density standard deviation percentage variation is within 20% for all the probes and below 10% for most of them.
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-5
Avino, Fasoli, and Furno
Rev. Sci. Instrum. 85, 033506 (2014)
FIG. 8. (a) Toroidal current signal for the #55066 with the red and blue time traces corresponding to the time windows where the HEXTIP experimental data are analyzed. (b) Continuous lines: radial cuts at z 0 cm of the 2D time averaged density profiles displayed in (c) and (e). Dashed and dotted lines: cuts at z 0 cm of the 2D profiles of the density signal standard deviation displayed, respectively, in (d) and (f). In white on (c)–(f) the simulated magnetic field lines on the poloidal cross section. The black crosses correspond to the positions of the single HEXTIP probes, while the black central circle represents the TC.
V. PLASMA PRODUCTION AND CONFINEMENT
In the new closed field line configuration, we investigated whether the steady state plasma profiles are different if the plasma is generated before or during the current flat-top, when the required closed field lines geometry is obtained. A very high similarity of 2D steady-state plasma profiles in those two cases is verified for several plasma parameters: ITC (400, 620) A, rEC ( − 15, −12.5, −10, −7.5) cm, Pmag 300 W, and pn 1.8 × 10−4 mbar. In Fig. 9, we show the results for ITC 620 A and rEC −10 cm. Figure 9(a) shows the current time trace of a shot where the magnetron is triggered during the TC current flat-top at about 700 ms, as can be seen in Fig. 9(c). The time averaged density signals in the time window corresponding to the red time traces in Figs. 9(a) and 9(c) between 1100 and 1250 ms, are provided in Fig. 9(e). Figures 9(b) and 9(d) provide an example where the current is driven at about 80 ms, when the magnetron low power phase is already started, while the corresponding densities time averaged between 700 and 850 ms are visible in Fig. 9(f). The relative difference of the time-averaged signals of 83 out of 85 probes between the two shots is within 10%. Averaging over a sequence of short time windows starting from the magnetron trigger, during the current flat-top, the plasma density time evolution from the breakdown to the steady state condition can be obtained. Data shown in Fig. 10 correspond to ITC 620 A and an EC layer moved towards the high field side (HFS), at rEC −12.5 cm. The density layers corresponding to the EC and Upper Hybrid (UH) resonance are reported, respectively, in magenta and gray in
Fig. 10. The density nUe H corresponding to the UH resonance is given by28 nUe H =
(2π )2 me 0 eB 2 f , − rf e2 2π me
(3)
FIG. 9. (a) and (b) TC current time traces for shots #58397 and #58839, corresponding to the magnetron triggered, respectively, during and before the current flat-top. (c) and (d) Injected magnetron power signals; in red the time windows when the ne time average shown in (e) and (f) has been performed.
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
Avino, Fasoli, and Furno 15
(a)
10 z [cm]
2 0
x 1015 10
t = (663.5−664.5)ms
(b)
6 4
1 −10
−20 −20 20
-10
0
10
20
t = (668.5−669.5)ms
(c)
z [cm]
-10
0
10
20
t = (694.5−695.5)ms
(d)
10
0 16
x 10 4
10
3
0
2
1.5 0 1 −10 −20 −20
0.5
-10
0
r [cm]
10
20
16
10
2
0 −20 16 −20 x 10 2.5 20 2
(e)
8
10 0
−10
17
10
−10
0 −20 −20
n e [m -3 ]
x 10 3 20
t = (662.5−663.5)ms
n e [m -3 ]
20
Rev. Sci. Instrum. 85, 033506 (2014)
n e [m -3 ]
033506-6
15
10
14
10
1 13
-10
0 r [cm]
10
20
10 −20
−10
0
r [cm]
t = (662.5−663.5)ms t = (663.5−664.5)ms t = (668.5−669.5)ms t = (694.5−695.5)ms 10 20
FIG. 10. (a)–(d) Time averaged 2D plasma density at different times during the plasma formation; (e) Radial cuts at z = 0 cm for the four chosen times with the central conductor in black at z = 0 cm. The EC resonance layer and the upper hybrid resonance layer correspond, respectively, to the magenta and gray curves.
where me the electron mass, 0 the permittivity of free space, e the elementary charge, and frf = 2.45 GHz is the magnetron frequency. The particle confinement time τ can be evaluated by switching off the power source and fitting the decaying probe signals with an exponential Aexp (−t/τ ), for a variety of conditions. We fix Pmag 600 W, rEC −15 cm, and ITC 620 A, and vary the hydrogen neutral pressure in the range (0.4–7) × 10−4 mbar. As it can be seen in Fig. 11 for the case at 1.9 × 10−4 mbar, a few general features characterize the signal time-decay. Two characteristic times can be inferred, based on two different fits performed on two separate timewindows. The first one just after the switching off of the magnetron (zero of the time-axis of Fig. 11) provides a value in the range τ1 = 0.3–0.4 ms. The second one after about 0.50 ms gives τ2 = 1.2–1.6 ms. As we are acquiring Isat , both density
and temperature are responsible for the signal decay and they could vary on different time-scales. In particular, as reported for the SMT,28 a fast cooling of the plasma could be responsible for the fast decay of τ 1 . Increasing the toroidal current to ITC 920 A, with rEC −10 cm, Pmag 300 W, and pn 1.8 × 10−4 mbar has not given any significant improvement on τ 1 and τ 2 . In conclusion, the measured particle confinement time for the explored parameters results to be of the same order of magnitude as observed for the optimized SMT.28 Considering the Bohm time τ B for a cylindrical column of radius a, defined as29 τB =
a2 8a 2 eB = 5.1 ms, 2DB k B Te
(4)
where Te 5 eV and B 80 mT were used, we find that τ B is factor of five larger than the experimental τ . In a weakly ionized plasma, other processes such as recombination could take place and play a significant role in the losses. Others losses channels are the TC surface, which amounts to 1/20 the vacuum chamber surface, and the supports, even if electrically floating. We note that for the sake of the experiments to be conducted on TORPEX, a confinement time comparable with the growth rate of the expected plasma instabilities is sufficient. VI. CONCLUSIONS AND OUTLOOK
FIG. 11. Isat time signals for the shot #57838, taking the switching off of the magnetron as zero of the time axis. (a) and (b) Signals of probes located, respectively, at (z = 0, r = 3.5) cm and (z = 0, r = 7) cm. In red and green the data corresponding to the two time windows where different exponential fits are performed.
A new in-vessel copper TC has recently been installed on TORPEX. The poloidal magnetic field generated by the current driven in the TC combines with the toroidal component, closing the magnetic field lines and introducing a rotational transform. The main technical features of the system are reported, together with the first 2D measurements of plasma parameters in the new configuration. The plasma
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-7
Avino, Fasoli, and Furno
shape changes when the poloidal component is included in the magnetic geometry, resulting in density profiles peaked at the center of the vacuum chamber, close to the TC where the poloidal field is higher. However, although the presence of closed magnetic field lines eliminates the parallel losses, the plasma density is the same order of magnitude with respect to the SMT since different loss channels are still present. Similar to the SMT case,28 the maximum of the density is located between the EC and UH layers. As already assessed for the SMT configuration,30 the injected microwaves at frf = 2.45 GHz in O-mode provide the plasma breakdown at the EC layer, accelerating the free electrons, which ionize part of the neutrals. The partial change to X-mode of the wave polarization after multiple reflections on the chamber wall, allows the activation of the UH, which then absorbs most of the microwave power. Dedicated experiments will be conducted in the future to obtain a deeper insight of plasma production mechanisms in the presence of a poloidal field in different plasma conditions. In the range of the explored experimental parameters in the new closed flux surfaces configuration, the measured particle confinement time results to be comparable with the one of the optimized SMT. Experiments dedicated to the investigation of electrostatic fluctuations will be performed with closed flux surfaces.
FIG. 12. Example of a single-null X-point on the LFS at z = 0 cm. (a) ne time averaged profiles. (b) ne standard deviation.
Rev. Sci. Instrum. 85, 033506 (2014)
Configurations of increasing complexity will be explored in future experimental campaigns, focusing on several fusion relevant topics, from plasma turbulence studies to fast ions interaction with turbulence in closed field lines.31 An example of the first experimental measurements performed on TORPEX in presence of a single-null X-point on the LFS at z = 0 cm is reported in Fig. 12. The HEXTIP ne time-averaged signals are shown in Fig. 12(a), while the standard deviation of density signals is illustrated in Fig. 12(b). A current of ITC 880 A and a vertical field obtained with a current in the VFC of IBv −54 A is used, with Pmag 250 W and rEC 1 cm.
ACKNOWLEDGMENTS
This work was supported in part by the Swiss National Science Foundation. The authors wish to thank A. Bovet and P. Ricci for the very helpful discussions. We also acknowledge the fundamental support of the CRPP technical team, in particular J. Theile and R. Chavan. 1 K.
L. Wong, M. Ono, and G. A. Wurden, Rev. Sci. Instrum. 53, 409 (1982). 2 D. Bora, Phys. Lett. A 139, 308 (1989). 3 K. Rypdal, E. Grønvoll, Å. Fredriksen, R. J. Armstrong, J. Trulsen, and H. Pécseli, Plasma Phys. Controlled Fusion 36, 1099 (1994). 4 N. Lemoine and D. Grésillon, Phys. Plasmas 12, 092301 (2005). 5 A. Fasoli, B. Labit, M. McGrath, S. H. Müller, G. Plyushchev, M. Podestà, and F. M. Poli, Phys. Plasmas 13, 055902 (2006). 6 K. W. Gentle and H. Huang, Plasma Sci. Technol. 10, 284 (2008). 7 Th. Pierre, Rev. Sci. Instrum. 84, 013504 (2013). 8 C. Riccardi, D. Xuantong, M. Salierno, L. Gamberale, and M. Fontanesi, Phys. Plasmas 4, 3749 (1997). 9 A. Fasoli, M. Fontanesi, A. Galassi, C. Longari, and E. Sindoni, Plasma Phys. Controlled Fusion 31, 313 (1989). 10 P. Ricci, C. Theiler, A. Fasoli, I. Furno, B. Labit, S. H. Müller, M. Podestà, and F. M. Poli, Phys. Plasmas 16, 055703 (2009). 11 P. Ricci, C. Theiler, A. Fasoli, I. Furno, K. Gustafson, D. Iraji, and J. Loizu, Phys. Plasmas 18, 032109 (2011). 12 A. Fasoli, A. Burckel, L. Federspiel, I. Furno, K. Gustafson, D. Iraji, B. Labit, J. Loizu, G. Plyushchev, P. Ricci, C. Theiler, A. Diallo, S. H. Müller, M. Podestà, and F. M. Poli, Plasma Phys. Controlled Fusion 52, 124020 (2010). 13 S. H. Müller, A. Fasoli, B. Labit, M. McGrath, M. Podestà, and F. M. Poli, Phys. Rev. Lett. 93, 165003 (2004). 14 I. Furno, B. Labit, M. Podestà, A. Fasoli, S. H. Müller, F. M. Poli, P. Ricci, C. Theiler, S. Brunner, A. Diallo, and J. Graves, Phys. Rev. Lett. 100, 055004 (2008). 15 M. Podestà, A. Fasoli, B. Labit, P. Ricci, F. M. Poli, A. Diallo, S. H. Müller, and C. Theiler, Phys. Rev. Lett. 101, 045001 (2008). 16 C. Theiler, A. Diallo, A. Fasoli, I. Furno, B. Labit, M. Podestà, F. M. Poli, and P. Ricci, Phys. Plasmas 15, 042303 (2008). 17 C. Theiler, I. Furno, P. Ricci, A. Fasoli, B. Labit, S. H. Müller, and G. Plyushchev, Phys. Rev. Lett. 103, 065001 (2009). 18 P. Ricci and B. N. Rogers, Phys. Rev. Lett. 104, 145001 (2010). 19 B. Labit, I. Furno, A. Fasoli, A. Diallo, S. H. Müller, G. Plyushchev, M. Podestà, and F. M. Poli, Phys. Rev. Lett. 98, 255002 (2007). 20 S. Skellett, Cryogenics 15, 563 (1975). 21 B. Lehnert, Plasma Phy. 10, 281 (1968). 22 A. Fasoli, F. Avino, A. Bovet, I. Furno, K. Gustafson, S. Jolliet, J. Loizu, D. Malinverni, P. Ricci, F. Riva, C. Theiler, M. Spolaore, and N. Vianello, Nucl. Fusion 53, 063013 (2013). 23 S. H. Müller, Ph.D. thesis, EPFL (CRPP), 2006. 24 F. Piras, S. Coda, I. Furno, J.-M. Moret, R. Pitts, O. Sauter, B. Tal, G. Turri, A. Bencze, B. P. Duval, F. Felici, A. Pochelon, and C. Zucca, Plasma Phys. Controlled Fusion 51, 055009 (2009). 25 D. D. Ryutov, R. H. Cohen, T. D. Rognlien, and V. Umansky, Phys. Plasmas 15, 092501 (2008).
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
033506-8 26 C.
Avino, Fasoli, and Furno
Schott, R. S. Popovic, S. Alberti, and M. Q. Tran, Rev. Sci. Instrum. 70, 2703 (1999). 27 S. H. Müller, A. Fasoli, B. Labit, M. McGrath, O. Pisaturo, G. Plyushchev, M. Podestà, and M. F. Poli, Phys. Plasmas 12, 090906 (2005). 28 M. Podestà, Ph.D. thesis, EPFL (CRPP), 2007.
Rev. Sci. Instrum. 85, 033506 (2014) 29 F.
F. Chen, Introduction to Plasma Physics and Controlled Fusion, 2nd ed. (Springer, New York, 2006), pp. 190–191. 30 M. Podestà, A. Fasoli, B. Labit, M. McGrath, S. H. Müller, and F. M. Poli, Plasma Phys. Controlled Fusion 48, 1053 (2006). 31 A. Bovet, I. Furno, A. Fasoli, K. Gustafson, and P. Ricci, Nucl. Fusion 52, 094017 (2012).
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: 216.165.95.69 On: Tue, 09 Dec 2014 21:42:17
