EX/P4-22

First Formation in Glass Spherical (GLAST)

S Hussain1, G M Vorobyov2 and GLAST Team3

1,3National Tokamak Fusion Program (NTFP), P.O. Box 3329, Islamabad, Pakistan;

2Saint Petersburg University, Saint Petersburg Russia [email protected]

Abstract: The first plasma formation in GLAST is presented. GLAST is a small having vacuum vessel of dielectric material with aspect ratio of 1.7. A plasma current of 2kA was produced for about 0.5msec using ECR assisted plasma startup. A small vertical field was then added to enhance the peak of plasma current up to 5kA. The variation of plasma current with the applied vertical field was also studied and the optimum values of the vertical were experimentally determined to be between 40-50 gauss. A high speed camera was used to study the plasma behavior during the whole discharge scenario. The diagnostic systems such as Rogowski coil, , fast photo-diode, and spectrometers were also used to record the signatures of plasma current and to estimate some plasma parameters. The edge electron temperature was estimate to be 13eV from the measurement of the variation in the loop voltage during the plasma current formation. 1. Introduction In spherical there is limited space available for the central solenoid due to their low aspect ratio. This reduces the availability of induced toroidal electric field required for the production of initial plasma. In addition, the presence of eddy currents in the vacuum vessel further suppress the generation of induced electric field. Therefore, in spherical tokamaks the effect of eddy currents has great concern while evaluating the induced electric field for breakdown. The induced electric field requirement for breakdown can be reduced in different ways either by preionization with electron cyclotron (EC) waves or by applying additional volt-second with vertical field coils or by suppressing the eddy currents or all of these. GLAST is a small spherical tokamak indigenously developed in Pakistan having vacuum vessel of dielectric material (Pyrex glass) with aspect ratio of 1.7 (see FIG. 1). The main design parameters of GLAST are R= 15cm, a= 9 cm, κ=2.5, Ip=50kA, BT= (0.1- 0.4) Tesla, p=10ms, and Te = (300-500) eV. This device is designed to study the breakdown avalanche, plasma startup and current formation using inductive and non- inductive processes in the presence of dielectric walls of the vessel having no issues of the eddy currents. We expect to have extremely useful and important experimental finding from our glass vessel tokamak. As a starting point, we studied the preionization process, the transition from pre-ionization to breakdown avalanche and then to the plasma current column formation. We first obtained the preionization using ECR and then build up plasma current by central solenoid and also with the application of small vertical FIG. 1. Glass Spherical magnetic field. Tokamak (GLAST)

1 EX/P4-22

2. Experimental Setup A schematic of the experimental arrangement is shown in Fig. 2. Three set of coil systems have been used on GLAST. The ohmic heating system that consists of double layer central solenoid (0.4 mH, 0.28 ohms) of length 53 having 190 turns per layer and a pair of coils (three turns each) connected in series with it. The central solenoid was operated at a current of 2.5kA to generate the induced toroidal loop voltage of about 10 volts.

FIG. 2. Schematic experimental arrangement The toroidal field coil system consists of 16 coils connected in series in such a way that the inner legs are twisted at certain angle. This arrangement generates additional vertical field effect that helps to stabilize the plasma. A vertical field coil system consisting of three pairs of coils having 30-turns each with inductance of 7mH was used to enhance the plasma current by vertical field drift effect. Three capacitor banks were used to energize each of these coil systems through sequential firing using thyristor based switching circuitry. The relative delays between different coil systems were applied with the help of a small 12-channel trigger control unit. 3. Vertical Flux Null Generation The generation of flux null is extremely important for the plasma startup in small tokamaks. The presence of net vertical magnetic field (error field) due to the central solenoid and the other coil systems greatly affect the requirements for the induced electric field during the startup phase. There are different approaches for the generation of vertical field null inside the vacuum vessel. One of the methods is the use of extra coil system and then current in these coils is controlled in a complicated way to generate null point at a particular point inside the vessel. We have adopted a simple technique to minimize the net vertical magnetic flux inside the whole vacuum vessel. Two pairs of vertical field coils were used in series with the central solenoid to generate null magnetic flux inside the vessel at the center of the resonance layer (R=6cm). A differential loop (one part around the central solenoid while other around the vessel) was used to observe net change in flux inside the vessel. The minimum net flux at a particular instant was achieved by varying the number of turns and also the position of the compensation coils connected in series with the central solenoid.

4. Tuning of Vertical Field Coil System The tuning of vertical field coil system i.e. to minimize the induced effect of central solenoid on the vertical coils system is also necessary for the tokamak startup. This was done by firing the central solenoid at small voltage (~100 volts) and measuring the induced voltage across the vertical system. The induced voltages were reduced to the minimum possible value by adjusting the number of turns and the direction of current in each of the coils.

2 EX/P4-22

5. Pre-ionization The electron cyclotron resonance absorption plays a vital role in reducing the requirements for induced toroidal loop voltages in spherical tokamaks. A domestic 2.45 GHz magnetron based waveguide system has been used to facilitate the startup of plasma current on GLAST spherical tokamak. The continuous microwave output of the magnetron was converted into pulses of 5msec with peak power of 1.5kW. The neon was used as working gas after achieving base vacuum of 8×10-7 mbar and glow discharge conditioning of the glass vessel. The applied toroidal field (~ 0.1 Tesla) was so adjusted for the ECR layer (corresponding to 2.45GHz) to form at the inboard side close to the central glass tube. The values of toroidal magnetic field at positions R=5cm (inboard side near central glass pipe), 15cm (plasma center) and 25cm (outboard side near vessel wall) are respectively 0.1T, 0.04T and 0.02T. 6. Diagnostic Systems The diagnostics installed on the device are the differential loops, magnetic probes, flux loops, spectroscopy, high speed imaging and mass spectrometery to estimate the magnetic and plasma parameters. 7. Results and Discussion In the first step, the base vacuum in the glass vessel was achieved up to 8×10-7mb. The inside of the chamber was then cleaned using under vacuum heating at 70 oC for many hours and also repeatedly performing high pressure argon glow discharge. The temporal evolution of the impurity concentration was monitored with a residual gas analyzer RGA300. 7.1 Optimization of ECR absorption Firstly, the electron cyclotron absorption was optimized to provide a base for plasma current startup. The microwave pulse was launched into the vessel filled with neon gas pressure of about 10Pa in the presence of toroidal magnetic field (875gauss) to generate the ECR layer at the inboard side. The width and the intensity of resonance layer were optimized by changing the gas pressure, orientation of the waveguide and relative delay between TF and the microwave pulse. A microwave photodiode was used to record the temporal intensity changes in the microwave pulse while an optical photodiode was used to record the light emission from plasma as a result of ECR absorption. The temporal response of both the diodes at different gas pressures is shown in Figure 3. The width and also the intensity of the optical emission decrease with the decrease in neon gas pressure. The operating pressure range for neon gas was found to be 3×10-3 mbar---5×10-4 mbar from these studies.

0.30 2.25 0.7 Pa vacuum 2.00 1.5 Pa 0.25 0.3 Pa 1.75 0.15Pa 1.50 0.20 1.25 0.15 Pa 0.15 1.00

0.75 0.3Pa 0.10 0.7Pa

Relative Intensity of MW of RelativeIntensity 0.50 Relative Intensity (Photodiode) Intensity Relative

0.25 0.05 15Pa 0.00 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 Time (msec) Time (msec) (a) (b) FIG. 3. Temporal variations in the a) light emission b) in microwave pulse

3 EX/P4-22

7.2 Plasma Current Studies In the next step, it was tried to generate plasma current by applying loop voltages in the presence of microwave pulse and the toroidal magnetic field. The three pulses were fired at fixed neon gas pressure of 3×10-3 mbar with suitable relative delays between them. A small value of plasma current (0.15kA) was observed at loop voltages of 11volts with charging the central solenoid at 3.5kV. The plasma current was increased with increase in the applied toroidal loop voltages. The temporal evolution of plasma current at three values of loop voltage along with the corresponding changes in light emission signals are shown in Fig. 4.

1.0 10 0.45 4.5kV 4.5kV 5 0.40 0.8 4kV 0.35 0 0.6 3.5kV 0.30 -5 0.4 4kV 0.25 -10 3.5kV 3.5kV 0.20 0.2 4kV -15 0.15 4.5kV

PlasmaCurrent(kA) 0.0 LoopVoltage (volts) -20

0.10 Relative(Photodiode) Intensity -0.2 -25 0.05 -30 0.00 -0.4 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 Time (msec) Time (msec) Time (msec) (a) (b) (c) FIG. 4. Effect of Loop voltages on plasma current a) Loop voltage b) light signal c) plasma current onTemporal variations in the a) light emission b) in microwave pulse A plasma current of about 2kA was produced for about 0.5msec. A fast photo-diode and a flexible Rogowski coil of about 1.7m long encircling the whole cross section of vessel was used to record the signatures of plasma current. A small vertical field was then added with the help of vertical field coil system without changing other operating parameters. We were succeeded to enhance the plasma current with the maxima at 5kA by scanning the field in both directions and also by varying the vertical field from 10 gauss to 100 gauss. The optimum values of the vertical magnetic field were found to be between 40-50 gauss. A change in the direction of plasma current was observed with changing the direction of vertical field that confirmed the enhancement of plasma current due to the vertical field drift effect. A high speed camera with a frame rate of 5000fps was used to record the temporal formation of resonance layer and also the whole discharge scenario. The first visible layer appears at time 1.2ms after the firing of microwave at t=0. The intensity and the size of resonant layer increase with time (Fig. 5) upto about t=2.2ms. The plasma current formation starts at about t=2.4ms, maximum is at t=2.6ms and then starts to decay and ends at t=3ms while the visible glow ends at t=4ms (Fig. 6).

1.2ms 1.4 ms 1.6 ms 1.8 ms 2.0 ms 2.2 ms FIG. 5 Temporal evolution of ECR layer

2.4 ms 2.6 ms 2.8 ms 3.0 ms 3.2 ms 3.4 ms 3.6 ms 3.8 ms 4.0 ms

FIG. 6 Temporal evolution and decay of plasma current

4 EX/P4-22

With vertical field Without vertical field FIG. 7 Images at maximum plasma current with and without vertical field

The Fig. 7 shows the comparison between the central frames recoded at the peak of plasma current. The effect of the vertical field is obvious that fills the whole volume of the vessel. The recorded experimental signals are shown in Fig. 8. The microwave absorption and the variations in the loop voltage signals correspond to the appearance of current and the light emission peaks. The Fig.4 shows two images recorded with high speed camera at the time of peak of plasma current with and without vertical magnetic field. The increase in the light emission intensity and the vertical expansion of plasma is evident from figure.

Rogowsli coil Photo-diode Loop voltage Microwave pulse FIG. 8 Experimental signals We have also estimated edge electron temperature of plasma from the variation in the loop voltage and also from the spectroscopic data. The average value of the electron temperature is about 13eV.

8. Conclusion In conclusion, we have successfully generated the plasma current in our device. The future plan is to increase the width and amplitude of plasma current by effectively utilizing the vertical field drift effect with the addition of one or more vertical field capacitor banks. The improvement of the existing diagnostics and the installation of some new diagnostic systems like H, impurity measurements, x-rays and interferometry are also planned. Since the vacuum vessel is of dielectric material so the research aspect regarding the adiabatic compression is also a target of experimental activities in the near future. We are hoping to contribute significantly in the worldwide experimental fusion research.

9. References [1] Spitzer L and Harm R, Phys. Rev. 89 977 (1953)

[2] Woods L C, Theory of Tokamak Transport (WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim), (2006). [3] ANISIMOV A. I., VINWRADOV N. I. and POLOSKIBN. P., Socier Phys. Tech. Phys. 18, 459, (1973).

5 EX/P4-22

[4] ATKINSODN. W., BRADLEJY. E ., DELLIAS. N., JOHNSONP. C., LEES D. J., LLOYDB . et al., Plasma Physics and Controlled Research 1980, Proc 8th Int Conf, Brussels, Vol I. p. 153, IAEA, Vienna, (1981).

[5] BULYGINSKDII. G., LARIONOVM . M.. LEVINL . S., MIKLUKHO0 . V., TOKUNOAV. I. and SHUSTOVA N. V., Societ J . Plasma Phys. 6, 11, (1980).

[6] CHO T., KUBOS ., IKEDAM ., SAITOT ., TERUMICHYI. , HAMADAY . and TASAKAS., Phys. Lett.77A. 318, (1980).

[7] COSTLEYA. E., HASTIER . J., PAULJ . W. M. and CHAMBERLAJ.I N, Phys. Rec. Lett. 33. 758, (1 974).

[8] FISCH N. J., Phys. Rec. Lett. 41, 873, (1978).

[9] HOLLYD .J ., PRAGESR. C .. SHEPARDD A . and SPROTTJ . C., Xucl. Fusion 21, 1483, (1981).

[10] JOBES F. et al., Phys. Reo. Lett. 52, 1005, (1984).

[11] KUBO S. et al., Phys. Rec. Lett. 50. 1994, (1983).

6