Electrode discharge assisted electron cyclotron wave current startup on the CT-6B tokamak

Electrode discharge assisted electron cyclotron wave current startup on the CT-6B tokamak Shaobai Zheng, Xuanzong Yang, Diming Jiang, Xinzi Yao, Chunh...
Author: Jonah Boyd
2 downloads 0 Views 217KB Size
Electrode discharge assisted electron cyclotron wave current startup on the CT-6B tokamak Shaobai Zheng, Xuanzong Yang, Diming Jiang, Xinzi Yao, Chunhua Feng, Deyi Jiang, Tongzhen Fang, Long Wang Institute of Physics, Chinese Academy of Sciences, Beijing, China Abstract. A novel method of electron cyclotron wave current startup has been proposed, in which a discharge between a pair of electrodes is introduced and forms a magnified toroidal plasma current in a strong toroidal field and a weak vertical field. The method has been demonstrated on the CT-6B tokamak. The experimental results are consistent with a model that is suggested.

1.

Introduction

Electron cyclotron wave current startup (ECCS) is a scheme for generating toroidal plasma currents in tokamaks. Following this scheme, a high power microwave beam in the electron cyclotron frequency range is launched into the toroidal field in the device, generating a plasma and inducing a plasma current under appropriate conditions. This current startup scheme has been demonstrated on a number of tokamaks [1–6], and has improved the startup performance of tokamaks. In particular, it showed a substantial reduction of loop voltage of the ohmic discharges and a saving of magnetic flux of the transformer in the initial phase of the discharge. The ECCS concept may be also used to build an RF tokamak in which lower hybrid waves or other driven waves are injected into the initial plasma carrying the seed current generated by electron cyclotron waves, to sustain and ramp up the plasma current without the action of an ohmic transformer. The initial plasma current in the ECCS scheme may be driven by different mechanisms as discussed in Ref. [6]. Usually, a weak vertical field is externally applied or introduced by the toroidal coils as an error field. In such a field configuration a vertical component of parallel velocity for electrons moving in one toroidal direction is produced and may cancel the toroidal drift and form closed trajectories. On the contrary, the electrons moving in the opposite direction are lost through contact with the vessel walls. The difference in confinement times between the motions in the two directions results in a net toroidal current [2]. This drive mechanism is valid for energetic electrons with a longer mean free path in comparison with the toroidal scale of the vessel. These electrons are generated through resonance absorption of the microwaves. Nuclear Fusion, Vol. 40, No. 2

The second drive mechanism related to a weak vertical field was suggested by Forest et al. [5]. The charge separation caused by ∇B drifts can be shorted out by parallel currents and in the process generates a toroidal current. Unfortunately, it is difficult to distinguish between these two drive mechanisms experimentally. In the CT-6B tokamak, an initial plasma current up to 600 A was achieved with a microwave power of 60 kW [6]. The current drive efficiency defined by [7] η = ne Ip R/Pw

(1)

is of the order of 10−4 W−1 m−2 , rather low in comparison with that of ECCD in several tokamaks. In (1), ne is the averaged electron density, Ip is the driven plasma current, R is the major radius and Pw is the drive power. In order to increase the drive efficiency in the ECCS scheme, a novel method, similar to the mechanism suggested by Forest et al. [5], has been suggested, in which a weak vertical field is also needed, and a pair of electrodes is installed on the top and on the bottom of the vacuum vessel and a voltage is applied to these during discharges. As the microwave beam is injected and a plasma is generated, a discharge between the two electrodes is triggered and this produces toroidal current additional to the wave driven currents generated by other mechanisms. The electrode discharge assisted ECCS method has been experimentally demonstrated on the CT-6B tokamak. The present article is organized as follows. In Section 2 a simplifying model is suggested for the method. In Section 3 the experimental arrangement is presented. In Section 4 experimental results are presented and discussed. Conclusions are drawn in Section 5.

c

2000, IAEA, Vienna

155

Shaobai Zheng et al. increase in discharge current and introduced toroidal plasma current is not limitless. Various factors affect the maximum plasma current. The voltage applied to the electrodes is Vd = Vp − V+ + V− Figure 1. Magnetic field configuration and electron trajectories for electrode discharge assisted ECCS.

2.

Formulation

The basic idea of the ECCS scheme with an electrode discharge is shown in Fig. 1. If a DC discharge occurs between two electrodes, one located at the top and the other at the bottom of the vessel, the electrons carrying the current will follow an oblique field composed of a strong toroidal field and a weak vertical field, and move around the torus many times until they arrive at the other electrode. In other words, the electron trajectories form helices. Then the discharge current between the electrodes is magnified in the toroidal direction. If toroidal drift is ignored, the pitch of the helical trajectories can be expressed as 2πRBv d= Bφ

(2)

where Bv and Bφ are the vertical and toroidal fields, respectively. The height of the electrodes should be higher than d. The parameter values are R = 0.45 m, Bv = 20 G, Bφ = 0.5 T and d = 1.12 cm. The ratio of the plasma current Ip caused by the electrode discharge to the discharge current Id , i.e. the magnifying power M , is M=

Ip 2a aBφ = = Id d πRBv

(3)

where a is half of the distance between the electrodes, and should be the minor radius of the plasma column. For the above mentioned d value and a = 10 cm, the magnifying power M is 20. The voltage to generate a plasma current Ip is Vp =

2πRρ Ip Dd

(4)

where D is the transverse width of the electrodes and ρ is the plasma longitudinal conductivity. If D = 10 cm, Te = 20 eV, Ip = 1 kA and ρ is taken to be the Spitzer resistivity, then the voltage is obtained to be Vp = 6.6 × 10−4 V. Expression (4) shows that the plasma current is proportional to the applied voltage. However, the 156

(5)

where V+ and V− are the sheath potentials with respect to the positive and negative electrodes, respectively. The maximum current is the ion saturated current on the negative electrode. For plasma parameters of Te = 20 eV and ne = 2 × 1012 cm−3 , the maximum current is Id,max = j+ hD = 8.88 A

(6)

where h = d = 1.12 cm is the effective height of the electrodes. The corresponding plasma current is 178 A. The current drive mechanism occurs in an open force line configuration. Therefore the maximum poloidal field produced by the total plasma current is equal to the vertical field. The maximum total plasma current is then Ip,max = kBv /µ0

(7)

where k is a constant coefficient determined by the current profile. Combining (3) and (7), the optimal vertical field for given Id and Bφ is derived to be r µ0 aBφ Id Bv = . (8) πkR The corresponding plasma current is s kaBφ Id Ip = . πµ0 R

(9)

Since Id is proportional to the width of the electrodes, and so roughly to the minor radius of the plasma, we have the following scaling law: q Ip ∝ a Bφ /R. (10)

3.

Experimental arrangement

The experiment was conducted on the CT-6B tokamak. This is a small device with an iron core transformer. The major radius is R = 45 cm and the minor plasma radius a = 12.5 cm, defined by a fixed four block poloidal limiter in normal tokamak discharges. The toroidal field can be varied in the range of Bt = 5–13 kG. The stainless steel welded vacuum vessel has a minor radius of b = 15 cm and two toroidal breaks. The maximum saturated magnetic Nuclear Fusion, Vol. 40, No. 2 (2000)

Article: Electrode discharge assisted ECCS on CT-6B Two resistors are connected in series and in parallel for measurements of the discharge current and the electrode voltage, respectively.

4.

Figure 2. Discharge electrodes in a transverse crosssection of the vessel together with the power supply for the electrode discharge: C, storage capacitor bank; SCR, switch; R0, serial resistor; R1 and R2, resistors for measurements of discharge current and electrode voltage.

flux of the transformer in one magnetized direction is 0.14 V s. The primary winding of the transformer with 280 turns is shorted and the iron core is not premagnetized in the ECCS experiment. The vacuum vessel is continuously filled with the working gas of hydrogen. Two gyrotron systems are used as the power sources of the electron cyclotron waves. The first gyrotron is of type GY-25 operating at 20.14 GHz and in TE01 mode. The output microwaves are injected into the vacuum vessel through a circular waveguide and a Vlasov antenna installed on the top of the vessel. The microwave beam is launched at 60◦ to the toroidal axis, and then has a wavevector component parallel to the magnetic field. The maximum output power of the gyrotron is 80 kW. The second gyrotron is of type D4032 operating at 34.34 GHz and in TE03 mode. The output microwaves are transmitted by a waveguide with a 90◦ bend and also perpendicularly injected into the vessel through a horn antenna from the weak field side. The maximum output power of the gyrotron is 150 kW, but the injected power is only about 120 kW due to the losses in the bent waveguide. Two plate molybdenum electrodes of 0.5 mm thickness, 10 cm width and 1.5 cm height are installed on the top and on the bottom at the same toroidal location in the vacuum vessel as shown in Fig. 2. Their radial positions can be adjusted to control the distance between them, or the minor radius of the plasma, as the electrodes act as a limiter. An energy storage capacitor bank for ohmic heating in normal tokamak discharges is used for the electrode discharges. The maximum charging voltage is 400 V. A silicon control rectifier switch and a 20 Ω resistor are connected in series in the discharge circuit. Nuclear Fusion, Vol. 40, No. 2 (2000)

Experimental results

Most of the discharges in the experiment are conducted with the GY-25 gyrotron of 20.14 GHz frequency. The toroidal field is set to be 0.585 T, corresponding to low field injection in the fundamental resonance region. The resonance layer is located at R = 36.7 cm. The maximum started current was achieved in the toroidal field region in the previous experiment [6]. The filling gas pressure is (2–8) × 10−5 t. The induced plasma current has only a very weak dependence on the pressure in this pressure range. The toroidal field is always fixed to be clockwise as viewed from the top. This toroidal direction is called positive. We also define the vertical field needed to maintain the equilibrium of a positive plasma current in normal tokamak discharges (upwards) positive. Then a positive vertical field will induce a positive plasma current through the drive mechanism without electrode discharges. We also call an applied voltage on the electrodes positive if the upper electrode is at the lower potential (upward electric field). A positive vertical field and a positive electrode voltage will drive a positive plasma current through the magnifying effect. In other words, a positive electrode voltage will drive a current in the same direction as that driven by the previously discussed mechanisms involving a weak vertical field. A set of typical discharge waveforms is shown in Fig. 3. These are, from top to bottom, plasma current, loop voltage, horizontal displacement of the plasma column, electrode voltage, electrode discharge current and injected microwave power. The applied vertical field in this discharge is −20.5 G. It can be seen from the waveforms that when the microwave beam is injected, an electrode discharge is triggered and a plasma current is also created. The electrode voltage, which is applied 4 ms in advance of the discharge, drops slightly due to the series resistor. Figure 4 shows the dependence of the driven plasma currents on the applied vertical field for the electrode voltages Vd = 0 and 100 V. The data for Vd = 0 are similar to the previous experimental results [6]. The function curve is composed of one straight line and two hyperbolas. The straight line section is determined by the maximum poloidal field 157

Shaobai Zheng et al.

Figure 3. Set of typical discharge waveforms. These are, from top to bottom, plasma current Ip , loop voltage Vl , horizontal displacement ∆x, electrode voltage Vd , electrode current Id and microwave power Pw . The microwave frequency was f = 20.14 GHz and the applied vertical field Bv = −20.5 G.

Figure 4. Driven plasma currents as functions of applied vertical field for electrode voltages Vd = 0 and 100 V.

generated by the drive current, which should be less than or equal to the vertical field. The curve is asymmetric with respect to the abscissa because a current in the negative toroidal direction is driven by the 158

Figure 5. Plasma current introduced by the electrode discharge with Vd = 100 V as a function of applied vertical field.

parallel component of the injected microwave, which is independent of the vertical field. The current component is estimated to be −19 A from the data shown in Fig. 4. The curve is also asymmetric with respect to the ordinate because there is an error field of about 10 G introduced by the toroidal field. The appearance of the function curve for Vd = 100 V is very similar to that in the case without an applied electrode voltage. The toroidal plasma currents introduced by the electrode discharge can be obtained by removing the plasma current without an electrode discharge from the total current with an electrode discharge, i.e. calculating the difference between the two sets of data shown in Fig. 4. The results are shown in Fig. 5. The function curve is also composed of one straight line and two hyperbolas. The hyperbolic dependence is qualitatively consistent with the relation (3). The linear dependence section indicates that the coefficients in (7) are different for the discharges with and without electrode discharges as they have different cross-section or current profiles. Figure 6 shows the plasma currents as functions of electrode voltage for different applied vertical fields. The curves indicate that the induced currents change their signs when the electrode voltage or the vertical field changes its polarity if the vertical error field and the plasma current driven by other mechanisms are taken into account. The plasma currents produced Nuclear Fusion, Vol. 40, No. 2 (2000)

Article: Electrode discharge assisted ECCS on CT-6B

Figure 6. Driven plasma currents as functions of electrode voltage for different applied vertical fields.

Figure 7. I–V characteristics of electrode discharges for different applied vertical fields.

with Vd > 0 and Vd < 0 enhance and reduce the original plasma currents for Vd = 0, respectively. The behaviour is also consistent with the previous analyses. The functions are linear for low values of electrode voltage and saturated for values higher than a critical value. The saturations are probably due to the limited saturation ion currents. Figure 7 shows the electrode discharge currents as functions of electrode voltage, i.e. the I–V Nuclear Fusion, Vol. 40, No. 2 (2000)

Figure 8. Plasma currents introduced by the electrode discharge as functions of electrode discharge current for different applied vertical fields.

characteristics, for different applied vertical fields. The dependences are composed of one linear and two saturated sectors. The currents for zero electrode voltage (shorted circuit) have finite values (−2 to −5 A). These are drift currents and make no contributions to toroidal plasma current. The total drift current over the width of the electrodes (10 cm) is about 29 A for the plasma parameters Te = 20 eV and ne = 5 × 1012 cm−3 with an isotropic electron velocity distribution. The ion drift current is neglected due to the low temperature. The calculated current value is much higher than the measured value as some of the drift electrons return along the oblique magnetic field, or are lost by other mechanisms such as diffusion and electric field drift. There is a vertical error field of about 10 G in the case of Bv = 0. The existence of the measured drift currents indicates that the pressure drive mechanism [5] works in this experiment. It drives toroidal plasma currents of at least several deca-amperes. We subtracted the values of plasma currents without an electrode discharge from those with applied electrode voltages shown in Fig. 6, and obtained the plasma current introduced by the electrode discharge. The results are shown in Fig. 8 as functions of the electrode discharge current for different applied vertical fields. The data show linear dependences with different slopes, or magnifying powers. Each function crosses over the abscissa at a value 159

Shaobai Zheng et al. Table 1. Measurement and calculated values of magnifying powers for different vertical fields Bv (G)

Mm

Mc

−67.5 −36 −20.5 46.5

8.4 ± 1.2 15.4 ± 3.5 16.4 ± 4.2 8.7 ± 0.9

7.1 15.9 39 7.3

a (cm)

Figure 9. Plasma currents introduced by the electrode discharge with Bv = 70.4 G as functions of electrode discharge current for different interelectrode distances.

of discharge current above −2 A, which is the measured value of drift current as mentioned above. Only the function for Bv = 0 appears in a different way from the others. Its slope takes different values for positive and negative plasma currents, with a large scatter. In this case the vertical error field introduced by the toroidal field acts. Its configuration may be very complicated, and is very probably axially asymmetric. It is believed that the different magnifying powers for the opposite directions are caused from some asymmetry of the error field configuration. The measured values (slopes) Mm and calculated values Mc of the magnifying power from (3) for four applied vertical fields, except for Bv = 0, are summarized in Table 1. This table reveals that three sets of experimental data are consistent with the calculated values. The only deviation of the experimental results from the theoretical prediction occurs for Bv = −20.5 G, which is located in the straight 160

Table 2. Measured and calculated values of magnifying power for different interelectrode distances and Bv = 70.4 G

7 9 11

Mm

Mc

4.8 ± 0.4 6.4 ± 0.2 8.4 ± 0.4

4.8 6.2 7.5

section of Fig. 5. In this case the total plasma current is limited by a maximum poloidal field and does not follow (3). Figure 9 shows the dependences of the plasma currents introduced by the electrode discharge on the discharge current for different distances between the electrodes and an applied vertical field Bv = 70.4 G. The magnifying powers for different interelectrode distances can be obtained from the diagrams and compared with the calculated values, as shown in Table 2. The measurement data are consistent with the calculated values. An essential question is whether electrode discharges introduce more impurities into the plasma. We have monitored some impurity spectral lines (oxygen and carbon ion lines) during the discharges and compared their intensities in the cases with and without electrode discharges. These experimental results show that no obvious impurity effects have been observed in the discharge current level. Secondary processes on the electrodes may not be important in these experiments. Another gyrotron, of type D4032, operating at 34.34 GHz, is also used in these experiments. Figure 10 shows a set of discharge waveforms with this gyrotron. The broken lines in Fig. 10 indicate waveforms without an applied electrode voltage. The toroidal field is 1.08 T and the applied vertical field is −70 G. The scenario is of low field side injection in the fundamental frequency region with the resonance layer located at R = 40 cm. In this discharge, the electrode voltage is applied 2 ms after microwave injection. Before the electrode discharge, a voltage of about 10 V on the electrodes is produced by electron drift in the toroidal field, but no drift current is measured in Id as the circuit is open. A feature of these waveforms different from those of Fig. 3 obtained with the first gyrotron is an ‘afterglow’, lasting for about 2 ms, in the plasma current and the discharge current after the microwave has been turned off, due to higher microwave power. The maximum plasma Nuclear Fusion, Vol. 40, No. 2 (2000)

Article: Electrode discharge assisted ECCS on CT-6B The fundamental problem concerns the limited discharge current determined by the saturated ion currents on the electrodes. In order to enhance the plasma current, multiple electrode pairs may be used to increase the area of the electrodes. A hot cathode, or other electron sources, may be introduced to provide the original electrons. Furthermore, use of a more powerful microwave beam will be effective. The scaling law (10) for the maximum started current obtained by means of the suggested method indicates that it is advantageous for a tokamak with low aspect ratio A = R/a. In such a device, current startup is very difficult with an ohmic transformer because of the poor efficiency of the central solenoid. This difficulty may be solved with the suggested method.

Acknowledgement Figure 10. Set of typical discharge waveforms obtained with the second gyrotron operating at 34.34 GHz. They are, from top to bottom, plasma current Ip , loop voltage Vl , electrode voltage Vd , electrode discharge current Id and microwave power Pw . The applied vertical field was Bv = −70 G.

This work was supported by the National Natural Science Foundation of China under Grant No. 19685009.

current achieved with this gyrotron is 780 A, including the current of 600 A introduced by the electrode discharge. The higher initial current achieved with the gyrotron results from the stronger toroidal field and the higher plasma parameters (density and temperature) produced with the higher microwave power.

[1] Kubo, S., et al., Phys. Rev. Lett. 50 (1983) 1994. [2] Shimozyma, T., et al., J. Phys. Soc. Jpn. 54 (1985) 1360. [3] Lloyd, B., et al., Nucl. Fusion 31 (1991) 2031. [4] Tanaka, S., et al., Nucl. Fusion 33 (1993) 505. [5] Forest, C.B., et al., Phys. Plasmas 1 (1994) 1568. [6] Han, G., et al., Nucl. Fusion 38 (1998) 287. [7] Erckmann, V., Gasparino, U., Plasma Phys. Control. Fusion 36 (1994) 1869.

5.

References

Conclusions

A novel method of ECCS has been proposed, in which a discharge introduced between a pair of electrodes forms a magnified toroidal plasma current. The method has been demonstrated on the CT-6B tokamak and the experimental results are consistent with the simple model proposed.

Nuclear Fusion, Vol. 40, No. 2 (2000)

(Manuscript received 2 August 1999 Final manuscript accepted 11 November 1999) E-mail address of Long Wang: [email protected] Subject classification: H1, Te

161

Suggest Documents