Formation of a 100-Ka Tokamak Discharge in the Princeton Large Torus by Lower Hybrid Waves

VOLUME 52, NUMBER 12 PHYSICAL REVIEW LETTERS 19 MARCH 1984

Formation of a 100-kA Tokamak Discharge in the Princeton Large Torus by Lower Hybrid Waves

F. Jobes, J. Stevens, R. Bell, S. Bernabei, A. Cavallo, T. K. Chu, S. Cohen, B. Denne, P. Efthimion, E. Hinnov, W. Hooke, J. Hosea, E. Mazzucato, R. McWilliams, R. Motley, S. Suckewer, G. Taylor, J. Timberlake, S. von Goeler, and R. Wilson Princeton Plasma Physics Laboratory, Princeton University, Princeton, 1Vew Jersey 08544 (Received 19 December 1983)

The development of noninductive current drive is of great importance in establishing the tokamak as a long-pulse or steady-state fusion reactor. Lower hybrid waves, carrying 200 kW of power at 800 MHz, have been launched into the Princeton Large Torus tokamak to initiate and drive the discharge current to a level in excess of 100 kA.

PACS numbers: 52.55.Gb, 52.80.-s

Up to the present time tokamak devices have the lower hybrid waves alone or by electron cyclo- been inherently pulsed, because they rely on a tron resonance. ' " In this Letter we report suc- transformer to drive the plasma current. This cessful efforts to create a target plasma and to raise current, in interaction with a transverse magnetic the plasma current to 100 kA by lower hybrid field, provides the restoring force needed to hold waves, unassisted by induction from the primary the plasma column in equilibrium away from the transformer or by the application of power at the vacuum walls. Prospects for the tokamak as a electron cyclotron frequency. fusion reactor would be greatly enhanced by the The experiment was carried out in the PLT development of a reliable method of steady-state tokamak, ' a torus with minor radius 0.4 m and ma- current drive. jor radius 1.32 m, at a toroidal field of 20 kG. Up A number of methods for achieving noninductive to 200 kW of rf power at 800 MHz was injected into current drive, all employing beams of radiation or the plasma from a six-element waveguide array in particles, have been suggested. The only method an outside port. Current drive was achieved with which has succeeded in sustaining all of the plasma two different couplers, one a narrow, 2-cm in = — current the absence of induction has been waveguide grill (n tt ck~~/cu 3.5) and the other a lower-hybrid current drive: Space-charge waves, broad 3.55-cm grill (n~~ —2). In each case the excited in the outer plasma layers by phased background gas, deuterium at a pressure of waveguide arrays, propagate to the center of the (1 —2) x10 5 Torr, was broken down by the rf torus, where they interact resonantly with the hot fields and a plasma of line-average density = 10'2 plasma electrons, forming a unidirectional hot- cm was formed. The phase angle between the in- electron tai1, which can carry a11 of the plasma dividual guides during this 10—30 msec initial stage current. This method has been used to supplement was set equal to 0' to minimize wave reflection at the transformer drive in the JFT-2, WT-2, Versator the mouths of the waveguides prior to plasma for- II, and JIPP-T-II tokamaks' and to maintain high mation. currents (up to 400 kA) for times long compared Following the breakdown stage the waveguide with typical current decay times in the Princeton phase was electrically switched to 90' or 135' so Large Torus (PLT) and Alcator C tokamaks. that the waves would propagate toroidally in one On the basis of present experimental results, direction and drive an electron current. With the maintenance of a steady-state plasma current in narrow grill, it was then possible to initiate and raise large future tokamaks by rf power alone appears to the plasma current with the lower hybrid waves; be difficult because of the high power require- simultaneously the vertical field, necessary to main- ments, but rf power should play a useful role in ex- tain the plasma torus in equilibrium, was raised by tending the current capabilities of present and fu- the normal PLT feedback control. With the broad ture tokamaks by assisting the transformer during coupler, however, careful programming of the ver- 7 8 current ramp-up and transformer recycling. With tical field proved ne-essary. As shown in Fig. 1 the these applications in mind, groups on PLT, WT-2, rf power was turned on with the vertical field biased and JIPP T-II have recently dern onstrated that at —30 6, at a level somewhat higher than that es- lower hybrid waves can drive the plasma current timated to cancel the stray fields from the toroidal above 4—10 kA in target plasmas created either by field. The vertical field was then slowly decreased @1984The American Physical Society 1005 VoLUMF 52,52 NUMBER 12 PHYSI CAL REVIEW LETTERS 19 MARcH 1984

—I20 0 1Q3 (a) M z t=0.2-0.3 sec Ct t=0.7-0.8 sec ~ Cl t =2.4-2.5 sec o 0 1.5

cr) ~ N ~4k j IT TW W O ~o = V ~ %T T 0 200- Ch 10 Ct W 0 250 500 750 he(keV)

0.0012 ~ LLJ 0.I-I I 0— .I ... l ..Ll. LL Y~fm~~pt l O) o OCL (d) O -0.5—

+ CL I &C 0 10 20 40 TIME {sec) MINOR RADIUS (cm) ' FIG. 1. (a) Time evoluvo u 1 ()' aver ' ' ield intensity; (d) loopl voltage; intensity. The re- ect variations th d'iTi erent timime groups.

towards zero an d ram p ed p sh py1 over 20 collisional issipation a e application ma r ier experiments') shs ould absorb les rgy' energy exchan ge with the , fo11o d b 2- 0 kA/ . D g e rf power must theerefore flo w into unmoni- ere was no ga toredt channelne s, such as rf 1 e plasma,

] Th e one-turn loooop volta e ig. 1, is The ff y is only 0 —0 meas d acuum a igher current le ve 1s in inductiv 1 d of this volt age and the plasma ischarges. '3 ' current is pro portional to thee net Poyntin r —x-ray measurem owing into the vacu th ive, ut in this exxper- ive, so that the at the current i ux flows ou e plasma into primarily by a fast electron ta'ail similar to that a poloidal fields. e in earlier ntenance pproximately 4% (— gy input to PLT wa ield ,ere d on axis and are Of this ener spread out over a 1 h 1 2(b)] Th a a yieldiel an upperr 1imit to the h ra ius since th e degree of contaminatio mission from th ' 4 rom Fig . 1 ; g(01 o t e rf in put energy: ing ul evaluation ofo tht etimeevevevolution of thee 1006 VOLUME 52, NUMBER 12 PHYSICAL REVIEW LETTERS 19 MARCH 1984

l50 tion, accelerated by the negative 0.1- to 0.5-V loop voltage arising from the current rise. This work represents the first use of lower hybrid waves for the dual purpose of creating a target plas- CQ ~ I 00- ma in a tokamak and driving the plasma current well tokamak l— into the range of typical operation for CA which closed magnetic surfaces are assured. One of I— the couplers —2 at 90' phase) requires special (n ~~ 50— programming of the vertical magnetic field. We be- O lieve that the initial boost of 0.01 V-sec by the vert-

C3 ical field is needed to generate a seed current of

Ct electrons in the & 20-keV energy range. At these

t f I t l t I t S I ~ r I I I f f f energies there is sufficient power in the rf wave 0.5 I.O l.5 2.0 spectrum to accelerate the electrons to higher ener- TlME (seel gies before they are lost along poorly developed magnetic-flux surfaces. The quasilinear theory of rf current drive, '6 which has motivated and guided much of the experimental effort in this field, is inadequate to explain the results of this experiment. According to this I- theory, current drive can be successful only if the

C3 number of fast electrons near the low-velocity end of the wave-number spectrum is sufficient to carry

I I 0 I the current. In the experiments on WT-2 the fast 0 0.5 l.o l.5 electrons are provided by cyclotron resonance heat- TlME (secI ing. In our experiment there is no obvious source variation the total microwave FIG. 3. (a) Time of of fast electrons in the 1—30-keV range except dur- emission (5—24 GHZ) near the electron plasma frequen- ing the first 20 msec of the discharge. The wave cy; (b) time history of the microwave emission near the spectrum launched by the grill may possibly be second and higher harmonics of the electron cyclotron broadened by repeated reflections of the lower hy- frequency ( ~ 117 GHz). brid ~aves at the surface' or by nonlinear wave in- teractions. ' ' Alternatively, the energy density of x-ray profiles. The very soft (& 10 keV) emission the waves may build up to such a level that trapping from the main body electrons shows a slow rise in and other strongly nonlinear processes occur. the central electron temperature from 3QO to 400 This work indicated that tokamak designers now eV during the discharge. The electron temperature have new options for plasma current drive, so that at r =10 cm varies from 250 to 350 eV. Evidence machine design and operation can be radically al- from carbon-ion C V emission indicates T, & 200 tered in an effort to achieve a more efficient and eV at r = 10 cm. Spectroscopic measurements of practical reactor, Doppler broadening of C V show that the carbon It is a pleasure to acknowledge the invaluable ion temperature is —100 eV, at r = 10 cm. contribution of the High Power rf Group, the PLT Fast electrons also radiate in the microwave range technical staff, and the Data Acquistion staff. We near the plasma frequency'5 and the harmonics of are grateful for continued support from Dr. H. P. the electron cyclotron frequency (Fig. 3).z The in- Furth. tensity of this radiation rises gradually with the This work was supported by the U.S. Department current until t =0.6—0.8 sec, when strong oscilla- of Energy under Contract No. DE-AC02-76- tions begin. A second tail in the x-ray spectrum CH03073. [Fig. 2(a)] appears at the onset of these oscillations. In past work on PLT this behavior has been associ- ated with velocity-space instabilities in the fast elec- tT. Yamamoto et at. , Phys. Rev. Lett. 45, 716 (1980). tron '~ the simultaneous tail, triggered by application zT. Maekaewa et al. , Phys. Lett. 85A, 339 (1981). of rf power and electric fields internal to the plasma S. C. Luckhardt et al. , Phys. Rev. Lett. 48, 152 column. In these experiments the instabilities may (1982). be driven by a fast electron tail in the reverse direc- 4K. Ohkubo et a/. , Nucl. Fusion 22, 203 (1982).

1Q07 VOLUME 52, NUMBER 12 PHYSICAL REVIEW LETTERS 19 MARCH 1984

sS. Bernabei et al. , Phys. Rev. Lett. 49, 1255 (1982). ternational Atomic Energy Agency, Vienna, 1977), Vol. M. Porkolab et al. , in Proceedings of the Ninth Interna 1, p. 21. tional Conference on Plasma Physics and Controlled Nuclear '3J. Stevens et al. , in Proceedings of the Third Joint Fusion Research, Baltimore, 1-8 September 1982 (Interna- Varenna Gre-noble International Symposium (Commission tional Atomic Energy Agency, Vienna, 1983), Vol. I, p. of the European Communities, Brussels, 1982), Vol. 2, 227. p. 455. 7N. J. Fisch, in Proceedings of the Third Joint Varenna '4S. von Goeler et al. , in Proceedings of the Fifth Topical Grenoble International Symposium (Commission of the Conference on Radio Frequency Plamsa Heating (Universi- European Communities, Brussels, 1982), Vol. 3, p. 841. ty of Wisconsin, Madison, Wisconsin, 1983), p. 96. W. Hooke et al. , in Proceedings of the Ninth Interna ~5P. C. Efthimion et al. , Bull. Am. Phys. Soc. 26, 975 tional on Plasma Physics and Controlled Nuclear Fusion (1981). Research, Baltimore, 1-8 September 1982 (International t6N. J. Fisch, Phys. Rev. Lett. 41, 873 (1978). Atomic Energy Agency, Vienna, 1983), Vol. I, p. 239. P. T. Bonoli, R. C. Engloda, and M. Porkolab, in sS. Bernabei, Bull. Am. Phys. Soc. 27, 960 (1982). Proceedings of the Fifth Topical Conference on Radio Fre tttS. Kubo et al. , Phys. Rev. Lett. 50, 1994 (1983). quency Plasma Heating (University of Wisconsin, Madi- ~ ~K. Toi, K. Ohkubo, K. Kawahata et al. , private com- son, Wisconsin, 1983), p. 72. munication. 8V. V. Parail and O. P. Pogutse, Sov. j. Plasma Phys. ' D. Grove et al. , in Proceedings of the Sixth International 2, 125 (1968). Conference on Plasma Physics and Controlled Nuclear '9C. S. Liu et al. , Phys. Rev. Lett. 48, 1479 (1982). Fusion Research, Berchtesgaden, West Germany, 1976 (In-

1008