There is currently a need for megawatt average power sources of 100–600 GHz radiation for electron cyclotron heating of fusion plasmas. One of the leading candidates for such a source, the conventional wave guide cavity gyrotron,(1) has produced impressive output powers and efficiencies at frequencies up to about 300 GHz. However, this gyrotron configuration is limited at high frequencies by high ohmic heating and problems with transverse mode competition due to the highly overmoded configuration, and with beam collection, since the beam must be collected along a section of the output waveguide. The quasi-optical gyrotron (QOG), first proposed in 1980 by Sprangle, Vomvoridis, and Manheimer,(3) features an open resonator formed by a pair of spherical mirrors instead of a waveguide resonator and has the potential for overcoming each of these limitations. The resonator mirrors can be well removed from the beam-wave interaction region, allowing a large volume for the interaction and low ohmic heating densities at the mirrors. The beam direction is transverse to the resonator so that beam collection is separate from the output waveguide. This geometry is particularly well suited to the use of a depressed collector for electron beam energy recovery. The QOG operates in the lowest-order transverse (TEMool) Guassian mode of the resonator, higher-order transverse modes being effectively suppressed by higher diffraction losses. This paper reviews recent progress toward the development of high-power quasi-optical gyrotrons for ECRH of fusion plasmas. It includes an overview of gyrotron theory in terms of normalized variables as they apply to the quasi-optical gyrotron for operation both in the fundamental and the higher harmonics. Scaling equations for the output power and resonator mirror heating by the RF are given. The design tradeoffs between annular and sheet electron beams are discussed as is the issue of beam space-charge depression in the open resonator. Recent advances in the analysis and design of QOG configurations capable of efficient and stable single-mode operation are discussed, showing the possibility of achieving 50% transverse efficiency in highly overmoded resonators. The application of a depressed collector is discussed as a means of recovering the energy in the axial motion of the spent electron beam and, thus, raising the output efficiency to near the transverse electronic efficiency. The problem of high field magnet design is addressed, for both fundamental and higher harmonic operations, the latter being necessary at very high frequencies. The design equations and tradeoffs are applied to the design of 1-MW, CW quasi-optical gyrotrons operating at 120 GHz, in the first and second harmonic at 280 GHz and in the second harmonic at 560 GHz. The output coupling for these 1 MW designs is 5–7% showing the potential for even higher powers per tube if sheet-beam electron guns can be developed. The estimated electronic efficiency of the fundamental harmonic designs is 23%, which leads to an output efficiency of 47% with the use of a depressed collector with a modest collection efficiency. The peak ohmic heating density is 500 kW/cm2 in all the designs. This leads to resonator mirror separations ranging from 127 cm for 120-GHz design, to 232 cm for the 560-GHz, second harmonic design. Finally, a simple output system composed of'elliptical and parabolic mirrors is described that converts the output radiation from the resonator into a parallel, quasi-Gaussian beam. Experimental programs are reviewed as well, including the recent experiment at the Naval Research Laboratory that produced frequencies ranging from 95–130 GHz and powers up to 150 kW. Operation in a single mode was observed at powers up to 125 kW despite the resonator being highly overmoded. Comparison is made with the theoretically-predicted region of single-mode operation. Recent progress in the experimental characterization of QOG resontors is summarized.
Read full abstract