A Coaxial Magnetron Injection Gun

- recent experimental results and an improved gun design -

B. Piosczyk

Forschungszentrum Karlsruhe, Association EURATOM-FZK,

Institut für Hochleistungsimpuls- und Mikrowellentechnik, D-76021 Karlsruhe, Germany

e-mail: bernhard.piosczyk@ihm.fzk.de; Tel.: + 49 7247 82 3541; Fax.: + 49 7247 82 4874

1. Introduction

Coaxial cavity gyrotrons with an RF output power of 2 MW, CW, operated at 170 GHz are of interest for the ITER tokamak. Development work performed at the Forschungszentrum Karlsruhe has demonstrated the feasibility of manufacturing such a gyrotron and information necessary for a technical design has been obtained. As part of the development detailed investigations have been performed on problems related to the coaxial arrangement [1,2]. Experiments have been performed on a coaxial gyrotron operated in the TE_31,17 mode at 165 GHz. This gyrotron is of demountable type with no sweeping of the electron beam along the collector surface. The power density at the collector limits the maximum achievable pulse length.

2. Recent experimental observations

When the pulse length was extended for the first time over about 10 ms a fast jump of the current I_ins to the insert was observed. I_ins rose from a value in the mA range up to very large values (> 20 A), limited only by the external circuit. The jump of I_ins occurred independently of the value of the beam current I_b. Even at I_b £ 1 A it has been observed when the pulse length was extended to about 30 ms. Finally, the coaxial insert was damaged by the related power loading. In detailed investigations it has been found that the effect is due to the build up of a Penning discharge in the rear part of the electron gun. In the cylindrically symmetric part between the coaxial insert and the cathode an electron trap is created by the electrostatic and magnetic fields as shown in Fig. 1. Thus a Penning type discharge may build up. In agreement with numerical simulations electrons from this discharge may diffuse along the magnetic flux surfaces towards the coaxial insert as observed experimentally. In order to verify the hypothesis of the Penning discharge, the geometry of the cathode and the insert has been modified such that trapping of electrons should not occur. Experiments performed with the modified geometry showed no limitation due to the above described effect. At I_b » 50 A a maximum pulse length of 22 ms has been achieved with an RF-output power of about 1 MW. At this current the pulse limitation is due to the temperature rise of the collector surface by about 600⁰C. At reduced beam currents the pulse length has been up to above 100 ms without problems. This confirms the given explanation and the suggested method of suppression.

After having suppressed the build up of a Penning discharge, the origin of the body current I_body has been investigated. Both the current to the insert, I_ins and the current to the outer wall contribute to I_body. In the measurements it has been found that (1) I_body is dominated by I_ins with a negligible contribution of the current to the outer wall: I_body » I_ins, and (2) the value of I_ins increases approximately linearly with the positive body voltage U_body which is applied to the anode, the gyrotron body and the coaxial insert (Fig. 2). The collector is kept at ground potential. At operation without depressed potential (U_body = 0 kV) I_ins is only about 2 mA and rises to about 35 mA at U_body = 27 kV as shown in Fig. 2. Furthermore the value of I_ins increases with time and becomes stationary after about 5 to 7 ms. The values given in Fig. 2 correspond to the stationary values. The background pressure has been measured to be between 10^-8 and few times 10^-7 mbar depending on current and pulse length. At the end of longer pulses with a strong pressure rise an additional increase of I_ins has been observed before a voltage breakdown.

To explain this behavior the following mechanism is suggested. By applying a body voltage a negative potential barrier arises in front of the collector. Thus electrons created by ionization of the background gas become trapped axially between the negative cathode and collector potential and radially by the strong axially symmetric magnetic field. Low energy electrons generated by ionizing the background gas are only able to escape either by radial diffusion or by diffusion in velocity space. In case of operation with U_body = 0 kV these electrons may drift along the magnetic field lines towards the collector. The current to the insert is assumed to be mainly due to diffusion of the trapped electrons across the magnetic field. For beam electrons with an energy around 90 keV the ionization has been estimated to be only about 10^-6 if the path is 1.5 m at a pressure of 10^-7 mbar. This means that at I_b = 50 A an equivalent electron current of only about 50 mA is generated by ionization. For U_body » 27 kV the measured value of I_ins is about 1000 times higher, however. In order to be able to explain this an additional ionization of the background gas due to the trapped electrons oscillating between the cathode and the collector is suggested. Under stationary conditions which are established after a few ms the electron rate generated by ionization has to be in equilibrium with I_ins. This means that along the beam path a background plasma is created in which an electron current is oscillating between the cathode and the collector. In the investigated case the amount of the trapped and oscillating electron current must be significantly larger than I_b. Whether the residual charge of the plasma is positive (necessary for compensation of the beam space charge [3]) or negative should depend on the balance between the ionization rate, the ion drain current towards the cathode and collector and the diffusion rate of electrons. As a summary, it follows that in operation with depressed collector a significantly more dense plasma along the beam path may be created. More detailed investigations are needed to understand the consequences on gyrotron operation.

3. Improved design of a coaxial magnetron injection gin (CMIG)

Based on the results concerning the build up of a Penning discharge due to trapping of electrons, a new CMIG gun has been designed. By conical shaping of the insert and the corresponding part of the cathode, electron trapping has been avoided. The gun will be fabricated for use in a 2 MW, 170 GHz coaxial gyrotron operated in the TE_34,19 mode.

References:

[1] B. Piosczyk et al., "Coaxial cavity gyrotron - recent experimental results", IEEE Trans.

Plasma Science, vol. 30, 2002, 819-827.

[2] B. Piosczyk, “A novel 4.5 MW electron gun for a coaxial gyrotron” IEEE Trans. Electron

Devices, vol. 48, 2001, 2938-2944.

[3] G. Dammertz et al., "Long-pulse operation of a 0.5 MW TE_10,4 gyrotron at 140 GHz",

vol. 24, 1996, 570-578.