Numerical and experimental analysis of THz sheet beam traveling wave tube amplifier (TWTA)

Abstract

Summary form only given. We report on the RF design, modeling, and particle-in-cell (PIC) analysis for the half vane staggered 0.22 THz traveling wave tube amplifier circuit. The simulations were conducted for the full length circuit (~ 27 mm) consisting of 60 periods without input or output couplers. Extensive cold-test (MWS) simulations were conducted for an efficient input/output coupler design for broadband matching with the slow wave TWT circuit. The back to back coupler response gives an S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sub> ~ -15 dB and S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">21</sub> ~ -1.3 dB at 0.22 THz that matched very well with the RF measurements with a 1 dB bandwidth of ~57 GHz. Mode conversion from the operating mode (TE <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">01</sub> ) to higher order modes was also studied and found to be at least down to -12 dB. The cold test simulation response of the TWT circuit including couplers (total length 37.644 mm) showed S11~ 14.66 dB and S21 ~ -3.07 dB at 220 GHz with a 1dB bandwidth of 44 GHz. This paper also presents the sever design proposed to be compatible with the micro-fabrication of the sheet beam circuits. The thickness of coating was varied (tapered) in the vanes for gradual change in wave impedance to minimize reflections. The thickness of coating of loss material in vanes (75 μm wide and 400 μm deep) and the conductivity of the coating material was varied to optimize the response. S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">21</sub> was decreased to less than -50 dB while keeping S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sub> varying around -5dB. Detailed PIC simulations were conducted to elucidate the device output response for the full TWT circuit incorporating the broadband couplers. For an electron beam of 20 kV and 0.25A, the output power was ~350 W for an input power of 50 mW with a gain of 38.4 dB. At the input port, reflections amounted to a power of 2.5W. A sever (distributed loss for length = 1/6 of total circuit length) was also incorporated in the full model and conductivity was swept from 1/10 to 100 S/m (that is possible by changing the doping levels of lossy materials e,g SiC). The maximum output power this distributed sever predicted was ~25 W and gain ~27 dB while keeping the reflections significantly low to about ~28 mW.