Photoconductive Switching of a High Voltage Spark Gap

Abstract

Summary form only given. Laser wakefield acceleration promises the production of high energy electrons from table-top accelerators. External injection of a relativistic electron bunch into a laser wakefield requires bunches of the order of the plasma wavelength, typically 100 fs. Acceleration fields necessary to create these bunches have to be of the order GV/m. RF technology has reached its limit of acceleration with fields of the order 100 MV/m, but pulsed DC acceleration can go up to GV/m gradients. Compact pulsed DC acceleration in the GV/m region is possible if high voltage pulses of the order MV can be switched on ps timescales with ps timing precision. Presently rise time and jitter of high voltage pulses in laser-triggered spark gaps are limited to the (sub)-nanosecond regime by the initial, stochastic breakdown processes in the gap. Picosecond switching precision can only be achieved if these stochastic breakdown processes, like avalanche- and streamer formation, are omitted. At laser intensities above approximately 1018 W/m2, tunneling ionization causes near-instantaneous ionization of a complete plasma channel between the electrodes. Because of the instantaneous ionization and high degree of ionization in the plasma channel, jitter is reduced significantly and ps switching precision can be achieved. We have demonstrated photoconductive switching of an atmospheric high voltage spark gap. A 200 femtosecond, 1-35 mJ Ti:sapphire laser pulse is cylindrically focused into a 1 mm, air filled, spark gap biased at 4.5 kV. A clear transition is measured between triggering, when the gap is only partially ionized, and photoconductive switching, when the entire gap is almost instantaneously ionized by the laser. The measured rise time of the photoconductively switched high voltage pulse is smaller than 100 ps and the time jitter is less than 15 ps. We also measured at a smaller gap distance and with a flow of nitrogen in the gap. From measured Vapplied-Vout curves and preliminary simulation results, a qualitative description of the plasma behavior is deduced.