The physics of photoconductive spark gap switching : pushing the frontiers

Abstract

Photoconductive switching of an atmospheric, air-¯lled spark gap by a high-power fem- tosecond laser is a novel approach for switching high voltages into pulses with a very fast rise time (order ps) and almost no shot-to-shot time variation (jitter). Such a switch makes it possible to synchronize high-voltage pulses more accurately than presently possi- ble. The goal of this research was to create ultrafast high-voltage pulses in order to develop a new type of electron accelerator and its diagnostics. An interesting future application of photoconductively switched ultrashort pulses is the creation of broadband, high-intensity terahertz radiation, which is a harmless alternative to X-rays for a number of medical and security purposes. Photoconductive spark gap switching in air combines the bene¯ts of two ¯elds of high- voltage switching: First, laser-triggered spark gap switching where the switching medium is either gas or liquid and a laser is used to initiate the breakdown of the gap. Secondly, photoconductive switching where the switching medium is a semiconductor device, which is completely illuminated by a short pulsed laser. If a complete (gas-¯lled) gap is su±ciently ionized by a femtosecond, high-power laser, stochastic breakdown processes (dominating in the laser-triggered switch) no longer determine the actual breakdown-behavior of the gap. The rise time of the photoconductively switched pulse is then determined by the geometry of the gap. The time jitter is limited only by the jitter of the switching laser (as in the semiconductor switch) and, because the switching medium is a gas, high currents can be switched. The principle of photoconductive switching was demonstrated in air and in nitrogen (Chap- ter 3 and 4). A femtosecond Ti:Sapphire laser was cylindrically focused in a 1 mm spark gap biased at 4.5 kV. When su±cient laser power was used (> 0.1 TW) the spark gap switched photoconductively. The measured rise time and jitter of the switched pulses were both below the resolution of the measurement equipment, i.e., better than 100 ps and 15 ps, respectively. Measurements at lower applied voltages but with the same gap distance showed that it was possible to switch voltages as low as 10% of the self-breakdown volt- age. However, a voltage drop over the gap was observed, which became more pronounced when switching lower voltages. Transient-plasma simulations (Chapter 6) explained this behavior by showing that the conductivity of the plasma is a function of the current that runs through the plasma. Together with an interferometric study of the switching plasma (Chapter 5), the simulations also revealed that photoconductive switching does not require full ionization of the switching plasma. The voltage drop can be reduced when more laser power is used to create a switching plasma column that has the same initial electron den- sity but a larger diameter. A three-dimensional electrodynamic model to simulate a photoconductively switched high- voltage spark gap was developed (Chapter 7). This model describes and monitors the elec- tromagnetic ¯eld-propagation in a coaxial spark gap setup after switching. It reveals also the in°uence of discontinuities, such as viewing ports, on the pulse shape and the rise time. The rise time is determined by the time it takes for a stable TEM mode to build up in the gap region. Commonly used zero-dimensional lumped element and one-dimensional trans- mission line models for laser-triggered spark gap optimization are shown to be insu±cient for optimizing the geometry of the photoconductive switch, because the electromagnetic ¯eld-propagation in three dimensions is neglected. We developed an optimization procedure for spark gap geometries based on our fully three-dimensional electrodynamic simulations (Chapter 8). In conclusion, we proved the principle of photoconductive switching of high voltages in air. The shot-to-shot time stability and the voltage working range of a spark gap are greatly enhanced, compared to conventional laser triggering, and high-voltage pulses with a very fast rise time can be made and modelled. New possibilities for compact pulsed DC electron acceleration are opened up by photoconductive switching, as well as numerous applications in other areas of research.