Ionization waves and streamer-to-spark transition in in-liquid discharges: sub-ns time-resolved imaging and Bayesian analysis

In-liquid sparks are transient stochastic plasmas ignited with a strong electric field. Those discharges showed a promising future for several applications (e.g. NP synthesis). An understanding of the underlying mechanisms during discharge is far from being understood. A spark starts as a streamer that propagates toward the counter electrode and then transits to a hot plasma channel. Such transition was studied in gas phase discharges, and different kinds of ionization waves (IWs) are highlighted such as return streamers or return strokes (~10⁶-10⁸ m/s), or filaments (~10⁵ m/s). The observation of such transitions in liquid is challenging due to the small size, high propagation speed, and stochasticity of the discharge. Herein, we used the two immiscible liquids setup to enhance the generation of in-liquid discharges. Indeed, long (several mm) discharges are successfully produced. Therefore, we could perform time-resolved imaging with a focus on the streamer-to-spark transition. The results are analyzed using a Gaussian process method, and three distinct IWs were identified: the positive streamer (~3×10⁴ m/s), a return streamer (~2×10⁶ m/s) and then a filament (~3×10⁵ m/s). Further analysis/experiments are performed to unveil such unique properties of in-liquid discharges.