[EN] In recent years, electric discharges have been found to be prolific sources of high-energy radiation [Abbasi et al., Dwyer et al., Enoto et al.]. In particular, terrestrial gamma ray flashes are observed in correlation to lightning activity [Østgaard et al.] and bursts of X-rays are measured in experiments with laboratory sparks [Kochkin et al.]. The gamma/X-ray photons are produced from Bremsstrahlung (braking radiation) of fast electrons scattering from atomic nuclei present in the vicinity of the discharge. The fast electrons may themselves either be spawned by cosmic radiation or accelerated in very intense and localised electric fields in the discharge.

The latter mechanism (acceleration) is the one studied in this thesis and denominated as “thermal electron runaway”. In the current state of knowledge, it poses many challenges to our understanding of discharges. For instance, the fluence of X-rays (number of photons traversing a unit area) surpasses the one predicted by current models of thermal runaway in discharges. Concretely, the probability of accelerating a thermal electron to high energies, where it radiates through bremsstrahlung, is found to be too low at typical electric fields encountered at the head of streamers (self-sustained ionisation waves). Conversely, the plausibility of very high electric fields which enable thermal runaway is uncertain.

From the modelling perspective, the abundance of thermal runaway electrons has been found to depend strongly on the model chosen to represent electron scattering with molecules [Chanrion et al., Diniz et al., Moss et al.]. In particular, in a comparative study, we showed that various models of elastic scattering give significantly different distributions of high-energy electrons. Hence, we identify two necessities precluding a sound study of thermal runaway:

1. Find a physical environment, composed of the electric field and the state of air, that fosters the conditions which enable thermal runaway. Then, in this environment, be able to model events of extreme rarity, down to arbitrarily low probabilities.
2. Possess a highly reliable and accurate modelling of electron-molecule collisions, consistent throughout a broad energy range: from zero to several MeV.

The first necessity is answered by the first part of the thesis where we investigate the abundance of high-energy electrons obtained in Monte Carlo simulations under various conditions of the electric field, the air composition and temperature. In a second article, we adapted the Monte Carlo importance sampling methodology into a “compaction” algorithm which enhances the statistics of high-energy electrons to an arbitrarily low probability of occurrence, however, at the cost of deteriorating the resolution of low-energy electrons.

The second necessity is addressed in the second part of the thesis, where an almost complete set of electron-molecule cross sections has been assembled independently from the databases that are currently in use. The assembly combined an exhaustive gathering (up to 2022) of experimental cross sections, accurate quantum mechanical calculations and simple analytical representations. The modelling of elastic scattering is based on our third article for calculating differential cross sections of electrons scattering elastically from diatomic molecules.

Furthermore, this thesis also contains a third part which supplements the first two parts with a thorough documentation of the process for constructing the new cross section database. It provides an overview of techniques for fitting experimental data and comparisons of various electron-molecule cross section databases currently in use.

So far, most of the literature has focussed on plausible mechanisms which lead to formation of intense electric fields in ionisation fronts. In this thesis, we turned toward a less frequented perspective by considering the change in chemical composition of air due to the discharge activity preceding thermal runaway. Contrary to the wont applied to dissertations in the sciences, this thesis is an original work which does not include text extracted from the publications written during the doctoral program. It is therefore not to be regarded as a reformulation of the content of these articles, but as the prime continuation thereof.

With the wherewithal that we developed – the compaction algorithm and the new set of cross sections – we have probed preliminarily the phenomenon of electron thermal runaway in hitherto understudied territories; low electric fields and varying gaseous compositions. As an open conclusion, we fancy that preconditioning of the gaseous medium by streamer coronas is relevant to unveil some of the mysteries shrouding our current understanding of thermal runaway.