**4. Discussion**

Information was obtained from spatially and temporally resolved video spectroscopy using HSC. That comprised the different phases of discharge and the occurrence of Swan band emission from C2 molecules. These Swan bands could be observed under varying conditions. Different amount of ablated PTFE from the nozzle wall and plasma temperature were generated depending on nozzle geometry and current density. According to the equilibrium composition calculations by Yang et al. [11], a considerable radiation of the C2 molecule indicates plasma temperatures in the range from 4000–6000 K. Firstly, there was an occurrence very close to the nozzle walls as typical behavior for cases of moderate PTFE influence, i.e., when the current density was not too high and the temperature close to the wall was rather low, allowing the existence of carbon dimers. Although it might be often neglected when the temperature distribution in the arc is investigated, the Swan bands represent the existence of carbon molecules due to wall ablation and thus, an important effect of cooling and change of plasma composition. Secondly, with higher current densities, the Swan band patterns were also distributed over the full vertical axis of the side-on 2D-spectra. However, it was found by Abel inversion that the Swan bands are emitted in a thin sheath at the nozzle wall. Thirdly, a different distribution was found under extreme conditions, i.e., with single long PTFE nozzle and high peak currents. The arc plasma was completely dominated by PTFE material and temperatures were moderate in the arc center, proved by weak or non-visible ionic and atomic carbon line emission. The Swan band pattern was emitted with the highest intensity in the central position though emission was extended to radial positions of the nozzle wall. Finally, Swan bands also appeared as an absorption pattern at moderate currents with setup (b). Emission from hot plasma in the arc center (proved by C II line emission) served as an internal background radiator that was absorbed by the much cooler carbon dimers near to the nozzle wall.

A considerable amount of CuF molecules in the high-current arc as well as around CZ was found from absorption spectra. This was not expected before, for several reasons, especially regarding that no other molecules were observed close to current-zero. A possible explanation is as follows: Around CZ it is expected that convective fluxes are significantly reduced due to equalization of pressures. As a consequence, copper atoms from still hot electrodes may expand diffusively along the nozzle and reach the position of OES (nozzle slit). In parallel, fluorine-containing molecules are still released from the nozzle wall. CuF molecules could be formed by chemical reaction of atomic F and Cu either at the hot W–Cu electrode surface followed by CuF evaporation or in the gas phase with Cu evaporated from the electrode. Similarly, the observed absorption during the high-current phase might be explained by the gas flow out of the nozzle into the heating chamber. In this case, copper atoms eroded or evaporated

from the W–Cu electrode might be flushed with the stream towards the heating channel, reacting on its way with fluorine from the wall, and being detected at the observation slit by absorption with the arc plasma as background radiator. However, during the time immediately after flow reversal, i.e., about 1 ms before CZ, the situation is very different: the gas flow is directed from the heating channel towards the electrodes. Thus, no copper from the electrodes should reach the observation area with the slit and react with fluorine. That means that probably no CuF should be produced at this period; any detected CuF should be a survivor from the heating chamber. As a pity, at the moment database is not sufficient to answer the question if there is a lower CuF concentration after flow reversal or not. In the video spectra there is simply not enough background emission to enable sufficient signal for an absorption.

Within the described experiments, limitations of reproducibility, fluctuation in transmission due to particles, film layers on windows, and dust did not allow temporally and spatially resolved determination of the absorption by CuF, e.g., using two-dimensional inverse Abel transformation of video spectra with higher spectral resolution. Nevertheless, this would be the next step if significant technical improvements were done. On the one hand, further optimization of the nozzle slit, its position and manufacturing technique might provide even fewer changes of the gas and droplet flow conditions, thus allowing measurements still closer to the undisturbed conditions at the nozzle. On the other hand, the observation technique itself might be improved, too. Nowadays, the advantages of intensified and high-speed video cameras can be combined in new generations of cameras or boosters. The background illumination could be improved, too. Beside improvements in the optical path in order to enhance the intensity and homogeneity, the pulsed xenon lamp might also be replaced by a laser-driven light source with extended pulse duration. As a consequence, quantification of the CuF absorption after CZ as well as during the arc discharge might be possible. Furthermore, OAS regarding Swan bands could be tackled. Last not least, tests with other electrodes should be carried out to finally prove the origin of absorption by CuF-molecules, e.g., made of pure tungsten.

Altogether, the possibilities of a recording of molecule radiation emission and absorption in the visible spectral range have been demonstrated for the case of high-current ablation dominated arcs. Using PTFE nozzles, tungsten–copper electrodes and operation in air or CO2, the Swan bands of the carbon dimer C2 and absorption of the CuF molecule were the only detectable radiation patterns. However, these patterns open up ways for a study of interesting ranges in high-current breaking processes like the colder plasma ranges near the nozzle walls and the time around CZ.
