**1. Introduction**

Self-blast circuit breakers represent one of the main technologies for high-current interruption at high voltage. After contact separation, intense radiation emitted from the high-current arc leads to a considerable photo-ablation of the surrounding nozzle which causes a pressure buildup and finally a strong gas flow necessary for arc quenching around current zero (CZ) [1,2]. Polytetrafluoroethylene (PTFE) is typically used as the nozzle material and SF6 as the filling gas. However, the substitution of the greenhouse gas SF6 by more environmentally-friendly gases like CO2 is an actual trend. The pressure buildup due to strong arc radiation and nozzle ablation, as well as the arc quenching processes, are key issues of the successful current breaking and have been subject to a large number of scientific studies. The main questions concern the properties of the arc and the hot gas regions like temperatures and species densities which are required for a sufficient understanding of the processes. Optical methods, like emission and absorption spectroscopy, can provide such quantities under the demand that arc and hot gas regions are optically accessible. However, an optical access can only be realized by adapted construction of specific model circuit breakers (MCB) or by appropriate model experiments [3–5].

Meanwhile, a sufficiently good knowledge of the arc properties during the high-current phase and in the high-temperature regions (above 6000 K) of the arc has been developed from spectroscopic studies of switching arc experiments and MCBs (see e.g., [3]). This is because atomic and ionic species dominate in the high-temperature regions and generate an intense spectral line radiation which can be well used for the determination of temperature and species densities [6–8]. However, the analysis of low-temperature regions of the arc fringes, of the regions near nozzle walls and of the temporal phase of arc quenching is much more challenging due to low line radiation intensities.

The investigation of the phase around current zero by optical emission spectroscopy (OES) and the determination of arc temperatures during the arc quenching as close as possible to CZ was a topic of our accompanying paper [9]. An MCB using CO2 as a filling gas and a PTFE-nozzle experiment under ambient air were used for the analysis of line radiation of oxygen and fluorine atoms as well as of carbon ions. Both setups will also be used in this study and explained shortly in Section 2.

It is well-known from composition calculations of thermal plasmas that the dissociation of filling gases like SF6 and CO2, reactions with the ablation product C2F4 and metal vapor from electrode erosion can produce a number of molecular species in an intermediate temperature range before an almost complete dissociation of atoms occurs at higher temperatures (see e.g., [10]). Mixtures of CO2 with higher amount of C2F4 are expected to contain considerable amounts of molecules at temperatures above 3000 K, namely CF4, CF3, CF2, C2F, C3, C2, CF, and CO (in order of dissociation with increasing temperatures) [11]. Hence, the study of molecule radiation can help to analyze the interesting ranges of lower temperatures near the nozzle boundaries and in the arc quenching phases. Unfortunately, there is a very low number of such studies for arcs in corresponding gas mixtures and particularly for switching arcs.

Interesting candidates for the study of molecule radiation are the Swan bands of the C2 molecule (around 500 nm) or the violet band of CN (around 385 nm) because of the relatively intense radiation in the optical range. Emission and absorption spectroscopy of the C2 radiation have been used for example to study the structure of carbon arcs for nanoparticle synthesis [12–14]. The radiation of CN was analyzed in a study of the arc ablation of organic materials in ambient air with close relation to low-voltage switching [15]. Furthermore, both molecules have been more intensely studied in plasmas produced by laser ablation or in the laser-induced breakdown [16–20].

The occurrence of C2 molecules is expected in switching arcs in CO2 atmosphere or in the case of ablation of PTFE or organic wall materials. However, most of the recent research on Swan bands C2 was carried out by laser-induced breakdown spectroscopy. In case of lower laser irradiance, the production of C2 molecules is dominated by excitation of larger molecules like C3, C4 with electrons followed by photo-defragmentation, delivering exited C2 molecules. In case of higher power, excitation resulting from electron–ion and ion–ion recombination dominates [16]. The intensity distribution of the emission pattern varies depending on pressure and temperature. Thus, an estimation of the vibrational temperature can be realized by comparison of measured and simulated spectra [17,19]. Temperatures in a thermal argon plasma interacting with various insulating plastic materials at magnetically-forced arc movement [21] and temperature decay of thermal plasmas caused by polymer ablation using inductively coupled plasma irradiation [22] were investigated experimentally and numerically. As an example of a switching arc study, the absorption spectrum of the C2 Swan bands was analyzed in a low-voltage circuit breaker model [23]. An arc moving between polyethylene walls was considered, and the density and the rotational temperature of the C2 molecules were determined from the absorption spectrum, which indicates the ablation of the plastic walls. Reports on the analysis of molecule radiation, the C2 Swan bands in particular, in high-voltage switching experiments as representative for high-voltage circuit breakers are missing so far.

During the OES study of an MCB and a nozzle experiment described in our first paper [9], molecule radiation of C2 and CuF was recorded under different conditions and to some extent in unexpected ranges of the arc. The occurrence of strong temperature gradients in the arc are already known as well as a number of molecules that might be expected to appear favorably at lower temperatures, i.e., either in the vicinity to the nozzle walls or at low currents. However, the occurrence of molecular species has been described by theoretical models (see, e.g., [11]), with a lack of experimental confirmation in many cases. In the present paper, it should be shown that some molecules are detectable under strongly varying conditions. The molecule CuF is expected when copper vapor from the electrode erosion is mixed with the dissociated PTFE vapor from the nozzle ablation [10]. The results for molecule emission and absorption should be given in this second paper in detail. The aim is to demonstrate the occurrence of molecule radiation as a possible candidate to characterize low-temperature regions in self-blast circuit breakers as well as ablation processes. However, the determination of quantities like rotational temperatures and densities is out of the scope of the present paper. The MCB and nozzle experiment setups will be presented shortly in Section 2 together with the setup for spectroscopic measurements because details can be found in [9]. Results are given in Section 3 followed by a discussion in Section 4.

### **2. Materials and Methods**

Two setups of electrodes and nozzles were used. They are described in detail in an accompanying paper [9]; basic features are sketched in Figure 1. Actually, the majority of experiments described in this paper were carried out with setup (b) and only a few with setup (a). The electrodes were made of W–Cu with a 10 mm diameter and had a fixed distance of 40 mm. Nozzles made of PTFE doped with <0.5 wt% molybdenum disulfide (MoS2) with an inner diameter of 12 mm were placed around the electrodes: Either setup (a) was applied with one 126 mm long, tubular-shaped nozzle of 50 mm outer diameter for strong ablation and high pressure built-up or setup (b) was used with two nozzles of about 50 mm length and 104 mm outer diameter separated by 4 mm distance to form a heating channel. At the electrode positions, the nozzle diameter was increased to about 16 mm for an exhaust gas flow.

The arcs were operated either under ambient conditions (setup (a)) or in a vessel filled with 1 bar CO2 (setup (b)) as part of a model circuit breaker similar to [3]. Windows in both the model chamber and the vessel allowed a free view through the nozzle and hence absorption experiments.

**Figure 1.** Setups (**a**) with a closed, long polytetrafluoroethylene (PTFE) nozzle for experiments with strong ablation and high pressure built-up and vertical observation slits in the middle and (**b**) with two separated PTFE nozzles forming a heating channel for plasma flow into a heating volume as used for the model circuit breaker.

Sine-like currents were applied for setup (a) with about 100 Hz frequency and 11 kA peak current. For setup (b) with 50 Hz and 5.3 kA. Thin Cu wires were used to initiate the arc discharges. Currents were measured using Rogowski coils. In case of setup (b), a pressure sensor (603 A from Kistler) was placed in the heating volume of the model circuit breaker.

Optical access was realized by vertical slits of 2 mm width that were sealed by 2 mm-thick quartz plates, ranging over the complete nozzle diameter. After each shot, the sealing plates were checked visually and exchanged; the transmission was measured regularly. Pairwise placement at opposite positions enabled background illumination and absorption measurements. For setup (a) the observation slits were placed in the middle between both electrodes. In setup (b), the observation point was positioned in one of the nozzles at half distance between electrode tip and nozzle exhaust, i.e., ∼9 mm away from both.

Different methods were applied for the optical analysis. Firstly, high-speed cameras (HSC) from Integrated Design Tools (IDT) were used to observe the general discharge behavior: Y6 with 24-bits color or Y4 with 10-bits monochrome. Secondly, optical emission spectroscopy was carried out by means of an imaging spectrograph with 0.5 m focal length (Roper Acton SpectraPro SP2500i). The nozzle slit was imaged on the entrance slit of the spectrograph to spectrally investigate arc cross sections, i.e., perpendicular to the arc axis. Using the spectrograph with Y4 HSC enabled to record series of 2D-spectra with typical repetition rates of 100 μs (frame rate 10 kfps), allowing rather long exposure times up to 98 μs that were necessary due to limited camera sensitivity. Alternatively, the HSC could be replaced by an intensified CCD camera (PI-MAX4 from Princeton Instruments) with higher sensitivity, allowing single frame acquisition of shorter exposure times even at lower intensities, e.g., around current zero. In a compromise between light intensity, spectral resolution, and exposure time, the entrance slit of the spectrograph was set to 50 μm. With gratings of 150 lines per mm for overview and 1800 L/mm for detailed spectra, the spectral range was 150 nm and 10 nm and the spectral resolution 0.3 nm and <0.1 nm, respectively. The intensity of side-on spectra was calibrated in units of spectral radiance by means of a tungsten strip lamp (OSRAM Wi 17/G) at the arc position. The window transmission of 50–70% was taken into account, mainly resulting from the coating of the quartz plates at the nozzles.

Thirdly, broadband absorption spectroscopy was carried out around CZ. Therefore, a background illumination was required with radiances higher or comparable to the emission of the arc. It was supplied by a pulsed high-intensity xenon lamp with a radiance similar to a Planckian radiator of 12,000 K [24]. The square-shaped pulse had about 1 ms-width at about 1 MW electric power, delivering a nearly constant emission intensity during the plateau phase.

Figure 2 shows exemplary current waveforms of the arc discharge around current zero (top, offset after CZ is caused by the Rogowski coil) and the quasi-rectangular pulsed current of the xenon lamp (bottom, red) as well as the spectrally integrated intensity measured by video spectroscopy (spectral range 400–800 nm). Since the electric pulse feeding the Xe-lamp was not perfectly rectangular, a heating phase of the xenon lamp could be observed. Thus, several Xe atomic lines were found in the first 100–200 μs of the 1 ms-pulse before a transition towards the 12,000 K-continuum emission. Additionally, with decreasing current also the emission intensity decreased. Hence, for the OAS analysis only the lamp's plateau phase was applied with a duration of about 700 μs. This relatively long, stable phase allowed for temporal investigation of absorption, e.g., compared to Z-pinches with some 10 μs of varying radiation intensity as used in [23].

**Figure 2.** (**top**) Arc current at the end of discharge. (**bottom**) Xe lamp current (red) and development of its emission intensity (blue squares).
