Decomposition Characteristics of SF₆ under Arc Discharge and the Effects of Trace H₂O, O₂, and PTFE Vapour on Its By-Products

Abstract

The research on decomposition characteristics of SF₆ and its by-products have great significance to the operation, maintenance, condition assessment and fault diagnosis of power equipment. In this paper, the particle composition models of SF₆, SF₆/polytetrafluoroethylene (PTFE), SF₆/PTFE/O₂, SF₆/PTFE/H₂O, and SF₆/PTFE/O₂/H₂O were established by using Gibbs free energy minimization method, and the effects of trace H₂O and O₂ impurities and PTFE vapour on SF₆ by-products were studied by the models. In order to verify the correctness of the simulation results, a series of breaking experiments were carried out on a 40.5 kV SF₆ circuit breaker, and a gas chromatograph was used to detect and analyse the SF₆ by-products. It was found that when PTFE vapour is involved in the arc plasma, the main by-product after arc quenching is CF₄, and the molar fractions of C₂F₆ and C₃F₈ are very low. When O₂ is involved, the main by-products are SOF₂, SO₂ and SO₂F₂, and a small amount of CO and CO₂ was also produced. When H₂O is involved, the main by-products in simulation are SOF₂, SO₂ and HF, and a small amount of SO₂, CO₂, CO, SO₂F₂ and H₂ was also produced. The experimental results are in good agreement with the above results.

1. Introduction

SF₆ is a greenhouse gas. It is one of the six gases included in the Kyoto Protocol aimed at reducing greenhouse gas emissions. The global warming potential (GWP) of SF₆ is 23,900 times greater than CO₂ according to the 2013 report of Intergovernmental Panel on Climate Change (IPCC), and it has a lifetime in the atmosphere of 3200 years. However, SF₆ has excellent insulation performance, arc quenching performance, and molecular stability. It is widely used in high voltage circuit breakers [1,2]. The high voltage SF₆ circuit breaker will inevitably produce a high temperature and high energy arc in the process of breaking currents. It will cause SF₆ to decompose and form SF₆ arc plasma. Many sulfur fluorides such as SF₅, SF₄, SF₃, SF₂, SF, S, and F will be produced in the arc plasma. As the arc extinguishes, the arc temperature will gradually decrease, and these species will recombine into SF₆ molecules [3,4]. However, there are trace H₂O and O₂ impurities in the circuit breaker, and the nozzle made by polytetrafluoroethylene (PTFE) material will be ablated by arc to produce trace PTFE vapour during the breaking process. The H₂O, O₂, and PTFE impurities will participate in the reaction of SF₆ arc plasma to form various by-products such as SOF₂, SO₂, SO₂F₂, CF₄, C₂F₆, C₃F₈, CO₂, CO, and HF [5,6].

On the one hand, some by-products such as SOF₂ and HF are corrosive and toxic, thus posing considerable threat to the safe and stable operation of equipment and health of the operation and maintenance personnel [7]. On the other hand, there is a close relationship between some by-products and the discharge faults in the equipment, so the decomposition component analysis (DCA) technology can be used in fault detection and condition assessment of SF₆ power equipment [8,9]. Therefore, the decomposition characteristics of SF₆ under arc discharge and the effects of trace H₂O, O₂, and PTFE vapour on its by-products have attracted extensive attention.

Many scholars have carried out a lot of research on this problem and achieved rich results. In terms of the experiments, Boudene et al. studied the decomposition products in detail under the condition of voltage of 60 kV, a current of 4.5 kA and arcing time 40–80 ms. They found that the gas production rate of SOF₂ and SO₂F₂ and arc energy is almost linear [10]. Belmadani and Casanvas examined the SF₆ byproducts of power arc discharges using gas chromatography [11,12]. The maximum arc current reached 8.3 kA, and the primary byproducts were SOF₂, SO₂F₂, SO₂, and CF₄. The concentrations of these byproducts decreased in the following order: SOF₂+SO₂ > CF₄ > SO₂F₂. Our previous work used a circuit breaker to study the influence of trace H₂O and O₂ on SF₆ by-products [13]. We found that the increase of the concentration of H₂O and O₂ will increase the production of SOF₂+SO₂, and the concentration of CF₄ is hardly affected by the concentration of H₂O and O₂. Andrzej pelc studied the generation of negative ions from SF₆ gas by means of hot surface ionization. He found eight ion species: SF₅^–, F^–, SF₆^–, SF₄^–, SF₃^–, SF₂^–, SF^– and F₂^–, with ion current intensities ratios of 1000:200:100:10:5:0.5:0.5:0.05. He also found the optimal temperatures at which the maximum of the ion current intensity is observed were estimated in the 1830–2000 ± 10 °C range [14].

In terms of the simulations, Brand et al. studied the particle compositions of SF₆ arc plasma at standard atmospheric pressure [15]. Chervy and Gleizes considered the influence of copper vapour on the particle compositions and proposed a particle composition model of an SF₆/Cu mixture [16]. Coufal studied the effect of PTFE ablation on particle compositions, and the Gibbs free energy minimization method was first used in the calculation [17]. Wang et al. studied the arc plasma particle compositions of an SF₆/CF₄ gas mixture [18]. Up to now, there are few reports about SF₆/PTFE/H₂O, SF₆/PTFE/O₂, and SF₆/PTFE/H₂O/O₂ gas mixture arc plasma. The arc plasma model cannot be used to study the comprehensive effects of trace H₂O, O₂, and PTFE vapour on SF₆ by-products. It seriously limits the reliability of the calculation results.

In order to solve this problem, the arc plasma models of SF₆, SF₆/PTFE, SF₆/PTFE/H₂O, and SF₆/PTFE/O₂ are established by using the Gibbs free energy minimization method in this paper. The effects of trace H₂O, O₂, and PTFE vapour on SF₆ by-products was studied by these models. On this basis, a comprehensive model of SF₆/PTFE/H₂O/O₂ gas mixture was finally established. The research on the particle composition of SF₆ arc plasma is improved. In order to verify the correctness of the calculation results, a series of breaking experiments were carried out on a 40.5 kV SF₆ circuit breaker, and a gas chromatograph was used to detect and analyse the SF₆ by-products. It was found that the simulation results are in good agreement with the experimental results.

2. Methods

Under the assumption of local thermodynamic equilibrium (LTE), the physical parameters of the system are only functions of temperature and pressure. The minimization of Gibbs free energy means that the Gibbs function of the equilibrium state is the smallest when the temperature and pressure are constant [17]. In this paper, the particle composition of SF₆ arc plasma is studied by this method. In pure SF₆ arc plasma, 12 kinds of particles such as SF₆, SF₅, SF₄, SF₃, SF₂, SF, S, S₂, F, F₂, S₂F₁₀, and FSSF were considered. In SF₆/PTFE arc plasma, 12 kinds of particles such as C₃F₈, C₂F₆, C₂F₅, C₂F₄, C₂F₃, C₂F₂, C₂F, C, CF, CF₂, CF₃, and CF₄ were added. In SF₆/PTFE/O₂ arc plasma, 10 kinds of particles such as O, O₂, SOF₂, SO₂F₂, SO, SO₂, CO₂, CO, COF, and COF₂ were added. In SF₆/PTFE/H₂O arc plasma, 9 kinds of particles such as H₂O, OH, H, HF, H₂, CH₄, CH₃, CH₂, and CH were added. In SF₆/PTFE/H₂O/O₂ arc plasma, all the 43 particles mentioned above were considered. The thermodynamic data of all particles, such as standard enthalpy of formation, entropy, and specific heat at constant pressure can be obtained in the JANAF database [19]. All calculations were completed in the Chemkin software.

In order to verify the correctness of the calculation results, a series of breaking experiments were carried out on a 40.5 kV SF₆ circuit breaker. The rated voltage was 40.5 kV, the rated current was 2.5 kA, the rated frequency was 50 Hz, the rated air pressure was 0.6 MPa, the rated short circuit breaking current was 31.5 KA, and the volume of single-phase air chamber was 30.0 L. Figure 1 shows the experimental arrangement. Before the experiment, SF₆ gases with different concentrations of H₂O and O₂ were filled into three chambers respectively. Sinusoidal current was generated by an L-C circuit with pre-charged capacitor banks. Arc current was about 10 kA. Arcing time with one current half-wave was 6.5 to 10.0 ms due to the mechanical dispersion of the actuator. Arc voltage and arc current were measured by a high voltage probe and Rogowski coil respectively, and recorded by oscilloscope. After each breaking experiment, the SF₆ by-products in the chamber were analysed by gas chromatograph.

This gas chromatograph was equipped with a hydrogen ionisation detector (FID) and a pulsed flame photometric detector (PFPD) to detect SF₆ by-products. The carrier gas was He with purity over 99.99%, the output pressure was 0.5–0.6 MPa, the injection temperature was 150 °C, the injection volume was 250 μL, the split ratio was 20:1, the column temperature was 50 °C for 5 min, and 10 °C/min is raised to 200 °C for 10 min. Figure 2 shows the test result of gas chromatograph. The detected gases in channel A were H₂, O₂, N₂, CO, CH₄, CO₂, CF₄, and C₂F₆. The detected gases in channel B were SF₆, SOF₂, SO₂F₂, SO₂, COS, C₃F₈, and CS₂. Trace H₂O content was measured using a SF₆ mirror dew point instrument. This instrument is often used in the field detection of trace H₂O content in SF₆ circuit breaker, with the advantages of fast response and high precision.

The specific experimental settings are shown in

Table 1. Setting of breaking experiment with different H₂O concentration.

Phase	H2O Concentration	Breaking Current	Breaking Times
A	106 ppm	~10 kA	5
B	748 ppm	~10 kA	5
C	1131 ppm	~10 kA	5

and

Table 2. Setting of breaking experiment with different O₂ concentration.

Phase	O2 Concentration	Breaking Current	Breaking Times
A	59 ppm	~10 kA	5
B	736 ppm	~10 kA	5
C	1202 ppm	~10 kA	5

. In order to research the effect of trace H₂O on SF₆ by-products after the arc was extinguished and verify the correctness of the simulation results, the three gas chambers of the circuit breaker were filled with SF₆ gases with different H₂O concentration. The concentration of H₂O before the breaking experiment in A, B, and C chambers were 106 ppm, 748 ppm, and 1131 ppm, respectively. The breaking current was 10 kA and the times of breaking is 5. The setting of the breaking experiment with different O₂ concentration was similar to that of H₂O.

3. Results and Discussion

3.1. Particle Composition of SF₆ and SF₆/PTFE

Figure 3 shows the particle composition of pure SF₆ arc plasma at 0.6 MPa. It can be found that SF₆ molecules begin to decompose at about 1000 K and forms SF₅, SF₄ and F atoms first. With the increase of arc temperature, SF₃, SF₂, and SF begin to form. It should be noted that the temperature required for the formation of SF₅, SF₄, SF₃, SF₂, and SF increases step-by-step, and the maximum mole fractions of SF₄ and SF₂ is significantly greater than that of SF₅, SF₃, and SF. The main reason is that the decomposition process of SF₆ is a process of gradually breaking the S-F bond to form a lower-level low fluorine sulfide, and the S-F bond energy of SF₄ and SF₂ is higher than that of SF₅, SF₃, and SF. When the arc temperature is greater than 3000 K, the mole fractions of all sulfur fluorides begin to decrease, and gradually decompose into S and F atoms. It should be noted that the arc temperature decreases gradually during the arc decay process. When the arc temperature decreases below 1000 K, all sulfur fluorides and atoms recombine into SF₆. The results of this part show good agreement with previous work [20], which proves the reliability of our calculation results.

Figure 4 shows the particle composition of SF₆/PTFE arc plasma at 0.6 MPa. Compared with pure SF₆ arc plasma, CF₄, CF₃, CF₂, S₂F₁₀, etc. appear in particle composition. CF₄ and CF₃ are formed by the combination of CF₂ radicals produced by the decomposition of PTFE and the F atom produced by SF₆ decomposition. The formation of CF₄ will lead to the lack of F atoms in the arc plasma, so some SF₅ radicals cannot obtain the F atoms to form SF₆, and these SF₅ radicals will combine with each other to form S₂F₁₀. It should be noted that as the arc temperature gradually decreases to room temperature, almost all PTFE vapour is converted into CF₄, and C₂F₆ and C₃F₈ are hardly generated. The molar fraction of C₂F₆ and C₃F₈ are both less than 10⁻⁶.

Figure 5 shows the change of CF₄, C₂F₆, and C₃F₈ concentrations with the breaking times. The breaking current is about 10 kA. It can be found that the concentration of CF₄ is much higher than that of C₂F₆ and C₃F₈, and shows an obvious increasing trend with the breaking times, while C₂F₆ and C₃F₈ have no obvious change. After five breaking experiments, the concentration of CF₄ reached 289.6 ppm, while the concentration of C₂F₆ and C₃F₈ was only 2.1 ppm and 0.2 ppm, respectively. It can be concluded that CF₄ is the main by-product produced by the ablation of PTFE, and the concentrations of C₂F₆ and C₃F₈ are very low. The experimental results are in good agreement with the simulation results.

3.2. Particle Composition of SF₆ and SF₆/PTFE/O₂

Figure 6 and Figure 7 show the particle composition of SF₆/PTFE/O₂ arc plasma with different O₂ concentrations. Compared with SF₆/PTFE arc plasma, a large number of particles containing O element such as SOF₂, SO₂F₂, COF₂, SO₂, CO, and CO₂ appear in arc plasma. CO appears in the temperature range of 2000 K–5000 K. CO₂ and SO₂ appear in the temperature range of >1500 K. COF₂ appears in the temperature range of 1000 K–4500 K. As the arc temperature gradually decreases to room temperature, the molar fraction of these species will be reduced to <10⁻⁶. SOF₂ and SO₂F₂ can appear in arc plasma when the temperature is less than 1000 K, and their molar fraction is much higher than that of CO₂, SO₂ and COF₂. Therefore, it can be concluded that the participation of a small amount of O₂ will promote the formation of SOF₂, SO₂F₂, COF₂, SO₂, CO, and CO₂. After the arc is extinguished, the concentration of SOF₂ and SO₂F₂ will be higher than that of COF₂, SO₂, CO, and CO₂.

Figure 8 shows the changes of SOF₂, SO₂, SO₂F₂, and CO concentrations with the breaking times under different O₂ concentration. As shown in Figure 8a, the concentration of SOF₂ increases with the breaking times. When the O₂ concentrations are 59 ppm, 748 ppm and 1130 ppm, respectively, the concentrations of SOF₂ can reach 761 ppm, 983 ppm and 1121 ppm, respectively, after five breaking experiments. This indicates that the increase of O₂ concentration can promote the formation of SOF₂. As shown in Figure 8b, the concentration of SO₂ increases with the breaking times. The concentrations of SO₂ were 3.1 ppm, 9.0 ppm and 9.7 ppm, respectively, after five breaking experiments. As shown in Figure 8c, the concentration of SO₂F₂ increases with the breaking times. After five breaking experiments, the concentrations of SO₂F₂ are 0.4 ppm, 7.1 ppm and 9.8 ppm, respectively. This indicates that the increase of O₂ concentration can significantly promote the formation of SO₂F₂. Figure 8d shows the change of CO with the breaking times. It can be seen that the concentration of CO is very low, and the highest is only 1.0 ppm. Due to the limitation of experimental conditions and the accuracy of testing equipment, when O₂ are 736 ppm and 1202 ppm, CO has no significant regularity. When O₂ is 59 ppm, the concentration of CO is obviously lower than the former two cases. This indicates that the increase of O₂ concentration can also promote the formation of CO, but this phenomenon is not obvious.

In comparison, the concentration of SOF₂ is much higher than that of SO₂, SO₂F₂ and CO. After five breaking experiments, the concentration of SOF₂ can reach 1000 ppm, and the concentrations of SO₂ and SO₂F₂ are both about 10 ppm, while the concentration of CO is lower than 2 ppm. This phenomenon indicates that when O₂ is involved in the arc plasma reaction, SOF₂ will be the main by-product after the arc is extinguished, and the concentrations of SO₂ and SO₂F₂ will also increase significantly.

3.3. Particle Composition of SF₆ and SF₆/PTFE/H₂O

Figure 9 and Figure 10 show the particle composition of SF₆/PTFE/H₂O arc plasma with different H₂O concentrations. Compared with SF₆/PTFE/O₂ arc plasma, some particles containing H such as HF, H₂, CH, and H appear in arc plasma, and the molar fractions of SOF₂ and SO₂F₂ were significantly affected. H₂ appears in the temperature range of 3500 K–5000 K. H appears in the temperature range of 2500 K–5000 K. HF can appear in the whole temperature range and its molar fraction is always greater than 10⁻¹. SOF₂ can appear in the temperature range of 300 K–5000 K, and its molar fraction increases with the decrease of arc temperature. When the temperature decreases to room temperature, the molar fraction of SOF₂ can reach 10⁻¹. SO₂F₂ can appear in the temperature range of 300 K–3000 K, and its molar fraction relatively low. It can be concluded that the participation of a small amount of H₂O will promote the formation of HF, H₂, SOF₂, SO₂F₂, COF₂, SO₂, CO, and CO₂. After the arc is extinguished, the concentration of HF and SOF₂ will be significantly higher than that of SO₂F₂, COF₂, SO₂, CO, and CO₂.

Figure 11 shows the changes of SOF₂, SO₂, CO₂, and CO concentrations with the breaking times under different H₂O concentration. It can be found that the concentration of SOF₂, SO₂, CO₂ and CO increase with the breaking times. The increase of H₂O concentration can promote the production of SOF₂, SO₂, CO₂, and CO. For example, when the H₂O concentrations are 106 ppm, 748 ppm and 1130 ppm, respectively, the concentrations of SOF₂ are 1097 ppm, 1130 ppm and 1253 ppm, respectively, after five breaking experiments. This conclusion is consistent with the calculation results of particle composition of SF₆/PTFE/H₂O arc plasma.

In comparison, the concentration of SOF₂ is much higher than that of SO₂, CO₂ and CO. After five breaking experiments, the concentration of SOF₂ can reach 1000 ppm, while the concentration of SO₂ and CO₂ is lower than 20 ppm, and the concentration of CO is lower than 2 ppm. This phenomenon indicates that when H₂O is involved in the arc plasma reaction, SOF₂ will be the main by-product after the arc is extinguished. This conclusion also shows good agreement with the calculation result.

Contrary to the simulation results, SO₂F₂ and HF are hardly detected in this experiment. We believe the main reason is that H₂O does not promote the formation of SO₂F₂ obviously, and HF can react with the metal materials and solid insulation materials in the circuit breaker quickly, which makes it hardly detected. The particle composition model of arc plasma needs to be further improved.

It can be found by comparing the breaking experiments under different O₂ and H₂O concentrations that both H₂O and O₂ can promote the formation of SOF₂. Under similar experimental conditions, the promotion effect of H₂O on the formation of SOF₂ is more obvious than that of O₂. For example, when the concentration of H₂O is 748 ppm, the concentration of SOF₂ can reach 1130 ppm after five breaking experiments, with an average of 226 ppm each time. When the concentration of O₂ is 736 ppm, the concentration of SOF₂ is 983 ppm after five breaking experiments, with an average of 197 ppm each time. It also can be found that both H₂O and O₂ can promote the formation of SO₂ and CO, but the concentrations are relatively low. O₂ can significantly promote the formation of SO₂F₂, while H₂O has no obvious effect on SO₂F₂.

3.4. Particle Composition of SF₆ and SF₆/PTFE/O₂/H₂O

Figure 12 shows the particle composition of 60% SF₆ + 20% CF₂ + 10% H₂O + 10% O₂ arc plasma at 0.6 MPa. A total of 43 kinds of particles are considered in this model. The molar fractions of 33 kinds of particles are more than 10⁻⁶. The SF₆ by-products detected after the breaking experiment all appeared in this model, and the molar fractions of these species are relatively high. CO appears in the temperature range of 2000 K–5000 K, and the maximum molar fraction is 4.6%. CO₂ and SO₂ appear in the temperature range of 1250 K–5000 K, and the maximum molar fraction is 0.27% and 0.19%, respectively. CF₄ appears in the temperature range of 300 K–3500 K, and the maximum molar fraction is 20.0%. SOF₂ appears in the temperature range of 500 K–4500 K, and the maximum molar fraction is 13.8%. SO₂F₂ appears in the temperature range of 300 K–2750 K, and the maximum molar fraction is 15.0%. HF appears in the whole temperature range, and the maximum molar fraction is 20.2%. It should be noted that SOF₂, SO₂F₂, CF₄, HF, O₂, and SF₆ are the main particles when the arc temperature drops below 1000 K.

4. Conclusions

This paper established the arc plasma models of SF₆ (12 particles are considered), SF₆/PTFE (24 particles are considered), SF₆/PTFE/O₂ (34 particles are considered), SF₆/PTFE/H₂O (43 particles are considered), and SF₆/PTFE/O₂/H₂O (43 particles are considered) by using the Gibbs free energy minimization method. The effects of trace H₂O and O₂ impurities and PTFE vapour on the SF₆ by-products were studied. In order to verify the correctness of the simulation results, a series of breaking experiments were carried out on a 40.5 kV SF₆ circuit breaker, and a gas chromatograph was used to detect and analyse the SF₆ by-products. It was found that the experimental results are in good agreement with the simulation results. The primary conclusions are summarized below:

(1)

SF₆ molecules began to decompose at about 1000 K and form SF₅, SF₄, SF₃, SF₂, SF, S, and F with the increase of arc temperature. The maximum molar fractions of SF₄ and SF₂ were higher than SF₅, SF₃, and SF. As the arc temperature gradually decreases to room temperature, all low fluorine sulphides recombine into SF₆.

(2)

When PTFE vapour was involved in the arc plasma, the main by-product after arc quenching was CF₄, and the molar fractions of C₂F₆ and C₃F₈ were very low. After five breaking experiments, the concentration of CF₄ can reach 289.6 ppm, while the concentration of C₂F₆ and C₃F₈ was only 2.1 ppm and 0.2 ppm, respectively. The simulation results were in good agreement with the experimental results.

(3)

When O₂ was involved in the arc plasma, the main by-products were SOF₂, SO₂, and SO₂F₂. At the same time, a small amount of CO, and CO₂ was produced. After five breaking experiments, the concentration of SOF₂ can reach 1100 ppm, and the concentrations of SO₂ and SO₂F₂ were both about 10 ppm, while the concentration of CO was lower than 2 ppm.

(4)

When H₂O was involved in the arc plasma, the main by-products were SOF₂, SO₂, SO₂F₂, and HF. At the same time, a small amount of CO₂, CO, and H₂ was produced. After five breaking experiments, the concentration of SOF₂ can reach 1200 ppm, and the concentration of SO₂ and CO₂ was lower than 20 ppm, while the concentration of CO was lower than 2 ppm. Contrary to the simulation results, SO₂F₂ and HF were hardly detected in this experiment.

(5)

When H₂O and O₂ impurities and PTFE vapour were involved in the arc plasma together, the main by-products were SOF₂, CF₄, SO₂F₂, and HF. At the same time, a small amount of SO₂, CO₂, CO, and H₂ was be produced.