Interaction Mechanism between the C4F7N–CO2 Gas Mixture and the EPDM Seal Ring

C4F7N (fluorinated nitrile) has been introduced as a remarkable substitute gas for the greenhouse gas SF6 (sulfur hexafluoride) which is used in gas-insulated equipment (GIE). Intensive investigations about the compatibility between C4F7N and materials used in GIE are required to decide their long-term behavior. In this paper, the interaction mechanism between EPDM, used as a sealing ring in GIE, and C4F7N–CO2 was explored. The composition and morphology properties of EPDM were first revealed based on scanning electron microscopy and X-ray photoelectron spectroscopy. It was found that EPDM rubber is incompatible with the C4F7N–CO2 gas mixture at temperatures higher than 70 °C. There exist chemical reactions between EPDM and C4F7N, resulting in the generation of gaseous byproducts including C3F6, CF3H, and C2F5H and corrosion of EPDM. DFT calculation also shows that the interaction between C4F7N and EPDM could cause the dissociation of C4F7N. Relevant results provide important guidance for the engineering application of the C4F7N gas mixture.


INTRODUCTION
Sulfur hexafluoride (SF 6 ) has been widely used as gas insulating medium in all kinds of gas-insulated equipment (GIE) since the 1970s because of its excellent insulation and arc-quenching properties. 1−4 However, SF 6 is one of the greenhouse gases with a global warming potential (GWP) 23 500 times higher than that of CO 2 (over a time horizon of 100 years) and an atmospheric lifespan of 3200 years. 5−7 The power industry accounts for 80% of the worldwide sale of SF 6 . 8 Therefore, seeking an eco-friendly gas insulating medium to replace SF 6 used in GIE has become a hot spot in recent years.
At present, C 4 F 7 N ((2,3,3,3-tetrafluoro-2-(trifluoromethyl)-2-propanenitrile, fluorinated nitrile) has been introduced as the remarkable substitute gas to SF 6 . 9,10 C 4 F 7 N has a GWP value of 2090, a dielectric strength twice that of SF 6 , and a boiling point of −4.7°C. 10 It needs to be mixed with CO 2 , N 2 , or air for engineering application because of its high liquefaction temperature. The GWP of the C 4 F 7 N gas mixture with 4, 6, and 10% C 4 F 7 N is 327, 462, and 690, respectively. In addition, the LC50 of 10% C 4 F 7 N gas mixture is in the range of 95 500 to 100 000 ppm (parts per million, μL/L), which is classified as a nontoxic substance according to the CLP regulation 1272/ 2008. 10 Several studies have been conducted on the performance evaluation of C 4 F 7 N including insulation properties, 11−14 decomposition characteristics, 15−20 arc extinguishing performance, 21 and toxicity 22,23 over the past three years, which confirm that C 4 F 7 N can be used as a substitute to SF 6 for medium-voltage and high-voltage applications.
It has been proven that GIE including gas-insulated switchgear (GIS), gas-insulated bus, and gas-insulated current transformer filled with C 4 F 7 N gas mixture could pass relevant electrical standard tests. 9 GIE is designed for a lifetime of 30 years or more. During this period, the effort for maintenance should be kept on a low level. Thus, intensive investigations about the compatibility between C 4 F 7 N and the materials used in GIE are required to decide about a long-term behavior. The compartments of GIE include metal (copper, aluminum, and steel), polymers (epoxy resin, rubber, and thermoplastic materials), oils, and desiccants, which are all (at least partly) in contact with the insulation gas. There might be aging processes caused by the combination of the used materials and the insulation gas mixture. The interaction might lead to changed properties of the material itself, but can also influence the quality of the insulation gas. Both effects must be investigated for all used materials. Studies on the decomposition properties of C 4 F 7 N−CO 2 or C 4 F 7 N−N 2 gas mixtures under AC breakdown, thermal and arc-quenching conditions indicate that CF 4 , C 2 F 6 , C 3 F 6 , C 3 F 8 , CF 3 CN, C 2 F 5 CN, C 2 N 2 , COF 2 , CO, and HF are the main generated byproducts. 16−19 The generation of C 3 F 6 and CO is common for thermal decomposition conditions and CF 4 and CF 3 CN are the main products after breakdown. 18,19 The decomposition mechanism of C 4 F 7 N was also explored by several researchers using density functional theory (DFT) and reactive force field (ReaxFF). 18,20 Our group also explored the compatibility of C 4 F 7 N−N 2 gas mixture with copper and aluminum. It was found that C 4 F 7 N−N 2 gas mixture has better compatibility with heated aluminum (at 120−220°C) than that of copper. The decomposition of C 4 F 7 N and generation of C 3 F 6 , CF 3 H was confirmed. 15 In addition, it was reported that leakage of SF 6 because of the sealing ring failure accounts for 30−40% GIE defects. 24 Thus, the compatibility of the sealing ring and gas is an important factor to determine the reliability of equipment. At present, the seal materials mostly used in GIE are ethylene propylene diene monomer (EPDM) and nitrile butadiene rubber. In addition, there are few studies on the compatibility between EPDM and C 4 F 7 N gas mixture.
In this paper, we conducted aging tests for EPDM using as a sealing ring in GIE under the C 4 F 7 N−CO 2 environment first. The gas components of C 4 F 7 N−CO 2 gas mixture and the morphology of EPDM was revealed based on the gas chromatography−mass spectrometry (GC−MS), scanning electron microscope (SEM) and X-ray photoelectron spectroscopy (XPS). Then, the interaction mechanism between EPDM and C 4 F 7 N was analyzed based on the DFT. The possible reaction pathways were considered and the reaction enthalpy was also calculated. Relevant results not only reveal the compatibility between EPDM and C 4 F 7 N gas mixture first but also provide reference for the engineering application of the C 4 F 7 N gas mixture.

RESULTS AND DISCUSSION
2.1. Compatibility between C 4 F 7 N−CO 2 and EPDM. 2.1.1. Gas Components Analysis. The EPDM samples were put in the holder in the test chamber. Then, the chamber was filled with 0.3 MPa C 4 F 7 N−CO 2 gas mixture and heated at 70 and 80°C for 90 h. A control group was also set without the EPDM sample to exclude other possible factors on the results. The gas mixture was detected by the GC−MS, and the EPDM was analyzed by SEM and XPS at the end of test. Detailed information on the test process can be found in the Method section.
According to the gas chromatogram of the 10%C 4 F 7 N−90% CO 2 gas mixture after aging tests, as shown in Figure 1, the characteristic peak of CF 3 H exists for both the control group (80°C) and the test group (70, 80°C). As for the aging test at 80°C, another gaseous byproduct C 3 F 6 can be found. In addition, the generation of C 2 F 5 H is also confirmed considering the existence of the characteristic mass charge ratio (m/z) of 101 (C 2 F 4 H group). Figure 2 shows the content of detected byproducts under different conditions. The interaction between EPDM and C 4 F 7 N−CO 2 gas mixture generated 0.95 ppm (control group, 80°C), 0.33 ppm (test group, 70°C), and 13.01 ppm (test group, 80°C) C 3 F 6 , respectively. It should be noted that the initial decomposition temperature of the C 4 F 7 N−CO 2 gas mixture is higher than 350 or 650°C. 10,15 Therefore, there exists chemical reaction between EPDM and C 4 F 7 N at 80°C. The peak area integral calculation shows that the content of CF 3 H and C 2 F 5 H increased sharply when the gas temperature reached 80°C, indicating that the temperature has quite the effect on the reaction between the C 4 F 7 N−CO 2 gas mixture and EPDM. In addition, the content of C 3 F 6 , CF 3 H at 70°C is lower than the control group and the difference is relatively small. This is due to the test temperature of the control group being set to 80°C and the detection limit of GC−MS at low content.
Overall, the gas composition analysis confirms that the chemical reaction between EPDM and C 4 F 7 N−CO 2 gas mixture occurs at temperatures around 80°C, resulting in the generation of characteristic byproducts C 3 F 6 , CF 3 H, and C 2 F 5 H. The duration time will also have influence on the test results because of the cumulative effect considering the long term operation of the equipment.
2.1.2. Morphology Analysis. Figure 3 shows the morphology of the EPDM rubber before and after interaction. It is apparent that the surface morphology of the untreated EPDM sample was homogeneous and smooth, and no obvious defects can be found at the magnification of 5000× and 10 000×. The appearance of white spots exists (see the red circle in Figure  3b) for the EPDM exposed to C 4 F 7 N−CO 2 gas mixture at 70°C for 90 h, indicating that C 4 F 7 N causes corrosion of EPDM. When the gas temperature reaches 80°C, there exist white spots randomly distributed on the EPDM surface (as shown in Figure 3c). The EDPM surface is corroded seriously at this condition because strong chemical reaction occurs between them.
It should be noted that the maximum operating temperature of electrical equipment is 40°C according to IEC 61869-1-2007. Considering the temperature-rise effect during the normal operation of GIE, the temperature of the sealing material will be in the range of 40−70°C. Therefore, the service life of EPDM for GIE using the C 4 F 7 N gas mixture as the insulating medium reduces.
2.1.3. XPS Analysis. Table 1 shows the XPS survey analysis results. It can be found that the composition of untreated and treated EPDM at 70°C is similar, which is quite different to the treated EPDM at 80°C. In particular, the composition of the F element increased from 1.52% (70°C) to 4.67% (80°C ), confirming that the fluorine accumulation process happens during the interaction. The atomic composition of N also increased by 0.97% as the temperature changes from 70 to 80°C. The content of C and O element has a little difference for the untreated EPDM and treated EPDM at 70°C . Figure 4 gives the high-resolution XPS spectra of the EPDM rubber treated with 10%C 4 F 7 N−90%CO 2 gas mixture. All the peaks were fitted based on XPS Peak software employing Gauss−Lorentz curves after subtraction of a Shirleytype background. The carbon calibrations were also conducted. As shown in Figure 4a, the characteristic peak located at 284.8 eV belongs to C−H or C−C bonds. In addition, the other two  25,26 As for the O 1s spectra shown in Figure 4b, the peaks located at 532.3 and 533.3 eV correspond to the CO and C−O−C component, respectively. 27 Overall, the composition of C and O for the untreated and treated EPDM is similar. In addition, the high-resolution spectra of F 1s includes two characteristic peaks located at 684.5 and 688.3 eV. The binding energy of F 1s close to 684.5 eV belongs to the F that is ironically fixed to the C atom. 28,29 This characteristic peak is not found for the untreated EPDM (quite low peak intensity). Moreover, the peak area of the C−F component increased obviously with temperature, indicating that the accumulation of fluorine occurs severely because of the reaction between EPDM and C 4 F 7 N−CO 2 . The other peak located at 688.3 eV is assigned to the CF x CH x group, corresponding to organic fluorine components. It is reported that there exists a significant shift of about 4 eV between the organic and inorganic fluoride components. 30 Thus, the two peaks of F 1s confirms that the generation of the inorganic fluoride component occurs after the aging tests.
Overall, XPS results indicates that EPDM rubber undergoes accumulation of fluorine when interacted with C 4 F 7 N−CO 2 gas mixture. This process is accelerated with the increase of temperature, which is consistent well with the GC−MS and SEM test results.
2.2. Interaction Mechanism between C 4 F 7 N and EPDM. 2.2.1. Structure Properties and Decomposition of EPDM. In order to understand the interaction mechanism between EPDM and C 4 F 7 N, we performed DFT calculations to reveal the structure properties of the interaction system.
The typical chemical structure of EPDM consists ethylene, propylene, and diene units. Ethylidene norbornene (ENB) is usually used as diene, as shown in Figure 5. 31 The ethylene− propylene−ENB chain was chosen as the EDPM polymer model for DFT calculations. This scheme is also used by    32 Considering the complex molecular structure of C 4 F 7 N and EPDM, the interaction between them discussed in this paper is mainly concentrated on the reaction of EPDM defects or particles and C 4 F 7 N. The EPDM defects and particles have unsaturated chemical bonds, which have stronger reactivity than the EPDM molecule. That is to say, the reaction between C 4 F 7 N and EPDM defects or particles occurs easier than the EPDM molecule. In this section, we first analyzed the structural properties of the EPDM monomer. Then, the possible defect formation pathways of EPDM were analyzed. In addition, the possible interaction between the C 4 F 7 N and several kinds of EPDM defects was explored. Figure 6 shows the calculated Mayer bond order (MBO) of EDPM. MBO is widely used to describe the relative strength of chemical bonds. 33 It can be found that the CC bond has the largest MBO value (1.763) and the MBO of the C−H bond in EDPM is smaller than most of the C−C bond. Moreover, four main dissociation paths were analyzed to reveal the formation of EDPM defects (as shown in Figure 7). Figure 8 demonstrates the calculated reaction enthalpy of considered EPDM dissociation paths. The dissociation of EPDM through path A requires adsorption of 85.37 kcal/mol. Moreover, the bond-breaking process of EPDM through path B to generate C 9 H 13 and C 5 H 11 has the lowest reaction enthalpy among all the considered paths (83.76 kcal/mol),    Reaction Mechanism between C 4 F 7 N and EPDM. The degradation of EPDM occurs during the aging process, which could generate several kinds of particles or defects on its surface. Moreover, the decomposition of C 4 F 7 N starts at temperatures higher than 350°C. 10, 19 Thus, we mainly considered the reaction between C 4 F 7 N and EPDM decomposed particles or defects to investigate the reaction mechanism between them. Table 2 gives the reaction paths and enthalpy between EPDM-decomposed particles through path A and C 4 F 7 N. All the proposed reaction paths have a negative reaction enthalpy, indicating that the reactions are thermodynamically supported. For the situation that the N atom in C 4 F 7 N close to the C 11 H 17 and C 3 H 7 (path A1 and A3), optimized structures show that the distance between N and C atoms is shortened. The distance of C atoms in C 11 H 17 and N atoms in C 4 F 7 N changes to 1.473 Å after interaction. The bond angle of the CN group which connected to the central C atoms of C 4 F 7 N also changes to 140.55°. As for reaction paths A2 and A4, the F atom linked to the central C atom of C 4 F 7 N is composed of a new chemical bond with the C atom in C 11 H 17 and C 3 H 7 , and another product C 4 F 6 N is produced after interaction. The reaction enthalpy of path A2 (−29.39 kcal/mol) and A4 (−23.44 kcal/ mol) is also lower than that of A1 (−16.19 kcal/mol) and A3 (−15.84 kcal/mol), which are more likely to occur. The accumulation of fluorine might occur through these two reaction paths.
The considered reaction paths between EPDM decomposed particles through path B and C 4 F 7 N is given in Table 3. The reaction enthalpy of all the paths is also negative. Similar to paths A1 and A3, the N atom in C 4 F 7 N tends to form new chemical bonds with EPDM defect particles. The distance between the N atom and C atom in EPDM is shortened. As for reaction paths B2 and B4, the C−F bond in C 4 F 7 N breaks and a new bond between F and C in the EPDM is produced after interaction. These two paths are more likely to occur because of the lower enthalpy, which is another possible fluorine accumulation paths. Table 4 gives the interaction paths between EPDM decomposed particles through path C, D, and C 4 F 7 N. We can find that paths C2 and C1 have the lowest reaction enthalpies of −46.31 and −40.97 kcal/mol, respectively. Thus, the above two reactions are more likely to occur. In addition, the reaction mechanism for the N atom or F atom in C 4 F 7 N closed to the EPDM defect particles is similar to that of Tables  2 and Table 3.
According to the above results, the interaction between C 4 F 7 N and EPDM defects and particles will result in the adsorption and dissociation of C 4 F 7 N. The N atom in the CN

ACS Omega
http://pubs.acs.org/journal/acsodf Article group of C 4 F 7 N has the tendency to form new bonds with EPDM defects or particles. The F atom connected to the central C atom of C 4 F 7 N could react with the EPDM particles, causing the dissociation of C 4 F 7 N to generate C 4 F 6 N as well as forming a new F−C bond with EPDM. These reaction paths could explain the accumulation of fluorine on the EPDM after aging. 2.2.3. Discussion. According to the above-mentioned results, EPDM rubber is incompatible with C 4 F 7 N−CO 2 gas mixture at temperatures higher than 70°C. The interaction between them will result in the decomposition of C 4 F 7 N, generating CF 3 H, C 3 F 6 , and C 2 F 5 H. The corrosion of the EPDM surface occurs and the accumulation of fluorine exists.
At present, there is no general standard defined on how compatibility for materials in contact with C 4 F 7 N gas mixture has to be tested. The test procedure must be defined by the manufacturer in a way to confirm that the impact of the material on the gas and the impact of the gas and possible decomposition products on the material does not affect the safety and performance of equipment in an unacceptable level. 34 In addition, Kieffel pointed out that gas permeability of the elastomer material used for gaskets should be considered because of the CO 2 molecule is smaller than C 4 F 7 N or SF 6 . The designed O-ring for GIS should be tight with enough reliability during all its lifetime and meet the maximum allowed leakage rate of 0.5% per compartment per year as specified in IEC 62271-203. 10 Butyl rubber is known to be tight toward low-molecular weight gases such as CO 2 and it has been widely used in the automotive industry for tires. 10 Halogenated butyl rubber which has no double bond on the main molecular chain and is not sensitive to ozone could be a candidate for C 4 F 7 N−CO 2 gas mixture equipment. It was reported that the permeation rate coefficient of halogenated butyl rubber under C 4 F 7 N−a CO 2 environment at 20°C is lower than that of EPDM. 10 While its stability under working temperature conditions needs to be further tested.

CONCLUSIONS
In this paper, we explored the compatibility between EPDM and C 4 F 7 N. The composition and morphology properties of EPDM and C 4 F 7 N−CO 2 gas mixture after aging tests were obtained based on the GC−MS, SEM, and XPS. The interaction mechanism between EPDM and C 4 F 7 N is also revealed based on the DFT calculations. The following useful conclusions can be obtained, (1) EPDM exposed under a C 4 F 7 N−CO 2 environment at temperatures around 80°C will result in the decomposition of C 4 F 7 N, producing several characteristic byproducts including C 3 F 6 , CF 3 H, and C 2 F 5 H. The EDPM surface will be corroded seriously at temperatures around 80°C. The fluorine accumulation process also exists during the interaction. The aging tests were conducted to explore the compatibility between EPDM and C 4 F 7 N. The test platform mainly consists of the heating system, the temperature control system and the gas chamber (as shown in Figure 9). The gas chamber with a volume 2 L is made of 304 L stainless-steel, which could withstand a high pressure of 0.6 MPa. The stainless-steel sample holder is placed at the bottom of the gas chamber to carry the EDPM rubber sample. The gas chamber is sealed using the fluororubber, which generally has wide chemical resistance. All the test chambers were sealed tightly, and no gas leakage was found during the test. The heating element is put in the center of the heat transfer bushing, which is installed on the top cover plate of the gas chamber. The temperature control system consists of the temperature sensor, the switch power supply and the solid-state relay. The K-type temperature sensor is used to monitor the temperature of the C 4 F 7 N−CO 2 gas mixture. The gas sensor sends signals to the proportion integration differentiation controller. The controller could compare the actual temperature with the set value and send a control signal to the solid-state relay to realize the switching power supply control.
4.1.2. Materials and Test Conditions. We used the C 4 F 7 N− CO 2 gas mixture to carry out thermal aging tests. Relevant studies reported that CO 2 has a great synergism effect with C 4 F 7 N, and CO 2 also shows better arc quenching capabilities compared with N 2 . 8 It was pointed out that the concentration of C 4 F 7 N in the gas mixture should be less than 10% for the working pressure of 0.5 MPa. Thus, we used 10%C 4 F 7 N−90% CO 2 gas mixture to conduct relevant tests. C 4 F 7 N is supplied by 3 M China with a purity higher than 99.2%. In addition, the standard gas for 10%C 4 F 7 N−90%CO 2 is provided by Wuhan Newradar Special Gas Co., Ltd. The EPDM rubber O-rings is supplied by Xu Ji Group Corporation, State Grid.
The EPDM O-ring we obtained is cylindrical as shown in Figure 10. The sealing ring was cut into the semi-cylindrical structure for the convenience of relevant SEM and XPS testing. The semi-cylindrical O-ring was put on the sample holder in the chamber. The gas chamber was evacuated using the vacuum pump to 10 −4 to 10 −5 Pa first and then filled with 10% C 4 F 7 N−90%CO 2 to 0.3 MPa. This procedure is repeated 3 times to remove gas impurities. As for the control group, EPDM was not put in the test chamber. The gas pressure of all the test groups is set to 0.3 MPa. The using of 0.3 MPa is for the safety consideration because the heating of the gas mixture will result in the gas pressure increase. In addition, it was  15 While the temperature of the equipment shell and O-rings is usually around 40−70°C (absolute temperature). Thus, the test temperature condition in this paper was set to 70 and 80°C. The duration time also has influence on the test results because of the cumulative effect. Here, the exposure time is set to 90 h to explore the interaction between EPDM and C 4 F 7 N−CO 2 gas mixture. There are currently no compatibility testing standards for the non-SF 6 gas insulating medium with materials. The select of 90 h is based on the time cost. 4.1.3. Morphology and Component Analysis. The gas mixture after aging tests were collected and analyzed using GC−MS (Shimadzu QP2010 Ultra). The column type is CP-Sil5CB (60 m × 8 μm × 0.32 mm). Both the SCAN and SIM method were used to detect the gas components. The heating scheme of the GC is listed as follows: (1) Keep the column at 32°C for 10 min. (2) Heat the column to150°C at the rate of 60°C/min and retain it for 2 min. The gas component is confirmed based on the NIST (National Institute of Standards and Technology) standard database and standard gases.
The morphology of the EPDM was analyzed using a Zeiss SIGMA field-emission scanning electron microscope manufactured by Carl Zeiss. Moreover, the components of the EPDM surface were detected by XPS (ESCALAB250Xi electron spectrometer manufactured by Thermo Fisher Scientific of United States).
4.2. Theoretical Methods. In order to investigate the interaction mechanism between C 4 F 7 N and EPDM, we carried out spin-polarized DFT calculations based on the Dmol 3 The generalized gradient approximation with the Perdew− Burke−Ernzerhof functional [generalized gradient approximation (GGA)-PBE] method was applied to treat the electron exchange and correlation. 37 We chose the double numerical plus polarization (DNP) as the atomic basis set and all electron core treatment was selected. 38 The convergence tolerance for geometry optimization calculation was listed as follows: (1) 1.0 × 10 −6 Ha for energy, (2) 0.005 Å for displacement, and (3) 0.002 Ha/Å for gradients. Frequency analysis was also conducted for all the reactants and products. The reaction enthalpy is defined as follows The negative value of reaction enthalpy indicates that the reaction is thermodynamic spontaneous and the positive value of the reaction enthalpy means the reaction belongs to endothermic process. Zero-point vibration energy and enthalpy correction (at 298.15 K) were both considered for reaction enthalpy calculations. It should be noted that the numerical basis sets implemented in Dmol 3 code are more complete than the traditional Gaussian functions, thereby minimizing or even eliminating basis set superposition error. 39