The decomposition mechanism of C4F7N-Cu gas mixtures

C4F7N is one of the most remarkable replacements for SF6, and its decomposition mechanism has a great influence on insulating performance and environmental properties. It is noteworthy that discharges or high temperature also evaporates metal electrodes (e.g., Cu) in the equipment, and the generated metal gases interact with C4F7N and thus affect the C4F7N decomposition mechanism, but the decomposition mechanism is still not clear. In this paper, therefore, the B3LYP method in conjunction with 6-311G(d, p) basis set (for C, N, and F atoms) and Lanl2DZ basis set (for Cu atom) combining transition state theory is used to study the decomposition mechanism of C4F7N-Cu mixtures. 31 reactions are determined in decomposition pathways of C4F7N-Cu mixtures, and their potential energy surface as well as reaction mechanisms are obtained. The rate constants over 300 K–3500 K relevant to the insulation breakdown temperature are calculated based on the above calculations, and dominant reactions in different temperature regions are selected. The results show that (R14) C4F7N + Cu → CF3CFCN + CF2CuF plays a major role in the decomposition of C4F7N-Cu mixtures below 1500 K, while (R1) C4F7N + Cu → TSa1 → CuCN + C3F7 and (R21) C4F7N + Cu → TSc1 → CF3CF(CN)CF2 + CuF are dominant above 1500 K; (R23) CF3CF(CN)CF2 + Cu → CF2(Cu)C(F)CN + CF3 is the most important reaction leading to the generation of CF3 below 1500 K with the overwhelming rate constant, but other reactions also generating CF3 are dominant above 1500 K.


I. INTRODUCTION
SF 6 is widely used as an insulating and arc-quenching medium in power equipment, but its global warming potential (GWP) is 23 900 times as much as CO 2 and its decomposition products (e.g., SO 2 F 2 , SO 2 , and SOF 2 ) under discharges or high temperature are poisonous. Therefore, great efforts have been taken to explore environmental-friendly alternative gases for SF 6 in consideration of environment and safety issues. [1][2][3][4][5] As one of the most promising candidates for SF 6 gas, C 4 F7N with excellent insulation properties and low GWP (i.e., 1/6 of SF 6 ) has attracted widespread attention. 1 In practice, a small proportion of C 4 F7N is usually mixed with buffer gases to decrease the condensing point, but its insulating capacity is still comparable to SF 6 . The decomposition products of C 4 F7N and its mixtures also meet safety requirements. In recent years, environmental-friendly power equipment filled with C 4 F7N have been developed with excellent insulating behavior.
To the best of our knowledge, extensive research has been carried out on the insulation properties and decomposition characteristics of C 4 F7N. Taking into account detachment from negative ions, Chen 2 calculated a corrected value of the density-reduced critical electric field strength (E/N)(crit) * = 785 ± 15 Td, which is verified by breakdown voltage measurements in homogeneous electric fields in C 4 F7N at 5-65 kPa. Chen 3 investigated the electron rate, transport coefficients, and the density-reduced critical electric field of C 4 F7N and its mixtures with N 2 and CO 2 . A synergy effect is observed in the mixtures, and ion kinetics plays a major role in C 4 F7N discharges. Li 4 calculated thermophysical properties, including composition, thermodynamic properties, transport coefficients, and net emission coefficients, of thermal plasmas formed from pure C 4 F7N and C 4 F7N-CO 2 mixtures to investigate the thermal breakdown properties between 300 and 30 000 K. Li 5 found that the breakdown voltage of C 4 F7N-N 2 -O 2 gas mixtures increases with the O 2 content (0%-8%). The addition of O 2 could effectively suppress the precipitation of the solid by-product on the electrode surface and reduce the discharge dispersity. Li 6 studied the dielectric properties of C 4 F7N-N 2 mixtures under nonuniform electric field and found that an increasing gas pressure or a mixing ratio can improve the insulation performance of the C 4 F7N/N 2 gas mixtures, making its negative partial discharge inception voltage reach 68.8% of pure SF 6 .
It should be mentioned that C 4 F7N decomposes into molecules with low carbon and low fluorine under discharges or high temperature and a small fraction of them could not recombine to form C 4 F7N and thus remains in the equipment. During the longterm operation of the equipment, the content of C 4 F7N decreases while the decomposition products with low insulation capacity and unknown-toxicity accumulate to threaten the safety operation of the equipment. Besides, the rate constants of C 4 F7N decomposition also affect the varying characteristics and the recombination properties of C 4 F7N decomposition components. Therefore, the decomposition mechanism of C 4 F7N (i.e., decomposition pathways and rate constants) has a great influence on the insulating and environmental performance. Current studies on the decomposition mechanism of C 4 F7N are carried out with the quantum chemistry method, which is an effectively widely used method in the investigation of the decomposition mechanism of SF 6 and its replacements. Hoesl 7 studied ion kinetics influencing C 4 F7N discharges with a pulsed Townsend experiment and the results show the evidence of long-lived anion, non-detaching anion and a short-lived anion. Chachereau 8 employed total internal reflection (TIR)/smog chamber experiments and ab initio quantum calculations to investigate the atmospheric chemistry of C 4 F7N and found that the sole atmospheric degradation products of (CF 3 )(2)CFCN appear to be NO, COF 2 , and CF 3 C(O)F. Li 5 found that decomposition of C 4 F7N-N 2 gas mixtures after breakdown mainly produces CF 4 , C 2 F 6 , C 3 F 8 , CF 3 CN, C 2 F 4 , C 2 F5CN, (CN) 2 , and C 3 F 6 . Another by-product COF 2 is found, and the amount of CF 4 is detected after the addition of O 2 . The self-recovery characteristics of the gas mixtures is promoted first and then reduced as the oxygen content increases. Yu 9 studied the thermal decomposition mechanism of C 4 F7N and pointed out that the decomposition species of i-C 4 and C 2 F 6 are characteristics to reflect the insulation properties of the equipment. Our previous work 10 studied the decomposition mechanism of C 4 F7N, and the results indicated that 16 reactions are included in its decomposition pathway and the dominant reactions and species are selected by analysis of the rate constant. Li 11 reported the rate coefficient for the OH-C 4 F7N reaction and determined the radiative efficiency with infrared absorption cross sections. Li 12 employed density functional theory (DFT) calculation, ReaxFF molecular dynamics simulation, and gas chromatographymass spectrometer to study the decomposition components of C 4 F7N-N 2 mixtures and found that C 2 F 6 , CF 4 , and CF 3 CN are main decomposition products with a relatively high content and N 2 has a buffering effect to avoid the massive decomposition of C 4 F7N, so C 4 F7N-N 2 mixtures have great self-recovery performance. It is noteworthy that discharges or high temperature also evaporates metal electrodes (e.g., Cu) in the equipment, and the generated metal gases interact with C 4 F7N and thus affect the C 4 F7N decomposition mechanism and the insulation properties of the equipment. However, the decomposition mechanism of C 4 F7N-Cu mixtures is still not clear. Therefore, this paper is devoted to investigating the decomposition mechanism of C 4 F7N-Cu mixtures. The decomposition pathways as constructed by 31 reactions and the potential energy surface (PES) are studied by density functional theory (DFT). The reactants, products, and transition states (TSs) in a considered reaction are optimized, and their energetic profiles as well as vibrational frequencies are also calculated with the same method. Using the above data, the rate constants are computed with transition state theory (TST) and the dominant reactions in C 4 F7N-Cu decomposition are selected based on the rate constant. This work is hopeful to lay a theoretical basis to evaluate the insulation behavior of the C 4 F7N-insulated power equipment and explore novel gas sensors. [13][14][15][16][17]

II. CALCULATION METHOD
As introduced in our previous work, 10,18-21 quantum chemistry calculations on geometric structures, harmonic frequencies, and energetic information of reactants, products, and TS in each reaction are carried out as the prerequisite for C 4 F7N-Cu mixtures decomposition analysis using the DFT/B3LYP method, [22][23][24][25] in conjunction with the 6-311G(d, p) basis set (for C, N, and F atoms) 18,19,26 and Lanl2DZ basis set (for Cu atom) 27,28 as implemented in the Gaussian09 packages. 29 Products, reactants, and intermediate products (i.e., particular configurations to reflect the reaction process of reactants along the reaction coordinate, IM) are characterized as stationary points on PES, shown by their real harmonic vibrational frequencies. 30,31 TS is a particular configuration with only one imaginary vibrational frequency, referring to the saddle point on the reaction coordinate, 32 and it is confirmed by intrinsic reaction coordinate (IRC) calculations 33 showing each TS connecting reactants and products. For those reactions without a TS, relax-scan calculations are carried out to determine PES by optimizing the molecular geometry with a selected single constant bond. 34 The scaling factors of the B3LYP/6-311G(d, p) basis set for frequencies and energies are 0.97 and 0.99, respectively. 35 Using above results, conventional transition state theory (CTST) 36 is used to calculate the rate constant of reactions with a TS and variational canonical transition state theory (VTST) is adopted to compute that of barrierless reactions. [37][38][39][40][41][42]

A. Decomposition pathways of C 4 F 7 N-Cu mixtures
Ionized components in the C 4 F7N + Cu mixture arc are important to evaluate the insulation behavior, and the arc plasma is usually studied with thermal equilibrium assumptions other than a chemical kinetic model. The reaction rate constants can be studied by collision cross section parameters. For low temperature and nonequilibrium processes such as PD and decaying-arc, neutral particles are dominant with an overwhelming composition 26,43 and chemical kinetic models are usually adopted. Therefore, the possible reaction positions of C 4 F7N-Cu mixtures tested and the detailed decomposition pathways consisting of 31 reactions are finally obtained and are shown in Fig. 1 to make this article easy to read. These decomposition reactions are also listed in Table I. Basic atoms (C, N, Cu, and F) and charged species will be theoretically generated after multiple chemical reactions, but the overall rate constants are usually very low and thus their contents are much less than other decomposition products under the rapid decreasing rate of temperature in electric breakdown. Therefore, dissociations into basic atoms and charged species are not considered, while the decompositions to simple products (e.g., CF 3 , CuF, CN, and CF) are fully investigated in this paper.
The optimized molecular structures of reactants, products, and TS in C 4 F7N-Cu decomposition pathways are shown in Fig. 2. The B3LYP method in conjunction with the 6-311G(d, p) basis set (for C, N, and F atoms) and Lanl2DZ basis set (for Cu atom) is verified in our previous work 10,18,21 that the calculated structural information and harmonic vibrational frequencies of the C/F/N/Cu chemical system are consistent well with experimental data from CCCBDB

No.
Reaction formula   Table II.
The products of C 3 F7 + CuCN and CuF + CF 3 CF(CN)CF 2 can be obtained via TS a1 and TS c1 with the obvious barrier heights and 11.276 kcal/mol and 33.107 kcal/mol, respectively. In reaction R1, a C(11)− −C(1) stretching vibration in TS a1 connects reactants of C 4 F7N + Cu and products of C 3 F7 + CuCN. In reaction R21, atom F(6) emigrates from atom C(2) to atom Cu(13) and the Cu(13)− −C(11) bond stretches at the same time in TS c1 to generate products of CuF + CF 3 CF(CN)CF 2 . Due to the different reaction positions and structures of TS a1 and TS c1 , the barrier height of reaction R1 is 21.831 kcal/mol lower than that of reaction R21, causing reaction R1 more likely to occur. In reaction R14, an intermediate product with atom F(5) in C 4 F7N replaced by radical CuF is first generated and then bond C(1)− −C(2) breaks to generate products of CF 3 CFCN + CF 2 CuF absorbing an energy of 20.047 kcal/mol studied by relaxed-scan calculation. The complex dissociation products of C 3 F7, CF 3 CFCN, and CF 3 CF(CN)CF 2 will continue to decompose and are discussed in Secs. III B 2, III B 3 and III B 4.
As another decomposition product of C 3 F7 + Cu generated in reaction R3, CF 3 CFCF 2 reacts with Cu through reaction (R10) CF 3 CFCF 2 + Cu → TS a4 → CF 3 + CF 2 CFCu. The vibration mode of TS a4 consisted of a stretching between bond C(1)− −C(5), and the barrier height of reaction R10 is 36.663 kcal/mol. As CF 2 CFCu consisted of bond C= =C which makes CF 2 CFCu hard to dissociate, its decomposition mechanism is not discussed in this paper.
In reaction R15, the structural changes in TS b1 show that this reaction consisted of the emigration of atom Cu(9) from N(8) to C(1) overcoming a barrier height of 46.498 kcal/mol. IRC calculations confirmed that TS b1 connects the reactants of CF 3 CFCN + Cu and product of CF 3 CF(Cu)CN. Then, CF 3 CF(Cu)CN isomerizes to CN-C(F)CuCF 3 with bond C(1)− −C(3) stretching in TS b2 in reaction R16 with a barrier height of 80.000 kcal/mol. CN-C(F)CuCF 3 continues to dissociate to products of CNCF + CF 3 Cu and CNCCuF + CF 3 through reactions R17 and R19, respectively. The absorbed energies for reaction R17 and R19 are 101.179 kcal/ml and 94.685 kcal/mol, respectively. CNCF dissociates to products of CN + CF in reaction R18, and the other product in reaction R14, CF 2 CuF, dissociates to products of CF 2 + CuF in reaction R20. Reactants of CF 3 CF(CN)CF 2 + Cu also evolve to products of CF 3 CFCF 2 + CuNC and CF 3 CF(CN)CuCF 2 by overcoming a barrier height of 46.460 kcal/mol (reaction R22) and 37.253 kcal/mol (reaction R27), respectively. The vibration mode of TS c2 in reaction R22 shows that radical CN is stretching between atom C(1) and atom Cu (12): as radical CN approaches atom Cu (12), bond length of C(2)− −Cu(12) stretches out to generate products of CF 3 CFCF 2 + CuNC; while radical CN emigrates to atom C(1), bond C(2)− −Cu(12) also shortens to a normal bond in reactants. The decomposition of CF 3 CFCF 2 with Cu has been discussed in Sec. III B 2. The vibration mode of TS c4 in reaction R27 shows that atom C(1) emigrates between atom C(2) and atom Cu (12)  and bond Cu(12)− −C(2) breaking along the reaction coordinate in reaction R28 and R31, respectively. The energies for the above bond-breaking processes are 111.168 kcal/mol and 43.117 kcal/mol, respectively. As the product in reaction R28, CF 2 CuCFCF 3 dissociates into products of CF 2 Cu + CFCF 3 and CF 2 + CF 3 CF(Cu) through reactions R29 and R30, and its decomposition PES is shown in Fig. 10. The decomposition mechanism of CN-C(F)CuCF 3 , the product in reaction R31 have been discussed in Sec. III B 3.

C. Rate constant of C 4 F 7 N-Cu mixtures decomposition
The logarithm of rate constants of C 4 F7-Cu mixtures decomposition reactions R1-R31 is shown in Fig. 11. The unit of rate constant for a unimolecular reactions is s −1 , while it is cm 3 /mol/s for a bimolecular reaction. The rate constants of reactions R14, R15, R22, R23, and R27 decrease dramatically over 300 K-3500 K temperature range, while the rate constants of other reactions increase with temperature rising. It can be inferred that reactions R1-R31 should be considered in modeling a decaying arc or investigating the insulation performance 10 because their rate constants are higher than 10 −8 cm 3 /molec/s or s −1 in some temperature regions. The rate constants differ significantly over the temperature range because of the difference between the reaction mechanisms. For example, the rate constant of reaction R30 is higher than that of reaction R29 because the PES of reaction R30 is lower than that of reaction R29, making it play a major contribution in CF 2 CuCFCF 3 decomposition. The original data of rate constants of 31 reactions are shown in the supplementary material.

ARTICLE
scitation.org/journal/adv By evaluating the contributions of reactions in loss of C 4 F7N, C 3 F7, CF 3 CF(CN)CF 2 , CF 3 C(CuF)CF 3 , and CF 3 CFCuCF 3 as well as the generation of CuF, CN, CF 2 , and CF 3 in C 4 F7N-Cu mixtures, this paper selects the main reactions over different temperature regions as listed in Table III.

IV. CONCLUSION
In this paper, the B3LYP method in conjunction with the 6-311G(d, p) basis set (for C, N, and F atoms) and Lanl2DZ basis set (for Cu atom) combining TST method is used to study the decomposition mechanism of C 4 F7N in the presence of Cu. 31 reactions are found in the decomposition of C 4 F7N-Cu mixtures. PES of these reactions is obtained, and the reaction mechanisms are analyzed. The rate constants over 300 K to 3500 K relevant to the insulation breakdown temperature are calculated based on above quantum chemistry calculations, and dominant reactions in different temperature regions are selected. For example, reaction R14 plays the major role in the decomposition of C 4 F7N-Cu mixtures below 1500 K, while reactions R1 and R21 are dominant above 1500 K; reaction R23 is the most important reaction leading to the generation of CF 3 below 1500 K with the overwhelming rate constant, but reactions R7, R8, R9, R12, R13, R19, and R23 are dominant above 1500 K. This work is hopeful to lay a theoretical basis to evaluate the insulation behavior of C 4 F7N-insulated power equipment and explore novel gas sensors.

SUPPLEMENTARY MATERIAL
See the supplementary material for the comparison of geometric structures of common species such CF 2 , CN, and CuF as presented in Fig. S1 and the original data of rate constants of 31 reactions as shown in Table S1.