Reactive molecular dynamics study of the decomposition mechanism of the environmentally friendly insulating medium C₃F₇CN

Abstract

The extensive use of sulfur hexafluoride (SF₆) gas in the power industry has a strong greenhouse effect. Hence, many scholars are committed to studying SF₆ alternative gases to achieve green power development. In the past two years, C₃F₇CN (heptafluoroisobutyronitrile) has attracted the attention of many scholars due to its excellent insulation and environmental protection characteristics as a potential alternative gas. This study theoretically explores the decomposition characteristics of C₃F₇CN and the C₃F₇CN/CO₂ gas mixture based on the reactive molecular dynamics method and density functional theory. The main decomposition pathways of C₃F₇CN and the enthalpy of each path at different temperatures were analyzed. The yield of the main decomposition products was obtained under several temperature conditions. The decomposition of C₃F₇CN mainly produced CF₃, C₃F₇, CN, CNF, CF_2, CF, F, and other free radicals and a few molecular products, such as CF₄ and C₃F₈. The C₃F₇CN/CO₂ gas mixture has more excellent decomposition characteristics than that of the pure C₃F₇CN. The addition of CO₂ effectively ensures that the gas mixture has a low liquefaction temperature, which is considerably suitable for use as a gas insulation medium. The relevant research results provide guidance for the further exploration on the electrical properties and practical engineering application of the C₃F₇CN gas mixture.

---

1 Introduction

SF₆ (sulfur hexafluoride) is the most common insulation medium used in high-voltage (HV) electrical equipment such as gas insulated switchgear (GIS).¹ However, SF₆ is a greenhouse gas (GHG) regulated under the Kyoto Protocol with a global warming potential (GWP) of 22 800.^2,3 Although the proportion of greenhouse effect caused by the use of SF₆ is not as large as the amount of fossil fuel energy consumed, the global atmospheric content of SF₆ has increased by 20% in the past five years and the atmospheric mole fraction of SF₆ is 7.28 ppq currently corresponding to a radiative forcing of 0.0041 w m⁻².^4,5 THe greenhouse effect has caused widespread public concern in recent years.⁶ Climate deterioration caused by SF₆ has gradually become one of the contradictions between power industry development and environmental protection. Hence, the power industry should gradually limit the use of SF₆ and search for an environmentally friendly alternative gas for SF₆ in gas-insulated equipment.

C₃F₇CN (heptafluoroisobutyronitrile) is a potential alternative gas for SF₆ (see Fig. 1 for its molecular structure) with a GWP value of 2100 and Ozone Depression Potential (ODP) value of 0.⁴ The dielectric strength of C₃F₇CN is 2.74 times that of SF₆.⁷ It should be used in combination with CO₂ or other buffer gases with low liquefaction temperature due to its high liquefaction temperature (i.e., −4.7 °C). In addition, the toxicity of the C₃F₇CN/CO₂ gas mixture with C₃F₇CN volume fraction below 10% is lower than that of pure SF₆. Thus, the former is harmless to humans.⁸ These physical and chemical properties indicate that C₃F₇CN is excellent in terms of environmental protection, insulation and safety, and has immense potential to replace SF₆.

Fig. 1 Molecule structure of C₃F₇CN.

In the past two years, many scholars have proven that the C₃F₇CN/CO₂ gas mixture has excellent insulation and arc extinguishing performance through numerous tests.^3,7–10 For engineering applications, a gas insulated line (GIL) with a rated voltage of 420 kV and a current transformer (CT) with a rated voltage of 245 kV using the C₃F₇CN/CO₂ gas mixture as insulating medium were tested in 2015.^3,8 In addition, the investigation of the decomposition properties of a gas-insulating medium is an important component to comprehensively evaluate its performance. On the one hand, the insulation defects caused by aging under long-term operating conditions in electrical equipment will lead to partial discharge (PD) or flashover and is accompanied by the decomposition of the insulating medium. The local overheating fault can also lead to the decomposition of the insulating medium. On the other hand, the decomposition characteristics of the insulating medium are closely related to its self-recovery characteristics and arc-extinguishing properties.

At present, researches on the decomposition characteristics of C₃F₇CN have achieved noteworthy results. Kieffel et al. investigated the pyrolysis characteristics of the C₃F₇CN/CO₂ gas mixture. They found the initial decomposition temperature of C₃F₇CN was 650 °C, thereby mainly producing decomposition products, such as CO, CF₃CN, and C₂F₅CN.⁸ Andersen et al. investigated the atmospheric chemistry of C₃F₇CN and evaluated its environmental effects based on density functional theory and infrared spectroscopy.¹¹ Our team has explored the decomposition mechanism of C₃F₇CN in the presence of trace water impurity through quantum chemical calculation. The ionization parameters of C₃F₇CN and various decomposition products were analyzed from the perspective of molecular structure, thereby providing a theoretical reference for the further study of the C₃F₇CN decomposition characteristics.¹²

At present, only a few studies have been conducted on the decomposition characteristics of the C₃F₇CN/CO₂ gas mixture. Our previous studies on the decomposition mechanism of C₃F₇CN were from the perspective of individual molecule and did not consider the decomposition process in multi-molecular and gas mixture systems. In this paper, we constructed the multi-molecular and C₃F₇CN/CO₂ gas mixture model to study the decomposition process of C₃F₇CN using the reactive force field molecular dynamics (ReaxFF-MD) method and density functional theory (DFT). The decomposition paths, reaction enthalpy, and products distribution of the C₃F₇CN gas mixture at different temperatures were firstly investigated. Relevant results provide guidance for further exploration on electrical properties and practical engineering application of C₃F₇CN gas mixture.

2 Theoretical methods and computational details

2.1 ReaxFF-MD simulation details

ReaxFF is based on bond order to describe the dissociation and formation of covalent bonds at the atomic level¹³ and is a beneficial tool for studying chemical reactions and predicting the performance of new materials. During the past more than ten years, ReaxFF has been used to treat a variety of elements and multifunctional systems and is suitable for describing polyatomic systems.^14–16 The energy expression of ReaxFF is presented as follows:

E_system = E_bond + E_over + E_under + E_val + E_pen + E_tors + E_conj + E_vdWaals + E_Coulomb, (1)

where E_bond corresponds to bond energy; E_over and E_under denote the over and under coordinated atoms, respectively, in the energy contribution; E_val, E_pen, E_tors, E_conj, E_vdWaals, and E_Coulomb represent the valence angle term, penalty energy, torsion energy, conjugation effects to energy, nonbonded van der Waals interaction, and Coulomb interaction, respectively.

Two periodic cubic boxes were built to explore the decomposition of C₃F₇CN and the C₃F₇CN/CO₂ gas mixture (see Fig. 2). The C₃F₇CN system contains 100 C₃F₇CN molecules with a density of 0.00811 g cm⁻³. The length of the box is 159 Å. At present, many scholars have tested the insulation characteristics of C₃F₇CN/CO₂ mixture and found that mixtures with 20% C₃F₇CN displayed dielectric strengths comparable to SF₆.^4,7 In order to reveal the decomposition properties of mixtures in this scale, we built the C₃F₇CN/CO₂ system which contains 100 C₃F₇CN molecules and 400 CO₂ molecules with a density of 0.00351 g cm⁻³. The length of the box is 260 Å. The initial density corresponds to the actual density of C₃F₇CN under 25 °C and 0.1 Mpa. The system was minimized with the NVE ensemble for 5 ps at 5 K and equilibrated with the NVT ensemble thereafter for 10 ps at 1000 K using a time step of 0.1 fs.¹⁶ The NVT simulations were then conducted at different temperatures for 1000 ps to explore the decomposition process of C₃F₇CN and C₃F₇CN/CO₂ gas mixture. The temperature was controlled by the method of a Berendsen thermostat with a 0.1 ps damping constant.¹⁷ All the ReaxFF-MD simulations in this paper were performed based on the ReaxFF module of ADF (Amsterdam Density Functional).¹⁸

Fig. 2 Configurations of the C₃F₇CN and C₃F₇CN/CO₂ systems.

2.2 DFT calculation details

The first principle of calculations based on the density functional theory (DFT) was performed to obtain the precise dissociation energy of C₃F₇CN at different temperatures. This theory is widely used in the study of chemical reaction mechanisms, and can describe any given chemical system with high accuracy.^19–21

The spin unrestricted DFT calculation in the current study was performed based on the Dmol 3 module of the Materials studio 8.0. The meta-generalized approximation (mGGA) method treated by M06L functions is used to describe the exchange-correlation energy. This local density functional does well for predicting geometries and vibrational frequencies and gives great performance for a combination of main-group thermochemistry and thermochemical kinetics.^22,23 The double numerical atomic orbital augmented by d-polarization (DNP) is used as the basis set. Geometry optimizations were performed with the following convergence parameters: (1) energy convergence tolerance of 1.0 × 10⁻⁵ Ha, (2) maximum force of 0.002 Ha Å⁻¹, and (3) maximum displacement of 0.005 Å. It should be pointed that spin contamination (unrestricted DFT is much less affected by it than unrestricted Hartree–Fock theory²⁴) and wave function unstability may occur in spin unrestricted calculation. The self-consistent-field (SCF) solution stability is checked to ensure the accuracy of calculation results.

The enthalpy correction is presented as follows to obtain the precise energy of the system at a given temperature:²⁵

ΔE = ∑(E_product + ΔE_product) − ∑(E_reactant + ΔE_reactant), (2)

where E_product and E_reactant correspond to the energy of the product and reactant, respectively, at 0 K and ΔE_product and ΔE_reactant correspond to the correcting values of the enthalpy of product and reactant, respectively, at a given temperature.

3 Results and discussion

3.1 Decomposition of the pure C₃F₇CN

3.1.1 Decomposition process and main reaction paths. The three main types of discharges in electrical equipment are the corona (partial discharge), arc, and spark discharges, wherein the corona and arc discharges are the most common.²⁶ The maximum temperature in the central region of the corona discharge is in the range of 700–1200 K and the highest temperature in the arc discharge region can reach 12 000 K.^26–28 The current study analyzed the decomposition mechanism of C₃F₇CN and its gas mixture at different temperatures to clarify the particle composition and product change. It should be noted that the time scale of ReaxFF MD simulation is normally limited to several dozens of nanoseconds due to the computational cost.^28,29 Hence, we enhance the temperatures to accelerate the simulation process to allow reactions can be observed.

First, the multi-molecular ReaxFF molecular dynamic simulations of C₃F₇CN were performed and demonstrated that the C₃F₇CN molecule began to decompose at 1900 K. Fig. 3 shows the time evolution of the major species of the C₃F₇CN decomposition at 1900 K. Table 1 shows the generation time of the major species and corresponding decomposition pathways. At 1900 K, the C₃F₇CN decomposition mainly produces free radicals, such as CF₃, C₃F₇, CN, CNF, CF₂, CF, CF₃CFCN(C₃NF₄), and F. It is important to note that some free radicals, such as CN, are harmful free radicals and could produce toxic molecules. The content and toxicity of the decomposition products of C₃F₇CN need further test study before engineering application.

Fig. 3 Time evolution of the major species at 1900 K.

Table 1 Generation time of the major species and corresponding decomposition pathways

Major species	Generation time (ps)	Reaction		ΔE^a (kcal mol^−1)
a T = 1900 K, at mGGA-M06L level with ZPE and enthalpy corrections.
C3NF4, CF3	615	A	C3F7CN → CF3 + CF3CFCN	72.67
C3F7	702.5	B	C3F7CN → C3F7 + CN	101.33
C4F6N	—	C	C3F7CN → (CF3)2CCN + F	89.17
CN, CF	685.625	D	CF3CFCN → CF3 + CF + CN	174.66
CF2, F	748.125	E	CF3 → CF2 + F	79.04
CNF	748.125	F	CN + F → CNF	−118.48
CF4	845.625	G	CF3 + F → CF4	−107.64
C	—	H	CF → C + F	121.23

---

C₃F₇CN began to decompose at 615 ps, thereby producing CF₃CFCN and CF₃. This reaction path is required to absorb 72.67 kcal mol⁻¹. Another decomposing path began at 702.5 ps, thereby producing C₃F₇ and CN and needs to absorb 101.33 kcal mol⁻¹. In addition, the decomposition of CF₃CFCN is at 685.625 ps, thereby generating CF₃, CF, and CN. CF₃ can further decompose to CF₂ and F by absorbing 79.04 kcal mol⁻¹, CF can further decompose to C and F by absorbing 121.23 kcal mol⁻¹. CF₄ is one of the stable decomposition products, which is formed by CF₃ and F and releases −107.64 kcal mol⁻¹. In addition, the DFT calculation results show that the energy required for path A is lower than that for path B, which means that path A is more likely to occur than path B. And the ReaxFF-MD results show that the number of generated CF₃ increases rapidly during 700–1000 ps, which indicates the results of ReaxFF are consistent with DFT calculations and the performance of ReaxFF is reliable.

3.1.2 Effect of temperature on the decomposition process. To further analyze the influence of temperature on the decomposition process of C₃F₇CN, the molecular dynamic simulation of the C₃F₇CN system was performed at five temperature zones, that is, 1900 K, 2000 K, 2100 K, 2200 K, and 2400 K, for a total simulation time of 1000 ps.

Fig. 4 shows the time evolution of the potential energy between 1900 K and 2400 K. The potential energy of the system increases during the reaction process at different temperatures, thereby indicating that the entire decomposition process of C₃F₇CN needs to absorb energy. The growth rate of the potential energy is evidently accelerated with the increase of temperature, indicating that the increase of the ambient temperature will lead to the accelerated decomposition of C₃F₇CN. When the ambient temperature reaches 2400 K, the growth rate of the potential energy in the time range of 125 ps to 600 ps is significantly higher than that of 600–1000 ps. This result indicates that the reactions in the system before 600 ps are mainly absorbed energy. Moreover, exothermic reactions may occur in the system after 600 ps, thereby leading to the considerably gradual growth of the overall potential energy.

Fig. 4 Time evolution of the potential energy between 1900 K and 2400 K (C₃F₇CN system).

Fig. 5 illustrates the time evolution of the C₃F₇CN molecules between 1900 K and 2400 K. The decomposition rate and quantity of the C₃F₇CN molecules are evidently accelerated with the increase of temperature. For example, only 8 C₃F₇CN molecules decomposed at 1900 K, whereas 96 C₃F₇CN molecules decomposed at 2400 K at the end of the molecular dynamic simulation.

Fig. 5 Time evolution of the C₃F₇CN molecules decomposition when temperature is increased from 1900 to 2400 K (C₃F₇CN system).

To further analyze the effect of temperature on the decomposition of C₃F₇CN, the enthalpy of the main reaction paths in the temperature range of 300–3000 K is calculated based on DFT (see Fig. 6). As the temperature increases, the reaction enthalpy of paths A, B, D, and E decreases, that is, the temperature increase is favorable for the reaction progress. The enthalpy change of path D is the most evident, thereby indicating that the increase of temperature has a significant effect on the decomposition of CF₃CFCN. The reaction enthalpy change of paths C and H is not significantly with increasing temperature. Paths F and G are exothermic reaction processes and the energy released decreases as the temperature increases.

Fig. 6 Enthalpy change of the reaction pathways from 300 K to 3000 K.

3.1.3 Intermediates and main products. Fig. 7 illustrates the time evolution of the major species of C₃F₇CN decomposition over the temperature range from 1900 K to 2400 K. Fig. 8 shows the maximum number of produced species. The statistical results in Fig. 7 shows that the amount of C₃F₇, CN, CNF, CF₃, CF₂, CF, CF₃CFCN(C₃NF₄), F, and other free radicals and CF₄ produced by the decomposition of C₃F₇CN increased with the increase of temperature.

Fig. 7 Time evolution of the major species between 1900 K and 2400 K (C₃F₇CN system).

Fig. 8 Number of produced species in C₃F₇CN at different temperatures.

The statistical results in Fig. 8 show that the content of CF₃ at each temperature is the highest among all the products, while the contents of CF₂, CN, CNF, and F are similar. The contents of C₃F₇, C₃NF₄, and C₄NF₆ are lower compared with those of other products. C₄NF₆ is generated by the dissociation of the F atom in the C₃F₇CN molecule from the central C atom (C2 atom as shown in Fig. 1), thereby requiring the absorption of 89.17 kcal mol⁻¹. In addition, a few small stable molecules, such as CF₄ and C₃F₈, are found during the simulation. The generation rate of CF₄ in the time range of 600–1000 ps at 2400 K increased and the amount of CF₃ and F during the time range reduced (as shown in Fig. 7). Given that the generation path of CF₄ (path G in Table 1) is exothermic, the gradual growth of the overall potential energy of the system is actually due to the large amount of CF₄ produced.

3.2 Decomposition of the C₃F₇CN/CO₂ mixture

3.2.1 Effect of temperature on the decomposition process. In order to obtain the decomposition properties of the insulating medium in a practical application condition, the decomposition characteristics of C₃F₇CN/CO₂ gas mixture over the temperature range from 2000 K to 3000 K were discussed. Fig. 9 shows the time evolution of the potential energy of the C₃F₇CN/CO₂ gas mixture system. Fig. 10 provides the decomposition amount of C₃F₇CN.

Fig. 9 Time evolution of the potential energy between 2000 K and 3000 K (C₃F₇CN/CO₂ system).

Fig. 10 Time evolution of the C₃F₇CN molecules decomposition when temperature is increased from 2000 to 3000 K (C₃F₇CN/CO₂ system).

The potential energy of the system shows an increasing trend in the range of 2200–3000 K. When the ambient temperature is above 2400 K, the potential energy of the system increases. When the ambient temperature is 2000 K, the potential energy of the system does not change substantially, thereby exhibiting a relationship with the decomposition degree when C₃F₇CN is low. The decomposition rate of C₃F₇CN is accelerated with the increase of temperature, thereby showing consistency with the conclusion of the pure C₃F₇CN system.

3.2.2 Intermediates and main products. Fig. 11 describes the time evolution of the major decomposition species of C₃F₇CN over the temperature range from 2000 K to 3000 K. Fig. 12 presents the maximum number of produced species.

Fig. 11 Time evolution of the major species between 2000 K and 3000 K (C₃F₇CN/CO₂ system).

Fig. 12 Number of produced species in C₃F₇CN/CO₂ at different temperatures.

Similar to the pure C₃F₇CN system, the content of the major decomposition products shows an increasing trend with the increase of temperature. The content of CF₃ in the products is the highest at each temperature. When the temperature is below 2800 K, the yields of CN, CNF, CF, and F are similar. When the temperature is above 2800 K, the content of F in the system increases. A new decomposition of products, namely, CO, is produced due to the addition of CO₂. The decomposition of CO₂ accelerated at temperatures over 2600 K and 40 molecules of CO are produced at 3000 K. In the C₃F₇CN/CO₂ system, the yield of CF₄ reduces. Only 12 CF₄ molecules were produced at 3000 K at the end of the simulation, which is lower than that of the pure C₃F₇CN system. In addition, the generation rate of CF₃ and F after 800 ps at 3000 K slowed down, which is related to the formation of CF₄.

3.2.3 Effect of CO₂ on the decomposition process. Fig. 13 shows the final decomposition amount of C₃F₇CN in C₃F₇CN and C₃F₇CN/CO₂ system at the same temperature. The decomposition amount of C₃F₇CN in the C₃F₇CN/CO₂ gas mixture system is lower than that in the pure C₃F₇CN system at the same temperature. For example, the decomposition amount in the pure C₃F₇CN system is 96 at 2400 K, whereas that of the C₃F₇CN/CO₂ gas mixture system is 58. The addition of the buffer gas CO₂ can effectively inhibit the decomposition of C₃F₇CN.

Fig. 13 Final decomposition amount of C₃F₇CN at different temperatures.

Table 2 shows that the final amount of the decomposition products in the C₃F₇CN/CO₂ gas mixture system is lower than that in the pure C₃F₇CN system at the same temperature. Moreover, the addition of CO₂ inhibits the formation of CF₄ and C, thereby making it difficult for the system to precipitate carbon and produce products with relatively poor insulation properties. This result is beneficial to protect the insulation properties of the system.

Table 2 Final amount of the decomposition products at different temperatures

Major species	Temperature
	2000 K		2200 K		2400 K
	C3F7CN	C3F7CN/CO2	C3F7CN	C3F7CN/CO2	C3F7CN	C3F7CN/CO2
C3F7	8	1	11	2	11	3
CF3	31	6	76	25	83	52
CF2	4	1	16	2	27	7
CF	9	4	33	8	48	32
CN	12	2	35	11	49	21
CNF	18	2	39	5	49	25
F	17	3	46	11	51	46
C3NF4	2	2	3	4	4	4
C4NF6	0	1	2	0	3	3
CF4	1	0	14	0	44	0
C	15	0	46	5	68	16

---

Therefore, the C₃F₇CN/CO₂ gas mixture is relatively suitable for using as a gas-insulated medium relative to pure C₃F₇CN. The mechanism of the action of CO₂ over C₃F₇CN to avoid its decomposition process in function of temperature needs further study.

4 Conclusion

This study analyzed the decomposition characteristics of C₃F₇CN and its gas mixture based on ReaxFF-MD method and density functional theory. The main decomposition pathways, reaction enthalpy, and the composition of the main decomposition products of C₃F₇CN at different temperatures were revealed. The decomposition characteristics of C₃F₇CN and C₃F₇CN/CO₂ gas mixture were compared and analyzed.

The decomposition of C₃F₇CN mainly produces CF₃, C₃F₇, CN, CNF, CF₂, CF, F, and other free radicals and a few molecular products, such as CF₄ and C₃F₈. The decomposition rate of C₃F₇CN and generation rate of the products increase with the increase of ambient temperature. The C₃F₇CN/CO₂ gas mixture has more excellent decomposition characteristics than that of pure C₃F₇CN at the same condition, which is relatively suitable for use as a gas-insulated medium in electrical equipment.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The current work is supported by the science and technology project of China Southern Power Grid (No. CSGTRC-K163010)

References

1. A. Beroual and A. Haddad, Recent Advances in the Quest for a New Insulation Gas with a Low Impact on the Environment to Replace Sulfur Hexafluoride (SF₆) Gas in High-Voltage Power Network Applications, Energies, 2017, 10(8), 1216.
2. X. Zhang, Y. Li and S. Xiao, et al., Decomposition mechanism of C₅F₁₀O: An environmentally friendly insulation medium, Environ. Sci. Technol., 2017, 51(17), 10127–10136.
3. J. Reilly, R. Prinn and J. Harnisch, et al., Multi-gas assessment of the Kyoto Protocol, Nature, 1999, 401(6753), 549.
4. Y. Kieffel, T. Irwin and P. Ponchon, et al., Green Gas to Replace SF₆ in Electrical Grids, IEEE Power and Energy Magazine, 2016, 14(2), 32–39.
5. M. P. Sulbaek Andersen, M. Kyte and S. T. Andersen, et al., Atmospheric Chemistry of (CF₃)₂CFCN: A Replacement Compound for the Most Potent Industrial Greenhouse Gas, SF₆, Environ. Sci. Technol., 2017, 51(3), 1321–1329.
6. Z. Lv, D. Zhao and S. Xu, Facile synthesis of mesoporous melamine-formaldehyde spheres for carbon dioxide capture, RSC Adv., 2016, 6(64), 59619–59623.
7. J. G. Owens, Greenhouse gas emission reductions through use of a sustainable alternative to SF₆, IEEE Electrical Insulation Conference (EIC), Montréal, Canada, 2016, pp. 535–538.
8. Y. Kieffel, Characteristics of g3-an alternative to SF₆, IEEE International Conference on Dielectrics(ICD), Montpellier, France, 2016, vol. 2, pp. 880–884.
9. H. E. Nechmi, A. Beroual and A. Girodet, et al., Fluoronitriles/CO₂ Gas Mixture as Promising Substitute to SF6 for Insulation in High Voltage Applications, IEEE Trans. Dielectr. Electr. Insul., 2016, 23(5), 2587–2593.
10. C. Preve, R. Maladen and D. Piccoz, Method for validation of new eco-friendly insulating gases for medium voltage equipment, IEEE International Conference on Dielectrics (ICD), Montpellier, France, 2016, vol. 1, pp. 235–240.
11. M. P. S. Andersen, M. Kyte, S. T. Andersen, C. J. Nielsen and O. J. Nielsen, Atmospheric Chemistry of (CF₃)₂CF–CN: A Replacement Compound for the Most Potent Industrial Greenhouse Gas, SF₆, Environ. Sci. Technol., 2017, 51(3), 1321–1329.
12. X. Zhang, Y. Li and S. Xiao, et al., Theoretical study of the decomposition mechanism of environmentally friendly insulating medium C₃F₇CN in the presence of H₂O in a discharge, J. Phys. D: Appl. Phys., 2017, 50(32), 325201.
13. A. C. T. V. Duin, S. Dasgupta and F. Lorant, et al., ReaxFF: A Reactive Force Field for Hydrocarbons, J. Phys. Chem. A, 2001, 105(41), 9396–9409.
14. J. Zhang, Y. Si and C. Leng, et al., Molecular dynamics simulation of Al–SiO₂ sandwich nanostructure melting and low-temperature energetic reaction behavior, RSC Adv., 2016, 6(64), 59313–59318.
15. S. Bhoi, T. Banerjee and K. Mohanty, Insights on the combustion and pyrolysis behavior of three different ranks of coals using reactive molecular dynamics simulation, RSC Adv., 2016, 6(4), 2559–2570.
16. D. Hong, L. Liu and Y. Huang, et al., Chemical effect of H₂O on CH₄ oxidation during combustion in O₂/H₂O environments, Energy Fuels, 2016, 30(10), 8491–8498.
17. H. J. C. Berendsen, J. P. M. Postma and W. F. V. Gunsteren, et al., Molecular dynamics with coupling to an external bath, J. Chem. Phys., 1984, 81(8), 3684–3690.
18. G. Te Velde, F. M. Bickelhaupt and E. J. Baerends, et al., Chemistry with ADF, J. Comput. Chem., 2001, 22(9), 931–967.
19. T. C. Ngo, D. Q. Dao and M. T. Nguyen, et al., A DFT analysis on the radical scavenging activity of oxygenated terpenoids present in the extract of the buds of Cleistocalyx operculatus, RSC Adv., 2017, 7(63), 39686–39698.
20. Q. Zhou, X. Su and W. Ju, et al., Adsorption of H2S on graphane decorated with Fe, Co and Cu: a DFT study, RSC Adv., 2017, 7(50), 31457–31465.
21. P. Zhang, X. Hou and J. Mi, et al., Curvature Effect of SiC Nanotubes and Sheet for CO₂ Capture and Reduction, RSC Adv., 2014, 4(90), 48994–48999.
22. Y. Zhao and D. G. Truhlar, The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals, Theoretical Chemistry Accounts: Theory, Computation, and Modeling, 2008, 120(1), 215–241.
23. Y. Zhao and D. G. Truhlar, A new local density functional for main-group thermochemistry, transition metal bonding, thermochemical kinetics, and noncovalent interactions, J. Chem. Phys., 2006, 125(19), 194101.
24. A. S. Menon and L. Radom, Consequences of spin contamination in unrestricted calculations on open-shell species: effect of Hartree-Fock and Møller-Plesset contributions in hybrid and double-hybrid density functional theory approaches, J. Phys. Chem. A, 2008, 112(50), 13225–13230.
25. X. Zhang, S. Xiao and J. Zhang, et al., Influence of humidity on the decomposition products and insulating characteristics of CF₃I, IEEE Trans. Dielectr. Electr. Insul., 2016, 23(2), 819–828.
26. R. Ono and T. Oda, Measurement of gas temperature and OH density in the afterglow of pulsed positive corona discharge, J. Phys. D: Appl. Phys., 2008, 41(3), 035204.
27. Y. Fu, M. Rong and K. Yang, et al., Calculated rate constants of the chemical reactions involving the main byproducts SO₂F, SOF₂, SO₂F₂ of SF₆ decomposition in power equipment, J. Phys. D: Appl. Phys., 2016, 49(15), 155502.
28. H. Tanaka, D. Tanahashi and Y. Baba, et al., Finite-difference time-domain simulation of partial discharges in a gas insulated switchgear, High Voltage, 2016, 1(1), 52–56.
29. Z. Chen, W. Sun and L. Zhao, High-temperature and High-pressure Pyrolysis of Hexadecane: A Molecular Dynamic Simulation Based on Reactive Force Field (ReaxFF), J. Phys. Chem. A, 2017, 121(10), 2069.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ra09959b

---