Adsorption Properties of ZSM-5 Molecular Sieve for Perfluoroisobutyronitrile Mixtures and Its Fluorocarbon Decomposition Products

Abstract

Perfluoroisobutyronitrile (C₄F₇N), an environment-friendly insulating gas, has excellent insulating properties and has the potential to be used in gas-insulated equipment when mixed with CO₂. Selecting suitable adsorption materials to adsorb the decomposition products of the C₄F₇N mixture can ensure the safe and stable operation of the gas-insulated equipment and the personal safety of the operators in the electric power industry. The adsorption characteristics of the ZSM-5 molecular sieve on C₄F₇N and its five fluorocarbon decomposition products were investigated by adsorption experiments. The results show that the ZSM-5 molecular sieve has a certain adsorption effect on six fluorocarbon gases; the adsorption performance of C₃F₆ and C₃F₈ are the best, with an adsorption efficiency over 85%, while the concentration of CO₂ and C₄F₇N is affected by the ZSM-5 molecular sieve. At the same time, the paper based on the Metropolis Monte Carlo simulation of Materials Studio software found that the ZSM-5 molecular sieve has the strongest adsorption effect on C₄F₇N molecules and the weakest adsorption effect on CO₂ molecules. The stronger the polarity of the gas molecule, the more obvious the adsorption effect of molecular sieve structure on it. As a result, the ZSM-5 molecular sieve could be used in tail gas purification of insulated equipment, as well as to provide solutions for the development and production of protective equipment.

1. Introduction

SF₆ is a stable gas with excellent insulation and arc extinguishing performance, which is widely used in various gas insulation equipment However, the global warming potential (GWP) value of SF₆ is high, about 23,500 times that of CO₂. In the “Kyoto Protocol” promulgated in 1997, it was clearly stated that SF₆ is a greenhouse gas [1,2,3,4]. With the increasing global greenhouse effect on the environment, the control of greenhouse gas emissions has become the primary task of protecting the global environment. Some countries and regions have introduced policies and regulations to limit the usage and emissions of SF₆. Particularly, China has pledged to reduce SF₆ emissions to 55~60% of 2005 levels by 2019 [5]. It is one of the research hotspots to find a new type of environment-friendly insulating medium that can replace SF₆ in the electric power industry. 3M, in conjunction with General Electric, introduced the “g3”, an environmentally friendly insulating gas, which is mainly 4~10% perfluoroisobutyronitrile (C₄F₇N) and CO₂. The insulation performance of C₄F₇N is about twice that of SF₆, and the GWP of C₄F₇N is only 2090. Although the liquefaction temperature is high (−4.7 °C), it has potential application in engineering when mixed with a buffer gas (CO₂, N₂) [6,7].

At present, a new type of environment-friendly insulating gas has been put into trial operation in a variety of gas-insulated equipment, and good progress has been made in the development of insulating equipment using C₄F₇N/CO₂ as the medium. However, when the gas-insulated equipment is operated for a long time, the insulating medium will decompose due to discharge and overheating [8,9,10]. Simka et al. have found that C₄F₇N mixtures can produce C₂F₃N and C₂N₂ toxic gases in partial discharge [11]. Radisavljevic et al. analyzed the gas after C₄F₇N arc discharge and found cyanides such as CF₃CN and C₂F₅CN [12]. Li et al. studied the decomposition products of C₄F₇N/CO₂ mixtures under different operating conditions; the results show that CO, CF₄, C₂F₆, C₂F₄, CF₃H, C₃F₈, C₃F₇H, C₂F₃N, C₃F₆, C₄F₆, C₄F₁₀, C₃F₅N, C₂N₂, C₄F₈, and HCN [13] are the possible decomposition products under different fault conditions. To ensure that some of the decomposition products will not interfere with the insulation performance of the insulating medium and ensure the long-term stable operation of the equipment, the equipment will be equipped with porous absorbent materials to adsorb the decomposition products.

Many studies have been carried out on the adsorption of decomposition products of C₄F₇N mixtures, and it is considered that molecular sieve and activated alumina materials have the potential to be used in gas-insulated equipment [14,15]. F. Meyer et al. studied the adsorption properties of 3A, 4A, and 5A molecular sieves for decomposition products CF₃CN, C₂F₅CN, and COF₂ [16]. Han et al. compared and analyzed the adsorption characteristics of γ-Al₂O₃ and molecular sieves (3A, 4A, and 5A) on C₄F₇N/CO₂ mixtures and their decomposition products. The experimental results show that γ-Al₂O₃ and molecular sieves have weak adsorption capacity for buffer gas CO₂ and decomposition products of CO and perfluorocarbon gases, but they can selectively adsorb nitrile gases, and the other four adsorbents can effectively adsorb the CNCN. γ-Al₂O₃ and 5A molecular sieve can effectively adsorb the CF₃CN [17]. The team of Hua simulated the adsorption performance of the C₄F₇N/CO₂ mixtures and its 13 decomposition products in Na-4A molecular sieve through molecular dynamics simulation methods. They calculated saturated adsorption capacity, adsorption isotherm, adsorption free energy, and other thermodynamic parameters and found that the Na-4A molecular sieve has weak adsorption capacity for perfluoroalkanes and olefins, and can efficiently adsorb various highly toxic decomposition products [18]. Based on the existing research, the adsorption of carbon and fluorine decomposition products of the C₄F₇N/CO₂ mixtures is not clear. Some fluorocarbon gases, hexafluoroethane (C₂F₆) and octafluoropropane (C₃F₈) have high greenhouse effect potential and atmospheric lifetime [19]. Some fluorocarbon gases, fluoroform (CF₃H), Carbon tetrafluoride (CF₄), and hexafluoropropylene (C₃F₆) are toxic which would cause adverse reactions after human inhalation [20,21,22]. Therefore, it is necessary to select suitable adsorbents to efficiently adsorb and treat the above-mentioned gases. In the selection of adsorbents, the pore diameter of the selected molecular sieve materials is mostly no more than 0.5 nm, and the adsorption effect of these materials on the larger gas molecules is not obvious. In this paper, the ZSM-5 molecular sieve with 0.55 nm pore diameter was selected to investigate the adsorption properties of five fluorocarbon decomposition products (CF₃H, CF₄, C₂F₆, C₃F₆, and C₃F₈) for the C₄F₇N mixtures and the main insulating gases (C₄F₇N and CO₂). Then the adsorption experiment of the standard gases of the above eight gases was carried out, and the concentrations were measured before and after the experiment by gas chromatography-mass spectrometer (GCMS) and Fourier infrared spectrometer (FTIR). Finally, the adsorption properties of ZSM-5 molecular sieve for all gases were evaluated by the reduction of concentration. In addition, the adsorption heat and other parameters are calculated by the method of molecular dynamics simulation in this paper. The results provide a reference for the selection of the adsorption materials in environmentally friendly insulated gas power equipment.

2. ZSM-5 Adsorption Experiment

2.1. Adsorption Experiment Method

Considering the actual concentration of C₄F₇N mixtures and common fluorocarbon gases, the relevant parameters of the adsorbate gas concentration used in this paper are shown in

Table 1. The concentration of adsorbate gas in the experiment.

Type of Gas	CF3H	CF4	C2F6	C3F6	C3F8	C4F7N Standard Gas	CO2 Standard Gas	Gas Mixtures
concentration	200 ppm	200 ppm	200 ppm	200 ppm	200 ppm	15%C4F7N/85%He	15%CO2/85%He	15%C4F7N/85%CO2

. The dilution background of each low-concentration standard gas was He, and the concentration error was ±3 ppm. The three adsorbate gases with high concentrations were made up of 99.2% C₄F₇N, 99.999% CO₂, and 99.999% He. The ZSM-5 molecular sieve material was provided by Pioneer Nano Company, and the adsorption experimental device is shown in Figure 1. The detailed experimental procedure is shown in

Table 2. Experimental procedure.

Step	Operation Content
①	Checking the tightness of the device
②	Loading with 3 g of adsorption material
③	Pumping of the chamber to vacuum
④	Heating activates the adsorbent
⑤	Vacuuming for five minutes to remove the residual impurity gas in the chamber
⑥	Filling the chamber with gas to an absolute pressure of 0.15 MPa, and then letting it stand for 48 h.
⑦	Detecting the concentration of the gas in the chamber after 48 h

.

2.2. Concentration Detection

For the concentration detection of low-concentration gases (CF₃H, CF₄, C₂F₆, C₃F₆, and C₃F₈), the instrument used was a Shimadzu QP-2020NX gas chromatography-mass spectrometer, equipped with a special capillary column (CP-Sil5CB) for gas qualitative determination detection, using 99.999% He as carrier gas. The column temperature adopted the programmed heating method: 0–9 min maintained at 32 °C unchanged, then the temperature was raised to 120 °C at a constant rate for 3 min after 9 min, and finally the temperature was kept unchanged to maintain the temperature at 120 °C. The inlet temperature was 200 °C, the split ratio was 50:1, the 200 °C EI ion source was used as the MS ion source, and the sample inlet temperature was 200 °C. After gas sampling, ion scanning was used to quantitatively analyze the characteristic ions of each fluorocarbon gas. The corresponding characteristic ion mass-to-charge ratios (m/z) of the six gases were 69, 69, 119, 131, 69, respectively.

For the detection of high-concentration gases (C₄F₇N and CO₂), the instrument used was a Thermo Scientific Nicolet iS50 infrared spectrometer. The optical path length of the gas cell in the infrared spectrometer was 2.4 m, the volume was 200 mL, and the material of the windows at both ends of the gas cell was potassium bromide. The number of scans was set to 1 and the resolution to 4 before collecting the absorption spectrum signal. A specific absorption peak was selected and the intensity of each gas absorption peak was compared before and after the experiment to quantitatively analyze the change in concentration. For the analysis of C₄F₇N concentration, the characteristic peaks in the range of 2220~2290 cm⁻¹ were selected. For CO₂ concentration analysis, characteristic peaks with wavenumbers ranging from 3650 to 3755 cm⁻¹ were selected.

2.3. Analysis of Results

Figure 2 shows the concentration changes of six low-concentration fluorocarbon decomposition products before and after the adsorption experiment.

It can be seen from the figure that the concentration of each gas decreased after the experiment. The adsorption properties of the ZSM-5 molecular sieve for five kinds of fluorocarbons were different, and the adsorption properties for C₃F₆ and C₃F₈ were the best. The concentrations of the three gases decreased from 200 ppm to 25 ppm, and 6 ppm respectively, and the adsorption efficiency was over 85%. The concentration of C₂F₆ decreased from 200 ppm before the experiment to 61 ppm after the experiment, and the adsorption efficiency was over 65%, but it was weaker than that of C₃F₆ and C₃F₈. The concentration of CF₃H and CF₄ decreased from 200 ppm before the experiment to 161 ppm and 156 ppm after the experiment, and the adsorption efficiency did not exceed 25%. In general, the ZSM-5 molecular sieve adsorbed most of the fluorocarbon gases, although the adsorption performance of some fluorocarbon gases was not outstanding, and the concentration of fluorocarbon decomposition products produced by the equipment due to various failures generally did not exceed the concentration parameters set in this paper. Thus, ZSM-5 molecular sieve could adsorb fluorocarbon decomposition products in electrical insulation equipment.

Figure 3a shows the characteristic peaks corresponding to the 15% C₄F₇N standard gas before and after the experiment. The figure shows that the characteristic peak of C₄F₇N had no obvious change after the experiment, and the two characteristic peak curves coincide. Comparing the absorbance in the 2260~2280 cm⁻¹ band, the absorbance of the characteristic peak before the experiment was slightly higher than the absorbance after the experiment, which shows that the ZSM-5 molecular sieve exhibited a partial adsorption capacity for C₄F₇N. Figure 3b shows the characteristic peaks corresponding to the 15% CO₂ standard gas before and after the experiment. The curve of absorbance changes with the number before and after the experiment was roughly the same in this band. When the wave number was constant, the absorbance before the experiment was slightly higher than that after the experiment, so the ZSM-5 molecular sieve also shows partial adsorption capacity to CO₂.

To calculate the change of main gas concentration before and after the experiment, this paper used Beer-Lambert law to calculate the two concentrations after the experiment through the relevant parameters of the standard gas concentration and the light absorption degree before the experiment [23]. To ensure the accuracy of the calculation results, this paper used two parameters, the peak area of the characteristic peak and the intensity of the absorbance, to estimate the concentration reduction after the experiment. Figure 4 shows the changes of the two parameters of the two gases before and after the experiment.

In the infrared absorption spectrum of C₄F₇N, the curve in the 2260~2290 cm⁻¹ band was integrated to calculate the peak area of the characteristic peaks before and after the experiment. The intensity of the absorbance corresponding to the wave velocity of 2272 cm⁻¹ was selected to quantitatively analyze the concentration of C₄F₇N before and after the experiment. In the infrared absorption spectrum of CO₂, the curve in the band of 3650~3755 cm⁻¹ was integrated, the peak area of the characteristic peak was calculated, and the intensity of the absorbance corresponding to the wave velocity of 3725 cm⁻¹ was selected to quantitatively analyze the CO₂ concentration before and after the experiment. Figure 4a shows the corresponding characteristic peak areas of the two gases before and after the experiment, and Figure 3b shows the corresponding absorbance intensities of the two gases before and after the experiment. It can be seen from Figure 4 that the parameters of the two gases after the experiment were lower than before the experiment. To reflect the adsorption efficiency of the ZSM-5 molecular sieve material for the two gases, the adsorption rate was calculated according to the characteristic peak area before and after the experiment and the decrease in absorbance intensity.

Table 3. Adsorption rates of CO₂ and C₄F₇N standard gases by ZSM-5 molecular sieve.

	Peak Area	Absorbance Intensity
adsorption rate of CO2	2.3%	1.9%
adsorption rate of C4F7N	5.6%	3.4%

shows the absorption rate of CO₂ and C₄F₇N in the ZSM-5 molecular sieve calculated with two parameters. Comparing the calculation results of the adsorption rate of the two gases, the adsorption rate calculated by the characteristic peak area was slightly higher than the result calculated by the intensity of the absorbance. From the perspective of the adsorption effect, the error calculated by the two methods can be ignored. According to the calculated adsorption rate of the two gases, the adsorption effect of ZSM-5 molecular sieve on CO₂ and C₄F₇N was not obvious, the average absorption rate of CO₂ was 2.1%, and the average adsorption rate of C₄F₇N was 4.5% [23].

Figure 5 shows the infrared spectra of C₄F₇N and CO₂ in the gas mixtures before and after the adsorption experiment. The concentration of C₄F₇N and CO₂ in the gas mixtures before and after the experiment is calculated by the characteristic peak area and peak intensity.

Table 4. Adsorption rates of CO₂ and C₄F₇N in gas mixtures by ZSM-5 molecular sieve.

	Peak area	Absorbance intensity
adsorption rate of CO2	−1.8%	−2.3%
adsorption rate of C4F7N	3.9%	8.0%

shows the adsorption rates of CO₂ and C₄F₇N in ZSM-5 molecular sieve for different parameters in the gas mixtures. It can be seen from Table 3 that the concentration of CO₂ after the adsorption experiment is slightly higher than that before the experiment, and the average adsorption rate is −2.05%. The concentration of C₄F₇N is lower than that before the experiment, and the average adsorption rate is 5.9%. It shows that the ZSM-5 molecular sieve in the mixed gas mainly adsorbs C₄F₇N gas, and C₄F₇N inhibits the adsorption of CO₂.

ZSM-5 molecular sieve has good adsorption capacity for most of the fluorocarbon decomposition products of the environmentally friendly insulating medium C₄F₇N/CO₂ mixtures and can achieve complete adsorption for a few fluorocarbon gases. It has the potential to be used in the treatment and purification of exhaust gas from gas-insulated equipment, and it can also provide solutions for the development and production of related protective equipment. However, the ZSM-5 molecular sieve also has a certain adsorption effect on the main insulating medium; specifically, it would affect the concentration of C₄F₇N in the mixtures. If the ZSM-5 molecular sieve can not only adsorb the fluorocarbon decomposition products but also ensure that the insulation performance of the gas is not affected, the proportion of C₄F₇N in the mixture can be increased to 1.04~1.08 times.

3. Adsorption Simulation

3.1. Model Building and Parameter Setting

To clarify the adsorption mechanism and real adsorption capacity of ZSM-5 zeolite porous materials for several fluorocarbon molecules and main insulating gases from the microstructure, the adsorption process was simulated by using the adsorption module in Materials studio simulation software. The initial model was built according to the ZSM-5 molecular sieve in the International Molecular Sieve Association (IZA) database. The simulation was performed using 2 × 2 × 2 cells with cell parameters: a = 4.00 nm, b = 3.98 nm, c = 5.34 nm, α = β = γ = 90°. To be close to the actual experimental molecular sieve structure composition, the Si atoms in the initial model were randomly replaced with Al atoms, and then the Na⁺ equilibrium charge was introduced. The optimized structure of NaZSM-5 with a Si/Al ratio of 47 is shown in Figure 6, which consists of vertical straight channels (5.3 × 5.6 Å) and transverse S-shaped channels (5.5 × 5.1 Å).

According to reference [24], the atomic distribution charges in adsorption are as follows: O(−0.4 e), Si(+0.766 e), Al(+0.575 e), Na (+1 e). The model structures of the five carbon-fluorine decomposition products and two main insulating gas molecules were optimized using the GGA and PBE basis groups in the Dmol3 module. The lowest energy configuration obtained is shown in Figure 7, and the corresponding atomic charge distribution and dynamic diameter are shown in

Table 5. Atomic charge distribution in gas molecules.

Molecule	Charge (e)					Dynamic Diameter (Å)
	C	F	N	H	O
CF4	+1.66	−0.41; −0.41; −0.41; −0.41	-	-	-	4.89
C2F6	+1.35; +1.35	−0.45; −0.45; −0.45; −0.45; −0.45; −0.45	-	-	-	6.24
C3F6	+1.11; −0.23; −1.65	−0.38; −0.40; −0.45; −0.41; −0.48; −0.41	-	-	-	7.25
C3F8	+1.37; +1.02; +1.37	−0.50; −0.46; −0.50; −0.46; −0.46; −0.46; −0.46; −0.46	-	-	-	7.44
CF3H	+0.98	−0.38; −0.38; −0.38	-	+0.16	-	6.21
CO2	+0.60	-	-	-	−0.30; −0.30	5.15
C4F7N	−0.14; +1.64; +1.64;+0.98	−0.43; −0.45; −0.46; −0.47; −0.45; −0.42; −0.46	−0.98	-	-	7.45

. The seven gas molecules were introduced into NaZSM-5 separately for kinetic simulations.

The simulation is based on the Metropolis sampling method, and the non-bond interaction is described by the Lennard-Jones potential energy and Coulomb interaction:

$$U\left(r\right)={\displaystyle \sum _{i,j}{\displaystyle \sum _{i\ne j}4{\epsilon }_{ij}}}\left[{\left(\frac{{\sigma }_{ij}}{r_{ij}}\right)}^{12}-{\left(\frac{{\sigma }_{ij}}{r_{ij}}\right)}^6\right]+\frac{q_i{q}_j}{4\pi {\epsilon }_0{r}_{ij}}$$

(1)

In Equation (1), U is the interaction potential energy between two molecules, i and j are different atoms, σ is the position diameter, ԑ is the depth of the potential well, q is the charge parameter set for each atom, and ԑ₀ is the vacuum dielectric constant. Combined with the actual operating environment of the equipment, the pressure was set to 0.15 MPa and the temperature was set to 300 K in the simulation process. The Universal Force Field was selected for simulation. Both the molecular sieve model and the adsorbate gas molecules were regarded as rigid structures. Electrostatic and van der Waals interactions were calculated by Ewald and atom-based methods respectively [24,25]. The 1 × 10⁶ step of the Monte Carlo method process was used to keep the balance, and the 1 × 10⁵ step was used to calculate the heat of adsorption of the gas molecules in the porous structure. The Lennard-Jones parameters of molecular sieve and adsorbent atom based on universal force field are shown in

Table 6. Atomic Lennard-Jones parameters in molecular sieves.

Atom	σ/Å	ε/kJ·mol−1
Si	4.27	1.30
Al	4.39	1.30
O	3.40	0.40
Na	3.14	2.10
C	3.89	0.39
N	3.66	0.32
F	3.47	0.30
H	3.19	0.06

.

3.2. Analysis of Simulation Results

Figure 8 shows the distribution of seven gas molecules in the structure of the NaZSM-5 molecular sieve after the adsorption simulation, and the red dots in the molecular sieve structure indicate the possible locations of gas molecules being adsorbed. From the Figure, the adsorption sites of CF₃H and CF₄ molecules in the carbon-fluorine decomposition product molecules are widely distributed in the molecular sieve structure, with adsorption sites in both S-type pore channels and straight pore channels. This indicates that CF₃H and CF₄ molecules are small in size and can diffuse freely in the molecular sieve structure, and the skeleton of the molecular sieve does not have a significant inhibitory effect on the two gas molecules. The adsorption sites of C₂F₆, C₃F_6, and C₃F₈ molecules are mainly distributed at the intersection of S-type pore channels and straight pore channels, and a few adsorption sites are in the longitudinal straight pore channels. This shows that the volume of these three gas molecules is larger than the diameter of the S-shaped pore channel, and it is difficult to enter the pore channel and interact with the internal atoms, and the molecular sieve skeleton has a binding effect on the movement of the three gas molecules. From Figure 8f, the adsorption sites of CO₂ molecules are distributed in two different pore channels in the molecular sieve structure. On the one hand, the CO₂ molecules are small in size, and the molecules can move across the pore channels from one pore cage to another when they diffuse; on the other hand, the atoms on the molecular sieve skeleton have an average interaction potential for CO₂ molecules, and there is no significant difference in the adsorption produced by the pore cages and pore channels. The adsorption sites of C₄F₇N molecules in the molecular sieve structure are mainly concentrated at the pore crossings. The range of C₄F₇N adsorption sites is smaller than the range of the three bulkier fluorocarbon gases. This can indicate that C₄F₇N molecules are larger than the molecular sieve pore size, and it is difficult to cross the pores to enter the inside of the molecular sieve structure, so intermolecular forces are generated mainly with the atoms on the pore cage to bind the C₄F₇N molecules.

The adsorption heat is an important parameter reflecting the strength of adsorption when gas is adsorbed in the structure of the molecular sieve. The adsorption heat parameters of seven gas molecules in the molecular sieve structure were calculated. The adsorption heat probability curves of seven gas molecules in the NaZSM-5 molecular sieve structure are shown in Figure 9. The results show that the adsorption heat of the seven gas molecules in the Nazsm-5 molecular sieve is −124.0~−49.3 kJ/mol.

Table 7. Average adsorption heat of seven gas molecules in NaZSM-5 molecular sieve.

Type of Gas	CF3H	CF4	C2F6	C3F6	C3F8	C4F7N	CO2
average adsorption heat (kJ/mol)	−89.2	−60.5	−83.5	−113.1	−96.5	−124.0	−49.3

shows the average adsorption heat of seven gas molecules in the Nazsm-5 molecular sieve.

Among the five-carbon and fluorine gas molecules, CF₄ is a non-polar molecule with a small and stable structure, and the molecular sieve structure exhibits a low heat of adsorption of −60.5 kJ/mol for it. The average adsorption heat of CF₃H is −89.2 kJ/mol and is slightly higher than that of CF₄. Although CF₃H molecule volume is smaller than CF₄, it is not easy to be bound by the molecular sieve skeleton, but its polar molecules are more likely to interact with the atoms on the skeleton to be bound, and when the molecular sieve skeleton and atomic force field act simultaneously, the two gases are subject to the same adsorption when they do molecular motion, so the adsorption rate of molecular sieve for the two gases in the adsorption experiment is approximately equal. The adsorption heat of C₂F₆ in the molecular sieve structure is between non-polar molecules (CF₄) and polar molecules (C₃F₆ and C₃F₈). Due to the symmetric structure of the C₂F₆ molecule, the molecular polarity is weak, but its larger size is more likely to interact with the atoms on the skeleton to produce strong induced attraction and then be bound in the molecular sieve skeleton, so in practice the molecular sieve material shows a certain adsorption capacity for C₂F₆. The two polar molecules C₃F₆ and C₃F₈ have larger average adsorption heats in the molecular sieve, −113.1 kJ/mol and −96.5 kJ/mol, respectively, and produce stronger interactions with the atoms in the molecular sieve. At the same time, the molecular volume of the two is large and also bound by the molecular sieve skeleton, so the NaZSM-5 molecular sieve shows a better adsorption capacity for it, and the adsorption rate of the molecular sieve for the two in the adsorption experiment is up to 80% or more.

For the two main insulating gas molecules, CO₂ and C₄F₇N, CO₂ has the smallest average adsorption heat in the molecular sieve structure at −49.3 kJ/mol, for which the NaZSM-5 molecular sieve shows a weak interaction. C₄F₇N has the largest adsorption heat in the molecular sieve structure at −124.0 kJ/mol, for which the molecular sieve is between strong physical adsorption and weak chemisorption. The NaZSM-5 molecular sieve has a strong adsorption effect on C₄F₇N. This could explain the phenomenon that the molecular sieve has a more pronounced effect on the concentration of C₄F₇N in the adsorption experiment.

The results of theoretical calculation correspond to the results of adsorption experiments. The results show that the adsorption of small non-polar gas molecules on the NaZSM-5 molecular sieve is weak, followed by small positive gas molecules. The larger chemical polarity molecules can interact with the atoms inside the molecular sieve to create an induced attraction that allows the gas to be partially absorbed. The stronger the molecular polarity, the more obvious the adsorption effect of molecular sieve structure on gas molecules, and the better the actual adsorption effect on gas.

4. Conclusions

In this paper, the adsorption properties of the ZSM-5 molecular sieve on the C₄F₇N/CO₂ mixtures and its five common fluorocarbons were investigated by gas adsorption experiments and molecular dynamics simulation. The following conclusions were obtained:

(1)

ZSM-5 molecular sieve has a certain adsorption effect on six kinds of fluorocarbon gases, of which the adsorption performance of C₃F₆, and C₃F₈ is the best, with an adsorption efficiency of over 85%.

(2)

The concentration of the two main insulating gases CO₂ and C₄F₇N will be affected by the ZSM-5 molecular sieve; if the insulation performance of the gas is not affected in practical application, the proportion of C₄F₇N in the mixture can be increased 1.04~1.08 times.

(3)

NaZSM-5 molecular sieve has the strongest adsorption to C₄F₇N and the weakest adsorption to CO₂. The stronger the polarity of the gas molecule, the more obvious the adsorption effect of molecular sieve structure.