Xianqiang Zheng, Yanlong Shen, Shitao Wang,*, Ke Huang, Dapeng Cao,*

1 State Key Laboratory of Organic-Inorganic Composites, Beijing University of Chemical Technology, Beijing 100029, China

2 State Key Laboratory of Laser Interaction with Matter, Northwest Institute of Nuclear Technology, Xi’an 710024, China

Keywords:Sulfur hexafluoride (SF6)Metal-organic frameworks Molecular simulation Covalent-organic frameworks Adsorption Separation

A B S T R A C T Sulfur hexafluoride (SF6) is an extremely severe greenhouse gas. It is an urgently important mission to find excellent candidates for selective adsorption of SF6, in order to reduce the emission of SF6 facilities.Here,we adopt the molecular simulation method to systematically explore the selective adsorption of SF6 in 22 kinds of representative covalent-and metal-organic frameworks. Results indicate that COF-6 is a promising candidate for the SF6 adsorption at low pressure P ＜20 kPa because of its small pore size,while MOF-180 and PAF-302 are excellent candidates at high pressure P = 2 × 103 kPa due to their large Brunauer-Emmett-Teller specific surface area(BET SSA)and pore volumes.For the two cases of the power industry(XSF6=0.1)and the semiconductor industry(XSF6=0.002)environments,COF-6 and ZIF-8 are fairly promising candidates for selective adsorption of SF6 from the SF6/N2 mixtures,because they not only present the high selectivity, but also the large adsorption capacity at ambient environment, which can be considered as potential adsorbents for selective adsorption of SF6 at ambient conditions.

1. Introduction

Sulfur hexafluoride (SF6) is a non-polar gas with a highly symmetrical octahedral geometry [1]. Due to its high dielectric strength, chemical stability, non-toxicity and incombustibility[2,3],it is widely used in industry[4-7].In spite of these characteristics,SF6is also a very severe greenhouse gas,whose global warming potential is twenty thousand times higher than CO2[8]. More seriously, its atmospheric lifetime is about three thousand years[1].Due to the long-term impact of SF6on the environment,many attempts have been made for the purpose of reducing emissions from SF6facilities. Even SF6was written into the Kyoto Protocol[8],which aims to control the world’s major man-made emissions.In addition, the F-Gas directive also leads to the use limitation of SF6in Europe [4].

It is a common method to reduce the use of SF6by mixing with nitrogen(N2)while still maintaining the required properties of the pure component [9], which can make the whole process cheaper and more environmentally friendly [10]. However, the mixture of SF6and N2increases the difficulty of separating SF6. In order to selectively adsorb SF6, membrane separation [11-14] and adsorption separation [7,15-21] technologies are often used due to their low operating cost and energy saving[22].Adsorption of SF6on different materials, such as zeolites [15,23], porous organic cages[18],metal-organic frameworks(MOFs)[7,18]and even columnar clay [24] have been reported experimentally and theoretically. By investigating MCM-41 and two zeolite template carbon materials,Buileset al. [2] found that the material with a pore size of about 1.1 nm shows the best selectivity. Baeet al. [25] addressed that the introduction of tertiary amine groups will increase the isosteric heat of adsorption,thereby improving the performance of the DCXbased POP framework for the adsorption of SF6. Takaseet al. [17]used single-walled carbon nanohorns(CNHs)for SF6/N2separation and confirmed that wide carbon nanopores are effective for separating SF6from mixtures. By introducing a series of polar functional groups into UIO-66, Kimet al.[26] proved that it is possible to systematically control the interaction between SF6molecule and the adsorbent surface by using polar functional groups and confirmed that the material with stronger polarizability exhibits higher selectivity for SF6/N2. Skarmoutsoset al. [27]adopted Monte Carlo and molecular dynamics (MD) simulation to investigate SIFSIX-2-Cu and its interpenetrated polymorphisostructural SIFSIX-2-Cu-i for SF6/N2separation,and found that controlling the inter-molecular permeation is an efficient way to adjust the performance of strategic gas mixtures separation.

However,there are few reports on covalent-organic frameworks(COFs) for selectively adsorbing of SF6. Different from MOFs, COFs hold strong covalent bonds (C-C, B-O, B-C,etc.) connected by light elements (such as C, H, and O) [28], so they have higher hydrothermal stability than MOFs. Moreover, 3D-COFs typically have ultra-high specific surface area (SSA) or large pore volume for gas adsorption, while 2D-COFs possess small pore size for separation of gas mixtures, as previously reported in literatures[29,30].Zeolitic imidazolate frameworks(ZIFs)are tetrahedral network structures formed by transition metal atoms and imidazole organic ligands [31]. Owing to the strong coordination bond between metal atoms and imidazole ligands,ZIFs also exhibit high thermodynamic and chemical stability in organic and aqueous solutions[32,33].Importantly,ZIFs usually exhibit small pore size,which is suited for gas separation. Zenget al. [34] adopted grand canonical Monte Carlo (GCMC) simulation to find that ZIF-12 is an excellent adsorbent for capturing Rn in indoor air. Therefore,it is of great significance to investigate the uptake and selective separation of SF6in COFs and ZIFs.

In current work,the GCMC simulations were carried out to systematically study the selective adsorption of SF6in 22 representative covalent- and metal-organic frameworks, including MOFs,COFs and ZIFs. Uptake and selectivity of SF6in these materials are evaluated and the potential candidates are recommended.Finally, some conclusions are drawn and discussed.

2. Computational Details

2.1. Porous materials

Previous investigations also reported adsorption and separation of SF6in some MOFs. Here, we did not consider these MOFs involved in previous investigations [5]. Instead, we considered other MOFs, including MOF-5, MOF-177, MOF-180, IRMOF-20,IRMOF-61, IRMOF-62, MIL-47, MIL-100, MIL-101 and UMCM-1,because these MOFs have large SSA and large pore volume such as MOF-180 [35] and MOF-177 [36], or small pore size like IRMOF-62 [37]. Besides MOFs, we also considered another two kinds of framework materials,COFs and ZIFs,because they are also excellent candidates in gas adsorption and separation process because of their large Brunauer-Emmett-Teller specific surface area (BET-SSA) or small pore size [38], including 2D-COFs (COF-5,COF-6, COF-8, COF-10) and 3D-COFs (PAF-302, COF-102 and COF-103), and ZIFs (ZIF-8, ZIF-10, ZIF-11, ZIF-67). Fig. 1 exhibits the structures of eight representative porous materials, and Table S1(Supplementary Material) lists the structure parameters of all these materials.

2.2. Models and methods

The structures of all materials were obtained from the Cambridge Crystal Database. The force field parameters of SF6and N2molecules represented by spherical Lennard-Jones(LJ)model were obtained from the literatures[39-40].The Dreiding force field[41]was used to describe the interaction between the skeleton atoms in the adsorbents and the UFF force field [42] was used only if the atoms not in the former.Because SF6is a non-polar molecule with a highly symmetrical octahedral structure,the adsorption of SF6in porous materials was mainly determined by van der Waals interactions. In addition, N2is also a non-polar molecule, so we did not consider coulomb interaction between gas molecules and framework atoms. In the simulation process, all the adsorbents were considered as a rigid material. Table 1 lists all the force field parameters of SF6and N2molecules and atoms in porous materials.

Table 1 Force field parameters for gases [39,40] and porous materials [41,42], where σ and ε are potential parameters reflecting the equilibrium distance between atoms and the depth of potential energy, and kb is Boltzmann constant

Here, the GCMC simulations were used to explore the selective adsorption of SF6in MOFs,COFs and ZIFs.Two mimicked industrial environments were considered, in which SF6:N2= 10:90 is for the power industry emission and SF6:N2=0.2:99.8 for the semiconductor industry emission. The calculation was carried out at 298 K with multipurpose simulation code (MUSIC 4.0) [43]. For a comparison to experimental data, the excess uptake is obtained by the absolute uptake minus the amount of bulk phase predicted by Peng-Robinson equation of state. The selectivity (Sabs（i/j）) of componentiagainst componentjis given bySabs（i/j）＝（xi/xj）/（yi/yj）, wherexandyare the molar ration of the components in adsorption and bulk phase [44],respectively.More details on the calculations can be found in the literatures [29,45-47] and the Supplementary Material.

3. Results and Discussion

3.1. Force field parameter verification

Fig.1. Structure illustrations of eight representative porous materials:(a)COF-6,(b)MOF-180,(c)ZIF-8,(d)COF-103,(e)UMCM-1,(f)IRMOF-62,(g)MOF-5,(h)PAF-302.The structures of all materials were obtained from the Cambridge crystal database.

In order to verify the accuracy of the force field parameters,first, we simulated the adsorption of SF6in DUT-9 and MIL-101(Cr),because these adsorption data is available from the literature of Senkovskaet al. [5]. Moreover, the two materials presented excellent performance for adsorption of SF6,and can be considered as the representative candidate. Fig. 2 shows the uptakes of SF6in the two materials, which is very consistent with the data from Senkovskaet al.[5], especially for the DUT-9 material. Although some errors appear for the MIL-101(Cr), the errors are still in the range of reasonable deviation.Therefore,we believed that the force field parameters can well evaluate the adsorption and separation of SF6in covalent- and metal-organic frameworks.

3.2. Adsorption of pure SF6

Fig. 3(a)and (b)show the SF6adsorption isotherms in 22 kinds of porous materials at 298 K at low pressure and high pressure,respectively. At low pressure area ＜20 kPa, the uptakes of SF6in COF-6, ZIF-8 and ZIF-67 increase rapidly and reach saturation due to their small pore sizes, because small pore would lead to strong affinity toward SF6owing to the overlap of potential energies within the small pores. At high pressure 1000-2058 kPa,MOF-180 and PAF-302 exhibit the extremely high uptakes due to their large pore volumes available for accommodating more SF6molecules.

Previous literatures [20,48] reported that the uptakes of SF6in zeolite-13X and activated carbon are 2.3 and 2.5 mmol·g-1at 298 K and 100 kPa, respectively. Interestingly, in the same condition, our calculation results indicate that the SF6adsorption capacities of COF-102, COF-103, IRMOF-20, IRMOF-62 and IRMOF-61 reach 12.36, 10.86, 10.71, 6.34 and 6.27 mmol·g-1,respectively, which are several times higher than zeolite-13X. At high pressure of 2058 kPa, the uptake order of SF6is MOF-180(28.95 mmol·g-1) ＞ PAF-302 (25.58 mmol·g-1) ＞ COF-103(22.22 mmol·g-1) ＞ UMCM-1 (20.13 mmol·g-1) ＞ COF-102(18.40 mmol·g-1),which is the completely same order of pore volume and BET SSA of these materials.It is found that SF6adsorption in MOF-180 requires a larger pressure to reach saturation compared to other materials.This is because MOF-180 has a very large pore size(1.522 nm)and pore volume(3.378 cm3·g-1),which lead to a weak affinity for SF6.Therefore,the adsorption capacity of SF6at low pressure is relatively small. However, at high pressure region,the SF6gas molecules were pressed into the relatively large pores of MOF-180, resulting in a larger adsorption capacity atP＞1 × 103kPa. The observation indicates that the adsorption of SF6at high pressure is in proportion to the BET SSA and pore volume of the porous materials. On the contrary, the COF-6, ZIF-8 and ZIF-67 with smaller pore size are excellent candidates for SF6adsorption at low pressureP＜20 kPa.

For the purpose of observing the adsorption behavior of SF6in the microstructure of COF-6 more intuitively, Fig. 4(a) and (b)show the snapshots of SF6in COF-6 at 10 and 2× 103kPa,respectively,which indicates that the optimal adsorption site of SF6molecule is located at the organic chain 1,3,5-benzenetriboronic acid(BTBA). The small pore size of COF-6 limits the entry of SF6molecules, leading to the saturation of adsorption at lower pressure about 10 kPa (Fig. 3(a)). Therefore, in the two snapshots of 10 and 2 × 103kPa, no apparent molecule number increases, due to its adsorption saturation. The density profiles of SF6in MOF-180 at 100 and 2×103kPa are shown in Fig.4(c)and(d),respectively.Most of the SF6molecules are adsorbed on the pore walls of the material, while almost no molecules are adsorbed in the middle of the pores at low pressure.SF6molecules fill the entire pore when the pressure is 2×103kPa,and the adsorption probability near the pore wall is apparently greater.The large pore volume of MOF-180 can adsorb more SF6molecules,which results in a great adsorption amount at high pressure.

Fig. 2. Adsorption isotherms of SF6 in (a) DUT-9 and (b) MIL-101(Cr) at 298 K.

Fig. 3. SF6 adsorption isotherms in 22 kinds of porous materials at 298 K: (a) at low pressure (0-100 kPa) and (b) high pressure (0-2000 kPa).

3.3. Separation of SF6 from the SF6/N2 mixture

Figs. 5 and 6 show the selectivities of 22 kinds of materials for SF6/N2mixture withXSF6= 0.1 andXSF6= 0.002 (mole fraction),respectively.At the case ofXSF6=0.1,COF-6 presents an extremely high selectivity for SF6/N2mixture. The selectivity is in a range of 460-850, and it shows a gradual decrease trend with the increase of pressure, indicating that COF-6 has a great potential in separating SF6. This possible reason is that the small pore size of COF-6(only 0.9 nm) exhibits a strong confinement effect on gas molecules, leading to a high selectivity. In addition, ZIF-8, ZIF-11, ZIF-67 and MIL-47 with small pore size also present relatively high selectivity, especially at low pressure. At 10 kPa, the selectivity of ZIF-11 is ～674, and the selectivities of ZIF-8, ZIF-67 and MIL-47 are in the ranges of 90-256, 70-222 and 43-310, respectively.Overall,the selectivity decreases with increasing pressure,because the high pressure would lead to the reduce of material affinity for target gas.

Fig.4. Snapshots of SF6in COF-6 at(a)10 kPa and(b)2×103kPa.Rose red balls represent SF6gas molecules.Density profiles of SF6in MOF-180 at(c)100 kPa and(d)2×103kPa.

In order to compare with the literature, we highlighted the selectivities of these 22 materials at 298 K and 100 kPa in Fig. 5(b), where the selectivities of COF-6, ZIF-8, ZIF-11, ZIF-67 and MIL-47 are 749.02,243.18,201.31,206.18 and 200.04,respectively,which are much bigger than the highest reported values of other adsorbents in the same conditions: Ca-A zeolite (S= 28) [24],SIFSIX-2-Cu (S= 25) [19], zeolite-13X (S= 44) [15], Zn-MOF-74(S= 46) [7], comparable to UiO-66-Br2(S= 220) [26]. Therefore,these five materials (red points in Fig. 5(b)) may be potential candidates for the separation of SF6.

AtXSF6= 0.002, the selectivity of COF-6 is still very high, and maintaining in the range of 560-720. Interestingly, the selectivity of ZIF-11 at 10 kPa reaches ～2325,which is much higher than other materials.With the pressure increases,it drops dramatically to 290 at 2×103kPa,but still only inferior to COF-6.In addition,we also found that IRMOF-62 presents high selectivity especially at low pressure, which is 654 at 10 kPa, indicating that ZIF-11, COF-6 and IRMOF-62 are potential candidates for selective adsorption of trace amount of SF6at ambient conditions. Similar to the case ofXSF6= 0.1, ZIF-8, ZIF-67 and MIL-47 also exhibit relatively high selectivity. The difference is that the selectivity changes more smoothly with pressure, because the SF6content is very low, and there is no large amount of SF6adsorption even at low pressure.The selectivities of ZIF-8, ZIF-67 and MIL-47 are in the ranges of 190-220, 170-196 and 170-300, respectively. The six materials mentioned above are considered as potential candidates for SF6separation atXSF6=0.002.Similarly,we presented the selectivities of 22 materials at 298 K and 100 kPa in Fig. 6(b), where the selectivities of COF-6, ZIF-11 and IRMOF-62 reach 591.74, 1446.32 and 435.93, respectively. Obviously, these three materials (red points in Fig. 6(b)) are better candidates for the separation of SF6, as discussed earlier.

Fig. 5. Selectivities of 22 kinds of materials for SF6/N2 mixtures with XSF6 = 0.1 at 298 K: (a) 0-2 × 103 kPa, (b) at 100 kPa.

Fig. 6. Selectivities of 22 kinds of materials for SF6/N2 mixtures with XSF6 = 0.002 at 298 K: (a) 0-2 × 103 kPa, (b) 100 kPa.

3.4. Uptakes of SF6 and N2 in SF6/N2 mixture

In order to further analyze the adsorption capacity of SF6and N2components in SF6/N2mixture, the uptakes of SF6and N2in 22 kinds of materials for SF6/N2mixture withXSF6= 0.1 andXSF6= 0.002 at 298 K are shown in Figs. 7 and 8, respectively. It can be seen from Fig.7(a)that COF-6 and ZIF-8 show high uptakes of SF6, which are 3.71 and 2.90 mmol·g-1at 100 kPa. Besides the two materials, MIL-47 and ZIF-67 also show higher uptakes for SF6component than other materials, which are 2.73 and 2.67 mmol·g-1at 100 kPa, respectively. The adsorption amount of N2component is very important for SF6/N2selectivity,so we also presented the uptakes of N2in Fig. 7(b). Unsurprisingly, the uptakes of N2in COF-6 and ZIF-8 are relatively low, leading to the high SF6/N2selectivity, as shown in Fig. 5(a).

Fig. 7. Uptakes of (a) SF6 and (b) N2 component in 22 kinds of materials for SF6/N2 mixture with XSF6 = 0.1 at 298 K.

Fig.8. Uptakes of(a)SF6 and(b)N2 component in 22 kinds of materials for SF6/N2 mixture with XSF6 =0.002 at 298 K.(c)The uptakes of N2 component in materials except MIL-101 and MIL-100.

At the case ofXSF6=0.002,MIL-101 has the highest uptake of SF6component among all materials,while its N2uptake is also the largest, resulting in a relatively low selectivity. ZIF-11 also presents high adsorption capacity of SF6, which is 0.37 mmol·g-1at 100 kPa, only inferior to MIL-101. However, the N2uptake of ZIF-11 is the lowest,which therefore leads to the high SF6/N2selectivity, as shown in Fig. 6(a). Similar to the case ofXSF6= 0.1, COF-6 shows not only high SF6uptake, but also relatively low N2adsorption. Therefore, COF-6 presents high SF6/N2selectivity.Except for MIL-101,ZIF-11 and MIL-100,the uptakes of SF6component in other materials increase almost linearly with pressure.Interestingly, we also found that materials with small pore size,such as COF-6 and ZIF-8,present not only high selectivity,but also large SF6adsorption capacity at ambient environment. Therefore,these two materials may be excellent candidates for adsorption and separation of SF6at ambient conditions.

Finally, for the mixture withXSF6= 0.1, we also discussed the influence of temperatures on the performance of SF6adsorption and separation in five representative materials, as shown in Fig.9.It can be seen that as the temperature rises,both the uptake and selectivity show a decreasing trend, which indicates typical physical adsorption. Therefore, we confirm that the selective adsorption of SF6in covalent- and metal-organic frameworks is mainly affected by van der Waals interactions,which is concordant with our simulation assumption.

Fig. 9. Uptakes and selectivities of five representative materials at different temperatures at 100 kPa (273, 283 and 298 K).

4. Conclusions

We have used the molecular simulation method to evaluate the uptakes and selectivities of 22 kinds of representative covalent-and metal-organic frameworks for SF6. Results indicate that COF-6 is an excellent candidate for selective adsorption of SF6at low pressureP＜ 20 kPa because of its small pore size, while MOF-180 and PAF-302 are excellent candidates at high pressureP= 2 × 103kPa due to their high BET SSA and large pore volumes.For the two mimicked industry environments, the power industry(XSF6=0.1) and the semiconductor industry (XSF6=0.002), the simulation results indicated that COF-6 is a very promising candidate to adsorb SF6selectively, whose selectivities for SF6/N2are in the ranges of 460-850 (XSF6= 0.1) and 560-720 (XSF6= 0.002), respectively. The small pore size of COF-6 exhibits a strong confinement effect on SF6gas molecules,leading to the high selectivity.Further investigations found that materials with small pore size, such as COF-6 and ZIF-8, present not only high selectivity, but also large SF6adsorption capacity at ambient environment. Therefore, these two materials may be promising candidates for selective adsorption of SF6at ambient conditions. It is expected that the present work is helpful to the study on the selective adsorption of SF6.

Acknowledgements

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We are greatly thankful to the support from the Open Fund of the State Key Laboratory of Laser Interaction with Matter(SKLLIM1710).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.03.010.