Encapsulation of [bmim⁺][Tf₂N⁻] in different ZIF-8 metal analogues and evaluation of their CO₂ selectivity over CH₄ and N₂ using molecular simulation†

Amro M. O. Mohamed‡ , Panagiotis Krokidas§ and Ioannis G. Economou *
Chemical Engineering Program, Texas A&M University at Qatar, P.O. Box 23874, Education City, Doha, Qatar. E-mail: ioannis.economou@qatar.tamu.edu

First published on 22nd June 2020

---

---

Design, System, Application The manuscript is entirely computational and refers to the design of new hybrid materials, which consist of ionic liquid (IL) pairs encapsulated in the cages of zeolitic imidazolate frameworks (ZIFs) with the aim to develop energy efficient and high performing processes for gas separation. This work focus on the effect of the metal center type of the framework on the performance of IL@ZIFs for the separation of CO2/CH4 and CO2/N2 mixtures. We use the highly studied ZIF-8 and prepare metal variants, by replacing the original metal Zn2+ with Co2+, Be2+ and Cd2+. The IL is [bmim+][Tf2N−]. Using Monte Carlo simulations, we predict the selectivity of the pristine ZIF-8 metal analogues and the IL@ZIF-8 metal analogues for the separation of the gas mixtures. It is shown that metal variation affects the ZIF performance and the best enhancements are those with Be2+ and Cd2+ replacement. In addition, for each IL@ZIF-8 case, there is an optimum IL composition that ensures maximum selectivity for the two mixtures. This optimum composition is related to the available pore volume parameter, which is common in all cases and for both mixtures. This finding can be used for the optimum design of the membrane materials.

1. Introduction

Increased anthropogenic CO₂ emission is considered to be the biggest environmental challenge of our century¹ that if not properly addressed, it is expected to have dreadful consequences on the ecosystem and on humans.² Carbon capture and storage (CCS) is a technology that is projected to support our efforts to achieve global warming control.³ However, the high energy and capital cost of capturing CO₂ from other mixtures with today's processes makes the design of energy-effective approaches for the separation of CO₂ an urgent research topic. Materials with well-controlled pores, whose dimensions range from a few nm to a few hundreds of nm, are in the forefront of research for CO₂ emission mitigation.⁴ Amongst these materials, metal–organic frameworks (MOFs)^5,6 and especially their sub-family, zeolitic imidazolate frameworks (ZIFs),^7,8 are considered as promising CO₂-selective materials for next generation separation processes. They exhibit a unique combination of characteristics of both MOFs and zeolites;^9,10 they have the chemical and thermal stability of zeolites, they undergo facile synthesis routines and they can be functionalized by properly modifying their structure at the molecular level. The latter property can potentially lead to new ZIFs with exceptional CO₂ selectivity.

Among the modification approaches that have been recently reported for enhancing the separation efficiency of ZIFs, such as substitution of building units^11–13 and structural manipulation with electric fields^14,15 and with heat,¹⁶ introduction of cavity occupants⁸ stands out as a very effective way to increase the selectivity towards CO₂. Ban et al.¹⁷ took advantage of this approach by encapsulating ionic liquid (IL) pairs in the cages of ZIF-8. The introduction of ILs in the framework created new adsorption sites, which increased the CO₂ sorption and enhanced the CO₂ selectivity over CH₄ and N₂. Moreover, the performance of this new type of material, called IL@ZIF, exceeds the performance of both the original ZIF-8 and the bulk IL as CO₂-selective materials. Since then, research has been carried out to study the effect of IL encapsulation on the CO₂ selectivity of these hybrid materials.^18–23 Moreover, in our previous work, we examined the role of both the IL type and composition and the distribution of IL pairs in ZIF cages in the performance of IL@ZIFs.²⁴ In this work, we show that IL@ZIFs can be further fine-tuned by properly replacing the metal center of ZIFs' tetrahedral building unit. The ZIF under study is ZIF-8 and the IL is methylimidazolium bis(trifluoromethylsulfonyl)imide ([bmim⁺][Tf₂N⁻]). The original metal center of ZIF-8, Zn²⁺, is replaced with Co²⁺, Be²⁺ and Cd²⁺ and the performance of both the pristine ZIF-8 metal analogues and the IL@ZIF-8 metal analogues is assessed in terms of separating CO₂/N₂ and CO₂/CH₄ mixtures.

This work was carried out exclusively with computational means. The adsorption of CO₂, N₂ and CH₄ was examined using Monte Carlo (MC) simulations, with the help of force fields developed recently by us. All the MC calculations were performed with the help of Cassandra code.²⁵ The results reveal that metal replacement benefits the CO₂ adsorption capacity of ZIF-8 and IL@ZIF-8. Moreover, the selectivity for CO₂/CH₄ and CO₂/N₂ mixtures in some instances is considerably increased when compared with the original, Zn-based, ZIF-8 and IL@ZIF-8. The results highlight the impact of metal substitution in combination with ionic liquid occupants for the design of high-performance CO₂-selective materials.

2. Methodology

Force fields and simulation details

The ZIFs studied in this work are the original ZIF-8 and metal analogues in which the original Zn²⁺ metal center has been replaced with either Co²⁺, Cd²⁺ or Be²⁺. The force fields for the original, Zn-based, ZIF-8 and its metal analogues were presented in our previous works.^12,26,27 Their performance were validated by reproducing several structural characteristics of the framework and by comparing diffusion/sorption of guest species with available literature data. The force field parameters for [bmim⁺][Tf₂N⁻] were developed by Economou and co-workers²⁸ and they have shown satisfactory agreement with experimental data for the volumetric, transport and structural properties.

During the Monte Carlo simulations, ZIFs are rigid, in both their pristine and IL encapsulated state. Therefore, we use only the Lennard-Jones (LJ) and electrostatic terms for non-bonded intra and intermolecular interactions (eqn (1)) of the force fields, according to the expression:

	(1)

where ε_ij and σ_ij are the depth of the LJ potential well and the distance at which the interparticle potential between i and j is zero, q_i and q_j are the partial charges of atoms i and j, respectively, ε_o is the vacuum permittivity and r_ij is the distance between the i and j atoms.

The force field for [bmim⁺][Tf₂N⁻] uses eqn (1) for inter and intramolecular non-bonded interactions. Moreover, ILs are flexible. To account for the flexibility, bond stretching and bond bending (eqn (2) and (3)), dihedral angle distortion (eqn (4)) and improper torsion (eqn (5)) potentials are used.

Bond stretching and angle bending are described through:

	(2)

	(3)

where k_l and k_θ are force constants, l and θ are the bond length and bond angle, and l_o and θ_o are the bond length and bond angle at equilibrium, respectively.

The dihedral angle distortion is expressed through the expression:

u(φ) = kφ[1 + cos(mφ − φo)]	(4)

where k_φ, φ and φ_o are the force constant, the dihedral angle and the dihedral angle at equilibrium.

Additionally, there are improper torsions whose interactions are expressed through:

u(ψ) = kψ(ψ − ψ0)2	(5)

where k_ψ, ψ and ψ_o are the force constant, the torsion angle and the torsion angle at equilibrium.

The transferable potentials for phase equilibria (TraPPE) were used to model CO₂ and N₂.²⁹ In TraPPE, CO₂ is modelled as a linear rigid triatomic molecule with three charged LJ interaction sites located at each atom. Nitrogen is described by two LJ interaction sites. The three charges required to model nitrogen's quadrupole moment are distributed among the two atoms and a third massless pseudo-atom placed at the center of the N–N bond. For the methane molecule, we used OPLS-AA,³⁰ which accounts for the hydrogen atoms explicitly and considers partial charges for C and H atoms. Interactions between unlike atoms were calculated using the Lorentz–Berthelot combining rule:

	(6)

	(7)

The van der Waals interactions were computed with a 14 Å cut-off radius. Electrostatic interactions were evaluated by the Ewald summation with a real space cut-off of 12 Å and an accuracy of 1 × 10⁻⁵; no tail corrections were used.

More details on the force fields and their parameter values can be found in the ESI† accompanying this publication. The accuracy of combining our IL-related and ZIF-related developed force fields to describe an IL@ZIF system was validated in our previous work, through successfully comparing our simulations²² with experimental data¹⁷ on the sorption of CO₂, N₂ and CH₄ and the selectivity of corresponding mixtures in both pristine ZIF-8 and IL@ZIF-8. This has been proven to work accurately in our recent works.

Four different ZIF-8 metal analogues are studied in this work: the original ZIF-8, where the metal center of the tetrahedral building (see Fig. 1) unit is Zn²⁺, and three variations that result from replacing Zn²⁺ with Co²⁺, Be²⁺ and Cd²⁺. This replacement results in three ZIF-8 metal analogues, namely ZIF-67, CdIF-1 and BeIF-1, respectively. Details on the reconstruction of the simulation boxes for these three ZIFs can be found elsewhere.^12,26

For each ZIF-8 metal analogue, we built one IL@ZIF analogue by introducing [bmim⁺][Tf₂N⁻] pairs in the framework's cavities (see Fig. 2). The IL@ZIF-8 metal analogues were created by inserting the [bmim⁺][Tf₂N⁻] IL pair in the cages of the four ZIFs with the use of the configurational bias Monte Carlo (CBMC) method,³¹ following a methodology that we reported in our recent works.^22,24 More specifically, we employed a pre-production step to sample the various coupled intramolecular degrees of freedom of the molecules under insertion (this step was applied for both IL insertion and gas penetrant insertion). The simulation was performed in the gas phase, where the molecules under investigation were segmented into fragments that are either branch points or ring groups. The code stores a large number of conformations produced in a library file. The library content was then used in the actual simulations. In the case of insertion, deletion or regrowth moves attempted, the specified fragments were used to establish new conformation for the molecules by progressively constructing and deconstructing the molecules.³²

For each of the four metal cases, we prepared simulation boxes with a varying number of encapsulated IL pairs to help us assess the effect of IL composition on the CO₂ selectivity of the IL@ZIF metal analogues in this work. Moreover, we have shown that the distribution of the ILs in the ZIF pore system affects the performance of these hybrid materials.²⁴ Thus, for each IL composition in the four IL@ZIF analogues, we created multiple initial configurations to account for the different distribution realizations of the IL pairs in the ZIF cages. Additional information about sampling different configurations and the convergence of the values of the properties calculated are provided in the ESI.†

Sorption simulations with Monte Carlo

The sorption of CO₂, CH₄ and N₂ in both the pristine ZIF-8 metal analogues and the IL@ZIF-8 metal analogues was simulated with the use of the grand canonical Monte Carlo (GCMC) simulation method, in which the chemical potential, μ, volume, V, and temperature, T, are kept constant. Preceding the sorption simulations, we followed the same pre-production steps mentioned in the previous section. In the sorption simulations, the guest molecules were inserted through insertion, deletion or re-growth steps. Furthermore, the IL molecules were fully flexible during the sorption. This is accomplished by allowing the translation, rotation and regrowth of the ILs during the gas sorption simulations. The adsorbed amount predicted by the simulations, N_absolute, is correlated with the experimentally measured amount, N_excess, through:

Nexcess = Nabsolute − ρBulkVpore	(8)

The product ρ_BulkV_pore expresses the amount of gas that would be present in the pores in the absence of gas–solid interactions, ρ_Bulk is the density of the gas and V_pore is the volume of the pores of the solid (ZIF-8 or IL@ZIF-8). The chemical potential was correlated with pressure through pre-production simulations for each gas in their bulk phase at the same temperature and pressure. The sorption MC simulation runs ranged from 0.7 to 3 × 10⁷ steps. More information on the simulation protocol used can be found in our previous works^22,24 and in the ESI.†

The effective volume of the cages of either pristine ZIFs or ZIFs with various encapsulated ILs was determined by inserting a spherical probe molecule with a radius of 1.84 Å that corresponds to N₂. This was accomplished in Materials Studio,³³ with the use of the Connolly surface task.

3. Results and discussion

The adsorption of CO₂ was simulated first in ZIF-8 and its metal analogues to isolate the effect of the metal center on the adsorption capacity and selectivity performance of ZIFs, at 298 K. Fig. 3 shows that the ZIF-8 metal analogues exhibit increased CO₂ capacity compared to ZIF-8. BeIF-1 exhibits the highest CO₂ sorption capacity and a steeper linear part at low pressure (Henry's law regime). The actual values and statistical uncertainties can be found in Table S14.

More detailed runs for CO₂, N₂ and CH₄ with ZIF-8 in the Henry's law regime (low pressure) helped us extract the isotherm slopes for all the guest molecules and estimate the so-called equilibrium selectivity:

	(9)

where K_H,i and K_H,j are the Henry's law solubility constants ^34,35 for components i and j. The Henry's law constant is evaluated by plotting the loading of the adsorbate per the total weight of the composite (weight %) as a function of the corresponding pressure (in bar).

The metal substitution effect on the performance of ZIF-8 is illustrated by the selectivity values for CO₂/CH₄ and CO₂/N₂ mixtures as shown in Table 1. The Be-based analogue (BeIF-1) exhibits the highest selectivity enhancement compared to the original ZIF-8; the initial CO₂/CH₄ selectivity goes from 7.4 to 11.7, while the CO₂/N₂ selectivity is almost doubled, increasing from 9.3 to 18.0.

---

Subsequently, four different IL-based systems were designed and developed, based on ZIF-8 and the three ZIF-8 metal analogues. Following the methodology described in the previous section, we inserted [bmim⁺][Tf₂N⁻] in the cages of the four ZIFs which resulted in a series of IL@ZIFs: [bmim⁺][Tf₂N⁻]@ZIF-8, [bmim⁺][Tf₂N⁻]@ZIF-67, [bmim⁺][Tf₂N⁻]@CdIF-1 and [bmim⁺][Tf₂N⁻]@BeIF-1. For the sake of space, we will use the general notation “IL” in place of “[bmim⁺][Tf₂N⁻]” in all instances. For each IL@ZIF case, we varied the insertion/encapsulation, starting from a low number of ILs and we increased gradually their number, in order to obtain IL@ZIF composites with a wide composition range. Then, the adsorption of the three gases under study was simulated at low pressures and at 298 K, in all IL@ZIF metal analogues and for each IL composition. The IL composition is given in weight fraction and is calculated as:

	(10)

Fig. 4 shows the results of MC simulations for the sorption of CO₂ on all the IL@ZIF metal analogues, as a function of IL composition (the reader can find the data points tabulated in the ESI†). It is clear that for all the ZIF-8 analogues, CO₂ capacity increases when the IL are included in the cages. Moreover, this increase is proportional to the IL composition. This is attributed to the creation of new CO₂ adsorption sites from ILs.^17,22,24 IL@BeIF-1 demonstrates the highest CO₂ adsorption among all the analogues. This can be explained by an advantage over the other metal analogues, since pristine BeIF-1 exhibits an already increased adsorption affinity with CO₂. Moreover, this analogue is the lightest of the four, and this favors the adsorption performance, measured in mmol of sorbed CO₂ per gram of IL@ZIF. The effect of IL composition on the CO₂ capacity in all the IL@ZIF-8 metal analogues is altered when the IL composition reaches a critical value and the capacity drops. In our previous work,²⁴ we showed that this is due to a competitive mechanism between the ILs' strong CO₂ affinity that increases the available sorption sites and the IL bulk that reduces the available pore space.

Finally, we want to draw the reader's attention to the larger statistical uncertainty bars, when compared to those of the pristine ZIF-8 metal analogues (Fig. 3). They are estimated by averaging over the adsorption values among all the IL configurations (conformation and distribution) for a given IL composition, for each IL@ZIF analogue. At this point, it's worth reminding that besides taking into consideration the flexibility conformations of ILs during sorption, we also consider the different possible distribution cases of the ILs in the cages of the ZIFs, at the IL composition point. Thus, for each composition, we simulated the sorption in multiple IL@ZIF realizations that correspond to different initial IL distributions. We want to underline the importance of sampling an adequate number of IL@ZIFs with different initial IL configurations in order to produce reliable results. The reader is referred to the last section of the ESI† (pgs. S-17–S-20) where we discuss the assessment of the IL distribution. In particular, we want to draw the attention to Fig. S-3,† which shows how the CO₂/CH₄ and CO₂/N₂ selectivities reach convergence by increasing the number of different initial IL configurations.

The performance of the analogues is further investigated by assessing their selectivity to CO₂ over CH₄ and N₂. MC single gas sorption simulations at low pressure for CO₂, CH₄ and N₂ provided their adsorption isotherm slopes in the Henry's law region of all the IL@ZIF-8 metal analogues, from which the selectivity is extracted. The estimated selectivities are plotted in Fig. 5 as a function of the IL composition. It is clear that for both CO₂/CH₄ and CO₂/N₂ mixtures, the IL encapsulation results in an enhanced hybrid ZIF selectivity performance, when compared with the performance of the pristine ZIF-8 metal analogues (Table 1). Moreover, the selectivity increases with increasing IL composition, which reaches a maximum value and then decreases. Finally, the reader should note that the high statistical uncertainties, which, as we stated in Fig. 4, are the product of averaging over all different IL distribution configurations in the cages. The results justify this choice and are indicative of the effect the IL distribution has on the separation efficiency for these hybrid materials and it should be taken into consideration when they are tested experimentally.

Fig. 5(a) and (b) reveal that the metal type affects considerably the selectivity of the IL@ZIFs when compared to the original (Zn-based) IL@ZIF-8. Table 2 summarizes the peak performance of all the IL@ZIF analogues, with the maximum selectivity and the corresponding IL wt. fraction, for each case. The maximum CO₂/N₂ selectivity of the IL@ZIF-8 is approx. 76. When Cd replaces Zn (IL@CdIF-1), the new analogue's selectivity reaches 95.2, the Co-based analogue exhibits a selectivity of 111.8, while the Be-based analogue (IL@BeIF-1) outperforms all four analogues, by exhibiting a CO₂/N₂ selectivity of 118.7. In the case of CO₂/CH₄, the improvement is less profound, with the Cd-based (CdIF-1) analogue demonstrating the highest enhancement (IL@CdIF-1: 54.1; IL@ZIF-8: 46.0). The values for the data in Fig. 5 can be found in Table S16.† However, we want to underline the impact of the uncertainty, which stems from sampling over different IL distributions at each IL composition. This is something that is neglected in works related to the encapsulation of ILs in MOFs. Our results show that these hybrid systems are significantly sensitive to the way the ILs are distributed in the available cages of the framework.

---

It is clear that addressing the main factor governing the performance of these materials is not straightforward. Increasing the number of IL pairs in the cages of the IL@ZIFs is regarded as a critical mechanism of sorption and selectivity control in these systems. At the same time, the increase of the number of IL molecules translates to a reduction of available free pore space. For each of the gases, there is a critical IL molecular number which marks the start of the decrease of adsorption. In our previous work,²⁴ extensive analysis with density functional theory calculations shed light on the underlying mechanism of selectivity enhancement: the IL anions dominate the adsorption and selectivity performance of these systems and their presence increases the overall affinity of the strongly sorbed species of a mixture, such as CO₂. In more detail, estimation of isosteric heats of adsorption and of binding energies revealed that this behavior is mainly attributed to the increased electrostatic interactions of CO₂ with the anions ([Tf₂N⁻]). In the case of CH₄ and N₂, as the number of IL molecules increases, the presence of ILs do not favor sorption and imposes a pore limiting factor. This behavior of ILs is governed by the bulky cations ([bmim⁺]): as the IL composition increases, the bulk of the cations gradually decreases the available pore space for CH₄ and N₂. The present results follow this pattern. In the case of CO₂, the presence of ILs stops being favorable when the IL composition reaches a critical value. At this point, the strong electrostatic interactions compete with the decreasing available pore volume. As the decreasing available pore volume dominates, the CO₂ capacity starts diminishing, and the CO₂ selectivity over CH₄ and N₂ drops.

In an effort to further assess the IL@ZIFs' selectivity, we use a parameter that we recently introduced, serving as a performance predictor in IL@ZIF systems. The so-called available pore volume (APV) is the ratio of the IL@ZIF's pore volume at a given IL composition to the pore volume of a given pristine ZIF analogue (the initial pore volume, before the ILs are inserted). The APV expresses the change of available space in the cages and depends on both the number of IL molecules and their density in a given composition.

In Fig. 6, the selectivity for the two mixtures of all the IL@ZIFs as a function of the APV is shown. The red curves correspond to a seven order polynomial fitted to the corresponding data. The results demonstrate that the APV can successfully describe the performance of the IL@ZIFs. A common trend is formed across the data points, which facilitates the assessment of the performance based on an individual structural characteristic, regardless of the metal variant or the gas mixture of interest. In both figures, the maximum performance corresponds to an APV of approx. 0.1. This is a rather interesting outcome, since this value is almost equal to the value we determined in a previous study, where we investigated the effect of varying the IL type in IL@ZIF-8 systems.²⁴ Thus, we can state that independent of the metal type or IL type, the peak performance of the IL@ZIFs is observed to be close to an APV value of 0.1. Thus, the APV facilitates the determination of the required number of IL molecules for maximum selectivity performance and can act as a guide for further developing these materials. Values for the data points in Fig. 6 are provided in Table S16,† along with the number of encapsulated IL pairs in each ZIF-8 metal analogue.

We close the investigation by including in our performance assessment the various conditions of binary adsorption. We chose IL@CdIF-1 as the analogue for the investigation. The IL composition in CdIF-1 corresponds to the highest selectivity (0.42). We consider three pressure values (0.5, 1 and 5 bars) and 50:50 and 10:90 volume ratios for CO₂/CH₄ and CO₂/N₂, respectively, as the conditions of industrial interest.

The results in Fig. 7 highlight two things: first, the ideal selectivity deviates from the one estimated in the binary mixture. This becomes more evident in the case of CO₂/N₂ (Fig. 7(b)), where CO₂ has a low concentration in the mixture. The impact of pressure on the performance of the IL@ZIFs is clear. The performance is high at low pressure and drops as the pressure increases. This is due to the very steep CO₂ adsorption isotherm, which favors the CO₂ selectivity at low pressures. Moreover, the impact of pressure is attributed to the gradual decrease of favorable adsorption sites for both species as pressure increases, and the adsorption is governed by the interactions of guest molecules with the non-selective parts of the hybrid material.

Metal replacement shows a significant potential as a way to tailor the adsorption-based properties of IL@ZIFs. Regarding the CO₂ storage, the approach is rather straightforward: a lighter metal, such as Be, will reduce the overall weight of the material. This result and the creation of new CO₂ sorption sites in the presence of encapsulated ILs will increase dramatically the CO₂ capacity in terms of sorbed CO₂ per weight of the material, as shown in both Fig. 3 and 4. Other metal candidates that will reduce the framework's overall weight are Fe (with an atomic mass of 55.85 amu), Mn (54.94 amu) and Mg (24.3 amu). Nevertheless, selecting the proper metal to enhance CO₂ selectivity is more complicated. This is due to the combined effects of reduced pore volume and favorable CO₂ sorption, which act competitively. IL molecules create favorable CO₂ sites and act as blocking entities for CH₄ and N₂, at the same time, which drives the selectivity enhancement. However, their increasing presence starts to act against CO₂ sorption at relatively high IL compositions.

This is easier to be realized by the assessment of the performance with the use of the APV. There is an optimal value for this parameter that ensures the highest possible CO₂ selectivity (over CH₄ or N₂) and highlights the effect of the available pore volume as the governing factor for the separation of species in these materials. Cd-, Co- and Be-based IL@ZIF analogues ensure a performance closer to the selectivity vs. the APV peak. What makes them better can be answered by a mere speculation, based on the set of results that we have: CdIF-1 has larger pores than ZIF-8, ensuring a larger number of encapsulated ILs and the creation of a larger number of preferable CO₂ sorption sites. On the other hand, ZIF-67 (Co-based) and BeIF-1 demonstrate an already enhanced selectivity in their pristine state. The packing of ILs in their cages acts cooperatively with their inherent higher CO₂ selectivity and their performance is further enhanced.

Although we consider the APV as a useful parameter, larger screening of metals is needed in order to provide a clear correlation and a trend that will explain the effect of metal substitution on the performance of IL@ZIFs.

4. Conclusions

IL@ZIFs are hybrid systems that combine the benefits of both ILs and ZIFs and are considered as highly promising materials for the development of CO₂-selective sorbents. In this work, we reported for the first time IL@ZIF-8 analogues designed by varying the metal center of the framework. We used ZIF-8 as our base material, in which the original Zn²⁺ center was replaced with Co, Cd and Be, resulting in ZIF-8 metal analogues, namely ZIF-67, CdIF-1 and BeIF-1, respectively. ZIF-8 and the metal analogues were computationally reconstructed and their CO₂/CH₄ and CO₂/N₂ sorption-based selectivity performance was studied with the use of Monte Carlo calculations and force fields developed by our group. Our results showed that the metal alteration affects the performance of the ZIF-8 analogues. ZIF-8 and the three metal analogues were then used to produce IL@ZIF-8 variants, by the encapsulation of [bmim⁺][Tf₂N⁻] in their cages. Subsequently, we simulated the sorption of CO₂, N₂ and CH₄ as a function of the IL composition. Our results show that IL@CdIF-1 and IL@BeIF-1 exhibit an exceptionally high selectivity for both mixtures. Moreover, our analysis reveals that the selectivity of all four cases increases as the IL composition increases. However, at higher IL composition, there is a strong trade-off between the increasing CO₂ affinity with the pore walls due to increasing CO₂–anion interactions and the gradually decreasing available pore space due to the bulky cations. This dictates an optimum number of IL molecules that must be inserted in the cages, which is different for each ZIF-8 metal analogue. The assessment and preparation of such materials is made easier by adopting the available pore volume (APV). For both mixtures and for all four IL@ZIFs, the selectivity is at maximum when the APV is approx. 0.1. Future work will focus on ZIFs with metal centers that are lighter or produce smaller pore volumes. Not only this would reduce the volume and number of IL molecules used (minimize structure weight), but also increase CO₂ adsorption and working capacity which are as important as the selectivity.

Conflicts of interest

No potential conflict of interest was reported by the authors.

Acknowledgements

This publication was made possible thanks to financial support from Texas A&M University at Qatar through the Responsive Research Seed Grant program and an award from the Qatar National Research Fund (a member of Qatar Foundation) with scholarship number GSRA5-1-0513-18051. The statements made herein are solely the responsibility of the authors. We are grateful to the High Performance Computing Center of Texas A&M University at Qatar for generous resource allocation.

References

1. M. Z. Jacobson, Review of Solutions to Global Warming, Air Pollution, and Energy Security, Energy Environ. Sci., 2009, 2, 148–173 RSC .
2. M. Oschatz and M. Antonietti, A Search for Selectivity to Enable CO2 Capture with Porous Adsorbents, Energy Environ. Sci., 2018, 11, 57–70 RSC .
3. M. Bui, C. S. Adjiman, A. Bardow, E. J. Anthony, A. Boston, S. Brown, P. S. Fennell, S. Fuss, A. Galindo and L. A. Hackett, et al., Carbon Capture and Storage (CCS): The Way Forward, Energy Environ. Sci., 2018, 11, 1062–1176 RSC .
4. R. L. Siegelman, P. J. Milner, E. J. Kim, S. C. Weston and J. R. Long, Challenges and Opportunities for Adsorption-Based CO2 Capture from Natural Gas Combined Cycle Emissions, Energy Environ. Sci., 2019, 12, 2161–2173 RSC .
5. C. A. Trickett, A. Helal, B. A. Al-Maythalony, Z. H. Yamani, K. E. Cordova and O. M. Yaghi, The Chemistry of Metal–Organic Frameworks for CO2 Capture, Regeneration and Conversion, Nat. Rev. Mater., 2017, 2, 17045 CrossRef CAS .
6. K. Adil, Y. Belmabkhout, R. S. Pillai, A. Cadiau, P. M. Bhatt, A. H. Assen, G. Maurin and M. Eddaoudi, Gas/Vapour Separation Using Ultra-Microporous Metal–Organic Frameworks: Insights into the Structure/Separation Relationship, Chem. Soc. Rev., 2017, 46, 3402–3430 RSC .
7. P. Krokidas, S. Moncho, E. N. Brothers, M. Castier, H. K. Jeong and I. G. Economou, On the Efficient Separation of Gas Mixtures with the Mixed-Linker Zeolitic-Imidazolate Framework-7-8, ACS Appl. Mater. Interfaces, 2018, 10, 39631–39644 CrossRef CAS .
8. J. Yu, L.-H. Xie, J.-R. Li, Y. Ma, J. M. Seminario and P. B. Balbuena, CO₂ Capture and Separations Using MOFs: Computational and Experimental Studies, Chem. Rev., 2017, 117, 9674–9754 CrossRef CAS PubMed .
9. K. S. Park, Z. Ni, A. P. Cote, J. Y. Choi, R. Huang, F. J. Uribe-Romo, H. K. Chae, M. O'Keeffe and O. M. Yaghi, Exceptional Chemical and Thermal Stability of Zeolitic Imidazolate Frameworks, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 10186–10191 CrossRef CAS PubMed .
10. C. Zhang and W. J. Koros, Zeolitic Imidazolate Framework-Enabled Membranes: Challenges and Opportunities, J. Phys. Chem. Lett., 2015, 6, 3841–3849 CrossRef CAS PubMed .
11. Y. Ban, Y. Li, Y. Peng, H. Jin, W. Jiao, X. Liu and W. Yang, Metal-Substituted Zeolitic Imidazolate Framework ZIF-108: Gas-Sorption and Membrane-Separation Properties, Chem. – Eur. J., 2014, 20, 11402–11409 CrossRef CAS PubMed .
12. P. Krokidas, S. Moncho, M. Castier, E. N. Brothers and I. G. Economou, Tailoring the Gas Separation Efficiency of Metal Organic Framework ZIF-8 through Metal Substitution: A Computational Study, Phys. Chem. Chem. Phys., 2018, 20, 4879–4892 RSC .
13. G. Chaplais, G. Fraux, J. L. Paillaud, C. Marichal, H. Nouali, A. H. Fuchs, F. X. Coudert and J. Patarin, Impacts of the Imidazolate Linker Substitution (CH 3, Cl or Br) on the Structural and Adsorptive Properties of ZIF-8, J. Phys. Chem. C, 2018, 122, 26945–26955 CrossRef CAS .
14. Q. Hou, Y. Wu, S. Zhou, Y. Wei, J. Caro and H. Wang, Ultra-Tuning of the Aperture Size in Stiffened ZIF-8_Cm Frameworks with Mixed-Linker Strategy for Enhanced CO2/CH4 Separation, Angew. Chem., Int. Ed., 2019, 58, 327–331 CrossRef CAS PubMed .
15. B. Tam and O. Yazaydin, Design of Electric Field Controlled Molecular Gates Mounted on Metal-Organic Frameworks, J. Mater. Chem. A, 2017, 5, 8690–8696 RSC .
16. D. J. Babu, G. He, J. Hao, M. T. Vahdat, P. A. Schouwink, M. Mensi and K. V. Agrawal, Restricting Lattice Flexibility in Polycrystalline Metal-Organic Framework Membranes for Carbon Capture, Adv. Mater., 2019, 31, 1900855 CrossRef PubMed .
17. Y. Ban, Z. Li, Y. Li, Y. Peng, H. Jin, W. Jiao, A. Guo, P. Wang, Q. Yang and C. Zhong, et al., Confinement of Ionic Liquids in Nanocages: Tailoring the Molecular Sieving Properties of ZIF-8 for Membrane-Based CO2 Capture, Angew. Chem., Int. Ed., 2015, 54, 15483–15487 CrossRef CAS PubMed .
18. Y. Lan, T. Yan, M. Tong and C. Zhong, Large-Scale Computational Assembly of Ionic Liquid/MOF Composites: Synergistic Effect in the Wire-Tube Conformation for Efficient CO 2/CH 4 Separation, J. Mater. Chem. A, 2019, 7, 12556–12564 RSC .
19. F. P. Kinik, C. Altintas, V. Balci, B. Koyuturk, A. Uzun and S. Keskin, [BMIM][PF6] Incorporation Doubles CO2 Selectivity of ZIF-8: Elucidation of Interactions and Their Consequences on Performance, ACS Appl. Mater. Interfaces, 2016, 8, 30992–31005 CrossRef CAS PubMed .
20. B. Koyuturk, C. Altintas, F. P. Kinik, S. Keskin and A. Uzun, Improving Gas Separation Performance of ZIF-8 by [BMIM][BF 4] Incorporation: Interactions and Their Consequences on Performance, J. Phys. Chem. C, 2017, 121, 10370–10381 CrossRef CAS .
21. M. Zeeshan, S. Keskin and A. Uzun, Enhancing CO2/CH4 and CO2/N2 Separation Performances of ZIF-8 by Post-Synthesis Modification with [BMIM][SCN], Polyhedron, 2018, 155, 485–492 CrossRef CAS .
22. A. Mohamed, P. Krokidas and I. G. Economou, CO2 Selective Metal Organic Framework ZIF-8 Modified through Ionic Liquid Encapsulation: A Computational Study, J. Comput. Sci., 2018, 27, 183–191 CrossRef .
23. M. Maurya and J. K. Singh, Effect of Ionic Liquid Impregnation in Highly Water-Stable Metal-Organic Frameworks, Covalent Organic Frameworks, and Carbon-Based Adsorbents for Post-Combustion Flue Gas Treatment, Energy Fuels, 2019, 33, 3421–3428 CrossRef CAS .
24. A. M. O. Mohamed, S. Moncho, P. Krokidas, K. Kakosimos, E. N. Brothers and I. G. Economou, Computational Investigation of the Performance of ZIF-8 with Encapsulated Ionic Liquids towards CO2 Capture, Mol. Phys., 2019, 117, 3791–3805 CrossRef CAS .
25. J. K. Shah, E. Marin-Rimoldi, R. G. Mullen, B. P. Keene, S. Khan, A. S. Paluch, N. Rai, L. L. Romanielo, T. W. Rosch and B. Yoo, et al., Cassandra: An Open Source Monte Carlo Package for Molecular Simulation, J. Comput. Chem., 2017, 38, 1727–1739 CrossRef CAS PubMed .
26. P. Krokidas, M. Castier, S. Moncho, E. Brothers and I. G. Economou, Molecular Simulation Studies of the Diffusion of Methane, Ethane, Propane, and Propylene in ZIF-8, J. Phys. Chem. C, 2015, 119, 27028–27037 CrossRef CAS .
27. P. Krokidas, M. Castier, S. Moncho, D. N. Sredojevic, E. N. Brothers, H. T. Kwon, H. K. Jeong, J. S. Lee and I. G. Economou, ZIF-67 Framework: A Promising New Candidate for Propylene/Propane Separation. Experimental Data and Molecular Simulations, J. Phys. Chem. C, 2016, 120, 8116–8124 CrossRef CAS .
28. E. Androulaki, N. Vergadou, J. Ramos and I. G. Economou, Structure, Thermodynamic and Transport Properties of Imidazolium-Based Bis(Trifluoromethylsulfonyl)Imide Ionic Liquids from Molecular Dynamics Simulations, Mol. Phys., 2012, 110, 1139–1152 CrossRef CAS .
29. M. G. Martin and J. I. Siepmann, Transferable Potentials for Phase Equilibria. 1. United-Atom Description of n-Alkanes, J. Phys. Chem. B, 1998, 102, 2569–2577 CrossRef CAS .
30. W. L. Jorgensen, D. S. Maxwell and J. Tirado-Rives, Development and Testing of the OPLS All-Atom Force Field on Conformational Energetics and Properties of Organic Liquids, J. Am. Chem. Soc., 1996, 118, 11225–11236 CrossRef CAS .
31. J. Ilja Siepmann and D. Frenkel, Conflgurational Bias Monte Carlo: A New Sampling Scheme for Flexible Chains, Mol. Phys., 1992, 75, 59–70 CrossRef .
32. E. J. Maginn, J. K. Shah, E. Marinrimoldi, S. Khan, N. Rai, T. Rosch and A. Paluch Computational Atomistic Simulation Software At Notre Dame for Research Advances, User Manual 1.2, 2014 .
33. Materials Studio, Dassault Systèmes BIOVIA, San Diego, 2017.
34. F. L. Smith and A. H. Harvey, Avoid Common Pitfalls When Using Henry's Law, Chem. Eng. Prog., 2007, 33–39 CAS .
35. R. Sander, Compilation of Henry's Law Constants (Version 4.0) for Water as Solvent, Atmos. Chem. Phys., 2015, 15, 4399–4981 CrossRef CAS .

---

Footnotes

† Electronic supplementary information (ESI) available: Force field parameters for the ZIF-8 metal analogues, IL and CO2, N2 and CH4. Data points for plots in Fig. 3–6. Information on the Monte Carlo simulation protocol for the preparation of the various IL@ZIFs and sorption simulations. Analysis of the effect of IL distribution on the convergence of sorption estimations. See DOI: 10.1039/d0me00021c

‡ Current address: Division of Sustainable Development, Hamad Bin Khalifa University, Education City, Doha, Qatar.

§ Current address: National Center for Scientific Research “Demokritos”, Institute of Nanoscience and Nanotechnology, Molecular Thermodynamics and Modelling of Materials Laboratory, GR – 153 10, Aghia Paraskevi Attikis, Greece.

This journal is © The Royal Society of Chemistry 2020