Highly Selective Separation of C 2 H 2 /CO 2 and C 2 H 2 /C 2 H 4 in an N-Rich Cage-Based Microporous Metal-Organic Framework

Abstract

The separation of acetylene (C₂H₂) from carbon dioxide (CO₂) and the purification of ethylene (C₂H₄) from C₂H₂ are quite essential processes for the chemical industry. However, these processes are challenging due to their similar physical properties, including molecule sizes and boiling points. Herein, we report an N-rich cage-based microporous metal-organic framework (MOF), [Cd₅(Tz)₉](NO₃) (termed as Cd-TZ, TZ stands for tetrazole), and its highly efficient separation of C₂H₂/CO₂ and C₂H₂/C₂H₄. Single-component gas adsorption isotherms reveal that Cd-TZ exhibits high C₂H₂ adsorption capacity (3.10 mmol g^-1 at 298 K and 1 bar). The N-rich cages in Cd-TZ can trap C₂H₂ with a higher isosteric heat of adsorption (40.8 kJ mol^-1) than CO₂ and C₂H₄ owing to the robust host-guest interactions between the noncoordinated N atoms and C₂H₂, which has been verified by molecular modeling studies. Cd-TZ shows a high IAST selectivity for C₂H₂/CO₂ (8.3) and C₂H₂/C₂H₄ (13.3). The breakthrough simulations confirm the potential for separating C₂H₂/CO₂ and the purification of C₂H₄ from C₂H₂.

1. Introduction

C₂H₂ and C₂H₄ are two of the most important chemical raw feedstocks for producing various commercial chemicals [1]. C₂H₂ is mainly obtained from coal or coal-derived coke with direct or indirect processes [2]. Thus, CO₂ is usually contained in the crude acetylene as an unavoidable impurity generally removed in an aqueous sodium hydroxide wash. C₂H₂ can also be purified by dissolving in organic solvents, such as N-methylpyrrolidone and dimethylformamide. C₂H₄ is typically produced by the pyrolysis of ethane gas or light naphtha, with a small amount of C₂H₂ generated. C₂H₂ in mixtures must be removed, reducing its content to below 5 ppm [3, 4] because of the effects of C₂H₂ in poisoning catalysis during the polymerization process of ethylene. Solvent extraction and partial hydrogenation are the main commercial techniques to separate trace C₂H₂ from C₂H₄. However, both the purification methods of C₂H₂ and C₂H₄ used in the chemical industry are with poor separation selectivity and are energy-intensive [5]. Alternatively, selective adsorption and separation processes based on porous sorbents are energy-efficient and sustainable.

In recent decades, MOFs [6–12] have been studied widely in separations and purifications for hydrocarbons [13–24]. However, the molecular sizes ( $3.3\times 3.3\times 5.7{\text{Å}}^3$ for C₂H₂, $3.2\times 3.3\times 5.4{\text{Å}}^3$ for CO₂, and $3.3\times 4.2\times 4.8{\text{Å}}^3$ for C₂H₄) and kinetic diameter (3.3 Å for C₂H₂, 3.3 Å for CO₂, and 4.2 Å for C₂H₄) of acetylene, carbon dioxide, and ethylene are very close [7, 25], making the efficient separation of C₂H₂/CO₂ and C₂H₂/C₂H₄ a challenge. Until now, MOFs [7, 26–31] with highly selective separation performance for both C₂H₂/CO₂ and C₂H₂/C₂H₄ are still rare. Currently, the reported MOF materials with record selectivity for C₂H₂/CO₂ and C₂H₂/C₂H₄ is an anion-pillared flexible MOF, UTSA-300a [26]. Because of the strong C–H···F hydrogen-bonding interaction between UTSA-300a and C₂H₂, C₂H₂ can easily open the pore and be adsorbed into it while the other gases are blocked out of the pore. Lee and co-workers [28] reported a cationic MOF, JCM-1, which possesses imidazolium and nitrate groups in the pore channel, exhibiting a higher selectivity and adsorption affinity for C₂H₂ than for CO₂ or C₂H₄. A broad strategy in MOFs to enhance the selectivity of C₂H₂/CO₂ and C₂H₂/C₂H₄ is to increase the polar groups or adsorption sites, such as fluorine/nitrogen/oxygen-containing functional groups, to strengthen the interaction with C₂H₂ over CO₂ or C₂H₄.

Herein, we report a microporous MOF [Cd₅(Tz)₉](NO₃) [32] with N-rich cavities for the preferential adsorption of C₂H₂ and the efficient separation for C₂H₂/CO₂ and C₂H₂/C₂H₄. At ambient conditions, Cd-TZ shows a higher adsorption capacity and isosteric heat of adsorption for C₂H₂ than for CO₂ and C₂H₄. Thus, Cd-TZ exhibits elevated C₂H₂/CO₂ and C₂H₂/C₂H₄ selectivity. The modeling studies reveal that the trapped C₂H₂ molecule is in the center of the cavity and strongly interacts with the surrounding N atoms. The simulated breakthrough curves demonstrate the effective separation of binary mixtures of C₂H₂/CO₂ and C₂H₂/C₂H₄.

2. Materials and Methods

2.1. Materials

All reagents and solvents were commercially available and, unless otherwise noted, were used without further purification. Cadmium nitrate tetrahydrate (Cd(NO₃)₂·4H₂O, AR) was purchased from Aladdin Reagent Co. Ltd. Ethyl tetrazole-5-carboxylate (>95%) was purchased from Bidepharm Co. Ltd.

High-purity CO₂ (99.999%), C₂H₄ (99.95%), and C₂H₂ (99.9%) were purchased from Qingdao Tianyuan Gas Co., Ltd. C₂H₂ was filtered through activated carbon to remove traces of acetone before use.

2.2. Synthesis of Cd-TZ

Cd-TZ was synthesized according to the methods reported in the literature [32]. A mixture of Cd(NO₃)₂·4H₂O and ethyl tetrazole-5-carboxylate was dissolved in 10 mL H₂O, then transferred to a 23 mL Teflon-lined autoclave and heated at 433 K for 72 h, followed by cooling to room temperature. Colorless crystals were collected by filtration and washed with H₂O and methanol. The activation process of the Cd-TZ sample was conducted under a vacuum at 423 K for 12 h.

2.3. Characterizations of Cd-TZ

Thermogravimetric analyses (TGA) were examined using a Netzsch STA 449C instrument under an N₂ atmosphere with a heating rate of 5 K min^-1. Powder X-ray diffraction (PXRD) data were performed on an X-ray diffractometer (Bruker D8 Adv., Germany) with Cu Kα radiation from 5 to 50° ( $\lambda =1.5406$ ) and a step size of 0.0167° in 2θ.

2.4. Adsorption Studies for CO₂, C₂H₂, and C₂H₄

The sample was degassed in a vacuum for 12 h at 423 K to remove guest molecules before adsorption. Equilibrium and kinetic adsorption experiments of CO₂, C₂H₂, and C₂H₄ were measured using a XEMIS magnetic suspension balance sorption analyzer (Hiden, UK) equipped with a circulating water bath at 273 K and 298 K, respectively.

2.5. Calculation of Brunauer-Emmett-Teller (BET) Surface Area and Langmuir Surface Area

Surface areas of Cd-TZ were calculated using the BET equation and Langmuir equation based on the CO₂ adsorption isotherm at 273 K.

BET surface area: $\begin{array}{}\text{(1)}& \frac{P}{n\left(P_0-P\right)}=\frac{1}{C{n}_m}+\frac{C-1}{C{n}_m}·\frac{P}{P_0},\text{(2)}& S_{\text{BET}}=n_m·A·{\sigma }_m.\end{array}$

Langmuir surface area: $\begin{array}{}\text{(3)}& \frac{P}{n}=\frac{P}{n_m}+\frac{1}{b{n}_m},\text{(4)}& S_{\text{Langmuir}}=n_m·A·{\sigma }_m,\end{array}$ where $P$ is the pressure (bar), $P_0$ is the saturated vapor pressure, $n$ is the adsorption amount under the corresponding pressure, $n_m$ is the amount related to the monolayer surface coverage, $C$ ( $C>0$ ) and $b$ are constants, ${\sigma }_m$ is the cross-sectional area of CO₂ ( $2.18\times {10}^{-19}{\text{m}}^2$ ), and $S_{\text{BET}}$ and $S_{\text{Langmuir}}$ are the calculated BET surface area and Langmuir surface area, respectively.

2.6. Calculation of Isosteric Enthalpy of Adsorption

Isosteric enthalpy of adsorption ( $Q_{\text{st}}$ ) was calculated from isotherms measured at 273 K and 298 K for CO₂, C₂H₂, and C₂H₄. The isotherms were firstly fit to a virial equation: $\begin{array}{}\text{(5)}& \ln P=\ln N+\frac{1}{T}\sum _{i=0}^m{a}_i{N}^i+\sum _{i=0}^n{b}_i{N}^i,\end{array}$ where $P$ is the pressure expressed in mbar, $N$ is the amount adsorbed in mmol g^-1, $T$ is the temperature in K, $a_i$ and $b_i$ are virial coefficients, and $m$ and $n$ represent the number of coefficients required to describe the isotherms adequately. To calculate $Q_{\text{st}}$ , the fitting parameters from the above equation were used for the following equation: $\begin{array}{}\text{(6)}& Q_{\text{st}}=-R\sum _{i=0}^m{a}_i{N}^i.\end{array}$

2.7. Ideal Adsorbed Solution Theory (IAST) Selectivity Calculation

The gas adsorption isotherms were firstly fitted to a dual-site Langmuir-Freundlich (DSLF) equation: $\begin{array}{}\text{(7)}& q=q_1\frac{b_1{p}^{n_1}}{1+b_1{p}^{n_1}}+q_2\frac{b_2{p}^{n_2}}{1+b_1{p}^{n_2}},\end{array}$ where $q$ (mmol g^-1) is the amount of adsorbed gas; $p$ (mbar) is the bulk gas phase pressure; $q_1$ (mmol g^-1) and $q_2$ (mmol g^-1) are the saturation capacities of sites 1 and 2, respectively; $b_1$ (mbar^-n1) and $b_2$ (mbar^-n2) are the affinity coefficients of site 1 and site 2, respectively; and $n_1$ and $n_2$ represent the deviations from an ideal homogeneous surface.

The adsorption selectivity is defined by $\begin{array}{}\text{(8)}& S_{\text{ads}}=\frac{x_1/y_1}{x_2/y_2},\end{array}$ where $x_i$ and $y_i$ are the mole fractions of component $i$ ( $i=1$ and 2) in the adsorbed and bulk phases, respectively.

2.8. Density-Functional Theory Calculations

All the geometry optimizations and binding energies were calculated by the periodic density functional theory (DFT) method using the DMol³ module. The host framework and the gas molecule were both regarded as rigid. The structures of the framework were first optimized. Then, guest gas molecules were introduced to the optimized framework, followed by a full structure relaxation. An isolated gas molecule placed in a supercell (with the same cell dimensions as the framework) was also relaxed as a reference to obtain the gas binding energy. The widely used generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) functional and the double numerical plus polarization (DNP) basis set, the Grimme method for DFT-D correction, and the DFT semicore pseudopots (DSPP) were used. The energy, force, and displacement convergence criteria were set as $1\times {10}^{-5}\text{Ha}$ , $2\times {10}^{-3}\text{Ha}$ , and $5\times {10}^{-3}\text{Å}$ , respectively. The following equations then calculated the static binding energy (at $T=0\text{K}$ ): $\begin{array}{}\text{(9)}& \Delta E_{\text{binding}}=E_{\text{MOF}}+E_{\text{gas}}-E_{\left(\text{MOF}+\text{gas}\right)},\end{array}$ where $\Delta E_{\text{binding}}$ is the static binding energy between the MOF and the gas molecule and $E_{\text{MOF}}$ , $E_{\text{gas}}$ , and $E_{\left(\text{MOF}+\text{gas}\right)}$ are the energies of the gas-free MOF, the free gas molecule, and the MOF-gas system, respectively.

2.9. Breakthrough Simulation

Breakthrough simulations were performed using the 3P-Sim software. In the simulated separation experiment, the Cd-TZ (1.0 g) was packed into the column with a length of 5 cm and an inner diameter of 0.45 cm. And the simulated operating condition is under 1 bar at 298 K with a continuous gas (C₂H₂/CO₂ (50/50, $v/v$ ) or C₂H₂/C₂H₄ (50/50, $v/v$ )) flow of 2 mL (STP) min^-1.

3. Results and Discussion

As shown in Figures 1(a) and 1(b), each Cd (II) atom in the framework is surrounded by different tetrazole ligands and coordinated with six nitrogen atoms to form an octahedral coordination structure, resulting in a cationic three-dimensional framework [Cd₅(Tz)₉]⁺ which is balanced by NO₃ ^-. Cd (II) atoms interconnect such six second-building units to form one-dimensional straight pore channels (4.6 Å), which have large cavities (5.3 Å) parallel to each other along the $c$ axis (Figures 1(c) and 1(d)). The surface of the cavities contains many uncoordinated tetrazole nitrogen atoms, which is conducive to capturing C₂H₂ molecules. Cd-TZ was successfully prepared by hydrothermal reaction at 433 K, and the phase purity of the material was identified by PXRD patterns which nicely match the calculated patterns from single crystal data (Figure 2(a)). To investigate the thermal stability of Cd-TZ, the synthesized Cd-TZ samples were subjected to TGA under a nitrogen atmosphere, as shown in Figure 2(b). The TGA curves show that the Cd-TZ sample lost most of the free water molecules in the pore channels as the temperature increased to 98°C. When the temperature exceeded 300°C, the sample lost weight rapidly. This phenomenon indicates that Cd-TZ has good thermal stability under a nitrogen atmosphere and can maintain the structure over a wide temperature range (<300°C). The specific surface area of Cd-TZ was characterized using CO₂ adsorption experiments performed at 273 K (Figure S1a). And the BET surface area (Figure S1b) and Langmuir surface area (Figure S2) were calculated to be 388 m² g^-1 and 403 m² g^-1, respectively.

Figure 1

The framework of Cd-TZ without hydrogen atoms and solvent molecules viewed along the $a$ axis (a) and $b$ axis (b) (Cd, C, N, and O are represented by light yellow, gray, lavender, and red, respectively). The pore structure of Cd-TZ without hydrogen atoms and solvent molecules viewed along the $a$ axis (c) and $b$ axis (d).

(b) (c) (d)

Figure 2

(a) PXRD patterns of as-synthesized Cd-TZ and activated Cd-TZ and (b) thermogravimetry curves of Cd-TZ.

(b)

Considering the suitable pore size and the presence of abundant uncoordinated nitrogen atoms in the pore surface of Cd-TZ, we were intrigued to explore its potential to capture C₂H₂ and, further, to separate C₂H₂/CO₂ and C₂H₂/C₂H₄. Therefore, single-component adsorption-desorption isotherms of C₂H₂, CO₂, and C₂H₄ were collected at 273 K and 298 K (Figures 3(a) and 3(b)). The adsorption capacity of C₂H₂ is higher than those of CO₂ and C₂H₄ in Cd-TZ under the same conditions. The C₂H₂ capacity is 3.10 mmol g^-1 at 298 K and 1 bar, higher than the capacity of CO₂ (2.07 mmol g^-1) and C₂H₄ (1.68 mmol g^-1) and outperforming many reported MOFs, such as ZU-62-Ni (3.0 mmol g^-1) [33], NKMOF-1-Ni (2.7 mmol g^-1) [34], Zn(ad)(int) (2.32 mmol g^-1) [35], CPL-1-NH₂ (1.84 mmol g^-1) [36], and Zn-FBA (1.03 mmol g^-1) [37]. Furthermore, the C₂H₂ adsorption isotherms rise rapidly in the low-pressure region at 273 K and 298 K. At 298 K and 0.01 bar, the uptake ratio of C₂H₂/CO₂ and C₂H₂/C₂H₄ in Cd-TZ is 1.8 and 3.0, respectively, indicating the stronger interaction between C₂H₂ and Cd-TZ. The isosteric enthalpies of adsorption for C₂H₂, CO₂, and C₂H₄ were calculated based on fitting using the virial method (Figure S3-S5) to evaluate the interaction between the framework and the guest molecule. As shown in Figure 3(c), the calculated value of $Q_{\text{st}}$ for C₂H₂, CO₂, and C₂H₄ at zero loading is 40.8 kJ mol^-1, 35.6 kJ mol^-1, and 31.8 kJ mol^-1, respectively. Furthermore, the $Q_{\text{st}}$ value of Cd-TZ for C₂H₂ is higher than those for CO₂ and C₂H₄ over the entire pressure range from 0 to 1 bar. All the results demonstrate that Cd-TZ prefers to adsorb C₂H₂ and has a stronger affinity for C₂H₂ than CO₂ and C₂H₄, which prompts us to explore the separation performance of Cd-TZ for the binary gas of C₂H₂/CO₂ and C₂H₂/C₂H₄.

Figure 3

Single-component adsorption-desorption isotherms of C₂H₂, CO₂, and C₂H₄ in Cd-TZ at (a) 273 K and (b) 298 K. (c) The isosteric enthalpies of adsorption for C₂H₂, CO₂, and C₂H₄ in Cd-TZ.

(b) (c)

The adsorption selectivity for C₂H₂/CO₂ (50/50, $v/v$ ) and C₂H₂/C₂H₄ (50/50, 1/99, $v/v$ ) was calculated based on the IAST model based on fitting using the DSLF equation (Figure S6-S8). The selectivity of Cd-TZ for C₂H₂/CO₂ (50/50, $v/v$ ) is 8.3 at 298 K and 1.0 bar (Figure S9), which is compared to the reported UTSA-74a (9.0) [38], and superior to some of the benchmark MOFs like SIFSIX-21-Ni (7.8) [39], TIFSIX-2-Cu-i (6.5) [40], FJU-90a (4.3) [41], and ZrT-1-tetrazl (2.8) [42] under similar conditions. As shown in Figure S10, the selectivity of Cd-TZ for the binary mixtures of C₂H₂/C₂H₄ (50/50, $v/v$ ) is up to 20.1 at 273 K and 1 bar, with the temperature increasing to 298 K, and the selectivity is about 13.3, which surpasses most of the reported MOFs, such as cPAF-28 (12.2) [43], MUF-17 (8.73) [30], NbU-1 (5.9) [44], CuTiF₆-TPPY (5.47) [45], and Fe-MOF-74 (2.1) [46]. Consider that C₂H₂ is a minor impurity in the crude C₂H₄ production by the cracking of ethane. Therefore, the IAST selectivity of C₂H₂/C₂H₄ (1/99, $v/v$ ) was also calculated in addition to that of C₂H₂/C₂H₄ (50/50, $v/v$ ). The calculated IAST selectivity value for the binary mixtures of C₂H₂/C₂H₄ (1/99, $v/v$ ) is 7.3 higher than those of SIFSIX-3-Ni (5.0) [47] and NUM-11a (1.65) [48]. These calculation results further indicate that Cd-TZ is a potential porous material in separating C₂H₂/CO₂ and removing C₂H₂ from C₂H₄.

The optimal adsorption sites of C₂H₂, C₂H₄, and CO₂ in the framework of Cd-TZ were elucidated by DFT calculations. The calculated static binding energy of C₂H₂ in Cd-TZ is 75.1 kJ mol^-1, much higher than those of C₂H₄ with 54.4 kJ mol^-1 and CO₂ with 35.6 kJ mol^-1. The lowest-energy binding configuration of C₂H₂, C₂H₄, and CO₂ in Cd-TZ is shown in Figure 4, indicating that the adsorbed C₂H₂ molecule is trapped in the center of the cavity and strongly interacts with the surrounding N atoms, which are highly electronegative. Figure S11 shows that the optimized distance between one H atom of C₂H₂ and the noncoordinated N atom in the cavity is 2.14 Å, which is smaller than the sum of the van der Waals radius of the H atom (1.20 Å) and N atom (1.55 Å). Moreover, the calculated distances between C atoms of C₂H₂ and H atoms of tetrazole are 2.72 Å and 2.74 Å, respectively, shorter than the sum of the van der Waals radius of H and C atoms, 2.90 Å. These results indicate the robust affinity of the framework to the C₂H₂ molecules. C₂H₄ has a larger size and planar configuration. Thus, only one end of the =CH₂ group is adsorbed in the cavity with side-on orientation and forms a weaker interaction with the surrounding tetrazoles (Figure S12). There are four C=C-H···N dipolar interactions (2.89 Å, 2.92 Å, 3.10 Å, and 3.10 Å) observed between C₂H₄ and the tetrazoles in the cavity. And the calculated H-C=C···H distances between C₂H₄ and the tetrazoles in the cavity are 2.70~2.89 Å (2.70 Å, 2.73 Å, 2.87 Å, and 2.89 Å). As shown in Figure S13, CO₂ has two weak C_CO2···N interactions and three weak O_CO2···H interactions with the tetrazoles. The calculated distance of C_CO2···N is 3.40 Å and 3.47 Å, while the estimated distance of O_CO2···H is 2.85 Å, 3.41 Å, and 3.49 Å, indicating the weak interaction between CO₂ and the framework. The simulation calculations confirm that the adsorption affinity of C₂H₂ is much stronger than C₂H₄ and CO₂, consistent with the experimental results.

Figure 4

Comparison of the optimal (a) C₂H₂, (b) C₂H₄, and (c) CO₂ adsorption sites observed by DFT calculations along the $b$ axis. Cd, C, N, and H in Cd-TZ are represented by dark blue, bright grey, lavender, and white, respectively. C in C₂H₂ and C₂H₄ is represented by bright yellow.

(b) (c)

The separation performance of Cd-TZ was investigated by breakthrough simulations with the binary gas mixture of C₂H₂/CO₂ (50/50, $v/v$ ) and C₂H₂/C₂H₄ (50/50, $v/v$ ) at 298 K and 1 bar. As shown in Figure 5(a), CO₂ gas was first eluted at 31 min from the column, whereas C₂H₂ was retained in the column until 50 min, demonstrating a stronger binding affinity of the framework to C₂H₂ molecules. Figure 5(b) shows that Cd-TZ can realize the complete separation of C₂H₂/C₂H₄ (50/50, $v/v$ ). The C₂H₄ broke through from the column after 30 minutes, and no C₂H₂ was eluted out before 53 minutes, indicating that high-purity C₂H₄ could be obtained with the framework. All the above results demonstrate the great potential of Cd-TZ in the separation of C₂H₂/CO₂ and C₂H₂/C₂H₄ in practice.

Figure 5

Simulated breakthrough curves for (a) C₂H₂/CO₂ (50/50, $v/v$ ) and (b) C₂H₂/C₂H₄ (50/50, $v/v$ ) at 298 K and 1 bar.

(b)

4. Conclusions

In summary, we have investigated the N-rich cage-based microporous metal-organic framework, Cd-TZ, showing a stronger affinity for C₂H₂ than CO₂ and C₂H₄, and efficiently separating binary mixtures of C₂H₂/CO₂ and C₂H₂/C₂H₄. Owing to the abundant uncoordinated nitrogen atoms on the pore surface and the cavities, Cd-TZ exhibits a high capacity for C₂H₂ (3.10 mmol g^-1 at 298 K and 1 bar). The calculated IAST selectivity of Cd-TZ for C₂H₂/CO₂ and C₂H₂/C₂H₄ is 8.3 and 13.3, respectively. The breakthrough simulation results well confirm the separation performance of Cd-TZ. In addition, the preferential binding of C₂H₂ over CO₂ and C₂H₄ is clearly demonstrated by DFT calculations. This study provides an exquisite example of MOF possessing abundant electronegative nitrogen sites on the pore surfaces and the cavities for the challenging separation of C₂H₂/CO₂ and C₂H₂/C₂H₄.