Understanding the Enhanced Separation Mechanism of C₂H₄/C₂H₆ at Low Pressure by HKUST−1

Abstract

The production of ethylene (C₂H₄) is typically accompanied by the formation of impurities like ethane (C₂H₆), making the separation of C₂H₄ and C₂H₆ crucial in industrial processes. Here, we investigated the S-shaped adsorption phenomenon of C₂H₆ on the metal–organic framework HKUST−1. The virial equation is used to fit the C₂H₆ and C₂H₄ adsorption isotherms under low coverage. The results showed that the repulsion energy between neighboring C₂H₆ molecules was significantly higher than that between neighboring C₂H₄ molecules, which was an important reason for the lower adsorption of C₂H₆ by HKUST−1 at low coverage. As more molecules are adsorbed, gas molecules aggregate within pores, leading to more hydrogen bonds formed between HKUST−1 and larger-sized C₂H₆ under high coverage conditions. This phenomenon plays a crucial role in the S-shaped adsorption behavior of HKUST−1 on C₂H₆. Additionally, this unique adsorption behavior allows for the efficient separation of C₂H₄/C₂H₆ mixtures at low pressures. The ideal adsorbed solution theory (IAST) selectivity of HKUST−1 for C₂H₄/C₂H₆ mixtures was 3.78 at 283 K and 1 bar, but increased significantly to 7.53 under low pressure. This unique mechanism provides a theoretical basis for the low-pressure separation of C₂H₄/C₂H₆ by HKUST−1 and establishes a solid foundation for future practical research applications.

1. Introduction

Ethylene (C₂H₄) is a crucial fundamental feedstock in the organic chemical industry, with an extensive array of downstream derivatives encompassing polyvinyl chloride, synthetic rubber, and other products [1,2]. Within the petrochemical industry, C₂H₄ is predominantly obtained through thermal cracking of ethane (C₂H₆), inevitably resulting in the presence of C₂H₆ impurities in the crude products. Currently, cryogenic distillation is the primary purification process for separating C₂H₄ from C₂H₄/C₂H₆ mixtures [3]. The cryogenic distillation process, however, is characterized by its high energy consumption and operation under harsh conditions, which imposes significant demands on equipment [4]. Therefore, it is imperative to explore a more cost- and energy-efficient approach for separating C₂H₄/C₂H₆ mixtures. In comparison, adsorption separation utilizing porous materials as physisorptive agents represents a highly efficient technology for achieving this goal [5,6,7,8,9].

Recently, numerous studies have documented the utilization of metal–organic frameworks (MOFs) in hydrocarbon separation, owing to their customizable structure and porosity, as well as their inherent porous nature [10,11,12]. The coordination bonds between organic linkers and metal ions/clusters give rise to the formation of MOFs, which represent a novel class of adsorbent materials. MOFs exhibit diverse functionalities, as well as highly adjustable pore structures and chemical properties, owing to the facile regulation or substitution of both metal ions/clusters and ligands. Among these, polymeric copper (II) benzene-1,3,5-tricarboxylate (HKUST−1), characterized by its exceptionally low mass density and extensive surface area, exhibits remarkable advantages as an adsorbent for gas adsorption [13] and separation [14,15]. HKUST−1 represents a pioneering MOF material employed for the separation of C₂H₄/C₂H₆. The adsorption isotherms for C₂H₆ and C₂H₄ on HKUST−1 at 295 K were investigated by Semanscin et al. in 2002. The results revealed a lower adsorption capacity of C₂H₆ compared to C₂H₄, particularly within the low-pressure range [16]. The interactions between C₂H₄ and C₂H₆ with HKUST−1 were investigated by Bhatia et al. through theoretical calculations, revealing that electrostatic interactions can induce selective adsorption of light hydrocarbons under low pressures [17]. Typically, MOFs that exhibit S-shaped adsorption profiles are usually flexible MOFs, with framework flexibility dependent on the structure of the MOFs and/or external stimuli [18,19,20]. Despite over two decades of extensive research on the highly crystalline material HKUST−1 [21,22,23], the mechanism behind the S-shaped adsorption phenomenon in HKUST−1 and its superior separation efficiency for C₂H₄/C₂H₆ at low pressures compared to high pressures remains unclear.

In order to investigate the underlying cause of the S-shaped adsorption phenomenon exhibited by HKUST−1 towards C₂H₆ and more efficient separation of C₂H₄/C₂H₆ under low-pressure conditions, we experimentally conducted measurements of the adsorption isotherms of C₂H₆ and C₂H₄ on HKUST−1, respectively, and accurately calculated the enthalpy of adsorption (Q_st) for C₂H₆ and C₂H₄ using both the virial equation and Clausius–Clapeyron equation. Under conditions of low coverage, there is a higher intermolecular repulsion between neighboring C₂H₆ molecules compared to neighboring C₂H₄ molecules, resulting in a lower adsorption capacity for C₂H₆ on HKUST−1 under low-pressure conditions. At high coverage, gas molecules aggregate within the pores, and C₂H₆ molecules with larger sizes can form more hydrogen bonds to the O atoms in HKUST−1, thereby increasing the adsorption capacity of C₂H₆. This elucidates the underlying reason for the S-shaped adsorption phenomenon exhibited by HKUST−1 for C₂H₆. In addition, this unique S-type adsorption phenomenon of C₂H₆ by HKUST−1 facilitates the separation of C₂H₄/C₂H₆ gas mixtures at low pressure. The ideal adsorbed solution theory (IAST) selectivity of HKUST−1 for C₂H₄/C₂H₆ at low pressure and 283 K reached 7.53, exhibiting a significant enhancement compared to the value obtained at 1 bar (3.78).

2. Materials and Methods

2.1. Material Synthesis

The reagents and solvents utilized in this study are readily available in the market without requiring additional purification. HKUST−1 was synthesized following a modified version of the literature method. In a typical synthesis, Cu(NO₃)₂·3H₂O (716 mg, 3 mmol) and 1,3,5-benzenetricarboxylic acid (421 mg, 2 mmol) were dissolved in a mixture of deionized water (6 mL) and ethanol (6 mL). Subsequently, the mixture solution was stirred at room temperature for 30 min before being subjected to the reaction for one day at 383 K. Finally, the resultant blue crystals were obtained through filtration and thoroughly washed with methanol.

2.2. Gas Sorption Measurements

The nitrogen adsorption isotherms at 77 K were determined using a Quantachrome Autosorb-iQ nitrogen volumetric adsorption apparatus. Prior to analysis, the samples underwent a 10 h vacuum degassing process at 423 K to eliminate any residual guest molecules. The adsorption isotherms for C₂H₄ and C₂H₆ on HKUST−1 were determined using an Intelligent Gravimetric Analyzer (Hiden, UK, IGA002) equipped with a circulating water bath. Before conducting the adsorption measurements, HKUST−1 underwent activation at a temperature of 423 K in an ultra-high vacuum environment (10⁻⁷ mbar) for a duration of 10 h.

2.3. Computational Simulation

The adsorption positions of different gas molecules were determined through Monte Carlo (GCMC) simulations in the Materials Studio’s adsorption module, utilizing experimental crystallographic data from the HKUST−1 structure. Before conducting the simulations, we optimized the guest gas molecules and structures using the Dmol³ module. The Mulliken charges of the framework atoms and ESP charges of all atoms in the gas molecule were assigned through charge assignment implemented in the Periodic Density Functional Theory (PDFT) of the Dmol³ module. The calculations employed a conformational bias method based on the COMPASSIII force field. Electrostatic and van der Waals interactions were handled using Ewald-based and atom-based summation methods, respectively, with a Lennard-Jones potential cut-off radius of 15.5 Å. The C₂H₆ molecule was loaded into an adsorption model at a fixed loading of 298 K using a 2 × 2 × 2 supercell as the simulation box. Both equilibrium and production steps were set to 1 × 10⁷.

3. Results and Discussion

3.1. Structural Analysis

The porous material HKUST−1 was initially reported by Williams and his collaborators, and has since been extensively investigated by numerous research teams [24,25,26]. The single-crystal X-ray diffraction analysis revealed that HKUST−1 is composed of dimerized copper tetracarboxylate paddle-wheel building blocks (Figure 1a). In the HKUST−1 structure, each 1,3,5-benzenetricarboxylic acid (BTC) ligand is bound to three dimeric copper wheels to form a highly porous framework with a face-centered cubic lattice with Fm3-m symmetry (Figure 1c). The pore network of HKUST−1 has two different sizes of pores after removing the water in the framework. The three-dimensional porous structure exhibits a square cross-sectional main channel with an approximate diameter of 0.88 nm (Figure 1c). Additionally, it possesses tetrahedral-shaped side pockets with diameters of approximately 0.55 nm (Figure 1b), which are interconnected to the main channel through triangular windows. The synthesis of HKUST−1 was accomplished through a hydrothermal reaction involving the combination of Cu(NO₃)₂·3H₂O, 1,3,5-benzenetricarboxylic acid, deionized water, and ethanol at a temperature of 383 K for 24 h. The PXRD pattern of HKUST−1 powder is shown in Figure S1, exhibiting a high degree of agreement with the simulated peaks. This indicates that the prepared HKUST−1 was a pure phase, and the sharp and clear peaks demonstrate its excellent crystal structure. The PXRD pattern of HKUST−1 after extraction by methanol (MeOH) remained unchanged compared to the untreated sample, suggesting that HKUST−1 retains an invariant phase composition and still possesses a microporous crystalline framework. The permanent porosity of HKUST−1 was investigated through a nitrogen adsorption and desorption curve at 77 K. According to Figure S2, HKUST−1 exhibits a type-I isothermal adsorption profile. In the low-pressure region, N₂ adsorption rapidly increased with increasing pressure, indicating that HKUST−1 possesses microporous characteristics. The BET surface area in the HKUST−1 was determined by analyzing nitrogen sorption isotherms collected at 77 K during adsorption studies, yielding a value of 1420 m² g⁻¹ (Figure S2a). The pore size distributions determined by the DFT (Density Function Theory) model based on 77 K N₂ adsorption showed values of approximately 0.55 nm and 0.90 nm (Figure S2b), which were consistent with the pore sizes obtained from crystal structure determination.

The BET surface area in the HKUST−1 was determined by analyzing nitrogen sorption isotherms collected at 77 K during adsorption studies, yielding a value of 1420 m² g⁻¹ (Figure S2a). The pore size distributions obtained from the nitrogen adsorption isotherm at 77 K were approximately 0.55 nm and 0.90 nm (Figure S2b), which were consistent with the pore size values determined from the crystal structure.

3.2. Adsorption Study of C₂H₆ and C₂H₄

The single-component adsorption isotherms of HKUST−1 for C₂H₆ and C₂H₄ were measured at various temperatures under the pressure of 0–1 bar (Figure 2a,b). The HKUST−1 framework exhibits an obviously preferential adsorption of C₂H₄ over C₂H₆. The adsorption capacities of C₂H₄ were 8.999 mmol g⁻¹, 8.651 mmol g⁻¹, 8.021 mmol g⁻¹, 7.161 mmol g⁻¹, and 6.372 mmol g⁻¹ at temperatures of 268 K, 273 K, 283 K, 298 K, and 313 K under atmospheric pressure, respectively, which were higher than the corresponding values for C₂H₆ (8.755 mmol g⁻¹, 8.349 mmol g⁻¹, 7.566 mmol g⁻¹, 6.245 mmol g⁻¹, and 4.793 mmol g⁻¹). In addition, HKUST−1 exhibited excellent thermal stability at the experimental temperature (Figure S3). The adsorption capacity of C₂H₄ by HKUST−1 surpasses that of C₂H₆ at low coverage. At high coverage, the larger size of C₂H₆ leads to an increased formation of hydrogen bonds between C₂H₆ and O atoms in the HKUST−1 framework, resulting in an enhanced adsorption capacity for C₂H₆ by HKUST−1. As can be observed from the C₂H₆ adsorption isotherm, HKUST−1 exhibits an S-shaped adsorption profile ranging from 40 mbar (268 K) to approximately 200 mbar (313 K), indicating its temperature independence and strong dependence on the concentration of C₂H₆.

In order to assess the adsorption affinity of C₂H₆ and C₂H₄ on HKUST−1, we calculated the isosteric heats of zero coverage (Q_{st, n=0}) for C₂H₆ and C₂H₄ on HKUST−1 based on their adsorption isotherms at various temperatures (Figures S4–S13). Analyzing isothermal data using the virial equation [27,28]:

ln(n/p) = A₀ + A₁n + A₂n² + A₃n³…

(1)

The virial coefficients, represented as A₀, A₁, etc., are temperature-dependent functions in this investigation. Specifically, n represents the adsorbed amount of gas, A₀ is related to the interaction between the adsorbate and the adsorbent, while A₁ quantifies the interactions among adsorbate molecules [29]. The weak interaction force between gas molecules and the material in the low-pressure region leads to a limited surface coverage. Consequently, for this experiment, the A₂, A₃,… virial coefficients are disregarded, thereby allowing equation 1 to be transformed into the following form:

ln(n/p) = A₀ + A₁n

(2)

The calculated value of Q_{st, n=0} for C₂H₄ was determined to be 39.341 kJ mol⁻¹ at near-zero coverage, which was higher than that of C₂H₆ (30.175 kJ mol⁻¹) (Figures S14 and S15 and Figure 2c), which is due to the stronger repulsion between C₂H₆-C₂H₆ molecules compared to C₂H₄-C₂H₄ molecules and the interaction between the double bond in C₂H₄ and the positively charged Cu atom. It is noteworthy that the absolute value of A₁ obtained from fitting HKUST−1 to the C₂H₆ isotherm surpassed that of the C₂H₄ isotherm at low coverage (Figure 2d, Figures S16 and S17). These differences are statistically significant, suggesting a stronger repulsive energy between neighboring substances for C₂H₆ compared to C₂H₄. This is one of the important reasons why the adsorption capacity of C₂H₆ is lower than that of C₂H₄ at low-coverage conditions. Under high-coverage conditions, the amount of C₂H₆ adsorbed by HKUST−1 increases with increasing pressure. This is due to the fact that C₂H₆ has a larger molecular size than C₂H₄, which occupies a larger volume in the pores of HKUST−1 and forms more hydrogen bonds with HKUST−1 at similar adsorption levels. These hydrogen bond formations are more likely to occur at similar adsorption levels than those observed for C₂H₄. This leads to the phenomenon of the S-shaped adsorption isotherm of C₂H₆ by HKUST−1.

The Q_st value determined from the Clausius–Clapeyron equation is another important parameter for assessing the adsorption affinity of C₂H₆ and C₂H₄ (Figure 3a, Figures S18 and S19). Figure 3a illustrates the Q_st curves of HKUST−1 for C₂H₆ and C₂H₄ calculated based on their isotherms. As expected, the Q_st values of C₂H₄ were significantly higher than those of C₂H₆ over the measured pressure range, suggesting that C₂H₄ has a stronger binding affinity for HKUST−1. This observation is consistent with the single-component adsorption isotherm. As the loading amount of C₂H₆ and C₂H₄ increased, the Q_st values for C₂H₆ and C₂H₄ in HKUST−1 decreased from 29.97 kJ/mol at a loading amount of 0.1 mmol g⁻¹ to 22.16 kJ/mol at a loading amount of 2 mmol g⁻¹ for C₂H₆, and from 37.13 kJ/mol at a loading amount of 0.6 mmol g⁻¹ to 23.70 kJ/mol at a loading amount of 7.2 mmol g⁻¹ for C₂H₄. This indicates that as the loading amount increases, HKUST−1 has a decreasing affinity towards C₂H₆ and C₂H₄. It is noteworthy that the Q_st value of HKUST−1 tends to increase with the increase of C₂H₆ and C₂H₄ adsorption under high-coverage conditions. This is due to the fact that as the number of aggregated C₂H₆ and C₂H₄ molecules gradually increases, the distance between them and the O atoms in the framework is shortened, resulting in the formation of more hydrogen bonds. In addition, under the identical adsorption capacity condition, C₂H₆ is more likely to form hydrogen bonds with O atoms in the framework because of its larger size. As depicted in Figure 3a, when the adsorption amount of C₂H₆ reached 2 mmol g⁻¹, HKUST−1 formed additional hydrogen bonds with C₂H₆, resulting in an increase in its Q_st value on C₂H₆; whereas, it was only when the adsorption amount of C₂H₄ reached 7.2 mmol g⁻¹ that HKUST−1 began forming more hydrogen bonds and caused an elevation in its Q_st value on C₂H₄. Overall, the larger molecular size of C₂H₆ compared to C₂H₄ makes it more likely to form additional hydrogen bonds with HKUST−1 as the amount of adsorbed C₂H₆ and C₂H₄ increases. This phenomenon serves as one of the key explanations for why the adsorption isotherm of C₂H₆ on HKUST−1 exhibits an S-shaped behavior.

3.3. Gas Separation Properties

The C₂H₄/C₂H₆ adsorption selectivity of HKUST−1 was assessed using the ideal adsorption solution theory (IAST) at temperatures of 268 K, 273 K, 283 K, 298 K, and 313 K (Figure 3b and Figures S20–S29). The IAST selectivity of C₂H₄/C₂H₆ (1/1) for HKUST−1 at 268 K, 273 K, 283 K, 298 K, 313 K, and 1 bar are shown in Figure 3b as 3.81, 3.83, 3.78, 3.57, and 3.45, respectively. It is noteworthy that in the low-pressure region, the IAST of C₂H₄/C₂H₆ (1/1) selectivity increased significantly. Among them, the IAST selectivity of C₂H₄/C₂H₆ (1/1) increased to 8.89 at 273 K and 30 mbar. In addition, the IAST selectivity of C₂H₄/C₂H₆ (1/1) increased to 7.53 at 283 K and 50 mbar. This enhancement can be attributed to the higher repulsive force between neighboring C₂H₆ molecules compared to that between neighboring C₂H₄ molecules under low-coverage conditions, thereby rendering HKUST−1 more effective in separating C₂H₄/C₂H₆ under low pressure. However, with increasing adsorption of C₂H₆ and C₂H₄ molecules, they tend to aggregate within the pores of HKUST−1. These aggregated C₂H₆ molecules occupy a larger volume and exhibit a higher propensity for forming hydrogen bonds with O atoms on the pore walls of HKUST−1 compared to the aggregated C₂H₄ molecules, thereby significantly enhancing the adsorption capacity of HKUST−1 for C₂H₆. Consequently, under high-pressure conditions, HKUST−1 exhibits unsatisfactory performance in separating C₂H₄/C₂H₆.

3.4. Molecular Simulations

The separation behaviors of HKUST−1 for C₂H₆ were investigated through Grand Canonical Monte Carlo (GCMC), aiming to identify the adsorption sites and interactions of HKUST−1 on C₂H₆. The density distribution map obtained from GCMC simulation revealed the presence of two different adsorption sites for C₂H₆ (Figure S30). Specifically, one set of sites is situated within tetrahedral-shaped side pockets (Figure 4a), while the other site is located within square cross-sectional channels (Figure S31). This observation suggests that the binding of C₂H₆ molecules may be attributed to the abundance of oxygen sites within the pores. The preferential adsorption site of HKUST−1 for C₂H₆ is located in the tetrahedral-shaped side pockets, where each C₂H₆ molecule interacts with the benzene ring in three 1,3,5-benzenetricarboxylic acids through the formation of four C-H⋅⋅⋅ᴨ interactions with a distance ranging from 3.091 to 3.547 Å. As the amount of C₂H₆ adsorbed increased, two C₂H₆ molecules were aggregated within the tetrahedral-shaped side pockets. Since C₂H₆ has a large volume, it occupies relatively more space within the tetrahedral-shaped side pockets and is closer to the HKUST−1 pore surface. As shown in Figure 4b, two C₂H₆ molecules formed six C-H⋅⋅⋅O hydrogen bonds with five O atoms in tetrahedral-shaped side pockets at distances ranging from 2.767 to 3.450 Å. The increase in the number of hydrogen bonds resulted in the enhancement of C₂H₆ adsorption capacity of HKUST−1, which is consistent with the experimentally observed increase in C₂H₆ adsorption by HKUST−1 at high C₂H₆ coverage.

4. Conclusions

In summary, we experimentally investigated the adsorption isotherms of C₂H₆ and C₂H₄ on HKUST−1 at different temperatures and investigated the disparities in affinity between HKUST−1 and C₂H₆, as well as C₂H₄, using both the virial equation and the Clausius–Clapeyron equation. It was observed that the absolute value of A₁ obtained from fitting HKUST−1 to C₂H₆ isotherms was higher than that obtained from fitting HKUST−1 to C₂H₄ isotherms at low coverage, indicating a stronger repulsive force between C₂H₆-C₂H₆ molecules compared to that between C₂H₄-C₂H₄. This difference in repulsive forces contributes significantly to the lower adsorption capacity of HKUST−1 for C₂H₆ at low coverage. At high coverage, due to the larger molecular size of C₂H₆, aggregation occurs within the pore channels leading to closer proximity between individual molecules and the pore wall of HKUST−1. Consequently, more hydrogen bonding takes place, resulting in an increased amount of C₂H₆. Under low and high coverage conditions, HKUST−1 exhibits distinct adsorption mechanisms for C₂H₆, resulting in the observation of an S-shaped adsorption phenomenon on HKUST−1. Moreover, it provides more favorable conditions for achieving efficient separation of C₂H₄/C₂H₆ under low pressure. The selectivity of HKUST−1 for C₂H₄/C₂H₆ was found to be 3.78 at 283 K and 1 bar; however, this selectivity significantly increased to 7.53 under low pressure, confirming the potential feasibility of employing HKUST−1 for efficient separation of C₂H₄ from C₂H₄/C₂H₆ mixtures under low pressure. Overall, this study not only elucidates the S-shaped adsorption isotherm of C₂H₆ by HKUST−1, but also highlights the significant potential for separating complex gas mixtures based on differences in repulsive forces among different guest molecules and interactions between different guest molecules and the framework. This unique mechanism is expected to advance the field of C₂H₄/C₂H₆ separation technology and contribute to the realization of a more energy-efficient and cleaner environment.