The synthesis of 2-allylterephthalic acid ligand was performed with reference to reported methods [22] (Scheme 1). 2-Bromoterephthalic acid (6.2 g, 25.3 mmol) and 5 mL of H₂SO₄ were added to 100 mL of CH₃OH solution. The mixture was heated under reflux for 24 hours. After the reaction, NaHCO₃ solution (50ml) was added to the mixture, and white precipitate of dimethyl 2-bromoterephthalate was obtained by filtration. And then dimethyl 2-bromoterephthalate, Pd (pph₃)₄, and allyltributyltin were dissolved in 200 mL of toluene, and reflux for 4 days under N₂ atmosphere. After the reaction, the product was purified by silica gel column chromatography and distilled under reduced pressure to obtain a colorless oil. The intermediate product was dissolved in 10 mL of THF, and NaOH solution (1 mmol/L, 65 mL) was added. The mixture was stirred at room temperature for 12 hours. After the reaction, the organic solvent was removed under vacuum, the residue was acidified with 1 M HCl solution, and white precipitate was obtained by filtration (yield 81%). Figure 1 shows the ¹H NMR spectrum of 2-allylterephthalic acid (BDC-allyl). ¹H NMR (400 MHz, DMSO-d6, ppm): 13.24 (s, 2H), 7.86 (d, 3H), 5.97 (ddt, 1H), 5.03 (m, 2H), 3.75 (s, 2H)

Scheme 1. Synthetic procedures of the BDC-allyl ligand.

BDC-allyl (5 mg) and zirconocene dichloride (15 mg) were dissolved in 1 mL of N,N'-dimethylacetamide (DMA), and four drops of deionized water were added. After full dissolution by ultrasound, the mixture was transferred to an electric thermostatic blast drying oven and reacted at 65˚C for 10 hours to obtain colorless and transparent cubic crystals (ZrT-1-allyl) with a yield of 73%.

To characterize the structure of ZrT-1-allyl, X-ray diffraction (PXRD) and high-resolution electrospray ionization time-of-flight mass spectrometry (ESI-TOF-MS) were performed. The PXRD results show that the diffraction peak positions of ZrT-1-allyl are consistent with those of the parent structure, indicating that ZrT-1-allyl is isostructural with ZrT-1 (Figure 2(b) and Table 1). The ESI-TOF-MS results confirm the existence of cages in ZrT-1-allyl (Figure 2(a)). The +2, +3, and +4 ion peaks in the ZrT-1-allyl spectrum are 1682.8120, 1122.1914, and 841.8985, corresponding to [M-4Cl-2H]²⁺, [M-4Cl-H]³⁺, and [M-4Cl]⁴⁺ ions, respectively.

The average measured molecular weight after m/z deconvolution is 3509 Da, which is 240 Da higher than 3269 Da of ZrT-1. This difference is exactly equal to the molecular weight difference of functional groups between terephthalic acid (H₂BDC) and BDC-allyl in six ligands ([M(C₃H₅)-M(H)]), further confirming that ZrT-1-allyl has the same structure as ZrT-1, which is a tetrahedral cage structure. Meanwhile, due to the weak diffraction of ZrT-1-allyl, no single crystal data was obtained, so we simulated the structure of ZrT-1-allyl (Figure 2(c)). ZrT-1-allyl has a tetrahedral cage structure, arranged symmetrically in the tetragonal space group I41/a. Each cage is composed of BDC-allyl occupying the six edges, and four Cp₃Zr₃O(OH₃) units at the four vertices. Cages were connected by H-Cl bonds, forming a tetrahedral structure inside the cavity and wavy pores between the cages.

To further analyze the porosity and separation performance, N₂ adsorption isotherms were obtained at 77 K. The isotherms are consistent with Type I isotherms, and the N₂ adsorption capacity is 173.295 cm³/g (Figure 2(d)). The Brunauer-Emmet-Teller (BET) specific surface area was calculated to be 508.7899 m²/g. Subsequently, the adsorption curves of ZrT-1-allyl for C₂H₂ and CO₂ were measured at 273 K and 298 K (Figure 2(e) & Figure 2(f)) At 298 K and 1 bar, the adsorption capacity of ZrT-1-allyl for C₂H₂ (42.54 cm³/g) is significantly higher than that for CO₂ (24.89 cm³/g), with an adsorption difference of 17.65 cm³/g, indicating that ZrT-1-allyl has certain potential for C₂H₂/CO₂ separation. To further evaluate the affinity of the material for C₂H₂ and CO₂, the IAST selectivity and adsorption enthalpy (Qst) were calculated according to the Ideal Adsorbed Solution Theory (IAST) and Clausius-Clapeyron equation, respectively (Figure 3(a) & Figure 3(b)). The Qst value of C₂H₂ (20.79 kJ/mol) is higher than that of CO₂ (14.62 kJ/mol), indicating that the material has higher affinity for C₂H₂ than CO₂. At 298 K and 100 kPa, the selectivity of ZrT-1-allyl for C₂H₂/CO₂ (50/50) is 2.8, which shows that the introduction of allyl functional groups effectively enhances the selectivity of the material for C₂H₂/CO₂. Collectively, ZrT-1-allyl has high C₂H₂ adsorption capacity, moderate separation selectivity, and low adsorption enthalpy, indicating its potential for separating C₂H₂/CO₂ mixed gases.

To validate the practical separation capability of the material, breakthrough experiments were performed on C₂H₂/CO₂ (Figure 3(c) & Figure 3(d)). The results demonstrate that ZrT-1-allyl can effectively separate C₂H₂/CO₂. At 298 K and 1 atm, the C₂H₂/CO₂ (50/50, v/v) mixed gas was passed through the sample tube at a flow rate of 1 mL/min, and the outlet gas was monitored by a mass spectrometer. Due to the weak affinity of ZrT-1-allyl for CO₂, it was desorbed first from the sample tube, after a period of time C₂H₂ was desorbed after reaching adsorption saturation, with a separation time of 24 min/g (Figure 3(d)). The breakthrough experiment proves that ZrT-1-allyl possesses practical C₂H₂/CO₂ separation capability and can serve as an ideal adsorbent separation material. Additionally, after three rounds of cycle tests, the separation performance of the material did not degradation significantly, indicating that the ZrT-1-allyl has excellent cycling stability.

The interaction between the framework and gases was further clarified and the adsorption behavior of gas molecules on the framework was determined via theoretical simulation. The H in C₂H₂ can form weak interactions with allyl groups through C-H・・・π (3.13 - 3.33 Å), and form C-H・・・O hydrogen bonds (3.03 - 3.2 Å) with carboxyl oxygen. Furthermore, multiple Cδ-・・・Hδ+ dipole-dipole interactions (3.05 - 3.23 Å) exist between adjacent C₂H₂ molecules adsorbed within the pores, which exert a synergistic effect to enhance the stability of C₂H₂ adsorption. This indicates that strong interaction between the cage and C₂H₂ (Figure 4(a) &

Figure 4(b)). GCMC simulation results show that gas molecules are preferentially adsorbed in the ZrT-1-allyl cages, and the adsorption density of C₂H₂ is stronger than that of CO₂ (Figure 4(c) & Figure 4(d)). Theoretical simulations and gas penetration experiment results mutually corroborate each other, proving that ZrT-1-allyl can effectively separate C₂H₂/CO₂ mixed gases and has potential for practical application.