Separation of CH₄/N₂ mixtures in metal–organic frameworks with 1D micro-channels

Abstract

This work aims to study the features of CH₄ and N₂ adsorption inside 1D micro-channels and develop the best suitable MOF candidates for the adsorptive separation of CH₄ against N₂. For this purpose, four MOFs ([Ni₃(HCOO)₆], [Cu(INA)₂], Al-BDC and Ni-MOF-74) with similar network topology and single 1D micro-channel have been systematically investigated via structure characterization and selective gas adsorption and separation. The selected MOFs are classified into three groups. Thereinto, Ni-MOF-74, with coordinatively unsaturated metal sites, is considered as strong polar adsorbent, whilst Al-BDC is treated as moderate polar adsorbent owing to the polar linkers. However, [Ni₃(HCOO)₆] and [Cu(INA)₂] are regarded as apolar (weak polarity) adsorbents because of lack of any polar functional groups inside the frameworks. The adsorption potential of CH₄ follows the trend Ni-MOF-74 > [Ni₃(HCOO)₆] > [Cu(INA)₂] > Al-BDC, while Ni-MOF-74 > Al-BDC > [Ni₃(HCOO)₆] > [Cu(INA)₂] for the potential of N₂. This implies that pore size and electrostatic interactions have different contributions to the CH₄ or N₂–MOFs interactions, resulting in excellent CH₄/N₂ selectivity more than 6 on [Ni₃(HCOO)₆] and [Cu(INA)₂].

---

Introduction

Effective separation of methane (CH₄) from unconventional natural gases (UNG) like coal-bed methane, landfill gas and shale gas is a significant supplement for the rapidly increasing demand for natural gas. Nevertheless, separation of CH₄/N₂ emerges as a bottleneck problem because of the very similar physical and chemical properties of CH₄ and N₂. Considerable efforts have been devoted to developing effective methods. Thereinto, the adsorption-drived separation methods like Pressure Swing Adsorption (PSA) are commonly accepted as the most potential candidate for the removal of nitrogen owing to its high energy efficiency, intrinsic economic feasibility and operating flexibility.¹ However, the vital step in designing PSA processes for the CH₄/N₂ mixture is the development of efficient adsorbents with appropriate pore network and pore aperture. Within the PSA process, a large majority of separations operate through the equilibrium adsorption of low CH₄ concentration mixture. Therefore, the goal of development and design of adsorbents focuses on the high CH₄ adsorption capacity, desired selectivities and excellent regeneration properties. In the literature, several available adsorbents such as carbon-base materials,^2–4 silica gels,⁵ zeolites,^6–9 metal–organic frameworks (MOFs)^10–15 and covalent organic frameworks(COFs)¹⁶ have been prepared and applied into the separation of CH₄/N₂. Among these adsorbent materials, MOFs, as a relatively new class of inorganic and organic hybrid porous materials, are constructed by coordination chemistry of metal ion clusters and organic linkers. The MOFs materials exhibit outstanding potential in applications of adsorption separation due to their flexibility to design through control of the size and shape of pore and cavity dimensions within the framework and surface chemistry (acid–base properties, functional groups).

Up to date, the mechanism for gas adsorption and separation inside channels has been investigated on some MOFs.^10,17–22 For equilibrium separation, the equilibrium effect is related to various structural properties of the adsorbents, such as network topology, surface area, free volume, pore size distribution, pore shape, pore connectivity and polarity of channels. These structural features are thought to have significant influence on diffusion pathways and adsorption potential. However, the structure and property of MOFs are quite complex, making it difficult to predict the equilibrium adsorption and transport properties. In order to unambiguously understand how to improve adsorbate–MOF interactions through tuning the MOFs structures and properties, MOFs including [Ni₃(HCOO)₆], [Cu(INA)₂], Al-BDC and Ni-MOF-74 are selected. These MOFs have similar network topology, do not interpenetrate and only contain a single 1D micro-channel. The simple pore structure will help to explain the specific role of pore structure in the separation process. In addition, to distinguish the large diversity of MOFs surface properties, the selected MOFs are classified into three main groups, based on electrostatic interactions: (a) MOF with strong polarized micropores because of coordinatively unsaturated metal sites, i.e., Ni-MOF-74, (b) MOF with moderate polarized micropores because of the polar groups of linkers, i.e., Al-BDC, (c) the MOFs with weak polarized micropores considered as apolar (weak polarity) adsorbents, i.e., [Ni₃(HCOO)₆] and [Cu(INA)₂].

On the basis of the above analysis, the MOFs are chosen to get an insight into molecular level details of the features for separation inside channels. In this work, [Ni₃(HCOO)₆], [Cu(INA)₂], Al-BDC and Ni-MOF-74 are prepared by using a modified solvent-thermal method, and the structures and surface properties are carefully characterized and analyzed. In order to understand the adsorption interaction between the inner surface of narrow pore and gas molecules, the pure gas adsorption and dynamic binary gas breakthrough separation methods have been employed to evaluate the CH₄ adsorption capacity and selectivity of CH₄/N₂ on selected MOFs. Since the selected MOFs adsorb nonpolar CH₄ and N₂ molecules mainly through the dispersion and polarization interactions, the adsorbate–adsorbent potential is ϕ = ϕ_D + ϕ_R + ϕ_Ind + ϕ_Eμ + ϕ_FQ. This equation provides theoretical prediction in factors which have an important effect on performance of adsorption and separation. These studies are conducive to better understand the gas–host interactions and practical guidance for screening of suitable MOF materials for enhanced separation processes.

Experimental section

1. Synthesis of the MOFs

All reagents and solvents for synthesis were obtained from commercial vendors and of analytical grade without further treatment prior to use unless otherwise stated.

[Ni₃(HCOO)₆] and Al-BDC were prepared and fully activated as described in our previous publications.^15,23 [Cu(INA)₂]: 0.1 mol of copper(II) sulfate pentahydrate and 0.2 mol of isonicotinic acid were suspended in 500 mL of a 1/1 v/v solution of ethanol/water in a glass flask and stirred under reflux (90 °C) for 12 hours. The blue product was filtered off and washed with 3 × 100 mL of ethanol. The sample was treated at 80 °C for 10 hours and subsequently dried at 130 °C in a vacuum drying oven for 12 hours. 31.5 g of powder were obtained.

Ni-MOF-74, were synthesized by scaling up the synthesis methods described by Dietzel et al.²⁴ In this work, the compound was prepared in a similar synthetic process with minor modifications. 22.5 mmol of 2,5-dihydroxyterephthalic acid was dissolved in THF (300 mL) under sonication. 300 mL of a solution composed of nickel acetate tetrahydrate (45.0 mmol) was added drop-wise into this solution over 5 min while stirring. After stirring at room temperature for 30 min to form a homogenous mixture, the resulting stock solution was decanted into a 1 L Teflon-lined stainless steel autoclave and sealed. Then, the reaction was carried out at 110 °C for 72 h under autogenous pressure by solvent-thermal synthesis. After cooling to room temperature, the yellow precipitate was obtained and then immersed in 300 mL of fresh methanol to exchange the DMF over 48 h and then washed with 2 × 100 mL of acetone. The as-synthesized sample was evacuated at 160 °C under dynamic vacuum conditions for 12 h.

2. Materials and adsorbent characterization

To obtain crystallographic structure of the materials, powder X-ray diffraction (PXRD) patterns were measured using a PANalytical X'pert diffractometer with Cu Kα1 radiation, 40 kV/40 mA current. The X-ray scanning speed was set at 8° min⁻¹ and a step size of 0.02° in 2θ. Porosity measurements of the selected MOFs were confirmed by the Ar physisorption measurements at 87 K on a volumetric sorption analyzer Autosorb-iQ2 from Quantachrome Instruments. Prior to measures, the samples were outgassed at 150 °C under vacuum for 12 h. The pore structure characteristics including the Brunauer–Emmett–Teller (BET) surface area, pore volume, and pore size distribution (PSD) were calculated from the Ar adsorption. The micropore volume and PSD were calculated using Dubinin–Raduskevitch method and non-linear density-functional theory, respectively. Thermo-gravimetric experiments (TGA) were carried out in the temperature range of 30–800 °C on a NETZSCH STA 449 F3 analyzer under argon atmosphere at a heating rate of 10 °C min⁻¹.

3. Adsorption measurements

Sorption isotherm studies of methane and nitrogen on all adsorbents were performed volumetrically (up to 100 kPa) on a modified Quantachrome Autosorb-iQ2 gas sorption analyzer coupled with a recirculating thermostatic bath with an accuracy of 0.01 K. Adsorption equilibrium data were collected at 288, 298, and 308 K, respectively. The error of volumetric adsorption can be greatly reduced by using a mixture of ethylene glycol and de-ionized water as heating fluid and filling the sample cell with rod. Prior to measurements, samples were degassed at 150 °C for 12 h under vacuum condition to evacuate residual guest molecules. All adsorption isotherms have been fitted using Tòth equation.where, n is the gas uptake (mmol g⁻¹), n_max is the maximum gas uptake (mmol g⁻¹), b is the Langmuir parameter, t is the parameter characterizing energetic heterogeneity and fractality of the surface of the sorbent.

Breakthrough curve measurements of CH₄/N₂ were carried out at 298 K using a fixed bed apparatus (Fig. S1†), allowing one to conduct measurements in a pressure range from atmospheric pressure up to 1.0 MPa. Flow rates of He (pre-purified, 99.99%), CH₄ (99.95%), and N₂ (99.98%) were controlled using Sevenstar D07-11C mass flow controllers. Feed gas was generated in situ. For CH₄/N₂ breakthrough experiments, granular adsorbent (60–80 mesh) was loaded into a fixed-bed stainless steel column with inner dimensions of 4 × 250 mm and then installed onto the apparatus under flowing He. Before the breakthrough curve measurements, the adsorbents were outgassed and in situ activated at 160 °C under a He flow of 20 NL min⁻¹ for 2 h. Subsequently, the equimolar mixtures of CH₄–N₂ (flow rate of 20 mL min⁻¹) were switched into the adsorbent bed. The effluent was analyzed using a TOF-MS300 mass spectrometer (MS) made by DICP (Dalian Institute of Chemical Physics). The signal strength of the mass spectrometer had a good linear relationship with the molar concentration of gas in the range of 0–10%. In the figures, gas ion intensities were in the form of the normalized I_i/I₀, where I_i, I₀ represented the measured ion intensity at the column outlet and the feed gas intensity of component i, respectively. After the adsorption step in every measurement was finished, the sample column was purged with He to regenerate for another measurement. The pressure drop over the column was always less than 0.005 MPa.

The equilibrium selectivities of adsorption are calculated by the equation:

where, n_i is the adsorbed amount of component i (i = A or B) in fixed bed, and y_i is the mole fraction of components i in gas phases at equilibrium.

Results and discussion

1. Synthesis and characterization

The PXRD measurements of activated [Ni₃(HCOO)₆], [Cu(INA)₂], Al-BDC and Ni-MOF-74 are carried out to confirm the crystal structures and phase purity (Fig. 1 and S2†). As shown in Fig. 1, the PXRD peak positions of these materials agree well with the simulated patterns (Fig. S2†), which confirms that the synthesized materials have the same structure as patterns reported in the literature.^25–28 Moreover, the sharpness and high intensity indicate that all materials are of pure-phase structure and high crystallinity.

Fig. 1 PXRD patterns of the as-synthesized compounds.

The thermal stabilities of selected MOFs are investigated by using thermal gravimetric analysis (TGA) between 30 °C and 800 °C, and weight loss profiles are exhibited in Fig. 2. For as-synthesized [Ni₃(HCOO)₆], the TGA curve shows two main events between 30 °C and 800 °C. The first gradual weight-loss step of 13.8% loss from 100 °C to 185 °C is observed, corresponding to liberation of DMF molecules trapped in the pores. After removing the guest molecules, the significant decomposition is observed until temperatures up to 240 °C. In general, the thermal behaviors of [Cu(INA)₂] and Ni-MOF-74 are similar to that of [Ni₃(HCOO)₆]. As for the Al-BDC sample, all the guests (CH₃OH, DMF, and BDC ligands) are gradually liberated from the pores up to 420 °C in two step processes with the observed total loss (20.6%), and the framework decomposition starts from 520 °C. On heating, Ni-MOF-74 immediately starts to lose water molecules coordinating the metal until it is completely dehydrated at 230 °C. Subsequent sample weight stays constant until 330 °C. These MOFs all can maintain the stability of the structure above 200 °C, indicating that they are thermally stable enough to be used as adsorbents in PSA processes.

Fig. 2 TGA plots of the as-synthesized compounds in argon atmosphere with the rate of 10 °C min⁻¹.

Fig. 3a shows the argon adsorption and desorption isotherms of the activated MOFs under liquid argon bath (87.3 K) to investigate the architectural rigidity and the permanent porosity. The resulting isotherms on evacuated sample are of a typical type-I isotherm with a sharp increases at very low relative pressures, corresponding to a permanent micro-porosity. All the porosity parameters are summarized in Table 1. The BET specific surface areas of the [Ni₃(HCOO)₆], Al-BDC and Ni-MOF-74 studied here are similar to those reported previously.^15,23,29 It is noted that the specific surface areas of [Cu(INA)₂] is ca. 251.8 m² g⁻¹, which is firstly given in the literature. Furthermore, the pore size distributions of the selected MOFs are depicted in Fig. 3b, which are based on the NLDFT model for cylinder pore geometry. Obviously, the results show that the pore size distributions of selected MOFs are narrow, and very close to that estimated from crystal structure. Moreover, the pore size of the frameworks continuously augment from [Ni₃(HCOO)₆] to Ni-MOF-74.

Fig. 3 Gas adsorption analysis. (a) Argon adsorption isotherms of activated Ni-MOF-74 (A), Al-BDC (B), [Cu(INA)₂] (C) and [Ni₃(HCOO)₆] (D). (b) Pore diameter distributions calculated by using NLDFT model. (Solid and open marks represent adsorption and desorption points, respectively.)

Table 1 Properties for the materials studied in this work

	[Ni3(HCOO)6]	[Cu(INA)2]	Al-BDC	Ni-MOF-74
a The density of MOFs is obtained by calculated.b SSA (specific surface area) calculated by BET method.c Vt (total pore volume) calculated by Gurvich-rule at P/P0 = 0.95.d Vm (micropore volume) calculated by D–R method.
Formula	C6H6Ni3O12	C12H8CuN2O4	C8H5AlO5	Ni2(C8H2O6)
Polarity	Weak	Weak	Moderate	Strong
ρ^a/g cm^−3	1.86	1.70	0.72	1.19
Pore shape	1D zigzag-shaped channel	1D rectangular channel	1D diamond-shaped channel	1D hexagonal channel
Pore size/nm	0.43	0.47	0.78	1.20
SSA^b/m^2 g^−1	232.6	251.8	926.1	1407.1
Vt^c/cm^3 g^−1	0.10	0.12	0.43	0.60
Vm^d/cm^3 g^−1	0.09	0.09	0.35	0.53

---

2. Comparison of the MOF structures

Theoretically, the higher adsorptive separation efficiency can be generally achieved through controlling pore network and pore aperture, tuning internal surface properties, increasing specific surface area and taking advantage of structural flexibility of MOFs. Therefore, in order to facilitate the interpretation of selective adsorption and separation, we comparatively analyze the selected MOFs structures and properties. Their structures are shown in Fig. 4.

Fig. 4 Crystal structures of metal organic frameworks: (a) [Ni₃(HCOO)₆]; (b) [Cu(INA)₂]; (c) Al-BDC and (d) Ni-MOF-74 (Ni, olive; O, red; C, gray; Cu, cyan; N, blue; Al, violet; H atoms are omitted for clarity).

[Ni₃(HCOO)₆] is constructed by the shortest carboxylate connector, HCOO–, which has the highest metal and HCOO– proportion compared with other MOFs. The coordination geometry of framework consisting of the apex-sharing M-centered MM4 tetrahedron nodes contains four crystallographically independent M(II) ions, each of which is octahedrally coordinated by six formate anions. Each formate ligand bridges three metal ions in syn–syn/anti mode, leading to a three-dimensional (3D) framework. The metal network exhibits diamondoid connectivity and forms zigzag-shaped channels with cross-section of ca. 4 × 5 Å running along the b axis. The channels are in an approximately honeycomb arrangement, and lined by alternate arrays of C–H groups and exposed oxygen atoms.^15,25

In [Cu(INA)₂], Cu(II) cations are coordinated to five IN units, resulting in the square pyramidal and octahedral Cu centers. Specifically, square pyramidal Cu(II) bonds to two nitrogen atoms from pyridyl groups and three oxygens from carboxylate groups of five IN units in a monodentate fashion. A five-connected three dimensional network is generated by the connection of square pyramidal Cu centers coordinated with five two-connected tridentate IN units with rectangular 1D channels of approximately 4 × 6 Å dimension that run parallel to the crystallographic a axis.^26,30

Al-BDC (MIL-53-Al) is generated from the interconnection of infinite trans chains of corner-sharing AlO₄(OH)₂ octahedra which is formed by coordination of Al(III) by two OH– groups and four terephthalate ligands in four directions. In this connection mode, the metal network forms 1D diamond-shaped channel with diameters of ca. 8.5 Å taking into account the van der Waals radius of the surface atoms. This metal–organic framework exhibits excellent gas adsorption performance on account of its structural flexibility and the aromatic walls of the channels.^27,31

Ni-MOF-74 (CPO-27-Ni) is built up by the connection of helical chains interconnected by deprotonated dihydroxy-terephtalate with three adjacent chains, resulting in a honeycomb motif with pores diameter of 11 Å. In framework, Ni²⁺ center adopts octahedral symmetry upon coordinating five carboxylate oxygens and one oxygen atom from solvent water molecule. The coordinatively unsaturated metal sites (CUS) are created when the water molecules are removed by heating. Compared with other classical MOFs, the most intriguing characteristic of this framework is exceptionally large densities of open metal sites (∼4.5 sites per nm³).^32,33

3. Pure gas adsorption on the MOFs

To investigate and analyze the adsorption and separation properties of selected MOFs for CH₄–N₂ mixture, the pure CH₄ and N₂ gas adsorption experiments were performed by volumetric method. The adsorption isotherms of CH₄ and N₂ on the selected MOFs at 288 K, 298 K, and 308 K were plotted in Fig. 5. Meanwhile, the adsorption data are fitted using Tòth equation to predict the adsorption behavior of the gas mixtures. The parameters of the fitting procedures are summarized in Table 2. As depicted in Fig. 5, the rapid rise of CH₄ uptakes at low pressures indicates that adsorption of CH₄ on selected MOFs is preferred to that of N₂. This fact is also expressed by the Henry's law constants, which is calculated from the product of Tòth equation constants (K_H = n_maxb). The bigger value means the higher adsorption potential, so the adsorption of CH₄ on selected MOFs is stronger than that of N₂ because of the higher K_H for CH₄ against N₂.

Fig. 5 Pure gas adsorption experiments: CH₄ (squares) and N₂ (circles) sorption isotherms for ultra-microporous materials at 288 K (black), 298 K (red) and 308 K (blue) (solid line: Tòth equation fitting curve).

Table 2 Parameters from pure gas adsorption isotherms for CH₄ and N₂ adsorption on the microporous materials at 298 K and the ideal selectivities calculated from Henry's law constants

	CH4		N2		αCH4/N2
	KH/mmol (g Pa)^−1	na/mmol g^−1	KH/mmol (g Pa)^−1	na/mmol g^−1
[Ni3(HCOO)6]	1.077 × 10^−5	0.791	1.789 × 10^−6	0.173	6.02
[Cu(INA)2]	9.673 × 10^−6	0.827	1.160 × 10^−6	0.121	8.34
Al-BDC	7.873 × 10^−6	0.728	2.212 × 10^−6	0.224	3.56
Ni-MOF-74	2.640 × 10^−5	2.537	2.012 × 10^−5	1.921	1.35

---

As expected in most adsorbents, CH₄ is more strongly adsorbed than N₂ which can be seen from the higher CH₄ uptakes against N₂. The difference in adsorption behaviors of CH₄ and N₂ in selected MOFs was explained based on the polarizability of the adsorbates. It is known that N₂ possesses quadrupole moment whereas CH₄ has a zero quadrupole moment (Table 3), but this difference has not correspondingly reflected into their adsorption capacity, suggesting influence of quadrupole moment on the final adsorption capacity is less.

Table 3 Physical properties of CH₄ and N₂

Adsorbate	Kinetic diameter/Å	Dipole/×10^−18 esu	Quadrupole moment/×10^40 cm^2	Polarizability/×10^−25 cm^3
CH4	3.80	0	0.00	26.0
N2	3.64	0	−4.91	17.6

---

Another important parameter that determines sorbate–sorbent interaction potential is the polarizability which may be responsible for the difference. CH₄ molecules exhibit a higher polarizability of CH₄ (26.0 × 10⁻²⁵ cm³) vs. N₂ (17.6 × 10⁻²⁵ cm³) which leads to higher adsorption capacity. According to this fact, we can infer that total adsorption potentials of CH₄ and N₂ are dominated by their polarizability rather than quadrupole moment values.

Since only the dispersion and polarization interactions are involved in the adsorption by selected MOFs, it can be sure that the effects of pore structures and surface properties will play the important roles in adsorption performance. At 100 kPa and 298 K, the adsorption potential of CH₄ (Table 2) follows the trend Ni-MOF-74 > [Ni₃(HCOO)₆] > [Cu(INA)₂] > Al-BDC. As for N₂, the adsorption potential shows interesting order Ni-MOF-74 > Al-BDC > [Ni₃(HCOO)₆] > [Cu(INA)₂]. The different order between CH₄ and N₂ confirms that the adsorbate–adsorbent potential is contributed by different forms of potential. As we all know, the adsorbate–adsorbent potential is ϕ = ϕ_D + ϕ_R + ϕ_Ind + ϕ_Eμ + ϕ_FQ for physical adsorption (where ϕ_D = dispersion energy, ϕ_R = close-range repulsion energy, ϕ_Ind = induction energy (interaction between electric field and an induced dipole), ϕ_Eμ = interaction between electric field (E) and a permanent dipole (μ), ϕ_FQ = interaction between field gradient (F) and a quadruple (with quadrupole moment Q)). For CH₄ and N₂, the polarizability has an effect on the first three contributions (ϕ_D + ϕ_R + ϕ_Ind), while the quadruple moment affects the last contribution (ϕ_FQ).³⁴ According to the analysis of the selected MOFs structures and pore properties, the Ni-MOF-74 with coordinatively unsaturated metal sites and Al-BDC with carboxylic functional group are classified into polar adsorbent, while the [Ni₃(HCOO)₆] and [Cu(INA)₂] (not carrying any polar functional groups) behave like apolar adsorbents. Therefore, the proportion of ϕ_Ind and ϕ_FQ contributes in the total energy for the former should be higher than the latter due to electrostatic interactions. Especially, for the adsorption of CH₄ and N₂ molecules on the Ni-MOF-74, the ϕ_Ind and ϕ_FQ interactions dominate in the total energy due to large densities of open metal sites. In addition, it can be found that the trend of adsorption potential for CH₄ on the [Ni₃(HCOO)₆], [Cu(INA)₂] and Al-BDC is consistent with that of the pore size distribution. Thus the more constricted pores give rise to the enhancement of the potential fields due to the increase of ϕ_D + ϕ_R. In the case of N₂, the pore size is not a predictor for the adsorption potential because of the inherent quadrupolar moment of N₂ molecule. Thus we can reach the conclusion that quadrupole moment interaction effect of N₂ has an essential effect upon the adsorption potential when the MOFs are adsorbents with highly polar surfaces. Besides, these facts also confirm that variation on the polar surfaces of the MOFs (that is, variation of adsorbate–adsorbent potential) arising from coordinatively unsaturated metal sites, functional groups, or charged atoms really makes a difference to the adsorption potentials of the gas molecules.

The adsorption isotherms of selected MOFs show that adsorption capacity of CH₄ and N₂ are different on these adsorbents. At 100 kPa and 298 K, Ni-MOF-74 exhibits the highest adsorption capacity of CH₄ (2.537 mmol g⁻¹) and N₂ (1.921 mmol g⁻¹), but the MOFs with least adsorbance of CH₄ (0.728 mmol g⁻¹) and N₂ (0.121 mmol g⁻¹) is Al-BDC and [Cu(INA)₂], respectively. Comparing the data given in Tables 1 and 2, there is no linear correlation between adsorption capacity and specific surface area or pore size. In comparison with [Ni₃(HCOO)₆] and [Cu(INA)₂], the interaction between the gas molecules and Al-BDC sharply decreases because of larger pore size and thereby the adsorption capacity of CH₄ is lower. But the polarity of Al-BDC is higher than that of [Ni₃(HCOO)₆] and [Cu(INA)₂], which leads to the higher adsorption capacity of N₂ owing to the enhancement of the interactions arising from quadrupole moment. Overall, the different adsorption capacity of CH₄ and N₂ on the selected MOFs is mainly attributed to the differences in synergistic effect between constricted pores and surface polarity.

Table 2 shows the ideal adsorption selectivities αCH₄/N₂ = K_CH4/K_N2 on the selected MOFs at 298 K, which are calculated from the Henry's law constants of the respective pure gas adsorption isotherms. Remarkably, the high selectivity of CH₄/N₂ is obtained on weak polar ultra-microporous materials, i.e., [Ni₃(HCOO)₆] and [Cu(INA)₂]. The selectivities for CH₄ adsorption over N₂ on [Ni₃(HCOO)₆] and [Cu(INA)₂] are as high as 6.02 and 8.34, respectively, which are higher than that of ever reported porous materials such as zeolites, AC and most of MOFs in CH₄/N₂ separation.^10,15,35,36 For polar adsorbent, i.e., Ni-MOF-74 and Al-BDC, the ideal selectivities αCH₄/N₂ are 1.35 and 3.56, respectively, relatively lower values compared with [Ni₃(HCOO)₆] and [Cu(INA)₂]. This exhibits that the electric field created by crystal structure can increase the ϕ_FQ contribution in adsorption potential of N₂ due to the quadrupolar moment, resulting in a negative contribution to the selectivity.

4. Breakthrough experiments

Breakthrough separation experiments were carried out at 298 K to better evaluate the CH₄ adsorption capacity and separation performance against N₂ of the prepared adsorbents in the equimolar mixture of CH₄/N₂, and to confirm the effect of pore structures and surface properties on CH₄/N₂ mixture separation. CH₄/N₂ mixture have been measured under different pressures from 0.1 MPa to 1.0 MPa on activated adsorbents, and the thermodynamic selectivities are calculated according to eqn (2). The selectivities of CH₄/N₂ on the selected MOFs are collected in Table 4. The thermodynamic selectivities calculated from the breakthrough curves of selected MOFs are in agreement with the ideal selectivities determined by ratio of the Henry's law. The discrepancy obtained performing selectivity and ideal selectivity is due to different sample activations or methods to calculate/determine the selectivity.³⁷

Table 4 Adsorptive separation performance of CH₄/N₂ on activated the microporous materials at 298 K

Adsorbent	Pressure/MPa
	0.1	0.2	0.4	0.6	0.8	1.0
[Ni3(HCOO)6]	6.2	6.4	6.5	6.4	5.9	6.2
[Cu(INA)2]	6.9	7.3	7.6	7.4	7.3	7.2
Al-BDC	3.0	2.9	3.6	3.5	3.7	3.4
Ni-MOF-74	1.4	1.5	1.7	1.6	1.8	2.0

---

Owing to the difference in density, the filling quantities of the selected MOFs are different (see Table S1†). The adsorption bed with the same volume packed with 2.53 g of [Ni₃(HCOO)₆], 2.45 g of [Cu(INA)₂], 1.75 g of Al-BDC and 2.24 g of Ni-MOF-74. Fig. 6 shows the breakthrough curves of CH₄ and N₂ at 298 K and 0.4 MPa. As shown in the Fig. 6, the breakthrough curves were plotted by the sharp breakthrough fronts and the marked “roll-up” of N₂, suggesting that the separation mechanisms of CH₄ from N₂ on all samples were assigned to thermodynamic separation.³⁸ When starting form clean adsorbents, initially the CH₄ was removed from the feed gas, before the first component N₂ eluting from the column. In the early process, the weaker adsorbed species N₂ molecules diffuse faster than the stronger adsorbed species CH₄ molecules into the MOFs pores, which leaded to a transient N₂ rich adsorbed phase.³⁹ Then, partial desorption of N₂ because of the displacement of CH₄ produces the momentary higher concentration of N₂ in the eluting gas than in the feed gas.⁴⁰ The above results confirmed that CH₄ had stronger affinity with these adsorbents compared to N₂. As we expected from the pure gas adsorption isotherms, all samples exhibited a CH₄ selective adsorbed phase at equilibrium. From the shapes of breakthrough curves, it can be easily found that there were significant differences in N₂ “roll-up” effect and elution time. The N₂ “roll-up” effect on [Cu(INA)₂] is the significantly biggest while that on Ni-MOF-74 is the smallest. As for elution time, the CH₄ on selected MOFs decreases in the following order: [Cu(INA)₂], Ni-MOF-74, [Ni₃(HCOO)₆] and Al-BDC, while the N₂ on selected MOFs ranks roughly in the order of Ni-MOF-74 > [Ni₃(HCOO)₆] > [Cu(INA)₂] > Al-BDC. It should be noted that the order of the elution time doesn't agree with their CH₄ and N₂ adsorption capacities owing to the difference in filling quantity of selected MOFs. But the difference in elution time between CH₄ and N₂ can partly reflect a material's gas-separation performance.³⁵

Fig. 6 Breakthrough curves of the CH₄–N₂ equimolar mixture at 298 K and 0.4 MPa on the activated microporous materials.

The adsorption capacities of N₂ on [Ni₃(HCOO)₆], [Cu(INA)₂], Al-BDC and Ni-MOF-74 were found to be ca. 0.16 mmol g⁻¹, 0.16 mmol g⁻¹, 0.15 mmol g⁻¹ and 0.76 mmol g⁻¹, respectively, at 0.4 MPa and 298 K. In the case of CH₄, the adsorption capacities of corresponding gases were ca. 1.09 mmol g⁻¹, 1.27 mmol g⁻¹, 0.49 mmol g⁻¹ and 1.28 mmol g⁻¹. Obviously, the adsorbance of CH₄ is more than that of N₂ and follows the order, Ni-MOF-74 > [Cu(INA)₂] > [Ni₃(HCOO)₆] > Al-BDC, which is similar to results obtained from the adsorption isotherms of CH₄ and N₂. The results indicate that CH₄ dynamic adsorption capacity on the selected MOFs at high pore filling is also mainly controlled by the methane affinity towards the framework likely caused from the pore size distribution and surface polarity of the channels.

For the selected MOFs in this study, they all have a single 1D channel system. Therefore, the uniform and single pore structure makes it easy to reveal the relationship between gas separation selectivity and the structural properties. Throughout the entire pressure range, the selectivities of [Ni₃(HCOO)₆], [Cu(INA)₂] and Al-BDC are 6.0–6.5, 7.0–7.6 and 3.0–3.7, respectively, which are in good agreement with the ideal selectivities calculated from of the Henry's law and nearly constant. However, the selectivity of Ni-MOF-74 shows slight increase with the increase of pressure, which is mainly considered that the unsaturated metal sites are occupied, resulting in the weakening of electric field. Moreover, the equilibrium selectivity from Ni-MOF-74 to [Cu(INA)₂] has a significant difference. For Ni-MOF-74, the selectivity is so low that it is not suitable for the separation of CH₄/N₂ mixture. But the selectivities of [Ni₃(HCOO)₆] and [Cu(INA)₂] are exceptionally high compared with most MOFs, zeolites and AC. As shown in the results, although the ultra-microspores can enhance the separation efficiency, the highly polar surfaces with high electric field gradients have a negative effect on the selectivity of CH₄/N₂. Thus, the high selectivity over [Ni₃(HCOO)₆] and [Cu(INA)₂] are mainly attributed to ultra-microspores with weak polarity.

Conclusions

In summary, four nanoporous MOFs with similar pore structure were prepared via a facile method and characterized. The frameworks exhibit good quality crystallization, high thermal stability and permanent porosity. The structural analysis and experimental investigations contribute to our understanding of the importance of pore size and the polar pore surface in CH₄ adsorption capacity and the separation performance against N₂. The experiment results show that pore size and electrostatic interactions have different contributions to the CH₄ or N₂–MOFs interactions. The more constricted pores mainly affect the increase of ϕ_D + ϕ_R, while the electrostatic interactions dominate the contribution of ϕ_Ind and ϕ_FQ. Moreover, the effect of gas polarizability on adsorption potential tends to be more profound with decreasing pore size. In case of effect of quadrupolar moment, it is more sensitive to polarity of pore surface. The equilibrium selectivities are determined to be αCH₄/N₂ = 6.0–6.5 for [Ni₃(HCOO)₆] and αCH₄/N₂ = 7.0–7.6 for [Cu(INA)₂], which indicates that the MOFs with weak polarized micropores have a very good selectivity in the separation of the CH₄/N₂ mixtures. In a word, the simple rules of thumb and guidelines will help us to choose the type of MOF that is best suitable for CH₄ adsorption and separation of CH₄/N₂.

Acknowledgements

This work was supported by grants from the National Natural Science Foundation of China (No. 21476231). We are also grateful to Prof. Li Haiyang and Hou Keyong for their advanced mass spectrometer technology.

Notes and references

1. D. M. Ruthven, S. Farooq and K. S. Knaebel, Pressure Swing Adsorption, Wiley-VCH, New York, 1994.
2. B. Yuan, X. F. Wu, Y. X. Chen, J. H. Huang, H. M. Luo and S. G. Deng, Environ. Sci. Technol., 2013, 47, 5474–5480.
3. H. H. Yi, F. R. Li, P. Ning, X. L. Tang, J. H. Peng, Y. D. Li and H. Deng, Chem. Eng. J., 2013, 215, 635–642.
4. S. J. Bhadra and S. Farooq, Ind. Eng. Chem. Res., 2011, 50, 14030–14045.
5. J. A. Delgado, M. A. Uguina, J. L. Sotelo and B. Ruíz, Sep. Purif. Technol., 2006, 50, 192–203.
6. A. Jayaraman, R. T. Yang, D. Chinn and C. L. Munson, Ind. Eng. Chem. Res., 2005, 44, 5184–5192.
7. D. M. Ruthven, AIChE J., 1976, 22, 753–759.
8. P. Y. Li and F. H. Tezel, Microporous Mesoporous Mater., 2007, 98, 94–101.
9. S. M. Kuznicki, V. A. Bell, S. Nair, H. W. Hillhouse, R. M. Jacubinas, C. M. Braunbarth, B. H. Toby and M. Tsapatsis, Nature, 2001, 412, 720–724.
10. J. Möllmer, M. Lange, A. Möller, C. Patzschke, K. Stein, D. Lassig, J. Lincke, R. Glaser, H. Krautscheid and R. Staudt, J. Mater. Chem., 2012, 22, 10274–10286.
11. P. Rallapalli, K. P. Prasanth, D. Patil, R. S. Somani, R. V. Jasra and H. C. Bajaj, J. Porous Mater., 2010, 18, 205–210.
12. Q. M. Wang, D. M. Shen, M. Bulow, M. L. Lau, S. G. Deng, F. R. Fitch, N. O. Lemcoff and J. Semanscin, Microporous Mesoporous Mater., 2002, 55, 217–230.
13. D. Saha, Z. B. Bao, F. Jia and S. G. Deng, Environ. Sci. Technol., 2010, 44, 1820–1826.
14. J. F. Yang, Q. H. Yu, Q. Zhao, J. M. Liang, J. X. Dong and J. P. Li, Microporous Mesoporous Mater., 2012, 161, 154–159.
15. X. Y. Ren, T. J. Sun, J. L. Hu and S. D. Wang, Microporous Mesoporous Mater., 2014, 186, 137–145.
16. Q. Chen, D. P. Liu, M. Luo, L. J. Feng, Y. C. Zhao and B. H. Han, Small, 2014, 10, 308–315.
17. J. J. Perry, S. L. Teich-McGoldrick, S. T. Meek, J. A. Greathouse, M. Haranczyk and M. D. Allendorf, J. Phys. Chem. C, 2014, 118, 11685–11698.
18. J. Li, G. P. Yang, S. L. Wei, R. C. Gao, N. N. Bai and Y. Y. Wang, Cryst. Growth Des., 2015, 15, 5382–5387.
19. J. R. Li, R. J. Kuppler and H. C. Zhou, Chem. Soc. Rev., 2009, 38, 1477–1504.
20. S. T. Meek, S. L. Teich-McGoldrick, J. J. Perry, J. A. Greathouse and M. D. Allendorf, J. Phys. Chem. C, 2012, 116, 19765–19772.
21. M. V. Parkes, C. L. Staiger, J. J. Perry, M. D. Allendorf and J. A. Greathouse, Phys. Chem. Chem. Phys., 2013, 15, 9093–9106.
22. S. T. Meek, J. J. Perry, S. L. Teich-McGoldrick, J. A. Greathouse and M. D. Allendorf, Cryst. Growth Des., 2011, 11, 4309–4312.
23. T. J. Sun, X. Y. Ren, J. L. Hu and S. D. Wang, J. Phys. Chem. C, 2014, 118, 15630–15639.
24. P. D. C. Dietzel, B. Panella, M. Hirscher, R. Blom and H. Fjellvag, Chem. Commun., 2006, 9, 959–961.
25. Z. M. Wang, B. Zhang, Y. J. Zhang, M. Kurmoo, T. Liu, S. Gao and H. Kobayashi, Polyhedron, 2007, 26, 2207–2215.
26. C. Z. J. Lin, S. S. Y. Chui, S. M. F. Lo, F. L. Y. Shek, M. M. Wu, K. Suwinska, J. Lipkowski and I. D. Williams, Chem. Commun., 2002, 15, 1642–1643.
27. T. Loiseau, C. Serre, C. Huguenard, G. Fink, F. Taulelle, M. Henry, T. Bataille and G. Ferey, Chem.–Eur. J., 2004, 10, 1373–1382.
28. E. J. Garcia, J. P. S. Mowat, P. A. Wright, J. Perez-Pellitero, C. Jallut and G. D. Pirngruber, J. Phys. Chem. C, 2012, 116, 26636–26648.
29. M. H. Rosnes, M. Opitz, M. Frontzek, W. Lohstroh, J. P. Embs, P. A. Georgiev and P. D. C. Dietzel, J. Mater. Chem. A, 2015, 3, 4827–4839.
30. J. Y. Lu and A. M. Babb, Chem. Commun., 2002, 13, 1340–1341.
31. S. Bourrelly, P. L. Llewellyn, C. Serre, F. Millange, T. Loiseau and G. Ferey, J. Am. Chem. Soc., 2005, 127, 13519–13521.
32. P. D. C. Dietzel, P. A. Georgiev, J. Eckert, R. Blom, T. Strassle and T. Unruh, Chem. Commun., 2010, 46, 4962–4964.
33. H. Wu, W. Zhou and T. Yildirim, J. Am. Chem. Soc., 2009, 131, 4995–5000.
34. R. T. Yang, Adsorbents: Fundamentals and Applications, John Wiley & Sons, Inc, Hoboken, New Jersey, USA, 2003, pp. 8–10.
35. L. B. Li, J. F. Yang, J. M. Li, Y. Chen and J. P. Li, Microporous Mesoporous Mater., 2014, 198, 236–246.
36. B. Liu and B. Smit, J. Phys. Chem. C, 2010, 114, 8515–8522.
37. L. Zhou, W. C. Guo and Y. P. Zhou, Chin. J. Chem. Eng., 2002, 10, 558–561.
38. D. Peralta, G. Chaplais, A. Simon-Masseron, K. Barthelet, C. Chizallet, A. A. Quoineaud and G. D. Pirngruber, J. Am. Chem. Soc., 2012, 134, 8115–8126.
39. R. J. Mohr, D. Vorkapic, M. B. Rao and S. Sircar, Adsorption, 1999, 5, 145–158.
40. U. Bohme, B. Barth, C. Paula, A. Kuhnt, W. Schwieger, A. Mundstock, J. Caro and M. Hartmann, Langmuir, 2013, 29, 8592–8600.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra12280a

---