Selective separation of CO₂–CH₄ mixed gases via magnesium aminoethylphosphonate nanoparticles

Abstract

Magnesium 2-aminoethylphosphonate (Mg₂O(2AEP) × 4H₂O) nanoparticles (particle diameter: 20–30 nm; specific surface area: 360 m² g⁻¹) are presented for selective separation of CO₂ and CH₄. Due to the base amino function, the nanoparticles can reversibly absorb CO₂ with a maximal uptake of 153 mg g⁻¹. Absorption and desorption are studied by infrared spectroscopy as well as by gravimetric sorption analysis. Furthermore, Mg₂O(2AEP) × 4H₂O shows reversible selective separation of CO₂ from CH₄. Here, pure and mixed gas adsorption isotherms (25 °C, 25 bar) of CO₂ and CH₄ show maximal uptakes of 153 mg g⁻¹ (CO₂) and 15 mg g⁻¹ (CH₄). Especially, data of mixed gas isotherms are comparably rare, but highly relevant for material characterization. Experimental isotherms were fitted by a dual-site Langmuir isotherm model (CO₂) and a Tòth model (CH₄). Mixed adsorption isotherms were modelled by volumetric-chromatographic methods resulting in a selectivity of α = 8 to 20.

---

Introduction

CO₂ is often present in natural gas reserves as a contaminant and needs to be removed as it contributes to the reduction of the energy content of methane and as it increases pipeline corrosion.¹ A routinely utilized strategy for separating CO₂ from natural gas (the so-called natural gas sweetening process) relates to chemical solvents as sorbents, including the technically most widely applied aqueous diethanolamine.² This process, however, has two major disadvantages: (i) liquid sorbents are highly corrosive (especially at increased temperature) and require special construction materials; (ii) liquid sorbents require expensive regeneration after each sorption cycle.³ Consequently, solid adsorbents are promising alternatives for CO₂ sorption and selective separation of CO₂/CH₄ mixtures.

Three different types of solid adsorbents have been already proposed for CO₂/CH₄ separation: zeolites,³ activated carbons⁴ and metal–organic frameworks (MOFs).⁵ Hereof, zeolites, in particular zeolite 13X and zeolite 5A, show the highest selectivity during equilibrium conditions.³ However, it must be emphasized that the experimental data for these adsorbents are only available for very low CO₂ concentrations (gas phase CO₂ molar fraction y_CO2 = 0.014). Natural gases and biogases, in contrast, often contain more than 40% CO₂ and N₂.⁶ Moreover, zeolites have the disadvantage of rapid saturation due to the water vapour being present in natural gas, reducing the CO₂ adsorption capacity.⁷ Finally, they need high temperatures (>300 °C) for regeneration, which increases the total-energy demand for the separation process. Nowadays, the investigation of gas sorption and separation processes is dominated by MOFs as powerful adsorbent materials.^6,8 Whereas high selectivity typically cannot be achieved just based on the unprecedented specific surface area, especially, amino-functionalized MOFs exhibit high selectivity for CO₂ separation.^6,8 Most often, studies relate to pure gases at low pressure (1 bar and below) and low temperature (room temperature and below). Moreover, the presence of humidity is a serious limitation since the availability and adsorption of H₂O drastically reduces the CO₂ uptake.

Here, we present the inorganic–organic hybrid nanomaterial magnesium aminoethylphosphonate (Mg₂O(2AEP) × 4H₂O) as a new adsorbent for selective adsorption of CO₂ and separation from CH₄. The adsorption of CO₂ on Mg₂O(2AEP) × 4H₂O and its reversibility are investigated in detail. Moreover, the CO₂/CH₄ selectivity is studied based on experimental pure-gas and mixed-gas isotherms. Moreover, gas sorption is modelled based on a dual-site Langmuir isotherm model (CO₂) and a Tòth model (CH₄). Especially, mixed-gas isotherms are rarely reported in the literature but highly relevant for evaluating the sorption characteristics and selectivity of new materials.

Experimental

Synthesis

For preparing Mg₂O(2AEP) × 4H₂O nanoparticles, Mg(n-C₄H₉)₂ solution in n-heptane (>99%, Sigma-Aldrich, Germany) and 2-aminoethyl phosphonic acid (H₂(2AEP), 99%, Acros Organics, Germany) were used as the starting materials. A transparent microemulsion was prepared by admixing of 70 mL of toluene as the non-polar phase, 3 mL of a solution of 0.27 mM of H₂(2AEP) in water as the polar phase, 1.82 g of cetyltrimethylammonium bromide (CTAB, 99%, Sigma) as the surfactant and 5 mL of n-hexane as the co-surfactant. This microemulsion system was stirred vigorously for 30 min at 35 °C (heating via oil bath). Subsequently, a solution of 1.6 mL of 0.5 M Mg(n-C₄H₉)₂ in n-heptane was injected. The formation of nanoparticles started immediately. The resulting suspension was left under stirring at room temperature for additional 12 hours. Thereafter, the nanoparticles were washed three times by sequential centrifugation/redispersion in/from ethanol. Further details regarding synthesis and fundamental materials characterization can be found elsewhere.⁹

Analytical methods

Scanning electron microscopy (SEM) was conducted on a Zeiss Supra 40 VP, using an acceleration voltage of 4 kV and a working distance of 4 mm. SEM samples were prepared by evaporation of an ethanolic suspension of the as-prepared powder. The mean particle diameter was calculated by statistical evaluation of at least 500 particles.

X-ray powder diffraction (XRD) was carried out with a Stoe Stadi-P diffractometer using Ge-monochromatized Cu-Kα1 radiation.

Fourier-transformed infrared spectroscopy (FT-IR) was performed on a Bruker Vertex 70. Samples were prepared by pestling of 2 mg of Mg₂O(2AEP) × 4H₂O with 300 mg of dried KBr in a glove-box.

Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was performed by a commercial analytical laboratory (Microanalytical Laboratory Pascher, Germany) to determine the magnesium and phosphorus content of Mg₂O(2AEP) × 4H₂O. Moreover, the nitrogen content was determined by combustion analysis.

Thermogravimetry (TG) was performed with a Netzsch STA 409C instrument, applying α-Al₂O₃ as crucible material as well as reference sample. The samples were heated in air up to 1300 °C with a heating rate of 1 K min⁻¹. The total sample weight was 30 mg.

Elementary analysis (C, H, N) was performed with an Elementar Vario EL device (Elementar, Hanau, Germany).

Gas sorption techniques

Volumetric N₂ sorption analysis of as-prepared Mg₂O(2AEP) × 4H₂O nanoparticles was carried out with a Belsorp mini II. Specific surface and pore volume were determined according to the formalisms given by Brunauer–Emmett–Teller and Barrett–Joyner–Halenda.

Gravimetric CO₂ and N₂ sorption analysis were carried out with a magnetic suspension balance (Rubotherm) that can be operated up to 200 bar. The significance of the balance is ≤0.1 mg. Thus, a stainless steel sample holder was filled with the as-prepared powder sample and the balance was evacuated for 6 h at 333 K and 10⁻³ mbar until constant mass was achieved. Afterwards, the gas was dosed into the balance chamber to elevated pressures. Equilibrium was achieved within 30 min to 2 h and identified by constant weight and pressure. The temperature was kept constant with an accuracy of ±0.5 K for each measurement. Additionally, a helium buoyancy correction was used to calculate the surface excess mass from the measured values. A detailed description of this procedure can be found elsewhere.^4–10

To investigate the uptake of CO₂/CH₄ gas mixtures, a volumetric-chromatographic method was employed.^10–17 A home-made apparatus was used to measure the adsorption equilibria at room temperature and at 1, 5 and 20 bar of pressure under different gas-phase compositions (y_CO2 = 0.05, y_CO2 = 0.25, y_CO2 = 0.75). The apparatus contained different vessels that were filled with the pure gases. To homogenize the gas mixture, a circulation pump was used. A more complete description of the apparatus can be found elsewhere.^11–18 The Mg₂O(2AEP) × 4H₂O nanoparticles were filled into the sample chamber and activated in vacuum at 80 °C for 12 hours. Afterwards, the temperature was fixed at 25 °C, and both pure gases were filled sequentially into the manifolds. In order to ensure homogeneity, a gas-chromatographic analysis was performed prior to starting the adsorption experiment. Subsequently, the valves to the sample cell were opened, and the adsorption took place while circulating the gas phase. After finishing the adsorption step, the circulation was terminated and the sample cell was closed. Gas phase concentrations were then analysed by gas chromatography (Chrompack GC CP 9001; separation column CarboPlot P7; 25 m to 0.53 mm) with sampling from the manifold. With the knowledge of pressure, temperature and gas-phase concentration, the surface-excess amount of both gases, CO₂ and CH₄, was calculated via the mass balance.

Results and discussion

Synthesis of Mg₂O(2AEP) × 4H₂O nanoparticles

To obtain Mg₂O(2AEP) × 4H₂O nanoparticles with high specific surface area as needed for efficient gas sorption and separation, we have performed the particle nucleation in the volume-restricted water pool of a water-in-oil microemulsion.⁹ Based on this measure, the particle size of Mg₂O(2AEP) × 4H₂O results in about 20–30 nm at narrow size distribution (Fig. 1). This diameter is in accordance with the data of scanning electron microscopy (SEM) and dynamic light scattering (DLS).

Fig. 1 Chemical composition (a) and particle size (according to electron microscopy (b) and dynamic light scattering (c)) of the as-prepared Mg₂O(2AEP) × 4H₂O nanoparticles.⁹

The essential functions of Mg₂O(2AEP) × 4H₂O comprise the base amino group of the organic 2-aminoethylphosphonate anion that guarantees for effective adsorption of the acidic CO₂. The phosphonate group of the [AEP]²⁻ anion, furthermore, allows obtaining an insoluble compound in water, which is prerequisite for preparing nanoparticles. Mg²⁺ was selected in view of its light weight and its frequently discussed second-order effects regarding the absorption and coordination of CO₂. Previous studies have already addressed aminoalkylphosphonates exhibiting different alkyl-chain lengths.⁹ This included aminomethyl phosphonate [AMP]²⁻, 1-aminoethyl phosphonate [1AEP]²⁻, 2-aminoethyl phosphonate [2AEP]²⁻, aminopropyl phosphonate [APP]²⁻, and aminobutyl phosphonate [ABP]²⁻. Hereof, Mg₂O(2AEP) × 4H₂O turned out as the optimal compound for CO₂ sorption.⁹ To this concern, longer alkyl chains increase the basic character of the amino group, which is preferred for CO₂ sorption, in principle. By lengthening the alkyl-chain, on the other hand, the molecular weight of the relevant compound is increased, which reduces the uptake of CO₂ in relation to the weight of the sorbent. As a consequence of these contradicting trends, Mg₂O(2AEP) × 4H₂O results as the most promising nanomaterial for selective CO₂–CH₄ separation.

According to X-ray powder diffraction Mg₂O(2AEP) × 4H₂O turned out as totally non-crystalline, which is an advantage and a disadvantage at the same time. Due to the non-crystallinity the porosity and diffusion of small molecules into the nanoparticle core can be expected to be much higher as compared to a periodically ordered, thus crystalline nanomaterial. This is definitely preferred for gas sorption and separation. On the other hand, the chemical characterization of the non-crystalline Mg₂O(2AEP) × 4H₂O is more difficult since X-ray diffraction is not applicable. To determine the chemical composition of the Mg₂O(2AEP) × 4H₂O nanoparticles, proving the presence of the functional organic [2AEP]²⁻ anion and of the inorganic [Mg₂O]²⁺ cation is naturally most important. Here, the presence of [2AEP]²⁻ was proven by infrared spectroscopy (FT-IR) (Fig. 2a). The characteristic vibrations of aminoethylphosphonate are clearly visible: ν(O–H): 3300–2900 cm⁻¹; ν(C–H): 2950–2850 cm⁻¹; ν(C–N)/δ(C–H): 1600–1200 cm⁻¹; ν(PO₃): 1200–850 cm⁻¹, δ(PO₃): 750–400 cm⁻¹. Moreover, the presence of magnesium as well as a Mg : P ratio of 2.4 : 1.0 were validated by inductively coupled plasma atomic emission spectroscopy (ICP-AES). The total thermal weight loss as determined by thermogravimetry (TG) (experimental: 40%; calculated: 41%, Fig. 2b) as well as elemental analysis (EA) (experimental: C: 10 wt%, H: 5.3 wt%, N: 4.1 wt%; calculated: C: 9 wt%, H: 5.4 wt%, N: 5.4 wt%), finally, lead to the composition [Mg₂O]²⁺[2AEP]²⁻ × 4H₂O. Details regarding the chemical characterization were reported elsewhere.⁹

Fig. 2 Infrared spectroscopy (a) and thermogravimetry (b) of the as-prepared Mg₂O(2AEP) × 4H₂O nanoparticles (FT-IR spectra with H₂(2AEP) as a reference).

In addition to the particle size and the chemical composition of the Mg₂O(2AEP) × 4H₂O nanoparticles, the fundamental material characterization addressed the specific surface area and porosity of the sample. To this concern, volumetric nitrogen sorption studies were performed and evaluated according to the Brunauer–Emmett–Teller (BET) and the Barrett–Joyner–Halenda (BJH) formalism. As a result, a specific surface area of 360(10) m² g⁻¹ and a pore volume of 0.9(1) cm³ g⁻¹ were determined.

CO₂ sorption on Mg₂O(2AEP) × 4H₂O

Mg₂O(2AEP) × 4H₂O already showed promising CO₂ adsorption (up to 150 mg g⁻¹) at high pressure (up to 150 bar) and high temperature (up to 80 °C), whereas the N₂ adsorption is negligible (<1 mg g⁻¹) over the complete pressure range (1–150 bar) (Fig. 3).¹² Although promising, however, pure-gas data are of limited significance for the much more relevant gas mixtures since interaction and interdiffusion of different molecules in mixed gases can dramatically affect the sorption processes in comparison to pure gases. Studies on mixed gases, on the other hand, are generally limited in the literature, which can be ascribed to the more complex sorption equipment and the more advanced analysis.

Fig. 3 CO₂ sorption cycles of Mg₂O(2AEP) × 4H₂O nanoparticles (gravimetric analysis at 353 K).

In general, the CO₂ sorption on Mg₂O(2AEP) × 4H₂O nanoparticles can be understood based on the following equilibrium:

R–NH₂ + CO₂ ⇆ [R–NHCOO]⁻ + [R–NH₃]⁺, (R: alkyl group) (1)

This reaction of acidic CO₂ with the base amino functions of the nanoparticles is well-known from conventional liquid-amine CO₂ absorbers (e.g. 2-aminoethanol/MEA, N-methyldiethanolamine/MDEA, 2-amino-2-methyl-1-propanol/AMP) as well as from post-synthesis amine-functionalized MOFs.^6,13 Following this reaction, Mg₂O(2AEP) × 4H₂O shows a maximum uptake of 153 mg g⁻¹ at high pressure (150 bar) (Fig. 3). This value even exceeds the expected amount of 85 mg g⁻¹ that was calculated according to eqn (1). The gravimetric CO₂ adsorption and desorption isotherms, moreover, show the principle reversibility of Mg₂O(2AEP) × 4H₂O. The thermal stability of Mg₂O(2AEP) × 4H₂O, in addition, is validated by thermogravimetry showing no significant weight loss in the 20–100 °C temperature range (Fig. 2b), which is more relevant for reversible CO₂ sorption.

Reversibility and stability of the Mg₂O(2AEP) × 4H₂O nanoparticles were furthermore tested before and after pure-gas CO₂ sorption. In view of the significance (≤0.1 mg) of the magnetic suspension balance, a certain mismatch between the 1st adsorption isotherm (starting at 0 mg g⁻¹) and the 1st desorption isotherm (finished at a residual uptake of about 19 mg g⁻¹) becomes indicative (Fig. 3). Obviously, a certain amount of CO₂ remains adsorbed on Mg₂O(2AEP) × 4H₂O subsequent to the 1st sorption cycle. This view is validated by performing a 2nd sorption cycle instantaneously after the 1st cycle. The 2nd adsorption isotherm starts at 19 mg g⁻¹ and also ends at 19 mg g⁻¹ with the 2nd desorption isotherm (Fig. 3). Hence, the 2nd CO₂ sorption cycle is fully reversible. Further sorption cycles show an identical behavior as the 2nd sorption cycle.

Based on infrared spectra, the origin of the residual weight of 19 mg g⁻¹ after the 1st CO₂ sorption cycle can be clearly attributed to remaining CO₂ (Fig. 4). Hence, the intensity of ν(C=O) vibration (1650–1500 cm⁻¹) is significantly increased after the 1st sorption cycle as compared to the as-prepared Mg₂O(2AEP) × 4H₂O nanoparticles (Fig. 4). It is to be noted that the spectra were normalized on the phosphonate valence vibration with its maximal intensity at 1099 cm⁻¹ since ν(PO₃) is independent of the CO₂ sorption. Based on this normalization, the intensity of ν(C=O) vibration can be directly compared and is now indicative for the CO₂ content in Mg₂O(2AEP) × 4H₂O before and after the CO₂ sorption. The finding of CO₂ remaining on Mg₂O(2AEP) × 4H₂O as well as the excess CO₂ uptake (85 mg g⁻¹) can be rationalized based on the following reaction (eqn (2)), which is also well-known from conventional liquid-amine CO₂ absorbers (e.g. 2-aminoethanol/MEA, N-methyldiethanolamine/MDEA, 2-amino-2-methyl-1-propanol/AMP) as well as from post-synthesis amine-functionalized MOFs.^6,13 Molar quantities of water that are intrinsically available in Mg₂O(2AEP) × 4H₂O support the following reaction and reformation of amine-functionalities.

[R–NHCOO]⁻ + H₂O ⇆ [HCO₃]⁻ + R–NH₂, (R: alkyl group) (2)

Fig. 4 FT-IR spectra of Mg₂O(2AEP) × 4H₂O nanoparticles before and after CO₂ adsorption (spectra normalized on ν(PO₃) at 1099 cm⁻¹).

Taking eqn (1) and (2) together, a maximum CO₂ uptake of 170 mg g⁻¹ is possible on Mg₂O(2AEP) × 4H₂O and almost realized with the here observed 153 mg g⁻¹. As a matter of fact, however, total CO₂ release from hydrogen carbonate ([HCO₃]⁻, eqn (2)) requires more energy as compared to the carbamate ([R–NHCOO]⁻, eqn (1)), which also explains the higher energy (i.e., evacuation or heating) that is necessary for complete reversibility.

Complete reversibility of Mg₂O(2AEP) × 4H₂O and complete release of all absorbed CO₂ – including the 19 mg g⁻¹ of CO₂ that remain absorbed after the 1st sorption cycle – is nevertheless possible. Considering the fact that the Mg₂O(2AEP) × 4H₂O nanoparticles are still in contact with a CO₂ atmosphere at 1 bar after the 1st sorption cycle, removing the CO₂ atmosphere and evacuation might be one option. As the decomposition of [CO₃]²⁻ according to eqn (2) requires a higher energy for activation, heating of the Mg₂O(2AEP) × 4H₂O nanoparticles in nitrogen could be another option.

Indeed evacuation as well as heating allow desorbing largely all CO₂ from the Mg₂O(2AEP) × 4H₂O nanoparticles. To this concern, evacuation was performed at elevated temperature (10⁻³ mbar, 333 K, 6 h) subsequent to the 1st sorption cycle (Fig. 5a). Alternatively, the CO₂ atmosphere can be replaced by helium after the 1st sorption cycle with the Mg₂O(2AEP) × 4H₂O nanoparticles heated thereafter (1 bar He atmosphere, 353 K, 12 h) (Fig. 5b). Via both measures 13 mg g⁻¹ (evacuation) and 16 mg g⁻¹ (heating in He atmosphere) of CO₂ can be removed. Notably the pressure reduction during evacuation needs to be performed slowly (30 min) to avoid dusting inside of the suspension balance. In sum, 96–98% of CO₂ (i.e. 146 to 149 mg g⁻¹ of 153 mg g⁻¹) adsorbed on the Mg₂O(2AEP) × 4H₂O nanoparticles can be reversibly desorbed when including additional evacuation or heating. At least 88% of CO₂ (i.e. 134 mg g⁻¹ of 153 mg g⁻¹) can be reversibly adsorbed and desorbed just by pressure swing cycles between 1 and 110 bar.

Fig. 5 Release of residual CO₂ from Mg₂O(2AEP) × 4H₂O nanoparticles after the 1st sorption cycle upon (a) evacuation (10⁻³ mbar, 333 K, sample weight 92.4 mg) or (b) heating in helium atmosphere (1 bar, 353 K, sample weight 100.8 mg).

CO₂–CH₄ pure gas isotherms

Aiming at CO₂ and CH₄ sorption and separation, first of all, pure-gas adsorption isotherms of Mg₂O(2AEP) × 4H₂O were measured gravimetrically at 25 °C (Fig. 6). At 25 bar, the nanoparticles adsorb 3.47 mmol g⁻¹ (153 mg g⁻¹) of CO₂, whereas they only absorb 0.9 mmol g⁻¹ (14.5 mg g⁻¹) CH₄ under similar conditions. Acidic CO₂ is preferably adsorbed on Mg₂O(2AEP) × 4H₂O nanoparticles over CH₄ due to chemisorption on the base amino-groups present in Mg₂O(2AEP) × 4H₂O (eqn (1)). In contrast to CO₂ but similar to N₂, CH₄ is excluded from this sorption mechanism.

Fig. 6 CO₂ (blue circles) and CH₄ (red circles) isotherms at 25 °C on Mg₂O(2AEP) × 4H₂O nanoparticles. Modelling of the experimental data with dual-site Langmuir model (CO₂, blue solid line) and Tòth model (CH₄, red solid line). Calculation of CO₂/CH₄ mixed isotherms performed according to IAST (solid green line: y_CO2 = 0.75; dotted green line: y_CO2 = 0.25; dashed green line: y_CO2 = 0.05).

Although the adsorption characteristics of pure CO₂ and pure CH₄ are promising, data of CO₂/CH₄ gas mixtures are needed for applied process design. To model and predict binary adsorption equilibria from pure gas adsorption isotherms, we have applied the ideal adsorbed solution theory (IAST) first.¹⁴ To this regard, a Tòth model¹⁵ turned out to be suitable for the description of pure CH₄ isotherms, whereas the experimental data for CO₂ were fitted using a dual-site-Langmuir model¹⁶ (Fig. 6). These two models were used within the IAST, and the calculated data were compared with experimental sorption isotherms from CO₂/CH₄ mixed gas systems. In Fig. 6, the green lines represent the IAST calculation of three different CO₂/CH₄ mixtures with y_CO2 = 0.05, y_CO2 = 0.25 and y_CO2 = 0.75.

CO₂–CH₄ mixed gas isotherms

Next, the adsorption of CO₂ and CH₄ mixtures was experimentally investigated at room temperature and at three different constant pressures (1, 5, 20 bar) by volumetric-chromatographic measurements. Here, the prediction from IAST and the experimental data are in excellent agreement (Fig. 7). Over the whole range of pressures and gas mixtures, adsorption of CO₂ on Mg₂O(2AEP) × 4H₂O is preferred over CH₄, confirming the interaction of CO₂ with the amino-groups available in the nanomaterial.⁶ To evaluate the potential of Mg₂O(2AEP) × 4H₂O regarding CO₂/CH₄ separation, the selectivity (α) was considered which is defined by:
herein, x is the molar fraction of the respective gas in the adsorbate; y is the molar fraction of the respective gas in the gas phase. The dimensionless selectivity indicates how much CO₂ is enriched in the adsorbate (x_CO2) as a function of the gas-phase composition at constant pressure and temperature. If α is higher than 1, CO₂ is preferentially adsorbed. On the contrary, CH₄ adsorption is preferred if α is lower than 1. For Mg₂O(2AEP) × 4H₂O nanoparticles, the experimental as well as the theoretical adsorption show nearly constant selectivity at low and high CO₂ gas phase molar fractions, ranging from α = 8 to α = 20 (Fig. 8). The IAST predictions are well in agreement with the experimental data and show values of α between 11 and 15.

Fig. 7 x,y-Diagram of binary adsorption of CH₄ and CO₂ on Mg₂O(2AEP) × 4H₂O nanoparticles at 25 °C and different pressures (blue circles: 1 bar, red triangles: 5 bar, green diamonds: 20 bar). IAST calculation from pure gas isotherms added as lines (blue: 1 bar, red: 5 bar, green: 20 bar).

Fig. 8 Selectivity of CO₂ adsorption from binary CO₂/CH₄ mixtures as a function of gas phase molar fraction of CH₄ during adsorption on Mg₂O(2AEP) × 4H₂O nanoparticles at different pressures (blue circles: 1 bar, red triangles: 5 bar, green diamonds: 20 bar). IAST calculation from pure gas isotherms added as lines (blue: 1 bar, red: 5 bar, green: 20 bar).

When evaluating the sorption properties of the Mg₂O(2AEP) × 4H₂O nanoparticles, the CO₂ selectivity for CO₂/CH₄ mixtures at high CO₂ content (>20%) exceeds many known materials, including zeolites, activated carbons and MOFs (Table 1). With regard to MOFs, only the amino-functionalized MOFs MIL-101(Al) and NH₂-MIL-53(Al), especially at low pressure (<10 bar) show a selectivity higher than Mg₂O(2AEP) × 4H₂O.^5e,f On the other hand, MOFs often need advanced synthesis. Moreover, they are sensitive to humidity (i.e. blocking the CO₂ sorption) and partly hydrolyzed by water (i.e. restricting the chemical stability).¹⁷ On the contrary, Mg₂O(2AEP) × 4H₂O nanoparticles are prepared in water. In contrast to zeolites and activated carbons, moreover, Mg₂O(2AEP) × 4H₂O can be easily regenerated since the absorbed CO₂ is desorbed either by slight heating (80 °C) and or by decompression.

Table 1 Adsorption selectivity of CO₂ over CH₄ of selected sorbent materials^a

Adsorbent	Adsorption conditions	Gas mixture composition	Selectivity α (CO2/CH4)	Pure CO2 uptake	Ref.
a † = excess adsorption; ‡ = absolute adsorption; †‡ = reference is not clear in whether absolute or excess adsorption is given.
Mg2O(2AEP) × 4H2O	1–20 bar, 25 °C	yCO2 = 0.05	8–20	153 mg g^−1, (25 °C, 25 bar)^†	This study
		yCO2 = 0.25
		yCO2 = 0.75
HKUST-1, (Cu3(BTC)2)	1–10 bar, 30 °C	yCO2 = 0.25	5–12	616 mg g^−1, (30 °C, 40 bar)^‡	5a
		yCO2 = 0.50
		yCO2 = 0.75
Zn2(ndc)2(dpni)	0.1–10 bar, 23 °C	yCO2 = 0.5	4–30	220 mg g^−1, (23 °C, 18 bar)^†	5b
MIL-53(Cr)	10 bar, 30 °C	yCO2 = 0.25, 0.50, 0.75	2–15	500 mg g^−1, (30 °C, 2 bar)^†‡	5c
MIL-100(Cr)	1–10 bar, 30 °C	yCO2 = 0.25, 0.50, 0.75	6–8	350 mg g^−1, (30 °C, 10 bar)^†‡	5d
NH2-MIL-53(Al), (amino-funct.)	1/5/45 bar, 30 °C	yCO2 = 0.4	60–100/20/2–3	310 mg g^−1, (30 °C, 40 bar)^†‡	5e
MIL-101(Al), (amino-funct.)	1 bar, 25 °C	yCO2 = 0.2	15–60	620 mg g^−1, (25 °C, 30 bar)^†‡	5f
TB-MOF, ([Me2NH2][Tb3L3(HCOO) × DMF × 15H2O])	0.1–1 bar, 20 °C	yCO2 = 0.5	11–17	78 mg g^−1, (20 °C, 1 bar)	18a
[Cu3(CPT)4(μ3-OH)] × NO3 × 7H2O	0.1–1 bar, 25 °C	yCO2 = 0.5	18–36	118 mg g^−1, (25 °C, 1 bar)	18b
Zeolite 13X, (NaX)	10–40 bar, 25 °C	yCO2 = 0.014	140–160	145 mg g^−1, (50 °C, 1 bar)	3 and 6
				325 mg g^−1, (25 °C, 32 bar)^†‡
Zeolite 5A, (CaA)	10–40 bar, 25 °C	yCO2 = 0.014	250–300	150 mg g^−1, (30 °C, 1 bar)	3 and 19
Active carbon, (AC A35/4)	1–15 bar, 25 °C	yCO2 = 0.2	3–9	352 mg g^−1, (20 °C, 10 bar)	4
		yCO2 = 0.5
		yCO2 = 0.9
Boron nitride	>227 °C	CO2, CH4, H2	>10	80 kcal mol^−1 adsorption energy	20

---

Conclusions

In conclusion, the inorganic–organic hybrid nanomaterial Mg₂O(2AEP) × 4H₂O is a highly promising alternative to existing adsorbents for CO₂ and selective separation from CH₄. The experimental data of different CO₂/CH₄ gas mixtures and theoretical adsorption models show nearly constant selectivity at low as well as at high CO₂ gas phase molar fractions ranging from α = 8 to 20. For the first time, adsorption and selectivity of CO₂/CH₄ mixtures (y_CO2 = 0.05, 0.25, 0.75) are experimentally determined and modelled based on a dual-site Langmuir isotherm model (CO₂) and a Tòth model (CH₄). The results from both approaches are in good agreement. Based on these results, Mg₂O(2AEP) × 4H₂O nanoparticles may gain considerable interest for selective separation of CO₂ as a contaminant in natural gas. The solid nanomaterial possesses basic properties and can be used in alternative to the widely applied but corrosive liquid amines. Mg₂O(2AEP) × 4H₂O is even more interesting due to its uncomplex synthesis in water, its low sensitivity to water, and its straightforward regeneration via gentle heating and/or decompression.

Acknowledgements

The authors are grateful to the Deutsche Forschungsgemeinschaft (DFG) for funding of analytical equipment.

Notes and references

1. Y. S. Bae, O. K. Farha, A. M. Spokoyny, C. A. Mirkin, J. T. Hupp and R. Q. Snurr, Chem. Commun., 2008, 35, 4135.
2. G. Astarita, D. W. Savage and A. Bisio, Gas treating with chemical solvents, Wiley, Chichester, 1983.
3. P. D. Rolniak and R. Kobayashi, AIChE J., 1980, 26, 616.
4. M. Heuchel, G. M. Davies, E. Buss and N. A. Seaton, Langmuir, 1999, 15, 8695.
5. (a) L. Hamon, E. Jolimaitre and G. D. Pirngruber, Ind. Eng. Chem. Res., 2010, 49, 7497; (b) Y. Bae, K. Mulfort, H. Frost, P. Ryan, S. Punnathanam, L. Broadbelt, J. T. Hupp and R. Q. Snurr, Langmuir, 2008, 24, 8592; (c) L. Hamon, P. L. Llewellyn, T. Devic, A. Ghoufi, G. Clet, V. Guillerm, G. D. Pirngruber, G. Maurin, C. Serre, G. Driver, W. van Beek, E. Jolimaıtre, A. Vimont, M. Daturi and G. Ferey, J. Am. Chem. Soc., 2009, 131, 17490; (d) L. Hamon, N. Heymans, P. L. Llewellyn, V. Guillerm, A. Ghoufi, S. Vaesen, G. Maurin, C. Serre, G. De Weireld and G. D. Pirngruber, Dalton Trans., 2012, 41, 4052; (e) S. A. Peter, G. V. Baron, J. Gascon, F. Kapteijn and J. F. M. Denayer, Adsorption, 2013, 19, 1235; (f) P. Serra-Crespo, E. V. Ramos-Fernandez, J. Gascon and F. Kapteijn, Chem. Mater., 2011, 23, 2565.
6. D. M. D'Alessandro, B. Smit and J. R. Long, Angew. Chem., Int. Ed., 2010, 49, 6058.
7. G. Li, P. Xiao, P. Webley, J. Zhang and R. Singh, Energy Procedia, 2009, 1, 1123.
8. (a) S. Li and F. Huo, Nanoscale, 2015, 7, 7482; (b) Z. Zhang, Z.-Z. Yao, S. Xiang and B. Chen, Energy Environ. Sci., 2014, 7, 2868; (c) S. Chaemchuen, N. A. Kabir, K. Zhou and F. Verpoort, Chem. Soc. Rev., 2013, 42, 9304; (d) X. Wang, H. Li and X.-J. Hou, J. Phys. Chem. C, 2012, 116, 19814.
9. J. G. Heck, J. Napp, S. Simonato, J. Möllmer, M. Lange, H. R. Reichardt, R. Staudt, F. Alves and C. Feldmann, J. Am. Chem. Soc., 2015, 137, 7329.
10. (a) J. U. Keller and R. Staudt, Gas Adsorption Equilibria, Springer, Berlin, 2005; (b) F. Dreisbach and H. W. Losch, J. Therm. Anal. Calorim., 2000, 62, 515; (c) J. Moellmer, A. Moeller, F. Dreisbach, R. Glaeser and R. Staudt, Microporous Mesoporous Mater., 2011, 138, 140.
11. J. Möllmer, M. Lange, A. Möller, C. Patzschke, K. Stein, D. Lässig, J. Lincke, R. Gläser, H. Krautscheid and R. Staudt, J. Mater. Chem., 2012, 22, 10274.
12. P. Leidinger, S. Simonato and C. Feldmann, Chem. Commun., 2012, 48, 7046.
13. (a) S. Choi, J. H. Drese and C. W. Jones, ChemSusChem, 2009, 2, 796; (b) S. Ma and H. C. Zhou, Chem. Commun., 2010, 46, 44; (c) D.-Y. Hong, Y. K. Hwang, C. Serre, G. Ferey and J.-S. Chang, Adv. Funct. Mater., 2009, 19, 1537.
14. A. L. Myers and J. M. Prausnitz, AIChE J., 1965, 11, 121.
15. J. Tòth, Adv. Colloid Interface Sci., 1995, 55, 1.
16. P. M. Mathias, R. Kumar, J. D. Moyer, J. M. Schork, S. R. Srinivas-an, S. R. Auvil and O. Talu, Ind. Eng. Chem. Res., 1996, 35, 2477.
17. J. Liu, P. K. Thallapally, B. P. McGrail, D. R. Brown and J. Liu, Chem. Soc. Rev., 2012, 41, 2308.
18. (a) Y. Yang, F. Jiang, L. Chen, J. Pang, M. Wu, Y. Wan, J. Pan, J. Qian and M. Hong, J. Mater. Chem. A, 2015, 3, 13526; (b) D. M. Chen, X. P. Zhang, W. Shi and P. Cheng, Inorg. Chem., 2015, 54, 5512.
19. Z. Liu, C. A. Grande, P. Li, J. G. Yu and A. E. Rodrigues, Sep. Sci. Technol., 2011, 46, 434.
20. Q. Sun, Z. Li, D. J. Searles, Y. Chen, G. Q. Lu and A. J. Du, J. Am. Chem. Soc., 2013, 135, 8246.

---