CO₂ capture using silica-immobilized dicationic ionic liquids with magnetic and non-magnetic properties

Abstract

The need to find alternative materials to replace aqueous amine solutions for the capture of CO₂ in post-combustion technologies is pressing. This study assesses the CO₂ sorption capacity and CO₂/N₂ selectivity of three dicationic ionic liquids with distinct anions immobilized in commercial mesoporous silica support (SBA- 15). The samples were characterized by UART-FTIR, NMR, Raman, FESEM, TEM, TGA, Magnetometry (VSM), BET and BJH. The highest CO₂ sorption capacity and CO₂/N₂ selectivity were obtained for sample SBA@DIL_2FeCl₄ [at 1 bar and 25 °C; 57.31 (±0.02) mg CO₂/g; 12.27 (±0.72) mg CO₂/g]. The results were compared to pristine SBA-15 and revealed a similar sorption capacity, indicating that the IL has no impact on the CO₂ sorption capacity of silica. On the other hand, selectivity was improved by approximately 3.8 times, demonstrating the affinity of the ionic liquid for the CO₂ molecule. The material underwent multiple sorption/desorption cycles and proved to be stable and a promising option for use in industrial CO₂ capture processes.

1. Introduction

Greenhouse gas (GHG) emissions are a growing concern, primarily due to high anthropogenic emissions from the burning of coal, natural gas, and oil [1]. The increase in the greenhouse effect has environmental, social, and economic impacts. Humanity is already experiencing the rising of land and ocean temperatures, resulting in more frequent cyclones, floods, heat waves, droughts, forest fires, dust storms, and desertification [2,3]. The main gases contributing to the increase in the greenhouse effect are carbon dioxide (CO₂), responsible for 75 % of emissions, methane (CH₄) 17 %, nitrous oxide (N₂O) 6 %, and fluorinated gases 2 % [4]. Energy production and industry are responsible for about 64 % of CO₂ emissions [5].

CO₂ emissions have increased by more than 50 % since the beginning of the industrial era and reaching a global record of 421 ppm in May 2022 [6]. The increase in atmospheric CO₂ concentration is responsible for two-thirds of the total energy imbalance, leading to a rise in the temperature of the planet [7]. A portfolio of options is available to mitigate CO₂ concentrations in the atmosphere. CO₂ capture and storage (CCS) is one of the most promising technologies to mitigate CO₂ emissions from industry and power plants. The main drawback of CCS is the high cost of the capture step [8]. In the Carbon Capture and Utilization (CCU) process, CO₂ is captured and used as raw material to produce fuels and chemicals [9]. To mitigate CO₂ emissions, a capture step can be inserted in the existing pre-combustion, oxy-combustion and post-combustion processes. In post-combustion, the capture step can be easily adapted to existing power plants without the need for radical changes, obtaining high efficiency in the carbon capture process [10], removing CO₂ after the burning of the fossil fuel [11].

Chemical absorption by amine aqueous solutions is the benchmark technology for CO₂ capture [12], with a high absorption capacity, fast reaction kinetics, and low cost [13]. However, some drawbacks are considered, such as the high energy requirement for the solvent regeneration process, the corrosivity of the equipment, and the high degradability, making the process more expensive [14].

Aiming to overcome the problems related to the use of amines in CO₂ capture, other technologies are studied and applied, such as polymer-based membranes [[15], [16], [17]] and adsorbents, mainly zeolites [[18], [19], [20]], silicas [[21], [22], [23]] and activated carbon [[24], [25], [26]]. Adsorbent materials are good candidates for carbon capture processes, due to their porosity, high specific surface area, besides the competitive advantages related to the low energy consumption in the regeneration process, fast adsorption kinetics, non-corrosivity and good regenerability in the adsorption/desorption process. In general, solid adsorbents have low selectivity to CO₂ [[27], [28], [29]]. Higher selectivity can be achieved by supporting CO₂ selective molecules in solid support [30].

Ionic liquids (ILs) are formed by organic cations and organic or inorganic anions. The myriad of available anions and cations allows for a large number of combinations and the design of ILs with desired properties [31]. It is well known from the literature that ILs are selective for CO₂ [32]. Immobilization of ILs in solid supports can improve the affinity for CO₂, in addition to solving the drawbacks of ILs related to the high viscosity and low adsorption kinetics. The formation of thin ILs films in the solid supports enhances the mass transfer process [33].

The application of dicationic ionic liquids (DIL) and tricationic ionic liquids (TIL) is being investigated in various research fields, such as extraction [34], liquid chromatography [35], catalysis [36] and gas separation [37]. Multicationic ILs are seen as a potential alternative that could enhance the functionality of ILs, compared to traditional monocationics [[38], [39], [40]].

When an IL contains metal, such as transition metals or rare earth ions, in the anion, it is called a magnetic ionic liquid (MIL) [41]. These materials possess the properties of ionic liquids and the ability to respond magnetically to an external field, making them magneto-responsive materials [42]. The viscosity of MILs could decrease when subjected to an external magnetic field, which is a crucial characteristic for their use in gas separation processes [43].

The synthesis and immobilization of three imidazolium-based dicationic ionic liquids on a commercial silica support were described in this study. A polyethyleneglycol chain was used to connect the cations and different anions. CO₂ sorption capacity and CO₂/N₂ selectivity were evaluated for all samples and compared to the monocationic ILs described in the literature.

2. Experimental

2.1. Materiais

Triethylene glycol (TEG, ≥99.0 %, Sigma-Aldrich, China), sodium p-toluenesulfonate (TsOCl, ≥98.0 %, Sigma-Aldrich, China), potassium hydroxide (KOH, ≥85 %, VETEC, Brazil), (3-Chloropropyl) triethoxysilane (CPTES, ≥95.0 %, Sigma-Aldrich, U.S.A), imidazole (≥99.0 %, Sigma-Aldrich, Germany), iron (III) chloride (FeCl₃, ≥97.0 %, Sigma-Aldrich, Germany), lithium bis(trimethylsilyl)amide (LiNTf₂, ≥99.0 %, Sigma-Aldrich, China), SBA-15 (≥99.0 %, Sigma-Aldrich, China), pellets of sodium hydroxide (NaOH, ≥95.0 %, Sigma-Aldrich, U.S.A), magnesium sulfate (MgSO₄, ≥98.0 %, VETEC, Brazil), toluene (≥99.0 %, VETEC, Germany), dichloromethane (CH₂Cl₂, P.A, Synth, Brazil), chloroform (CHCl₃, P.A, Synth, Brazil) and acetonitrile (CH₃CN, ≥99.0 %, Merck, U.S.A). The organic solvents were previously purified prior to the reactions.

2.2. Syntheses of dicationic IONIC liquids

2.2.1. Synthesis of triethylene glycol di(p-Toluenesulfonate) (TsOTEGTsO)

Synthesis was performed following the procedure described in the literature [44]. In a round bottom flask, a solution of 16.55 g of TEG in 80 mL of dichloromethane was added to 53.60 g of TsOCl at room temperature. This mixture was cooled to 0 °C, then 58.20 g of KOH was added, in portions, and the reaction was stirred for 10 min, then allowed to warm to room temperature and kept under stirring for 12 h obtaining a heterogeneous solution. 50 mL of deionized water was added, diluting the previously formed white solid (TsOTEGTsO). The product was then extracted with chloroform, dried with magnesium sulfate, filtered and kept under vacuum, yielding 88 %. A schematic representation of the reaction synthesis is presented in Fig. 1. Characterizations are in agreement with the literature [44]. ¹H RMN (CDCl₃, 400 MHz, δ em ppm): 2.46 (6H, s, CH₃), 3.67 (4H, t, OCH₂CH₂O), 3.54 (4H, s, CH₂), 4.15 (4H, t, CH₂), 7.35 (4H, d, CH), 7.81 (4H, d, CH). 13C RMN (CDCl₃, 400 MHz, δ em ppm): 21.62 (2C, s, -CH₃), 68.75 (2C, s, S–O–CH₂–CH₂–O), 69.21 (2C, s, S–O–CH₂–CH₂–O), 70.68 (2C, s, O–CH₂–CH₂–O), 127.95 (4C, s, CH_arom.), 129.86 (4C, s, CH_arom.), 133.03 (2C, s, CH_arom.), 144.86 (2C, s, CH_arom.).

Fig. 1.

Schematic representation of the synthesis of TsOTEGTsO.

2.2.2. Synthesis of 1,8-diimidazole-3,6-dioxaoctane (ImTEGIm)

Synthesis was performed following the procedures described elsewhere [45]. In a three-neck flask, 13.07 g of imidazole was melted at approximately 100 °C and then 6.98 g of NaOH was added. After complete dissolution, 50 mL of dry toluene was added to form an azeotropic mixture with the water generated by the reaction, stirred for 3 h, and the toluene/water mixture was removed under vacuum. To the imidazole sodium obtained, 40 g of TsOTEGTsO diluted in 100 mL of acetonitrile was added and refluxed for 24 h. The solvent was removed by simple distillation and the remaining product in the flask was washed in dichloromethane separating Im TEGIm and sodium p-toluenesulfonate. The liquid extracted with dichloromethane was washed with deionized water, and the aqueous solution was kept under vacuum under heating at 60 °C to remove the solvent, remaining only a viscous liquid of light brown color, yielding 73 %. The schematic representation is given in Fig. 2. The characterizations are in agreement with the literature [44,45]. 1H RMN (D₂O, 400 MHz, δ em ppm): 3.51(4H, s, CH₂), 3.66 (4H, t, OCH₂CH₂O), 4,07 (4H, t, CH₂), 6.95 (4H, d, CH), 7.51 (2H, s, CH). 13C RMN (CDCl₃, 400 MHz, δ em ppm): 46.81 (1C, d, N–CH₂– CH₂–O), 68.30 (1C, d, N–CH₂CH₂–O), 69.69 (2C, q, OCH₂CH₂O),120.35 (2C, d, -CH CH-), 129.43 (2C, s, –CH CH-), 138.13 (2C, s, N–CH N).

Fig. 2.

Schematic representation of the synthesis of ImTEGIm.

2.2.3. Synthesis of dichlorate 1,8-diimidazole-3,6-dioxaoctane-bis- (propyl)triethoxysilane (DIL_2Cl)

The experimental procedure was performed according to the literature [46]. 7.10 g of ImTEGIm, 20 mL of dry toluene and 14.06 g of CPTES (added slowly under an inert medium) were added in a three-neck flask. The mixture was then stirred for 24 h at 65 °C. After completion, two phases appeared, attributed to DIL_2Cl and toluene. DIL_2Cl was cleaned with anhydrous toluene and evaporated under vacuum, resulting in a highly viscous liquid with a yield of 87 %. The illustration of the synthesis process can be viewed in Fig. 3, and the results of the characterizations match the findings previously documented in the literature [[44], [45], [46]]. 1H RMN (D₂O, 400 MHz, δ em ppm): 0.60 (2H, s, -Si-CH₂-CH₂-), 1.97 (t, 4H, -Si–CH₂–CH₂-CH₂-), 2.39 (s, 12H, CH₃–CH₂–O–Si-), 3.59 (s, 9H, CH₃–CH₂–O–Si-), 3.63 (t, 8H, -CH₂CH₂OCH₂CH₂OCH₂ CH₂-), 3.78 (t, 4H, N–CH₂–CH₂-), 4.21 (q, 2H, -Si–CH₂–CH₂-CH₂-N-), 7.07 (d, 1H, -N- CH CH–N-), 7.20 (d, 1H, -N-CH CH–N-), 7.70 (s, 1H, -N-CH N-). 13C RMN (D₂O, 400 MHz, δ em ppm): 8.40 (2C, d, O–Si–CH₂–CH₂–CH₂–N), 20.52 (6C, s, CH₂–CH₂–O–Si), 23.39 (6C, s, CH₂–CH₂–O–Si), 46.81 (2C, d, O–Si–CH₂–CH₂–CH₂–N), 49.15 (2C, q, O–Si–CH₂–CH₂–CH₂–N), 51.67 (2C, t, N–CH₂–CH₂–O), 68.35 (2C, d, N–CH₂– CH₂–O), 69.63 (2C, s, O–CH₂–CH₂–O), 120.46 (2C, d, N–CH CH–N), 129.43 (2C, s, N–CH CH–N), 137.95 (2C, d, N–CH N).

Fig. 3.

Schematic representation of the synthesis of DIL_2Cl.

2.3. Chemical immobilization OF IONIC liquids in solid supports and anions exchange

The procedure for immobilization of 10 % DIL_2Cl on solid support (SBA-15) was based on the literature [47]. The solid support was previously dried in an oven at 110 °C for 2 h. Then 1.55 g was weighed directly into a three-neck flask and 30 mL of dry toluene was added and the mixture was stirred at room temperature for 2 h. 0.17 g of DIL_2Cl was diluted in 5 mL of deionized water and added in toluene at 95 °C and kept for 24 h, in a reflux system, producing the SBA@DIL_2Cl, as seen in Fig. 4.

Fig. 4.

Schematic representation of the synthesis of hydrolysis and condensation of DIL_2Cl for immobilization on the adsorbent support, forming SBA@DIL_2Cl.

The SBA@DIL_2Cl anion exchange reaction (Fig. 5) was performed according to a procedure adapted from the literature [48,49]. The mixture of LiNTf₂ salt (in 50 mL of distilled water) and immobilized DIL_2Cl (2:1 ratio) was stirred at room temperature for 78 h. Subsequently, the mixture was filtered, repeatedly washed with deionized water, and extracted with soxhlet using first deionized water and dichloromethane, which removed the unreacted products and produced SBA@DIL_2NTf₂.

Fig. 5.

Anion exchange of the SBA@DIL_2Cl for SBA@DIL_2NTf₂, SBA@DIL_2NTf₂/FeCl₄ and SBA@DIL_2FeCl₄.

To obtain FeCl₄‾ as an anion, FeCl₃ (dissolved in acetonitrile) and DIL_2Cl (2:1) were added to SBA@DIL_2Cl and stirred in an inert medium at 60 °C for 48 h, obtaining SBA@DIL_2FeCl₄. The product was filtered, washed with deionized water and soxhlet extracted with deionized water to remove unreacted products.

To obtain a DIL containing FeCl₄/NTf₂ as an anion, a partial exchange was first performed with the LiNTf₂ salt, in a 1:1 ratio with the SBA@DIL_2Cl, following the same procedures reported above to obtain the NTf₂ anion. The remaining chloride anions were exchanged for the FeCl₄‾ anionic complex, using the same procedure described before, except for FeCl₃: SBA@DIL_2Cl ratio (1:1) obtains SBA@ DIL_NTf₂/FeCl₄.

2.4. Samples characterization

The structures of the dicationic ionic liquids (DILs) were analyzed using a range of techniques. An infrared spectrometer, FTIR PerkinElmer Spectrum One, with a Universal Attenuated Total Reflectance sensor (UATR-FTIR) was used for the analysis. Additionally, a Bruker Advance DRX-400 spectrometer operating at 400 MHz was used to perform the 1H and 13C NMR analyses in the liquid state and the ²⁹Si MAS-NMR in the solid state. The Brunauer-Emmett-Taller (BET) method was used to analyze the specific surface area, and the Barrett-Joyner-Halenda (BJH) method was used to determine the pores volumes and sizes, with the aid of the Quantachrome equipment New - Surface Area & Pore Size Analyzer. To determine the thermal stability and the percentage of DIL and magnetic dicationic ionic liquid (MDIL) grafted onto the support, a thermogravimetric analysis (TGA) was performed under a nitrogen atmosphere between 25 and 800 °C and a heating rate of 20 °C/min, using TA Instruments Discovery SDT 650. Equation 1: IL % = [(W₁₆₇–W₈₀₀)/W₁₆₇] × 100 was used to calculate the actual percentage of IL grafted on the support, where W₁₆₇ and W₈₀₀ are the sample weights at 167 °C and 800 °C, respectively [50]. To gain further insight into the mesopore materials, a range of microscopy techniques were utilized. The Field Emission Scanning Electron Microscope (FESEM) was used, with the FEI Inspect F50 equipment in secondary electron mode (SE), to perform microscopy analyses. The hexagonal structures of the support with and without DILs were confirmed using transmission electron microscopy (TEM) with aTecnai G2 T20 FEI equipment operating at 200 kV. The anionic iron complexes contained in the supports were characterized using Raman spectroscopy, which was performed with a Horibam LabRamHr Evolution Laser Raman Spectrometer, model DXR (laser excitation wavelength of 532 nm), and Access alpha 300 (632.8 nm - micro- Raman single-spot analysis and mapping microscope). Finally, the magnetic properties were analyzed using a Vibrant Sample Magnetometer (VSM), model EZ9 by MicroSense, to evaluate the impact of an external magnetic field on the samples.

2.5. CO₂ sorption measurements

The CO₂ sorption capacity was determined using the pressure decay method, as described in the literature [51]. The experiment was carried out three times to ensure accuracy. The sample, which weights approximately 0.6 g, was dried in an oven at 80 °C for 1.5 h prior to the test. The CO₂ sorption tests were carried out at 25 °C and the amount of adsorbed CO₂ was calculated using the method described in our previous works [52,53].

2.6. CO₂/N₂ selectivity tests

Selectivity experiments were carried out at 25 °C and 20 bar in a dual-chamber gas sorption cell based on Koros et al. [51] using a binary mixture (15.89 mol % of CO₂ and a balance of N₂). Gas chromatography with a thermal conductivity detector was used to determine the composition at the exit of the sorption system, obtaining the non-adsorbed CO₂/N₂ fraction. To calculate the separation efficiency, the procedure described in the literature was followed [52,54]. The molar fractions were determined in the gas phase (Yi) and the sample phase (Xi) according to equation (2):

$$S=\frac{XCO2/YCO2}{XN2/YN2}$$ (2)

3. Results and discussion

3.1. Precursors and IONIC liquids characterization

The IR spectra in Fig. 6 provide valuable information about the molecular structure of the synthesized molecules, confirming the success of the synthesis and supporting the NMR results. In Fig. 6a, TEG, the starting molecule of the reaction, shows characteristic absorption of O–H stretching bands (3388 cm⁻¹), aliphatic C–H stretching (2925 - 2869 cm⁻¹), folding of aliphatic C–H (1455 - 1351 cm⁻¹), and C–O–C stretching of ether group (1116 and 1056.9 cm⁻¹) [46]. Fig. 6b presents the TsOTEGTsO spectrum that presents characteristic band vibrations of aromatic C–H elongation (2959 - 2929 cm⁻¹), C C aromatic (1596 - 1348 cm⁻¹) and aliphatic S O elongation (1169 cm⁻¹), indicating the presence of the p-toluenesulfonate group [55]. The aliphatic C–H elongation band (2899 - 2872 cm⁻¹), and the stretching C–O–C of the ether group (1133–1092 cm⁻¹) are alsoobserved. The disappearance of the O–H stretching indicates the formation of TsOTEGTsO (3388 cm⁻¹).

Fig. 6.

FT-IR spectra of a) TEG (pure), b) TsOTEGTsO, c) Im TEGIm, d) DIL_2Cl.

From Fig. 6C, one can see the typical imidazolium bands that replace the p-toluenesulfonate. Aromatic = C–H elongation (3111 cm⁻¹), aliphatic C–H elongation (2911 - 2869 cm⁻¹), C N elongation (1673 cm⁻¹), C C elongation (1506–1440.5 cm⁻¹) and C–N elongation (1229 cm⁻¹). Also, the characteristic band of the C–O–C of TEG is maintained (1077 cm⁻¹).

Fig. 6d, after CTPES addition in ImTEGIm an intense band characteristic of Si–O–C (1035 cm⁻¹) and Si–C elongation (817 cm⁻¹) can be seen, confirming the obtainment of DIL_2Cl.

3.2. Supported dicationic IONIC liquids - characterization

The immobilization of the ionic liquid on the SBA-15 mesoporous support was evaluated by ²⁹Si MAS-NMR (Fig. 7). The pristine mesoporous silica (Fig. 7a) spectra show three typical resonance frequencies at −110 ppm, - 100 ppm and −92 ppm, assigned to [Si (OSi)₄] (Q⁴), [Si (OSi)₃ OH] (Q³) and [Si(OSi)₂ (OH)₂] (Q²) groups of the SBA-15. After immobilization of the DILs [10 % (w/w)] (see Fig. 7 b, c and d) a reduction in the bands Q², Q³ and Q⁴ was observed, indicating the reaction of SBA-15 hydroxyl groups with the silane groups present in the DILs that confirms the immobilization of the DIL in the solid support.

Fig. 7.

²⁹Si MAS-NMR spectra of a) SBA-15 (pure), b) SBA@DIL_2NTf₂ 10 %, c) SBA@DIL_NTf₂/FeCl₄ 10 %, d) SBA@DIL_2FeCl₄ 10 %.

Fig. 8 presents the ²⁹Si MAS-RMN spectra for a sample with 20 % of DIL immobilized in SBA-15. This sample with a higher DIL content was synthesized to improve the signal related to IL and confirm the immobilization of DIL. A reduction in Q² and Q³ bands and the appearance of T¹ and T³ bands (bands at −48 ppm and −68 ppm attributed to [R–Si (OSi) (OH)₂] (T¹) and [R–Si (OSi)₃] (T³) suggest that DIL has reacted with the hydroxyl groups of the SBA-15 and formed a covalent bond with the mesoporous material [56].

Fig. 8.

²⁹Si MAS-NMR spectra of SBA-15 (pure) and SBA@DIL_2Cl 20 %.

Fig. 9 show the results of the TGA analysis of pristine SBA-15, DILs. And MDILs that were immobilized in the solid support. All of the samples exhibited a first stage of mass loss between 25 and 82 °C attributed to the moisture adsorbed by hygroscopic SBA-15 [57]. Pristine SBA-15 showed a more significant mass loss compared to the supports with immobilized DILs or MDILs. This behavior is probably associated with the filling of pores and the presence of ILs on solid surfaces reducing the availability of hydroxyl groups and inhibiting the presence of moisture [58]. The second stage of mass loss is related to the silane ligands present on the surface of the support, starting at around 167 °C for all three samples containing DILs. The third stage of mass loss is associated with less accessible organic fragments remaining from DlLs and can be used to calculate the percentage of DIL and MDIL grafted to the support [59,60]. The mass loss of SBA@DIL_2NTf₂, SBA@DIL_2FeCl₄ and SBA@DIL_NTf₂/FeCl₄ was 11.3 %, 9.9 % and 11.6 %, respectively, confirming that the immobilized DIL content is closely in agreement with the theoretical value of 10 %.

Fig. 9.

Thermogravimetric analysis (TGA).

The specific surface areas of the pristine SBA-15, DILs and MDILs grafted on SBA-15 were determined by the Brunauer-Emmett-Teller (BET) method (Fig. 10). The results showed that pristine SBA-15 (Fig. 10a) has type IV isotherms and a type H1 hysteresis loop, which are characteristic of regular cylindrical pores and a high degree of uniformity of pore size, a typical feature of mesoporous materials [61]. The supports SBA@DIL_2FeCl₄ (Fig. 10c), SBA@DIL_2NTf₂ (Fig. 10b) and SBA@DIL_NTf₂/FeCl₄ (Fig. 10d) showed type IV isotherms and a type H3 hysteresis loop, which is related to the presence of open and partially blocked mesopores due to the addition of DILs and MDILs.

Fig. 10.

Nitrogen adsorption/desorption BET isotherms at 77 K, for a) SBA-15 (pure), b) SBA@DIL_2NTf₂, c) SBA@DIL_2FeCl₄ e d) SBA@DIL_ NTf₂/FeCl₄.

It can be observed that samples containing the NTf₂‾ anion resulted in a more pronounced change in the hysteresis loops due to the anion size. The NTf₂‾ molecule is larger than that of FeCl₄‾, thus filling the pores more extensively and generating greater difficulty in filling with N₂ during the test, as also described by Sing and coworkers [62]. The highest specific surface area was observed for SBA-15 (777 m²/g), which then decreased with the immobilization of DILs and MDILs [SAB@DIL_2FeCl₄ (347 m²/g), SAB@DIL_NTf₂/FeCl₄ (263 m²/g) and SAB@DIL_2NTf₂ (258 m²/g)].

The pore size distribution and cumulative volume of SBA-15 and SBA-15 with immobilized DILs and MDILs were determined using the Barrett-Joyner-Halenda (BJH) model on isothermal adsorption data (Fig. 11). The results showed that the pore volume of the solid support decreased after immobilization with DILs and MDILs. The cumulative pore volume of SBA-15 was 1.021 cm³/g, while it decreased to 0.802 cm³/g for SBA@DIL_2NTf2, 0.899 cm³/g for SBA@DIL_2FeCl₄, and 0.870 cm³/g for SBA@DIL_NTf₂/FeCl₄. The average pore radius was calculated from the pore distribution curves and was found to be in the range of 4.216–5.366 nm.

Fig. 11.

BJH pore size distribution (red) and cumulative pore volume (blue) of the nitrogen isotherm at 77 K generated on a) SBA-15 (pure), b) SDIL_2NTf₂, c) SDIL_2FeCl₄, d) SDIL_NTf₂/FeCl₄.

The Raman spectra of the samples were analyzed as shown in Fig. 12. Only samples SBA@DIL_2FeCl₄ and SBA@DIL_NTf₂/FeCl₄ show peaks at 292 and 388 cm⁻¹, which are attributed to the symmetrical bending and stretching vibrations of the Cl–Fe–Cl bonds. These peaks are characteristic of the FeCl₄‾ anionic complex. Table 1 compares the data obtained in this work with published studies, evidencing that the complex has been successfully formed.

Fig. 12.

Raman spectra of SBA-15 (pure), SBA@DIL_2NTf₂, SBA@DIL_2FeCl₄ and SBA@DIL_NTf₂/FeCl₄.

Table 1.

Overview of the major peaks and assigned vibrations of the reference Raman spectra FeCl₄^‾_.

Chemical specie	Peaks (cm−1)		References
FeCl4‾	200	330	[63]
	134	334	[64]
	333	384	[65]
	333	384	[66]
	292	388	This Work

The more intense peaks in SBA@DIL_2FeCl₄ than in SBA@DIL_NTf₂/FeCl₄ may be related to the greater amount of FeCl₄^‾ anions present in the sample.

The SEM and TEM images of the samples are shown in Fig. 13. The pristine SBA-15 (see Fig. 13a) has a worm-like appearance in the form of cylindrical tubes composed of rope-like agglomerates of short rods as described in the literature [67]. Similar characteristics can be observed in their counterparts immobilized with ionic liquids. However, a slight alteration in surface morphology, such as reduced agglomeration of the rope-like aggregate, was observed on the immobilized supports, particularly those containing the NTf₂ anions, as depicted in Fig. 13 (b) and (d). This could be attributed to the immobilization of liquids, which does not solely occur within the pores. TEM images showed that the pore structures remained unchanged despite the slight morphological change.

Fig. 13.

SEM (10 μm) and TEM (200 nm) of a) SBA-15 (pure), b) SBA@DIL_2NTf₂, c) SBA@DIL_2FeCl₄ and d) SBA@DIL_NTf₂/FeCl₄.

The magnetic properties of the samples containing the anion FeCl₄^‾ (SBA@DIL_2FeCl₄ and SBA@DIL_NTf₂/FeCl₄) were evaluated (see Fig. 14). The capability of an external magnetic field to influence materials containing this type of anion was analyzed. However, it was observed that it was not feasible to measure the response curves and generate the graphs δM_R, indicating very weak magnetic interactions. This was attributed to the superposition of the diamagnetic signal from the samples themselves (SBA-15) and from the glass in the cylinder (sample holder).

Fig. 14.

Magnetization hysteresis loops of the SBA@DIL_2FeCl₄ and SBA@DIL_NTf₂/FeCl₄ samples. The insets show the row data in the whole magnetic-field range before performing diamagnetic correction.

Despite the ferromagnetic signal being superimposed on the diamagnetic signal, the SBA@DIL_2FeCl4 support shows a ferromagnetic signal 1.9 times greater than that of SBA@DIL_NTf₂/FeCl₄ (loops with coercivities of about 350 and 180 Oe). This indicates that the observed data was influenced by the 10 % (m/m) concentration of the ionic liquid.

3.3. CO₂ sorption and CO₂/N₂ selectivity tests

The CO₂ sorption performance of pristine SBA-15, supported DIL and MDILs can be seen in Fig. 15. In this type of material, the gas fills the pores, hence mesoporous supports with a high surface area tend to exhibit greater CO₂ sorption capacity, as observed in previous studies [50,56,68,69]. The data presented in Fig. 15 follow these characteristics, with the pristine SBA-15 showing the highest CO₂ capture capacity, followed by SBA@DIL_2FeCl₄ > SBA@DIL_NTf₂/FeCl₄ > SBA@DIL_2NTf₂. Consistent with the reduction in specific surface areas of the samples, fewer free pores result in lower CO₂ capture capacity. However, at the equilibrium pressure of 1 bar, it can be seen that pristine SBA-15 and SBA@DIL_2FeCl₄ have similar CO₂ sorption capacities, [58.61 (±4.48) mgCO₂/g] and [57.31 (±0.02) mgCO₂/g], respectively. This similarity is attributed to the fact that at higher pressures, CO₂ physisorption in the hexagonal pores of SBA-15 becomes the determining factor. The increase in CO₂ sorption at higher pressures is more strongly related to the available surface area and pore volume. Consequently, due to its lower porosity compared to unmodified silica supports, the CO₂ sorption of the grafted samples decreases, as reported by Mohamedali and coworkers [70]. This trend is evident from the behavior of the sorption curve when the pressures vary from 1 to 30 bar.

Fig. 15.

CO₂ sorption at 25 °C at different pressures.

The CO₂/N₂ selectivity of the supports was evaluated (Fig. 16), with pristine SBA-15 showing the worst performance [3.22 (±0.45)], which can be attributed to its limited ability to interact with CO₂ molecules, primarily through interactions with the hydroxyls present on the support surface [71]. The support containing DIL with only NTf₂^‾ anions, SBA@DIL_2NTf₂, demonstrated an improvement in selectivity [9.92 (±0.72)], improving the selectivity of the material by approximately threefold. NTf₂^‾ is one of the most commonly used fluorinated anions for CO₂ capture, and the charge delocalization and low electrical polarizability resulting from fluorination have corresponding effects on the electrostatic interactions and dispersion forces influencing CO₂ sorption, as noted by Zheng et al. [72]. However, fluorinated compounds are associated with high cost and negative environmental impact [73]. The mesoporous supports containing FeCl₄^‾ anions performed most effectively in this study, with SBA@DIL_NTf₂/FeCl₄ [11.17 (±0.44)] and SBA@DIL_2FeCl₄ [12.27 (±0.72)] exhibiting notable performance. For the FeCl₄^‾ anion, the interaction is attributed to coordination induced by metal ions, involving a coordination interaction between weakly electrophilic CO₂ and the metal center [74,75].

Fig. 16.

CO₂/N₂ selectivities at 25 °C and 20 bar.

Although the effect of anions on CO₂ sorption is considered to be more significant than that of cations [76,77], the increased selectivity may also be related to the interactions between the imidazole cations and CO₂, including van der Waals forces, hydrogen bonding, and the steric effect of the imidazole cations [78,79]. In addition, the presence of an ether group as a ligand between the cations allows for Lewis acid-base interactions with CO₂ acting as a Lewis acid due to its low electron density compared to TEG [80,81]. This may explain the 3.8-fold increase in selectivity observed on the SBA@DIL_2FeCl₄ support.

Table 2 compares the CO₂ sorption and CO₂/N₂ selectivity results obtained in this work with those published in previous studies with mesoporous materials containing monocationic ionic liquids, impregnated and grafted with the same proportion of ionic liquid as in this study [10 %(w/w)], under conditions as close as possible. The data lead to the conclusion that the incorporation of MDILs into the SBA-15 support (SBA@DIL_NTf₂/FeCl₄ and SBA@DIL_2FeCl₄) shows a better CO₂/N₂ selectivity than the monocationic ionic liquids. This significant increase can be attributed to the greater number of coordination sites available [82], which doubles in the presence of two cations and two anions that have been shown to exhibit a strong interaction with CO₂. The use of the FeCl₄ anion, with its coordination interaction between the weakly electrophilic CO₂ and the metal center, outperformed the interaction promoted by the NTf₂ anion. The delocalization and the low electrical polarizability promoted by the fluorination have corresponding effects on the electrostatic interactions and dispersion forces influencing CO₂ absorption, as previously mentioned in the CO₂/N₂ selectivity. These factors contribute to the significant improvement, as illustrated in Table 2.

Table 2.

CO₂ sorption and CO₂/N₂ selectivity, materials measured in this work compared to reported values in the literature.

			CO2 sorption		Experimental Conditions
Support	Funcionalization	Ionic Liquid	mgCO2/g	mmol/g	PCO2(bar)	T (°C)	CO2/N2 Selectivity	Ref.
Silica particles	Grafted	[P8883 ]TFSI 10 %	–	0.99	1	40	6.0	[83]
ZIF-8	Impregnation	[Emim][Ac] 10 %	–	0.3	1	30	∼7.5	[84]
Commercial Silica	Impregnation	mbmim[Tf2N]10 %	52.1	–	4	45	6.9	[68]
Commercial Alumina	Impregnation	mbmim[Tf2N]10 %	56.0	–	4	45	3.7	[68]
Commercial Silica	Grafted	[i-C5TPIm][Cl] 10 %	60.68	–	4	45	4.34	[56]
Commercial Silica	Grafted	[i-C5TPIm][Tf2N] 10 %	69.38	–	4	45	4.38	[56]
Activated Carbon	Impregnation	[bmpy][Tf2N] 10 %	–	0.62	1	25	–	[85]
SiO2	Grafted	[bmim][CF3SO3] 10 %	11.90	–	1	25	–	[86]
SBA-15	Grafted	DIL_2FeCl4 10 %	57.30	1.30	1	25	12.3	This work
SBA-15	Grafted	DIL_NTf2/FeCl4 10 %	54.42	1.24	1	25	11.2	This work
SBA-15	Grafted	DIL_2FeCl4 10 %	90.33	2.05	5	25	12.3	This work
SBA-15	Grafted	DIL_NTf2/FeCl4 10 %	80.48	1.83	5	25	11.2	This work

3.4. Stability tests

The stability test was carried out under the same conditions as the CO₂ capture capacity tests, with a selected equilibrium pressure of 10 bar, using the sample that showed the best CO₂ capture capacity and selectivity against CO₂/N₂, which was SBA@DIL_2FeCl₄. The sample underwent 10 sorption and desorption cycles, resulting in a consistent sorption value of [90.17 (±0.37)], as shown in Fig. 17.

Fig. 17.

CO₂ sorption/dessorption tests on SBA@DIL_2FeCl₄ (25 °C; 5 bar).

The data allows us to conclude that both SBA-15, and MDIL (with FeCl₄ anion only) are stable. This suggests that chemically, the MDIL molecules and their active sites continue to facilitate the expected physisorption, while mechanically, the pores remain unaffected during retests. Enhancing competitiveness with aqueous amine solutions is intriguing, as it has the potential to lower process costs. Despite their affordability, aqueous amine solutions degrade over time due to chemisorption processes and exposure to high temperatures [14,[87], [88], [89]]. By utilizing less ionic liquid more efficiently, particularly through grafting onto supports, it could become more competitive. This approach, as suggested by Hiremath et al. [33] enhances adsorption kinetics and minimizes mass transfer resistance.

4. Conclusion

To expand the options for CO₂ capture and separation, we developed imidazolium-based dicationic ionic liquids (DILs) linked by TEG and combined with different anions (2NTf₂^‾, 2FeCl₄^‾ and NTf₂^‾/FeCl₄^‾). The DILs were successfully supported in the SBA-15 10 % (w/w). MDIL with FeCl₄^‾ anion presented the best results both for CO₂ sorption and CO₂/N₂ separation compared to fluorinated anions. Yet, SBA@DIL_2FeCl₄ proved to be chemically and mechanically stable, making it a promising option for use in post-combustion processes.

Compared to the results obtained with monocationic ionic liquids described in the literature using mesoporous supports and the same percentage of IL content, our results show that the use of DILs leads to improved CO₂ capture and selectivity due to increased availability of coordination sites.

CRediT authorship contribution statement

Evandro Duarte: Writing – original draft, Methodology, Data curation. Franciele Bernard: Validation, Methodology, Formal analysis, Conceptualization. Leonardo Moreira Dos Santos: Validation, Resources, Methodology, Data curation. Barbara B. Polesso: Methodology, Validation. Rafael Duczinski: Methodology, Formal analysis, Conceptualization. Vitor Forneck: Methodology, Formal analysis. Julian Geshev: Methodology, Formal analysis. Sandra Einloft: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.