Hybrid Ionic Liquid–Silica Xerogels Applied in CO₂ Capture

Abstract

The imidazolium-based ionic liquids (ILs) are solvents known for selectively solubilizing CO₂ from a gas CH₄/CO₂ mixture, hence we have produced new hybrid adsorbents by immobilizing two ILs on xerogel silica to obtain a solid–gas system that benefits the ILs’ properties and can be industrially applied in CO₂ capture. In this work, the ILs (MeO)₃Sipmim.Cl and (MeO)₃Sipmim.Tf₂N were used at different loadings via the sol–gel process employing a based 1-methyl-3-(3-trimethoxysylilpropyl) imidazolium IL associated to the anion Cl⁻ or Tf₂N⁻ as a reactant in the synthesis of silica xerogel. The CO₂ adsorption measurements were conducted through pressure and temperature gravimetric analysis (PTGA) using a microbalance. SEM microscopies images have shown that there is an IL limit concentration that can be immobilized (ca. 20%) and that the xerogel particles have a spherical shape with an average size of 20 µm. The adsorbent with 20% IL (MeO)₃Sipmim.Cl, SILCLX20, shows greater capacity to absorb CO₂, reaching a value of 0.35 g CO₂ / g adsorbent at 0.1 MPa (298 K). Surprisingly, the result for xerogel with IL (MeO)₃Sipmim.Tf₂N shows poor performance, with only 0.05 g CO₂ / g absorbed, even having a hydrophobic character which would benefit their interaction with CO₂. However, this hydrophobicity could interfere negatively in the xerogel synthesis process. The immobilization of ionic liquids in silica xerogel is an advantageous technique that reduces costs in the use of ILs as they can be used in smaller quantities and can be recycled after CO₂ desorption.

1. Introduction

Excessive increase in greenhouse gas concentrations, especially carbon dioxide, stems from an energetic matrix based on the combustion of fossil fuels which causes global warming and has other environment impacts [1,2,3]. Mitigation of these impacts has been the subject of research for the development of adsorbents for CO₂ capture by the application of carbon capture and storage (CCS) or carbon capture utilization (CCU) technologies [4,5,6,7,8].

Ionic liquids (ILs) ionic liquids are a class of materials widely studied in recent times. It has been proposed as an alternative solvent for CCS or conversion [9,10,11,12,13,14,15]. These solvents are selective for CO₂ separation in CO₂/CH₄ gas mixtures [16,17]. As the ILs are non-volatile compounds, when CO₂ is desorbed under depressurized conditions, no loss of solvent occurs. Moreover, one can design the chemical structure of ILs, choosing the appropriate cation/anion to optimize the capacity of the ionic liquid in the CO₂ absorption [18,19,20,21].

Incorporating ILs in solid matrices emerged as a new and challenging field, allowing the use of heterogeneous systems for CO₂ capture [22,23,24,25]. The immobilization process is usually carried out by cationic or anionic incorporation, or by cationic anchorage, or through in-situ polymerization via a sol–gel process [26,27]. In the case of cationic or anionic incorporation, the IL is synthesized and then fixed into the inorganic material through an IL cation or anion anchorage. In the case of the in situ polymerization via sol–gel process, the IL is used as a templating agent for the synthesis of silica xerogels due to its surfactant character and the low surface tension which enables control of the silica particle size, gelation time, structures, and morphology [28,29].

In the classical sol–gel process, an alkoxide precursor is hydrolyzed leading to the formation of Si-OH bonds (silanol groups) that condensate forming an inorganic three-dimensional network. The targeted material is then obtained after drying—xerogel or aerogel [30].

The IL acts like a solvent, catalyst or template in in-situ polymerization, bringing with it the advantage of being easier to handle and having wider applicability in various fields. ILs improve the efficiency and selectivity and allow easy separation of the products and catalyst recycle, especially in catalysis [31,32].

With the immobilization of ILs in gels leading to ionogels, these materials are applied in electrochemistry devices to encapsulate functional molecules (catalysts, sensing molecules, fluorescent metal complexes) offering a new method to functionalize nanostructured inorganic materials [33,34]. The xerogel silica final structure is highly influenced by the nature of the alkyl-alkoxysilanes precursor such as the chain length, functionality, and its hydrophobic character. These properties can be modified adding a wide range of compounds to the alkyl-alkoxysilanes precursor, such as surfactants like ionic liquids [35].

The combination of the well-known affinity between pure ILs and CO₂ [9,11,16,17,18,19] and solid adsorbents leads to a synergic effect reported by Aquino et al. [25]. Therefore, this study aims to synthesize hybrid silica–IL xerogels via a sol–gel process using different alkyl-alkoxysilanes compounds: the tetramethoxysilane (TMOS), the methyltrimetoxysilane (MTMS), and a based 1-methyl-3-(3-trimethoxysylilpropyl) imidazolium IL associated to the anion Cl⁻ or Tf₂N⁻ incorporated in the range 1–40% IL/ alkyl-alkoxysilanes v/v. During the synthesis, TMOS and MTMS are hydrolyzed leading to the formation of silanols groups. Part of these silanol groups condensate producing a three-dimensional network and the remaining silanol groups will react with the IL ensuring its participation in the structure building.

These materials were evaluated for CO₂ adsorption. It is important to highlight that the use of IL as a precursor for the synthesis of xerogel is an innovative route for increasing the CO₂ adsorption of the material. The IL is responsible for the three-dimensional matrix formation as well as for increasing the affinity between the CO₂ and the adsorbent. Typical CO₂ adsorption values are approximately 0.09 g per g of neat ionic liquid [20].

2. Materials and Methods

2.1. Chemicals

Lithium bis(trifluoromethylsulfonyl) imide (LiTf₂N) (Alfa Aesar, 98.0%); (3-chloropropyl)trimethoxysilane (Alfa Aesar, 97.0%); sodium fluoride (NaF) (Sigma Aldrich, 99.0%); 1-methylimidazole (Sigma Aldrich, 99.5%); tetramethoxysilane (TMOS) (Sigma Aldrich, 98.5%); polyvinyl alcohol (PVA; Mn 15.000) (Sigma Aldrich, 88.0%); methyltrimetoxysilane (MTMS) (Sigma Aldrich, 98.5%); carbon dioxide (CO₂) (Air Liquide, 99.998%).

2.2. Ionic Liquid Synthesis

Firstly, 1-methyl-3-(3-trimethoxysilylpropyl) imidazolium chloride was synthesized through the reaction of 1-methylimidazole with (3-chloropropyl) trimethoxysilane (molar ratio 1: 1.5) at 363 K for 48 h. Then the reaction mixture was cooled down to room temperature and the organic upper phase separated. The resulting product was a yellow viscous ionic liquid. Then the IL phase was washed thoroughly with diethyl ether and residual ether was removed under vacuum.

An equimolar amount of LiTf₂N was added to a solution of (MeO)₃Sipmim.Cl in dry acetone and left under stirring for five days at room temperature. After concentrating under vacuum, the solid was solubilized in dichloromethane, dried, filtered over a celite bed, and washed with water until AgNO₃ test was negative. The organic phase was dried over anhydrous MgSO₄, filtrated and the solvent was removed under vacuum [36,37,38]. The structure of the ILs (MeO)₃Sipmim.Cl and (MeO)₃Sipmim.Tf₂N and was confirmed by ¹H-NMR and FTIR.

(MeO)₃Sipmim.Cl: ¹H-NMR (400 MHz, CDCl₃, 298 K) δ (ppm): 0.58 (m, CH₂Si); 1.91 (m, CH₂CH₂N); 3.49 (s, SiOCH₃); 3.63 (s, CH₃N); 4.06 (t, CH₂N); 4.27 (t, CH₂); 7.39 (s, H5); 7.61 (s, H4); 10.39 (s, H2). FTIR ν (cm⁻¹): 3031, Si-O; 2944-2839 for aliphatic C-H stretching (methyl and methylene groups); 1570-1457 C=C stretching and C-N of the imidazolium ring; 1175-1071 Si-OCH₃; 805 from Cl⁻ anion.

(MeO)₃Sipmim.Tf₂N: ¹H-NMR (400 MHz, CDCl₃, 298 K) δ (ppm): 0.51 (m, CH₂Si); 1.85 (m, CH₂CH₂N); 3.47 (s, SiOCH₃); 3.83 (s, CH₃N); 4.25 (t, CH₂N); 7.70 (s, H5); 7.61 (s, H4); 8.99 (s, H2). FTIR ν (cm⁻¹): 3161-3115, Si-O; 2934-2853 for aliphatic C-H stretching (methyl and methylene groups); 1572-1460 C=C stretching and C-N of the imidazolium ring; 1182-1047 Si-OCH₃; 791-742 from Tf₂N⁻ anion.

2.3. Ionic Liquid Immobilization

The IL was dissolved in an aqueous solution containing NaF (0.20 g L⁻¹) and PVA (4.64 g L⁻¹). The amount of water in this solution was fixed (6.86 mmol). The solution was vigorously shaken on a vortex mixer. The precursors were then added (Figure 1) in amounts that yielded a water/silane molar ratio of 1:8, irrespective of the type and number of precursors used (e.g., 0.143 mmol of TMOS and 0.714 mmol MTMS in 1:5 TMOS/MTMS gels).

The % (v/v) of IL (1, 5, 10, 20, 30, and 40%) was determined by the IL volumes and the total volume of the solution containing water, NaF, PVA, TMOS, and MTMS.

The mixture was again vigorously shaken on the vortex mixer until it became homogeneous. It was then placed in an ice bath until gelation occurred (after a few seconds), and kept in the ice bath for an additional 10 min. The container with the obtained gel was kept at 277 K for 24 h, after which the gel was put in a kiln at 308 K for 24 h. The white gel obtained was crushed and washed (for about 10 min)/centrifuged (at 10,000 rpm), first using acetone then n-pentane (also 2 mL of each). The gel was left at 308 K for 24 h in a kiln [34].

The obtained samples were denominated according to the IL content and the anion present as described as supported ionic liquid (SIL)-Anion (TF2N or CL for (MeO)₃Sipmim.Tf₂N and (MeO)₃Sipmim.Cl respectively)–Xerogel (X)–% IL (number corresponding to the % of loaded IL)).

2.4. Characterization

The FTIR spectra were collected on a PerkinElmer Spectrum 100 spectrometer in KBr pellet form. NMR measurements were performed at room temperature using a Bruker Avance III 400 MHz spectrometer; solid-state NMR cross-polarization magic angle spinning (²⁹Si CPMAS) at 5 kHz was selected to record silicon spectra. SEM on Microscope Analytical JEOL, model 7001F FEG-SEM. The surface area and pore size were calculated from sorption analyses (N₂ at 77 K) using a Micromeritics Instrument Corporation, TriStar II 3020 V1.03 and Brunauer–Emmett–Teller (BET) method. The real density of xerogel samples (ρs) was determined on Ultrapycnometer 1000—Quantachrome Corporation, cell volume of 20.45 cm³, and pressure of 21.0 psi (1.45 bar).

2.5. CO₂ Adsorption Measurements

The sorption of CO₂ in the samples were gravimetrically assessed in a magnetic suspension balance (MSB), (Rubotherm Prazisionsmesstechnik GmbH, 35 MPa and 673 K) equipped with a single sinker device for adsorbate density determination and thermostatized with an oil bath (Julabo F25/± 273 K). The apparatus details are well described elsewhere. When compared to other gravimetrical sorption methods, the MSB device has the advantage of allowing high-pressure sorption measurements as the sample can be potted into a closed chamber coupled to an external precise balance (accuracy of ±10 μg).

The samples (0.06 to 0.09 g) were weighed and transferred to the MSB sample container, and the system was subjected to a 10⁻⁷ MPa vacuum at the temperature of the sorption measurement, 298 K, for 24 h (constant weight was achieved in this time). The CO₂ was admitted into the MSB pressure chamber up to the desired pressure, 0.1–2 MPa in this study, a pressure gauge with an accuracy of 10⁻³ MPa was used to control the system pressure.

The solubility of CO₂ in the samples for each isotherm and pressure considered was measured 3–4 hours after no more weight increasing for CO₂ sorption had been observed. At this step of CO₂ solubility in the samples, the weight reading from the microbalance at pressure P and temperature T is recorded as ${\mathrm{W}}_{\mathrm{t}}\left(\mathrm{P},\mathrm{T}\right)$. The mass of dissolved CO₂ in the sample $\left({\mathrm{W}}_{absolute}\right)$ was calculated using Equation (1)

$${\mathrm{W}}_{absolute}=\left[{\mathrm{W}}_{\mathrm{t}}\left(\mathrm{P},\mathrm{T}\right)-{\mathrm{W}}_{\mathrm{sc}}\left(\mathrm{P},\mathrm{T}\right)+{\mathsf{\rho }}_{\mathrm{g}}\left(\mathrm{P},\mathrm{T}\right)·\left({\mathrm{V}}_{\mathrm{sc}}\left(\mathrm{T}\right)+{\mathrm{V}}_{\mathrm{s}}\left(\mathrm{T}\right)+{\mathrm{V}}_{\mathrm{ads}}\right)\right]-{\mathrm{W}}_{\mathrm{s}}\left(\mathrm{vac},\mathrm{T}\right)$$

(1)

where ${\mathrm{W}}_{\mathrm{sc}}\left(\mathrm{P},\mathrm{T}\right)$ is the weight of sample container, ${\mathsf{\rho }}_{\mathrm{g}}\left(\mathrm{P},\mathrm{T}\right)$ stands for CO₂ density, directly measured with the MSB coupled single-sinker device, ${\mathrm{V}}_{\mathrm{sc}}\left(\mathrm{T}\right)$ is the volume of the sample container, determined from a buoyancy experiment when no sample is charged into the sample container, ${\mathrm{V}}_{\mathrm{s}}\left(\mathrm{T}\right)$ the specific solid sample volume, ${\mathrm{W}}_{\mathrm{s}}\left(\mathrm{vac},\mathrm{T}\right)$ is the weight of samples under vacuum and the term ${\mathsf{\rho }}_{\mathrm{g}}\left(\mathrm{P},\mathrm{T}\right)·\left({\mathrm{V}}_{\mathrm{sc}}\left(\mathrm{T}\right)+{\mathrm{V}}_{\mathrm{s}}\left(\mathrm{T}\right)\right)$, represents the buoyancy force.

The volume of the adsorbed phase ${\mathrm{V}}_{\mathrm{ads}}$ must be taken into account in the buoyancy correction to determine the absolute gas adsorption. In this paper, the density of the adsorbed phase ${\mathsf{\rho }}_{\mathrm{ads}}$ was assumed to be the density of liquid CO₂ in a reference state (boiling point at 1 atm) and the ${\mathrm{V}}_{\mathrm{ads}}$ was obtained by dividing the adsorbed mass, $m_{ads}$ by the density of the adsorbed phase, ${\mathsf{\rho }}_{\mathrm{ads}}$. Further details on data handling from the adsorption isotherms measured may be found in literature [39,40].

3. Results

The FTIR spectra of the xerogel matrix and SILTF2NX20 (see Figure A1), respectively, show i) in the region between 3550–3416 and 3488 cm⁻¹ the presence of O-H bands of silanol groups; ii) at 2977 and 2975 cm⁻¹ C-H stretching of CH₃ groups; and iii) at 1638 and 1627 cm⁻¹ C-O stretching of methoxy groups. In these spectra, the presence of bands 1277 and 1275 cm⁻¹ characterize the Si-CH₃ bond and at 1037 and 1049 cm⁻¹ the Si-O-Si bond. The presence of the imidazole IL linked to the xerogel is mainly characterized by a band at 1577 cm⁻¹ corresponding to a C=C aromatic stretching and a band at 3167 cm⁻¹ corresponding to an aromatic C-H, both belonging to the imidazolium ring [37,41]. In addition, the FTIR spectrum of SILTF2NX20 also shows two signals at 1353 and 1196 cm⁻¹ corresponding respectively to the Si-O and Si-O-CH₃ stretching [37].

It is noteworthy that the siloxane bands of Si-O-Si, in the range of 1136–1186 cm⁻¹ and disiloxane band of R₃Si-O-SiR₃ in the range of 1049–1070 cm⁻¹ are ascribed to the chemical bonds between silicon and IL indicating a strong interaction between the IL and the gel matrix.

Table 1. ²⁹Si CPMAS results for supported ILs in xerogel matrix.

	XEROGEL			SILCLX10			SILCLX40			SILTF2NX10			SILTF2NX40
Form	δ (ppm)	intgr	%	δ (ppm)	intgr	%	δ (ppm)	intgr	%	δ (ppm)	intgr	%	δ (ppm)	intgr	%
Q2	/	/	/	/	/	/	/	/	/	/	/	/	/	/	/
Q3	/	/	/	−101.14	0.02	1.74	−102.28	0.02	1.47	/	/	/	/	/	/
Q4	−117.31	0.49	11.86	−110.21	0.13	11.30	−111.47	0.10	7.35	−110.41	0.13	10.48	−110.56	1.00	14.25
T2	−72.38	2.64	63.92	−66.06	1.00	86.96	−66.68	1.00	73.53	−66.03	1.00	80.56	−65.65	5.09	72.51
T3	−63.58	1.00	24.21	/	/	/	−58.10	0.24	17.65	−56.10	0.11	8.87	−56.49	0.93	13.25

presents data corresponding to the ²⁹Si CPMAS analysis of the xerogel matrix synthesized without IL, XEROGEL, and the xerogel containing IL. This data confirms, in the case of XEROGEL, the presence of chemical environments for Si at –63.58 and –72.38 ppm related to T² and T³, and at –117.31 ppm attributed to Q⁴, respectively. For all samples containing ILs, the largest composition is of T² corresponding to (-O-)₂Si-(O-R)₂, independently of the anion.

Only the samples with the IL have the Cl⁻ anion present Q³ form corresponding to (-O-)₃Si-OH. The forms T² and T³ indicate the immobilization of IL, therefore we observed that the major quantity of T²+T³ is that with a higher load IL (40%). The values of T² and T³ allow verification if there are free OH groups on the support or if the IL is bonded through the OH groups and then being part of the xerogel.

SEM images of the xerogel matrix and of the xerogels containing the (MeO)₃Sipmim.Tf₂N in different concentrations are represented in Figure 2. Based on these images it can be observed that the use of 1% IL (MeO)₃Sipmim.Tf₂N as precursor is not sufficient to form the typical spheres of the xerogel matrix. However, as the concentration of IL increases to 5, 10, and 20% the formation of the microspheres becomes increasingly visible. When the IL load reaches 30%, the collapse of the spheres can be observed indicating an over concentration of IL as precursor for the synthesis of the xerogel.

The images of the synthesized xerogel also demonstrate a peculiar feature related to the presence of the ionic liquids that acts as a surfactant resulting in the formation of spherical particles during the sol–gel process. As the (MeO)₃Sipmim.Cl and (MeO)₃Sipmim.Tf₂N ILs are amphiphilic structures composed of a polar head (the imidazolium ring) that is hydrophilic and carbon or siloxane chain that is hydrophobic, they form micelles during the sol–gel process synthesis [42].

The XEROGEL, SILTF2NX20, and SILTF2NX40 were characterized through N₂ sorption at 77 K and helium pycnometry. Specific surface area (S_BET), pore volume (V_p), pore and diameter size and density results are shown in

Table 2. Structural properties for ILs in xerogel silica

Samples	SBET (m2/g)	Vp (cm3/g)	Pore Size (nm) a	Diameter (nm) b	ρs (g/cm3) c
XEROGEL	9	0.016	7.1	8.2	1.80
SILTF2NX20	4	0.008	7.5	10.2	3.08
SILTF2NX40	1	0.002	4.9	5.6	---

^a Adsorption average pore width (by BET). ^b BJH adsorption average pore diameter. ^c Determined by helium pycnometer.

.

Data obtained from N₂ sorption (S_BET, V_p) shows that partially replacing MTMS with the IL decreases the specific surface area and the pore volume indicating that the support is more compact according to the density values.

CO₂ adsorption tests were performed immediately after the synthesis of the materials. The tested materials were the Xerogel, SILTF2NX20, SILCLX20, and SILCLX40. It is worth noting that for the calculation of the adsorption in excess and absolute from the PTGA data, the density value for the sample SILTF2NX20 obtained through helium pycnometry (see Table 2) was used. As the experimental CO₂ adsorption curves, absolute and in excess, were very similar due to employing the low working pressure, only the absolute adsorption data is presented Figure 3 as a function of CO₂ pressure.

It is evident that the sample SILCLX20 exhibits the greater capacity to CO₂ adsorption, reaching a value of 0.35 g CO₂ / g adsorbent at 1.0 MPa. The sample with the same IL, but with greater immobilized concentration SILCLX40 has a very low performance, only 0.07 g CO₂ / g adsorbent at the same pressure, which indicates that increasing the amount of IL immobilized in the xerogel hinders the adsorption capacity. This effect can be caused by both a block of the solid’s pores by the IL and IL trapped without accessibility or free volume to adsorb CO₂.

Surprisingly, the adsorption result obtained for SILTF2NX20 (0.05 g CO₂ / g adsorbent at 1.0 MPa) is very low. As the Tf₂N⁻ anion has a hydrophobic character, it was expected that it would improve the IL–CO₂ interaction. This result can be attributed to a low IL interaction with the xerogel silica networking.

The concentration of IL immobilized in the support as mentioned earlier has a slight adverse effect on adsorption efficiency, as the higher the concentration of IL support, the lower the mass of CO₂ adsorbed. Results obtained with SILCLX20 were distinguished from all the others and is an exceptional mark.

4. Discussion

The CO₂ sorption capacity was previously studied by Aquino et al. [25] through the same gravimetrical sorption methods and experimental conditions using mesoporous material MCM-41 (M) containing supported ionic liquids (SIL) with 50% load (50). The ionic liquids had the same (MeO)₃Sipmim cation and four different anions (${\mathrm{Cl}}^{-}$ (CL), ${\mathrm{BF}}_4^{-}$ (BF4), ${\mathrm{PF}}_6^{-}$ (PF6), and ${\mathrm{TF}}_2{\mathrm{N}}^{-} \left(\mathrm{TF}2\mathrm{N}\right)$). This study evaluated the influence of the anion nature on the CO₂ sorption capacity (see Figure 4).

The results represented in Figure 4 clearly show that the IL containing a chloride anion (SILCLM50) had better CO₂ adsorption. For comparison, SILCLX20 (xerogel process) from our study (see Figure 3) presents a greater capacity to adsorb CO₂, reaching a value of 0.35 g CO₂ / g adsorbent (at 1.0 MPa) than SILCLM50 (supported on MCM-41) which attains only 0.11 g CO₂ /g adsorbent (at 1.0 MPa) (see Figure 4). In any case, supported ILs (in xerogel or MCM-41) provide higher CO₂ adsorption capacity than when used as solvents due to the contribution of the support. The comparison of the results reported in Figure 3 and Figure 4 indicates clearly that when using xerogel support a lower amount of IL (20%) (see Figure 3) is required than when MCM-41(50%) is employed.

Additionally, the literature reports results that described the use of different materials, such as Zeolite 13X at 303 K [43], β-zeolite at 303 K [44], and MCM-41 at 298 K [25]. The best result obtained was by Jadhav, P.D., using the zeolite 13X, 0.0550 g CO₂ being adsorbed per g of adsorbent. The addition of amine, that has an affinity with CO₂ [5] to the zeolite does not increase the quantity of CO₂ adsorbed.

The best CO₂ adsorption amount obtained in this study (SILCLX20: 0.35 g CO₂ / g adsorbent) is approximatively 500 and 200% higher than when the zeolite 13X and (MeO)₃Sipmim.Cl supported in (MCM-41 (SILCLM50, see Figure 4) are used as an adsorbent, respectively.

After the CO₂ adsorption, the PTGA was depressurized without heating. It was verified that the sample weight remains equal to its value before the CO₂ adsorption process. This result indicates that CO₂ was physically adsorbed, and that the adsorbent material was not modified during the adsorption experiments enabling its recovery.

The higher thermal stability of the hybrid material and the possibility of their reuse in physical separation systems are properties that indicate these materials have the potential for CO₂ capture.

5. Conclusions

The sol–gel immobilization of ionic liquids is an advantageous technique that reduces costs in the use of ILs, as they can be used in smaller quantities. When ILs (MeO)₃Sipmim.Cl and (MeO)₃Sipmim.Tf₂N were supported at different concentrations of IL via the sol–gel process it was noted that there is an ideal IL concentration that can be immobilized in this matrix type (ca. 20%). The CO₂ adsorption capacity results show that these materials can be applied as adsorbents in a capture system, as they offer satisfactory CO₂ sorption capacity, the best result being obtained by SILCLX20 that reached 0.35 g CO₂ / g adsorbent at 1.0 MPa pressure.