Impregnation of ionic liquids in mesoporous silica using supercritical carbon dioxide and co-solvent

Abstract

Ionic liquids (ILs) have been considered to be attractive alternatives for CO₂ capture. The impregnation of ILs on porous substrates can significantly reduce the ILs consumption, enhance the mass transfer of CO₂ in the ILs and make their recycling easier. However, due to the high viscosity of the ILs, the impregnated ILs were often poorly dispersed on the substrates or blocked the nanochannels when the traditional impregnation method was used. In this study, a new, effective and environmentally benign technique to impregnate ILs onto mesoporous substrates was proposed. The impregnation of [Bmim]BF₄ on SBA-15 was successfully achieved using supercritical carbon dioxide (scCO₂) as the solvent and methanol as co-solvent. The ILs were proved to be impregnated into the nanochannels rather than on the outside of the substrate by N₂ adsorption–desorption and SEM analysis. The operating pressure, temperature and time were investigated to find out the optimum parameter to synthesis these nanocomposites. Finally, the CO₂ adsorption capacity of the ILs@SBA-15 composites was performed on a dynamic adsorption apparatus.

---

1. Introduction

Carbon dioxide (CO₂) is the primary greenhouse gas emitted mainly from the combustion of carbon-based fossil fuels through human activities. One promising solution to reduce CO₂ emission from large emission sources (e.g. coal fired power plants) is to capture the generated CO₂ with sorbents before it is emitted in the atmosphere.^1,2 Aqueous ammonia is the widely used adsorbent to remove CO₂ gas.^3,4 However, the disadvantages of the amine aqueous-based CO₂ capture technique mainly include relatively high amine loss and degradation, high energy consumption for regeneration, and unavoidable equipment corrosion. In recent years, amine functionalized solids, activated carbon, carbon molecular sieves, and zeolites have also been extensively studied as adsorbents for CO₂ gas.^5,6 However, these adsorbents also have shortfalls such as low capacities and difficult regeneration processes. Therefore, novel sorbent materials and technologies for efficient and economical CO₂ capture have attracted increasing attention from both academia and industry.

Most recently, ionic liquids (ILs) have been considered to be attractive alternatives for the capture of CO₂ because of their negligible vapor pressures, high thermal stability, tunable physicochemical properties and ability to capture CO₂.^7–9 To extend the applications of ionic liquids in this area, different types of composites have been tried such as ILs/amine mixtures¹⁰ and ILs/polyethylene glycol mixtures.¹¹ What's more, the impregnation of ionic liquids on different porous supports (such as zeolites, silica gel, metal–organic frameworks) have also been investigated.^12,13 There are many advantages of supporting ILs on porous substrates. First, it is desirable to minimize the amount of utilized ionic liquid on the basis of economic criteria and possible toxicological concerns. Second, the mass transfer of CO₂ in the ILs is greatly enhanced because of the significant increase of the mass transfer area. Third, the ease of recycling and the possibility to use a fixed-bed or a fluid-bed separation equipment.

The impregnation of ionic liquid on nanostructured mesoporous materials can be achieved by many methods such as impregnation,¹⁴ sol–gel (for inorganic substrates),¹⁵ grafting,¹⁶ polymerization^17,18 and so on. Compared to the impregnation method, the impregnation of ionic liquid using grafting method and sol–gel method are more stable because of the formation of covalent bonds between the ionic liquid and the surface of the substrates. However, the disadvantages of these two method are the low liquid loadings. The impregnation method is the easiest way for the preparation of an impregnated ionic liquid. However, in an impregnation method, because of the high viscosity of the ILs, the impregnated ILs are often dispersed poorly, sometimes parts of the ILs locate at the entrance of the substrates pores and block the pores.

In this paper, we report a new method of supporting ionic liquids within the nanochannels of mesoporous silica in supercritical carbon dioxide (scCO₂) and co-solvent. It is a new, effective and environmentally benign technique to synthesize supported nanocomposites. CO₂ is nontoxic, nonflammable, inexpensive, and has moderate critical temperature and pressure (T_C = 31.1 °C, P_C = 7.38 MPa). As it was reported that scCO₂ was highly soluble in ILs, while ILs were almost insoluble in scCO₂.¹⁹ So the mixture solvents of IL–CO₂ can avoid cross-contamination of solvents and simplify the process of chemical reaction and separation.^19,20 However, it was reported that ionic liquid can dramatically dissolve in scCO₂ with polar organic compounds (e.g. acetone) especially as the concentration of the compounds in scCO₂ exceeds 10 mol%.²¹ In this study, the impregnation of ionic liquid [Bmim]BF₄ on SBA-15 was prepared using scCO₂ as solvent. Three different co-solvents including methanol, acetone and hexane were used respectively to find out a most proper one. The nanocomposites were characterized by Energy Dispersive X-ray (EDX), Scanning Electronic Microscopy (SEM), N₂ adsorption–desorption and Thermogravimetric Analysis (TGA). After comparison with the traditional impregnation method, a series of the experiments were conducted to investigate the influence of the parameters of interest including the initial amount of ILs, the deposition pressure, temperature and time on the ILs loading of the composites. Finally, the CO₂ capture capacity of the ILs@SBA-15 composites were performed on a dynamic adsorption apparatus.

2. Experiments

2.1. Materials

Mesoporous SBA-15 was prepared according to the published literature.²² The synthesized SBA-15 has a BET surface area of 836 m² g⁻¹, total pore volume of 1.05 m³ g⁻¹ and average pore diameter of 6.1 nm. [Bmim]BF₄ with a purity of more than 99% was purchased from Shanghai Cheng Jie Chemical Co. LTD. Carbon dioxide (>99%) was obtained from DaLian GuangMing gas Co. Ltd. Methanol, acetone and n-hexane of analytical grade were all supplied by TianJin FuYu fine chemicals Co. Ltd.

2.2. Experimental method

(1) Supercritical fluid deposition (SCFD) method. The experimental apparatus and the process were similar to those used for preparing SBA-15 supported metal/metal oxide nanocomposites, which were shown in our previous reports.^23–26 In brief, a certain amount of ILs and co-solvent were placed at the bottom of a high-pressure stainless steel reactor. A stainless steel basket containing a certain amount of substrate was fixed in the upper part of the reactor to avoid the direct contact of ILs and the substrate. Then the reactor was connected to the lines, preheated to the experimental temperature and charged into CO₂ using a piston pump until the pressure of the system got up to the required pressure. The reactor was slowly depressurized after it was kept at the desired pressure and temperature for a certain time. CO₂ and co-solvent were removed during the decompression process and SCFD-ILs@SBA-15 nanocomposites were obtained.

(2) Impregnation method. In an impregnation method, a certain amount of ILs was first dispersed in 5–10 mL acetone and then 200 mg SBA-15 was added into this solution. After stirred for 9–16 h at room temperature, the excess solvent was removed by filtration and then the nanocomposites were dried at 80 °C until the weight becomes constant. Then the IMP-ILs@SBA-15 nanocomposites were obtained.

(3) Apparatus and procedure of CO₂ absorption. The apparatus used for CO₂ absorption experiments are illustrated in Fig. 1, which mainly includes a fixed-bed adsorption column (internal diameter 10 mm, length 180 mm) and a gas chromatograph (FuLi GC9790). The operation temperature in the adsorption was controlled by a heat-controlling device including a temperature programmer to monitor equipped in the inlet gas line. The inlet CO₂ concentration was 1.1% mol (N_Co2 : N_N2 = 1 : 9) and the flow rate of the mixed gas was controlled by the mass flow controller and fixed at 40 mL min⁻¹ in this experiment. For a typical test, 0.2 g adsorbent was loaded into the fixed-bed adsorption column, and breakthrough curves were measured at 25 °C and 0.2 MPa atmospheric pressure. The experiment wasn't stopped until the effluent concentration of CO₂ reached the specified influent concentration (100% breakthrough), the dynamic adsorbed amount was obtained from the breakthrough curve and the formula is as follows:Where q is the adsorbed amount, mg g⁻¹; V is the volume flow of the feed gas, mL min⁻¹; M is the molar mass of CO₂; C₀ refers to the CO₂ concentration of the feed gas, mol%; C refers to the CO₂ concentration of the end gas, mol%; t is the adsorption time, min; M_ad refers to the mass of the adsorbent, g; V_m refers to the molar volume of the gas, mL mol⁻¹.

Fig. 1 Apparatus of CO₂ dynamic adsorption.

2.3. Characterizations

The resulting composites were characterized by Energy Dispersive X-ray (EDX), Scanning Electronic Microscopy (SEM), N₂ adsorption–desorption and Thermogravimetric Analysis (TGA). EDX was performed on an Energy Dispersive X-ray Detector (Optima 2000DV, Perkin-Elmer, USA) and used to analyze the element composition of the composites. The microscopic morphology of the substrate and the ILs@SBA-15 composites were investigated by SEM (QUANTA 450, FEI, USA). The surface area, pore volume and the pore size of the substrates and the nanocomposites were determined from N₂ adsorption–desorption isotherms obtained (at 77 K) on a Micromeritics ASAP2010 analyzer. The amount of the impregnated ILs was determined by the weight change of the substrate using an analytical balance (Adventue model AR2140) with the accuracy of 0.1 mg. TGA was conducted on a thermogravimetric analyzer TGA 2050 (Thermal Analysis Instruments, USA) with an air flow rate of 100 ml min⁻¹ and a temperature ramp of 10 K min⁻¹ from 20 to 800 °C to confirm the amount of the impregnated ILs.

Here we define some parameters to illustrate the impregnation results. ILs loading was determined as the ratio of the weight of the impregnated ILs to the total weight of the composites. The impregnation ratio was determined as the ratio of the weight of impregnated ILs to the weight of the ILs initially put in the reactor.

3. Results and discussion

3.1. Influence of the co-solvent

SCFD method involves the dissolution of precursor in scCO₂, the diffusion and impregnation of the supercritical solution into the pores of the substrates and adsorption of the precursor onto the walls of the channels, followed by depressurization. Therefore, the dissolution of ILs in scCO₂ is essential to the successful preparation in this method. However, it was reported that scCO₂ was soluble in the ILs, while the solubility of the ILs in scCO₂ was extremely low.¹⁹ As we know that, scCO₂ has a strong ability to dissolve organic solutes and these organic solutes may act as co-solvent to enhance the ability of scCO₂ to dissolve ILs. Wei-Ze Wu et al.²¹ reported that ILs could dramatically dissolve in scCO₂ with polar organic compounds (e.g. acetone) especially as the concentration of the compounds in scCO₂ exceeds 10 mol%, while the effect of a nonpolar organic compound (e.g. n-hexane) in scCO₂ on the solubility was very limited even when its concentration was as high as 30 mol%.

In this study, methanol, acetone and n-hexane were tried to use as co-solvents respectively. Three samples were prepared using 200 mg SBA-15, 300 mg [Bmim]BF₄, 6 mL co-solvent at deposition temperature of 32 °C, pressure of 16 MPa and deposition time of 24 h. The ILs loading were 49.8%, 9.4% and 0% for the three samples prepared using methanol, acetone and n-hexane as co-solvent respectively. The first thinking coming to our mind was that the different solubilities led to the different ILs loadings about the three samples. Wei-Ze Wu et al.²¹ reported that the polarity of the organic co-solvent was the dominant factor in the solubility enhancement of the ILs in scCO₂. As we know that the polarity of acetone (dipole moment = 2.89 D) was much stronger than that of methanol (dipole moment = 1.69 D), so it can be speculated that the solubility of ILs in acetone–CO₂ system may be much larger than that in methanol–CO₂ system. However, it was strange that the ILs loading of the sample prepared in the acetone–CO₂ system was much smaller than that prepared in the methanol–CO₂ system. We hypothesize that the ILs loading was not only dependent on the solubility of the ILs in the system, but also greatly influenced by the interaction between the co-solvent and the substrates. This phenomenon was also observed when the metal supported nanocomposites were prepared in scCO₂ and co-solvent system.²⁷ Hence, methanol was chosen as co-solvent for all the following experiments.

3.2. Characterization of the impregnated ionic liquids

To further confirm the feasibility and the advantages of preparing impregnated ionic liquids using SCFD method, two samples with the same ILs loading were prepared using SCFD method and traditional impregnation method respectively. The SCFD-ILs@SBA-15 sample was prepared using 100 mg [Bmim]BF₄ as precursor, 200 mg SBA-15 as substrate and 6 mL methanol as co-solvent at pressure of 14 MPa, temperature of 32 °C and deposition time of 24 h. The ILs loading was 13.4% according to the weight difference before and after deposition, and the loading was also confirmed based on the residue mass of the composite materials estimated using the TGA analysis. To confirm whether the ILs were impregnated on the substrate, SCFD-ILs@SBA-15 was analyzed by EDX and the results were shown in Table 1. It can be seen that besides Si and O, there are F element which was derived from [Bmim]BF₄, indicating that ILs have been impregnated on the substrate.

Table 1 Elemental analysis of the SCFD-ILs@SBA-15

Element	Weight percent (%)	Molar percent (%)
C	12.25	18.54
O	45.80	52.04
F	7.59	7.27
Si	32.95	21.33
P	1.39	0.82
Totals	100.00	100.00

---

The N₂ adsorption–desorption isotherms of the mother SBA-15 and the two samples were shown in Fig. 2a. It can be seen that, the isotherms of both SCFD-ILs@SBA-15 and IMP-ILs@SBA-15 showed typical steps in the range of 0.5–0.8 relative pressure (p/p₀) characteristic of a highly ordered mesoporous structure. Compared with the mother SBA-15, the BET surface area of SCFD-ILs@SBA-15 was decreased from 836 m² g⁻¹ to 599 m² g⁻¹, and the total pore volume which was directly derived from the adsorption isotherm (p/p₀ = 0.995) was decreased from 1.05 cm³ g⁻¹ to 0.89 cm³ g⁻¹. The average pore diameter was decreased from 6.1 nm to 5.7 nm. For the IMP-ILs@SBA-15 sample, the surface area, pore volume and average diameter were 599 m² g⁻¹, 0.89 cm³ g⁻¹ and 5.7 nm. All the three parameters decreased significantly after impregnation of ILs for the two samples, indicating effective filling of the ILs into the nanochannels.

Fig. 2 (a) N₂ adsorption–desorption isotherms and (b) pore size distribution of SBA-15 and ILs@SBA-15 prepared by different methods.

Fig. 2b indicates the size distribution of SBA-15 and ILs@SBA-15 derived from the desorption branch, it can be clearly seen that the most probable pore sizes of the samples were decreased after impregnation of ILs for the two samples. It could also be seen from Table 2 that the surface area and the pore volume of the sample prepared using SCFD method were larger than those of the sample prepared using impregnation method. It seems that more ILs were confined within the nanochannels of the substrates using traditional method. Herein we define two parameters to analyze the pore filling of the samples. The actual filling ratio was defined as the ratio of the pore volume decrement of the composite to the pore volume of the substrate. And the theoretical filling ratio was defined as the ratio of the volume of impregnated ILs to the pore volume of the substrate. So the theoretical filling ratio for the two samples were the same, while the actual filling ratio of IMP-ILs@SBA-15 (36.2%) was much larger than that of IMP-ILs@SBA-15 (15.2%). This indicated that for IMP-ILs@SBA-15, a large part of ILs were on the outside of the nanochannels or blocked the pores, which led to the decrease of the surface area and the pore volume. However, the actual filling ratio of the sample prepared using SCFD method was comparable to the theoretical filling ratio, indicating that most of the impregnated ILs were inside the nanochannels of the substrates. The zero surface tension and the gas-like diffusivity of scCO₂ favored the impregnation and dispersion of the ILs into the nanochannels.

Table 2 The N₂ adsorption–desorption analysis results of SBA-15 and ILs@SBA-15 samples

Sample	Surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Average diameter (nm)	Actual filling ratio (%)	Theoretical filling ratio (%)
SBA-15	836	1.05	6.1	—	—
IMP-ILs@SBA-15	466	0.67	5.1	36.2	11.3
SCFD-ILs@SBA-15	599	0.89	5.7	15.2	11.3

---

The SEM images of the mother SBA-15 (Fig. 3a), IMP-ILs@SBA-15 (Fig. 3b and c) and SCFD-ILs@SBA-15 (Fig. 3d–f) were shown in Fig. 3. It can be seen that the structures of the substrates were destroyed after impregnation process (b and c) mainly due to the vigorous agitation, while they kept very well in the SCFD method (d and e). In the SCFD method, the substrates were held in a basket in the upper part of the reactor, which was far from the stirrer. The ILs were transferred into the nanoscale channels with the help of scCO₂ and the co-solvent. In addition, it can be seen from Fig. 3e and f that there were no ILs outside the nanochannels or at the entrance of the pores.

Fig. 3 SEM images of SBA-15 and ILs@SBA-15: (a) SBA-15; (b) and (c) IMP-ILs@SBA-15; (d)–(f) SCFD-ILs@SBA-15.

3.3. Investigation of the operating parameters

In a typical SCFD process, ILs first dissolve in scCO₂ with the assistance of co-solvent, then the mixture of the ILs, co-solvent and CO₂ diffuse into the nanochannles and adsorb onto the substrates. After depressurization, CO₂ and co-solvent are removed and ILs@SBA-15 are obtained. Thus, there are several steps in the whole preparation process and all of them will influence the experimental results. Besides, for each step, there is more than one parameter to be controlled. Generally speaking, there are four important parameters including the deposition pressure, temperature, time and the amount of precursors influencing the experimental results in the SCFD method.²³ In this study, a series of experiments were performed to investigate the influence of the four parameters on the ILs loading and impregnation ratio, in the temperature range of 32–50 °C, pressure range of 9–18 MPa, and deposition time of 6–24 h. The weight of the substrates was fixed at 200 mg and the amount of the co-solvent methanol was fixed at 6 mL for all the experiments.

As can be seen in Table 3 (sample 1–4), the reaction pressure had significant effect on the ILs loading at lower pressure. It was 11.9% at 9 MPa, 35.9% at 14 MPa and 63.2% at 16 MPa. However, the ILs loading for the sample prepared at 16 MPa was comparable to that prepared at 18 MPa. For the scCO₂ at a certain temperature, the pressure determines the solvent power of scCO₂ and thus influences the dissolution of ILs. However, the increment of the solvent power will become smaller at higher pressure. To ensure the large scCO₂ density and its solvent power, as well as consider the equipment capacity and the economic cost, 16 MPa was used as the deposition pressure in the next experiments.

Table 3 Effect of the operating parameters on the deposition results

Sample	Weight of ILs (mg)	Pressure (MPa)	Temperature (°C)	Time (h)	ILs loading (%)	Impregnation ratio (%)
1	500	9	32	24	11.9	5.4
2	500	14	32	24	35.9	22.6
3	500	16	32	24	63.2	69.6
4	500	18	32	24	62.8	67.5
5	100	16	32	24	16.7	43.0
6	200	16	32	24	33.7	49.5
7	300	16	32	24	49.8	66.2
8	400	16	32	24	57.2	67.1
3	500	16	32	24	63.2	69.6
7	300	16	32	24	49.8	66.2
9	300	16	40	24	49.2	63.8
10	300	16	50	24	49.8	65.1
11	300	16	32	6	36.9	38.5
12	300	16	32	12	46.8	58.5
13	300	16	32	18	48.2	59.8
7	300	16	32	24	49.8	66.2

---

As can be seen in Table 3 (sample 5, 6, 7, 8, 3), the ILs loading increased with the increase of the amount of ILs put into the reactor initially. It was 16.7% for 100 mg ILs, 33.7% for 200 mg ILs, 49.8% for 300 mg ILs, 57.2% for 400 mg ILs and 63.2% for 500 mg ILs. It can be seen that the increment of the ILs loading was decreased after the ILs amount was larger than 400 mg. It can be speculated that large amount of ILs won't lead to increase of ILs loading infinitely due to the solubility limitation of ILs in the co-solvent–CO₂ system and the adsorption capacity of the substrates. Herein, the amount of ILs was fixed at 300 mg for the other experiments in the following.

As far as the deposition temperature was concerned, the ILs loading almost kept constant in the temperature range of 32 °C to 50 °C in our experiment. As can be seen from sample 7, 9 and 10, temperature had only a little effect on the ILs loading from 32 to 50 °C. High temperature is beneficial to the diffusion but it is not helpful to the adsorption of the ILs on the substrate. Therefore, the temperature was fixed at 32 °C in the following experiments since it was mild and beneficial to the adsorption.

Another important parameter for SCFD method is the deposition time. It was found that the ILs loading increased with the increase of deposition time. However, the increment of the ILs loading was smaller and smaller after 12 h. It can be seen that the ILs loading increased by 9.9% when the deposition time increased from 6 h to 12 h, while it was increased by only 1.4% when the deposition time increased from 12 h to 18 h and 1.6% from 18 h to 24 h. It indicated that the adsorption equilibrium was obtained after 12 h. Hence the deposition time was suggested to be limited within 6 h. Yin et al.²³ deposited AgNO₃ into mesoporous SBA-15 using scCO₂ as the solvent, the mixture of ethanol and ethylene glycol as the co-solvent, and found that the metal loading changed from 3.99 wt% to 19.6 wt% as the deposition time varied from 3 h to 12 h while it kept almost constant after 12 h (19.6 wt% for 12 h, 19.2 wt% for 19 h and 20.4 wt% for 28 h). They proposed that the adsorption reached equilibrium after 12 h of deposition and the adsorbed amount wouldn't increase with deposition time after this equilibrium.

The results of the N₂ adsorption–desorption isotherms of the ILs@SBA-15 with different loadings of 16.7% (sample 5), 36.9% (sample 11) and 49.8% (sample 7) prepared by SCFD method were shown in Fig. 4a. And the N₂ adsorption–desorption isotherms of the samples with the similar loadings prepared using impregnation method were given in Fig. 4c. It can be seen that, for all the SCFD samples, the nitrogen adsorption isotherms were of type IV isotherms according to IUPAC classifications, and the isotherms displayed a pronounced steep H1 hysteresis loop at relative pressures ranged from 0.5 to 0.8, which were characterized by 2D hexagonal structure of ordered mesoporous materials. It indicated that the ordered mesoporous structures of SBA-15 weren't changed after impregnation of ILs even with a high ILs loading of 49.8%. But for the IMP sample with high loading of 50.3% (the experimental conditions and ILs loading of the impregnation sample are shown in Table 4), the N₂ adsorption isotherms were quite different from the others, almost no volume adsorbed was observed. The size distribution of SCFD samples were shown in Fig. 4b and that of IMP samples were shown in Fig. 4d. The decrease of the most probable pore size indicated the effective filling of ILs into the nanochannels of the substrates.

Fig. 4 (a) and (c) N₂ adsorption–desorption isotherms and (b) and (d) pore size distribution of SBA-15 and ILs@SBA-15 with different ILs loading.

Table 4 The experimental conditions and results of the impregnation method

Sample	Dispersing agent	Dispersing agent amount (mL)	ILs amount (mg)	Time (h)	ILs loading (%)
14	Acetone	5	200	9	13.4
15	Acetone	10	400	16	34.9
16	Acetone	5	800	9	50.3

---

The BET surface area, pore volume, average diameter, actual filling ratio and theoretical filling ratio of the samples prepared by SCFD and impregnation methods were shown in Table 5. It was interesting to find that a sample with the loading of 50.3% prepared using impregnation method had the smallest BET surface area of only 36 m² g⁻¹ while the SCFD sample with the similar loading of 49.8% had a BET surface area of 156 m² g⁻¹. In addition, the actual filling ratios of SCFD-ILs@SBA-15 were similar to the theoretical filling ratios especially with a higher loading, while for the IMP-ILs@SBA-15, there was a big difference between the two parameters. All these phenomena demonstrated that SCFD method made the ILs penetrate into the nanochannels more easily and led to more uniform dispersion compared with the impregnation method.

Table 5 The N₂ adsorption–desorption analysis results of the ILs@SBA-15 prepared by SCFD and impregnation method with different ILs loadings

Sample	Surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Average diameter (nm)	Actual filling ratio (%)	Theoretical filling ratio (%)
SBA-15	836	1.05	6.1	—	—
SCFD-ILs@SBA15-16.7	579	0.71	5.8	32.4	16.6
SCFD-ILs@SBA15-36.9	347	0.56	5.3	46.7	43.9
SCFD-ILs@SBA15-49.8	156	0.29	5.5	72.4	77.9
IMP-ILs@SBA15-13.4	466	0.67	5.1	36.2	11.6
IMP-ILs@SBA15-34.9	246	0.31	5.0	64.8	40.1
IMP-ILs@SBA15-50.3	36	0.07	8.7	93.3	75.7

---

3.4. CO₂ absorption test

The breakthrough curves of CO₂ adsorption of the ILs@SBA-15 samples prepared by the two methods were shown in Fig. 5 and their adsorption capacities calculated from formula (1) were given in Table 6. As can be seen from Fig. 5a that the breakthrough time of the three SCFD samples increased with the increase of the ILs loading, indicating the increasing absorption capacities. When the ILs loading was 36.9%, the adsorption capacity was 51.3 mg g⁻¹ while the adsorption capacity was 29.6 mg g⁻¹ for the mother SBA-15 and 7.0 mg g⁻¹ for the pure [Bmim]BF₄ (ref. 28 and 29) respectively. The main reason for the increase of the adsorption capacity was that the ILs were uniformly dispersed inside the channels of the substrates and the surface area of the ILs was significantly enlarged compared with the liquid ILs. However, the adsorption capacity wouldn't increase with the increase of the ILs loading infinitely, for example, when the ILs loading increased from 36.2% to 49.8%, the adsorption capacity was deceased from 51.3 mg g⁻¹ to 49.5 mg g⁻¹. It seems that the adsorption capacity of the nanocomposite was influenced by both the impregnated ILs and the nanochannels of the substrates. There is an appropriate ILs loading for the maximum of the adsorption capacity. When the ILs loading is small, CO₂ adsorbed by ILs will dominate the whole adsorption capacity of the composites. However, as the filling of the nanochannles completed, firstly, the loaded ILs will be outside the nanochannels and won't contribute to the adsorption capacity;³⁰ secondly, the decrease of the control nanochannels will decrease the adsorption capacity. Xianfeng Wang³⁰ prepared impregnated amino acid ionic liquids into nanoporous microspheres as robust sorbents for CO₂ capture, and determined the effect of IL loadings (i.e. 0, 20, 40, 50, 60, and 100 wt%) on CO₂ capture performance. They reported that the capacity increased with increasing [EMIM][Gly] loading until 50 wt% loading followed by a decrease in capacity. They figured out that the decrease of adsorption capacity at 60 wt% was that the large amount of [EMIM][Gly] could have blocked some of the pores of substrate particle. In our experiment, as can be seen in Table 5, when the loading was 49.7%, the actual filling ratio was 72.4% while the theoretical filling ratio was 77.9%. Therefore, too large loading will lead to block of the substrates and thus reduced the accessible specific surface area of the sorbents, leading to reduced adsorption capacity.

Fig. 5 Breakthrough curves of CO₂ adsorption of (a) SCFD-ILs@SBA-15; (b) IMP-ILs@SBA-15.

Table 6 CO₂ adsorption capacity of the ILs@SBA-15 with different ILs loadings

SCFD	ILs loading (%)	0	16.7	36.9	49.8
	Adsorption capacity (mg g^−1)	29.6	42.5	51.3	49.5
IMP	ILs loading (%)	0	13.4	34.9	50.3
	Adsorption capacity (mg g^−1)	29.6	35.9	40.6	27.1

---

As for the IMP samples, the adsorption capacity was also increased with the increase of the ILs loading which can be seen from Fig. 5b and Table 6, but when the ILs loading was 50.3%, the adsorption capacity was only 27.1 mg g⁻¹ which was comparable to the mother SBA-15. It can be speculated that most of the impregnated ILs were on the outside of the nanochannels or even blocked the pores. And this result was consistent with the N₂ adsorption–desorption isotherm shown in Fig. 4c. In addition, for the samples with the similar ILs loading, the adsorption capacity for the SCFD sample was much larger than that of the IMP sample, indicating more effective filling of the nanochannles and more uniform dispersion of the ILs on the substrates when SCFD method was used.

4. Conclusions

In summary, [Bmim]BF₄ was successfully impregnated into the nanochannels of mesoporous SBA-15 using scCO₂ as solvent and methanol as co-solvent. This novel method has the advantage of uniform dispersion of the impregnated ILs compared to traditional impregnation method. The substrates retained the highly ordered mesoporous structures after the impregnation of ILs into the nanochannels even when the loading was as high as 49.8%. Methanol was found to be an effective co-solvent among the three co-solvents tested in this study. The ILs loading of the nanocomposites were found to increase with the increase of deposition time and pressure, while the deposition temperature had only a little effect on the deposition process in this study. When the ILs loading was 36.2%, the maximum adsorption capacity of 51.3 mg g⁻¹ was obtained. However, the maximum adsorption capacity was not obtained when the ILs loading was the largest, indicating that the CO₂ sorption capacity was partially dependent on the ILs loading. The presented pathway may pave the way for developing similar ILs functionalized porous nanocomposites for CO₂ capture.

Acknowledgements

The work was financially supported by the National Natural Science Foundation of China (21376045, 21506027), Chinese Postdoctoral Science Foundation (2015M571307) and the Fundamental Research Funds for the Central Universities.

References

1. E. D. Bates, R. D. Mayton, I. Ntai and J. H. Davis, J. Am. Chem. Soc., 2002, 124, 926–927.
2. B. B. Jiang, V. Kish, D. Fauth, M. L. Gray, H. W. Pennline and B. Y. Li, Int. J. Greenhouse Gas Control, 2011, 5, 1170–1175.
3. A. Yamasaki, J. Chem. Eng. Jpn., 2003, 36, 361–375.
4. J. T. Yeh, K. P. Resnik, K. Rygle and H. W. Pennline, Fuel Process. Technol., 2005, 86, 1533–1546.
5. K. T. Chue, J. N. Kim, Y. J. Yoo, S. H. Cho and R. T. Yang, Ind. Eng. Chem. Res., 1995, 34, 591–598.
6. E. Diaz, E. Munoz, A. Vega and S. Ordonez, Chemosphere, 2008, 70, 1375–1382.
7. Y. Y. Jiang, G. N. Wang, Z. Zhou, Y. T. Wu, J. Geng and Z. B. Zhang, Chem. Commun., 2008, 505–507,  10.1039/b713648j.
8. B. E. Gurkan, J. C. de la Fuente, E. M. Mindrup, L. E. Ficke, B. F. Goodrich, E. A. Price, W. F. Schneider and J. F. Brennecke, J. Am. Chem. Soc., 2010, 132, 2116–2117.
9. J. Ren, L. B. Wu and B. G. Li, Ind. Eng. Chem. Res., 2012, 51, 7901–7909.
10. Q. Huang, Y. Li, X. B. Jin, D. Zhao and G. Z. Chen, Energy Environ. Sci., 2011, 4, 2125–2133.
11. X. Y. Li, M. Q. Hou, Z. F. Zhang, B. X. Han, G. Y. Yang, X. L. Wang and L. Z. Zou, Green Chem., 2008, 10, 879–884.
12. G. Q. Lai, J. J. Peng, J. Y. Li, H. Y. Qiu, J. X. Jiang, K. Z. Jiang and Y. J. Shen, Tetrahedron Lett., 2006, 47, 6951–6953.
13. L. J. Han, S. J. Zhang, Y. B. Wang, X. J. Yan and X. M. Lu, Inorg. Chem., 2009, 48, 786–788.
14. M. H. Valkenberg, C. deCastro and W. F. Holderich, Green Chem., 2002, 4, 88–93.
15. C. E. Song, W. H. Shim, E. J. Roh and J. H. Choi, Chem. Commun., 2000, 1695–1696,  10.1039/b005335j.
16. N. M. Patil, T. Sasaki and B. M. Bhanage, ACS Sustainable Chem. Eng., 2016, 4, 429–436.
17. H. Ohno, M. Yoshizawa and W. Ogihara, Electrochim. Acta, 2004, 50, 255–261.
18. L. B. Wu, D. An, J. Dong, Z. M. Zhang, B. G. Li and S. P. Zhu, Macromol. Rapid Commun., 2006, 27, 1949–1954.
19. L. A. Blanchard, D. Hancu, E. J. Beckman and J. F. Brennecke, Nature, 1999, 399, 28–29.
20. A. M. Scurto, S. Aki and J. F. Brennecke, J. Am. Chem. Soc., 2002, 124, 10276–10277.
21. W. Z. Wu, J. M. Zhang, B. X. Han, J. W. Chen, Z. M. Liu, T. Jiang, J. He and W. J. Li, Chem. Commun., 2003, 1412–1413,  10.1039/b302311g.
22. D. Y. Zhao, J. L. Feng, Q. S. Huo, N. Melosh, G. H. Fredrickson, B. F. Chmelka and G. D. Stucky, Science, 1998, 279, 548–552.
23. Q.-Q. Xu, Y.-Q. Wang, A.-Q. Wang and J.-Z. Yin, Nanotechnology, 2012, 23, 285602.
24. Q.-Q. Xu, C.-J. Zhang, X.-Z. Zhang, J.-Z. Yin and Y. Liu, J. Supercrit. Fluids, 2012, 62, 184–189.
25. J.-Z. Yin, Q.-Q. Xu and A.-Q. Wang, Chem. Eng. Commun., 2010, 197, 627–632.
26. Q. Q. Xu, G. Xu, J. Z. Yin, A. Q. Wang, Y. L. Ma and J. J. Gao, Ind. Eng. Chem. Res., 2014, 53, 10366–10371.
27. Q. Q. Xu, Y. L. Ma, X. Gang, J. Z. Yin, A. Q. Wang and J. J. Gao, J. Supercrit. Fluids, 2014, 92, 100–106.
28. M. B. Shiflett and A. Yokozeki, Ind. Eng. Chem. Res., 2005, 44, 4453–4464.
29. L. M. G. Sanchez, G. W. Meindersma and A. B. de Haan, Chem. Eng. Res. Des., 2007, 85, 31–39.
30. X. F. Wang, N. G. Akhmedov, Y. H. Duan, D. Luebke and B. Y. Li, J. Mater. Chem. A, 2013, 1, 2978–2982.

---