Plasma–ionic liquid-assisted CO₂ capture and conversion: A novel technology

Since the onset of the industrial revolution, there has been a steady increase in CO₂ emissions into the environment. This poses a significant global climate challenge, prompting the need to find effective strategies for minimizing such greenhouse gas emissions. ^1–3) Ionic liquids (ILs) have recently gained considerable interest in CO₂ capture. ^3–5) The absorption of CO₂ by ILs results from physical interactions between the anions and cations of the ILs and CO₂ molecules. ⁶⁾

An alternative approach to reducing the CO₂ concentration in the atmosphere involves the conversion of CO₂ into valuable products. In recent years, non-thermal plasma (NTP) has experienced substantial growth to enhance the breakdown and conversion of CO₂. ⁷⁾ Our prior research successfully demonstrated that ILs can protect the protein structure in the presence of NTP. ^8–10) These studies showed that ILs are stable in the presence of NTPs and do not denature the proteins. Nevertheless, no study to date has demonstrated the capture and conversion of CO₂ using a plasma–IL combination. In a previous study, Li and Gallucci ¹¹⁾ reported the capture of CO₂ using hydrotalcite pellets and later the desorption and CO₂ conversion using an Ar feed gas coaxial DBD reactor. Moreover, Masaaki et al. reported that plasma-based desorption was higher than thermal desorption of CO₂. ¹²⁾

Hence, in this pioneering effort, we employed an IL for CO₂ capture [Fig. 1(a)] and storage [Fig. 1(b)] followed by the conversion of the captured CO₂ into CO using NTP [Fig. 1(c)]. This study used 1-Butyl-3-methylimidazolium chloride ([Bmim]Cl) IL, obtained from Sigma-Aldrich. We implemented streamer plasma (as described in our previous articles ^13,14 ) to convert captured CO₂ to CO, as shown in Fig. 1. In the prior investigation, Jang et al. showed the solubility of CO₂ in [Bmim]Cl IL at a temperature range of 353.15–373.15 K under high pressure conditions at approximately 40 MPa. ¹⁵⁾ [Bmim]Cl IL also showed versatility in applications such as supercritical carbon dioxide preparation, ¹⁶⁾ cellulose dissolution, ¹⁷⁾ etc. Therefore, we utilized [Bmim]Cl IL to effectively capture and store CO₂ at room temperature and converted the stored CO₂ into CO using NTP.

Zoom In Zoom Out Reset image size

In this study, we introduced CO₂ with a purity of 99.99% into a 45 ml solution of deionized (DI) water and DI water + [Bmim]Cl (4, 12, and 20 wt%) along with 5 ml of ionic strength adjuster for CO₂. CO₂ was introduced into the reactor at a flow rate of 0.372 l min⁻¹ through a gas tube with a diameter of 4.35 mm for 2 min. Subsequently, we measured the absorbed CO₂ concentration using the sensor (details provided in the supporting files). We used pure CO₂ in the present study to avoid the influence of other gases (diluting gas) on holding the CO₂ in the IL solution and the role of dilute gas in CO₂ conversion. After 10 min of CO₂ supply, the CO₂ concentration in the solution was 630 ± 50 mg l⁻¹, which later decreased to 9 ± 4 mg l⁻¹ after 1440 min, as displayed in Fig. 2(a). Hence, the loss of CO₂ concentration after 1440 min was 621 mg l⁻¹. On the other hand, for [Bmim]Cl (4 wt%), the CO₂ concentration was 860 ± 55 mg l⁻¹ after 10 min of CO₂ supply, although the final CO₂ concentration after 1440 min was 310 ± 20 mg l⁻¹ [see Fig. 2(a)]. Furthermore, with an increase in the [Bmim]Cl concentration to 10 wt%, the CO₂ concentration after 10 min was 830 ± 60 mg l⁻¹, and after 1440 min, it was 330 ± 11 mg l⁻¹ [Fig. 2(a)]. This observation indicates that CO₂ absorption decreases at higher CO₂ concentrations but sustains a substantial level over time.

Zoom In Zoom Out Reset image size

Later, we increased the [Bmim]Cl concentration to 20 wt%, and after 10 min of CO₂ supply, the measured CO₂ concentration was 730 ± 40 mg l⁻¹, which was lower than the concentrations observed with 4 wt% and 10 wt% [Bmim]Cl [see Fig. 2(a)]. However, after 1440 min, the captured CO₂ concentration was 370 ± 15 mg l⁻¹, which was higher than the concentrations observed with other wt% values of [Bmim]Cl. Note that the study did not explore concentrations of [Bmim]Cl higher than 20 wt% due to the elevated solution viscosity, making it challenging to measure CO₂ accurately with the sensor. The results suggest that the CO₂ absorption decreases as the [Bmim]Cl concentration increases. This trend occurs due to the rising viscosity of the solution, leading to a reduction in CO₂ solubility, and this correlation is also supported by previous studies. ^3,18) Meanwhile, the respective losses in CO₂ concentration after 1440 min were 550, 500, and 360 mg l⁻¹ for 4, 10, and 20 wt% [Bmim]Cl solutions. This study demonstrates that at higher concentrations of [Bmim]Cl, the IL in the solution can effectively retain the captured CO₂ for a longer time at atmospheric pressure and room temperature than lower concentrations of [Bmim]Cl solutions.

In the subsequent section, we measured the CO formation after 1440 min from the stored CO₂ in DI water and [Bmim]Cl solutions, which were exposed to streamer plasma for 1 min at atmospheric pressure without any feed gas (discharge occurs without gas flow, maintaining a relative humidity of approximately 45%. Following the plasma discharge, N₂ gas flowed to remove the CO or other gases produced during the discharge from the reactor). Five milliliters of the solution was transferred to the plasma reactor and then treated with plasma for 1 min (see Fig. 1). A 5 ml solution was used due to current reactor design limitations; however, we plan to treat a larger volume of the solution in future studies. The plasma discharge power was 1.3 W for DI water and [Bmim]Cl (2 wt%), while for [Bmim]Cl (10 wt% and 20 wt%), the discharge power was slightly decreased to 1.2 W due to an increase in the ionic state of the solution (see the supporting file for details). The solution's plasma treatment resulted in CO production in the gas phase, which was detected by a CO sensor (described in the supporting file). The plasma treatment on the DI water without CO₂ flow solution resulted in a concentration of 8.28 × 10⁻⁷ ± 0.85 × 10⁻⁷ moles of CO. Meanwhile, the CO concentrations for solutions of 4, 10, and 20 wt% [Bmim]Cl without CO₂ flow were 1.49 × 10⁻⁶ ± 0.71 × 10⁻⁷, 1.58 × 10⁻⁶ ± 0.47 × 10⁻⁷, and 1.62 × 10⁻⁶ ± 0.81 × 10⁻⁷ moles, respectively [see Fig. 2(b)].

On the other hand, plasma discharge on the CO₂ flow solution after 1440 min resulted in the following concentrations: 1.67 × 10⁻⁶ ± 0.21 × 10⁻⁶ moles for water, 1.85 × 10⁻⁶ ± 0.18 × 10⁻⁶ moles for [Bmim]Cl 4 wt%, 1.90 × 10⁻⁶ ± 0.12 × 10⁻⁶ moles for [Bmim]Cl 10 wt%, and 2.38 × 10⁻⁶ ± 0.19 × 10⁻⁶ moles for [Bmim]Cl 20 wt% [Fig. 2(b)]. These findings indicate that more CO was produced for higher concentrations of [Bmim]Cl, possibly because these solutions stored more CO₂. The CO generation from CO₂ is due to the electron-based reaction, and we assume that NTP treatment can break the physical binding of CO₂ with IL and convert it to CO [Eq. (1)]. The O generated from the reaction is later converted to O₂ [see Eq. (2)] and released into the air or bound with the IL solution. The fate of the O₂ bound with the IL solution is unknown. There are many possible CO₂ and H₂O dissociation products, as explained in our previous study. ¹³⁾ Some are mentioned below; refer to Eqs. (1)–(9).

$$\begin{eqnarray}&&{{CO}}_{2}+e\to {CO}+O\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&O+O+M\to {O}_{2}+M\end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray}&&{{CO}}_{2}+e\to e+C+{O}_{2}\end{eqnarray} \tag{ 3 }$$

$$\begin{eqnarray}&&{{CO}}_{2}^{+}+e\to C+{O}_{2}\end{eqnarray} \tag{ 4 }$$

$$\begin{eqnarray}&&{H}_{2}O+e\to {OH}+H+e\end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray}&&O+{O}_{2}+M\to {O}_{3}+M\end{eqnarray} \tag{ 6 }$$

$$\begin{eqnarray}&&{O}_{3}+H\to {OH}+{O}_{2}\end{eqnarray} \tag{ 7 }$$

$$\begin{eqnarray}&&{O}_{3}+{OH}\to {{HO}}_{2}+{O}_{2}\end{eqnarray} \tag{ 8 }$$

$$\begin{eqnarray}&&{OH}+{OH}+M\to {{H}_{2}O}_{2}+M\end{eqnarray} \tag{ 9 }$$

The detection of CO in the gas phase from those solutions that were not treated with CO₂ could be due to the decomposition of the [Bmim]Cl solution or the decomposition of captured CO₂ during solution preparation. However, we detected CO in the DI water (not treated with CO₂); this CO detection was due to the solubilized CO₂ in the DI water. Similarly, it is highly possible that the CO detection in the [Bmim]Cl solutions (not treated with CO₂) was due to captured CO₂ in the [Bmim]Cl during solution preparation. As shown in Fig. 2(a), [Bmim]Cl solution holds more CO₂ than water; therefore, more CO₂ is held in the [Bmim]Cl during solution preparation, resulting in higher CO formation. If we compare the data of [Bmim]Cl 4, 10, and 20 wt% (without CO₂ flow) with only water, more CO is detected. Although there is a slight increase in CO between [Bmim]Cl 4 wt% and 10 wt%, the increase in CO is negligible between the solutions of [Bmim]Cl 10 wt% and 20 wt%. Indeed, if CO were produced because of the degradation of [Bmim]Cl, its impact would be negligible. This conclusion arises from the observation that the CO concentration did not rise proportionally with the increase in [Bmim]Cl concentration. Additionally, our previous study showed that IL can protect the hemoglobin protein structure in the presence of DBD treatment using different feed gases for 10 min. ⁹⁾ Additionally, we checked the absorbance spectra of [Bmim]Cl IL (4, 10, and 20 wt%) solutions before and after plasma treatment. Figure 3(a) shows the absorbance spectra of 1 min NTP-treated water, [Bmim]Cl IL (4 wt%) control (without NTP treatment), and 1 min NTP-treated [Bmim]Cl IL (4 wt%). No peak shift was observed before and after NTP treatment. The absorbance spectra of the [Bmim]Cl IL control (without NTP treatment) were similar to the literature. ¹⁹⁾ This shows that a 1 min plasma treatment did not degrade the [Bmim]Cl. Similar results were also observed for the [Bmim]Cl IL (10 wt% and 20 wt%) control (without NTP treatment) and 1 min NTP-treated [Bmim]Cl IL (10 wt% and 20 wt%), as shown in Figs. 3(b) and 3(c). Although we are not neglecting the degradation of [Bmim]Cl IL caused by the plasma, the current study concludes that if the CO production were due to [Bmim]Cl IL degradation, then it would be a very small or insignificant amount.

Zoom In Zoom Out Reset image size

To address any possibility of [Bmim]Cl degradation with longer treatment durations, we measured the absorbance spectra after 3, 5, and 7 min, as shown in Fig. S1. Figure S1(a) represents the absorbance spectra of DI water after NTP treatment for different treatment times. Figures S1(b), S1(c), and S1(d) show a slight increase in intensity in the absorbance spectra for the 7 min NTP treatment to IL solution. However, this might be attributed to the accumulation of reactive species in the solution or the possibility of hypochlorite formation, as illustrated in Eqs. (10)–(12).

$$\begin{eqnarray}&&{{Cl}}^{-}+O\to {{ClO}}^{-}\end{eqnarray} \tag{ 10 }$$

$$\begin{eqnarray}&&{{Cl}}^{-}+{O}_{3}\to {{ClO}}^{-}+{O}_{2}\end{eqnarray} \tag{ 11 }$$

$$\begin{eqnarray}&&{{Cl}}^{-}+2{OH}\to {{ClO}}^{-}+{H}_{2}O\end{eqnarray} \tag{ 12 }$$

Figure 4 shows no visible change in the [Bmim]Cl solution before and after NTP treatment (1 min and 7 min). The abstraction of hydrogen or adding OH in 1-Butyl-3-methylimidazolium is possible, but it requires high energy to break the carbon chain from 1-Butyl-3-methylimidazolium to produce CO. However, we will conduct further detailed studies on the degradation of 1-Butyl-3-methylimidazolium in the future.

Zoom In Zoom Out Reset image size

Later, we calculated the percentage of energy efficiency for CO production (η %) using Eq. (13) from Ref. 11

$$\begin{eqnarray}&&{\eta } \% =\left(\frac{{\unicode{x02206}H}_{298}^{^\circ }\left(\frac{J}{{mol}}\right)\times {CO}\,{production}({mol})}{{plasma}\,{discharge}\,{power}\,\left(W\right)\times {time}(s)}\right)\times 100\end{eqnarray} \tag{ 13 }$$

For water, the η% was 0.60%, while for [Bmim]Cl IL 4 wt%, 10 wt%, and 20 wt%, the η% was 0.67%, 0.75%, and >0.94%, respectively. This indicates that a 60 s NTP treatment can generate CO, but simultaneously, plasma energy was also utilized in the CO₂ desorption, and production of reactive oxygen and nitrogen species (RONS) due to water dissociation [Eqs. (1)–(12)], heat, etc. There is no direct comparison of our work with any reported studies, as our CO₂ absorbent IL is mixed with water, and we have not used any other gas flow that can aid in desorption. Nevertheless, in a previous study, researchers reported a maximum energy efficiency of 0.98% for the first 400 s, which later decreased to 0.68% for 1000 s, using Ar plasma with hydrotalcite pellets as a CO₂ sorbent. ¹¹⁾ Compared to the reported work by Li and Gallucci, ¹¹⁾ our energy efficiency for CO production was slightly less for [Bmim]Cl IL 20 wt%. However, there are significant differences in the treatment conditions and the target of the work (use of water-soluble sorbent).

Furthermore, we conducted non-reactive molecular dynamics (MD) simulations to calculate the free energy profiles (FEPs) of CO₂ and CO in both water and a mixture of water + [Bmim]Cl. MD simulations were carried out using umbrella sampling (US) simulations. ^20,21) The US simulation technique provides comprehensive FEPs along the reaction coordinates for CO₂ and CO. Details of the simulation procedures can be found in the supporting file. Figure 5 illustrates that the FEPs of CO₂ and CO decrease at the vacuum–water interface, indicating minimum values of ∆G that facilitate the accumulation of CO₂ and CO. This suggests an enhancement of the solute molecule concentration at the vacuum–water or vacuum–water + [Bmim]Cl interface.

Zoom In Zoom Out Reset image size

The ∆G_gs values were calculated to represent the free energy change for the vacuum–water interface, whereas ∆G_sl was associated with the free energy barriers transitioning from the vacuum–water interface into the liquid phase. Hydration-free energy (∆G_hydr) values were used to elucidate the transport of CO₂ and CO molecules from the gaseous phase into the solution (water or water + [Bmim]Cl). Comprehensive explanations of the ∆G_gs, ∆G_sl, and ∆G_hydr measurements are available in previous studies. ^22,23) The ∆G_gs, ∆G_sl, and ∆G_hydr values for CO₂ in water were determined to be −5.52, 8.97, and 3.45 kJ mol⁻¹, respectively [see Fig. 5(a) and Table I]. In contrast, the corresponding ∆G_gs, ∆G_sl, and ∆G_hydr values for CO₂ in water + [Bmim]Cl were −6.83, 8.47, and 1.64 kJ mol⁻¹, respectively [see Fig. 5(a) and Table I]. These simulation results closely align with our experimental findings, indicating that the ∆G_gs and ∆G_hydr values are lower in water + [Bmim]Cl compared to water, clearly demonstrating that CO₂ encounters less resistance in the IL solution, potentially increasing the amount of CO₂ molecules entering into the solution (water + [Bmim]Cl) compared to water.

For CO in water, the ∆G_gs, ∆G_sl, and ∆G_hydr values were −8.93, 13.36, and 4.43 kJ mol⁻¹, respectively (see Table I). Conversely, for CO in water + [Bmim]Cl, the ∆G_gs, ∆G_sl, and ∆G_hydr values were −1.11, 14.45, and 13.34 kJ mol⁻¹, respectively [see Fig. 5(b) and Table I]. These US simulations align with our experimental results, as the ∆G_gs and ∆G_hydr values were lower for CO permeation in water than in water + [Bmim]Cl. This implies that after CO production by plasma, there is a reduced chance of absorption in IL compared to water.

In summary, we have introduced a novel technology that integrates plasma and [Bmim]Cl IL for the simultaneous capture, storage, and conversion of CO₂ for the first time. Our experimental results indicate that the water + [Bmim]Cl solution can store CO₂ under atmospheric pressure and room temperature. Moreover, the release of CO₂ during plasma treatment produces CO. Our MD simulation supports our experimental findings, suggesting that CO₂ molecules easily transition from the gaseous phase into the water + [Bmim]Cl solution. In contrast, the penetration of CO molecules into the water + [Bmim]Cl solution is more challenging than water alone. This observation implies that once plasma produces CO, its solubility in the IL solution may be limited, showcasing the potential of this technology for efficient CO₂ capture and conversion.

Acknowledgments