Revealing how molten salts promote CO₂ capture on CaO via an impedance study and sorption kinetics simulation

Abstract

Electrochemical impedance spectroscopy analyses were utilised to explore the oxygen ion conductivity of alkali metal salts promoted CaO, which revealed that oxysalts, such as carbonate and sulfate with good ion conductivity, promote CO₂ capture.

---

Introduction

CaO as a high temperature CO₂ sorbent is promising for some CO₂-free sorption-enhanced steam reforming processes.^1,2 It has been proposed to be applied in both pre- and post-combustion capture processes based on the reversible carbonation/calcination reaction,³ and could potentially be applied in the sorption-enhanced steam reforming (SESR) reaction, which is a candidate technique for producing pure H₂via simultaneous reforming and adsorptive separation of CO₂.^4–6 By removing CO₂ from the reaction products, the balance is driven to the right-hand side toward a higher conversion and yields a pure hydrogen steam. Moreover, the reaction temperature can be reduced to around 500–600 °C, which is much lower than those used in conventional steam reforming (800 °C). Hence, CaO is regarded as one of the most promising sorbents.^7–9

Although CaO exhibits high CO₂ capture capacity under stoichiometric conditions (17.86 mmol g⁻¹), conventional CaO sorbents are not able to achieve complete conversion at the end of the first carbonation due to a diminution of availability of CaO active sites. Along with the carbonation reaction, the blockage of internal pores prevents CO₂ from reaching the remaining active surface area. This situation significantly increases the diffusion resistance of CO₂ throughout the CaCO₃ product layer, creating a slower diffusion-limited reaction phase.^10,11 Various techniques have been proposed to improve the CO₂ uptake of CaO or to reactivate the used CaO. Hu et al.,¹² for instance, employed a series of organic acids to modify limestone to produce CaO sorbents and the modified sorbent maintained a higher specific surface area (SSA) that contributed to the superior performance of the sorbent. Akgsornpeak et al.¹³ developed synthetic CaO via a sol–gel method and investigated the effect of the Ca²⁺/CTAB ratio. They found that the presence of CTAB can effectively prevent the agglomeration of CaO particles, which can greatly increase the BET SSA and pore volume of the resulting CaO sorbents. Apart from the structural improvement, an alternative strategy to increase the CO₂ uptake capacity of CaO is the doping of alkali metal salts. Reddy et al.¹⁴ reported that doping CaO with alkali metals can result in an improvement of CO₂ uptake capacity. The addition of alkali metals improved the CO₂ uptake capacity in the following order Cs > Rb > Na > K > Li. It was found that CO₂ is preferably adsorbed on Cs₂O rather than CaO, CsOH- and Cs₂CO₃-doped CaO sorbents due to the formation of a highly dispersed Cs₂O on the CaO support.

Our group previously reported that mixed alkali metal carbonate molten salts promoted CaO with superior CO₂ capture performance. However, there is still no clear understanding of how such alkali metal carbonates facilitate the CO₂ capture by CaO. According to the reports on alkali metal nitrite coated MgO,^15,16 we assume that the existence of O²⁻ in the liquid molten salt phase might generate CO₃²⁻ ions, which accelerates the CO₂ capture process. To demonstrate this assumption, we compared the CO₂ capture capacity of different mixed metal salts coated CaO and conducted an electrochemical impedance spectroscopy experiment to explore the ionic conductivity of each of the sorbents.

Results and discussion

Previously, a negative correlation between the CO₂ uptake and the melting point temperature was demonstrated.¹⁷ All of the selected molar ratios of M₁ : M₂, at which the melting point of (M₁–M₂)X was minimal, are listed in Table 1. Fig. 1 shows the CO₂ uptake of neat CaO, hydrated CaO, and 10 mol% binary alkali metal salts (M₁–M₂)X, (M = Li, Na, K; X = SO₄²⁻, CO₃²⁻, Cl⁻) coated CaO when exposed at 600 °C to 100% dry CO₂ at ambient pressure (1 bar) for 1 h. It is apparent that all samples show typical two differentiated phases, a fast reaction controlled stage followed by a much slower solid state diffusion stage, which was generally accepted for CaO carbonation. However, different trends could still be observed for neat CaO, hydrated CaO, and alkali metal molten salt coated CaO. For neat CaO, both the CO₂ sorption capacity and sorption kinetics were relatively poor. After 1 h of sorption, the CO₂ uptake was still less than 3.26 mmol g⁻¹, and in the first minute, only 31.81% of its maximum CO₂ uptake (60 min) was achieved. For hydrated CaO, it showed a slightly higher CO₂ uptake (6.44 mmol g⁻¹) than neat CaO. It was suggested that more cracks could be formed during the hydration process, creating more channels that extend to the interior of the particles, with a consequent improvement in the CO₂ capture capacity of CaO.^18,19 However, its sorption kinetics were still very low; only 29.82% of its maximum CO₂ uptake (60 min) was achieved within the first minute. In this process, a gas–solid reaction between CO₂ and CaO on the surface of the grains took place, leading to the formation of a thin layer of CaCO₃, which normally occurred in short periods of time and ended with saturation of the surface active sites.

Table 1 Molar ratio and melting point for metal salts mixtures (M₁–M₂)X, (M = Li, Na, K; X= SO₄²⁻, CO₃²⁻, Cl⁻) and CO₂ uptake capacity for coated sorbents

Component	Ratio (M1 : M2)^a	Melting point^a (°C)	Uptake capacity (mmol g^−1)
a The molar ratio and melting point were obtained from FactSage database.^21
(Na–K)2SO4/CaO	0.742 : 0.258	831	7.52
(Na–K)2CO3/CaO	0.555 : 0.445	710	8.55
(Na–K)Cl/CaO	0.506 : 0.494	657	8.02
(Li–Na)2SO4/CaO	0.617 : 0.383	591	9.73
(Li–K)2SO4/CaO	0.81 : 0.19	530	10.93
(Li–Na)2CO3/CaO	0.52 : 0.48	500	10.27
(Li–K)2CO3/CaO	0.428 : 0.578	498	10.38
(Li–Na)Cl/CaO	0.72 : 0.28	554	1.40
(Li–K)Cl/CaO	0.592 : 0.418	353	2.41
CaO	—	—	3.26
Hydrated CaO	—	—	6.44

---

Fig. 1 CO₂ uptake by (a) CaO; (b) alkali metal chloride coated CaO (MCl/CaO); (c) alkali metal carbonate coated CaO (MSO₄/CaO); and (d) alkali metal sulfate coated CaO (MCO₃/CaO).

While for (Li–K)Cl/CaO and (Na–K)Cl/CaO, the CO₂ uptakes for the fast reaction stage were even lower than neat CaO, demonstrating a decrease in the active sites on the surface. The melting points of (Li–K)Cl and (Na–K)Cl were below the sorption temperature (600 °C), leading the phase transformation from solid to liquid, which covered the surface of the active sorbents and hindered the reaction. This data suggested that the alkali chloride molten salts could not provide any promotion effect on the CO₂ capture on CaO, but, on the contrary, resulted in an inhibiting effect.

All other sorbents coated with alkali carbonates and sulphates showed much higher CO₂ uptake capacity during the carbonation. With 10 mol% (Li–K)₂SO₄, (Li–Na)₂SO₄, and (Na–K)₂SO₄, the CO₂ uptake of CaO was promoted to be 10.93, 9.73, and 7.52 mmol g⁻¹, respectively. With 10 mol% (Li–K)₂CO₃, (Li–Na)₂CO₃, and (Na–K)₂CO₃, the CO₂ uptake of CaO was promoted to be 10.38, 10.27, and 8.55 mmol g⁻¹, respectively. The sorption kinetics were also significantly improved by doping with alkali carbonates and sulphates. For instance, with 10 mol% (Li–K)₂SO₄, (Li–Na)₂SO₄, (Li–K)₂CO₃, and (Li–Na)₂CO₃ coating, 56%, 51%, 64%, and 54% of their maximum CO₂ uptakes (60 min) were achieved within the first minute, which were much higher than that for neat CaO. After the relatively fast initial reaction, a slower reaction stage controlled by diffusion in the pores or in the product layer takes place, resulting in a slower reaction rate. The increased carbonation rate resulted in fast cessation of carbonation, and the quickly formed CaCO₃ product layer may hinder further carbonation.²⁰

The components, molar ratio of molten salts, melting point and CO₂ uptake capacity of promoted CaO are listed in Table 1. As the melting point of the eutectics can be tuned by changing the molar ratio to influence the CO₂ uptake of the promoted CaO sorbents, the minimum melting point eutectics were optimized for loading. When the sorption temperature was kept at 600 °C, different phases of salt mixtures were formed; for example, (Na–K)₂SO₄ which has a melting point of 831 °C will be in solid phase, which showed limited promotion on CO₂ uptake, maybe due to the formation of a chemically bound Ca–O–K surface species, the same as the K₂CO₃-promoting mechanism in Layered Double Hydroxide (LDH).^22,23 In contrast, for (Li–Na)₂SO₄/CaO or those listed downward, which were used as a sorbent, the coating mixture salts melted under identical conditions and led to a significant promoting effect on CO₂ uptake. To explore the intrinsic reason for this promotion effect, (Li–K)₂CO₃/CaO, (Li–K)₂SO₄/CaO, and (Li–K)Cl/CaO were chosen as the representative molten salts.

To explore the impact of structure and morphology, XRD, SEM, and BET analyses were measured (Fig. S2–S4, ESI†). The results showed that all samples consisted of alkali metal salts and Ca(OH)₂. Upon calcination, Ca(OH)₂ transformed into crystalline CaO and no other impurities were detected. The grain diameters of these three modified CaO samples were similar in shape, which led to an analogous microstructure and surface area (Table S1, ESI†).

Ion-diffusion efficiencies can be measured through a determination of mobility or more directly through an electrical measurement of ion conductivity.²⁴ Impedance is the response of an electrochemical system to an applied alternating voltage. The frequency dependence of impedance can reveal the underlying processes in electrochemical systems. Fig. 2 shows the impedance spectra in Nyquist form obtained from CaO, (Li–K)Cl/CaO, (Li–K)₂SO₄/CaO, and (Li–K)₂CO₃/CaO. For CaO and (Li–K)Cl/CaO, the impedance spectra were totally irregular, which showed similar trends with an open circuit (Fig. 2(a) and (b)) corresponding to the CV test (Fig. S5a and b†). The real part (Z′) in the Nyquist plot coordinated values was around 10⁵ Ω, corresponding to an impedance of about 10⁵ Ω in magnitude. Fig. 2(c) and (d) show typical impedance spectra for (Li–K)₂SO₄/CaO and (Li–K)₂CO₃/CaO; the resistance was calculated from the intersection with the real axis at high frequency by using ZSimDemo software. (Li–K)₂SO₄/CaO and (Li–K)₂CO₃/CaO showed much lower resistance (72.36 Ω and 27.09 Ω), which corresponds to a high ionic conductivity according to the formula σ = L/RA, where σ is the ionic conductivity, R is the bulk resistance, and L and A are the thickness and sectional area of the pellets.²⁴L and A are constants when a sample is prepared, thus σ was numerical inversely proportional to A. It can be seen that the oxygen ion conductivity for (Li–K)₂SO₄/CaO and (Li–K)₂CO₃/CaO was several orders of magnitude higher than that for CaO and (Li–K)Cl/CaO. These results are in concordance with CO₂ sorption tests that (Li–K)₂SO₄ and (Li–K)₂CO₃ coated CaO resulted in excellent CO₂ uptakes due to good O²⁻ conductivity, which offered a continuous stream of CO₃²⁻ for the carbonation reaction.

Fig. 2 Impedance spectroscopy of (a) CaO; (b) (Li–K)Cl/CaO; (c) (Li–K)₂SO₄/CaO; and (d) (Li–K)₂CO₃/CaO.

To understand the detailed reaction process for CO₂ capture on neat and promoted CaO samples, the dynamic variations during the reaction of the particles with CO₂ were further examined by fitting to the well accepted double exponential model (Fig. S6, ESI†) as reported for other sorbents.^25,26 The double exponential model is:

y = A exp(−k₁t) + B exp(−k₂t) + C (1)

In eqn (1), y represents the CO₂ uptake of sorbents, t is the sorption time in minute, and k₁ and k₂ are the kinetic parameters in the surface chemical reaction and bulk diffusion process, respectively. Constants A and B are the coefficient of each process that controls the whole CO₂ sorption process, and C was the y value when t tending to infinity indicates the maximum sorption capacity. For (Li–K)Cl/CaO and hydrated CaO which did not show O²⁻ transmittance capacity, A is larger than B (see Table S2, ESI†), suggesting that the diffusion stage has a great influence on the overall CO₂ sorption process. By changing the coating mixtures to (Li–K)₂CO₃ and (Li–K)₂SO₄, B became larger than A, demonstrating that the diffusion stage was no longer a limiting factor, the chemisorption process with a fast reaction rate contributed more to the total uptake capacity.

The possible mechanism for the effect of different alkali metal coatings was illustrated schematically in Fig. 3, in regard to (Li–K)₂CO₃, (Li–K)₂SO₄ coated CaO particles (Fig. 3(a)), the alkali metal mixtures show a lower melting point indicating that it will be in a molten phase at 600 °C with good oxygen ion conductivity, the adsorbed CO₂ combined with O²⁻ to generate CO₃²⁻, which formed CaCO₃ by reaction with CaO at the CaO–CaCO₃ interface as shown in eqn (2)–(4).¹¹

Fig. 3 Schematic diagrams of the reaction for CO₂ with coated CaO particle at 600 °C, (a) CaO coated with low melting point oxysalt eutectic: (b) CaO coated with anaerobic salts and (c) CaO coated with high melting point mixtures.

At the CaO–CaCO₃ interface:

CO₃²⁻ + CaO → CaCO₃ + O²⁻ (2)

At the pore surface:

CO₂(g) ⇄ (CO₂)_ads (3)

(CO₂)_ads + O²⁻ → CO₃²⁻ (4)

The generated O²⁻ interacted with adsorbed CO₂ in the carbonates molten salts to generate the carbonate ions (CO₃²⁻).²⁷ (Li–K)₂CO₃ and (Li–K)₂SO₄ possessed a good O²⁻ transmission ability (as discussed in Fig. 2), which ensured a continuous counter diffusion of CO₃²⁻ and O²⁻ and made a great contribution to the diffusion stage, leading to a complete carbonation. While (Li–K)Cl does not possess any O²⁻ transmission capacity at 600 °C, even when it was in the liquid phase. On the contrary, the (Li–K)Cl molten mixtures lay over the surface of the CaO sorbents and cover the basicity, leading to a low rate of carbonation in the first fast reaction stage. In (Li–K)Cl molten mixtures, the generated O²⁻ and CO₃²⁻ cannot diffuse from the gas–liquid surface to the liquid–solid and solid–solid interface, and the interrupted counter-diffusion of CO₃²⁻ and O²⁻ hindered the continuous reaction of CaO and CO₂ as shown in Fig. 3(b). For (Na–K)X (X= SO₄²⁻, CO₃²⁻, Cl⁻) that have a higher melting point, the mixtures will be in the solid state at the CO₂ adsorption temperatures. This solid mixture could not make a contribution to the transmission of CO₃²⁻ and O²⁻. However, the K ions dispersed into the CaO sorbents and created more Ca–O–K bonds to enhance the promotion effect, the same as the LDH reveals.²² This explains the reason that (Na–K)X (X= SO₄²⁻, CO₃²⁻, Cl⁻) only showed a limited promotion on CO₂ uptake as shown in Fig. 3(c).

Conclusions

This study demonstrates that the uptake of CaO can be influenced by different mechanisms. By loading the eutectic mixtures with a low melting point, but no O²⁻ conduction capability, such as (Li–K)Cl, (Li–Na)Cl, the CO₂ uptake decreased dramatically due to the reduction of the basic active sites. By loading the eutectic mixtures with a high melting point, the CO₂ uptake was only slightly improved. By loading the eutectic mixtures with a low melting point (lower than the sorption temperature) and a good O²⁻ transmission capacity, such as (Li–K)₂CO₃ or (Li–K)₂SO₄, the CO₂ uptake can be improved significantly. The highest capacity of over 10.9 mmol g⁻¹ at 600 °C can be attained. The peculiar effect of alkali metal salts was attributed to the presence of a high concentration of oxygen ions in the molten alkali metal salts that facilitated the generation of carbonate ions (CO₃²⁻), resulting in the rapid formation of CaCO₃ and ease of regeneration of the particles at high temperatures. This combined effect of high O²⁻ concentration and O²⁻ conductivity was seen to have the most benefit for the whole carbonation process.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Fundamental Research Funds for the Central Universities (2016ZCQ03), the National Natural Science Foundation of China (51622801, 51572029, and 51308045), and the Beijing Excellent Young Scholar (2015000026833ZK11).

Notes and references

1. A. Perejón, L. M. Romeo, Y. Lara, P. Lisbona, A. Martínez and J. M. Valverde, Appl. Energy, 2016, 162, 787–807.
2. Z. H. Lee, S. Ichikawa, K. T. Lee and A. R. Mohamed, J. Energy Chem., 2015, 24, 225–231.
3. J. Blamey, M. Zhao, V. Manovic, E. J. Anthony, D. R. Dugwell and P. S. Fennell, Chem. Eng. J., 2016, 291, 298–305.
4. L. Barelli, G. Bidini and F. Gallorini, Appl. Energy, 2015, 143, 110–118.
5. A. Di Giuliano, J. Girr, R. Massacesi, K. Gallucci and C. Courson, Int. J. Hydrogen Energy, 2017, 42, 13661–13680.
6. M. Z. Memon, X. Zhao, V. S. Sikarwar, A. K. Vuppaladadiyam, S. J. Milne, A. P. Brown, J. Li and M. Zhao, Environ. Sci. Technol., 2017, 51, 12–27.
7. A. Antzara, E. Heracleous, D. B. Bukur and A. A. Lemonidou, Int. J. Greenhouse Gas Control, 2015, 32, 115–128.
8. I. Zamboni, C. Courson and A. Kiennemann, Appl. Catal., B, 2017, 203, 154–165.
9. G. Ji, X. Xu, H. Yang, X. Zhao, X. He and M. Zhao, Environ. Sci. Technol., 2017, 51, 11484–11492.
10. A. Ebrahimi, M. Arab, M. Saffari, A. I. Minett and T. Langrish, Chem. Eng. J., 2016, 291, 1–11.
11. S. K. Bhatia and D. D. Perlmutter, AIChE J., 1983, 29, 79–86.
12. Y. Hu, W. Liu, J. Sun, M. Li, X. Yang, Y. Zhang, X. Liu and M. Xu, Fuel, 2016, 167, 17–24.
13. A. Akgsornpeak, T. Witoon, T. Mungcharoen and J. Limtrakul, Chem. Eng. J., 2014, 237, 189–198.
14. E. P. Reddy and P. G. Smirniotis, J. Phys. Chem. B, 2004, 108, 7794–7800.
15. W. Gao, T. Zhou, Y. Gao, B. Louis, D. O'Hare and Q. Wang, J. Energy Chem., 2007, 26, 830–838.
16. T. Harada, F. Simeon, E. Z. Hamad and T. A. Hatton, Chem. Mater., 2015, 27, 1943–1949.
17. L. Huang, Y. Zhang, W. Gao, T. Harada, Q. Qin, Q. Zheng, T. A. Hatton and Q. Wang, Energy Technol., 2017, 5, 1328–1336.
18. Y. Wu, J. Blamey, E. J. Anthony and P. S. Fennell, Energy Fuels, 2010, 24, 2768–2776.
19. N. Phalak, N. Deshpande and L. S. Fan, Energy Fuels, 2012, 26, 3903–3909.
20. H. Chen, P. Zhang, Y. Duan and C. Zhao, Appl. Energy, 2016, 162, 390–400.
21. FactSage FTsalt Salt Database List of Systems and Phases, http://www.factsage.cn/fact/documentation/FTsalt/FTsalt_list.htm.
22. S. Li, Y. Shi, Y. Yang, Y. Zheng and N. Cai, Energy Fuels, 2013, 27, 5352–5358.
23. S. Walspurger, L. Boels, P. D. Cobden, G. D. Elzinga, W. G. Haije and R. W. van den Brink, ChemSusChem, 2008, 1, 643–650.
24. X. Wang, W. Zhou, J. B. DeLisio, G. C. Egan and M. R. Zachariah, Phys. Chem. Chem. Phys., 2017, 19, 12749–12758.
25. H. P. R. Rodríguez-Mosqueda, J. Phys. Chem. A, 2010, 114, 4535–4541.
26. H. Kim, H. D. Jang and M. Choi, Chem. Eng. J., 2015, 280, 132–137.
27. R. I. Olivares, Sol. Energy, 2012, 86, 2576–2583.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7se00502d

---