Operando modeling and measurements: Powerful tools for revealing the mechanism of alkali carbonate-based sorbents for CO₂ capture in real conditions

Abstract

Alkali carbonate-based sorbents (ACSs), including Na₂CO₃- and K₂CO₃-based sorbents, are promising for CO₂ capture. However, the complex sorbent components and operation conditions lead to the versatile kinetics of CO₂ sorption on these sorbents. This paper proposed that operando modeling and measurements are powerful tools to understand the mechanism of sorbents in real operating conditions, facilitating the sorbent development, reactor design, and operation parameter optimization. It reviewed the theoretical simulation achievements during the development of ACSs. It elucidated the findings obtained by utilizing density functional theory (DFT) calculations, ab initio molecular dynamics (AIMD) simulations, and classical molecular dynamics (CMD) simulations as well. The hygroscopicity of sorbent and the humidity of gas flow are crucial to shifting the carbonation reaction from the gas−solid mode to the gas−liquid mode, boosting the kinetics. Moreover, it briefly introduced a machine learning (ML) approach as a promising method to aid sorbent design. Furthermore, it demonstrated a conceptual compact operando measurement system in order to understand the behavior of ACSs in the real operation process. The proposed measurement system includes a micro fluidized-bed (MFB) reactor for kinetic analysis, a multi-camera sub-system for 3D particle movement tracking, and a combined Raman and IR sub-system for solid/gas components and temperature monitoring. It is believed that this system is useful to evaluate the real-time sorbent performance, validating the theoretical prediction and promoting the industrial scale-up of ACSs for CO₂ capture.

Graphical abstract

Keywords

CO₂ capture / carbonation / theoretical modeling / operando techniques / reaction visualization

1 Introduction

To avoid the serious effects of global warming on both human society and the ecosystem, CO₂ reduction has become a consensus worldwide. CO₂ capture is a vital to achieving this goal, especially in the energy section. Toward the end of the 20th century, a supported potassium carbonate (K₂CO₃) solid sorbent was first proposed to successfully recovery carbon dioxide (CO₂) from flue gases, demonstrating a thermal-swing approach for CO₂ capture [1,2]. CO₂ is chemically absorbed by K₂CO₃, which is dispersedly attached on activated carbon, following the reaction of K₂CO₃ + CO₂ + H₂O = 2KHCO₃ to form potassium bicarbonate species. The entrapped CO₂ is released by the decomposition of KHCO₃ at the elevated temperature, and afterwards the solid sorbent is regenerated for the next cyclic operation. Before long, sodium carbonate (Na₂CO₃) was also proved to be another promising candidate for CO₂ capture from the humid flue gases [3,4]. The concept map for low-temperature CO₂ capture by alkali carbonate-based sorbents (ACSs), including K₂CO₃-based and Na₂CO₃-based sorbents, is illustrated in Fig.1.

ACSs are feasible for low-temperature CO₂ reduction from stationary sources, especially from fossil−fuel fired power plants because they are:

1) Convenient to retrofit the existing power plants: The nature of the reversible carbonation reaction determines that the forward step prefers a high moist atmosphere with a mild temperature, which is attainable after the wet flue gas desulfurization process [5]. In addition, the higher temperature required for the backward step, the regeneration of ACSs, is easily achieved by suction heating from steam turbine [6]. Dual fluidized-bed reactor is demonstrated to be suitable for the dry regenerable CO₂ capture process using ACSs [7,8]. Therefore, the CO₂ capture unit is individually incorporated with the existing power plants without making many changes.

2) Cost-effective and energy-efficient to operate: First, the minor retrofit to the power plants costs less than other CO₂ capture technologies. In addition, the active components, K₂CO₃ and Na₂CO₃, are abundant and cheap, as well as various common supports, for instance, alumina, silica, and activated carbon. Moreover, the heat demand for the sorbent regeneration and the pressure drop for the sorbent fluidization are less than that of the liquid scrubber technologies, such as the monoethanolamine-based method [9]. Furthermore, the efficiency penalty can be further reduced by incorporating with heat recovery facilities [10], system integration optimization [11], and the solar thermal power system [12,13].

3) Avoidable from the secondary pollution: According to the decomposition of the bicarbonate species, CO₂ and H₂O are the only two gaseous products which are both non-toxic, non-corrosive, and easy-to-collect. No secondary pollution, such as ammonia, VOC, and NO_x, is generated during the regeneration period compared to amine-based solvents or sorbents [14].

However, some drawbacks of the alkali carbonate-based dry carbonate process should be overcome to facilitate its commercial applications:

1) Non-sufficient CO₂ capacity and sorption/desorption kinetics: Owing to the poor pore structure and scarce surface exposure, non-supported carbonates are commonly not feasible to be used for CO₂ capture, despite of their high theoretical CO₂ capacity [15,16]. Porous materials are applied to decrease the size of active components (K₂CO₃ or Na₂CO₃) and increase the CO₂ capacity and sorption/desorption kinetics. However, the dispersion of active components on the supports is influenced by many factors, such as the support structure, support chemical properties, and sorbent preparation methods, resulting in different CO₂ reduction performances [17]. In general, the CO₂ capacity and sorption/desorption kinetics of ACSs are slightly worse than those of amine-based solvents or sorbents.

2) Hypersensitive to the operation parameters: Aside from the physical/chemical properties of ACSs, the sorbent performance is highly dependent on the operation parameters, which affects the real environment of the gas-solid reaction. For a dual fluidized-bed reactor system, gas inlet velocity, solid circulation rate, temperature and water vapor content are the four factors related to the CO₂ removal performance [18].

Scientists made great efforts to improve the sorbent performance experimentally in the past two decades. Four key factors were mostly regulated in the design of sorbent, including active components, supports and additives, and preparation method (Fig.2(a)). Crystal structure and dosage of active components were commonly investigated. Zhao et al. reported that CO₂ capture on hexagonal K₂CO₃ was more kinetically favorable than that on monoclinic K₂CO₃ under the same carbonation condition [19]. Ryu et al. manufactured different alumina-supported sorbents with various loadings of K₂CO₃ and found the CO₂ capacity of PAl40 was around 124.5 mgCO₂/g sorbent, which is much higher than that of other sorbents with lower K₂CO₃ loadings [20]. The effects of types, pore structures, and the surface area of support were extensively studied as well [21‒23]. Bararpour et al. investigated the effect of textual properties, including surface area, pore size, and pore volume, on the capturing performance of mesoporous alumina supported K₂CO₃, and concluded that the larger pore size and pore volume facilitate K₂CO₃ penetration into the pores of supports during impregnation, providing more active sites for carbonation reactions and thereby achieving a better CO₂ capturing performance [21]. Khuong et al. highlighted that the coexistence of micropores and macropores in activated carbons contributed to great CO₂ sorption on K₂CO₃-activated samples by forming solutions for carbon sequestration [22]. Kazemi et al. successfully improved the CO₂ capacity of a γ-Al₂O₃ supported K₂CO₃ sorbent by increasing its specific surface area, reaching 131.6 mgCO₂/g sorbent [23]. Additives were also investigated for the specific aims, like enhancing the kinetics of carbonation [24,25], improving the regeneration kinetics, and decreasing the regeneration heat consumption [26]. The wet impregnation method [27,28], the spray-dryer method [29], and even direct mixing [17] were studied in the ACSs preparation. Some theoretical modeling work were done as well to facilitate sorbent design from the atomic level by rationally simplifying the sorbent structure and the reaction conditions. The density functional theory (DFT) was first introduced to study the crystal structures of both alkali carbonate and bicarbonate species, including Li₂CO₃/LiHCO₃, Na₂CO₃/NaHCO₃, and K₂CO₃/KHCO₃. The thermodynamic properties of each crystal were calculated and the CO₂ capture by the thermal swing mode using Na- or K- carbonate/bicarbonate pair was predicted to be feasible under the post-combustion condition [30,31]. Potassium carbonate sesquihydrate (K₂CO₃·1.5H₂O) was investigated under the same computational condition and it was proved to be more energy-efficient using K₂CO₃·1.5H₂O to capture CO₂ than that of K₂CO₃, provided that the operation conditions are carefully maintained to reduce the dehydration of K₂CO₃·1.5H₂O [32]. Besides the CO₂ sorption/desorption capacity and reaction rate, SO₂ and abrasion resistance are the other two key assessment criteria to evaluate sorbent (Fig.2(a)) [24,26,29,33]. Moreover, unraveling the reaction mechanism is also important to the sorbent development. At present, the intrinsic reaction mechanism is mainly validated by offline characterization methods, such as X-ray diffraction (XRD), N₂ adsorption/desorption curve, Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), and electron microscopy techniques (Fig.2(b)) [28]. However, the ex-situ characterization results may not reveal the real species, especially the intermediates, owing to the inconsistency between the environment (such as temperature, pressure and the relative humidity) of the sorbent performance tests and that of the characterization process. For example, Luo et al. indicated that K₂CO₃·1.5H₂O was the intermediate during K₂CO₃ carbonation based on their XRD results [34] while Jayakumar et al. suggested that K₂CO₃ hydration and carbonation occurred in parallel with no direct conversion from K₂CO₃·1.5H₂O to KHCO₃ via thermo-gravimetric analysis (TGA) [35]. This inconsistency of the role of K₂CO₃·1.5H₂O may result from improper preservation before the ex-situ XRD test. Furthermore, the intrinsic kinetics of sorbents are commonly investigated in a particle-motionless reaction cell like thermal gravimetric analyzer, ignoring the effect of particle motion. However, the sorbent performance was proved to be notably influenced by flow patterns in a fluidized reactor, indicating that the reaction mechanism is dependent on momentum/heat/mass transfer [36] (Fig.2(b)). Currently, intrusive pressure sensor is applied to reflect the average flow field in the reactor [37], nevertheless the accurate particle behaviors are hard to analysis. DFT modeling were used to reveal the reaction mechanism as well. The surface carbonation steps on different facets of K₂CO₃ were simulated by DFT calculations [38,39]. The results indicate that at least two surficial carbonation pathways exist, depending on the surface textural structure [39].

By far, neither the experimental characterization nor the theoretical simulations can unfold what exactly happen during the ACSs working period. In this perspective paper, multi-scale theoretical modeling and a compact operando measurement system are introduced to understand the operando behaviors on the sorbent chemically and physically, contributing to the design of sorbents and reactors.

2 Multi-scale theoretical modeling

CO₂ capture using ACSs belongs to non-catalytic gas-solid reaction. The operando modeling is very challengeable. First, the microstructure of solid sorbents is complex. Due to the heavy loading of alkali carbonates, the surface-active sites are diverse and the surface textual are complicated. Some active components assume to form clusters rather than be atomically dispersed on the supports [28]. Moreover, the sorbent structure can be dynamic fluxional under an operando condition. The carbonation reaction happens at a relative low temperature (40‒70 °C for Na₂CO₃-based sorbents, and 50‒90 °C for K₂CO₃-based sorbents, respectively) with a high humidity (it is sometimes beyond a relative humidity of 90%). As a result, deliquescence and hydration, can happen, further complicating the reaction mechanism. Multi-scale theoretical simulations are a plausible understanding of a working heterogeneous catalyst system at the atomic scale [40]. Here, DFT, ab initio molecular dynamics (AIMD), classic molecular dynamics (CMD), and machine learning (ML) technologies are proposed to realize the operando modeling of CO₂ capture on ACSs.

2.1 DFT calculations

The DFT method can be used to deal with some small systems (100‒300 atoms), which can accurately predict the surficial process (step 1 in Fig.3(a)) and bulk properties (steps 2/3 in Fig.3(a)) involving charge transfer in the thermodynamic steady-state and the quasi-steady-state (e.g., transition states).

Crystal structure refinement, surface bicarbonate formation, bulk defect pair formation, and bulk defect diffusion on Na₂CO₃ were conducted for the first time via DFT approaches, indicating that the surficial carbonation reaction was facile without the obstruction from the high surface humidity and that the generation of the ionic defect pairs (I_H⁺/V_Na⁻) in Na₂CO₃ was the rate-limiting step in the kinetics of bulk carbonation of Na₂CO₃, followed by the Na⁺ cation diffusion in the NaHCO₃ layer [41,42]. Furthermore, the Li-doping effect on the enhancement of the I_H⁺/V_Na⁻ defect pairs generation was predicted by DFT calculations. The foreseen acceleration of the initial CO₂ sorption on Na₂CO₃ was experimentally validated to be over 125% [41]. The effect of the texture of support on the regeneration of bicarbonates was investigated at an atomic level as well. It was predicted that the surficial basic OH groups of ZrO₂ and TiO₂ were beneficial to the bicarbonate decomposition, causing the energy reduction of the sorbent regeneration. It was proved by TGA that the activation energy of KHCO₃ decomposition was decreased by 46.6% and 23.3% if adding 5% ZrO₂ or 5% TiO₂, respectively [26,43].

2.2 AIMD/CMD simulations

The AIMD method combines the advantages of both the first principle methods and the CMD methods. Finite-temperature dynamical trajectories were generated by using forces computed “on the fly” from electronic structure calculations, and hence, the subsequent molecular dynamics were much more “realistic” [44]. The AIMD method was first introduced to solve the surface chemical bonding and bond-breaking between CO₂ and H₂O species on Na₂CO₃ (001) and Na₂CO₃ (−402) [42]. It was revealed that a high coverage of H₂O molecules on Na₂CO₃ was unable to stop CO₂ binding excess H₂O or asymmetric OH groups (Fig.3(b)). To simulate the surface chemo-physical changes involving thousands of atoms is time-consuming and unaffordable by utilizing the DFT method. Moreover, AIMD simulations can give dependable dynamical trajectories in several or tens of picosecond, facilitating the analysis of some dynamic parameters. It facilitates the dynamic modeling of reaction environment and takes insights into the fluxionality of the gas-solid interface. However, AIMD methods are barely applied in the development of ACSs by far.

The CMD method based on molecular mechanics force fields was performed to demonstrate the different hydrophilicity and deliquescence between Na₂CO₃ and K₂CO₃ (Fig.3(c))). The simulation duration is longer than 500 picoseconds, which is enough to tell the discrepancy between Na₂CO₃ and K₂CO₃. Although Na₂CO₃ deformed earlier than K₂CO₃, Na⁺ cations did not separate from the main cluster before 200 picoseconds. K₂CO₃ maintained a relatively regular crystal structure for about 50 picoseconds and then the crystal structure collapsed rapidly. At 500 picoseconds, most K⁺ cations and {\text{CO}}_3^{2-}anions are separately distributed around by H₂O molecules. Based on the different deliquescence of Na₂CO₃ and K₂CO₃ predicted by CMD, a strategy was proposed to increase the CO₂ sorption of ACSs by tuning the hydrophilicity of sorbent. The experimental results confirmed that Na₂CO₃ mixed with w(K₂CO₃) = 1.3% (w is a mass fraction) could absorb 50% more CO₂ than pure Na₂CO₃, which was less hydrophilic [45]. Similarly, Gayan and Vicki found that the adsorbed water could enhance surface reactivity by providing a medium for ionic dissociation [46]. CMD methods has been used to solve many other problems, such as, particle sintering [47], crystal thermal decomposition [48], and molecular diffusion in pore channels [49].

For the complicated sorbent system, which may involve active components, additives, supports and gas molecules, it is preferred to construct the more realistic particle model (beyond thousands of atoms) and perform the dynamic simulation under reasonable operation conditions (e.g., temperature and pressure) via AIMD and CMD methods.

2.3 ML approaches

As illustrated in Fig.3(d), high throughput screening (HTS) with ML approaches can contribute to figuring out candidate materials for CO₂ capture using ACSs. The accuracy of the ML algorithm relies heavily on the selection of gross descriptors. Here, the results from the theoretical simulations (DFT, AIMD and CMD) are preferred, considering the fact that uniform settings can be achieved easily on simulation while the experiments may induce a lot of anthropogenic errors. Due to the fact that the massive data are needed for ML approaches, the descriptors should be attained time-efficiently. HTS is widely implemented in the fields of battery electrolytes [50], membranes [51], and catalysts [52]. However, it is still not induced into the development of ACSs. It is prospected that ML approaches can provide help in quickly finding the optimal materials and the optimal operation parameters.

3 Compact operando measurement system

The term “operando” was coined by Miguel A. Bañares, which means “operating” in Latin. Operando methods require the data from the real working condition, including temperature, pressure, and atmosphere. Besides, the data should be obtained inline or real-time. Hence, all operando methods belong to in situ methods. The most commonly used operando method is operando spectroscopy, mainly composed of an infrared spectroscopy (IR), a Raman spectroscopy, and an X-ray absorption spectroscopy. An in situ environmental transmission electron microscopy and an in situ environmental scanning electron microscopy are included in operando methods though the operating pressure is currently lower than the atmospheric pressure. Operando characterization have been successfully applied for electrocatalysis [53,54], photocatalysis [55], thermocatalysis [56], and gas-solid interfacial adsorption [57,58]. Nevertheless, the reaction cell is designed for non-motion particles, gas, or liquids. Therefore, the existing operando methods are not suitable for gas-solid reactions in fluidized reactors. A compact operando measurement system is proposed to reveal the real kinetics of CO₂ sorption on ACSs (Fig.4).

3.1 Micro fluidized bed (MFB)

An MFB reactor is often less than 20 mm in diameter and filled with only a few to a few hundred milligrams of sample, ranging from several micrometers to several hundred micrometers in diameter. The operating gas velocity is often between 3 times to 7 times of the critical fluidization wind speed, providing an ideal mode of plug flow and reducing the effects of gas backmixing and diffusion. Thanks to the slight temperature difference between samples, the minimization of the gas backmixing, and the miniaturization of the reaction area, MFB is an ideal differential chemical reactor. The kinetics of redox performance on oxygen carrier in chemical looping combustion [59,60], and the kinetics of carbonation on calcium carbonate [61] and calcium oxide [62] are successfully obtained by MFB reactors. Two technical routes are alternatively used to determine the kinetics, including an online process mass spectrometer set at the gas outlet of the MFB reactor [61] or a TGA holding the entire MFB reactor to monitor the weight change [62]. The later technical route (MFB-TGA) is adopted in the proposed compact operando measurement prototype system for intrinsic kinetics analysis (Fig.4).

3.2 Multi-camera 3D imaging

Although the ideal plug flow should be kept during the experiment, the undesirable operation parameters (e.g. gas feed rate, gas composition, operating temperature) and the varying sorbent properties (e.g. particle components, size and number, moisture content, particle sphericity factor, abrasion rate) can alter the flow pattern, inducing diverse CO₂ sorption performance. A 3D particle tracking velocimetry by defocused imaging technique [63] is proposed to online monitor particle fluidization behavior (Fig.4). Under the backlighting condition (Light source 1), the illuminated particles are imaged on two charge coupled device (CCD)-cameras (Camera 1 and Camera 2). The optical paths, which are adjusted by the light splitter, from particles to the two CCDs are unequal. Hence, the extent of defocus is dichotomous, giving rise to the determination of the particle positions. The particle velocity can be calculated from two sequential frames by cross-correlation algorithm. Besides, multi-camera tomography, can be alternatively used to obtain 3D particle tracking velocimetry as well. The particle concentration can be adjustable from dilute to dense by using particle tracking velocimetry or particle imaging velocimetry.

3.3 Operando infrared/Raman measurements

Operando spectrum technologies, including IR and Raman, are extensively implemented in food and medicine industry for process analysis. Direct spectrum detection for solids in flow is difficult since particles may not run through the detection area (the focused light spot), causing a bad signal intensity or a poor resolution. In the prototype, IR source (Light source 2 in Fig.4) is used as the backlighting for the IR Camera or IR spectrometer. Due to the strong infrared absorption effect of water and CO₂, the signals received on the IR CCD are diminished and the concentrations of gases are acquired by the inverse algorithm. Moreover, the spontaneous far IR radiation of solid sorbents and gas components (wavelength: 8‒10 μm) can be recorded by the IR camera and used to calculate temperature theoretically. However, the weak signal, the stray light, and the relatively poor photon absorption efficiency and the small number of pixels of the far IR CCD or the complementary metal oxide semiconductor (CMOS) may affect the results. Light source 3 at 785 nm is used to stimulate Raman scattering from the solid sorbents and the Raman signals are analyzed by the Raman spectrometer to monitor the generation of the solid products and the consumption of the solid reactants. The time resolution of the operando Raman measurement is supposed to be worse than that of the other measurements, because of the weak Raman effect. However, some enhancement means, including surface enhanced Raman scattering, stimulated Raman scattering, wide-angle Raman scattering collection, and optical path enlargement can be applied to improve the time resolution to seconds [64]. In virtue of the combined Raman and IR sub-system, the new solid products (bicarbonate or carbonate hydrate) and the gas composition (mainly H₂O concentration) are monitored simultaneously. The mechanism of carbonation and hydration can be deduced, and these values can be validated by the weight fluctuation obtained by the MFB sub-system as well. By adjusting the flow pattern, the sorbent performance under the various operating condition is monitored. Moreover, the 3D particle imaging sub-system can capture some characteristic flow area, and the detailed local reaction intensity can be calculated by the combined Raman and IR sub-system. Hence, some hints for reactor design and operation condition optimization are obtained, facilitating the industrial scale-up of ACSs for CO₂ capture.

4 Conclusions and perspectives

CO₂ capture using dry regenerable ACSs are competitive and promising among various CO₂ reduction technologies. In this perspective, this paper first briefly introduced the challenges of developing more efficient ACSs: ① solid ACSs are component-complicated and sensitive to the operation parameters, resulting in massive factors to be considered in the development of sorbent, ② current techniques, including offline characterizations and particle-motionless reactor validation, are neither sufficient to reveal the intrinsic reaction mechanism nor to decouple the kinetics with the fluidization. Then, it proposed multi-scale theoretical simulations to understand the nature of reaction at the micro-level. DFT methods were demonstrated to reveal both the surface and bulk chemical behavior assuming that the sorbent model was simple. AIMD and CMD methods could be used to investigate the dynamics of sorbent in a realistic environment. ML approaches, involving DFT/AIMD/CMD results, were illustrated to facilitate material screening and ACSs development. Finally, it proposed a compact operando measurement system as well. The MFB-TGA technique allows the kinetic analysis to decouple from fluidization effects. 3D imaging method assists the understanding of the fluidization effects on chemical kinetics performance. Operando IR and Raman tools can reveal the gas composition, solid components and temperature, simultaneously. With the aid of these new research tools, it is believed that the complicated phenomena, for instance, the multi component synergistic effects, the steric-hinerance effects, and the effects of the intermediate products, the dynamic interface structure between sorbents and atmosphere can be systematically studied, thereby some inconsistency in literature can be well discussed, more convinced results can be concluded, and more efficient sorbents can be designed and manufactured. Furthermore, new ACSs are ready to be produced not only for CO₂ capture from flue gas, but direct from air. Combined Operando modeling and measurements are also very useful to investigate other heterogeneous reactions which are highly sensitive to atmosphere and flow conditions, such as the surface chemistry of environmental interfaces [46].

References

[1] Hirano S, Shigemoto N, Yamada S. . Cyclic fixed-bed operations over K₂CO₃-on-carbon for the recovery of carbon dioxide under moist conditions. Bulletin of the Chemical Society of Japan, 1995, 68(3): 1030–1035

[2] Hayashi H, Taniuchi J, Furuyashiki N. . Efficient recovery of carbon dioxide from flue gases of coal-fired power plants by cyclic fixed-bed operations over K₂CO₃-on-carbon. Industrial & Engineering Chemistry Research, 1998, 37(1): 185–191

[3] Liang Y, Harrison D P, Gupta R P. . Carbon dioxide capture using dry sodium-based sorbents. Energy & Fuels, 2004, 18(2): 569–575

[4] Green D A, Turk B S, Gupta R P. . Capture of carbon dioxide from flue gas using solid regenerable sorbents. International Journal of Environmental Technology and Management, 2004, 4(1/2): 53–67

[5] Zhao C, Chen X, Anthony E J. . Capturing CO₂ in flue gas from fossil fuel-fired power plants using dry regenerable alkali metal-based sorbent. Progress in Energy and Combustion Science, 2013, 39(6): 515–534

[6] Wu Y, Chen X, Ma J. . System integration for coal-fired power plant with post combustion CO₂ capture: Comparative study for different solid dry sorbents. Fuel, 2020, 280: 118561

[7] Ju Y, Lee C H. Dynamic modeling of a dual fluidized-bed system with the circulation of dry sorbent for CO₂ capture. Applied Energy, 2019, 241: 640–651

[8] Ma J, Zhong J, Bao X. . Continuous CO₂ capture performance of K₂CO₃/Al₂O₃ sorbents in a novel two-stage integrated bubbling-transport fluidized reactor. Chemical Engineering Journal, 2021, 404: 126465

[9] Luis P. Use of monoethanolamine (MEA) for CO₂ capture in a global scenario: Consequences and alternatives. Desalination, 2016, 380: 93–99

[10] Xie W, Chen X, Ma J. . Energy analyses and process integration of coal-fired power plant with CO₂ capture using sodium-based dry sorbents. Applied Energy, 2019, 252: 113434

[11] Wu Y, Chen X, Ma J. . System integration optimization for coal-fired power plant with CO₂ capture by Na₂CO₃ dry sorbents. Energy, 2020, 211: 118554

[12] Bonaventura D, Chacartegui R, Valverde J M. . Carbon capture and utilization for sodium bicarbonate production assisted by solar thermal power. Energy Conversion and Management, 2017, 149: 860–874

[13] Bonaventura D, Chacartegui R, Valverde J M. . Dry carbonate process for CO₂ capture and storage: Integration with solar thermal power. Renewable & Sustainable Energy Reviews, 2018, 82: 1796–1812

[14] Reynolds A J, Verheyen T V, Adeloju S B. . Towards commercial scale postcombustion capture of CO₂ with monoethanolamine solvent: Key considerations for solvent management and environmental impacts. Environmental Science & Technology, 2012, 46(7): 3643–3654

[15] Zhao C, Chen X, Zhao C. . Carbonation and hydration characteristics of dry potassium-based sorbents for CO₂ capture. Energy & Fuels, 2009, 23(3): 1766–1769

[16] Dong W, Chen X, Wu Y. . Carbonation characteristics of dry sodium-based sorbents for CO₂ capture. Energy & Fuels, 2012, 26(9): 6040–6046

[17] Bararpour S T, Karami D, Mahinpey N. Post-combustion CO₂ capture using supported K₂CO₃: Comparing physical mixing and incipient wetness impregnation preparation methods. Chemical Engineering Research & Design, 2018, 137: 319–328

[18] Chaiwang S C, Sema T. . Statistical experimental design for carbon dioxide capture in circulating fluidized bed using computational fluid dynamics simulation: Effect of operating parameters. Journal of Applied Science and Engineering, 2020, 23(2): 303–317

[19] Zhao C, Chen X, Zhao C. Effect of crystal structure on CO₂ capture characteristics of dry potassium-based sorbents. Chemosphere, 2009, 75(10): 1401–1404

[20] Ryu D Y, Jo S, Kim T Y. . CO₂ sorption and regeneration properties of K₂CO₃/Al₂O₃-based sorbent at high pressure and moderate temperature. Applied Sciences (Basel, Switzerland), 2022, 12(6): 2989

[21] Bararpour S T, Karami D, Mahinpey N. Utilization of mesoporous alumina-based supports synthesized by a surfactant-assisted technique for post-combustion CO₂ capture. Journal of Environmental Chemical Engineering, 2021, 9(4): 105661

[22] Khuong D A, Trinh K T, Nakaoka Y. . The investigation of activated carbon by K₂CO₃ activation: Micropores-and macropores-dominated structure. Chemosphere, 2022, 299: 134365

[23] Kazemi H, Shahhosseini S, Bazyari A. . A study on the effects of textural properties of γ-Al₂O₃ support on CO₂ capture capacity of Na₂CO₃. Process Safety and Environmental Protection, 2020, 138: 176–185

[24] Dong W, Chen X, Yu F. . Na₂CO₃/MgO/Al₂O₃ solid sorbents for low-temperature CO₂ capture. Energy & Fuels, 2015, 29(2): 968–973

[25] Chen X, Dong W, Yu F. CO₂ capture using dry TiO₂-doped Na₂CO₃/Al₂O₃ sorbents in a fluidized-bed reactor. Journal of Southeast University (Natural Science Edition), 2015, 31(2): 200–225

[26] Wu Y, Cai T, Zhao W. . First-principles and experimental studies of [ZrO(OH)]⁺ or ZrO(OH)₂ for enhancing CO₂ desorption kinetics–imperative for significant reduction of CO₂ capture energy consumption. Journal of Materials Chemistry. A, Materials for Energy and Sustainability, 2018, 6(36): 17671–17681

[27] Nasiman T, Kanoh H. CO₂ capture by a K₂CO₃–carbon composite under moist conditions. Industrial & Engineering Chemistry Research, 2020, 59(8): 3405–3412

[28] Cai T, Chen X, Zhong J. . Understanding the morphology of supported Na₂CO₃/γ-AlOOH solid sorbent and its CO₂ sorption performance. Chemical Engineering Journal, 2020, 395: 124139

[29] Kim K, Yang S, Lee J B. . Analysis of K₂CO₃/Al₂O₃ CO₂ sorbent tested with coal-fired power plant flue gas: Effect of SO_x. International Journal of Greenhouse Gas Control, 2012, 9: 347–354

[30] Duan Y, Zhang B, Sorescu D C. . CO₂ capture properties of M–C–O–H (M=Li, Na, K) systems: A combined density functional theory and lattice phonon dynamics study. Journal of Solid State Chemistry, 2011, 184(2): 304–311

[31] Duan Y, Zhang B, Sorescu D C. . Density functional theory studies on the electronic, structural, phonon dynamical and thermo-stability properties of bicarbonates MHCO₃, M = Li, Na, K. Journal of Physics Condensed Matter, 2012, 24(32): 325501–325516

[32] Duan Y, Luebke D R, Pennline H W. . Ab initio thermodynamic study of the CO₂ capture properties of potassium carbonate sesquihydrate, K₂CO₃·1.5H₂O. Journal of Physical Chemistry C, 2012, 116(27): 14461–14470

[33] Liu W, Wu Y, Cai T. . A molding method of Na₂CO₃/Al₂O₃ sorbents with high sphericity and low roughness for enhanced attrition resistance in CO₂ sorption/desorption process via extrusion-spheronization method. Powder Technology, 2020, 366: 520–526

[34] Luo H, Chioyama H, Thürmer S. . Kinetics and structural changes in CO₂ capture of K₂CO₃ under a moist condition. Energy & Fuels, 2015, 29(7): 4472–4478

[35] Jayakumar A, Gomez A, Mahinpey N. Post-combustion CO₂ capture using solid K₂CO₃: Discovering the carbonation reaction mechanism. Applied Energy, 2016, 179: 531–543

[36] Jaiboon O, Chalermsinsuwan B, Mekasut L. . Effect of flow patterns/regimes on CO₂ capture using K₂CO₃ solid sorbent in fluidized bed/circulating fluidized bed. Chemical Engineering Journal, 2013, 219: 262–272

[37] Nimvari M I, Zarghami R, Rashtchian D. Experimental investigation of bubble behavior in gas-solid fluidized bed. Advanced Powder Technology, 2020, 31(7): 2680–2688

[38] Gao H, Pishney S, Janik M J. First principles study on the adsorption of CO₂ and H₂O on the K₂CO₃(001) surface. Surface Science, 2013, 609: 140–146

[39] Liu H, Qin Q, Zhang R. . Insights into the mechanism of the capture of CO₂ by K₂CO₃ sorbent: A DFT study. Physical Chemistry Chemical Physics, 2017, 19(35): 24357–24368

[40] Shi X, Lin X, Luo R. . Dynamics of heterogeneous catalytic processes at operando conditions. JACS Au, 2021, 1(12): 2100–2120

[41] Cai T, Chen X, Johnson J K. . Understanding and improving the kinetics of bulk carbonation on sodium carbonate. Journal of Physical Chemistry C, 2020, 124(42): 23106–23115

[42] Cai T, Johnson J K, Wu Y. . Toward understanding the kinetics of CO₂ capture on sodium carbonate. ACS Applied Materials & Interfaces, 2019, 11(9): 9033–9041

[43] Zhao W, Wu Y, Cai T. . Density functional theory and reactive dynamics study of catalytic performance of TiO₂ on CO₂ desorption process with KHCO₃/TiO₂/Al₂O₃ sorbent. Molecular Catalysis, 2017, 439143–439154

[44] Iftimie R, Minary P, Tuckerman M E. Ab initio molecular dynamics: Concepts, recent developments, and future trends. Proceedings of the National Academy of Sciences of the United States of America, 2005, 102(19): 6654–6659

[45] Cai T, Chen X, Tang H. . Unraveling the disparity of CO₂ sorption on alkali carbonates under high humidity. Journal of CO₂ Utilization, 2021, 53: 101737

[46] Rubasinghege G, Grassian V H. Role(s) of adsorbed water in the surface chemistry of environmental interfaces. Chemical Communications (Cambridge), 2013, 49(30): 3071–3094

[47] Zhang L, Lu Y, Rostam-Abadi M. Sintering of calcium oxide (CaO) during CO₂ chemisorption: A reactive molecular dynamics study. Physical Chemistry Chemical Physics, 2012, 14(48): 16633–16643

[48] Lan G, Li J, Zhang G. . Thermal decomposition mechanism study of 3-nitro-1,2,4-triazol-5-one (NTO): Combined TG-FTIR-MS techniques and ReaxFF reactive molecular dynamics simulations. Fuel, 2021, 295: 120655

[49] Yu K, Tang H, Cai T. . Mechanism of kaolinite’s influence on sodium release characteristics of high-sodium coal under oxy-steam combustion conditions. Fuel, 2021, 290: 119812

[50] Borodin O, Olguin M, Spear C E. . Towards high throughput screening of electrochemical stability of battery electrolytes. Nanotechnology, 2015, 26(35): 354003

[51] Wang M, Xu Q, Tang H. . Machine learning-enabled prediction and high-throughput screening of polymer membranes for pervaporation separation. ACS Applied Materials & Interfaces, 2022, 14(6): 8427–8436

[52] Lv X, Wei W, Huang B. . High-throughput screening of synergistic transition metal dual-atom catalysts for efficient nitrogen fixation. Nano Letters, 2021, 21(4): 1871–1878

[53] Li J, Gong J. Operando characterization techniques for electrocatalysis. Energy & Environmental Science, 2020, 13(11): 3748–3779

[54] Yang Y, Xiong Y, Zeng R. . Operando methods in electrocatalysis. ACS Catalysis, 2021, 11(3): 1136–1178

[55] Dong X, Cui Z, Sun Y. . Humidity-independent photocatalytic toluene mineralization benefits from the utilization of edge hydroxyls in layered double hydroxides (LDHs): A combined operando and theoretical investigation. ACS Catalysis, 2021, 11(13): 8132–8139

[56] Chakrabarti A, Ford M E, Gregory D. . A decade+ of operando spectroscopy studies. Catalysis Today, 2017, 283: 27–53

[57] Peltzer D, Múnera J, Cornaglia L. The effect of the Li:Na molar ratio on the structural and sorption properties of mixed zirconates for CO₂ capture at high temperature. Journal of Environmental Chemical Engineering, 2019, 7(2): 102927

[58] Braglia L, Fracchia M, Ghigna P. . Understanding solid-gas reaction mechanisms by operando soft X-ray absorption spectroscopy at ambient pressure. Journal of Physical Chemistry C, 2020, 124(26): 14202–14212

[59] Yan J, Shen T, Wang P. . Redox performance of manganese ore in a fluidized bed thermogravimetric analyzer for chemical looping combustion. Fuel, 2021, 295: 120564

[60] Liu L, Li Z, Li Z. . Fast redox kinetics of a perovskite oxygen carrier measured using micro-fluidized bed thermogravimetric analysis. Proceedings of the Combustion Institute, 2021, 38(4): 5259–5269

[61] Yu J, Zeng X, Zhang G. . Kinetics and mechanism of direct reaction between CO₂ and Ca(OH)₂ in micro fluidized bed. Environmental Science & Technology, 2013, 47(13): 7514–7520

[62] Li Y, Li Z, Wang H. . CaO carbonation kinetics determined using micro-fluidized bed thermogravimetric analysis. Fuel, 2020, 264: 116823

[63] Zhou W, Zhang Y, Chen B. . Sensitivity analysis and measurement uncertainties of a two-camera depth from defocus imaging system. Experiments in Fluids, 2021, 62(11): 224

[64] Hess C. New advances in using Raman spectroscopy for the characterization of catalysts and catalytic reactions. Chemical Society Reviews, 2021, 50(5): 3519–3564

RIGHTS & PERMISSIONS

Higher Education Press 2023