Decrease in the Adsorption Capacity of Adsorbents in the High-Temperature Carbonate Loop Process for CO₂ Capture

Abstract

In this study, the sorption capacity of limestone samples for CO₂ was investigated to determine the conditions under which they can be used in the high-temperature carbonate loop process. For the work, limestone samples from the Czech Republic were used, which contained a high proportion of CaO (more than 97 wt.%). A total of 20 cycles of calcination (950 °C) and subsequent CO₂ sorption–carbonation (650 °C) were performed for each limestone sample tested. The sorption capacity towards CO₂ in the 20th cycle was less than 10% of the value determined in the first carbonation cycle of the samples and the most significant decrease was observed between the first and second cycles. The highest sorption capacity was determined for the Branžovy sample, which captured 268 mL of CO₂/per 1 g of sorbent by chemisorption. Only 15 mL of carbon dioxide per 1 g of sorbent was bound by physisorption. However, in repeated use, the Vitošov limestone had the highest sorption capacity for CO₂. For all samples, the amount of carbon dioxide bound by physisorption was in the range of 4 to 10% of the amount bound by chemisorption. Due to sintering of the material, the BET specific surface area decreased by 95 to 96%.

1. Introduction

The carbonate loop is a post-combustion technology. Carbon dioxide, due to chemisorption, bonds to calcium oxide, forming calcium carbonate. Subsequently, calcium carbonate undergoes calcination, resulting in the release of carbon dioxide at elevated temperatures. The regenerated calcium oxide is reused in the carbonation process to bind carbon dioxide again. The process is repeated multiple times. During the carbonation process, heat is released. The heat released in the carbonator can be used for steam production. A similar amount of heat has to be supplied to the calcinator to provide the endothermic calcination reaction [1,2,3,4].

The main advantages of the high-temperature carbonate loop compared to other carbon capture technologies are lower operating cost, wide availability of natural limestone, and high carbon dioxide capture efficiency [5,6,7]. Limestone or dolomite can be used not only to capture carbon dioxide, but also other substances that are present in flue gases, such as SO₂ [8,9]. Ströhle et al. tested a limestone carbonate looping process for a 1 MW_th pilot plant and the total carbon dioxide capture was 92% [10]. Based on the study by Ströhle et al., Hilz et al. presented a publication on carbonate loop utilization in 2019 that showed a long-term project for a 20 MW_th pilot unit that was intended for industrial use in an existing 600 MW_el power plant [11]. An evaluation was carried out on a power plant burning pulverized coal with a capacity of 600 MW, which was equipped with a calcium carbonate loop. The simulation result showed that the net efficiency of the process was 33.8% with a carbon dioxide capture ratio of 94% [12].

In order for the process to be financially profitable, it is necessary for one batch of raw materials to realize, if possible, the largest number of calcination (CaCO₃ decomposition) and carbonation (chemisorption of CO₂ on CaO) work cycles. In the 1970s, Barker noted the thermal sintering problem, which limits the gradual reversibility of the reaction [13].

Wang et al. provided the carbonation reaction at the atomic level. The process is showed in Figure 1. Solid product molecules spread at the CO₂–CaCO₃ interface and connect in the existing C₂O₅²⁻ phase. Additionally, surface carbonates can combine with crystalline oxygen to form a metastable corner-bonded tetrahedron CO₄⁴⁻ and diffuse into CaO particles. The increase in temperature accelerates the formation of external CO₂-CO₃²⁻ phases and the diffusion of internal CO₃²⁻, causing the adsorption of more CO₂ molecules, and the speed of the carbonation reaction increases significantly [14].

The decrease in the sorption capacity of CaO used in the high-temperature carbonate loop process to capture CO₂ at different partial pressures was also studied [14]. The authors state that the main contribution to the deactivation of the sorbent is the sintering of the material, which occurs under both normal and elevated pressures. The activity of the deactivated sorbent can be partially increased by its reaction with water vapor [15].

Authors have also closely studied the issue of CaO deactivation in the high-temperature carbonate loop process [16]. They described potential options to suppress the deactivation process, such as doping the material with alkali metals/alkaline earth metals, reducing the pressure of CO₂, or injecting water vapor [16].

The carbonate loop generally consists of two interconnected fluidized bed reactors filled with limestone/calcium oxide [17]. The process diagram is shown in Figure 2.

Circulation in the fluidized bed equipment of limestone particles, which act as the sorption material, and mechanical stress, lead to a gradual size decrease. The decrease in particle size reduces the life time of the sorption material [18].

Fluidized bed technologies provide good contact between the gaseous phase and sorption material. In the case of the carbonate loop, limestone is the solid phase and the flue gas containing CO₂ is the gaseous phase. The carbon dioxide content in flue gas is approximately 15% [2,19].

Limestones have a very low specific surface area. Limestone porosity is affected by temperature conditions and partial CO₂ pressure. Natural limestone contains predominantly CaCO₃. Equation (2) describes the endothermic decomposition of CaCO₃ to porous CaO and gaseous CO₂. Calcium oxide, which is one of the calcination products, has a more developed porous structure and a higher specific surface area, which allows good CO₂ capture in the porous system, and subsequent reaction in the active sites of CaO is possible according to Equation (3).

The calcination is strongly influenced by the temperature in the range of 600–1000 °C, as shown in Figure 3. CaCO₃ conversion increases rapidly with increasing temperature [20].

CaCO₃ calcination (2) occurs at high temperature. The equilibrium pressure of CaCO₃ must be higher than the partial pressure of carbon dioxide in the system [21]. The equilibrium pressure is calculated according to the following equation:

$$p_{{CO}_2}=\mathrm{e}\mathrm{x}\mathrm{p}(17.74-0.00108T+0.332\ln T-\frac{\mathrm{22,020}}{T})$$

(1)

As the temperature increases, so does the equilibrium pressure. The concentrations of carbon dioxide in waste gases are around 15%. In order to achieve CO₂ enrichment, calcination must take place below a concentration of 70% CO₂. For this reason a higher calcination temperature than that of carbonation is needed. For example, to obtain a pure stream of CO₂ with a partial pressure of 1 bar, a temperature greater than 900 °C is necessary [16]. As Hu and Scaroni (1996) described [22], as the temperature increases, the equilibrium pressure increases, so the partial pressure of carbon dioxide surrounding the material should be lower for complete and rapid decarbonation. The calcination reaction will not start when the pressure of CO₂ is higher than the equilibrium pressure [22].

During calcination, calcium carbonate decomposition occurs at temperatures in the range of 850–950 °C, and thus carbon dioxide is released (Equation (2)). The regenerated calcium oxide is again utilized as a sorbent in the subsequent cycle. The released carbon dioxide is ready for further processing such as compression, transportation, and injection into the geological storage site [6,10,23,24]. During the CaO carbonation process, carbon dioxide chemisorption occurs (Equation (3)). Temperatures in the range of 550–750 °C are required for the reaction to proceed rapidly.

Calcination: CaCO₃ (s) → CaO (s) + CO₂ (g) ΔH₂₉₈ = +179.2 kJ mol⁻¹

(2)

Carbonation: CaO (s) + CO₂ (g) → CaCO₃ (s) ΔH₂₉₈ = −179.2 kJ mol⁻¹

(3)

Natural limestone also contains MgCO₃. The reversible chemical reaction for removal of CO₂ involving magnesium oxide is described by Equation (4). The thermal decomposition of magnesium carbonate produces magnesium oxide and carbon dioxide. This reaction occurs at a temperature greater than 650 °C [25].

MgO (s) + CO₂ (g) → MgCO₃ (s)

(4)

MgCO₃ (s) → MgO (s) + CO₂ (g)

(5)

Sulfur dioxide is a compound commonly present in flue gases. The presence of SO₂ has a negative effect on the sorption of carbon dioxide onto CaO. Sulfur dioxide at a temperature of around 850 °C reacts with CaO to form calcium sulfate. Calcium sulfate is a very thermally stable compound that does not decompose even at calcination temperatures as high as 1300 °C. The reaction of sulfur dioxide with calcined limestone takes place as described in Equation (6).

CaO (s) + SO₂ (g) + 1/2 O₂ (g) → CaSO₄ (s)

(6)

The reaction between SO₂ and CaCO₃, described in Equation (7), also reduces calcination efficiency.

CaCO₃ (s) + SO₂ (g) + 1/2 O₂ (g) → CaSO₄ (s) + CO₂ (g)

(7)

The reaction between CaCO₃ and gaseous SO₂ is mainly influenced by temperature. The sulfation reaction rate is the fastest in the temperature range from 900 to 1100 °C. At a temperature above 1100 °C, the reaction rate decreases [26]. Sulfation under the reaction conditions commonly applied during the calcination and carbonation processes is a non-reversible process, which occurs at a much faster rate than carbon dioxide sorption [27]. Calcium sulfate is the crystal substance that, once formed, clogs primarily the surface pores of the sorption material. Therefore, the adsorption capacity of CO₂ is reduced [28]. In combustion systems, it is possible to remove SO₂ using direct limestone sulfation as described by Equation (7).

The purpose of this work was to investigate the possibility of using Czech Republic limestones in the high-temperature carbonate loop process, as the relevant properties of Czech limestones in connection with this technology have not yet been investigated. In the case of the application of a high-temperature carbonate loop in the territory of the Czech Republic, the use of these limestones should be advantageous due to the high content of CaO and, finally, due to the low procurement costs of the input material, which would not have to be imported from abroad.

2. Materials and Methods

Several types of limestones commonly available in the Czech Republic were used for this work to determine their suitability in the carbonate loop. It was therefore necessary to specify the changes in their properties before calcination, after calcination, and after exposure to carbon dioxide.

The suitable materials chosen to be tested as adsorbents in the high-temperature carbonate loop were four commercially available limestones from different deposits in the Czech Republic. The limestone samples were selected on the basis of previous high-temperature carbon dioxide adsorption capacity tests [29]. The adsorption capacity of CaO for CO₂ depends primarily on its purity. Therefore, all limestone samples selected for testing contained a very high proportion of CaCO₃. The samples were from the following deposits in the Czech Republic:

• quarry Libotín
• quarry Branžovy
• quarry Vitošov
• quarry Tetín

The limestone samples were prepared for testing in the high-temperature loop by crushing in a jaw crusher and sieving into fraction categories. The fraction size of 0.2–0.5 mm was selected to carry out the tests. This fraction size is often used in fluidized bed high-temperature carbonation loops. All samples were calcined in a muffle furnace at a temperature of 950 °C for 12 h in an inert atmosphere. Afterwards, in order to determine the structural and surface properties of the sample, the following analyses were performed: XRF, specific surface area, pore distribution, and total pore volume. The same analyses were performed using limestone samples after repeated carbon dioxide chemisorption tests at a temperature of 650 °C.

2.1. X-ray Fluorescence Analysis

X-ray fluorescence analysis (XRF) was performed using an ARL 9400 analyzer. XRF analysis enables the determination of the elemental composition of each limestone sample. The content of CaO and MgO was determined. On the basis of the content of calcium and magnesium oxides, it was possible to estimate the content of calcium and magnesium carbonate under the assumption that all oxides were converted into carbonates. The XRF analyzer is equipped with a special program that makes it possible to calculate the carbonate content in the sample [30]. The standard error of measurement for CaO is 0.06%, for MgO is 0.04%, for Al₂O₃ is 0.02%, for SiO₂ is 0.03%, and for Fe₂O₃ is 0.02%. These analyses were performed to make it clear whether some of the tested samples do not contain significant amounts of impurities that could affect the deactivation of CaO in repeated calcination–carbonation cycles.

2.2. Specific Surface Area Measurement

The specific surface area was measured using a Coulter SA 3100 instrument (Beckman Coulter, Brea, CA, USA). The instrument works on the basis of the N₂ physical adsorption from the gaseous phase at a temperature of −196 °C. Each sample was weighed into a special sample cell and subsequently degassed under deep vacuum at a temperature of 150 °C for 240 min. The sampling cell was again weighed and placed in the instrument’s measuring port to undergo sample evacuation under deep vacuum. After evacuation, the sample cell was immersed in liquid nitrogen and was dosed with precisely calculated volumes of gaseous nitrogen. Once the adsorption equilibrium stabilized, the equilibrium pressure was measured to calculate the amount of adsorbed nitrogen. The described procedure was first applied to determine the adsorption isotherm and then the desorption isotherm. Both isotherms were determined at a relative pressure in the range of 0–0.99 at a temperature of −196 °C. A pressure decrease in the sample cell containing the sample was carried out by gradually aspirating known nitrogen volumes from the sample cell. The measurements were evaluated on the basis of the shape of the adsorption and desorption isotherms. For the pressure range corresponding to relative pressure of 0–0.3, the quantity of adsorbed nitrogen was evaluated based on the BET equation and the BET specific surface area of each sample material determined from the coefficient obtained. The total pore volume was determined on the basis of the quantity of adsorbed nitrogen at a relative pressure close to 1. The pore size distribution was calculated from the desorption isotherms based on the Barret–Joyner–Halenda model, which uses the Kelvin equation.

2.3. Carbon Dioxide Sorption Determination Using the Quantachrome ASiQ Instrument

Cyclic CO₂ sorption for each limestone sample was measured via the Quantachrome ASiQ sorption system. The sample weight was approximately 1 g. Initially, the sample cell containing the sample was heated to a temperature of 950 °C and the heating rate was 20 °C min⁻¹. This led to the calcination of the sample according to Equation (2). The sample was left at a temperature of 950 °C for 4 h, during which the carrier gas (helium) flowed throughout the sampling. The time of calcination was selected based on previous laboratory tests, where the CO₂ concentration was released by the CO₂ analyzer. Afterward, evacuation was carried out, and the sample cell was left to cool at laboratory temperature. Once the sample cell had cooled, it was then weighed and the sample weight loss after calcination was determined. The following step was carbonation at a temperature of 650 °C in a pure carbon dioxide atmosphere. The reaction that takes place during this process is described by Equation (3). After carbonation, the sample was cooled to laboratory temperature and weighed to determine the weight gain. The calcination and carbonation cycles were repeated 20 times until the constant carbonation efficiency was confirmed.

The equipment recorded the total gas volume of carbon dioxide at a temperature of 650 °C and pressures in the range of 100–800 Torr. This volume represents the total volume used for both chemisorption and physisorption. After adsorption, the saturated material was evacuated for 120 min to remove the physically trapped amount from the sample (evacuation of the sample is sufficient to break the van der Waals bonds and release carbon dioxide) and the equipment measured the amount of gas released. The molecules of CO₂ captured by chemisorption were firmly attached to the surface of the material, and to break the chemical bonds, more energy was needed in the form of an increase in temperature to 950 °C. The amount of CO₂ released by heating the sample was also recorded by the equipment.

3. Results and Discussion

In

Table 1. Selected oxide content in limestone samples via XRF analysis.

Oxide Content [wt.%]
Limestone Sample Identification	CaO	MgO	Al2O3	SiO2	Fe2O3
Libotín	97.94	0.88	0.10	0.61	0.06
Branžovy	98.41	0.70	0.13	0.25	0.21
Vitošov	97.93	0.59	0.37	0.72	0.21
Tetín	97.06	1.52	0.22	0.47	0.20

a comparison of the XRF elemental analysis is depicted for the four selected samples. The highest CaO content of 98.41 wt.% was found in the Branžovy sample. The lowest CaO content of 97.06 wt.% was found in the Tetín sample. Furthermore, the Tetín sample had the highest MgO content of 1.52 wt.%.

The samples were analyzed using a Coulter SA 3100 (i) before calcination, (ii) after calcination, and (iii) after undergoing the calcination-carbonation cycles in the Quantachrome ASiQ instrument. These analyses were performed to determine the changes in surface area and total pore volume as a result of the calcination process and repeated carbonation.

As can be seen from the results shown in

Table 2. Comparison of specific surface area and pore volume.

Limestone Sample Identification	BET Specific Surface Area [m2 g−1]			Total Pore Volume [mL g−1]
	Before	After Calcination	After 20 Cycles	Before	After Calcination	After 20 Cycles
Libotín	0.677	3.516	0.157	0.011	0.015	0.005
Branžovy	0.380	4.075	0.170	0.004	0.026	0.001
Vitošov	0.455	7.769	0.492	0.004	0.045	0.013
Tetín	0.565	2.804	0.100	0.006	0.010	0.003

, the specific surface area increased for all samples after calcination at a temperature of 950 °C. This phenomenon can most probably be attributed to the development of internal structures due to CO₂ release. In contrast, after exposure to a carbon dioxide atmosphere, a decrease in specific surface area can be observed. The decrease in surface area is also partially caused by sample sintering. Note that sample sintering is a nonreversible process.

Pore size distribution using the Coulter SA 3100 was also determined before calcination, after calcination, and after twenty carbonation–calcination cycles. All acquired results are summarized in Figure 4, Figure 5, Figure 6 and Figure 7.

Comparing sample results in Figure 4, Figure 5, Figure 6 and Figure 7, it can be seen that the most prominent structural changes were in the mesopores and macropores. The only exception that demonstrated negligible changes was the Tetín sample, which was also the sample with the lowest and highest CaO and MgO content, respectively. The biggest structural changes were observed in the Libotín and Branžovy samples. In Figure 4 it can be observed that the pore size of the Libotín sample after calcination corresponded to the pore size after 20 cycles. In Figure 4, it can be seen that CaCO₃ does not have micropores. After calcination of the sample, which contains CaO, the pores increased to 6 nm, and therefore the pore volume increased, as shown in Table 2. The thermal decomposition of CaCO₃ increased the BET specific surface area due to pore formation in two size ranges of about 20–80 nm and below 6 nm. The same fact was published by Barker in 1973 [13]. He used mercury porosity to determine the pore volume distribution. During his study, calcium oxide showed a steep rise in pore volume in the 4 nm region and a large increase in the surface area due to the formation of pores in the two size distributions of 40 nm and smaller than 4 nm. This phenomenon was also noted by Han et al. 2022 [16], who found a decrease in the sorption capacity at a high calcination temperature. This was caused by an increase in CaO nanocrystals, and as soon as the conversion reached higher values, the sintering of the material also occurred [16].

The results agree with the research of Scaltsoyiannes and Lemonidou [31], who reported that CaO sintering during each calcination step affects porosity, and, conversely, CaCO₃ sintering does not affect material deactivation. They experimentally verified that after each calcination step, the size of the CaO crystals increased, leading to a gradual loss in porosity and surface area with the number of cycles. They proposed a model that, based on measured data, accurately predicted limestone deactivation under different conditions [31]. The phenomenon of sintering during the carbonation process has also been investigated in other studies [32].

In Figure 5 it can be observed that the pore size of the Branžovy sample before calcination corresponded to the pore size after calcination; however, the pore size changed after twenty cycles. There was a significant increase in small pores and a decrease in macropores in this limestone, while at the same time the total pore volume was reduced. This may be due to sintering of the material. A study that examined the reduction in material sintering was described in a previous article [33]. Staf et al. examined the hydration of calcinates, which reduced the sintering of the material and suppressed the decrease in sorption capacity, which was stabilized after four cycles.

Repeated calcination–carbonation cycles led to the loss of small pores and an increase in the volume of macropores. These structural changes caused the loss of specific surface area, as mentioned in various research articles and confirmed by the results shown in Table 2. Like Basinas and his team [28], we found that (i) loss of specific surface area due to structural changes leads to a decrease in the calcium oxide sorption capacity for CO₂, and (ii) sorption of carbon dioxide in CaO samples depends predominantly on material porosity and the volume of micropores.

The adsorption capacity of limestone was determined using Quantachrome ASiQ equipment under carbon dioxide atmosphere. All experimental results are depicted in Figure 8.

The theoretical sorption capacity for CO₂ chemisorption was calculated from the CaCO₃ content in the individual samples. For example, for the Libotín sample, the sorption capacity was 0.43 g CO₂ per 1 g of adsorbent.

In the first adsorption cycle, all samples showed a higher sorption capacity for CO₂ than their theoretical chemisorption capacity calculated from the CaO content in each sample. It is clear that a small part of CO₂ was also sorbed by physical adsorption.

In all samples, repeated calcination–carbonation cycles caused a clear loss in sorption capacity, with the most prominent change being between the first and second cycles. Sorption capacity loss was probably caused by a decrease in pore volume during the first calcination–carbonation cycle. Antzara et al. observed the same phenomenon in their research [34]. In the literature, the most common reason for the decrease in the adsorption capacity of CaO for CO₂ is the growth in the particle volume due to the formation of CaCO₃ in the pores on the outer surface of the particles, which limits the diffusion of CO₂ into the inner space of the particles. Another stated reason is the change in particle morphology associated with pore closure [35].

During the other cycles, the third to the fifteenth, a slight sorption capacity loss was also observed. Based on the results depicted in Figure 8, it can be stated that after the fifteenth cycle no additional significant sorption capacity loss was observed. Carbon dioxide sorption capacity loss can be primarily attributed to sorption material sintering as a result of long exposure to high temperatures during the calcination process. There was a significant decrease in the CO₂ adsorption capacity after the first cycle, which may be due only to the sintering of the material.

According to data from the literature, partially deactivated CaO particles can be reactivated by the reaction of CaO with water vapor [33].

The adsorption isotherms for individual samples are shown in the following figures. The graphs show the amount of carbon dioxide used for physisorption and chemisorption that occurred in the limestones during the measurement. Figure 9 shows how much carbon dioxide volume was consumed for physisorption at a temperature of 650 °C.

Figure 10 shows the volume of carbon dioxide that was used for chemisorption at a temperature of 650 °C.

4. Conclusions

In this work, the CO₂ adsorption capacity of CaO from various limestone samples was measured during 20 repeated cycles of calcination–carbonation. Experimental tests were carried out using four selected commercially available limestones from quarries in the Czech Republic. Carbon dioxide sorption capacity tests were performed using limestone particles with a fraction size of 0.2–0.5 mm. The properties and sorption capacities of the selected sorption materials were determined during repeated high-temperature carbon dioxide adsorption.

According to XRF analysis, Branžovy limestone had the highest CaO content (98.41 wt.%) and the Tetín limestone had the lowest (97.60 wt.%).

The assumption is that the adsorption of carbon dioxide in limestone takes place preferably as chemisorption at a temperature of 650 °C. After the first adsorption, all limestones showed a higher sorption capacity for CO₂ than their theoretical chemisorption capacity, corresponding to the fact that part of the carbon dioxide was bound to the material by physisorption. During the measurement of individual limestones, the physisorption measurement was also evaluated using a Quantachrome ASiQ instrument. The highest amount of carbon dioxide sorbed by physisorption 23 mL g¹ was consumed by the Vitošov sample, which contained the highest amount of Al₂O₃ and SiO₂. The amount of CO₂ consumed by physisorption ranged from 4 to 10% of the amount consumed by chemisorption for all samples.

The most noticeable loss in sorption capacity was observed between the first and second cycles. Sorption capacity loss was caused by material sintering. From the fifteenth cycle onwards, no significant sorption capacity loss was observed. The loss in carbon dioxide sorption capacity for CaO during repeated calcination–carbonation cycles was caused primarily by two factors. The first is a high carbon dioxide content, which causes fast saturation of the active layers in the CaO particles. The second is material sintering due to long exposure to high temperatures during the calcination process. The loss in sorption capacity was confirmed by the significant changes in the specific surface area and total pore volume in all sorption materials tested.

After 20 carbonation–calcination cycles, the pore volumes of the sorbents changed on average by 70%. This was probably caused by the closing of parts of the pore due to sintering, which limited the access of CO₂ to the inner space of the adsorbent particles, and therefore its sorption capacity for CO₂.

Even more significant was the drop in the BET specific surface of the tested samples, which was 95–96% after 20 carbonation–calcination cycles. This can also be attributed to the surface sintering of the adsorbent particles.

It should be noted that even a low sulfur dioxide content has a negative effect on the carbon dioxide sorption capacity of limestone. Therefore, it is recommended to perform flue gas desulfurization when limestone materials are used in a high-temperature–carbonate loop.

After the first adsorption, the highest CO₂ sorption capacity was found in the Branžovy limestone, but when using the limestones in the carbonation–calcination cycles, Vitošov limestone was found to be the limestone with the highest sorption capacity. Therefore, due to its higher sorption capacity in repeated use, Vitošov limestone would be the best of the samples tested for application in the high-temperature carbonate loop process.