Deactivation Patterns of Potassium-Based γ-Alumina Dry Sorbents for CO₂ Capture

Abstract

Gamma-alumina (γ-Al₂O₃) is an essential support material in dry sorbents used to capture CO₂ from flue gas. This study explores the deactivation of potassium-based γ-Al₂O₃ sorbents due to by-products such as KAl(CO₃)(OH)₂ during CO₂ capture. We synthesized sorbents with K₂CO₃ loadings of 5, 10, 20, and 30 wt% and subjected them to repeated capture and regeneration cycles. The results show significant variations in the deactivation degree: the sorbent with 5 wt% K₂CO₃ exhibited a 100% deactivation rate, while the 30 wt% variant showed a markedly reduced rate of 44.6%. These findings highlight the impact of the formation of KAl(CO₃)(OH)₂ at the interface between K₂CO₃ and γ-Al₂O₃ on sorbent deactivation. An equation that can be used to predict the final CO₂ capture capacity based on the ratio of active material to support was proposed using these results.

1. Introduction

Carbon dioxide (CO₂), a principal greenhouse gas, is predominantly released into the atmosphere through the combustion of fossil fuels [1]. Its accumulation is a significant contributor to global warming, presenting a grave environmental challenge with potentially catastrophic repercussions. As such, effective strategies for CO₂ capture from flue and waste gas streams are critically needed to mitigate its impact. Traditional methods, including membrane separation, solvent capture, and molecular sieve capture, have been extensively employed for this purpose, albeit with associated drawbacks such as high operational costs and considerable energy demands [2,3,4,5,6,7]. In response to these challenges, the chemical capture of CO₂ using dry regenerable sorbents has emerged as a promising alternative. The solid sorbent-based CO₂ capture process has been studied extensively in recent years as a cost-effective and efficient CO₂ capture technology [8,9]. This approach is heralded for its potential cost-effectiveness and energy efficiency, positioning it as a compelling option for CO₂ capture [8,10,11,12,13,14,15,16,17,18,19]. The CO₂ capture capability of the sorbent, which is based on 35 wt% K₂CO₃ as the active material and uses various raw materials including gamma alumina as supports, was measured over time during long-term operation in a 10 MW scale dry CO₂ capture facility. Initially, it demonstrated a high capture capacity of 66.7 mg CO₂/g sorbent, but after 150 days, this capacity decreased to 23.2 mg CO₂/g sorbent, indicating that more than 60% of the sorbent had been deactivated [20,21,22]. This suggests that the inclusion of gamma alumina among the various sorbents could have contributed to the deactivation of K₂CO₃. These findings provide an important benchmark for evaluating the long-term operational performance and stability of sorbents in dry CO₂ capture processes. Among various materials, γ-Al₂O₃ is recognized for its critical role as a support or additive in the development of dry sorbents for CO₂ capture from flue gases [15,23,24,25,26]. Nonetheless, the application of γ-Al₂O₃-based sorbents, particularly those incorporating K₂CO₃, is impeded by issues of deactivation through the formation of undesired by-products like KAl(CO₃)(OH)₂ during CO₂ capture [10,23]. This study is dedicated to exploring the deactivation mechanism of potassium-based γ-Al₂O₃ sorbents in the context of CO₂ capture. By preparing sorbents with varied K₂CO₃ loadings (5, 10, 20, and 30 wt%) and subjecting them to repeated capture–regeneration cycles, our aim was to dissect the relationship between K₂CO₃ loading and sorbent deactivation. This investigation is intended to shed light on the chemical intricacies at play, particularly focusing on the impact of potassium presence on the surface of γ-Al₂O₃ and its role in by-product formation and consequent sorbent deactivation. Our empirical findings underscore a notable correlation between potassium loading and the reduced incidence of sorbent deactivation during CO₂ capture. This observation suggests that side reactions involving potassium are instrumental in influencing by-product formation and thereby modulating sorbent deactivation. Extrapolating from these results, we propose a model for predicting the CO₂ capture capacity of potassium-based γ-Al₂O₃ sorbents based on the ratio of active material to support. The insights gleaned from this study are pivotal for the rational design and optimization of CO₂ capture materials, heralding advancements in the efficiency and sustainability of CO₂ capture technologies.

2. Materials and Methods

The alkali metal-based sorbent used in this study was prepared by the impregnation method. A typical preparation procedure for the sorbent supported on γ-Al₂O₃ (Aldrich (Seoul, Republic of Korea), 99%) was as follows: Specific amounts of γ-Al₂O₃ were added to a solution containing the desired ratio of anhydrous potassium carbonate (K₂CO₃, Aldrich) in 25 mL of deionized water. The amounts used for each compound were as follows: 9.5 g of γ-Al₂O₃ and 0.5 g of K₂CO₃ for 5KAlI, 9.0 g of γ-Al₂O₃ and 1.0 g of K₂CO₃ for 10KAlI, 8.0 g of γ-Al₂O₃ and 2.0 g of K₂CO₃ for 20KAlI, and 7.0 g of γ-Al₂O₃ and 3.0 g of K₂CO₃ for 30KAlI. Then, the content was mixed with a magnetic stirrer for 24 h at room temperature [7,9,18,23]. After stirring, the mixture was dried in a rotary evaporator at 60 °C. The dried samples were calcined in a furnace with N₂ flow (100 mL/min) for 6 h at 400 °C. The ramping rate of the temperature was maintained at 5 °C/min [27,28]. The sorbents are denoted as follows: 5KAlI, 10KAlI, 20KAlI, and 30KAlI where K represents K₂CO₃, Al represents γ-Al₂O₃ (Aldrich), I denote the impregnation method, and 5, 10, 20, and 30 represents the K₂CO₃ loaded. The amount of alkali metal impregnated was identified by using inductively coupled plasma–atomic emission spectrometry (ICP-AES; GBC Scientific (Keysborough, Australia), Integra KL) Power X-ray diffraction (XRD; Philips (Amsterdam, The Netherlands), X’PERT) at the Korea Basic Science Institute (Daegu, Republic of Korea) was also carried out in order to confirm the structure using Cu Kα radiation. The TPD experiment was conducted in a nitrogen atmosphere, and the sample was heated from 100 °C to 500 °C at a heating rate of 1 °C/min.

CO₂ capture and regeneration processes were performed in a fixed-bed quartz reactor with a diameter of 0.75 in, which was placed in an electric furnace under atmospheric pressure. Half (0.5) a gram of the sorbent was packed into the reactor, and space velocity (SV) was maintained at 3000 h⁻¹ to minimize severe pressure drops and channeling phenomena. All volumetric gas flows were measured under standard temperature and pressure (STP) conditions. The conditions of CO₂ capture and regeneration and the composition of mixed gases are shown in

Table 1. Experimental conditions.

	CO2 Capture	Regeneration
Temperature (°C)	60	200
Pressure (atm)	1	1
Flow rate (mL/min)	60	60
Gas composition (vol%)	CO2: 1, H2O: 10, N2: balance	N2: balance

. Outlet gases from the reactor were automatically analyzed every 4 min by a thermal conductivity detector (TCD; Donam Systems Inc., Deajeon, Republic of Korea), which was equipped with Porapak Q (1/8 in. stainless).

The CO₂ capture capacity was calculated from the CO₂ breakthrough curve, which indicates the amount of CO₂ captured until the output concentration of CO₂ reaches 10 vol%, which is the same as the inlet concentration. The CO₂ capture capacity is determined by the amount of CO₂ captured per 1 g of sorbent (mg CO₂/g sorbent). The CO₂ capture capacity was calculated according to Equation (1) as follows:

$$\begin{gathered} \begin{array}{c}{\mathrm{CO}}_2\mathrm{capture}\mathrm{capacity}=(\frac{P\times V_{{\mathrm{C}\mathrm{O}}_2}}{R\times T}\times M_{{\mathrm{C}\mathrm{O}}_2}\times t, \mathrm{mg})/(g \mathit{sorbent})\\ (P—\mathrm{Pressure}\mathrm{of}{\mathrm{CO}}_2(\mathrm{Pa},\mathrm{Pascal});V_{{\mathrm{C}\mathrm{O}}_2}—\mathrm{Volume}\mathrm{of}{\mathrm{CO}}_2\left({\mathrm{m}}^3\right);R—\mathrm{Gas}\mathrm{constant}(\mathrm{J}/(\mathrm{mol}·\mathrm{K}\left)\right); \\ T—\mathrm{Temperature}(\mathrm{K},\mathrm{Kelvin});M_{{\mathrm{C}\mathrm{O}}_2}—\mathrm{Molar} \mathrm{mass} \mathrm{of} {\mathrm{C}\mathrm{O}}_2(\mathrm{g}/\mathrm{mol}); t—\mathrm{Time} (\mathrm{s}, \mathrm{seconds}))\end{array} \end{gathered}$$

(1)

3. Results

3.1. CO₂ Capture Capacity of 5KAlI, 10KAlI, 20KAlI, and 30KAlI Sorbents

Figure 1 shows the breakthrough curves of the sorbents after one to five cycles. The 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents were prepared usingK₂CO₃ loading 5, 10, 20 and 30, respectively. In the cases of the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents, the breakthrough times were approximately 0, 4, 8, and 12 min, respectively. It can be seen that the breakthrough time of one cycle increases in proportion to the amount of K₂CO₃ loaded. However, a noticeable degradation in the capture performance was observed with subsequent cycles. Specifically, the 5KAlI and 10KAlI sorbents exhibited a negligible capture capacity beyond the second cycle. Furthermore, the breakthrough time of the 20KAlI sorbent significantly decreased to 4–8 min in subsequent cycles, down from an initial 16 min. Similarly, the 30KAlI sorbent experienced a reduction in its breakthrough time to approximately 12 min from an initial 28 min, indicating a loss of more than half of its capture capacity.

Figure 2 shows the CO₂ capture capacities of the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents during the multiple tests for sorption (60 °C) and regeneration (200 °C). The CO₂ capture capacities of the sorbents were calculated from the breakthrough curves during multiple tests. The CO₂ capture capacities of the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents decreased after two cycles. In the cycle results after two cycles, the reduction decreases and the capture capacity converges. The capture capacities after one cycle were 15.8, 30.5, 60.3, and 88.2 mg CO₂/g sorbent for the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents, respectively. The final capture capacities were 0, 2.0, 24.9, and 50.4 mg CO₂/g sorbents for the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents, respectively. And you can see that the capture capacity remains constant as the cycle continues. Therefore, it can be assumed that the capture capacity after five cycles is the final capture capacity. The 30KAlI sorbent’s performance is consistent with the results reported in references [6,10].

3.2. Deactivation Degree According to K₂CO₃ Loading Amount

Figure 3a shows the amount of CO₂ captured per 1 g of sorbent during 1 cycle (I) and 5 cycles (II) as a function of the loading amount of K₂CO₃. In both cases, the total CO₂ capture capacity of the sorbent increased with the loading amount of K₂CO₃. In order to investigate the effects of the loadings of K₂CO₃ on the total CO₂ capture capacities of the sorbents in detail, the amount of CO₂ captured per 1 g of K₂CO₃ was calculated from Figure 3a. These results are shown in Figure 3b. The theoretical value of the sorbent, which was calculated from moles of K₂CO₃ involved in the sorbent, was 3.18 mg CO₂/g K₂CO₃ as shown in Figure 3b. The theoretical value of the sorbent was calculated from the number of moles of K₂CO₃ contained in the sorbent when 1 mole of K₂CO₃ captures a stoichiometric amount equivalent to 1 mole of CO₂. The initial capture capacity was similar to the theoretical value, but the final capture capacity was found to fall short of the theoretical value. In particular, the initial capture capacity and the CO₂ capture capacity of the 5KAlI and 10KAlI sorbents were confirmed to be approximately 95 to 100% of the theoretical value (3.18 mg CO₂/g K₂CO₃). However, during regeneration, the sorbent was not regenerated and the capture ability was almost lost. The initial CO₂ capture capacity of the 20KAlI and 30KAlI sorbents was confirmed to be approximately 92–95% of the theoretical value, and the final capture capacities were 38% and 57%. The area (I) shown in Figure 2a,b shows the difference between the initial capture capacity and the final capture capacity. As shown in Figure 3a, the gap in the CO₂ capture amount decreased per 1 g of the 5KAl, 10KAl, 20Kal, and 30KAl sorbents. However, if you look at the graph in Figure 3b showing the decrease in the amount of CO₂ captured per g K₂CO₃, you can see that as the loading amount of K₂CO₃ increases, the CO₂ captured per g K₂CO₃ decreases. Based on this area (I), the deactivation degree of each sorbent was calculated and is shown in

Table 2. Deactivation degree according to K₂CO₃ loading amount.

Sorbent	Initial Capture Capacity (mg CO2/g Sorbents)	Final Capture Capacity (mg CO2/g Sorbents)	Degree of Deactivation (%)
5KAlI	15.8	0	100
10KAlI	30.5	2.0	26.7
20KAlI	60.3	24.9	62
30KAlI	88.2	50.4	44.6

. The formula for calculating the degree of deactivation is Equation (2) as follows:

$$\begin{array}{l}\mathrm{Degree} \mathrm{of} \mathrm{deactivation} (\%) = \\ \left[\right(\mathrm{Initial}\mathrm{capture}\mathrm{capacity}-\mathrm{Final}\mathrm{capture}\mathrm{capacity})/\mathrm{Initial}\mathrm{capture}\mathrm{capacity}]\times 100\end{array}$$

(2)

An analysis of Table 2 reveals that the deactivation degrees of the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents were 100%, 76.6%, 62.0%, and 44.6%, respectively. It can be seen that the degree of deactivation decreases depending on the K₂CO₃ loading amount, and the amount that it decreases also appears to gradually decrease. This observation hints at the layered deposition of K₂CO₃ on the γ-Al₂O₃ support during sorbent manufacturing. Moreover, it implies potential surface reactions between alumina and K₂CO₃, contributing to the regeneration process. Consequently, achieving complete regeneration at a temperature of 200 °C may be challenging due to the intricate nature of these surface interactions.

3.3. Structure Identification of the 5–30KAlI Sorbents

Figure 4 shows XRD patterns of structural changes in fresh 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents and the sorbents after being regenerated at 200 °C. In the case of the sorbents before the reaction, only peaks in K₂CO₃ (JCPDS No. 16-0820) and γ-Al₂O₃ (JCPDS No. 10-0425) were observed, and it can be seen that the peaks become sharper as the loading amount of K₂CO₃ increases. In the case of the sorbents that were regenerated at 200 °C, KAl(CO₃)(OH)₂ (JCPDS No. 15-3303) peaks can be seen in addition to the K₂CO₃ and γ-Al₂O₃ peals [24,29]. Also, in the case of K₂CO₃∙1.5H₂O (JCPDS No. 11-0655) and K₄H₂(CO₃)₃∙1.5H₂O, K₂CO₃ and water vapor will react and act like active species [28,30,31]. When CO₂ is captured, a phase change in K₂CO₃ and γ-Al₂O₃ occurs in the form of KAl(CO₃)(OH)₂. These results also imply that KAl(CO₃)(OH)₂ is not converted to its original active substance, K₂CO₃. Therefore, as the cycle test continues, the CO₂ capture capacity of the sorbents can be seen to decrease. These results suggest that by-products such as KAl(CO₃)(OH)₂ are influenced by the structure of alumina and that the regeneration ability of potassium-based alumina sorbents is directly related to the by-products.

3.4. Physical Properities of the 5–30KAlI Sorbents According to BET

The study rigorously examines the impact of varying K₂CO₃ loadings on the physical properties of K₂CO₃/γ-Al₂O₃ sorbents, as depicted in

Table 3. BET results of γ-Al₂O_3,5–30KAl sorbents and K₂CO₃.

	Surface Area (m2/g)	Pore Volume (cc/g)	Pore Diameter (nm)
γ-Al2O3	158.74	0.296	6.529
5KAlI	150.5	0.27	4.9
10KAlI	121.25	0.212	3.834
20KAlI	77.237	0.151	3.832
30KAlI	60.559	0.117	3.828
K2CO3	0.592	0.001	3.058

and Figure 5. This examination zeroes in on two crucial characteristics: the Brunauer–Emmett–Teller (BET) specific surface area and the Barrett–Joyner–Halenda (BJH) pore volume. A conspicuous trend that is discernible from Table 2 highlights a reciprocal relationship between the increase in K₂CO₃ sorbent loading and both the BET specific surface area and BJH pore volume. The initial results for the γ-Al₂O₃ support stood at 158.74 m²/g and 0.296 cc/g for BET surface area and BJH pore volume, respectively. It was observed that, with escalating K₂CO₃ loadings, these values inversely diminished. Specifically, the sorbent labeled as 5KAlI showcased the highest BET surface area and BJH pore volume at 150.5 m²/g and 0.270 cc/g, respectively. Contrary to the expected reduction in the capture capacity due to diminished surface area and pore volume, the 10KAlI sorbents showed enhanced CO₂ capture capabilities when compared to the 5KAlI samples. This paradoxical finding suggests that the concentration of K₂CO₃ plays a pivotal role in augmenting the sorbents’ capture efficiency, surpassing the effects of surface area and pore volume reductions. Moreover, the 30KAlI sorbents exhibited markedly reduced BET surface areas and BJH pore volumes, recorded at 60.56 m²/g and 0.117 cc/g, respectively. This indicates a shift in the capture mechanism towards chemical interactions at higher K₂CO₃ loadings. Further elucidation are provided by the pore size distribution analysis presented in Figure 6. A noticeable decline in porosity was observed with increasing K₂CO₃ loading, particularly impacting the medium to large pores within the size range of 3.5 to 5.0 nm. Among the analyzed variants, the 5KAlI sorbents demonstrated the most significant peak in the rate of pore volume change, underscoring their superior ability to retain larger pore volumes. This thorough analysis elucidates the complex interplay between K₂CO₃ loading and the physicochemical properties of K₂CO₃/γ-Al₂O₃ sorbents, offering insights into their intricate effects on CO₂ capture performance. Based on these results, the relationship between the BET surface area value and CO₂ capture capacity in various alumina phases will be revealed in a future paper.

3.5. Surface Morphology of the 5–30KAlI Sorbents According to SEM

Figure 6a shows an image of the fresh sorbents indicating that K₂CO₃ is uniformly distributed on the surface of γ-Al₂O₃ in a needle-like structure. As the amount of K₂CO₃ increases, these needle-like structures appear to be more densely packed. S. Toufigh Bararpour et al. (2019) [26] also showed that K₂CO₃ exhibits a needle-like structure as it is loaded. Figure 6b shows an image of the spent sorbents after use. Compared to the fresh sorbents in Figure 6a, the spent sorbents exhibit significant morphological changes. The needle-like structures of K₂CO₃ on γ-Al₂O₃ appear to be less dense and more irregular. These alterations in morphology suggest a decline in the sorbent’s reactivity and efficiency over time.

3.6. Regeneration Properties of the 5–30KAlI Sorbents

TGA and DTG were conducted on the spent 5–30KAlI sorbents after five cycles of CO₂ adsorption. These analyses were performed across a temperature spectrum of 20 to 500 °C at a ramp rate of 5 °C/min, as depicted in Figure 7. The initial reduction in weight is attributable to the desorption of H₂O. Both the TGA and DTG results identified two pronounced mass loss regions, at approximately 100–200 °C and 250–400 °C. According to prior studies, these regions are indicative of the desorption of KHCO₃ and the decomposition of KAl(CO₃)(OH)₂ [23,27]. Specifically, in the DTG curve (Figure 7b), there is an absence of peaks for the 5KAlI and 10KAlI sorbents within the 100–200 °C range, whereas the peaks for the 20KAlI and 30KAlI sorbents are prominent, with the 30KAlI peak being markedly larger than its 20KAlI counterpart. At elevated temperatures, ranging from 250 to 400 °C, all of the sorbents exhibited peaks, with the magnitude of mass loss being amplified in conjunction with increased K₂CO₃ loadings. This observation suggests the desorption of CO₂ and H₂O from the decomposition of KAl(CO₃)(OH)₂.

In this study, the results of the TPD experiments conducted using the 5–30KAlI sorbents are presented in Figure 8. These experiments facilitated a deeper understanding of the sorbent’s regeneration characteristics. The TPD curves were measured by heating at a rate of 1 °C per minute in a N₂ atmosphere to analyze the desorption characteristics of CO₂ and H₂O. This methodology enabled a precise determination of the decomposition temperatures of the KHCO₃ and KAl(CO₃)(OH)₂ structures, the desorption temperature of H₂O, and the amounts of CO₂ and H₂O desorbed at each temperature. The accuracy of the TG and DTG results was corroborated by a GC analysis, which confirmed the clear separation of the CO₂ and H₂O peaks. The data obtained from these experiments are invaluable for optimizing the regeneration process of the KAlI5–30 sorbents. The TPD results for the 5KAlI and 10KAlI sorbents displayed a single peak between 250 and 400 °C, whereas the 20KAlI and 30KAlI sorbents exhibited two peaks within the 100 to 400 °C range. The peak between 100 and 200 °C corresponds to CO₂ desorption from the decomposition of KHCO₃ (Equation (3)), and the peak between 250 and 350 °C results from the decomposition of KAl(CO₃)(OH)₂ (Equation (4)). In the case of sorbents with a lower K₂CO₃ content, such as 5KAlI and 10KAlI, the decomposition of KAl(CO₃)(OH)₂ structures leads to CO₂ desorption. For adsorbents with a higher K₂CO₃ content, like 20KAlI and 30KAlI, the decomposition of both the KHCO₃ and KAl(CO₃)(OH)₂ structures produced CO₂ peaks. Furthermore, the decomposition of the KHCO₃ and KAl(CO₃)(OH)₂ structures, as described by the following equations, resulted in H₂O peaks observed at 100 °C and between 100 °C to 200 °C and 250 to 400 °C, respectively:

2KHCO₃ → K₂CO₃+ CO₂ + H₂O

(3)

2KAl(CO₃)(OH)₂ → K₂CO₃ + Al₂O₃ + CO₂ + 2H₂O

(4)

The decomposition of KHCO₃ and KAl(CO₃)(OH)₂ occurs, respectively, at 100 °C to 200 °C and 250 °C to 400 °C. Notably, as K₂CO₃ loading increases, the CO₂ peak from KHCO₃ decomposition also increases proportionally; however, the difference in the size of the CO₂ peak from KAl(CO₃)(OH)₂ decomposition is relatively minor. This supports the observation that deactivation decreases with increasing K₂CO₃ loading at a regeneration temperature of 200 °C, as illustrated in Figure 3. Additionally, complete regeneration of γ-Al₂O₃-based sorbents is possible at temperatures above 400 °C. The desorption capacities obtained from these TPD results can be quantified as follows: the desorption capacity is 16.2, 30.5, 61.1, and 88.5 mg CO₂/g sorbent for the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents, respectively. These figures closely match the CO₂ capture capacities shown in Figure 2. The calculated desorption capacities between 250 °C to 400 °C for one cycle are 16.2, 30.5, 40.1, and 44.5 mg CO₂/g sorbent for the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents, respectively. These results are nearly identical to the difference between the initial and final capture capacities shown in Figure 3, indicating a decrease in the absorption capacity due to the non-decomposition of KAl(CO₃)(OH)₂ at 200 °C.

Table 4. Comparative analysis of weight loss and desorption capacities of CO₂ and H₂O at two temperature intervals for 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents after 5 CO₂ capture cycles below approximately 200 °C: (I) and between 250 and 400 °C: (II).

Sorbent		Weight Loss (%)	CO2 Desorption Capacity (%)	H2O Desorption Capacity (%)
5KAlI	(I)	5.21	0	5.15
	(II)	30.9	1.62	1.29
10KAlI	(I)	5.59	0	5.53
	(II)	5.21	3.05	2.51
20KAlI	(I)	6.94	2.1	4.82
	(II)	9.76	4.01	3.29
30KAlI	(I)	8.07	4.02	4.09
	(II)	7.49	4.45	3.66

presents a quantitative evaluation of the thermal stability and desorption properties of the 5KAlI, 10KAlI, 20KAlI, and 30KAlI sorbents evaluated after five CO₂ capture cycles. The data are categorized into two temperature ranges, below approximately 200 °C (I) and between 250 and 400 °C (II), with mass loss and CO₂ and H₂O desorption observed in the TGA and TPD analyses. The amount of desorption was expressed as a percentage of the amount of CO₂ and H₂O desorbed per g sorbent for easy comparison. In the lower temperature range (I), the observed mass loss was predominantly associated with the desorption of physically adsorbed water and the decomposition of KHCO₃. Conversely, in the higher temperature range (II), the desorption ratios of CO₂ to H₂O were approximately in a 2:1 molar ratio, suggesting that the mass loss was primarily due to the decomposition of KAl(CO₃)(OH)₂. Importantly, the sum of the desorption capacities for CO₂ and H₂O matches the total weight loss recorded in the TGA analysis, which underscores the accuracy of the thermogravimetric and TPD measurements

3.7. Derivation of the Predictive Model for Capture Capacity Based on K₂CO₃ Loading Amounts

Figure 9 illustrates the relationship between the initial and final CO₂ capture capacities as a function of the loading amount of K₂CO₃. Each graph represents the CO₂ capture capacities on the y-axis as related to the K₂CO₃ loading ratio on the x-axis, and the data points are explained visually through the fitting of linear and quadratic regression equations. Figure 9 (I) demonstrates that the initial CO₂ capture capacity increases linearly with the increase in the K₂CO₃ loading ratio. The formula for calculating the initial CO₂ capture capacity is expressed by Equation (5) as follows:

y = 3.18x

(5)

This linear relationship shows that the initial CO₂ capture capacity increases directly in proportion to the increase in the K₂CO₃ loading ratio, aligning well with the data points. Conversely, Figure 9 (II) depicts a nonlinear relationship for the final CO₂ capture capacity. The formula for the final CO₂ capture capacity is given in Equation (6) as follows:

y = 0.0296x² + 1.247x − 13.05

(6)

The model predicts an almost perfect accuracy for the data, showing that the final capture capacity is zero when the loading amount of K₂CO₃ is at or below 8 wt% and increases proportionally above this threshold. This formula allows for the prediction of CO₂ capture capacity based on the loading amount of K₂CO₃. To validate the model, the study measured the actual CO₂ capture capacities of sorbents loaded with varying weight percentages of K₂CO₃ (12, 25, 35 wt%), marked by red points, confirming a good fit with the regression models. Based on these data, these models can serve as crucial tools for optimizing the K₂CO₃ loading ratio and maximizing the efficiency of the CO₂ capture process.

4. Discussion

This study has delineated the complex interaction between K₂CO₃ loading and γ-Al₂O₃ supports and its effects on CO₂ capture, confirming that more K₂CO₃ loading initially enhances the sorption capacity of the tested sorbents. However, the formation of KAl(CO₃)(OH)₂ by-products during sorbent operation has been identified as a major impediment, challenging the sorbent’s long-term stability and functionality. Intriguingly, by-product formation does not increase proportionately with K₂CO₃ loading, suggesting that there are intricate chemical dynamics at play. Specifically, we observed a marked disparity in the deactivation rates between different loading levels. The 5KAlI demonstrates a deactivation rate of 100%, whereas the 30KAlI shows a significantly reduced rate of approximately 44.6%. The interface between K₂CO₃ and γ-Al₂O₃ emerges as a key factor in sorbent deactivation, with only the K₂CO₃ that is directly in contact with the γ-Al₂O₃ undergoing a transformation to KAl(CO₃)(OH)₂, while the remaining layers form KHCO₃ and retain their CO₂ absorption capacity. These findings have led to the development and validation of a predictive model for CO₂ capture capacity based on K₂CO₃ loading amounts, establishing a robust framework for future sorbent optimization strategies. Our findings are consistent with those reported for CO₂ capture performance observed using potassium-based γ-Alumina dry sorbents [6,16,23]. However, our study extends this knowledge by specifically examining the effect of varying K₂CO₃ loadings on deactivation mechanisms, providing a deeper understanding of these processes. Moving forward, further investigations are needed to refine the predictive model and explore alternative modifications to the sorbent materials that might inhibit or delay the formation of deactivating by-products. Additionally, assessing the scalability and economic viability of optimized sorbent systems in industrial applications will be crucial for advancing CO₂ capture technology.