Effect of Al₂O₃ Particle Addition on Fluidized Bed Thermochemical Heat Storage Performance of Limestone: From Instability Mitigation to Efficiency Enhancement

Abstract

This study elucidates the mechanism of fluidization instability during limestone carbonation under a 100% CO₂ atmosphere and determines the influence of Al₂O₃ fluidization aids (dosage and particle size) on exothermic performance. The experiments demonstrate that rapid CO₂ absorption in the emulsion phase, coupled with insufficient gas replenishment from the bubble phase, disrupts the balance between drag force and buoyancy, leading to localized defluidization. This instability impedes gas exchange between the bubble and emulsion phases, resulting in bubble coalescence and channeling across the bed. The fluidization instability reduces the maximum exothermic temperature and causes significant temperature heterogeneity in the bed. With repeated thermal cycles (20 cycles), the CO₂ absorption capacity of limestone diminishes (the effective conversion rate drops to 0.25), and the instability disappears. The addition of 5wt.% Al₂O₃ (particle size: 0.05–0.075 mm) stabilizes the fluidization state during carbonation, significantly homogenizing the bed temperature distribution, with maximum and average temperature differentials reduced by 63% and 89%, respectively, compared to pure limestone systems.

1. Introduction

Thermal energy storage (TES) technologies can be categorized into sensible heat storage (SHS), latent heat storage (LHS), and thermochemical heat storage (TCHS) based on their material characteristics [1,2]. Among these, TCHS has garnered significant attention due to its high energy density, achieved through reversible chemical reactions. The development of TCHS traces back to the 1970s oil crisis, when researchers proposed utilizing chemical reactions to store solar thermal energy [3]. Wentworth and Chen further established criteria for selecting TCHS reaction systems, including reversibility, temperature compatibility, high specific enthalpy, and environmental friendliness [4]. Among various TCHS systems, the CaCO₃/CaO cycle (CaL), characterized by its high operating temperature (650–1000 °C), large storage density (3.4 GJ/m³), cost-effective raw materials (natural limestone, industrial waste), and non-toxic/non-corrosive nature, is regarded as one of the most promising TES technologies for concentrated solar power (CSP) applications [5,6,7,8,9,10,11,12]. Particularly, calcium carbide slag (CCS)—an industrial byproduct that poses significant disposal challenges—has emerged as an alternative CaO source due to its calcium-rich composition, with studies demonstrating its potential to achieve carbonation conversion rates comparable to natural limestone under optimized conditions [13,14].

As illustrated in Figure 1 [15], a complete CaL cycle comprises two phases: calcination (charging) and carbonation (discharging). During the charging phase, CaCO₃ absorbs concentrated solar heat in a calciner, decomposing into CaO and CO₂, which are stored separately. For discharge, CaO and CO₂ are fed into a carbonator to release stored heat via the exothermic carbonation reaction. The resulting CaCO₃ is recycled to the solar calciner, while the heat from the carbonator is transferred to working fluids (e.g., CO₂, steam) through direct/indirect heat exchange to drive turbine generators. The core of CaL lies in the reversible calcination/carbonation reaction between CaO and CO₂, which exhibits an equilibrium temperature of 895 °C under atmospheric pure CO₂ conditions, with reaction rates tunable via a CO₂ partial pressure within 650–1000 °C [16,17]. Although initially developed for CO₂ capture, CSP-CaL systems operate under high-temperature (>850 °C) pure CO₂ environments to enhance thermoelectric efficiency [18,19,20]. Recent integration strategies include (1) closed CO₂ Brayton cycles, where unreacted CO₂ directly drives gas turbines with system efficiencies of 45–46% [21]; (2) heat exchanger-based configurations coupled with Rankine cycles or supercritical CO₂ recompression cycles [22,23]; and (3) hybrid gas–steam cycles combining Brayton and Rankine cycles [24].

Current research on CaL exothermic performance predominantly relies on thermogravimetric analyzers or fixed-bed reactors [25], while fluidized beds—with superior gas–solid mixing (uniform temperature fields or enhanced mass transfer)—are better suited for scaled TES systems [26]. However, carbonation under high-concentration CO₂ atmospheres often triggers fluidization instability. Similar defluidization phenomena have been observed in CO₂ hydrogenation-to-methane experiments by Abba et al. [27] and Kai’s group [28,29,30,31], confirming its prevalence in gas volume reduction reactions. The instability mechanism involves insufficient gas resistance in the emulsion phase due to reduced gas volume, leading to gravity–buoyancy imbalance, localized defluidization, and bubble phase disturbances [32,33]. Existing mitigation strategies fall into two categories—external field enhancement (e.g., vibration or magnetic fields), which faces scalability challenges due to complex equipment [34,35,36,37,38], and intrinsic fluidization optimization via fluidization aids (e.g., Al₂O₃) or surface modification to regulate particle behavior [39,40,41]. This study employs Al₂O₃ inert particles as fluidization aids, leveraging their mechanical anti-agglomeration effects (disrupting particle clusters via collisions) and industrial compatibility (requiring no reactor modifications).

From the dual perspectives of dynamic heat storage conversion rates and exothermic temperature fields, this work systematically investigates the effects of Al₂O₃ dosage and particle size on the fluidized exothermic performance of limestone, aiming to provide theoretical foundations for the efficient design and scale-up of CaL-CSP systems.

2. Materials and Methods

2.1. Experimental Materials

Natural limestone, sourced from a mining site in Zhengzhou, Henan Province, China, served as the calcium-based raw material. The limestone was crushed using a jaw crusher (Gilson LC-34, Shanghai, China) and sieved into three particle size ranges (0.25–0.355 mm, 0.18–0.25 mm, and 0.125–0.18 mm) via a vibrating sieve machine (Zhetai GZS-1, Zhengzhou, China). The limestone was pre-calcined at 850 °C for 30 min under a pure N₂ atmosphere in a bubbling fluidized bed reactor. The chemical composition of the calcined limestone (denoted as CaO) was analyzed using X-ray fluorescence spectroscopy (XRF), as summarized in

Table 1. Chemical Composition of Calcined Limestone (wt.%).

CaO	MgO	SiO2	Al2O3	Fe2O3	K2O	Others
95.28	1.53	2.25	0.38	0.28	0.25	0.03

.

Al₂O₃ particles (>93% purity) were employed as fluidization aids, and sieved into three size ranges (0.125–0.18 mm, 0.075–0.125 mm, 0.05–0.075 mm) using the same vibrating sieve equipment. The selection of Al₂O₃ particles as fluidization aids was carefully determined based on their unique combination of chemical stability under high-temperature carbonation conditions (600–850 °C), an optimal particle size distribution that effectively interferes with limestone particle agglomeration without causing self-aggregation, and inherent physical properties including appropriate hardness and density that collectively promote stable gas–solid fluidization dynamics by maintaining balanced interactions between the bubble and emulsion phases throughout the thermochemical cycling process.

2.2. Experimental Device

The CaCO₃/CaO cycle experiments were conducted in a custom-designed bubbling fluidized bed reactor (BFBR) system (Figure 2). The BFBR system comprised N₂/CO₂ gas cylinders, mass flow controllers, a gas mixing chamber, a quartz reactor (inner diameter: 34 mm, height: 900 mm), K-type thermocouples (±0.75%T, where T is the measured temperature in °C), a real-time temperature display, a precision electronic balance (Mettler Toledo-XS105DU, Greifensee, Switzerland, accuracy: 0.1 mg), and a data acquisition unit. The reactor’s central 200 mm zone maintained a temperature uniformity of ±5 °C via electric heating and a PID temperature controller.

For each test, 16g of CaO was weighed and loaded into the reactor, forming a 2 cm high bed. Three thermocouples positioned 0 cm, 1 cm, and 2 cm above the gas distributor measured temperatures at the bottom (T₃), middle (T₂), and top (T₁) of the bed. The reactor was heated to initial carbonation temperatures (550 °C, 600 °C, or 650 °C) under 100% N₂. Upon temperature stabilization, the atmosphere was switched to carbonation conditions (100% CO₂ or 70% CO₂/30% N₂). Temperature changes during carbonation were monitored until cooling to the initial temperature, after which the system was purged with 100% N₂ and heated to 850 °C for 20 min of calcination.

To determine the effective heat storage conversion rate, 1 g of post-carbonation sample was extracted, cooled in a N₂-filled desiccator for 3 min, and weighed. The sample was then calcined in a fixed-bed reactor under identical conditions (850 °C, 100% N₂) for 20 min, cooled, and reweighed. For the Al₂O₃-added samples, limestone particles were separated from Al₂O₃ via a vibrating sieve before weighing. The BFBR was cycled by resetting the temperature and gas atmosphere for subsequent tests. Apparent gas velocities during calcination (U_cal) and carbonation (U_carb) were 0.041 m/s and 0.041–0.062 m/s, respectively.

2.3. Performance Evaluation Metrics

The exothermic performance of limestone was evaluated using three key parameters: effective heat storage conversion rate, heat storage density, and average exothermic temperature.

Effective Heat Storage Conversion Rate (X_ef,N): Defined as the ratio of the mass of CaO that is actually reacted during each carbonation cycle to the total mass of the sample before carbonation, calculated by Equation (1). Heat Storage Density (H_g,N): Represents the maximum heat released per unit mass of CaO during each carbonation cycle, derived from Equation (2). Average Exothermic Temperature (T_a): Calculated as the arithmetic mean of temperatures measured at three monitoring points using Equation (3).

$$X_{\mathrm{ef},N}={\displaystyle \frac{m_{\mathrm{car},N}-m_{\mathrm{cal},N}}{m_0}}\cdot {\displaystyle \frac{M_{\mathrm{CaO}}}{M_{{\mathrm{CO}}_2}}}$$

(1)

$$H_{\mathrm{g},N}={\displaystyle \frac{\Delta m\cdot \Delta H^0}{M_{{\mathrm{CO}}_2}\cdot m_0}}=X_{\mathrm{ef},N}\cdot {\displaystyle \frac{\Delta H^0}{M_{\mathrm{CaO}}}}$$

(2)

$$T_{\mathrm{a}}={\displaystyle \frac{T_1+T_2+T_3}{3}}$$

(3)

$$T_{\mathrm{d}}={\displaystyle \frac{\left|T_1-T_2\right|+\left|T_1-T_3\right|+\left|T_2-T_3\right|}{3}}$$

(4)

$${\mathrm{R}}_{\mathrm{h}}={\displaystyle \frac{{{\mathrm{T}}_{\mathrm{a},\mathrm{m}\mathrm{a}\mathrm{x}}-\mathrm{T}}_{\mathrm{a},0}}{∆\mathrm{t}}}$$

(5)

Additional Exothermic Parameters: Temperature difference (T_d), evaluated via Equation (4) to quantify thermal uniformity across the bed. The maximum exothermic temperature (T_max) characterizes the peak temperature reached during the exothermic reaction process and reflects the limit value of thermal energy release. The duration of heat release (t_h) quantifies the time span during which the system temperature is maintained within the range of T_max ± 3 ° C, and is used to evaluate the stability of heat output. The heat release rate (R_h) characterizes the initial stage of heat release kinetics by calculating the differential quotient of the temperature increment (ΔT = T_maxial − T_initial) and time variation, and can be used as an important indicator of the propagation speed of the reaction front. These three parameters constitute a complete evaluation system, where T_max determines the energy grade, t_h controls the duration of heat release, and R_h regulates the power density, collectively describing the key performance characteristics of thermochemical energy storage systems.

3. Results and Discussion

3.1. Fluidization Instability and Its Impact on Limestone’s Exothermic Performance

Figure 3 illustrates the bed morphology at the end of the carbonation phase during the first heat storage cycle under a 100% CO₂ atmosphere. As shown, void channels formed within the carbonation bed under pure CO₂ conditions. This phenomenon arises because the carbonation reaction predominantly occurs in the emulsion phase of the fluidized bed. CO₂ gas in the emulsion phase is extensively absorbed by CaO, while gas from bubbles inadequately compensates for the absorbed CO₂. Consequently, the drag force in the emulsion phase gradually diminishes until it becomes insufficient to counteract the gravitational force of particles, triggering local defluidization.

The formation of local defluidization reduces the voidage in the emulsion phase, hindering gas exchange between bubbles and the emulsion phase. Additionally, exothermic carbonation elevates the emulsion phase temperature relative to the bubble phase, amplifying diffusion resistance for gas migrating from bubbles to the emulsion. This results in bubble coalescence and short-circuiting flow through the bed. Even when CO₂ absorption slows near the reaction endpoint, re-fluidization remains unachievable. In gas volume reduction reactions, elevated reaction rates exacerbate gas depletion, intensifying this instability.

Under 100% CO₂ carbonation, channeling (indicated by arrows in Figure 3) directs gas flow through void channels, effectively transitioning the bed to a fixed-bed state. Poor gas–solid mixing and inefficient heat/mass transfer in this state degrade the limestone’s exothermic performance and induce significant temperature differences across the bed. Thus, this study evaluates gas–solid mixing uniformity using CaO exothermic temperatures (T₁, T₂, T₃) and their spatial gradients (T_d,max, T_d,avg).

Figure 4 and

Table 2. Exothermic parameters of limestone at temperature measurement points under varying CO₂ concentrations during carbonation.

	Point	Tmax (°C)	th (s)	Rh (°C/s)	Td,max	Td,avg
100% CO2	1	838.0	21	7.12	60	38
	2	879.0	29	5.39
	3	825.0	32	5.05
70% CO2	1	853.0	134	3.95	18	7
	2	853.0	154	3.95
	3	845.0	144	3.67

illustrate the influence of CO₂ concentration on limestone’s exothermic temperatures and temperature differences during the first heat storage cycle. As shown in Figure 4a, under 100% CO₂ carbonation, channeling caused poor heat and mass transfer within the bed, leading to significant disparities in T_max (maximum exothermic temperature and R_h (exothermic rate) across the three measurement points. At measurement point 2, the limestone’s T_max reached 879.0 °C, 54 °C higher than at point 3. Similarly, R_h at point 1 (7.12 °C/s) exceeded that at point 3 by 41%. Under 100% CO₂, the maximum and average temperature differences (T_d,max and T_d,avg) reached 60 °C and 38 °C, respectively.

Diluting the carbonation atmosphere with N₂ achieves two effects: (1) reducing the carbonation reaction rate, and (2) mitigating excessive shrinkage of the emulsion phase caused by rapid CO₂ absorption, thereby stabilizing fluidization. As demonstrated in Figure 4b, under 70% CO₂, T_max and R_h at points 1 and 2 were identical (853.0 °C and 3.95 °C/s, respectively), while point 3 showed slightly lower values (845.0 °C and 3.67 °C/s). This discrepancy likely arises because point 3 measures temperatures near the gas distributor plate, where bed expansion during fluidization reduces particle density, and incoming gas cools the plate surface. Under 70% CO₂, T_d,max and T_d,avg decreased to 18 °C and 7 °C, respectively.

In the fluidized bed system, two distinct phases govern the reaction dynamics: (i) The bubble phase, characterized by gas-rich voids (typically 1–10 mm diameter) with minimal particle content, which act as bypass channels that reduce gas–solid contact efficiency. These bubbles coalesce vertically, exacerbating temperature gradients. (ii) The emulsion phase, constituting a dense particle suspension, facilitates the primary carbonation reaction CaO + CO₂→CaCO₃. Crucially, the Al₂O₃ additives (0.05–0.075 mm) modulate interphase exchange by increasing emulsion-phase voidage, thereby mitigating CO₂ depletion-induced contraction that triggers defluidization.

The N₂ dilution strategy was primarily adopted to modulate the carbonation kinetics by reducing CO₂ partial pressure (from 100% to 70%), thereby preventing excessive emulsion-phase contraction, while subsequent Al₂O₃ addition serves as a complementary approach to mechanically sustain interparticle spacing. These combined effects address defluidization at both the macroscopic (gas-phase composition) and microscopic (particle-level interaction) scales: N₂ dilution effectively decelerates the carbonation front propagation, whereas Al₂O₃ particles overcome the inherent limitation of gas-phase modulation—its inability to prevent microparticle cohesion—by functioning as spacers to maintain minimum fluidization voidage under rapid CO₂ absorption. Although lowering CO₂ concentration stabilizes fluidization and homogenizes the bed’s temperature field, it concurrently reduces limestone’s exothermic performance, underscoring the necessity for strategies that prevent fluidization instability in pure CO₂ atmospheres without compromising reaction efficiency.

Figure 5 and

Table 3. Exothermic parameters of limestone at temperature measurement points under different cycle numbers.

N	Point	Tmax (°C)	th (s)	Rh (°C/s)	Td,max	Td,avg
10	1	834.0	36	5.25	25	12
	2	882.0	34	5.81
	3	878.0	25	4.79
20	1	874.0	10	7.53	9	3
	2	873.0	9	7.27
	3	864.0	15	7.22

present the effects of cycle number on limestone’s exothermic temperatures and temperature differences. In gas-consuming reactions where gas volume decreases with progress, accelerated reaction rates exacerbate gas depletion, thereby intensifying fluidization instability. Notably, repeated heat storage cycles degrade limestone activity, reducing both CO₂ absorption capacity and reaction kinetics. This decay may suppress channeling phenomena after multiple cycles, as hypothesized. Figure 5a,b depict exothermic profiles during the 10th and 20th cycles, respectively. By the 20th cycle, T_max (maximum exothermic temperature) and R_h (exothermic rate) converged across all three measurement points, with T_d,max and T_d,avg reduced to 9 °C and 3 °C, respectively. This uniformity confirms stabilized fluidization and a homogeneous temperature field, indicating the absence of channeling. Concurrently, limestone’s CO₂ capture performance deteriorated significantly over cycles. The effective conversion rate (X_ef,20) and heat storage density (H_g,20) dropped to 0.25 and 0.79 MJ/kg, marking a 63% decline compared to initial values (X_ef,1 and H_g,1). These results validate the hypothesis that prolonged cycling mitigates fluidization instability at the cost of reduced reaction efficacy.

This study reveals a novel stabilization mechanism through comparative analysis of performance between the initial and 20th cycles. The results demonstrate that the intrinsic reduction in CO₂ capture capacity beyond 20 operational cycles spontaneously suppresses emulsion-phase contraction, thereby eliminating the need for continuous stabilizer additives. Experimental evidence confirms that while the thermochemical properties of CaCO₃ degrade with increasing cycles, this degradation process exhibits a definitive correlation with the progressive reduction in fluidization instability. Notably, the research identifies 20 cycles as a critical threshold—when material activity decays to this level, defluidization phenomena completely disappear. This discovery establishes a quantitative relationship between material deactivation degree and hydrodynamic stability, providing new theoretical foundations for optimizing thermochemical energy storage system design.

3.2. Effect of Al₂O₃ Particle Addition on Limestone’s Exothermic Performance

Figure 6 and

Table 4. Exothermic parameters of limestone at temperature measurement points with varying Al₂O₃ particle addition.

Additive Amount	Point	Tmax (°C)	th (s)	Rh (°C/s)	Td,max	Td,avg
1 wt.%	1	873.0	33	4.73	44	17
	2	850.0	28	3.74
	3	884.0	39	4.13
3 wt.%	1	893.0	63	4.33	67	17
	2	862.0	59	3.93
	3	888.0	74	4.67
5 wt.%	1	894.0	47	5.11	22	4
	2	887.0	43	4.72
	3	884.0	46	4.91

demonstrate the influence of Al₂O₃ particle addition (0.05–0.075 mm) on limestone’s exothermic temperatures and temperature differences during the first heat storage cycle. As shown in Figure 6a,b, Al₂O₃ doping within 1–3 wt.% exhibited limited improvements in fluidization quality, with significant T_d,max (maximum temperature difference) and T_d,avg (average temperature difference) persisting across measurement points. For instance, with 3 wt.% Al₂O₃, T_d,max and T_d,avg reached 67 °C and 17 °C, respectively. Increasing the Al₂O₃ doping ratio to 5 wt.% reduced T_d,max and T_d,avg by 67% and 76%, respectively. Notably, at 3 wt.% doping, the maximum T_max discrepancy between measurement points 1 and 2 was 31 °C, whereas 5 wt.% doping narrowed this gap to 10 °C across all three points. Additionally, exothermic rates (R_h) converged at all measurement points with 5 wt.% Al₂O₃, indicating enhanced fluidization homogeneity. The improved fluidization stems from inert Al₂O₃ particles dispersing around limestone grains, mitigating emulsion-phase voidage reduction caused by rapid CO₂ absorption. This mechanism facilitates gas exchange between bubbles and the emulsion phase, stabilizing bed dynamics and homogenizing the temperature field.

3.3. Effect of Al₂O₃ Particle Size on Limestone’s Exothermic Performance

Figure 7 and

Table 5. Exothermic parameters of limestone at temperature measurement points with varying Al₂O₃ particle sizes (Al₂O₃ 5 wt%).

Particle Size (mm)	Point	Tmax (°C)	th (s)	Rh (°C/s)	Td,max	Td,avg
0.125–0.18	1	890.0	60	3.19	51	15
	2	865.0	44	2.40
	3	885.0	52	2.88
0.075–0.125	1	889.0	51	4.47	53	12
	2	863.0	62	3.66
	3	886.0	56	4.56
0.05–0.075	1	894.0	47	5.11	22	4
	2	887.0	43	4.72
	3	884.0	46	4.91

illustrate the influence of Al₂O₃ particle size (5 wt.% addition) on exothermic temperatures and temperature differences in limestone at measurement points during the first heat storage cycle. As shown in Figure 7a,b, inert flow aid particles with sizes >0.075 mm provided limited fluidization improvement, with significant temperature differences (T_d,max and T_d,avg) persisting during both heating and stable exothermic phases. For example, with 0.075–0.125 mm Al₂O₃ particles, T_d,max and T_d,avg reached 53 °C and 12 °C, respectively. Reducing the particle size to 0.05–0.075 mm decreased T_d,max and T_d,avg by 60% and 67%, respectively. Smaller Al₂O₃ particles enhanced emulsion-phase voidage under identical doping ratios, mitigating CO₂ absorption-induced shrinkage and improving gas exchange between bubble and emulsion phases. This mechanism significantly stabilized fluidization. The optimal Al₂O₃ particle size in this study was 0.05–0.075 mm, achieving the most homogeneous temperature distribution and minimized fluidization instability.

4. Conclusions

This study investigated the impact of fluidization instability on limestone exothermic performance during carbonation under high-concentration CO₂ atmospheres, and explored the effects of Al₂O₃ particle addition as a flow aid. The key conclusions are as follows:

Mechanism of Fluidization Instability in Pure CO₂: Under 100% CO₂ carbonation, rapid CO₂ absorption in the emulsion phase leads to insufficient gas replenishment from bubbles. This imbalance between drag force, buoyancy, and particle gravity triggers local defluidization, which blocks gas exchange between bubbles and the emulsion phase. Consequently, enlarged bubbles channel through the bed, causing severe temperature heterogeneity (e.g., T_d,max = 60 °C, T_d,avg = 38 °C).

CO₂ Dilution with N₂ for Stabilization: Reducing the CO₂ concentration to 70% via N₂ dilution lowers the carbonation rate and prevents excessive emulsion-phase shrinkage. This strategy stabilizes fluidization and homogenizes the temperature field, reducing T_d,max and T_d,avg by 70% (to 18 °C) and 82% (to 7 °C), respectively.

Cycle-Dependent Degradation and Stability: After 20 heat storage cycles, limestone’s CO₂ capture capacity declines significantly, with effective conversion (X_ef,20) and heat storage density (H_g,20) dropping to 0.25 and 0.79 MJ/kg (a 63% reduction from initial values). However, this degradation eliminates fluidization instability, as evidenced by uniform temperature distribution (T_d,max = 9 °C, T_d,avg = 3 °C).

Optimized Al₂O₃ Addition for Fluidization Enhancement: Adding 5 wt.% Al₂O₃ particles (0.05–0.075 mm) stabilizes fluidization by preserving emulsion-phase voidage and enhancing bubble–emulsion gas exchange. This reduces T_d,max and T_d,avg to 22 °C and 4 °C, representing 63% and 89% improvements over pure limestone.

This study demonstrates that Al₂O₃ particle addition effectively mitigates fluidization instability in limestone-based thermochemical storage, reducing temperature variations by 63–89%. The findings offer practical solutions for calcium-looping systems in CO₂ capture and solar energy storage, where stable fluidization is critical. Future work should explore hybrid additives (e.g., Al₂O₃-MgO) and pilot-scale validation to advance industrial adoption.