Calcium Looping: On the Positive Influence of SO2 and the Negative Influence of H2O on CO2 Capture by Metamorphosed Limestone-Derived Sorbents.

The CO2 capture performance of sorbents derived from three distinct limestones, including a metamorphosed limestone, is studied under conditions relevant for calcium looping CO2 capture from power plant flue gas. The combined and individual influence of flue gas H2O and SO2 content, the influence of textural changes caused by sequential calcination/carbonation cycles, and the impact of CaSO4 accumulation on the sorbents' capture performance were examined using bubbling fluidized bed reactor systems. The metamorphosed limestone-derived sorbents exhibit atypical capture behavior: flue gas H2O negatively influences CO2 capture performance, while limited sulfation can positively influence CO2 capture, with space time significantly impacting CO2 and SO2 co-capture performance. The morphological characteristics influencing sorbents' capture behavior were examined using imaging and material characterization tools, and a detailed discussion is presented. This insight into the morphology responsible for metamorphosed limestone-derived sorbent's anomalous capture behavior can guide future sorbent selection and design efforts.

Introduction

Calcium looping (CaL) is a promising postcombustion CO2 capture technology that can potentially contribute to global decarbonization efforts. CaL is a high-temperature separation process that utilizes a dual fluidized bed system to continuously cycle solid CaO particles between carbonation conditions and calcination conditions, where carbonation conditions favor the forward and calcination conditions favor the reverse of the following reaction: CaO + CO2 ⇌ CaCO3.[1] During CaL, CO2 is captured from industrial or power plant flue gas by exothermic carbonation in the first fluidized bed reactor, the carbonator. Carbonation is performed at 650 °C to limit equilibrium constraints on efficient CO2 capture while maintaining a fast carbonation rate.[2] The resulting CaCO3 is transferred to the second fluidized bed reactor, the calciner, where CaO is regenerated and a highly enriched CO2 stream is produced by endothermic calcination. The energy required for calcination is often supplied by in situ oxy-combustion, which leads to a high CO2 partial pressure in the calciner and shifts the thermodynamic equilibrium discouraging calcination.[3,4] Calcination temperatures >900 °C are therefore required for efficient and fast sorbent regeneration.[5] Note that under CaL process conditions, CaO simultaneously irreversibly reacts with SO2 allowing CO2 and SO2 co-capture from sulfated flue gases. The resulting gas stream leaving the calciner therefore mainly comprises CO2 and H2O. High purity CO2, suitable for sequestration in appropriate geological formations, is obtained by simply dewatering the calciner effluent gas stream. CaL has been demonstrated at three pilot plants (≥1 MWth) achieving a technology readiness level of 6.[3,6−8]

The high-temperature nature of CaL allows high-grade heat recovery and auxiliary power generation downstream of the calciner and carbonator, which can improve the process energy efficiency relative to first generation capture technologies such as amine scrubbing.[9−11] The CaL process is most appealing when limestone-derived CaO is used as the sorbent. Limestone is a widely available natural CaCO3 source used in cement production, steel manufacturing, and flue gas desulfurization. The opportunity for synergy with industrial processes, coupled with the low cost of limestone, contributes to the economic viability of CaL, with competitive capture costs estimated by multiple technoeconomic studies.[12−14] While CaL with limestone-derived sorbents is an appealing approach to decarbonization, some technological concerns remain.

As limestone is exposed to calcination/carbonation cycles, its CO2 capture capacity, or “activity”, decays due to sintering, attrition, and exposure to deactivating flue gas constituents such as SO2.[15−17] This loss in sorbent activity, or deactivation, is offset by continuously adding fresh limestone to the system and purging the spent sorbent. Fresh limestone addition, however, increases the calcination load, reduces process energy efficiency, and increases costs.[18,19] Consequently, much research has focused on investigating different CaO sources, sorbent pretreatments, and approaches to spent sorbent reactivation; details can be found in recent reviews by Erans et al.[20] and Hu et al.[21]

Most studies on limestone-derived sorbent CO2 capture performance have focused on unmetamorphosed limestones and dolomites, with the influence of different limestone impurities, specifically in the form of inert supports, examined. In this study, we examine the capture performance of naturally occurring metamorphosed limestone-derived sorbents. Note that metamorphosed limestone, commonly referred to as marble, is widely available and priced comparably to unmetamorphosed limestone. Metamorphosed limestone is a limestone that has been subjected to elevated temperatures and pressures during natural geological processes. Under these conditions, the limestone experiences deformation and recrystallization, which lead to distinct metamorphic morphologies. While the specific metamorphic morphologies that evolve depend on the geological conditions of metamorphism, limestone metamorphism primarily results in increased grain sizes.[22] The CO2 capture performance of metamorphosed limestone-derived sorbents is expected to differ from that of unmetamorphosed limestone-derived sorbents because limestone-derived sorbent activities and reactivities are determined by sorbent morphology, specifically, crystalline structure, porosity, and surface area.[23−25]

The primary sintering mechanisms responsible for reducing unmetamorphosed limestone-derived sorbent activity with progressing calcination/carbonation cycles entail grain coarsening, or growth, by crystallite migration and coalescence, and particle shrinkage, or densification.[25] These morphological changes result in an increase in the sorbent’s average pore diameter and a reduction in the surface area and porosity.[25,26] A recent study by Pinheiro et al.[27] demonstrated that metamorphosed limestone-derived sorbents experience a reduction in mean pore diameters and therefore an increase in the CaO surface area with calcination/carbonation cycles. Moreover, Pinheiro et al.[27] correlated these observations with lower deactivation rates and improved CO2 capture capacity relative to unmetamorphosed limestone-derived sorbents.

The underlying cause of the anomalous behavior reported by Pinheiro et al.[27] merits further exploration. Deeper insights into this unexpected behavior can inform natural sorbent selection and synthetic sorbent design efforts. Moreover, flue gas H2O and SO2 content impacts CO2 capture by sorbents, but this impact has not been previously studied for metamorphosed limestone-derived sorbents. Examining capture in realistic chemical environments, by a sorbent that exhibits an atypical structure and capture behavior, can shed more light on the physiochemical characteristics and interactions that dominate CO2 and SO2 capture behavior.

In this manuscript, we address the underlying causes of the structural changes experienced by metamorphosed limestone due to cycling and we examine the influence of the CaO matrix and realistic chemical environments on capture. We use fluidized bed reactor systems to investigate the individual and combined influence of flue gas H2O and SO2 content on metamorphosed limestone-derived sorbent CO2 capture behavior. While H2O typically positively influences CO2 capture and SO2 negatively influences CO2 capture in unmetamorphosed limestone-derived sorbents,[16,28] we show that the metamorphosed limestone-derived sorbent unexpectedly demonstrates the opposite behavior. The roles of the chemical composition and structure are resolved by studying three limestones: a metamorphosed limestone, an unmetamorphosed limestone with a similar chemical composition, and a chemically distinct unmetamorphosed limestone. The influence of the particle structure is further explored by considering two distinct particle size distributions (PSDs) for each limestone. Furthermore, we examine the influence of multiple calcination/carbonation cycles and of CaSO4 accumulation on limestone-derived sorbent textural evolution and capture behavior. We use imaging and material characterization tools to examine the limestones and their calcines at different instances in the process and devote a discussion to the underlying factors that control the morphological changes observed in metamorphosed limestone-derived sorbents.

Experimental Section

Metamorphosed Saabar limestone, from the Red Sea coastal plane, and unmetamorphosed Riyadh limestone, from the Arabian platform, were provided by United Mining Investments Co., Saudi Arabia. These limestones were milled by BHS-Sonthofen GmbH and sieved by Allgaier Process Technology GmbH to “fine” and “coarse” size fractions: 100–400 and 400–1000 μm, respectively. Unmetamorphosed German limestone, from the Central Uplands, was also obtained as fine and coarse size fractions. PSDs were measured using a Malvern Mastersizer 3000, equipped with an Aero S dry particle dispersion unit. The elemental composition of each limestone fraction was determined by X-ray fluorescence (XRF), performed using an XGT-7000 (Horiba). A Bruker D8 ADVANCE equipped with a Cu Kα radiation source (λ = 1.5418 Å) and operated at 40 mA was used to acquire powder X-ray diffraction (PXRD) patterns on 5–80° 2θ range with 0.03°/step and 0.5 s/step. The crystalline phase identifications were carried out in Diffract.Eva software with the help of the PDF-4+ (2019) crystal database. The structural characteristics, unit cell parameters, estimated content, and degree of crystallinity were evaluated using Reflex from the Accelrys Material Studio software package. Elemental distribution within the limestone samples and limestone surface morphology were mapped and imaged using a Zeiss MERLIN field emission scanning electron microscope equipped with a Gemini II electron optical column and Oxford Instruments X-Max80 silicon drift detector for energy dispersive spectroscopy (EDS).

Two bubbling fluidized bed (BFB) reactor systems were used to study the individual and combined influence of flue gas H2O and SO2 content on CO2 capture by each of the limestone fractions. A 150 mm diameter, 3.5 m tall electrically heated 20 kWth BFB (BFB 1) was used to precalcine 5 kg samples of each limestone fraction. This BFB is equipped with a double cyclone system and candle filter for separating entrained fines from the reactor effluent before sampling by a gas analyzer (ABB Advance Optima 2020) for continuous CO2, SO2, and O2 monitoring. The limestones were calcined under N2 for 20 min at 850 °C. The capture performance of the precalcined samples was examined using the second 70 mm diameter 7 kWth BFB (BFB 2), described in detail elsewhere,[29] equipped with effluent line sampling for continuous CO2, SO2, and O2 gas analysis (X-STREAM Emerson Process Management GmbH & Co. OHG).

BFB 2 was preheated to 650 °C with an initial inert bed inventory composed of silica sand. Gas flow rates were set to achieve a fluidization velocity of 0.5 m/s, and depending on the experimental run, an influent gas composition of either (i) 12–14% CO2, (ii) 12–14% CO2, 10–14% H2O, (iii) 12–14% CO2, 2000 ppm SO2, or (iv) 12–14% CO2, 10–14% H2O, 2000 ppm SO2, with 3% O2 and the balance N2. Baseline gas concentrations were recorded. A sample of precalcined limestone was then dropped into BFB 2 and the semibatch carbonation/sulfation reaction was allowed to proceed for 30 min while gas concentrations were recorded. Four runs with variable initial space times were performed per gas atmosphere composition and per precalcined limestone fraction. The total BFB 2 bed inventory was held constant at 1 kg for all runs while the mass of precalcined limestone added into the reactor was varied to achieve different initial space times corresponding to 9.3, 4.7, 2.3, and 0.9 min. A number of random experimental repeats were performed with results used to calculate error bars.

The influence of limestone cycling and CaSO4 content on CO2 and SO2 co-capture was studied using the same two BFB systems. Fine Riyadh and Saabar limestones were cycled between carbonation and calcination conditions in BFB 1. BFB 1 was electrically preheated to 400 °C, and a 10 kg limestone sample was added in an inert atmosphere of N2 with fluidization maintained at ∼0.4 m/s. The temperature was then raised to calcination temperatures between 850 and 900 °C and held for 15 min. In addition to electrical heating, in situ methane combustion was employed to raise the reactor temperature from 750 to 900 °C. The resulting calcination atmosphere was ∼30% CO2 and 13% H2O with the balance nitrogen. The calcined limestone was then collected by means of a valve at the bottom of BFB 1, the reactor was cooled to 650 °C, and carbonation conditions were set. The calcined limestone was then added back into BFB 1 and allowed to carbonate for 20 min. Afterward, the recarbonated sorbent was reheated to 750 °C in an inert atmosphere in BFB 1, and calcination and carbonation was repeated with multiple cycles performed per experimental run. Two runs were performed per limestone source with different carbonation gas atmospheres: (i) 12% CO2, 10% H2O and (ii) 12% CO2, 10% H2O, 2000 ppm SO2, with 0.5% O2 and the balance N2. Calcined and recarbonated samples were collected after each calcination and carbonation. Fines generated by particle attrition and fragmentation were collected from the cyclones and weighed after each calcination. The limestones were cycled until sorbent activity, X, was reduced to 0.1 mol/mol. The co-capture performance of samples of the calcined limestones with X ≈ 0.6, 0.3, and 0.1 mol/mol was evaluated using BFB 2 and the experimental method previously described. A synthetic flue gas composed of 12% CO2, 10% H2O, 2000 ppm SO2, and 3% O2 with the balance N2 was used.

A custom simultaneous thermal analysis unit (Linseis GmbH) equipped with a gas mixing manifold was used to assess calcined limestone activity at different cycle numbers. Briefly, thermogravimetric analysis (TGA) was performed on 10 mg samples of calcined limestone by heating (200 °C/min) the sample to 650 °C in N2 and then isothermally carbonating the sample in a 12% CO2 atmosphere with the balance N2 for 30 min. The recorded changes in sample mass were used to calculate X. TGA was also used to support the capture trends observed in BFB 2 runs; the capture behavior of calcined limestone samples was examined with TGA carbonation atmospheres set to either (i) 12–14% CO2, (ii) 12–14% CO2, 10–14% H2O, (iii) 12–14% CO2, 2000 ppm SO2, or (iv) 12–14% CO2, 10–14% H2O, 2000 ppm SO2, with 3% O2 and the balance N2.

Imaging and characterization were performed on calcined and recarbonated limestone samples taken at different experimental stages. The effect of cycling on PSDs was monitored using the Mastersizer. Alterations to crystallinity and crystal composition were tracked using PXRD. Morphological changes were studied using scanning electron microscopy (SEM)–EDS. BFB 1 calcination and carbonation efficiencies and sorbent sulfur accumulation were monitored using XRF and a Flash 2000 CHNOS Organic elemental analyzer (Thermo Scientific) equipped with a thermal conductivity detector. Brunauer–Emmett–Teller (BET) theory was used to evaluate ASAP 2420 N2 adsorption data and determine the surface area of calcined limestones. Combined results from N2 adsorption and mercury intrusion porosimetry, performed using an Autopore IV 9510 (Micromeritics), were used to determine calcined limestone porosity and map pore size distribution.

Results and Discussion

Experimental Approach Theory and Validation

The capture performance of the calcined metamorphosed and unmetamorphosed limestones is compared using two key CaL system metrics that influence energy efficiency and thus financial viability: capture efficiency and capture capacity. Capture efficiency, Ep, is defined as the rate of pollutant (p) uptake by the sorbent, Ḟp,c, normalized to the rate of pollutants entering the carbonator, where p refers to either CO2 or SO2Sorbent capture capacity, Xmax, is pragmatically defined as the maximum achievable CaO conversion to CaCO3 and CaSO4 while maintaining a fast CO2 capture rate

Note that carbonation is a two stage process, with an initial fast reaction stage followed by a slow product layer solid-state diffusion-controlled stage. The initial fast reaction stage is controlled by CO2 diffusion into sorbent pores, CO2 surface diffusion, and carbonation kinetics.[30,31] In eq , tmax is the time at which the fast reaction stage ends. Sorbent capture efficiency and capture capacity are calculated using BFB 2 gas analyzer data and a mass balance across the gas phasewhere xi refers to the combined mole fraction of inerts in the gas phase.

Sorbent capture behavior is visualized by plotting capture efficiency versus CaO conversion, X. CaO conversion is calculated by solving eq at discrete time intervals, t. Note that CaO conversion refers to the combined conversion of CaO to CaCO3, XCaCO, and CaSO4, XCaSO. The extent of CaO conversion in the carbonator, or carbonator conversion (Xcarb), is a key CaL system design parameter. System cost and energy requirements are minimized while sufficient capture efficiency is maintained by operating at an appropriate Xcarb.[32] Therefore, monitoring changes to capture efficiency with respect to CaO conversion provides a valuable perspective.

The solid lines in Figure a show the CO2 capture behavior of once-calcined fine metamorphosed Saabar and unmetamorphosed Riyadh and German limestones when exposed to a 14% CO2 and 3% O2 in the N2 influent in BFB 2. Two unmetamorphosed limestones were studied to resolve the roles of the chemical composition and structure: while both metamorphosed Saabar and unmetamorphosed Riyadh limestones have similar impurity content, with a relatively high quartz (SiO2) content, unmetamorphosed German limestone reveals higher purity in terms of the carbonate phase composition (see Table ). The limestone impurity content is corroborated by PXRD (Figures S1–S3 and Table S1).

Figure 1

CO2 capture performance of once-calcined fine limestones at different BFB 2 influent gas compositions (τ0 = 4.7 min): (a) capture efficiency, and (b) conversion curves. Solid lines: 14% CO2 and 3% O2 in N2, dashed lines: 14% CO2, 14% H2O, 2000 ppm SO2 and 3% O2 in N2. The gray dashed line indicates Xmax.

Table 1

Chemical Composition of Fine Limestones (wt %)

	CaO	SiO2	MgO	Al2O3	Fe2O3	SO3	Na2O	K2O	LOIa
Saabar	51.2	3.91	0.64	0.25	0.18	0.20	0.05	0.02	41.6
Riyadh	51.8	3.17	0.36	0.33	0.21	0.51	0.05	0.07	42.6
German	54.2	0.61	0.61	0.09	0.07	0.14	0.01	0.04	43.2

LOI: loss on ignition.

Sorbent capture capacity, visualized in Figure b as the CaO conversion at the inflection point between fast and slow capture rates, coincides with capture efficiency reduction to ∼0.2 mol/mol (see Figure a). The capture capacity varies per limestone source, with the Riyadh sorbent exhibiting the greatest CO2 capture capacity followed by Saabar and German. The capture efficiency peak seen in Figure a coincides with the initial introduction of CaO into BFB 2. Relatively low initial molar ratios of CaO to CO2, or space times (τ) are used to ensure that the recorded capture efficiency is not reflective of the thermodynamic equilibrium of carbonation. Under these conditions, we propose that the recorded CO2 and SO2 capture behavior is diffusion- and kinetically controlled and is therefore correlated to the sorbent’s initial morphology and subsequent morphological changes with progressing sulfation and carbonation. The validity of our approach and the experimental methodology are supported by solving for the apparent carbonation kinetic constant.

The CaO fast carbonation rate can be calculated using the following semiempirical model adapted by Hawthorne et al.[33] from Nitsch’s[30] kinetic rate modelwhere m is a semi-empirical exponent that varies from 2/3 under kinetic-controlled conditions to 4/3 under diffusion-controlled conditions.[18] Bhatia and Perlmutter[2] noted that in accordance with a spherical grain model, the kinetic constant, k, is a function of the surface reaction rate constant, ks, the sorbent’s initial surface area, So, and porosity, εo, such that

The assumption that capture behavior recorded in BFB 2 is correlated with the sorbent morphology is supported by solving for the apparent kinetic constant using gas analyzer data and eq (kx,Exp) and evaluating the results against the kinetic constant calculated from the BET surface area and porosimetry data using eq (kx,Theo). Relevant sorbent morphological data is presented in Table .

Table 2

Relevant Sorbent Textural and Morphological Data

sorbent D [μm]	N		So [m2/cm3]	εo	kx,Theo × 103 [m3/mol·s]	Sm [m2/cm3·g CaO]	εm [g–1 CaO]	CaSO4 [wt %]
400–1000	1	Saabar	12.6	0.28	10.5	27.5	0.61
		Riyadh	12.8	0.52	15.8	15.3	0.61
		German	12.9	0.47	14.4	15.0	0.54
100–400	1	Saabar	33.4	0.37	31.7	38.8	0.43
		Riyadh	19.7	0.42	20.3	23.7	0.51
		German	11.8	0.48	13.4	13.3	0.53
	6	Saabar	7.46	0.26	6.01	9.21	0.32
		Riyadh	6.96	0.49	8.11	7.91	0.56
	13	Saabar	6.02	0.20	4.49	7.93	0.26
		Riyadh	5.52	0.30	4.70	6.98	0.38
	3 w/SO2	Saabar	8.72	0.30	7.45	11.8	0.41	1.4
		Riyadh	7.91	0.48	9.07	9.89	0.60	2.9
	6 w/SO2	Saabar	6.02	0.31	5.23	8.37	0.44	3.9
		Riyadh	5.40	0.41	5.43	7.01	0.53	5.2

The apparent kinetic constant, kx,Exp, is calculated for each of the once-calcined fine limestones (see Figure a). Combined CO2 capture data is used from all four BFB 2 runs performed at different initial space times when introducing a reactor influent of 14% CO2 and 3% O2 with the balance N2. Equation is fit to data corresponding to X > 0.3 and X < Xmax, where m ≈ 1.17. The CO2 concentration in BFB 2, CCO, is assumed uniform and is estimated using the logarithmic average of the inlet and outlet CO2 concentrations, as proposed by Shimizu et al.[1] While this simplifying assumption disregards the complex fluid dynamics of the BFB system, kx,Exp and kx,Theo are comparable when a simple correction factor is introduced (see Figure b). The good fit obtained in Figure b, with R2 = 0.94, suggests that this study’s approach and methodology are reasonable for the comparative evaluation of sorbents.

Figure 2

(a) Derivation of the apparent kinetic constant, kx,Exp, by fitting eq to experimental BFB 2 data. (b) Comparison of the experimental and theoretical kinetic rate constants (R2 = 0.94).

Kinetic constant calculations also indicate that assuming the sorbents exhibit similar m-values, once-calcined fine metamorphosed Saabar limestone should initially capture CO2 at a 20% faster rate than the Riyadh sorbent and a 70% faster rate than the German sorbent, with the capture rate discrepancy between the Riyadh and Saabar sorbent narrowing as conversion increases due to Saabar sorbent’s lower capture capacity. Considering that capture efficiency is an indicator of the capture rate, Figure a reveals that the initial CO2 capture rate of once-calcined fine Saabar limestone does not exceed the initial capture rate of Riyadh and German sorbents. This suggests that the Saabar sorbent’s initial carbonation rate is significantly more diffusion-controlled.

The once-calcined fine metamorphosed Saabar limestone demonstrates further anomalous capture behavior when carbonated in BFB 2 with an influent gas composition typical of coal flue gas, 14% CO2, 14% H2O, 2000 ppm SO2, and 3% O2 in N2 (see Figure a). While both unmetamorphosed limestone-derived sorbents demonstrate enhanced CO2 capture efficiency and capture capacity when H2O and SO2 are introduced, the metamorphosed limestone-derived sorbent experiences reductions in both CO2 capture efficiency and capture capacity. This anomalous behavior is better understood by examining the individual and combined effects of H2O and SO2 on the capture behavior of metamorphosed and unmetamorphosed limestone-derived sorbents.

Influence of Gas Composition on CO2 and SO2 Capture Behavior

The influence of reaction atmosphere composition on the CO2 capture performance of each of the once-calcined metamorphosed and unmetamorphosed limestones is presented in Figure . Note that all limestones were calcined at 850 °C in N2. Although calcination at 850 °C in N2 is not practical for CaL systems, these conditions allow baseline comparison of our limestone-derived sorbents’ morphology and capture behavior with much of the published data[12,20,34] since a bulk of CaL sorbents has been studied under these conditions. The calcined limestone capture behavior was studied with BFB 2 influent gas composition varied between (i) 12–14% CO2, (ii) 12–14% CO2, 10–14% H2O, (iii) 12–14% CO2, 2000 ppm SO2, and (iv) 12–14% CO2, 10–14% H2O, 2000 ppm SO2, with 3% O2 and the balance N2.

Figure 3

Influence of gas composition on the CO2 capture performance of once-calcined limestones (τo = 4.7 min): (a) fine Riyadh, (b) coarse Riyadh, (c) fine German, (d) coarse German, (e) fine Saabar, and (f) coarse Saabar.

The two once-calcined unmetamorphosed Riyadh and German limestones behave similarly when exposed to different influent gas compositions in BFB 2 (see Figure a–d). Relative to baseline conditions (CO2 and O2 in N2), H2O addition positively influences the once-calcined unmetamorphosed limestones’ capture efficiency and capture capacity. Conversely, SO2 addition reduces both capture efficiency and capture capacity. When both H2O and SO2 are added, the presence of H2O has a greater influence on the combined effect. These trends are observed for fine and coarse once-calcined Riyadh and German limestones, for all examined space times, and are in line with the results from previous studies.

The positive influence of H2O addition on CO2 capture efficiency and capacity is well documented.[28,35−39] This phenomena has been attributed to the formation of a hydroxylated layer on the sorbent surface, which is reported to enhance surface diffusion and reactivity with carbon dioxide.[28,35−39] The negative influence of SO2 addition has also been widely reported[16,34,40] and is expected since sulfation and carbonation are competitive reactions. Furthermore, CaSO4 has a larger molar volume than CaCO3 and can cause sorbent deactivation by pore blockage.[16] The stronger influence of H2O on capture capacity, in an atmosphere containing both H2O and SO2, has also been previously reported.[17,41] The hydroxylated layer formed on the sorbent surface in the presence of H2O is polar and causes pore narrowing.[41] Sorbent permeability to larger polar SO2 molecules is therefore reduced in the presence of H2O, while permeability to smaller nonpolar CO2 molecules remains undisrupted.[42]

The once-calcined metamorphosed Saabar limestone behaves unexpectedly when exposed to different influent gas compositions in BFB 2 (see Figure e,f). While H2O addition initially enhances CO2 capture efficiency, it ultimately has a negative effect on both CO2 capture efficiency and CO2 capture capacity. Once-calcined fine Saabar limestone experiences a shorter-lived positive effect due to H2O addition and a significantly greater negative effect when compared to once-calcined coarse Saabar limestone. When SO2 is added into the capture atmosphere, the impact on fine Saabar sorbent’s CO2 capture efficiency varies from negative to positive, depending on the initial space time (see Figure ). SO2 addition may also positively influence the CO2 capture capacity of fine Saabar sorbent; at τo = 9.3 min, fine Saabar sorbent experiences enhanced conversion to CaCO3 in the presence of SO2. Note that CaL systems typically operate at relatively high space times to ensure a sustained high capture efficiency. Unlike the fine Saabar sorbent, once-calcined coarse Saabar limestone’s capture capacity and efficiency are not positively influenced by SO2 addition; nonetheless, relative to the once-calcined Riyadh and German limestones, the coarse Saabar sorbent is less negatively impacted by SO2 addition. Furthermore, for both fine and coarse Saabar sorbents, superior CO2 capture capacity and capture efficiency are achieved in a capture atmosphere containing both H2O and SO2 when compared to a capture atmosphere with only H2O added. Limited sulfation of once-calcined Saabar limestones enhances this sorbent’s CO2 capture performance.

Figure 4

Influence of space time on the relative CO2 capture behavior of once-calcined fine Saabar limestone: (a) τo = 9.3 min, (b) τo = 4.7 min, (c) τo = 2.3 min, and (d) τo = 0.9 min.

The influence of sorbent sulfation on carbonation performance is examined using data from the experimental sets in which BFB 2 influent gas composition is set to either: (i) 12–14% CO2, 2000 ppm SO2, or (ii) 12–14% CO2, 2000 ppm SO2, 10–14% H2O, and 3% O2 with the balance N2. Note that the two once-calcined unmetamorphosed Riyadh and German limestones exhibit similar SO2 capture performance, and therefore, once-calcined Riyadh limestone is selected as a representative example with its SO2 capture performance discussed in more detail. Once-calcined German limestone’s SO2 capture performance is presented in Figure S4.

Once-calcined unmetamorphosed Riyadh and metamorphosed Saabar limestones exhibit similar SO2 capture efficiency under H2O-free conditions (see Figure ). Upon H2O addition, once-calcined unmetamorphosed limestones exhibit enhanced SO2 capture efficiency while the fine Saabar sorbent exhibits reduced SO2 capture efficiency. This observed divergence in behavior is analogous to the divergence in CO2 capture behavior observed upon H2O addition. Under reactor conditions, where the CO2 concentration is greater than the equilibrium CO2 concentration, the direct sulfation pathway (CaCO3 + SO2 + 1/2O2 → CaSO4 + CO2) is favored.[34] Since H2O addition enhances once-calcined unmetamorphosed limestones’ carbonation efficiency, sulfation efficiency is also enhanced. Since H2O addition reduces once-calcined fine Saabar limestone’s carbonation efficiency, sulfation efficiency is also reduced. For once-calcined coarse Saabar limestone, the SO2 capture efficiency is initially enhanced by H2O addition and then drops at X ≈ 0.5. Note that X ≈ 0.5 also corresponds to the point at which H2O addition begins to negatively impact coarse Saabar sorbent carbonation (see Figure f). A relationship between CaSO4 and CaCO3 accumulation is revealed by comparing results from runs at different initial space times.

Figure 5

SO2 capture performance of once-calcined Saabar and Riyadh limestones (τo = 4.7 min): (a) fine and (b) coarse limestones.

In this study, the initial space time is increased by introducing more CaO into BFB 2 without changing the influent gas flow rates. Increasing the initial space time therefore leads to reduced CO2 and SO2 concentrations in BFB 2 during the initial fast carbonation stage and, consequently, slower carbonation and sulfation reaction rates. In BFB 2 atmospheres that do not contain SO2, the once-calcined Saabar, Riyadh, and German limestones achieve a slightly higher CO2 capture capacity at higher space times; capture capacity is seen to increase by <0.1 mol/mol with a tenfold increase in space time (Figure S5). A similar effect was previously reported by Manovic and Anthony[43] and is caused by the slower carbonation rate allowing improved surface diffusion thereby reducing the probability of CaCO3 forming in a manner that blocks unreacted CaO.

In BFB atmospheres that contain SO2, experimental results indicate that the relative average concentration of SO2 in the BFB (CSO/CCO) and the relative average SO2 capture rate (ḞSO/ḞCO) decrease as the initial space time is increased (Figure S6). Consequently, the relative accumulation of CaSO4 versus CaCO3 diverges for different initial space times (see Figure ). For both fine and coarse metamorphosed and unmetamorphosed sorbents, conversion to CaCO3 is favored at higher initial space time. The slow accumulation of CaSO4, observed at high initial space time, is correlated with enhanced carbonation efficiency and CO2 capture capacity under H2O-free conditions for the fine Saabar sorbent (as seen in Figure ). The discrepancy in relative conversion (CaSO4 vs CaCO3) due to space time appears around X > 0.2 for the fine Saabar sorbent under H2O-free conditions (Figure c), and the effects on CO2 capture performance can be seen in Figure at X > 0.2. Moreover, both size fractions of the once-calcined metamorphosed limestone are less susceptible to the influence of space time on relative conversion in the presence of H2O. For Saabar sorbents at all examined space times, H2O addition causes the sulfation reaction to slow before carbonation transitions to its slow solid-state diffusion-controlled stage (Figure S7). Consequently, the slow CaSO4 accumulation has a dampened influence on CaCO3 accumulation.

Figure 6

Influence of space time on the relative conversion of the fine Riyadh sorbent in (a) H2O-free, and (b) 10% H2O atmospheres, and fine Saabar sorbent in (c) H2O-free and (d) 10% H2O atmospheres.

The TGA runs performed in parallel with the BFB runs support the observed positive influence of H2O addition on once-calcined unmetamorphosed limestone capture capacity and the negative influence of H2O addition on once-calcined metamorphosed limestone capture capacity (see Figure ). Note that the TGA runs are executed with lower initial space times (<0.2 min) than the BFB runs and in line with our observations on the influence of space time: (i) once-calcined Riyadh limestone experiences a greater negative impact on CO2 capture capacity due to SO2 addition in TGA versus in the BFB, and (ii) once-calcined metamorphosed limestone experiences a negative impact on CO2 capture capacity due to SO2 addition in TGA.

Figure 7

TGA CO2 capture results for once-calcined fine (a) Riyadh and (b) Saabar limestones.

Role of the Sorbent Morphology on Capture Behavior

Pore size distributions for the once-calcined Saabar, Riyadh, and German limestones are presented in Figure , and the BET surface area and other relevant textural data are summarized in Table . Elemental analysis (Table S2) reveals that once-calcined fine Saabar limestone and both size fractions of the two unmetamorphosed limestone-derived sorbents exhibit similar purity, 85–90% CaO on a mass basis, but once-calcined coarse Saabar limestone is only 46% CaO. Since SEM–EDS imaging (Figure ) reveals that the impurities are largely nonporous, porosimetry and surface area data are reported per gram of CaO allowing comparison.

Figure 8

Pore size distributions for once-calcined (a) fine, and (b) coarse limestones. Solid lines: differential pore volume, dashed lines: cumulative intrusion.

Figure 9

SEM images of the (a) Saabar, (b) Riyadh, and (c) German limestones, and of the once-calcined fine (d) Saabar, (e) Riyadh, and (f) German limestones, and of the once-calcined coarse (g) Saabar, (h) Riyadh, and (i) German limestones. White arrows indicate impurities.

Pore size distributions for the once-calcined Saabar, Riyadh, and German limestones are multimodal with an intense peak in the 50–200 nm range. The fine and coarse metamorphosed Saabar limestone-derived sorbents have smaller peak pore diameters relative to the unmetamorphosed Riyadh and German limestone-derived sorbents. Pinheiro et al.[27] similarly found that metamorphosed limestone-derived sorbents exhibit peak pore diameters in the mesoporous range (2–50 nm), and unmetamorphosed limestones calcined once under similar conditions typically have peak pore diameters in the macroporous range (>50 nm).[44] Moreover, once-calcined fine Saabar limestone has a high surface area and exhibits low porosity; these textural characteristics were also reported by Pinheiro et al.[27] and appear to be typical of calcined fine metamorphosed limestones. While once-calcined fine Saabar limestone’s pore size distribution is narrow and uniform with macropores contributing very little to the pore volume, a significant fraction of once-calcined coarse Saabar limestone’s pore volume is due to macropores. These textural differences are likely responsible for the different capture behavior displayed by the fine and coarse Saabar sorbents.

Once-calcined metamorphosed limestone mesopores are susceptible to pore blockage during carbonation. Alvarez and Abanades demonstrated that while capture capacity is typically directly proportional to the surface area and coincides with the formation of a critical product layer, ∼50 nm in thickness,[44] pore blockage is likely in “narrow pores”, defined as <150 nm in diameter.[25] The pore blockage experienced by once-calcined fine Saabar limestone results in unreacted CaO, and therefore, this sorbent’s relatively high surface area does not equate to high capture capacity when carbonated in BFB 2 with an influent gas concentration of 14% CO2 and 3% O2 in N2 (Figure ). Once-calcined fine Riyadh limestone also experiences some pore blockage during carbonation; although it has double the surface area of once-calcined fine German limestone, once-calcined fine Riyadh limestone does not exhibit a proportionally greater capture capacity. The wider pores exhibited by once-calcined German and coarse Riyadh limestones are less susceptible to pore blockage by carbonation. In the case of once-calcined coarse Saabar limestone, CO2 capture involves gas intrusion into the macropores followed by permeation into the mesopores. Consequently, once-calcined coarse Saabar limestone is initially exposed to lower concentrations of CO2 at mesopore openings. This shielding of the mesopores by the macropores reduces and postpones mesopore blockage, as seen in Figure f. While once-calcined coarse Saabar limestone’s baseline CO2 capture efficiency is relatively high at X < 0.5, a steep drop in capture efficiency is observed at X ≈ 0.5. We propose that this drop in capture efficiency coincides with the delayed mesopore blockage resulting from the macropore shielding effect.

The difference in CO2 capture performance between once-calcined fine and coarse Saabar limestones when H2O is added can also be attributed to the coarse Saabar sorbent’s macropores. In the case of once-calcined fine Saabar limestone, introducing a hydroxylated layer causes further mesopore narrowing and enhances CO2 reactivity. Since fast carbonation rates further increase the probability of bottleneck formation,[43] H2O addition increases fine Saabar sorbent’s susceptibility to pore blockage. For once-calcined coarse Saabar limestone, initial CO2 capture efficiency is controlled by gas intrusion into the sorbent’s macropores, and therefore, once-calcined coarse Saabar limestone initially displays a capture behavior similar to macroporous once-calcined unmetamorphosed limestones, with enhanced CO2 capture efficiency observed due to H2O addition. The subsequent restricted gas permeation into the once-calcined coarse Saabar limestone’s mesopores leads to a delayed and significantly reduced negative impact on the sorbent’s capture capacity relative to that experienced by once-calcined fine Saabar limestone. In Figure f, the drop in once-calcined coarse Saabar limestone’s capture efficiency due to H2O addition is seen at X ≈ 0.4. The once-calcined metamorphosed limestones’ mesopores appear to be the dominant textural feature controlling capture capacity in these sorbents, with H2O addition intensifying mesopore blockage. Moreover, mesopore narrowing due to H2O addition more significantly impacts sulfation; in an atmosphere containing both H2O and SO2, the once-calcined metamorphosed limestones’ sulfation rate slows before the carbonation rate enters its slow solid-state diffusion-controlled stage.

The mesoporous nature and low porosity of once-calcined fine metamorphosed limestones limit SO2 intrusion into these sorbents’ pore network and contribute to these sorbents’ reduced susceptibility to sulfation deactivation. For once-calcined fine metamorphosed limestones, sulfation likely proceeds through the “unreacted core” mode,[45] with a CaSO4 shell forming on the external surface of the sorbent. Since direct sulfation occurs, the CaSO4 shell that forms is porous.[46] While the wider pores and higher porosity exhibited by once-calcined unmetamorphosed limestones are less susceptible to pore blockage by carbonation, they exhibit less diffusional resistance to SO2 molecules, enhanced sulfation rates, and pore blockage by sulfation.[47] In the case of once-calcined fine German limestone, this translates to the high sulfation deactivation seen in Figure c. Once-calcined coarse German limestone, which has a similar porosity and pore size distribution to the once-calcined fine German limestone in the <200 μm range, experiences less sulfation deactivation than once-calcined fine German limestone. This is due to the coarse German sorbent’s >200 μm diameter macropores that are less susceptible to pore blockage by CaSO4 formation. Similarly, once-calcined coarse Saabar limestone’s high volume of macropores in the >200 μm range enhances this sorbent’s resistance to sulfation deactivation; relative to once-calcined Riyadh and German limestones, the coarse Saabar sorbent experiences the least sulfation deactivation (Figure ). While fine Saabar sorbent’s narrow pores influence this sorbent’s resistance to sulfation deactivation, the positive influence of SO2 on fine Saabar sorbent carbonation is not explained by pore size distribution alone. SEM–EDS and PXRD are used to further examine the sorbents’ morphology and structural and compositional characteristics.

SEM images of the limestones and their calcines are presented in Figure . While the impurities in the unmetamorphosed limestones are relatively evenly distributed throughout the limestones’ matrices (see Figure e,f,h,i), the metamorphosed Saabar limestone is primarily composed of large monomineralic phases separated by clear boundaries (see Figure d,g). This marbled distribution of monomineralic phases is due to grain size growth during limestone metamorphism. The grain size growth results in metamorphic differentiation or the redistribution of chemical components within the limestone without altering the overall chemical composition.[22] Note that the degree of metamorphic differentiation is a function of metamorphic conditions. SEM further reveals that once-calcined coarse Saabar limestone macropores are large fractures formed at phase boundaries and within the large monomineralic phases (see Figure g). This may indicate that Saabar limestone is more sensitive to thermal stress than Riyadh or German limestones.

While fine Riyadh and Saabar limestones have a similar bulk chemical composition (Table ), differences in the distribution of impurities and crystalline structure are likely responsible for the distinctive textural features of once-calcined Riyadh and Saabar limestones. Figure a–c shows the textural quality resulting from Saabar limestone’s large CaCO3 grains versus Riyadh and German limestones, which have more pronounced polycrystalline structures and exhibit lower crystallinity (Table S1). The structural refinements of the limestones suggest that their dominant phase is better described as calcium magnesium carbonate with a doping amount of magnesium (Table S1, Figure S8–S10). The large and active monomineralic grains of metamorphosed Saabar limestone have a composition of Ca0.91Mg0.09CO3, while Riyadh and German limestones contain larger amounts of magnesium and reveal compositions of Ca0.85Mg0.15CO3, and Ca0.84Mg0.16CO3, respectively. The higher CaCO3 purity of the metamorphosed limestone’s crystals may have resulted from metamorphic differentiation. The low incidence of structural defects in the metamorphosed limestone’s large, structured, highly crystalline, relatively pure CaCO3 phase likely leads to the evolution of the once-calcined metamorphosed limestone’s uniform mesopores. Unmetamorphosed limestones’ more polycrystalline structure and lower crystallinity likely leads to the evolution of the larger macropores and wider pore size distributions seen for these once-calcined unmetamorphosed limestones.

The bulk of once-calcined fine metamorphosed limestones’ reactive surface area is only accessible after reactant diffusion through the mesopores. Therefore, once-calcined fine metamorphosed limestone’s initial carbonation rate is more diffusion-limited than that of unmetamorphosed limestone-derived sorbents. Moreover, the higher CaO purity of once-calcined metamorphosed limestones translates to higher surface reactivity and increased susceptibility to unreacted CaO shielding and mesopore blockage by carbonation. We postulate that the introduction of a limited amount of impurities in the form of CaSO4 pacifies sections of the highly reactive once-calcined Saabar limestone’s matrix and introduces some resistance to mesopore blockage due to fast carbonation. Sulfation may also have an effect analogous to the introduction of an inert support into the relatively pure recarbonated sorbent surface; note that CaSO4 has a higher sintering or “Tammann” temperature (861 °C) than CaCO3 (533 °C) and is therefore less susceptible to sintering under carbonation conditions.[21,48] The larger molar volume of CaSO4 relative to CaCO3 may also lead to fractures, introducing defects in the Saabar sorbent’s uniform structure. These structural defects enhance reactant access to active (1, 1, 1) oriented surfaces[49] and reactant permeation into the particle’s core. The introduction of fractures in the CaO matrix due to sulfation is discussed in greater detail in Section .

Influence of Sorbent Cycling on Morphology and Capture Behavior

The influence of calcination/carbonation cycles, performed under practical conditions, on the textural evolution and co-capture behavior of metamorphosed and unmetamorphosed limestone-derived sorbents is examined. The fine Saabar sorbent, with its initial uniform mesoporous structure, is compared to the macroporous polycrystalline fine Riyadh sorbent. The sorbent cycled in a SO2-free atmosphere is compared to the sorbent cycled with SO2 in the carbonation atmosphere, and sorbents with activities X ≈ 0.1, 0.3, and 0.6 mol/mol are compared. We define X as the baseline CO2 capture capacity, when exposed to 12% CO2 in N2, for a sorbent calcined N times. Note that once-calcined limestone activity X1 ≈ 0.6, and sorbent residual activity after innumerable cycles is typically ≈0.07.[15] Since sulfation, predictably,[16] accelerates both metamorphosed and unmetamorphosed limestone-derived sorbent deactivation rates (Figure S14), sulfated and nonsulfated sorbents at different cycle numbers but equivalent activities are compared. This allows the comparison of textural changes experienced by sulfated and nonsulfated sorbents that result in similar activities but different co-capture behavior. Nonsulfated sorbents at N = 13 and sulfated sorbents at N = 6 have an activity, X ≈ 0.1; nonsulfated sorbents at N = 6 and sulfated sorbents at N = 3 have an activity, X ≈ 0.3.

Figure shows the CO2 and SO2 capture behavior of fine Saabar and Riyadh sorbents at different cycle numbers and activities, when the sorbents are recarbonated in BFB 2 with an influent gas composition of 12% CO2, 10% H2O, 2000 ppm SO2, and 3% O2, in N2 (hereafter referred to as “wet synthetic flue gas”). Similar trends are observed at all examined space times; results from τo = 4.7 min are presented. For the unmetamorphosed Riyadh limestone-derived sorbents, the deactivation mechanism does not significantly impact CO2 capture behavior from a gas atmosphere typical of coal or heavy fuel oil power plant flue gas; the nonsulfated and sulfated Riyadh sorbents with similar activities exhibit similar CO2 capture behavior (see Figure a). A slight reduction in fast stage (X < Xmax) capture efficiency is observed for the sulfated Riyadh sorbent versus the nonsulfated Riyadh sorbent with similar activity, with the divergence in capture efficiency growing with CaSO4 accumulation. A slight increase in slow stage (X > Xmax) capture efficiency and conversion is also observed when the Riyadh sorbent is deactivated by sulfation. This is due to the introduction of CaSO4 impurities, which increases lattice defects in the sorbent’s matrix and in turn enhances solid-state diffusion.[50] Unmetamorphosed Riyadh limestone-derived sorbent sulfation follows a similar trend to carbonation behavior as a result of the direct sulfation pathway (see Figure c).

Figure 10

Influence of sorbent activity and CaSO4 content on the CO2 capture performance of (a) Riyadh and (b) Saabar sorbent and on the SO2 capture performance of (c) Riyadh and (d) Saabar sorbent when exposed to wet synthetic flue gas in BFB 2 (τo = 4.7 min).

Sulfated metamorphosed Saabar limestone-derived sorbents outperform nonsulfated Saabar sorbents with similar activity in terms of co-capture efficiency and, in the case of X ≈ 0.3 sorbents, co-capture capacity, indicating that metamorphosed limestone sulfation significantly impacts the sorbent’s textural evolution (see Figure b). The superior slow stage capture efficiency and conversion exhibited by the sulfated Saabar sorbent is analogous to that observed for the sulfated Riyadh sorbent. The superior fast stage CO2 capture efficiency exhibited by sulfated Saabar sorbents indicates that these sorbents’ capture rates are less diffusion-controlled in a carbonation environment containing H2O than nonsulfated metamorphosed limestone-derived sorbents with similar activity. The superior co-capture capacity of the sulfated Saabar sorbent with X ≈ 0.3 indicates that this sorbent is less susceptible to pore blockage in a carbonation environment containing H2O than the nonsulfated Saabar sorbent with X ≈ 0.3. Metamorphosed Saabar limestone-derived sorbent sulfation follows a similar trend to carbonation behavior as a result of the direct sulfation pathway (see Figure d).

Figure presents the evolution of the pore size distributions of cycled fine Riyadh and Saabar sorbents, and BET surface area and other relevant textural data are provided in Table . Sintering, experienced with increasing calcination/carbonation cycles, causes surface area reduction and changes pore size distribution and porosity.[20] Contrary to Pinheiro et al.’s observations,[27] we find that the metamorphosed and unmetamorphosed limestone-derived sorbents experience a comparable decay in the activity (Figure S14) and surface area. Note that mild calcination conditions were employed in Pinheiro et al.’s study.[27] While the high crystallinity associated with Saabar limestone is expected to reduce susceptibility to sintering,[51] PXRD reveals that although the calcined Saabar’s CaO phase maintains a high crystallinity with cycling, its recarbonated CaCO3 phase loses crystallinity with the cycle number (Table S3).

Figure 11

Evolution of pore size distributions for cycled fine (a) Riyadh and (b) Saabar sorbents.

Porosity and pore size distributions for the sulfated and nonsulfated unmetamorphosed Riyadh limestone-derived sorbents with X ≈ 0.3 are very similar. The sulfated Riyadh sorbent that has undergone the same number of calcination/carbonation cycles as the nonsulfated Riyadh sorbent, N = 6, is relatively more sintered; the sulfated N = 6 sorbent has larger peak pore diameters, a smaller surface area, and lower porosity than its counterpart. These textural changes are primarily due to enhanced sintering during calcination caused by the lower Tammann temperature of CaSO4 (861 °C) versus CaO (1313 °C).[48] The enhanced sintering and CaSO4 deactivation are responsible for sulfated Riyadh sorbent’s low activity at N = 6 (X6w/SO ≈ 0.1). The nonsulfated Riyadh sorbent with X13 ≈ 0.1 has a similar surface area to its sulfated counterpart (X6w/SO ≈ 0.1) but lower porosity. Itskos et al. have previously reported on sulfation, especially direct sulfation, enhancing sorbent porosity.[46] While both the X ≈ 0.1 sulfated and nonsulfated unmetamorphosed limestone-derived sorbent pore diameters peak at ∼210 nm, the nonsulfated fraction has a secondary pore diameter peak emerging at ∼90 nm (see Figure a).

The emergence of a population of small pores has been attributed to sorbent regeneration, which is experienced by sorbents subjected to an extended carbonation that surpasses the fast reaction stage and allows recarbonation of previously occluded CaO.[51] During sorbent cycling in BFB 1, the carbonator space time and residence time result in a carbonator conversion Xcarb ≈ 0.15. Since Riyadh sorbent X12 < 0.15, the carbonation reaction proceeds into the solid-state diffusion-controlled stage, leading to some sorbent reactivation postcalcination. The presence of occluded active CaO with cycling is corroborated by PXRD, which reveals that crystalline CaO remains available after N = 12 sorbent recarbonation (Table S4). While sulfation deactivation is mainly responsible for the worse co-capture performance of the sulfated Riyadh sorbent with X13 ≈ 0.1 compared to the nonsulfated Riyadh sorbent with X13 ≈ 0.1 (Figure a), enhanced diffusional resistance to SO2 molecules due to the population of narrow pores emerging in nonsulfated Riyadh sorbent with N = 13, and this sorbent’s lower porosity may also play a role.

For the nonsulfated metamorphosed Saabar limestone-derived sorbent with X13 ≈ 0.1, the bimodal pore size distribution is more developed, with a large population of ∼50 nm pores and secondary populations of pores with diameters >250 nm (see Figure b). While X12 for the metamorphosed and unmetamorphosed limestone-derived sorbents is comparable, the enhanced regeneration experienced by the metamorphosed limestone-derived sorbent indicates that the nonsulfated Saabar sorbent retains its susceptibility to pore blockage in the presence of H2O. This premature pore blockage lengthens the solid-state diffusion-controlled stage thereby amplifying the population of small pores that emerge postcalcination. Note that the metamorphosed limestone-derived sorbent mesopores disappear with initial cycling and larger pore diameters evolve, peaking at ∼110 nm for the nonsulfated Saabar sorbent at N = 6. While nonsulfated Saabar sorbent’s pores are largely macroporous by N = 6, they remain <150 nm and are therefore susceptible to pore blockage. Additionally, nonsulfated Saabar sorbent’s negligible pore volume associated with pore diameters >200 nm indicates that negligible fractures or defects are present in the large dense metamorphosed limestone-derived sorbent’s monomineralic CaO structure, enhancing susceptibility to pore blockage. While once-calcined Riyadh and German limestones’ pore diameters also peak at <150 nm, these unmetamorphosed polycrystalline limestone-derived sorbents have pores with diameters >200 nm, have lower surface reactivity, and therefore do not experience pore blockage in the presence of H2O.

While the sulfated metamorphosed Saabar limestone-derived sorbent with X3w/SO ≈ 0.3 has a similar pore size distribution and activity as the nonsulfated Saabar sorbent with X6 ≈ 0.3, the sulfated Saabar sorbent’s higher porosity and possibly its slightly lower purity lead to reduced susceptibility to pore blockage. This translates to the higher co-capture efficiency and capacity exhibited by the sulfated Saabar sorbent when exposed to wet synthetic flue gas in BFB 2 (Figure b). The sulfated Saabar sorbent with X6w/SO ≈ 0.1 retains a higher porosity and larger average pore diameter than the nonsulfated Saabar sorbent with X13 ≈ 0.1. The sulfated Saabar sorbent’s textural quality enhances pore diffusion, causing initial fast stage carbonation to be less diffusion-controlled relative to that experienced by the nonsulfated Saabar sorbent with X13 ≈ 0.1. This is responsible for the sulfated Saabar sorbent’s superior CO2 capture efficiency when exposed to wet synthetic flue gas in BFB 2 (Figure b). Note that the sulfated and nonsulfated metamorphosed Saabar sorbents with X ≈ 0.1 exhibit similar capture capacity when exposed to wet synthetic flue gas in BFB 2. The emergence of large pores, >250 nm, in the N = 13 nonsulfated metamorphosed limestone-derived sorbent reduces this sorbent’s susceptibility to pore blockage.

The sulfated metamorphosed Saabar limestone-derived sorbent at N = 6 also has a higher porosity than the nonsulfated Saabar sorbent at N = 6. The higher porosity of the sulfated fraction is partially due to the emergence of large pores, >250 nm, analogous to those seen for the nonsulfated Saabar sorbent at N = 13. SEM imaging reveals that sulfation significantly enhances metamorphosed limestone-derived sorbent grain coarsening, favoring this mechanism of sintering over densification, which is seen in nonsulfated metamorphosed limestone-derived sorbents (see Figure a,b). Lattice defects introduced by sulfation have been reported[52] to accelerate lattice diffusion, grain coarsening, and pore diameter widening. Grain boundary stresses due to the accelerated textural changes caused by sulfation may lead to fractures by N = 6. In Figure c, a SEM image of the recarbonated sulfated metamorphosed limestone-derived sorbent is presented, and fractures are seen to propagate from phase boundaries with sorbent impurities. These large fractures play a similar role to the once-calcined coarse metamorphosed limestone-derived sorbent macropores, thereby enhancing co-capture from flue gas containing H2O. Moreover, reactivation experienced by the sulfated metamorphosed Saabar limestone-derived sorbent leads to the emergence of a secondary pore population peaking at ∼90 nm, similar to that seen for the Riyadh sorbent, versus ∼50 nm for the nonsulfated Saabar sorbent. Sulfation alters the metamorphosed limestone-derived sorbent morphology leading it to more closely resemble unmetamorphosed limestone-derived sorbents. Sulfation also preserves the porosity of the metamorphosed limestone-derived sorbent, and for metamorphosed limestone-derived sorbents with initially low porosity, high surface reactivity, and relatively small pore sizes, the preservation of porosity reduces susceptibility to CO2 diffusional resistance and pore blockage and leads to enhanced co-capture performance in an atmosphere containing H2O.

Figure 12

SEM images of the cycled Saabar sorbent: (a) nonsulfated with X13 ≈ 0.1, (b) sulfated with X6w/SO ≈ 0.1, (c) recarbonated sulfated sorbent with X6w/SO ≈ 0.1. White arrows indicate impurities.

The influence of different reaction atmospheres on unmetamorphosed limestone-derived sorbents’ capture behavior is relatively well understood. Unmetamorphosed limestone-derived sorbents in early calcination/carbonation cycle stages are typically polycrystalline and macroporous. These sorbents exhibit limited pore blockage due to carbonation, enhanced capture efficiency, and capture capacity in the presence of H2O and deactivation and pore blockage due to sulfation. In an atmosphere containing both H2O and SO2, the positive influence of H2O counteracts the negative influence of SO2 on capture capacity by limiting SO2 intrusion into the sorbent pores. Calcined metamorphosed limestones’ high purity CaO arranged in large structured and dense monomineralic phases, with relatively low porosity and small pores, is likely responsible for the atypical capture behavior displayed by these sorbents in response to different gas atmospheres (see Figure ). We propose that metamorphosed limestone-derived sorbents’ small pores are susceptible to pore blockage during carbonation. H2O addition further narrows the sorbents’ pores and increases reactivity with CO2 enhancing pore blockage. Limited sulfation deactivates sections of the sorbents’ surface and introduces some resistance to pore blockage. In an atmosphere containing both H2O and SO2, limited sulfation counteracts the enhanced reactivity caused by H2O addition and reduces its negative influence. The presence of relatively large macropores (>200 nm) in certain metamorphosed limestone-derived sorbents dampens the influence of the smaller pores on capture behavior. Furthermore, metamorphosed limestone-derived sorbent sulfation transforms the sorbent upon subsequent recalcination leading sulfated metamorphosed limestone-derived sorbents to behave more similarly to unmetamorphosed limestone-derived sorbents (see Figure ).

Figure 13

Illustration of the influence of reaction atmosphere gas composition and sorbent morphology and composition on capture behavior. The dashed lines illustrate the relative unreacted pore volume of the metamorphosed limestone-derived sorbent due to the different reaction environments.

Figure 14

Influence of sorbent activity on CO2 capture performance when exposed to wet synthetic flue gas in BFB 2 (τo = 4.7 min): (a) nonsulfated and (b) sulfated sorbents. Solid lines: X ≈ 0.1, dashed lines: X ≈ 0.3, dotted lines: X ≈ 0.6.

The different textural evolutions experienced by the unmetamorphosed Riyadh limestone versus the metamorphosed Saabar limestone result in the nonsulfated Saabar sorbent underperforming in terms of co-capture efficiency from wet synthetic flue gas when compared to the nonsulfated Riyadh sorbent (see Figure a). While the nonsulfated Saabar sorbent retains a higher surface area with cycling, nonsulfated Saabar and Riyadh sorbents exhibit similar CO2 capture capacities during co-capture from wet synthetic flue gas. The sulfated metamorphosed and unmetamorphosed limestone-derived sorbents perform similarly, with the Saabar sorbent exhibiting a slightly higher CO2 capture capacity and efficiency with increasing cycle number (see Figure b). This indicates that while metamorphosed limestone-derived sorbent’s susceptibility to CO2 diffusional resistance and pore blockage in the presence of H2O is carried through to sorbents with X ≈ 0.1 when cycled in a SO2-free environment, sorbent sulfation counteracts these propensities.

From a practical perspective, while metamorphosed limestone may not be preferred for CO2 capture from prescrubbed flue gas sources, this limestone can be used in CO2 and SO2 co-capture CaL systems. Metamorphosed limestone attrition and fragmentation rates should be considered when assessing this sorbent’s promise. Preliminary analysis indicates that the Saabar sorbent may be more susceptible to both attrition and fragmentation. While little change to the PSDs within BFB 1 are observed for any of the cycled sorbent fractions, about 15–20% more fines were collected from the reactor cyclones when Saabar was cycled versus Riyadh. Furthermore, while gas analyzer data indicate that the metamorphosed and unmetamorphosed limestone-derived sorbents capture the same amount of SO2 during cycling, elemental analysis reveals that the metamorphosed limestone-derived sorbent has a lower CaSO4 content (see Table ). The CaO purity of the Saabar sorbent also decreases with the cycle number (Table S2). This indicates that the Saabar sorbent is more prone to attrition than the unmetamorphosed limestone-derived sorbent, with the newly deposited surface CaSO4 especially vulnerable to attrition. Saabar limestone’s slightly elevated attrition rate may therefore be favorable under co-capture conditions.

Conclusions

This manuscript reports on the CO2 and SO2 capture behavior of metamorphosed and unmetamorphosed limestone-derived sorbents. The individual and combined influence of flue gas H2O and SO2 content on sorbent CO2 and SO2 capture performance is investigated. The influence of multiple calcination/carbonation cycles and of CaSO4 accumulation on the limestone-derived sorbents’ textural evolution and capture behavior are also examined. While we find that the unmetamorphosed limestone-derived sorbents’ capture performance aligns well with previous work, the metamorphosed limestone-derived sorbent exhibits anomalous behavior. Our experimental results reveal the following:

Contrary to expectations, calcined metamorphosed limestone CO2 capture performance is negatively influenced by flue gas H2O content and positively influenced by flue gas SO2 content.

The negative impact of flue gas H2O content on metamorphosed limestone-derived sorbent CO2 capture performance persists for sorbents with activities ranging from 0.6 to 0.1 mol/mol, when the sorbents have been cycled in a sulfur-free flue gas.

Metamorphosed limestone-derived sorbent that has been cycled in sulfated flue gas exhibits similar capture behavior to unmetamorphosed limestone-derived sorbents.

Space time has a significant impact on SO2 and CO2 co-capture performance.

Analyzing the limestone-derived sorbents using material characterization and imaging tools, we conclude that the distribution of impurities in the metamorphosed limestone-derived sorbent matrix is ultimately responsible for the sorbent’s capture behavior. The metamorphosed limestone is primarily composed of large monomineralic phases separated by clear boundaries. The large, relatively pure, and highly crystalline monomineralic CaCO3 grains in metamorphosed limestone are responsible for the evolution of relatively narrow uniform pores and high purity crystalline CaO with low overall porosity upon calcination. These morphological characteristics are likely responsible for the anomalous capture behavior of the metamorphosed limestone-derived sorbent. We postulate that the calcined metamorphosed limestone’s high CaO purity, narrow pores, and low porosity translate to higher surface reactivity and increased susceptibility to pore blockage during carbonation. As a result, flue gas H2O content, known to enhance CaO reactivity with CO2 and cause pore narrowing, exacerbates metamorphosed limestone-derived sorbent pore blockage. Furthermore, we propose that the introduction of a limited amount of impurities in the form of CaSO4 may pacify sections of the metamorphosed limestone-derived sorbent’s highly reactive matrix and introduce some resistance to mesopore blockage due to fast surface carbonation. We also find that

Cycling metamorphosed limestone in an atmosphere containing SO2 preserves the sorbent’s porosity and increases the sorbent’s grain and pore sizes, resulting in sulfated metamorphosed and unmetamorphosed limestone-derived sorbents exhibiting similar capture performance.

The influence of calcined metamorphosed limestones’ narrow pores on capture behavior is dampened in sorbents with a secondary population of large macropores (>200 nm).

Our findings indicate that the metamorphosed limestone may be more prone to attrition and fragmentation than the unmetamorphosed limestone. The mechanical stability of this sorbent requires further investigation. The influence of additional calcination/carbonation cycles and sulfation on the metamorphosed limestone-derived sorbent’s capture behavior also requires further investigation. The understanding gained regarding the underlying morphology that influences the anomalous capture behavior of the metamorphosed limestone-derived sorbent will help guide future sorbent selection and design efforts.