Article pubs.acs.org/IECR
High Temperature Carbonation of Ca(OH)2: The Effect of Particle Surface Area and Pore Volume V. Materic,*,† M. Hyland,‡ M. I. Jones,‡ and B. Northover† †
Advanced Materials Department, Callaghan Innovation, 69 Gracefield Road, 5040 Lower Hutt, New Zealand Department of Chemical and Material Engineering, University of Auckland, 20 Symonds Street, 1142 Auckland, New Zealand
‡
ABSTRACT: The high temperature carbonation of Ca(OH)2 forms part of a number of proposed schemes for CO2 capture and thermal energy storage. This reaction exhibits a number of unusual features under some conditions such as Ca(OH)2 superheating, i.e., the inhibition of the dehydration of Ca(OH)2. In this work the effect of the surface area and pore volume of Ca(OH)2 particles on the conversion and kinetics of high temperature carbonation is investigated. A range of Ca(OH)2 particles was prepared by hydration of CaO and then subjected to carbonation at high temperatures in nonisothermal thermogravimetric analysis experiments. Surprisingly, the surface area and pore volume of the sorbent, as measured by N2 adsorption, had little effect on the conversion and kinetics of carbonation. In contrast, macroporosity associated with pores > 100 nm was postulated to have a significant effect on carbonation conversion and kinetics, possibly due to a pore-plugging mechanism.
1. INTRODUCTION The aim of this work was to investigate the relationship between the surface area and porosity (pore volume) of Ca(OH)2 particles and the kinetics of their carbonation by CO2 at high temperature (>200 °C). This reaction can be expected to proceed via two distinct routesa “direct” carbonation route (eq 1) or an “indirect” route via the formation of a stable CaO phase intermediary (eqs 2 and 3) since Ca(OH)2 is known to decompose at temperatures above 400 °C (eq 2). Ca(OH)2 + CO2 = CaCO3
(1)
Ca(OH)2 = CaO + CO2
(2)
CaO + CO2 = CaCO3
(3)
Gaseous CO2 and solid Ca(OH)2 could dissolve in this layer, facilitating the recombination of the resulting Ca2+ and CO32− ions into precipitated CaCO3. With an RH of 0.88, Beruto and Botter achieved near total conversion of Ca(OH)2 to CaCO3 in 12 h at ambient temperature.16 In contrast, this study found that, in the absence of water vapor, the reaction was exceedingly slow and the conversion to CaCO3 limited. At 80 °C, under dry conditions (RH < 0.005) they reported a conversion of only 12% in 48 h.16 In the temperature range considered in this work (>200 °C), relative humidity is inherently low so that the carbonation of Ca(OH)2 cannot be expected to proceed via the water-catalyzed route described above. Recent kinetic studies have reported observing relatively rapid carbonation of Ca(OH)2 under isothermal conditions at temperatures between 200 and 425 °C,1,18,19 suggesting that a direct solid−gas carbonation mechanism (eq 1) is active at these temperatures since dehydration of Ca(OH)2 (eq 2) is not expected to occur at least until ≈400 °C. The rate and extent of the direct carbonation of Ca(OH)2 were found to be greater than those of CaO under identical conditions, which Nikulshina et al. attributed to a selfcatalyzing effect due to the release of water from Ca(OH)2.1 The focus of this work is the considerable differences in the kinetics of the high temperature carbonation of Ca(OH)2 that were revealed by nonisothermal experiments, both in terms of the carbonation conversion and the pattern of water release.6,13,14 With some materials, a near complete conversion to CaCO3 was observed before the material reached 560 °C and a single water release peak was observed, centered at ≈420 °C.14,19 With other materials, the conversion to CaCO3 at 560 °C was lower (20−60%) and a second water release peak was observed if CO2 was removed at 560 °C6,13 or heated further to 620 °C.13,14
The high temperature carbonation of Ca(OH)2 forms a part of a number of proposed schemes for CO2 capture and for thermal energy storage. So far, this reaction has been proposed as part of a Ca(OH)2−CaCO3−CaO chemical cycle for the separation of CO2 from atmospheric air,1 the storage of thermal solar energy,2 or CO2 capture from flue gases.3 In addition, this reaction is also used in a number of strategies aiming to increase the reactivity of CaO-based sorbents for Ca Looping systems.4−6 In these systems, CaO-based sorbents are subjected to cyclic carbonation/calcination (eq 3) allowing the separation, or capture, of CO2 from gas streams.7,8 However, repeated carbonation/calcination cycling also leads to a reduction in surface area and an increase in pore size of the sorbent due to the sintering occurring at each calcination step, which, in turn, leads to a progressive loss of reactivity of CaO sorbents toward CO2.9−12 Hydrating unreactive CaO to form Ca(OH)2 which is then carbonated at high temperature was shown to restore sorbent reactivity for carbonation.6,13,14 At low temperatures, the direct carbonation of Ca(OH)2 (eq 1) was found to be catalyzed by the presence of water vapor above a relative humidity (RH) of ≈0.6.15−17 It was postulated that liquid-like layers of water adsorbed from the atmosphere were formed and acted as a reaction interface. © 2014 American Chemical Society
Received: Revised: Accepted: Published: 2994
September 12, 2013 December 17, 2013 January 27, 2014 January 27, 2014 dx.doi.org/10.1021/ie403005s | Ind. Eng. Chem. Res. 2014, 53, 2994−3000
Industrial & Engineering Chemistry Research
Article
flow of N2 at 905 °C for 15 min, while carbonation reactions were performed in 37.5% CO2/N2 at 600 °C. Due to the absence of a cooling system, the heat released by the carbonation reaction would raise the bed temperature to approximately 640−660 °C. Once the cycling was completed, a final calcination was performed and the cycled CaO was discharged from the bed and kept in an airtight container to minimize interactions with atmospheric humidity and CO2. Samples of cycled CaO, thus produced, were then hydrated using three different techniques. Liquid Water Hydration. Liquid water hydration was performed by contacting a 4 g sample of cycled CaO with 5 mL of water (1:4 CaO/H2O molecular ratio) in a stirred beaker at room temperature for 30 min. No effort was made to control the temperature increase of the mixture as the reaction progressed. However, the beaker was covered with aluminum foil to minimize water evaporation. During liquid water hydration many CaO particles were pulverized, leading to a reduction in the average particle size. Nevertheless, the sorbent remained distinctly particulate. Immediately after the hydration step, the hydrated sample was dried overnight in a 140 °C oven prior to heating in CO2 and surface area measurements. Oven Steam Hydration. Oven steam hydration was performed by placing a glass vial containing a 1.5 g sample of cycled CaO in a loosely capped jar containing water which was then placed inside a laboratory oven. Hydration was performed at two different temperatures, at 90 and 120 °C, where the hydrating atmosphere contained saturated and unsaturated steam, respectively. Indeed, at 90 °C the equilibrium water vapor pressure is below 100 kPa so that the relative humidity in the loosely capped jar was at 100% and the jar was thus filled with saturated (or wet) steam, causing observable condensation on the inner walls of the vial containing the sample. In contrast, at 120 °C the equilibrium water vapor pressure is above 1 atm so that relative humidity was below 100% and the loosely capped jar was filled with unsaturated (or dry) steam, as confirmed by the observed absence of condensation on the inner walls of the vial. The progress of the hydration reaction was monitored by the expansion of the solid inside the glass vial using engraved distance markings, and sufficient time was allowed in each case to maximize the hydration conversion. The final conversion to Ca(OH)2 was evaluated by the weight gain after drying the hydrated sample. 2.2. Testing Methods. Surface Area and Porosity Measurements. The specific surface area and pore volume of different CaO and Ca(OH)2 samples were measured by N2 adsorption. Adsorption isotherms were measured on ≈1 g samples using a Micrometrics ASAP-2010 apparatus with helium as the carrier gas. Automated Brunauer−Emmett−Teller (BET) calculations were used to calculate surface area and pore volume of the samples. Note that due to the limitations of this technique, the volume of pores larger than ≈100 nm is typically not detected. High Temperature Carbonation Tests. In a TGA apparatus, 20 mg samples of Ca(OH)2 were first dried at 200 °C for 20 min and then heated in 100% CO2 at 15 °C.min−1 up to 520 °C. Once that temperature was reached, superheated dehydration was triggered by replacing CO2 by N213 and holding the temperature at 520 °C for 20 min. Finally, the remaining mixture of CaCO3/CaO was calcined by heating to 900 °C in N2. A mass vs time profile of one such experiment is plotted in Figure 1. The composition of the sample at 520 °C was then calculated with the knowledge of the mass of H2O (as Ca(OH)2) which is
This loss of water was attributed to the dehydration of Ca(OH)2 remaining in the sample at that point. Given that the normal dehydration temperature for Ca(OH)2 is usually ≈400−420 °C, this Ca(OH)2 is said to have been “superheated” and the loss of water is referred to as “superheated dehydration”.6,13 The reported differences in carbonation kinetics of the different Ca(OH)2 samples have been attributed to their morphology properties, i.e., surface area, porosity, and pore size distribution. Blamey et al. reported that increased particle sizes lead to a lower carbonation conversion and increased superheated dehydration effect.14 In contrast, powdered Ca(OH)2 samples achieved near full conversion and did not exhibit the superheated dehydration effect.14 They postulated that the formation of an impervious layer of CaCO3 on the surface of Ca(OH)2 had the potential to block the escape of H2O from the center of the particles and thus lead to the superheated dehydration effect.14 In contrast, Materić and Smedley proposed that the superheated dehydration effect was due to the interaction of gaseous CO2 with a water film formed on the surface of Ca(OH)2.13 Interestingly, in all cases an unusual kinetic behavior was observed when Ca(OH)2 was heated in the presence of CO2 the conversion to CaCO3 was found to be independent of the contact time between Ca(OH)2 and CO2 and instead be closely correlated with the temperature of the sample.13,19 In nonisothermal carbonation experiments, Montes-Hernandez et al. reported identical conversion vs temperature curves with varying heating rates (5 and 10 °C·min−1)19 while Materić and Smedley reported that using a sawtooth heating profile did not significantly affect the conversion vs temperature curves.13 The object of this work is to assess the effect of surface area and porosity (pore volume) of Ca(OH)2 particles on the kinetics and conversion of their high temperature carbonation in nonisothermal experiments. The specific surface area of particles can be expected to have a significant effect on the kinetics of carbonation and thus ultimately on the conversion achievable during the experiment. In contrast, particle porosity can be expected to have only a limiting effect on conversion due to the expansion of the product CaCO3 with a density of 37 cm3·mol−1 vs 33 cm3·mol−1 for Ca(OH)2. Namely, the reaction could not proceed further if the expanded CaCO3 product filled all the pore volume available; such a mechanism is referred to as pore plugging. In this study, a range of Ca(OH)2 particles of widely varying surface area and pore volume were prepared from raw limestone by subjecting it to a number of calcination/carbonation cycles and hydrating the resulting CaO with a variety of hydration techniques. Thus prepared, the Ca(OH)2 samples were heated to 520 °C in the presence of CO2 in a thermogravimetric analysis (TGA) apparatus allowing measurement of the carbonation conversion and the extent of the superheated dehydration effect achieved during the heating period.
2. MATERIALS AND METHODS 2.1. Sample Preparation. Raw Limestone. In this work a highly calcitic limestone (97% CaCO3) was used as the starting material for the preparation of Ca(OH)2 samples. This material was obtained from Taylor’s, Te Kuiti, New Zealand, and sieved to 300−600 μm prior to use. Calcination/Carbonation Cycles. Batches of 100 g of raw limestone were subjected to a varying number of calcination/ carbonation cycles (1, 7, and 37) in a bubbling fluid bed described elsewhere.6 Calcination reactions were performed in a 2995
dx.doi.org/10.1021/ie403005s | Ind. Eng. Chem. Res. 2014, 53, 2994−3000
Industrial & Engineering Chemistry Research
Article
several zones of two to three different particles and images were taken at different scales.
3. RESULTS 3.1. Surface Area and Porosity of Cycled CaO. N2 adsorption isotherms of spent sorbent samples, i.e., cycled CaO prior to hydration, were measured and the results of BET calculations are shown in Table 1. As expected, the surface area and pore volume are considerably reduced with increasing cycle numbers.9−12 It is worth noting at this point that the N2 adsorption technique is unable to detect pores larger than ≈100 nm and that extensively cycled spent sorbent samples (CaO) are known to develop a substantial number of pores in the 400−600 nm pore size range.12,20 The total volume of large pores (400−600 nm) was reported to grow with cycling number and exceed the volume available in the smaller pores after the 29th cycle. Given their size, such large pores are not expected to significantly contribute to the surface area of the samples so that the surface area values measured by N2 adsorption are assumed to be correct. However, such large pores do contribute to the porosity of the sample so that pore volume measurements obtained here via N2 adsorption underestimate the porosity of the more cycled samples (7 and 37 cycles) which are expected to have a substantial number of large pores (400−600 nm).12,21 The existence of such pores in the interior of spent sorbent particles was observed using SEM; see Figure 2. In the case of the once-cycled sample, CaO grains appear composed of an agglomeration of small crystallites leaving between them a network of small pores, Figure 2a. In contrast, in the 37-cycled sample, the crystallites are considerably larger and the existence of large pores (>100 nm) can be observed, Figure 2b. 3.2. Hydration Conversion and Times. As illustrated in Table 1, hydration conversions of ≈90% could be achieved with all the hydration methods employed. However, the time required
Figure 1. Mass vs temperature profile of a nonisothermal high temperature carbonation experiment. Note that sample mass is expressed as a percentage of a fully carbonated sample.
lost during superheated dehydration, the mass of CO2 (as CaCO3) lost during calcination, and the mass of CaO remaining after calcination; see Figure 1. SEM Analyses. Due to the inability of the N2 adsorption method to detect pores larger than ≈100 nm, the existence of such pores was investigated using scanning electron microscopy (SEM) in both spent (CaO) and 120 °C steam-hydrated (Ca(OH)2) sorbent samples. Sample particles were cut with a razor blade and metal coated with Au/Pd for 120 s. SEM images were taken on a Quanta 450 scanning electron microscope equipped with a tungsten filament. The secondary electron images were acquired by operating the SEM at 20 keV using a spot size of 3. The SEM beam was pointed at clean fracture surfaces. It was possible to identify such areas since outer surfaces of the particle were distinguishable from inner surfaces as they were covered with residual fines from the sieving process. For each sample, the beam was focused on
Table 1. Surface Area and Porosity Measurements and Hydration Time for Cycled CaO Samples As a Function of the Number of Calcination/Carbonation Cyclesa 1 cycle 7 cycles 37 cycles a
surface area (m2·g−1)
porosity (cm3·g−1)
saturated steam (90 °C)
unsaturated steam (120 °C)
8.1 3.8 1.8
0.08 0.03 0.01
92% in 170 min 90% in 200 min 92% in 240 min
88% in 35 min 90% in 45 min 91% in 45 min
Note: Times for hydration in liquid water are not reported because they were kept constant at 30 min.
Figure 2. SEM image of a fracture surface of (a) a once-cycled CaO particle and (b) a 37-cycled CaO particle. Scale bar: 5 μm. 2996
dx.doi.org/10.1021/ie403005s | Ind. Eng. Chem. Res. 2014, 53, 2994−3000
Industrial & Engineering Chemistry Research
Article
Figure 3. SEM image of a fracture surface of (a) a once-cycled, steam-hydrated Ca(OH)2 particle and (b) a 37-cycled, steam-hydrated Ca(OH)2 particle. Scale bar: 4 μm.
Table 2. Surface Area, Porosity, and Hydration Conversion of Ca(OH)2 Samples Prepared in This Work surface area (m2·g−1)
porosity (cm3·g−1)
hydration conversion (%)
34.4 10.6 11.3
0.2 0.1 0.09
85 93 88
32.9 9.5 7.5
0.2 0.07 0.06
87 92 85
25.5 7.2 8.1
0.16 0.06 0.06
88 93 92
1 cycle liquid 90 °C steam 120 °C steam 7 cycles liquid 90 °C steam 120 °C steam 37 cycles liquid 90 °C steam 120 °C steam
to reach the final hydration conversion level varied considerably and increased with decreasing surface area of the CaO sample. It took approximately 35 min to achieve maximum conversion for the once-cycled CaO in unsaturated steam, while 240 min was required to hydrate the 37-cycled CaO. Surface area and porosity measurements of different Ca(OH)2 samples produced in this work are reported in Table 2. Hydration of cycled CaO samples led to an increase in both the surface area and porosity of the particles; the most marked increase was observed with the 37-cycled CaO; compare Table 1 and Table 2. Samples hydrated with liquid water exhibited the highest surface area and porosity while these were similar for both steamhydrated sorbents. As discussed in preceding text, the limitations of the N2 adsorption technique suggest that the measured pore volume might be underestimated for the cycled and hydrated sorbents, assuming that hydrating the spent sorbents did not close the large pores initially present in the cycled CaO samples. The existence of such pores in the interior of spent sorbent particles, hydrated in steam at 120 °C, was observed using SEM; see Figure 3. Similarly to the parent CaO particles, the Ca(OH)2 grains appear as an agglomeration of crystallites which remain considerably smaller in the case of the once-cycled, steamhydrated sample. Similarly to the parent CaO particles, a network of large pores (420 °C, which were similar for all three samples. Therefore, it can be concluded that the surface area available in pores measurable by N2 adsorption, i.e.,
