CaCO3 Crystallite Evolution during CaO Carbonation: Critical

Oct 12, 2015 - Synopsis. From the analysis of the CaCO3 crystalline phase evolution during the CaO−CO2 reaction (CO2 capture), the concept of critic...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/crystal

CaCO3 Crystallite Evolution during CaO Carbonation: Critical Crystallite Size and Rate Constant Measurement by In-Situ Synchrotron Radiation X‑ray Powder Diffraction A. Biasin,† C. U. Segre,‡ and M. Strumendo*,† †

Department of Industrial Engineering, University of Padova, Via G. Gradenigo 6/a, Padova, 35131, Italy Department of Physics & CSRRI, Illinois Institute of Technology, 3101 S. Dearborn Street, Chicago, Illinois 60616, United States



S Supporting Information *

ABSTRACT: In this work, the evolution of the CaCO3 crystalline phase during the CaO−CO 2 reaction was investigated by means of in situ synchrotron radiation X-ray powder diffraction performed at the Advanced Photon Source (APS) facilities of the Argonne National Laboratory. CO2 absorption experiments were carried out in a high temperature reaction capillary with a controlled atmosphere of pure carbon dioxide (CO2 partial pressure of 1 bar) and in the temperature range between 450 and 750 °C, using CaO-based sorbents obtained by calcination of commercial calcium carbonate. The Rietveld refinement method was applied to estimate the average size of the CaCO3 crystallites formed during a carbonation time of 20 min, as a function of the carbonation temperature and of the initial calcination conditions. Local maxima were observed in the CaCO3 crystallite size versus time curves and were identified as the critical CaCO3 crystallite sizes, marking the transition between the first fast carbonation stage and the second reaction stage controlled by product-layer diffusion. A relationship between this parameter and the reaction temperature, as well as with the initial (at the beginning of carbonation) CaO crystallite size, were found. The CaCO3 critical crystallite sizes were used to estimate the active surface areas of the CaO sorbent particles where CaCO3 crystals form and grow. The computed active surface areas were utilized to calculate the kinetic parameters of the surface carbonation reaction: a reaction rate constant of 4.41 × 10−4 mol/m2 s, with zero-activation energy, was obtained.

1. INTRODUCTION CO2 is the primary anthropogenic greenhouse gas in the atmosphere, and it is widely recognized as the main contributor to global warming and climate change.1−3 Carbon dioxide emissions have increased rapidly due to technological and industrial development, and, in the past decade, the reduction of the CO2 emitted by the power generation industry (mainly from fossil fuel combustion and coal gasification-based power plants) has become an important worldwide scientific goal.2,3 Carbon capture and storage (CCS) is a range of technologies focused on affordably and efficiently removing CO2 from industrial gas streams and on transporting it to storage locations providing long-term CO 2 isolation from the atmosphere.4 Among these technologies, a promising and cost-effective CO2 capture technique under investigation involves the utilization of calcium-based solid sorbents and exploits the reaction of calcium oxide with CO2 to form calcium carbonate under operating conditions (especially at high temperatures) not reachable with other CCS technologies, such as amine-based and ionic liquid scrubbing systems or physical adsorbents like zeolite and metal organic frameworks,5−7 aimed at the sequestration of CO2 from currently existing industrial sources of carbon dioxide. The carbon © XXXX American Chemical Society

dioxide sequestration performed with CaO-based sorbents is widely discussed in the literature and recently has been reviewed by several authors.4,8−10 Calcium-based materials are suitable sorbent candidates for capturing CO2 due to a number of advantages, including a high sorption capacity for CO2 at high temperatures, a low cost of sorbent manufacturing and regeneration, and the abundance of their natural precursors. The chemical reaction CaO(s) + CO2(g) ↔ CaCO3(s) is reversible, and hence this technology is based on cyclic stages of carbonation and of calcination. The improvement of this technology and the development of new calcium-based solid sorbents are currently a matter of study, and, despite the apparent simplicity of the chemistry involved, several aspects of the carbonation reaction are still not clearly understood. It is currently accepted that, when the gas-phase mass transfer resistances can be neglected, the CaO−CO2 reaction is controlled by the surface chemical reaction and by the product layer diffusion.11−15 Diffusion through the solid product layer is relevant because at the interface between the solid and the gas Received: April 22, 2015 Revised: September 9, 2015

A

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

To summarize, thus far several contributions have related the transition from the fast regime to the slow product-layer diffusion controlled regime to the filling of small pores and/or to the development of a critical carbonate layer and have focused attention on the impact of the pore size distribution on the critical CaCO3 product layer thickness. However, an unambiguous value of such critical thickness has not been identified yet, and no direct measure of this parameter has been reported. Over the last 30−40 years several significant contributions were dedicated to the interpretation of the carbonation reaction and kinetics in terms of the sorbent structural properties (porosity, specific surface, structural parameter, or the whole pore size distribution) through the development and application of random pore/grain models.13−15,21,23,24 However, very few contributions have been focused on investigating the impact of CaO/CaCO3 crystallite size on the carbonation reaction with CaO-based solid sorbents. Molinder et al.25 measured the crystallite size of partially hydrated CaO sorbents in a study on simultaneous hydration and carbonation of CaO, partially hydrated CaO and of Ca(OH)2. However, even though the carbonate crystallite growth kinetics can be expected to play a significant role in understanding the kinetics of carbonation (a reaction with a crystalline product layer whose growth is a function of the crystallite growth), the influence of the CaCO3 crystalline domain size on the carbonation reaction and kinetics has never been (to the best of our knowledge) investigated. Another crucial aspect concerning the carbonation reaction is the determination of the surface reaction kinetic parameters (pre-exponential factor and activation energy of the reaction rate constant).13,14,26−28 In the literature, information regarding the activation energy of the carbonation reaction is often contradictory. On the basis of the argument that the activation energy of the calcination reaction is very close to the enthalpy of the reaction, Dennis and Hayhurst29 assumed a zeroactivation energy for the CaO carbonation, confirming the result that Bhatia and Perlmutter13 obtained studying TGA conversion over time data by means of a random pore model. More recently, Sun et al.14 argued that a zero-activation energy is rather rare and performed a detailed study of the carbonation reaction carrying out TGA experiments under a wide range of CO2 partial pressures. Applying a grain model on the early instants of the CaO versus time profiles, Sun et al.14 obtained that the activation energy of the surface CaO−CO2 reaction was small but not zero, namely, 29 kJ/mol. This value is lower than the activation energy of 78 kJ/mol determined by Kyaw et al.27 The quality of the estimated activation energies depends on the accuracy of the experimental data, specifically of the conversion versus time curves. Typical values of the initial and/ or maximum reaction rate of about 0.05 s−1 obtained through thermogravimetric measurements are reported in the literature13,14,27 and were used to compute the reaction rate constants thus far available in the literature. However, recently the authors (Biasin et al.30), using the synchrotron radiation in situ X-ray diffraction technique, measured much higher conversion rates with an average value of 0.280 s−1. Such novel results were obtained thanks to the flow-cell geometry utilized that permits minimization of the external mass diffusion but were not used to recalculate the surface reaction kinetic parameters.

reactant the formation and growth of a solid product layer of CaCO3 occur, making the contact between new molecules of carbon dioxide and a still existing portion of reactant CaO progressively more arduous. As a consequence, the carbonation is characterized by an initial very rapid reaction period, followed by a much slower stage limited by the diffusion of the reacting CO2 through the product layer of CaCO3. The transition between these two stages is sharp and has been related to the formation of calcium carbonate that progressively decreases the CaO reactivity toward the carbon dioxide until the thickness of CaCO3 has reached a critical value, beyond which the diffusioncontrolled process becomes slower as the conversion increases.16 The concept of the critical carbonate product layer thickness at which the reaction becomes controlled by CO2 diffusion was initially introduced by Barker,17 who, assuming a homogeneous distribution of CaCO3 over the entire internal surface area of calcium oxide sorbent particles, and taking into account the expansion of the solid associated with the variation in the molar volume between CaO and CaCO3 (from 16.9 to 36.9 cm3/ mol), estimated a critical carbonate thickness of 22 nm. He related the reactivity of fresh newly formed calcium oxide (mean particle size of about 10 μm) to the formation of small pores (in the range of about 40 nm and below 4 nm), which significantly increase the sorbent surface area. In order to exclude the effects arising from the complex morphology of porous CaO sorbent particles, Mess et al.16 performed carbonation experiments utilizing nonporous crystals of calcium oxide (particle size of about 15−20 μm) and measured uniform product layer thicknesses from about 1 μm up to values comparable with the whole particle size, after long carbonations. On the basis of the Mess et al. data,16 Abanades and Alvarez18,19 computed estimated thicknesses between 130 and 220 nm for the carbonate at the onset of the slow reaction period. As suggested by Barker,17 the CaO−CO2 reaction is sensitive to the pore size distribution of the fresh sorbent particles. The formation and growth of the calcium carbonate solid product cause the obstruction of the particle porous structure by filling the pores, making the diffusion of gaseous CO2 through and inside the particle porosities more difficult. Several other authors (including Bhatia and Perlmutter,13 Alvarez and Abanades,20 Grasa et al.,21 Sun et al.22) investigated this aspect combining pore models with experimentally determined pore size distributions (obtained by mercury porosimetry), CaO conversion profiles (from TGA experiments), and the qualitative analysis of the pore filling by means of optical and scanning electron microscopy. Bhatia and Perlmutter13 attributed the sharp transition from the first rapid regime to the product layer diffusion controlled regime in the carbonation conversion profiles of fresh calcines to the filling of the small pores, with a size below approximately 100 nm. Sun et al.22 applying a gas−solid reaction random pore model based on discrete pore size distribution measurements, proposed that the transition from fast to slow stages of carbonation should be ascribed to the consumption of all the pores smaller than 250−300 nm. Alvarez and Abanades20 computed an average critical product layer thickness of 49 nm (for a wide range of sorbents and carbonation conditions) interpreting with a simple pore model the differences obtained from the comparison of the mercury porosimetry curves of calcines and the carbonated counterparts. More recently, Grasa et al.21 reported critical product layer thicknesses between 30 and 40 nm (corresponding to CaO conversions of 0.55−0.6). B

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

CaCO3 crystalline domains (namely, absence of initial preferred orientations), and hence of a good quality of the diffraction data. 2.2. In-Situ SR-XRD Experiments. The in situ SR-XRPD investigation of the average CaCO3 crystallite size evolution during CaO carbonation was performed at the beamline 17-BM-B of the Advanced Photon Source (APS), at Argonne National Laboratory. The experiments consisted of monitoring X-ray powder diffraction data during a single stage of the carbon dioxide absorption process using CaO sorbents obtained from a preliminary thermal decomposition of CaCO3 samples. Calcium oxide particles produced through calcination stages of the carbonate precursor become porous thanks to the marked decrease in the molar volume that is achieved during the thermally induced CO2 release (theoretical free volume of pores up to 54%), and the resulting increase in specific surface area typically makes CaO particles more reactive toward the following CO2 absorption. The measurements were carried out using a versatile setup especially designed for conducting in situ X-ray diffraction studies in nonambient environments.31 The system is based on a flow cell/ furnace that allows the control of several environmental parameters including temperature (up to 900 °C) and atmosphere/gas flow conditions. A small amount of sample (about 5 mg) was loaded into a 1 mm quartz glass capillary (0.1 mm wall thickness, 75 mm long), filling approximately 3−5 mm of the capillary. The capillary was placed in the flow cell system, in turn mounted on the instrument using a standard goniometer head. Through a set of tubes and fittings, the gas was forced to flow into the capillary, through the sample particles. In order to prevent any change in the position of the powder during the experiments, mainly due to the gas flow, the sample inside the capillary was stabilized by means of quartz wool positioned nearby the powder, at both capillary extremities. The heating was provided by two resistive elements placed immediately above and below the sample capillary. A K-type thermocouple, placed within the capillary in close proximity to the powder, was used both for sample temperature monitoring and control. Experimental data were collected using an amorphous silicon-based flat plate detector from PerkinElmer (model XRD 1621 CN3-EHS, detector size of 2048 × 2048 pixels, pixel size of 200 × 200 μm). The wavelength (λ = 0.72959(3) Å with a focused beam of 300 μm) and the sample-to-detector distance (d = 450 mm, giving a usable range of 2θ of 1−43°) were calibrated using LaB6 as a reference material. The collection of the diffraction data was restricted to the first 20 min of carbonation, which include the first rapid initial period characterizing the CaO−CO2 reaction and the transition to the second slow stage. Specifically, in order to obtain results with a high time scale resolution (also suitable for kinetic studies), the data were recorded continuously with a frame rate of 0.25 s per frame during the first 5 min of the reaction (i.e., 1200 patterns), followed by a frame rate of 1 s per frame for the remaining 15 min (i.e., 900 patterns). The raw images were therefore reduced to one-dimensional diffraction patterns by means of FIT-2D software.32,33 Finally, the data were analyzed by the Rietveld profile refinement method using the General Structure Analysis System (GSAS) software.34 In particular, the sequential profile refinement approach was used to process consecutive analyses (time-resolved measurements) and hence to collect results as a function of the carbonation time. 2.2.1. Preparation of the CaO Samples and Carbonation Tests. In this study commercial calcium carbonate samples were first decomposed in order to produce fresh porous CaO samples to be afterward used to investigate the CaO carbonation. Calcination stages were carried out in N2 atmosphere (at a pressure of 1 bar), with a flow rate of approximately 10 N mL/min and a heating rate of 10 °C/min. The calcination temperature was varied in the range between 700 and 900 °C, and two different residence times were selected, namely, 5 and 60 min. These operational conditions were chosen in order to ensure the complete decomposition of CaCO3 samples, as well as to obtain fresh calcines with different average size of nascent CaO crystalline domains.30 N2 flow within the capillary was used to facilitate the removal of the CO2 released at the reaction zone during the calcium carbonate decomposition away from the solid surface, mainly to reduce its contribution to sintering processes that can affect the sample

Using the tapping mode of an atomic force microscope, Li et al.12 recently observed that the CaCO3 product grows with an island morphology on the CaO surface, and these islands appear not to be uniformly distributed. Specifically, they noted that the sorbent surface is characterized by terraces, steps, and kinks, and hence not all the specific surface area of the sorbent particles (as typically measured by N2 adsorption experiments) seems to be uniformly utilized during the carbonation reaction. However, a quantitative estimate of the active specific surface area was not provided by Li et al.,12 but this information is required in order to correctly estimate the surface reaction kinetic parameters and cannot be computed directly from N2 adsorption measurements. In this work, the CaO−CO2 reaction was studied through in situ X-ray powder diffraction using synchrotron radiation (SRXRPD). Synchrotron radiation was chosen because it produces X-ray data with excellent accuracy and a high time resolution that are essential to investigate rapid transformations such as those occurring during the first stage of the calcium oxide carbonation. The SR-XRPD data of the CaO carbonation were collected at the Argonne National Laboratory Advanced Photon Source. Such SR-XRPD data were previously analyzed by the authors30 in order to compute and monitor the CaO conversion over time and to investigate the dependence of the CaO conversion−time curves on the initial (at the beginning of carbonation) CaO crystallite size. Instead, in this investigation the same SR-XRPD data were used for a different goal, namely, to monitor and analyze the calcium carbonate crystallite size trends over time, and to correlate the calcium carbonate crystallite size time evolution with the CaO conversion.30 The link identified in this work between the CaCO3 crystallite size trends over time, and the CaO conversion curves over time allows a kinetic analysis of the CaO carbonation to be carried out. More specifically, the measurements of CaCO3 crystallite size versus time (a novel result for CaO carbonation) are used to propose a definition of the critical carbonate crystallite size (marking the beginning of the transition between the first fast carbonation stage and the second reaction stage controlled by product-layer diffusion) and a new method to measure such quantity directly. Relationships between the critical CaCO3 crystallite size and both the initial CaO crystallite size and the carbonation temperature are identified as well. Additionally, an equation is developed in this work to compute the active specific surface areas from the critical CaCO3 crystallite size. Finally, such quantities (not available thus far in literature for the CaO carbonation) are used to calculate the kinetic parameters of the surface CaO carbonation reaction from the SR-XRPD data.

2. EXPERIMENTAL DETAILS 2.1. Materials. High purity commercial calcium carbonate (marble granular A6297 purchased by AppliChem) was used as the starting material in all the experiments performed in this study. The material, characterized by an initial particle size distribution varying between few microns up to 300 μm, was previously sieved to collect a fraction of powder with a controlled average particle diameter in the range of 150−160 μm. Even though CaCO3 particles were polycrystalline and their size would not have affected the diffraction significantly, this size was selected to ensure both a low compaction of the material within the capillary used for conducting in situ X-ray powder diffraction experiments, and a minimal graininess in the 2-D diffraction raw images recorded, as indication of powder samples characterized by a nearly ideal random distribution of the initial coherently scattering C

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Table 1. Summary of Experimental Operating Conditions and Corresponding Results (Obtained by Rietveld Refinement): Final CaO Conversion,30 Initial CaO Crystallite Size30 and Final CaCO3 Average Crystallite Size calcination

carbonation

run

T [°C]

t [min]

T [°C]

final CaO conversion X [−]

initial CaO average crystallite size [nm]

final CaCO3 average crystallite size [nm]

A B C D E F G H I J K L M N O

900 900 700 700 700 900 900 900 900 900 750 800 900 900 700

5 5 60 60 60 5 5 60 60 5 5 5 60 5 60

450 550 650 450 550 500 750 650 450 650 650 650 550 600 500

0.63 0.26 0.72 0.63 0.58 0.21 0.57 0.42 0.07 0.64 0.75 0.71 0.53 0.66 0.19

39 (±2) 100 (±3) 29 (±2) 38 (±2) 61 (±3) 109 (±2) 156 (±3) 255 (±5) 124 (±2) 52 (±2) 31 (±2) 35 (±2) 44 (±1) 50 (±2) 63 (±3)

59 (±3) 42 (±6) 175 (±5) 65 (±3) 99 (±4) 27 (±5) 375 (±7) 327 (±7) 15 (±10) 134 (±5) 186 (±6) 181 (±5) 86 (±3) 118 (±3) 29 (±6)

during the calcination stage at high temperature.35−45 After the complete thermal decomposition of CaCO3, the new-formed CaO was cooled down (at 10 °C/min, still keeping the sample under N2) to the desired carbonation temperature. Afterward, the carbonation stage was carried out isothermally at several temperatures between 450 and 750 °C. As mentioned, the overall carbonation time was set at 20 min for each of the tests. An atmosphere of pure CO2 was used maintaining a constant flow rate of approximately 10 N mL/min at a pressure of 1 bar. The transition between N2 and CO2 was realized by means of an automatic switching valve. 2.2.2. Carbonation Conversion and CaCO3 Average Crystallite Size Estimation. Information about the structural changes of the sorbent samples, mainly CaCO3/CaO phases and CaCO3 crystallite size evolution over time, as a function of both calcination conditions and carbonation temperature, were obtained by means of the Rietveld refinement method. Calcium oxide mass fractions were estimated from the Rietveld quantitative phase analysis (QPA) of the diffraction data refining phase scale factors,34 and were used to calculate the CaO conversion X(t) during the CO2 absorption stages as30 X(t ) =

the peak shapes,34 while the temperature factors were adjusted to take into account the effect of the different carbonation temperatures on the crystalline structure evolution (namely, the attenuation of the Xray scattering caused by the thermal motion of the atoms within the crystal lattice46,47). A cosine Chebyshev function of 12 polynomial terms was applied to fit the profile background. The structure models for the starting material (CaO) and the crystalline reaction product (CaCO3) were taken from the literature; 261847-ICSD and 20179ICSD were selected for lime and calcite, respectively. The instrumental contribution to the line broadening was evaluated by refining the profile parameters for LaB6. Finally, to achieve a careful control of the sequential refinement, in the analysis of the time-resolved patterns only the angular range of 10.5−31.5° of 2θ was used for the refinement of the global and the peak shape broadening parameters, as well as for the phase specific scale factors. The quality of the fitting was then evaluated using the weighted residual factor (wRp) of the refined analysis of each diffraction pattern, in agreement with the method of the least-squares used by GSAS as main refinement technique.34 Satisfactory fittings for all the cases were obtained, with an average value of wRp of about 0.08 (±1% of standard deviation) for the different time-resolved experiments performed.

1 − wCaO(t ) 1 − wCaO(t ) + wCaO(t )

MWCaCO3 MWCaO

(1)

3. RESULTS AND DISCUSSION The data collected from the time-resolved measurements were used to calculate the calcium carbonate average crystallite size trends over time during the CO2 absorption process and were compared with the profiles of the CaO conversion30 in order to identify relationships between the carbonation rate/conversion and the structural properties of the crystalline species (CaCO3) produced by the reaction. The CaO conversion versus time curves, as well as the dependence of the carbonation kinetics/ final calcium oxide conversion from the initial (at the beginning of the carbonation) CaO average crystallite size, were obtained and discussed in Biasin et al.30 Instead, the focus of this work is on the carbonate crystallite size evolution during the reaction progress and its effect on the carbonation kinetics. A summary of all the experiments performed is reported in Table 1, underlining the operating conditions and the main results obtained by Rietveld refinements for each test. As mentioned, 20 min of carbonation was arbitrarily taken as reference time to estimate the final conversion for each experiment as well as the final CaCO3 average crystallite size. The errors in the average crystallite size calculation were estimated from the standard deviation values provided by GSAS for the LX parameters.

where wCaO(t) is the calcium oxide mass fraction evaluated at consecutive diffraction analyses/scans (namely, at a different time t) recorded during the experiments, while MW is the molecular weight of the species involved in the carbonation. The resulting conversion profiles versus time of these experiments and their dependence on the CaO crystallite size at the beginning of carbonation were presented and discussed in detail in Biasin et al.30 The average size of the CaCO3 crystallites in the material was also evaluated from the Rietveld refinements of the profile parameters. In particular, GSAS fits crystallite size broadening with the Lorentzian isotropic crystallite size-broadening factor (LX). The relationship between this (dimensionless) parameter and the average crystallite size (CS, in nm), for each crystalline phase within the material, was obtained using the following equation:34

CS =

18 × 103Kλ π LX

(2)

where λ is the X-ray wavelength (0.07296 nm) and K is a dimensionless constant related to the crystallite shape taken as 0.9. During the refinements, only the Lorentzian size-broadening profile parameter was refined, while the strain-broadening factor did not show a significant contribution. The Thompson-Cox-Hastings pseudoVoigt-type function was used to fit the Lorentzian term and describe D

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 1. CaCO3 average crystallite size time evolution for isothermal carbonation at different temperatures (calcination at 900 °C for 5 min). Profiles corresponding to 20 min (left) and 20 s (right) of carbonation.

Figure 2. CaCO3 average crystallite size time evolution for isothermal carbonation at different temperatures (calcination at 900 °C for 60 min). Profiles corresponding to 20 min (left) and 20 s (right) of carbonation.

Figure 3. CaCO3 average crystallite size time evolution for isothermal carbonation at different temperatures (calcination at 700 °C for 60 min). Profiles corresponding to 20 min (left) and 20 s (right) of carbonation.

3.1. CaCO3 Average Crystallite Size Profiles over Time. As for the curves of calcium oxide conversion versus time described in Biasin et al.,30 the corresponding trends of the average CaCO3 crystallite size as a function of the carbonation time were similarly obtained from the same experiments performed at different carbonation temperatures, using the CaO sorbents previously prepared through calcination stages

carried out at several operating conditions. Specifically, the carbonate crystallite size curves versus time were compared: (i) as a function of the carbonation temperature, with CaO sorbent samples produced at the same calcination conditions (Figures 1−3); (ii) as a function of the CaO initial crystallite size obtained from different calcination stages, at the same carbonation temperature (Figures 6−8). In each of the Figures E

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 4. Final and critical CaCO3 average crystallite sizes as a function of the carbonation temperature. A linear fitting for the final CaCO3 crystallite size is represented by the solid line.

Figure 5. Comparison between CaCO3 average crystallite size and CaO conversion30 versus time curves (carbonation at 450 °C): transition between the initial fast carbonation stage and the second slow carbonation regime.

the curves present a clear slow increase of the carbonate crystallite size over long times. In a few of these curves (namely, curve A in Figure 1, curves I and M in Figure 2, and curve D in Figure 3), the growth rate is so modest that they appear almost constant over time. These general features (initial fast growth stage, sudden transition phase, final slow growth stage) recall similar trends typically observed in the CaO conversion versus time profiles.13,14,21,30 According to these profiles, generally the higher the temperature maintained during the isothermal carbonation, the larger the final CaCO3 average crystallite size measured after 20 min of reaction. Calcium carbonate crystalline domains with an average size in the range between 120 and 180 nm were measured at the end of carbonations performed at 600 and 650 °C (curves N and J in Figure 1 and curve C in Figure 3), while final sizes smaller than 100 nm are always observed for lower carbonation temperatures, namely, at 450 °C (curves A, I and D in Figure 1, 2, and 3 respectively), 500 °C (curve F in Figure 1 and curve O in Figure 3), and 550 °C (curves B, M, and E in Figure 1, 2, and 3 respectively). Average crystallite sizes larger than 300 nm were identified in CaCO3 crystals produced after CO2 absorption experiments carried out at 650 °C (curve H in Figure 2) and at 750 °C (curve G in Figure 1).

1−3 and 6−8 the complete profiles (up to 20 min of carbonation) are shown on the left side, while a detailed view of the calcium carbonate crystallite size trend during the first instants of the reaction is plotted on the right side. Additionally (as previously noted in Biasin et al.30), it has to be underlined that the change in the accuracy observable in each profile after the first 5 min of carbonation reflects the change of the exposure time applied during the collection of the X-ray data, as described in Section 2. Figures 1, 2, and 3 show the CaCO3 average crystallite size profiles, as a function of the carbonation temperature, obtained from CO2 absorption experiments performed with calcium oxide particles produced respectively at 900 °C for 5 min (Figure 1), 900 °C for 60 min (Figure 2), and 700 °C for 60 min (Figure 3). In the curves shown in Figures 1, 2, and 3 three regimes can be distinguished: (1) a first, fast CaCO3 crystallite growth stage; (2) a transition phase; (3) a final stage in which the CaCO3 crystallite growth is slow. The first fast stage can be appreciated in the detailed views of Figures 1, 2, and 3 (on the right side), where the first instants of reaction are magnified to highlight the fast increase of the carbonate crystallite size and where it can be seen that the first stage is completed within a few seconds. The slow CaCO3 crystallite growth typical of the final stage can be seen in the left side of Figures 1−3. Most of F

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 6. CaCO3 average crystallite size time evolution for carbonation at 450 °C using sorbents with different initial CaO crystallite size. Profiles corresponding to 20 min (left) and 20 s (right) of carbonation.

Figure 7. CaCO3 average crystallite size time evolution for carbonation at 550 °C using sorbents with different initial CaO crystallite size. Profiles corresponding to 20 min (left) and 20 s (right) of carbonation.

Figure 8. CaCO3 average crystallite size time evolution for carbonation at 650 °C using sorbents with different initial CaO crystallite size. Profiles corresponding to 20 min (left) and 20 s (right) of carbonation.

The effect of the carbonation temperature on the final (at the end of each experiment) calcium carbonate average crystallite size is shown in Figure 4. From the plot it can be noted that the dependence of the CaCO3 crystallite size on the carbonation temperature is roughly linear. This result can be related with the findings of Li et al.,12 who investigated the effect of the reaction temperature on the carbonation reaction through

atomic force microscopy and noticed that the carbonate product island size increases with the carbonation temperature. As mentioned, all the curves in Figures 1−3 show a characteristic trend that reflects the typical evolution of the carbonation reaction, with a fast initial CaCO3 crystal growth stage that proceeds up to a point where a sharp change in the carbonate crystal growth rate marks the beginning of the G

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

transition to the final slow reaction regime. This feature suggested a comparison of the CaCO3 average crystallite size curves over time with the corresponding CaO conversion trends. Specifically, as shown in Figure 5 considering a carbonation experiment performed at 450 °C (run D), it can be observed that where the average CaCO3 crystallite size reaches a (local) maximum value, it is possible to identify a corresponding value in the CaO conversion versus time curve (dot in Figure 5) that can be reasonably associated with the point marking the end of the fast carbonation period. For this reason, such a maximum value of the CaCO3 crystallite size is defined as the critical carbonate crystallite size CScrit CaCO3. This feature is clearly observable in almost all the experiments during the early instants of carbonation in the form of a peak in the CaCO3 crystallite size versus time profile, with the exception of the tests in which the lowest values of the CaO conversion after 20 min of carbonation were measured (runs B, F, I, and O), and in the data recorded for the CO2 absorption carried out at 750 °C (run G). Nevertheless, also in these 5 (out of 15) cases an abrupt change in the crystal growth rate was observed and was associated with the transition from the initial fast reaction step to the product layer diffusion controlled stage. In order to emphasize the effect of the initial calcium oxide average crystallite size on the evolution of the average size of the CaCO3 crystalline domains, carbonate crystallite size versus time curves obtained performing isothermal carbonations at temperatures of 450, 550, and 650 °C, using CaO sorbent samples with different values of the initial crystallite size (produced through different calcination conditions) were plotted. The comparison is shown in Figures 6, 7, and 8, respectively. At a carbonation temperature of 450 °C (Figure 6), particles with CaO crystalline domains of about 40 nm (curves D and A) lead to carbonate products with final and critical values of their crystallite size larger than those achieved by sorbent samples with an initial CaO average crystallite size of 124 nm (curve I). Similarly, at a carbonation temperature of 550 °C (Figure 7), particles with crystalline domains between 44 and 61 nm (curves M and E) lead to carbonate products with final values of their crystallite size larger than that achieved by the sorbent sample with an initial CaO average crystallite size of 100 nm (curve B). The correlation between increasing CaO initial crystallite size and decreasing CaCO3 crystallite size is even clearer if referring to the critical CaCO3 crystallite size (CScrit CaCO3) rather than to the final CaCO3 crystallite size. Again, at a carbonation temperature of 650 °C (Figure 8), CaO sorbents with an initial average crystallite size of about 30 nm (curves C, K, and L) show a very similar trend, leading to CaCO3 crystallites with sizes of about 180 nm after 20 min of reaction, higher than the one (about 130 nm) characterized by a CaO initial average crystallite size of about 50 nm (curve J). Accordingly, curves C, K, and L show larger crystallites than curve J at the critical point. Instead, curve H shows an anomalous behavior; namely, using CaO sorbent particles with very large initial CaO crystallites, very large final carbonate crystallites are obtained (over 300 nm). 3.2. CaCO3 Critical Crystallite Size. In the previous section (Figure 5), it was shown that from the CaCO3 crystallite size (CS) versus time curves it is possible to identify a critical CaCO3 CS (CScrit CaCO3) corresponding to the beginning of the transition between the first fast reaction stage, and the final slow product layer diffusion controlled step. The list of all

the critical CaCO3 CS determined in this way is reported in Table 2, together with the corresponding CaO conversions Table 2. Critical Carbonate Crystallite Sizes and Corresponding CaO Conversions30 run

critical CaCO3 crystallite size [nm]

critical CaO conversion (Xcrit)[−]

A B C D E F G H I J K L M N O

71 (±3) 21 (±3) 158 (±5) 71 (±3) 75 (±3) 12 (±2) 282 (±8) 270 (±8) 12 (±8) 122 (±7) 179 (±5) 174 (±5) 92 (±3) 120 (±3) 14 (±3)

0.49 0.09 0.68 0.56 0.43 0.06 0.49 0.35 0.02 0.58 0.69 0.63 0.45 0.56 0.11

(from Biasin et al.30). To the best of our knowledge, these data are the first direct measurements of critical product properties during the carbonation of porous particles; in fact, previous estimates of the critical product layer thickness17 h available in the literature are typically based on ex-situ measurements of specific surface area or of pore size distributions (from mercury porosimetry) coupled with theoretical relationships used to compute h.17,19,22 Additionally, Figures 4 and 9 show that the critical CaCO3 CS (CScrit CaCO3) cannot be taken as a constant parameter, but it is dependent on the operating conditions. As the final CaCO3 CS (crystallite size), also the critical CaCO3 CS increases together with reaction temperature (Figure 4) and assumes small values at temperatures close to the CaCO3 Tammann temperature.13 Instead, Figure 9 shows that (with the exception of runs G and H, both performed at high carbonation temperature) the critical CaCO3 CS decreases when the initial CaO crystallite size increases. Specifically, a relationship of inverse proportionality between these two quantities seems to hold. An interesting and new (in the studies on the carbonation reaction) observation is the CaCO3 CS decrease (Figure 5) occurring after the critical CaCO3 CS has been reached. It can be noted, in a typical CaCO3 CS vs time curve, that the CaCO3 CS first slightly decreases until it reaches a local minimum, after which it increases again following the typical product layer diffusion trend. Further investigation is required in order to explain this observation. 3.3. Estimate of the Specific Surface Area of the CaO Sorbents. The experimental measurements of the critical CaCO3 crystallite size were used to estimate the specific surface area utilized for the carbonation during the first reaction stage. Barker,17 assuming a carbonate product layer of uniform thickness distributed over the entire initial sorbent surface area (as determined for example by nitrogen adsorption measurements) and taking into account the expansion of the solid sorbent particles associated with the molar volume increase from CaO to CaCO3, estimated the carbonate critical thickness h from the conversion achieved at the end of the fast reaction period Xh and the initial (at the beginning of the carbonation) H

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 9. Final and critical CaCO3 average crystallite sizes as a function of the initial (at the beginning of carbonation) size of the CaO crystalline domains. Results of all runs are reported on the left. The same data, except those referring to runs G and H, are magnified on the right.

specific surface area per unit of mass of CaO S0 through the following relationship:17,20 h=

m VCaCO 3 Xh × 103 MWCaO S0

observed that the CaCO3 product grows with an island morphology on the CaO surface. From the atomic force microscopy images presented in Li et al.12 the product island shape appears to be roughly conical. In Figure 10, a plot of Si values calculated by means of the eq 4 as a function of the measured critical CaCO3 crystallite size is shown. Figure 10 shows that, except for runs B, F, I, O, the critical CaCO3 crystallite size decreases for increasing values of the specific surface area Si. Such trend can be roughly expressed through a linear correlation (solid line in Figure 10), namely:

(3)

m

In eq 3, VCaCO3 is the calcium carbonate molar volume (in cm3/mol), h is expressed in nm and S0 in m2/g. Conversely, assuming that a number of conically shaped product islands12 are formed on the internal particle surface, and that these islands are characterized by an height hi estimated through the critical carbonate crystallite size CScrit CaCO3, a new equation (analogous to eq 3) can be developed to compute from CScrit CaCO3, and from the corresponding CaO conversion Xcrit the specific surface area which is effectively utilized to produce the carbonate islands (and which initially is a portion of the CaO−gas interface) Si: Si = 3 ×

m VCaCO 3

Xcrit

crit MWCaO CSCaCO 3

crit CSCaCO = − 1.84 × 101Si + 3.10 × 102 3

It should be noted that the relationship formalized by eq 5 is independent of the temperature applied during the isothermal CO2 absorption tests and from the operating conditions used to produce the starting calcium oxide sorbent particles. As shown in Tables 1 and 2, runs B, F, I, O (crosses in Figure 10) are characterized by very small values (much smaller than the values typical of the other runs) of the critical carbonate crystallite size and of the critical/final CaO conversions as well, and do not follow the trend indicated by eq 5. Different properties of the samples produced in such runs were already highlighted,30 and it was noticed that they do not follow the relationship between final CaO conversion and initial CaO crystallite size that instead holds for the other runs. 3.4. Determination of the Intrinsic Surface Chemical Reaction Kinetic Parameters. The estimate of the specific surface area of the CaO sorbent particles utilized during the carbonation represents an important element to complete the investigation of the CaO conversion rates, specifically to determine the kinetic parameters of the surface carbonation reaction. As previously discussed in Biasin et al.,30 in addition to monitoring the average size evolution of the CaO/CaCO3 crystalline domains during the carbonation reaction, the in situ X-ray powder diffraction experiments permitted a detailed analysis of the rapid first stage of the CaO−CO2 reaction and to identify high conversion rates never measured before for the carbonation reaction using CaO-based solid sorbents. In the initial part of the first rapid stage of the carbonation (at low conversions), the measured conversion rate of the calcium oxide carbonation is representative of the intrinsic reaction rate due to the surface chemical reaction, because the product layer diffusion can be neglected. In this regime, the

× 103 (4)

A summary of such calculated values is listed in Table 3. The assumption of conically shaped product islands is based on the work of Li et al.,12 who, using atomic force microscopy, Table 3. List of the Estimated CaO Specific Surface Areas Utilized during Carbonation run

CaO specific surface area Si [m2/g]

A B C D E F G H I J K L M N O

13.56 8.55 8.52 15.51 11.31 10.08 3.42 2.55 3.18 9.36 7.62 7.17 9.63 9.21 15.18

(5)

I

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 10. Relationship between the critical CaCO3 crystallite size and the specific surface area Si. Crosses refer to runs B, F, I, O. Linear correlation (solid line) is based on all runs except B, F, I, O.

Figure 11. Arrhenius plot for carbonation experiments performed in the temperature range between 450 and 750 °C with CO2 partial pressure of 1 bar. Crosses refer to runs B, F, I, O.

specific rate (in s−1) can be expressed as a function of the specific surface area S of the CaO sorbent particles, namely:14 dX 1 eq n = k(PCO2 − PCO ) SMWCaO 2 dt (1 − X )

initial conversion rates (using a linear fitting up to 25% of the final conversion) were obtained in Biasin et al.30 As far as the specific surface areas, recent work by Li et al.12 based on atomic force microscopy show that carbonate product islands (instead of a uniform layer) are formed on the particle surface, and that the particle surface is not uniformly covered by the carbonate. Therefore, the specific surface area to be used in eq 7 is not the total initial specific surface area (as can be measured by nitrogen adsorption), but the active initial specific surface area, this being defined as the portion, at the carbonation beginning (t = 0), of the total surface area (CaO−gas interface) that is af terward (after t = 0) utilized for the carbonation reaction and covered by the product islands. The initial active specific surface areas were estimated by the specific surface areas effectively utilized during the CaO carbonation at the end of the fast first carbonation stage (when the critical crystallite size was reached); these latter quantities were calculated by eq 4 and reported in Table 3. Placing:

(6)

In eq 6, k = k0 exp(−Ea/RT) is the rate constant expressed according to the Arrhenius law through a pre-exponential factor k0, and the activation energy Ea, and (PCO2 − Peq CO2) is the reaction driving force (n is the reaction order), namely, the difference between the operating CO2 partial pressure and its equilibrium value at the carbonation temperature of the CO2 absorption experiment. According to Sun et al.,14 the CaO carbonation can be described as a zero-order reaction with respect to the CO2 partial pressure, relative to the equilibrium partial pressure, when such quantity is equal or higher than 10 kPa. In this work, the experiments were carried at a CO2 partial pressure equal to 1 bar; therefore the reaction order n can be assumed equal to zero and eq 6 can be rewritten as dX 1 = k SMWCaO dt (1 − X )

(7)

K=

In order to evaluate the kinetic parameters (pre-exponential factor and activation energy), eq 7 was evaluated at time equal to zero, namely, placing X = 0. The experimental data of the

dX 1 dt SMWCaO

eq 7 at the initial time can be rewritten as J

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 12. Specific reaction rate vs estimated CaO sorbent specific surface area at the beginning of carbonation. Solid line represents eq 7 at the initial time with k0 = 4.41 × 10−4 mol/m2 s and Ea = 0 kJ/mol.

Figure 13. Initial reaction rate as a function of the carbonation temperature: comparison between experimental data and values calculated applying eq 7 with k = k0 = 4.41 × 10−4 mol/m2 s.

ln K = ln k 0 −

Ea RT

reaction rate is proportional to the specific surface area S. In Figure 12, a plot of the experimental specific reaction rates at time zero as a function of the initial active specific surface areas of the CaO sorbents is shown. Such experimental data confirm that higher specific reaction rates are achieved when the initial active specific surface area of the CaO sorbent particles is higher and that a linear dependence between such quantities exists and can be formalized through a linear fitting (with intercept equal to zero, according to eq 7), providing the following expression to compute the conversion rate at the carbonation beginning:

(8)

Therefore, the two kinetic parameters, namely, the activation energy Ea and the pre-exponential factor k0, were obtained by a linear fitting of the experimental data of K reported as a function of the carbonation temperature in a plot of ln K versus 1/T, as shown in Figure 11. A linear fitting based on eq 8, using all the runs except runs B, F, I, O, led to the parameters k0 = 6.54 × 10−4 mol/m2 s and Ea = 2.35 kJ/mol. Even though some authors assert that a zeroactivation energy is rare and that the activation energy for the carbonation is small but not zero, the very small value calculated in this work can be assumed, on the contrary, to imply a zero-activation energy. Despite values of 78 kJ/mol,27 29 kJ/mol,14 21.3 and 19.2 kJ/mol21 and, more recently, 30.2 kJ/mol28 which have been reported, our result is supported by Dennis and Hayhurst29 based on equilibrium analyses and calcination data, as well as by Bhatia and Perlmutter13 who, thanks to a kinetic study based on a random pore model, determined a zero-activation energy for the CaO−CO2 reaction in the regime controlled by the chemical kinetics. In addition, it can be observed that if the activation energy is assumed to be equal to zero, according to eq 7 the specific

dX = 2.47 × 10−2Si dt

(9)

The slope of the previous equation, scaled by the molecular weight of calcium oxide, defines a new reaction rate constant, namely, k = 4.41 × 10−4 mol/m2 s, independent of the carbonation temperature. This result can be compared with the reaction rate constant obtained by Sun et al.,14 who performed a detailed study of the CaO carbonation carrying out TGA experiments under a wide range of CO2 partial pressures, including 1 bar (namely, the CO2 partial pressure used in our carbonation experiments). Additionally, the kinetic analysis developed in this work follows the same assumption that the K

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 14. Effect of the carbonation temperature on the estimate of the initial active specific surface area.

islands initially form and afterward grow. In support to this interpretation it should be noted that Li et al.11,12 observed that when the carbonation temperature increases the product island density (approximately the number of carbonate molecule clusters per unit of CaO surface) decreases while the CaCO3 island size increases. The data presented in this work in Figure 4 and in Figure 14 show similar trends. In fact, the initial active surface area, which can be related to the nucleation points on the CaO sorbent surface where the product islands start to form, decreases with the reaction temperature (Figure 14), while the CaCO3 average crystallite size increases with the reaction temperature (Figure 4) as the average size of the product islands. These results provide support to the rate equation theory explanation of the CaO−CO2 reaction mechanism.

carbonation rate is zeroth order with respect to the CO2 partial pressure (at 1 bar of CO2 partial pressure) and the same rate expressions (eqs 6 and 7) used by Sun et al.14 They obtained the result that the pre-exponential factor k0 is equal to 1.67 × (10−3) mol/(m2 s) and the activation energy is 29 kJ/mol. Accordingly, the rate constant k is weakly dependent on the temperature and is equal to 1.34 × (10−5) mol/(m2 s) at 450 °C and equal to 5.52 × (10−5) mol/(m2 s) at 750 °C (in our experiments, the carbonation temperature was in the range between 450 and 750 °C). Therefore, the rate constant computed in this work is higher than the value computed by Sun et al.14 Such discrepancy may be explained by two reasons: (1) as detailed in Biasin et al.,30 in the in situ XRD experiments of this work the external mass diffusion was minimized using a capillary flow-cell obtaining conversion rates significantly higher than the typical values obtained in literature by TGA; (2) the rate constant of this work is computed with reference to the active specific surface areas, which are lower than the total specific surface areas used by Sun et al.14 Applying the estimated reaction rate constant in the eq 7, a comparison between the experimental and the computed reaction rates as a function of the reaction temperature was also performed. The result, shown in Figure 13, confirms a good correspondence between the experimental and calculated values of the initial reaction rates, and explains the (apparently) unexpected decreasing trend of the reaction rate with respect to the reaction temperature followed in most of the isothermal carbonation experiments (except in one run at a carbonation temperature of 450 °C). The specific surface areas calculated applying eq 4 and using the experimental data of the critical CaCO3 crystallite size provides an estimate of the active (not the total) specific surface area of the sorbent particles. The concept of active surface area can find support on the rate equation theory proposed by Li et al.,11,12 by means of which the authors investigated the nucleation and growth of CaCO3 product during the carbonation reaction of CaO with CO2. Specifically, they describe the formation of CaCO3 crystals through a mechanism based on the evolution of carbonate product islands, consisting in clusters of CaCO3 molecules/crystallites distributed over the surface of CaO particles, whose size is influenced by surface, grain boundary and lattice diffusion processes.11,12 The active surface area identified in this work can be related to the nucleation points where calcium carbonate crystallites/product

4. CONCLUSIONS In this work, the time-resolved synchrotron radiation X-ray powder diffraction (SR XRPD) technique was applied to investigate the chemical transformations occurring during the CO2 absorption process. Specifically, the evolution of the average size of the CaCO3 crystalline domains was monitored applying the Rietveld refinement method on the X-ray diffraction data collected at the beamline 17-BM-B of the Advanced Photon Source (APS) at the Argonne National Laboratory. Calcium carbonate crystallite size versus time curves show a characteristic trend that reflects the typical evolution of the carbonation reaction with a fast initial CaCO3 crystal growth stage that proceeds up to a turning point where a sharp change in the carbonate crystal growth rate determines the transition to a second slower reaction period. More specifically, local maxima were observed in the CaCO3 crystallite size versus time curves and were identified as the critical CaCO3 crystallite sizes, marking the transition between the first fast carbonation stage and the second reaction stage controlled by the product-layer diffusion. This approach permits a direct measurement of the critical product properties of carbonated porous particles, instead of estimating them (the critical product layer thickness) from ex-situ measurements of specific surface area or of pore size distributions coupled with theoretical relationships.17,19,22 Additionally, the concept of critical CaCO3 crystallite size does not require the assumption that the CaO particle internal surfaces are uniformly covered by the reaction product (such L

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

at Argonne and use of the Advanced Photon Source, an Office of Science User Facility operated for the U.S. Department of Energy (DOE) Office of Science by Argonne National Laboratory were supported by the U.S. DOE under Contract No. DE-AC02-06CH11357. The skillful assistance from the personnel of the 17-BM-B beamline during the synchrotron radiation experiments has been greatly appreciated. Our special thanks are directed to Dr. G. J. Halder and Dr. A. Yakovenko for their help during the work at the APS. Additional support for this work was provided in part by the National Science Foundation under Grant No. DMR-086935. Finally, special gratitude is also directed to Dr. James A. Kaduk for his valuable advices and suggestions regarding the Rietveld refinement method.

assumption is instead typically made when computing the critical product layer thickness17). Assuming that a number of conically shaped product islands12 are formed on the internal particle surface (rather than a uniform product layer), an equation was developed to compute the specific surface area effectively utilized to produce the carbonate islands (rather than the total specific surface area) from the critical CaCO3 crystallite size. The estimate of such active specific surface areas allowed the determination of the intrinsic rate parameters of the surface carbonation reaction. Specifically, combining recent reaction rate data30 not affected by external mass diffusion with the active specific surface areas computed in this work, a new reaction rate constant of 4.41 × 10−4 mol/m2 s was obtained, independent of the carbonation temperature. Namely, a zero-activation energy was found for the carbonation reaction. The measured critical CaCO3 crystallite sizes were compared with the average size of the initial (at the beginning of the carbonation) CaO crystalline domains previously calculated.30 The comparison shows that a relationship of inverse proportionality holds between such two quantities, so that the smaller is the size of the fresh calcium oxide crystallites, the larger is the critical carbonate crystallite size. Instead, the critical carbonate crystallite size was found to be linearly dependent on the carbonation temperature. Finally, it has to be noted that several findings of this work are in agreement with some observations (obtained through atomic force microscopy) of Li et al.,11,12 namely, (a) the reaction product is not distributed uniformly over the CaO porous particle internal surface, and only the active portion of the total specific surface area is utilized during the carbonation; (b) the active surface area, which can be related to the nucleation points on the CaO sorbent surface where the product islands start to form,11,12 decreases with the reaction temperature; (c) the CaCO3 average crystallite size increases with the reaction temperature, as the average size of the product islands.11,12





ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.cgd.5b00563. Evolution of the in situ SR-XRD patterns of CaO and CaCO3 as a function of the carbonation time during CaO carbonation at 450 °C and 1 bar of CO2. Typical profile refinement for a diffraction pattern obtained at a CaO conversion of 40% and at a carbonation temperature of 450 °C. Measured (black crosses) and calculated (red solid line) values are compared (their difference is plotted by the blue solid line). The position of the Bragg peaks for CaCO3 and CaO are shown by pink and cyan blue dots, respectively (PDF)



REFERENCES

(1) Metz, B. IPCC Special Report on Carbon Dioxide Capture and Storage; Cambridge University Press for the Intergovernmental Panel on Climate Change: Cambridge, 2005. (2) Bernstein, L.; Pachauri, R. K., et al. Climate Change 2007: Synthesis Report; IPCC: Geneva, Switzerland, 2008. (3) Ciferno, J.; Litynski, J.; Plasynski, S.; Murphy, J. DOE/NETL Carbon Dioxide Capture and Storage RD&D Roadmap, 2010. (4) Dean, C. C.; Blamey, J.; Florin, N. H.; Al-Jeboori, M. J.; Fennell, P. S. Chem. Eng. Res. Des. 2011, 89, 836−855. (5) Figueroa, J. D.; Fout, T.; Plasynski, S.; McIlvried, H.; Srivastava, R. D. Int. J. Greenhouse Gas Control 2008, 2, 9−20. (6) Wang, Q.; Luo, J.; Zhong, Z.; Borgna, A. Energy Environ. Sci. 2011, 4, 42. (7) Yu, C.-H. Aerosol Air Qual. Res. 2012, 745−769. (8) Stanmore, B.; Gilot, P. Fuel Process. Technol. 2005, 86, 1707− 1743. (9) Blamey, J.; Anthony, E. J.; Wang, J.; Fennell, P. S. Prog. Energy Combust. Sci. 2010, 36, 260−279. (10) Kierzkowska, A.; Pacciani, R.; Müller, C. ChemSusChem 2013, 6, 1130−1148. (11) Li, Z.; Sun, H.; Cai, N. Energy Fuels 2012, 26, 4607−4616. (12) Li, Z.; Fang, F.; Tang, X.; Cai, N. Energy Fuels 2012, 26, 2473− 2482. (13) Bhatia, S. K.; Perlmutter, D. D. AIChE J. 1983, 29, 79−86. (14) Sun, P.; Grace, J.; Lim, C.; Anthony, E. Chem. Eng. Sci. 2008, 63, 47−56. (15) Khoshandam, B.; Kumar, R. V.; Allahgholi, L. Korean J. Chem. Eng. 2010, 27, 766−776. (16) Mess, D.; Sarofim, A.; Longwell, J. Energy Fuels 1999, 13, 999− 1005. (17) Barker, R. J. Appl. Chem. Biotechnol. 1973, 23, 733−742. (18) Abanades, J.; Alvarez, D. Energy Fuels 2003, 17, 308−315. (19) Alvarez, D.; Abanades, J. Energy Fuels 2005, 19, 270−278. (20) Alvarez, D.; Abanades, J. Ind. Eng. Chem. Res. 2005, 44, 5608− 5615. (21) Grasa, G.; Murillo, R.; Alonso, M.; Abanades, J. C. AIChE J. 2009, 55, 1246−1255. (22) Sun, P.; Grace, J.; Lim, C.; Anthony, E. Chem. Eng. Sci. 2008, 63, 57−70. (23) Yu, Y. S.; Liu, W. Q.; An, H.; Yang, F. S.; Wang, G. X.; Feng, B.; Zhang, Z. X.; Rudolph, V. Int. J. Greenhouse Gas Control 2012, 10, 510−519. (24) Zhou, Z.; Xu, P.; Xie, M.; Cheng, Z.; Yuan, W. Chem. Eng. Sci. 2013, 95, 283−290. (25) Molinder, R.; Comyn, T. P.; Hondow, N.; Parker, J. E.; Dupont, V. Energy Environ. Sci. 2012, 5, 8958−8969. (26) Kyaw, K.; Kanamori, M. J. Chem. Eng. Jpn. 1996, 29, 112−118. (27) Kyaw, K.; Kubota, M.; Watanabe, F.; Matsuda, H.; Hasatani, M. J. Chem. Eng. Jpn. 1998, 31, 281−284. (28) Wu, S. F.; Lan, P. Q. AIChE J. 2012, 58, 1570−1577.

AUTHOR INFORMATION

Corresponding Author

*Tel.: +39 049 8275464; e-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors wish to acknowledge the Advanced Photon Source facility for the provision of beam time. Specifically, work done M

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

(29) Dennis, J. S.; Hayhurst, A. N. Chem. Eng. Sci. 1987, 42, 2361− 2372. (30) Biasin, A.; Segre, C. U.; Salviulo, G.; Zorzi, F.; Strumendo, M. Chem. Eng. Sci. 2015, 127, 13−24. (31) Chupas, P. J.; Chapman, K. W.; Kurtz, C.; Hanson, J. C.; Lee, P. L.; Grey, C. P. J. Appl. Crystallogr. 2008, 41, 822−824. (32) Hammersley, A. P.; Svensson, S. O.; Hanfland, M.; Fitch, A. N.; Hausermann, D. High Pressure Res. 1996, 14, 235−248. (33) HammersleyA. P.FIT2D: An Introduction and Overview, 1997 (34) Larson, A.; Von Dreele, R. GSAS, 1994, 748. (35) Ewing, J.; Beruto, D.; Searcy, A. W. J. Am. Ceram. Soc. 1979, 62, 580−584. (36) Beruto, D.; Barco, L.; Searcy, A. W. J. Am. Ceram. Soc. 1984, 67, 512−516. (37) Borgwardt, R. H.; Roache, N. F.; Bruce, K. R. Ind. Eng. Chem. Fundam. 1986, 25, 165−169. (38) Borgwardt, R. H. Ind. Eng. Chem. Res. 1989, 28, 493−500. (39) Borgwardt, R. H. Chem. Eng. Sci. 1989, 44, 53−60. (40) Mai, M. C.; Edgar, T. F. AIChE J. 1989, 35, 30−36. (41) Milne, C. R.; Silcox, G. D.; Pershing, D. W.; Kirchgessner, D. A. Ind. Eng. Chem. Res. 1990, 29, 139−149. (42) Fuertes, A. B.; Alvarez, D.; Rubiera, F.; Pis, J. J.; Marbán, G.; Palacos, J. M. Chem. Eng. Commun. 1991, 109, 73−88. (43) Fierro, V.; Adánez, J.; García-Labiano, F. Fuel 2004, 83, 1733− 1742. (44) Sun, P.; Grace, J.; Lim, C.; Anthony, E.AIChE J. 2007, 53.243210.1002/aic.11251 (45) Manovic, V.; Charland, J.-P.; Blamey, J.; Fennell, P. S.; Lu, D. Y.; Anthony, E. J. Fuel 2009, 88, 1893−1900. (46) Dinnebier, R. E.; Billinge, S. J. L. Powder Diffraction; The Royal Society of Chemistry: Cambridge, U.K., 2008. (47) Als-Nielsen, J.; McMorrow, D. Elements of Modern X-ray Physics; John Wiley & Sons: New York, 2011.

N

DOI: 10.1021/acs.cgd.5b00563 Cryst. Growth Des. XXXX, XXX, XXX−XXX