A new model of the carbonator reactor in the Calcium Looping technology for post-combustion CO2 capture

A new model of the carbonator reactor in the Calcium Looping technology for post-combustion CO2 capture C. Ortiza, R. Chacarteguia,*, J.M. Valverdeb, J.A. Becerraa, L. A. Perez-Maquedac 5 aEnergy Engineering Department, University of Seville, Camino de los descubrimientos s/n, 41092 Sevilla, Spain bFaculty of Physics, University of Seville, Avenida Reina Mercedes s/n, 41012 Sevilla, Spain cInstituto de Ciencia de Materiales de Sevilla (C.S.I.C.- Univ. Sevilla), Americo Vespucio 49, 41092 Sevilla, Spain * Corresponding author. Tel.: +34 954487615. E-mail address: [email protected] Abstract The Ca-Looping (CaL) process is considered as a promising technology for CO2 post-combustion capture in power generation plants yielding a minor penalty on plant performance as compared with other capture technologies such as conventional amine-based capture systems. This manuscript presents a new carbonator reactor model based on 15 lab-scale multicyclic CaO conversion results, which take into account realistic CaO regeneration conditions that necessarily involve calcination under high CO2 partial pressure and high temperature. Under these conditions, CaO conversion in the diffusion controlled stage is a relevant contribution to the carbonation degree in the typical residence times. The main novelty of the model proposed in the present work is the consideration of the capture efficiency in the diffusion controlled phase of carbonation. It is demonstrated that increasing the residence time by 20 a few minutes in the carbonator yields a significant improvement of the capture efficiency. Model predictions are shown to agree with experimental results retrieved from pilot-scale tests. The new model allows a more accurate evaluation and prediction of carbonator’s performance over a wider range of residence times. The results obtained may be relevant for the optimization of CaL operation parameters to be integrated in real power plants. Keywords: CO2 capture, Ca-looping, limestone calcination, diffusion phase, calcium looping, carbon dioxide 25 capture. 1. Introduction Taking into account the current energy system and its expected evolution, capture and permanent geological storage of CO2 from fossil fuel power plants is considered as a necessary technology to be developed at commercial scale in the short term for mitigation of global warming. The IPCC [1] has evaluated existing and emerging 30 technologies of CO2 capture and storage (CCS) and concludes that, in most scenarios, CCS should contribute by a 15-55% to the cumulative global mitigation effort until 2100. Amongst the CO2 capture technologies potentially foreseeable for commercial development, the Ca-Looping (CaL) process is highlighted as one with good prospects for post-combustion capture, mainly due to the possibility of being implemented in already existing power plants and the minor penalty on plant performance as compared to other capture technologies such as conventional amine-based capture systems [2,3]. A further advantage of the CaL technology is the low price, wide availability 5 and harmlessness towards the environment of the sorbent precursor raw material, namely natural limestone [2,4,5]. The CaL technology for CO2 post-combustion capture is based on the reversible chemical reaction of carbonationcalcination of CaO. The basic cycle involves CO2 capture from the flue gas stream of a power plant using lime (CaO) derived from calcination of natural limestone: 10 𝐶𝐶𝐶𝐶𝐶𝐶(𝑠𝑠)+𝐶𝐶𝐶𝐶2⇄𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶3(𝑠𝑠) ∆𝐻𝐻𝑟𝑟0=−178 𝑘𝑘𝑘𝑘 𝑚𝑚𝑚𝑚𝑚𝑚 (1) Flue gases from coal-fired power plants generally contain a mole fraction of CO2 in the range of 10-15% [6,7], whereas typical residence times in the carbonator reactor would be on the order of minutes. Taking into account these constraints, optimum carbonation temperatures are around 650°C for a quick enough reaction kinetics and low value of the equilibrium CO2 concentration in order to achieve significant efficiencies of CO2 capture (around 15 80-90%) [8]. Efficient gas-solid contact and heat/mass transfer would be ensured by the use of circulating fluidizedbed reactors (CBF). CFBs are typically operated at atmospheric pressure under the fast fluidization regime, with gas velocities of the order of 5-10 m/s [9,10]. The partially carbonated particles are driven into a second fluidized bed reactor (calciner), where CaO is regenerated by calcination under high temperatures and necessarily high CO2 concentration [11–14]. 20 Once CO2 is captured in the carbonator and heat from the exothermic reaction is recovered, the flue gas almost free of CO2 is released into the atmosphere. After a residence time in the carbonator of a few minutes, the carbonated solids are circulated into the calciner. Coal is burned into the calciner reactor using pure O2 (oxycombustion) to increase the temperature up to the value required for the endothermic calcination reaction to occur fast and avoid CO2 dilution at the same time. Thus, CO2 is retrieved from the calciner for compression, transport 25 and geological storage or other uses. The regenerated sorbent produced in the calciner is returned to the carbonator for a new cycle. In order to predict the efficiency of CO2 capture by integrating the CaL technology on existing power plants, a number of models have being proposed in the literature [15–18]. Besides, several pilot-medium scale plants (up to 1.7 MWth) demonstrate a high CO2 capture efficiency, which raises hopes on a successful scale-up of the technology [13,14,19–21]. Interestingly, the CaL technology is also suitable to be integrated with biomass combustion as shown in a pilot-scale plant (300 kWth), which would allow producing power with negative CO2 emissions [20]. 5 The multicyclic calcination/carbonation behavior of CaO used in current carbonator models has been inferred from thermogravimetric analysis (TGA) results in which calcination is usually carried out under low CO2 partial pressure mainly due to technical difficulties. Under these conditions, most of the carbonation reaction in short residence times occurs on the surface of the CaO particles through a kinetically controlled fast phase whereas the subsequent solid-state diffusion controlled phase is comparatively negligible as will be seen in detail below. The present 10 manuscript proposes a new carbonator model that does consider the relevant effect of realistic calcination conditions, necessarily involving calcination under high CO2 pressure, on the multicyclic CaO conversion behavior. In order to stress the relevant role of calcination conditions on the sorbent performance, the next section is devoted to briefly review recent TGA results that will be employed in the formulation of the new carbonator model. We follow up with the development of the carbonator model. Finally, capture efficiency results from the new model 15 are compared with predictions by models reported in the literature and pilot-scale results. 2. A brief review on the role of calcination conditions on the multicyclic CO2 capture behavior of limestone Results of TGA tests show that the multicyclic CO2 capture capacity of limestone derived CaO is strongly influenced by the conditions under which calcination is performed [22–24]. The most feasible method to generate the heat 20 required in the calciner while at the same time avoiding CO2 dilution is to burn coal by oxy-combustion [4,11–14]. Thus, calcination is carried out in an environment containing between 70% and 90% vol. concentration of CO2 at atmospheric pressure [11,23,25–27]. Calcination kinetics becomes extremely slow under high CO2 partial pressures nearby equilibrium (T ≈ 870-900ºC) [22]. This makes necessary to increase the temperature over 930ºC to increase the CO2 partial equilibrium pressure well above the CO2 partial pressure in the calciner and yield sufficiently fast 25 reaction kinetics in short residence times [14,21,28]. 2.1 Sorbent deactivation The combination of high temperature and high concentration of CO2 during calcination enhances the progressive deactivation of the sorbent as the number of cycles builds up due to enhanced sintering [29–32]. Enhanced sintering of CaO has a strongly negative effect on the active surface of the solids available for the gas-solid reaction in the kinetically driven fast phase. Grasa et al. [33] proposed to fit the observed evolution of CaO conversion 𝑋𝑋𝑁𝑁 (ratio of mass of CaO converted to CaCO3 to mass of CaO initial) with the number of cycles by means of the equation 5 (2): 𝑋𝑋𝑁𝑁=1 1 1−𝑋𝑋𝑟𝑟+κ𝑁𝑁+𝑋𝑋𝑟𝑟 (2) where κ=0,52 and 𝑋𝑋𝑟𝑟=0,075 are the deactivation constant and the residual conversion, respectively. This equation is however based on results from multicyclic calcination/carbonation TGA tests in which the sorbent is regenerated by calcination under low CO2 partial pressure. Thus, conversion after 1st calcination (X1) is rather high (around 0.7). 10 This high value of conversion is due to the relatively high porosity of the CaO particles derived by short-time calcination under low CO2 partial pressure. Despite that Equation (2) cannot reflect the effect of calcination under high CO2 pressure as would be the case in the CaL technology, it has been routinely used in theoretical studies as representative of the sorbent behavior in order to model the carbonator reactor [15–18]. In more recent TGA studies [34,35], a modified equation was proposed using data obtained from multicyclic 15 calcination/carbonation tests under more realistic calcination conditions involving high CO2 partial pressure (70% v/v CO2, 30% air) at temperatures of 950°C as well as fast transitions between the calcination and carbonation stages. The modified equation proposed is: 𝑋𝑋𝑁𝑁 𝑋𝑋1=𝑋𝑋𝑟𝑟 𝑋𝑋1+�1 κ(𝑁𝑁−1)+�1−𝑋𝑋𝑟𝑟 𝑋𝑋1�−1� (3) where κ= 0,776 and 𝑋𝑋𝑟𝑟= 0,077 and 𝑋𝑋1= 0,48 is CaO conversion after 1st calcination. Carbonation/calcination 20 conditions in these tests where 650ºC/5 min for carbonation under 15%CO2 and 950ºC/5 min for calcination under 70% CO2. Equation (3) looks similar to equation (2), but with the main difference that it takes into accounts the smaller value of conversion in the first cycle after calcination under high CO2 partial pressure. Besides, the deactivation constant is higher as compared to that derived from tests carried out calcining under low CO2 partial pressure. In the real 25 process, a makeup flow of fresh limestone must be continuously fed into the calciner in order to compensate for sorbent deactivation. Limestone is therefore first calcined in the stationary state under high CO2 partial pressure at the same conditions that the partially carbonated solids are calcined for CaO regeneration. Thus, the use of equation (3) for modelling the carbonator reactor might help simulating more realistic CaL conditions. 5 2.2. Fast and slow carbonation stages Two well differentiated stages can be observed during carbonation of CaO solid particles [30,36–38]. The first stage takes place on the free surface of the solid through the nucleation and growth of CaCO3 and is governed by the kinetics of the reaction between CaO and CO2. The end of the fast stage takes place when a 40-60 nm thickness product layer is formed, which makes inaccessible a large fraction of CaO in the interior of the particles for fast 10 kinetically controlled carbonation [39]. Once the CaO free surface directly available for fast carbonation has been covered the reaction continues in a second stage which is controlled by the solid-state diffusion of CO32- and O2- ions through the product layer, which is a much slower process [40,41]. TGA results show that the solid-state diffusion controlled carbonation is greatly enhanced when CaO is regenerated under high CO2 partial pressure as compared to calcination under low CO2 concentration [26,42,43]. Figure 1 shows 15 the time evolution of sorbent weight during the first calcination/carbonation cycles of two TGA tests in which calcination was carried out either under high CO2 partial pressure or air. The results show that carbonation in the fast initial phase is extraordinarily hindered when the sorbent is regenerated under high CO2 concentration. Even though the reaction rate is similar for the two tests, the maximum carbonation conversion in the fast phase is significantly hampered for the sorbent regenerated under high CO2 partial pressure essentially due to severe 20 sintering and drastic reduction of the surface area [22]. On the other hand, carbonation in the diffusion controlled phase is notably enhanced when calcination is carried out under high CO2 concentration especially in the first cycles. Figure 1: Time evolution of sorbent weight% during precalcination and subsequent carbonation/calcination cycles. Carbonation at 650ºC for 5 min (15% CO2/85% air vol/vol). Different calcination conditions are indicated. Reproduced from data reported in [35]. 5 A close examination of TGA test results reported in the literature for CaO derived from other precursors such as steel slag (Figure 2a) and results obtained from pilot scale tests (Figure 2b) reveals also that the relatively slower solid-state diffusion controlled phase contributes substantially to carbonation for residence times of a few minutes. 10 Figure 2: (a) Time evolution conversion of steel slag during the first, 5th, 10th, 15th, and 20th cycles (reproduced from [48] with permission). (b) Average CO2 carrying capacity (Xave) vs. time and reaction constant (ksB), derivation for sample removed from the CFB carbonator during operation in pilot-scale test (reproduced from [28] with permission). Figure 3 shows the multicyclic evolution of CaO conversion in the kinetically controlled (XNK) and diffusion controlled 15 (XND) carbonation stages derived from TGA tests reported in [35] in which calcination was carried out under high CO2 partial pressure. Both XNK and XND can be pretty well fitted by equation (3) using the parameters given in table 1. Table 1: Fitting parameters for sorbent deactivation curves of conversion in the kinetic and diffusion controlled stages (Fig. 3). 5 Figure 3: CaO conversion in the kinetically controlled carbonation stage (XNK) and in the diffusion controlled stage (XND) as a function of the cycle number. Pre-calcination and calcinations conditions: CO2 vol% =70%, t=5 min, T=950ºC. Carbonation conditions: CO2 vol% =15%, t=5 min, T=650ºC. The inset shows the time evolution of sorbent wt% in the first 3 cycles. Obtained from data originally reported in [42]. 10 As can be seen in Figure 3, the contribution of the diffusion controlled phase to CaO conversion is quite relevant being of the same order than CaO conversion in the kinetically controlled phase for 5 min overall carbonation periods. Moreover, an increase of the carbonation phase above 5 min would lead to higher conversion in the diffusive phase since carbonation in this stage grows roughly linearly with time until carbonation is completed [42]. Thus, it seems clear that a carbonator model should consider carbonation in the diffusion stage in order to predict 15 realistic values of efficiency and to draw useful conclusions on the optimum operating parameters in a commercial plant. This is the main goal of the present manuscript. 20 Carbonation stage: Kinetic Diffusion Time: 0,3 min 5 min Deactivation constant: κK =0,676 κD =0,871 Residual conversion: XrK=0,0296 XrD =0,0408 1st cycle conversion: X1,K=0,0218 X1,D=0,263 3. Carbonator model approach 3.1 Kinetic model As mentioned in the introduction, in order to predict the CO2 capture efficiency of the CaL technology, many authors [15–18] propose to simplify the CaO conversion behavior by assuming that it reaches a maximum value, XN, at a constant rate (proportional to the free surface available of CaO and the thickness of the CaCO3 layer formed) in a 5 time tlim, after which the reaction rate drops to zero (solid lines in Fig. 4a). This simplifying assumption would be reasonable if calcination was performed under low CO2 partial pressure. However, as pointed out above, the diffusive carbonation phase plays a relevant role for calcination under high CO2 partial pressure whereas the kinetically controlled carbonation stage is severely hampered. Using the multicyclic TGA results reported in [34,35,42], we develop below a modified carbonator model built upon the model proposed by Alonso et al. [18]. 10 The main novelty of our model is that carbonation in the diffusion controlled phase will be considered to predict the carbonation efficiency. Figure 4 (a) shows experimental results from TGA tests in which calcinations were carried out under low CO2 partial pressure. The solid lines are the kinetic curves assumed in the carbonator model described in [18]. As can be seen, the diffusion controlled stage is negligible in the first cycles but as the number of cycles is increased the diffusive 15 conversion becomes more important (see results for cycle N = 20 in Figure 4(a)). As seen in Figure 4 (a) the time for the kinetic stage (tk) decreases with the number of cycles as conversion in the diffusive phase is enhanced, which is however dismissed in the theoretical model (solid lines). Figure 4(b) illustrates the kinetic curves proposed in our work as representative of the experimental curves reported in [42] in which calcinations were carried out under high CO2 partial pressure. It is assumed that the time evolution of CaO conversion can be approximated by two 20 lines of constant slope, one corresponding to the kinetically controlled carbonation stage and another for the diffusive carbonation stage. In this way, the new kinetic model has into account the relevant role of the diffusion controlled phase. According to experimental measurements, conversion would be increased roughly linearly with time in the diffusion controlled stage against the widely accepted conception that once the fast kinetically controlled phase is completed, additional carbonation by solid-state diffusion is negligible. Thus, it may be 25 anticipated that an important parameter influencing the carbonation efficiency would be the residence time of the solids in the carbonator. Figure 4: CaO conversion as a function of time for different cycles showing in (a) the theoretical approximation for the kinetic 5 model (solid lines) to describe the progress of the carbonation reaction with time (reproduced from [18] with permission). Data points are experimental results from TGA tests in which calcination was carried under low CO2 concentration. In (b) the kinetic model curves proposed in the present article (solid lines) are shown superposed to experimental data from TGA tests reported in [49] in which calcination was carried out under high CO2 concentration (dotted lines). Note the different scales in (a) an (b) for the CaO conversion degree. 10 According to the kinetic model proposed in this work (solid lines in Fig 4(b)), the reaction rate can be expressed as a function of residence time in the carbonator by the following expression: 𝑟𝑟𝑁𝑁_𝑖𝑖= ⎩ ⎪ ⎨ ⎪ ⎧ 𝑟𝑟𝑁𝑁𝐾𝐾=𝑋𝑋𝑁𝑁𝑁𝑁 𝑡𝑡𝑁𝑁 𝑓𝑓𝑚𝑚𝑟𝑟 𝑡𝑡≤𝑡𝑡𝑁𝑁 𝑟𝑟𝑁𝑁𝐷𝐷=𝑋𝑋𝑁𝑁𝑁𝑁 𝑇𝑇0−𝑡𝑡𝑘𝑘 𝑓𝑓𝑚𝑚𝑟𝑟 𝑡𝑡𝑁𝑁<𝑡𝑡≤𝑡𝑡𝑚𝑚𝑚𝑚𝑚𝑚 ⎭ ⎪ ⎬ ⎪ ⎫ (4) where rN_i is the reaction rate in the i-phase (either kinetic or diffusion), tk is the time the kinetic phase takes, tmax is 15 the residence time of the particles in the carbonator, and XND/(T0 -tK) is the rate of diffusive conversion that will be derived from experimental results. Using experimental data reported in [42], the reaction rates in the kinetic phase (rNk) and diffusive phase (rND) can be adjusted as a function of the number of cycles N using Eq. (3) as shown in Figure 5. 20 (a) (b) theoretically allowed if carbonation in the diffusion controlled phase is fully neglected as assumed in previous modelling works [18]. In the reference model [18] the active fraction of solids drops significantly after a short time once the kinetic controlled phase is ended. In contrast, the solids are considered to remain active for a long time in the proposed model in the present work as long as the subsequent carbonation in the diffusion controlled phase is taking place. 5 Figures 7 and 8 demonstrate that our model predicts a greater capture efficiency for longer residence times of the particles in the carbonator. For example, Fig. (8) shows that for a ratio FR/FCO2=5, and a solids inventory of about 200 kg/MW, the proposed model predicts a higher capture efficiency than the reference model [18]. Figure 9 shows the capture efficiency calculated as a function of the residence time in the carbonator using FR/FCO2=5 and F0/FCO2=0.01 and varying the value of solids inventory, which changes the residence time τ (Eq. (13)). This figure 10 shows also the ratio of the carbonation efficiency in the kinetic controlled phase to that in the diffusion controlled phase (right axis). Figure 9: -Left axis: Carbonation efficiency as a function of the residence time in the reactor (modifying the solids inventory, Eq. (13)). –Right axis: Ratio of CO2 capture efficiency in the kinetic controlled phase to carbonation efficiency in the 15 diffusion controlled phase as a function of residence time. Calculations are made for fixed values of FR/FCO2 and F0/FCO2 as indicated Figure 9 shows that, according to the reference model [18], the capture efficiency is not improved further after a residence time of about 300 s since the fraction of active particles (considered as only those able to react in the fast 20 phase) diminishes with time quickly. On the other hand, the capture efficiency is further improved by taking into account enhanced conversion in the diffusive phase as shown by the results obtained from the model proposed. As can be seen in Fig. 9 the capture efficiency in the diffusive phase (ECO2,D) becomes the major contribution to the total carbonation efficiency just after tens of seconds. Besides, the capture efficiency continues to increase with the residence time and may reach a rather high value at residence times above 500 s. 5 Figure 10: Carbonation efficiency as a function of the residence time in the carbonator, which is varied by changing the FR/FCO2 ratio (upper horizontal axis). Calculations made for fixed values of Ws and F0/FCO2 as indicated. 10 Figure 10 shows the effect on the capture efficiency of increasing the residence time (horizontal bottom axis) by decreasing the flow rate of recirculated solids between the reactors (FR, horizontal top axis) while maintaining fixed a value for the solids inventory. As may be observed, by taking into account conversion in the diffusive phase, the capture efficiency is notably increased by increasing the value of the average residence time of the particles (τ) while the solids recirculation flow rate is decreased (with a constant value of solids inventory in the carbonator). 15 An additional benefit of reducing the recirculation flow rate can be that the solids activity would be extended for longer times since the frequency of calcinations is reduced, which would serve to minimize the makeup flow of fresh limestone needed. 2712172227 0.3 0.4 0.5 0.6 0.7 0.8 0.9 100 150 200 250 300 350 400 450 500 550 FR/FCO2 ECO2 τ(s) Ws=400 kg/MW F0/FCO2=0,01 ECO2 (calculated according [18]) ECO2 (proposed model) Figure 11: Capture efficiency as a function of the ratio of makeup flow rate to CO2 flow rate (F0/FCO2) at different solid inventories (Ws) and ratios of the recirculation flow rate to CO2 flow rate (FR/FCO2). Results from the proposed model in this work are compared with results from the model presented in a previous work [18]. 5 The results plotted in Figure 11 correspond to simulations in which the solids recirculation flow rate FR is kept fixed while the flow rate of limestone makeup into the calciner (F0) is varied. Interestingly, it is seen that high capture efficiency may be achieved by operating with low values of F0 when Ws is increased as a consequence of having extended the lifetime of the particles that remain active in the carbonator by virtue of diffusive carbonation. We must note, however, that neither sulphation nor ashes, which will reduce the activity of the sorbent, have been 10 considered in the model. An additional makeup flow of fresh limestone will have to be employed in practice due to irreversible sulphation and deactivation by ashes as reported in many studies [7,16,25,45–47]. Let us now compare predicted results by the proposed model with experimental results obtained from pilot-scale tests. Capture efficiency obtained in pilot-scale tests is usually reported as a function of an “active space time” [13,14,28], which is expressed as the product of the residence time times the fraction of active particles reacting in 15 the fast stage and the average conversion in this stage 𝜏𝜏𝑚𝑚 =𝜏𝜏𝑓𝑓𝑚𝑚𝑋𝑋𝑚𝑚𝑎𝑎𝑒𝑒,𝑁𝑁. Taking into account also carbonation in the diffusive phase, 𝜏𝜏𝑚𝑚 would be defined in our model as [28] 𝜏𝜏𝑚𝑚=𝜏𝜏�𝑓𝑓𝑚𝑚𝑋𝑋𝑚𝑚𝑎𝑎𝑒𝑒,𝑁𝑁+(1−𝑓𝑓𝑚𝑚)𝑋𝑋𝑚𝑚𝑎𝑎𝑒𝑒,𝑁𝑁� (27) Table 3 shows the operating conditions used in the simulations whose results are illustrated in Figure 12. This figure shows experimental values of the carbonation efficiency obtained from the pilot-scale INCAR plant and reported in 20 0.4 0.5 0.6 0.7 0.8 0.9 1 0.04 0.08 0.12 0.16 0.2 ECO2 F0/FCO2 FR/FCO2 = 5 Ws=200 kg/MW calculated using [18] Ws=200 kg/MW Extracted from data [18] Ws=200 kg/MW proposed model Ws=400 kg/MW calculated using [18] Ws=400 kg/MW proposed model [14,19,28]. The operating conditions used in the simulations have been chosen as representatives of those reported at INCAR pilot-scale tests [14,28]. Table 3: Operating conditions used in the simulations whose results are plotted in figure 12 5 Figure 12: Normalized capture efficiency as a function of the active space time. Results obtained from the proposed model in 10 this work using diverse operating conditions (displayed in Table 3) are compared with results obtained from previous models [16] [18] and with experimental data reported from pilot-scale INCAR tests [28]. As can be seen in Figure 12 the model developed in the present work provides results that conform better to experimental results as compared to the results obtained from the previously proposed model by Alonso et al. [18], 15 which predict higher capture efficiencies than the experimental ones. Results extracted from a model proposed by Romano [16], which takes into account the effect of ash, irreversible sulphation, and fluidized bed dynamics have been also plotted in Fig. 12. Romano also neglects carbonation in the diffusion phase and uses CaO multicyclic 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 110 100 ECO2/ECO2eq τa (s) INCAR Conditions 1 Conditions 2 Conditions 3 Conditions 4 Conditions 2. Calculated using [18] Model proposed in [16] Vgas= 1,15 (m3/s)/MW f0=0,15 Ws (kg/MW)= 100-800 Conditions 1 Conditions 2 Conditions 3 Conditions 4 FR/FCO2 5 10 5 10 F0/FCO2 0,1 0,1 0,01 0,01 conversion data derived from tests in which calcination was carried out under low CO2 partial pressure [33]. Although Romano’s model results fit to the experimental results within the data scatter, it may be seen that the carbonation efficiency is predicted to increase with the active residence time at a higher rate than the experimentally observed trend (Fig 14). Arguably, the key point enabling our model to reproduce the experimental trend is to include CaO conversion in the diffusion phase, which brings about a non-negligible contribution to the 5 carbonation efficiency. As pointed out by Charitos et al. [42], the particle fraction that would be reacting in the diffusion regime can be much higher than the active fraction reacting in the fast regime. Moreover, as reviewed in the introduction of the present work, carbonation in the diffusion regime is greatly enhanced when calcination is carried out under high CO2 concentration. The new model takes into account these relevant aspects in order to predict more realistic values for the capture efficiency. A subject for further work would be to incorporate into our 10 model the effect of irreversible sulphation, ashes and fluidized bed dynamics. 5. Conclusions Carbonator models formulated until now are based on lab-scale multicyclic CaO conversion results carried out by regenerating CaO under low CO2 partial pressure. Under these unrealistic conditions, carbonation in the kinetically controlled fast phase prevails on diffusive carbonation. Thus, these models neglect further carbonation of the 15 particles once the fast carbonation stage is ended. However, calcination under realistic CaO regeneration conditions, involving a high CO2 partial pressure in the calcination environment, leads to a significant enhancement of CaO carbonation in the diffusion controlled phase as compared to the kinetically controlled fast phase, which is severely hindered. The new model proposed in this work is based on lab-scale multicyclic CaO conversion results in which calcination is carried out under high CO2 partial pressure as expected in practice. Given the extraordinary 20 relevance of carbonation in the diffusive phase observed under these realistic conditions, the model incorporates in the calculation of the capture efficiency this stage of carbonation. Thus, the particles are considered to remain active in the carbonator beyond the kinetically controlled carbonation phase, which is quite reduced. A main conclusion from the simulation results is that the capture efficiency is improved by increasing the residence time in the carbonator. Thus, for typical inventories of solids in the carbonator used in previous models, we see that the 25 molar flow rate of CaO recirculated can be substantially decreased while a high capture efficiency is obtained. By increasing the residence time of the particles in the carbonator, deactivation due to sintering in the calciner would be mitigated. Furthermore, a smaller flow of recirculated solids would allow reducing energy penalties in the operation of power plants since the amount of heat required for calcination would be lowered and the energy necessary for material transport would be also reduced. In addition, a prolonged carbonation during extended residence times in the carbonator will allow extracting more heat from the reaction exothermicity. In a future work we plan to improve the proposed model by taking into account other aspects that will affect the carbonation 5 efficiency at practice such as irreversible sulphation, deactivation by ashes and the effect of fluidization dynamics. --------------------------------- Acknowledgements 10 This work was supported by the Andalusian Regional Government Junta de Andalucia (contract FQM-5735), Spanish Government Agency Ministerio de Economia y Competitividad (contract CTQ2014-52763-C2-2-R). Notation A carbonator section, m2 r N,K particle reaction rate in the kinetic regime, s-1 CCO2 average CO2 concentration, mol/m3 Save,K average reaction available surface, m−1 C CO2,eq equilibrium concentration of CO2, mol/m3 S N reaction available surface in the N cycle, m2/m3 Deff intrinsic diffusion constant, m3/(mol ·s) T temperature (ºC) D*eff equivalent diffusion constant, m3/(mol ·s) t time, s E CO2 carbon capture efficiency t K characteristic time at which kinetic phase end, s ECO2,eq maximum carbon capture efficiency tmax Particle residence time in carbonator, s E CO2,D carbon capture efficiency in the diffusion regime V gas volume flow rate per MW of a typical power plant in flue gas entering the carbonator, m 3 /s E CO2,K carbon capture efficiency in the kinetic regime V M,CaCO3 molar volume, m3/mol f 0 inlet molar fraction of CO 2 W s solid inventory in the carbonator per MW of a typical power plant , kg f a volumetric fraction of CaO that reacts in the kinetic reaction regime W s,c solid inventory in the calciner per MW of a typical power plant , kg fe equilibrium molar fraction of CO2 X carbonation degree of a CaO particle F0 mole flow of fresh makeup limestone, mol/s Xave average conversion of the sorbent F CO2 mole flow of CO 2 in flue gas entering the carbonator X ave,D average conversion of the sorbent in the diffusion phase F CO2,cal mole flow of CO 2 originating in the calciner, mol/s X ave,K average conversion of the sorbent in the kinetic phase F P mole flow of solids purge in the calciner, mol/s X max,ave maximum average carbonation degree of CaO in the solid population F R mole flow of CaO coming from the calciner, mol/s X max,ave,D maximum average carbonation degree of CaO in the solid population in the diffusion phase h thickness of the product layer of a sorbent particle, nm X max,ave,K maximum average carbonation degree of CaO in the solid population in the kinetic phase k s intrinsic kinetic constant m4/(mol ·s) X N maximum carbonation degree of CaO in the N cycle T 0 time of TGA multicyclic test, s X ND maximum carbonation degree of CaO in the N cycle in the diffusion phase M CaO Molar mass of CaCO 3 , g/mol X NK maximum carbonation degree of CaO in the N cycle in the kinetic phase m CO2 volume flow rate per MW of a typical power plant of CO2 in flue gas entering the carbonator, m 3 /s X r residual conversion capacity of a sorbent particle N calcination–carbonation cycles κ deactivation constant of a sorbent particle N Ca mol of Ca in the carbonator, mol Φ N particle fraction in the N cycle P pressure (bar) ρCaO CaO density, g/m3 r ave,D average reaction rate in the diffusion regime, s-1 ρ g density of gas phase, g/m3 r ave,K average reaction rate in the kinetic regime, s-1 τ average residence time in the carbonator, s rN,D particle reaction rate in the diffusion regime, s-1 τa active space time of the carbonator, s References 5 [1] Change WGI of the IP on C. IPCC,2005: Special Report on CARBON DIOXIDE CAPTURE AND STORAGE. 2005. [2] Romeo LM, Lara Y, Lisbona P, Martínez A. Economical assessment of competitive enhanced limestones for CO2 capture cycles in power plants. Fuel Process Technol 2009;90:803–11. doi:10.1016/j.fuproc.2009.03.014. [3] Dean CC, Blamey J, Florin NH, Al-Jeboori MJ, Fennell PS. The calcium looping cycle for CO2 capture from 10 power generation, cement manufacture and hydrogen production. Chem Eng Res Des 2011;89:836–55. doi:10.1016/j.cherd.2010.10.013. [4] Martínez A, Lara Y, Lisbona P, Romeo LM. Energy penalty reduction in the calcium looping cycle. Int J Greenh Gas Control 2012;7:74–81. doi:10.1016/j.ijggc.2011.12.005. [5] Cormos C-C. Economic evaluations of coal-based combustion and gasification power plants with post- 15 combustion CO2 capture using calcium looping cycle. Energy 2014;78:665–73. doi:10.1016/j.energy.2014.10.054. [6] Wang W, Ramkumar S, Wong D, Fan LS. Simulations and process analysis of the carbonation-calcination reaction process with intermediate hydration. Fuel 2012;92:94–106. doi:10.1016/j.fuel.2011.06.059. [7] Manovic V, Anthony EJ. Competition of sulphation and carbonation reactions during looping cycles for CO2 20 capture by cao-based sorbents. J Phys Chem A 2010;114:3997–4002. doi:10.1021/jp910536w. [8] Romano MC, Martínez I, Murillo R, Arstad B, Blom R, Ozcan DC, et al. Process simulation of Ca-looping processes: Review and guidelines. Energy Procedia 2013;37:142–50. doi:10.1016/j.egypro.2013.05.095. [9] Kim K, Kim D, Park Y, Soon K. International Journal of Greenhouse Gas Control A solid sorbent-based multistage fluidized bed process with inter-stage heat integration as an energy efficient carbon capture process. 25 Int J Greenh Gas Control 2014;26:135–46. doi:10.1016/j.ijggc.2014.03.012. [10] Ylätalo J, Ritvanen J, Tynjälä T, Hyppänen T. Model based scale-up study of the calcium looping process. Fuel 2014;115:329–37. doi:10.1016/j.fuel.2013.07.036. [11] Ylätalo J, Parkkinen J, Ritvanen J, Tynjälä T, Hyppänen T. Modeling of the oxy-combustion calciner in the post-combustion calcium looping process. Fuel 2013;113:770–9. doi:10.1016/j.fuel.2012.11.041. [12] Vorrias I, Atsonios K, Nikolopoulos A, Nikolopoulos N, Grammelis P, Kakaras E. Calcium looping for CO2 5 capture from a lignite fired power plant. Fuel 2013;113:826–36. doi:10.1016/j.fuel.2012.12.087. [13] Dieter H, Bidwe AR, Varela-duelli G, Charitos A, Hawthorne C. Development of the calcium looping CO2 capture technology from lab to pilot scale at IFK , University of Stuttgart. Fuel 2014;127:23–37. doi:10.1016/j.fuel.2014.01.063. [14] Arias B, Diego ME, Abanades JC, Lorenzo M, Diaz L, Martínez D, et al. Demonstration of steady state CO2 10 capture in a 1.7MWth calcium looping pilot. Int J Greenh Gas Control 2013;18:237–45. doi:10.1016/j.ijggc.2013.07.014. [15] Lasheras A, Ströhle J, Galloy A, Epple B. Carbonate looping process simulation using a 1D fluidized bed model for the carbonator. Int J Greenh Gas Control 2011;5:686–93. doi:10.1016/j.ijggc.2011.01.005. [16] Romano MC. Modeling the carbonator of a Ca-looping process for CO2 capture from power plant flue gas. 15 Chem Eng Sci 2012;69:257–69. doi:10.1016/j.ces.2011.10.041. [17] Romano M, Martínez I, Murillo R, Arstad B. Guidelines for modeling and simulation of Ca-looping processes. Prepared for the European Energy Research Alliance. SintefNo 2012. [18] Alonso M, Rodríguez N, Grasa G, Abanades JC. Modelling of a fluidized bed carbonator reactor to capture CO2 from a combustion flue gas. Chem Eng Sci 2009;64:883–91. doi:10.1016/j.ces.2008.10.044. 20 [19] Alonso M, Rodríguez N, González B, Grasa G, Murillo R, Abanades JC. Carbon dioxide capture from combustion flue gases with a calcium oxide chemical loop. Experimental results and process development. Int J Greenh Gas Control 2010;4:167–73. doi:10.1016/j.ijggc.2009.10.004. [20] Alonso M, Diego ME, Pérez C, Chamberlain JR, Abanades JC. Biomass combustion with in situ CO2 capture by CaO in a 300kWth circulating fluidized bed facility. Int J Greenh Gas Control 2014;29:142–52. 25 doi:10.1016/j.ijggc.2014.08.002. [21] Ströhle J, Junk M, Kremer J, Galloy A, Epple B. Carbonate looping experiments in a 1 MWth pilot plant and model validation. Fuel 2014;127:13–22. doi:10.1016/j.fuel.2013.12.043. [22] Valverde JM, Sanchez-Jimenez PE, Perez-Maqueda L a. Limestone Calcination Nearby Equilibrium: Kinetics, CaO Crystal Structure, Sintering and Reactivity. J Phys Chem C 2015;119:1623–41. doi:10.1021/jp508745u. 30 [23] Coppola A, Scala F, Salatino P, Montagnaro F. Fluidized bed calcium looping cycles for CO2 capture under oxy-firing calcination conditions: Part 1. Assessment of six limestones. Chem Eng J 2013;231:537–43. doi:10.1016/j.cej.2013.07.113. [24] Manovic V, Charland JP, Blamey J, Fennell PS, Lu DY, Anthony EJ. Influence of calcination conditions on carrying capacity of CaO-based sorbent in CO2 looping cycles. Fuel 2009;88:1893–900. 35 doi:10.1016/j.fuel.2009.04.012. [25] J.M. Cordero MA. Modelling of the kinetics of sulphation of CaO particles under CaL reactor conditions. Fuel 2015;150:501–11. doi:10.1016/j.fuel.2015.02.075. [26] Valverde JM. Relevant Influence of Limestone Crystallinity on CO2 Capture in The Ca-Looping Technology at Realistic Calcination Conditions 2014. [27] Martínez I, Grasa G, Murillo R, Arias B, Abanades JC. Modelling the continuous calcination of CaCO3 in a Calooping system. Chem Eng J 2013;215-216:174–81. doi:10.1016/j.cej.2012.09.134. [28] Charitos A, Rodríguez N, Hawthorne C, Alonso M, Zieba M, Arias B, et al. Experimental Validation of the 5 Calcium Looping CO2 Capture Process with Two Circulating Fluidized Bed Carbonator Reactors. Ind Eng Chem Res 2011;50:9685–95. doi:10.1021/ie200579f. [29] Shimizu T, Hirama T, Hosoda H, Kitano K, Inagaki M, Tejima K. A Twin Fluid-Bed Reactor for Removal of CO2 from Combustion Processes. Chem Eng Res Des 1999;77:62–8. doi:10.1205/026387699525882. [30] Abanades JC. The maximum capture efficiency of CO2 using a carbonation/calcination cycle of CaO/ CaCO3. 10 Chem Eng J 2002;90:303–6. doi:10.1016/S1385-8947(02)00126-2. [31] Abanades JC, Alvarez D. Conversion limits in the reaction of CO2 with lime. Energy and Fuels 2003;17:308– 15. doi:10.1021/ef020152a. [32] Borgwardt RH. Calcium oxide sintering in atmospheres containing water and carbon dioxide. Ind Eng Chem Res 1989;28:493–500. doi:10.1021/ie00088a019. 15 [33] Grasa GS, Abanades JC. CO2 Capture Capacity of CaO in Long Series of Carbonation/Calcination Cycles. Ind Eng Chem Res 2006;45:8846–51. doi:10.1021/ie0606946. [34] Valverde JM. A model on the CaO multicyclic conversion in the Ca-looping process. Chem Eng J 2013;228:1195–206. doi:10.1016/j.cej.2013.05.023. [35] Valverde JM, Sanchez-Jimenez PE, Perez-Maqueda L a. Role of precalcination and regeneration conditions 20 on postcombustion CO2 capture in the Ca-looping technology. Appl Energy 2014;136:347–56. doi:10.1016/j.apenergy.2014.09.052. [36] Arias B, Abanades JC, Grasa GS. An analysis of the effect of carbonation conditions on CaO deactivation curves. Chem Eng J 2011;167:255–61. doi:10.1016/j.cej.2010.12.052. [37] Mess D, Sarofim AF, Longwell JP. Product Layer Diffusion during the Reaction of Calcium Oxide with 25 Carbon Dioxide. Energy & Fuels 1999;13:999–1005. doi:10.1021/ef980266f. [38] Bhatia SK, Perlmutter DD. Effect of the product layer on the kinetics of the carbon dioxide-lime reaction. AIChE J 1983;29:79–86. doi:10.1002/aic.690290111. [39] Alvarez D, Abanades JC. Determination of the Critical Product Layer Thickness in the Reaction of CaO with CO 2. Ind Eng Chem Res 2005;44:5608–15. doi:10.1021/ie050305s. 30 [40] Li Z, Fang F, Xiao-yu T, Cai N. Effect of Temperature on the Carbonation Reaction of CaO with CO2. Energy and Fuels 2012;26:4607–16. doi:10.1021/ef300607z. [41] Sun Z, Luo S, Qi P, Fan LS. Ionic diffusion through Calcite (CaCO 3) layer during the reaction of CaO and CO2. Chem Eng Sci 2012;81:164–8. doi:10.1016/j.ces.2012.05.042. [42] Sanchez-Jimenez PE, Valverde JM, Perez-Maqueda L a. Multicyclic conversion of limestone at Ca-looping 35 conditions: The role of solid-sate diffusion controlled carbonation. Fuel 2014;127:131–40. doi:10.1016/j.fuel.2013.09.064. [43] Valverde JM, Sanchez-Jimenez PE, Perez-Maqueda L a. Calcium-looping for post-combustion CO2 capture. On the adverse effect of sorbent regeneration under CO2. Appl Energy 2014;126:161–71. doi:10.1016/j.apenergy.2014.03.081. [44] Grasa GS, Abanades JC, Alonso M, González B. Reactivity of highly cycled particles of CaO in a carbonation/calcination loop. Chem Eng J 2008;137:561–7. doi:10.1016/j.cej.2007.05.017. 5 [45] Symonds RT, Lu DY, Manovic V, Anthony EJ. Pilot-scale study of CO2 capture by CaO-based sorbents in the presence of steam and SO2. Ind Eng Chem Res 2012;51:7177–84. doi:10.1021/ie2030129. [46] Ryu HJ, Grace JR, Lim CJ. Simultaneous CO2/ SO2 capture characteristics of three limestones in a fluidizedbed reactor. Energy and Fuels 2006;20:1621–8. doi:10.1021/ef050277q. [47] Sun P, Grace JR, Lim CJ, Anthony EJ. Removal of CO2 by Calcium-Based Sorbents in the Presence of SO2. 10 Energy & Fuels 2007;21:163–70. doi:10.1021/ef060329r. [48] Tian S, Jiang J, Li K, Yan F, Chen X. Performance of steel slag in carbonation–calcination looping for CO2 capture from industrial flue gas. RSC Adv 2014;4:6858. doi:10.1039/c3ra47426g. [49] Valverde JM, Sanchez-Jimenez PE, Perez-Maqueda L a. High and stable CO2 capture capacity of natural limestone at Ca-looping conditions by heat pretreatment and recarbonation synergy. Fuel 2014;123:79– 15 85. doi:10.1016/j.fuel.2014.01.045.