Chem. Eng. Technol. 2009, 32, No. 3, 425–434 425 Robin W. Hughes1 Arturo Macchi2 Research Article Dennis Y. Lu1 Edward J. Anthony1 Changes in Limestone Sorbent Morphology during CaO-CaCO3 Looping at Pilot Scale 1 CANMET, Natural Resources Canada, Ontario, Canada. 2 University of Ottawa, Ottawa, Ontario, Canada. A pilot-scale dual fluidized bed combustion system was used for CO2 capture using limestone sorbent with CaO-CaCO3 looping. The sorbent was regenerated at high temperature using an air- or oxygen-fired fluidized bed calciner with flue gas recycle firing hardwood pellets. Two limestones were evaluated for CaOCaCO3 looping. Changes in the sorbent morphology during the tests were identified by scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDX). Changes in pore size distribution and sorbent surface area that occurred during reaction were determined by N2 BET porosimetry. Thermogravimetric analysis (TGA) was used to determine the activity of the sorbent after processing in the dual fluidized bed combustion system. It was found that oxygenfired calcination with high CO2 partial pressure reduced the effectiveness of the two limestone sorbents for CO2 capture when compared to material calcined under oxygen-enhanced air combustion. A shell 1–2 lm thick, with reduced porosity, was formed around the sorbent particle and is believed to be responsible for reduced conversion of CaO to CaCO3. It is believed that ash deposition contributes to the formation of the shell. Keywords: Ca Looping Cycles, Carbonation, Fluidized bed combustion, Oxyfuel, Sulphation Received: November 10, 2008; accepted: December 03, 2008 DOI: 10.1002/ceat.200800590 1 Introduction CaO-based CO2 looping combustion is expected to be an energy-efficient and cost-effective carbon dioxide capture process that can be retrofitted to existing combustion systems or operate in new, more tightly integrated systems. To date, the majority of research on the topic has been conducted using thermogravimetric analysis (TGA) with the CaCO3 calcined under a N2 atmosphere. Untreated limestones typically have reduced capacity for recarbonation after each cycle of reaction with an initial rapid decrease in capacity eventually levelling off at a molar conversion of around 8–15 % [1]. In a commercial process, the rate of decay in sorbent activity will increase as a result of a number of effects, such as increased temperature, increased CO2 concentration, and ash impurities. The reduction in capacity of the sorbent is due to sintering, resulting in reduced surface area and reduced suitable porosity for rapid gas diffusion into the sorbent interior. Three physical – Correspondence: Dr. E. J. Anthony (banthony@nrcan.gc.ca), CANMET, Natural Resources Canada, 1 Haanel Drive Ottawa, Ontario K1A 1M1, Canada. © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim changes take place during calcination: (1) formation of calcium oxide having a pseudo-lattice of calcium carbonate (rhombohedral), (2) recrystallization of the pseudo-lattice to a more stable calcium oxide lattice (face-centred cubic), and (3) sintering of the recrystallized CaO [2]. Sintering begins before recrystallization is complete; thus, curves of specific surface area vs. temperature of calcination show maxima at or near the temperature of calcination for which decomposition is just complete. The presence of both CO2 and water can increase the rate of sintering of CaO-based sorbents. Enhanced sintering in the presence of carbon dioxide may depend on the formation of carbonates at the surface of CaO by chemisorbed carbon dioxide [3]. H2O vapor enhances the rate of crystal growth of CaO at temperatures as low as 300–400 °C. Water may increase the mobility of ions by briefly forming surface hydroxyl groups [4]. Ions with a different valence from Ca2+, such as K+, Na+, and Mo3+, when incorporated in a CaO lattice result in defects that increase the rate of solid-state diffusion. Ions such as these have been shown to increase the rate of sintering of CaO [5]. In this work, we have operated a pilot-scale CaO-CaCO3 looping system [6] to determine the morphological changes that occur to the sorbent as a result of elevated CO2 concentra- http://www.cet-journal.com 426 Chem. Eng. Technol. 2009, 32, No. 3, 425–434 R. W. Hughes et al. tions, high bulk temperatures required for calcination, even higher localized temperatures resulting from the combustion of the fuel required to provide the heat of calcination with high oxygen concentrations, and ash effects. 2 Experimental 2.1 Pilot Operations The pilot operations for these tests were designed to investigate the changes in the morphology of limestone with the reversible CaO-CaCO3 reaction while minimizing side reactions such as direct or indirect sulphation that would obscure the effects of the primary reaction of interest. The operating conditions of the facility were intended to result in temperatures, partial pressures of CO2, O2 and H2O, and particle mechanical stresses that would be expected in commercially operating equipment. Matching these conditions in the pilot facility with those of an operating facility should provide similar sintering rates and, hence, changes in sorbent morphology and sorbent activity at the two equipment scales. The changes in sorbent morphology will provide the basis for selecting appropriate particle reaction models for calciner and carbonator reactor modeling in the future and may provide insights that will allow improved sorbent performance. The CANMET Energy Technology Centre-Ottawa (CETCO) dual fluidized bed combustion (FBC) system for CaOCaCO3 looping studies [6] was used to calcine and carbonate limestone sorbent in a sequential manner. The test conditions are summarized in Tab. 1. First the sorbent was fully calcined in the calciner and then the entire batch of sorbent was conveyed with air to the carbonator. After a carbonation period of 70 min, the sorbent was discharged from the carbonator through a transport line where the sorbent fell freely into the calciner. This cycle was repeated two or three times for each material and test condition, as time permitted. The heating rate of the sorbent was controlled to some extent in one of two ways. In the slow heating method, the sorbent was heated gradually with electrical heaters in the fluidized bed to Table 1. Test conditions for pilot-scale operations. 600–650 °C. Above this temperature, combustion was initiated and the sorbent bed continued to heat gradually until calcination began. In the rapid heating method, the sorbent was injected into the fluidized bed, which was operating at a temperature sufficient for calcination. Tests were completed for all test conditions with the slow heating rate method. In addition, the fast heating rate method was used for runs 1o–p, 1a–c, 1o– c, 1a–c–h, and 1o–c–h (see Tab. 1). The calciner was operated as a bubbling fluidized bed combustor firing one of two types of wood pellets – low-ash or high-ash. The ash fusion, ultimate, proximate, and major oxides analyses for the fuels and two limestone sorbents tested, where appropriate, are provided in Tab. 2. The calciner was fluidized with oxygen-enhanced air or by oxygen with recycled flue gas. In the oxygen-enhanced air firing mode, the fluidizing gas composition was around 46 vol.-% N2 and 54 vol.-% O2. The flue gas exiting the calciner typically had a dry basis composition of 43 vol.-% CO2, 2.5–7 vol.-% O2, and balance mainly N2. The fuel feed rate was maintained at 1.5 kg/h. Air and oxygen were mixed prior to entering the calciner windbox. In the oxygen-fired mode, the flue gas exiting the calciner passed through a baghouse that removed particulates. The gas was then cooled to approximately 15 °C in a tube-in-shell heat exchanger arranged in such a fashion that the condensate that was formed could be removed from the system using a peristaltic pump. The flue gas was pressurized using a rotary lobe blower which resulted in a flue gas temperature of approximately 30 °C prior to mixing with the oxygen. Recycled flue gas and oxygen were mixed prior to entering the calciner windbox. Oxygen firing in these tests resulted in fluidizing gas composition of around 36 vol.-% CO2, 60 vol.-% O2, and balance mainly N2. The flue gas exiting the calciner typically had a dry basis composition of 84–92 vol.-% CO2, 2.5–7 vol.-% O2, and balance mainly N2. The fuel feed rate in oxygen-enhanced air firing mode was maintained at 1.7 kg/h. The carbonator was fluidized with either 8 vol.-% CO2 with balance air or 8 vol.-% CO2, 17 vol.-% H2O(g) with balance air, as indicated in Tab. 1. Water has been shown to have an effect on the carbonation conversion [7] even at temperatures in excess of those where stable calcium hydrate is expected. The two fluidizing streams provided conditions in the carbonator that corresponded to wet and dry flue gases. Moisture content of the flue gas entering the carbonator may have an effect on how both retrofit Run ID and greenfield power plants are heat integrated with the CaO-CaCO3 looping CO2 capture system. The effect on heat integration may have a small, 1a–p but noticeable, effect on the net efficiency of the 2a–p system. Sorbent Calciner Fluidizing Gas Fuel Carbonator Fluidizing Gas (balance air) Katowice Air Low Ash 8 % CO2 Low Ash 8 % CO2, 17 % H2O(g) Low Ash 8 % CO2 1o–p Low Ash 8 % CO2, 17 % H2O(g) 2o–p Low Ash 8 % CO2 1a–c Low Ash 8 % CO2, 17 % H2O(g) 2a–c Oxyfuel Low Ash 8 % CO2 1o–c Air High Ash 8 % CO2 1a–c–h Oxyfuel High Ash 8 % CO2 1o–c–h Oxyfuel Cadomin Air © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 2.2 Sample Conversion Loss in sample weight during heating to 850 °C in an air atmosphere was used to determine the carbonate content of the sorbent from different locations in the pilot facility. The carbonate content was calculated via loss in mass assuming all changes in mass were due to the release of CO2. http://www.cet-journal.com Chem. Eng. Technol. 2009, 32, No. 3, 425–434 CO2 looping combustion Table 2. Ash fusion, proximate, ultimate, and major oxides analyses for fuels and sorbents. Katowice Limestone Cadomin Limestone 427 sults. The thermogravimetric analysis can increase the certainty in the EDX results by providing the degree of carbonation for the bulk of the particle. Analysis High-ash Wood Pellet Low-ash Wood Pellet Initial Softening, Oxidizing ASTM D1857 [°C] 1129 1215 2.4 Fluid, Oxidizing ASTM D1857 [°C] 1235 1349 A porosimeter was used with N2 to determine the pore volume and pore surface area distributions of the sorbent samples. Moisture 8.08 7.81 3 Results Ash 4.69 0.68 3.1 SEM with EDX Volatiles 68.32 75.54 Fixed Carbon 18.91 15.97 C 45.13 45.80 H 5.07 5.52 N 0.22 < 0.10 S < 0.05 < 0.05 O 36.76 40.04 CaO 21.25 16.64 54.10 MgO 2.13 3.86 0.89 SiO2 41.10 33.06 0.85 Al2O3 10.06 11.11 0.24 Fe2O3 3.78 3.16 0.09 TiO2 0.45 0.262 < 0.03 P2O3 1.10 2.00 < 0.03 SO3 0.65 1.78 < 0.100 Na2O 1.53 2.13 < 0.20 K2O 4.15 13.02 0.06 Loss on fusion 12.48 10.91 43.64 N2 BET Porosimetry Proximate [wt %] Ultimate [wt %] Major Oxides [wt %] 2.3 Scanning Electron Microscopy (SEM) with Energy Dispersive X-Ray Spectroscopy (EDX) Pilot plant samples were mounted on carbon stickers and analyzed by SEM with EDX with scale ranges between 10 and 60 lm. Magnification at this scale is suitable for seeing changes in the grain structure of the sorbent and surface structures of the particle. The EDX results provide semi-quantitative compositions for various surface and interior regions. The sum of component weight fractions for the analyses performed in this work range from about 0.2 to 0.55, so a significant portion of the sample is not accounted for. Where EDX compositions are provided, caution should be exercised in interpreting the re- © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim An SEM image of a particle from a sample taken at the end of the 2nd cycle of oxygen-enhanced calcination can be seen in Fig. 1. This particle was calcined by burning low-ash wood pellets at 860 °C with oxygen-enhanced air as fluidizing gas. The fluidizing gas was 53 % oxygen with balance nitrogen. The particle has been broken in order to see both the surface and the interior. Spectrum 2 in Fig. 1 shows the surface of the particle to be heavily sintered when compared to the interior of the particle in the area identified as Spectrum 1. The surface layer appears to be thin, 1–2 lm thick, with 50.64 greatly reduced porosity. This surface layer could pose significant resistance to mass transfer into 3.28 and out of the particle, resulting in increased time 1.30 required for calcination and reduced carbonation conversion in the fast reaction phase. The EDX re0.40 sults shown in Tab. 3 indicate that the interior of 0.11 the particle is composed mainly of CaO, with other 0.04 major oxides consistent with those in the parent limestone. The TGA results are in agreement with < 0.03 the EDX results for the particle interior indicating < 0.100 that there is almost no CaCO3 present. An SEM image of a particle that was calcined < 0.20 under the same conditions as the previously de0.14 scribed particle and then carbonated at 600 °C with a fluidizing gas of 8 % CO2 with balance dry 43.99 air is shown in Fig. 2. This particle has also been broken to see both the surface (Spectrum 1) and the interior (Spectrum 2). The region identified as Spectrum 1 appears to be completely nonporous and would prevent gas phase diffusion of CO2 into and out of the particle. The EDX results indicate elevated levels of Na2O and K2O in this area compared to the parent limestone, indicating that some fuel ash may have deposited on the surface. There is no evidence of these materials in the region identified as Spectrum 2. The levels of carbonation calculated for both regions (41– 52 %) are in excess of those determined for the bulk of the particle (16 %), so it would appear that carbonation occurs at or near the surface of the particle more so than deep within the particle. This can be seen more clearly in Fig. 3, which was drawn from the same sample as that shown in Fig. 2. The particle in this figure does not appear to have any ash deposition http://www.cet-journal.com 428 Chem. Eng. Technol. 2009, 32, No. 3, 425–434 R. W. Hughes et al. Figure 1. SEM image of a Katowice limestone particle calcined with a fluidizing gas of oxygen-enhanced air (53 % O2, balance N2) burning low-ash wood pellets (860 °C). Spectrum 1 – particle interior with partially carbonated CaO; Spectrum 2 – particle surface with major oxide composition similar to parent limestone. on it, as evidenced by both EDX and the absence of the nonporous layer seen in Fig. 2. It is believed that this is a particle that fragmented after transport from the calciner to the carbonator. The surface (Spectrum 2) is fully carbonated according to the EDX results, while the particle interior has approximately the same degree of carbonation as determined by loss-in-weight for the whole particle; 14 % vs. 16 %, respectively. Fig. 4 provides an image of a Cadomin limestone particle calcined with a fluidizing gas of oxygen mixed with recycled flue gas (60 % O2, 36 % CO2, balance mainly N2), burning high-ash wood pellets (910 °C). Here, again, we see a low-porosity surface layer, but it would appear that this is at least partially due to the surface being carbonated to various extents. The concentrations of Na and K are elevated in the regions identified as Spectrum 2 and Spectrum 3, so some ash deposition may have occurred here. Loss-in-weight analysis shows that material drawn with this particle is only 3.7 % carbonated. It would appear that the material near the surface is carbonated, while deep within the interior of the particle, it is nearly completed calcined. The variation in degree of carbonation is clear in Fig. 5, which shows an image of a Cadomin limestone particle carbonated (600 °C) with a fluidizing gas of 8 % CO2, balance air, after having been calcined in the same manner as the particle in Fig. 4. The outer surface appears to be heavily sintered and carbonated while the interior is somewhat less so. The degree of carbonation in the area identified as Spectrum 1 is 40 % while loss-inweight analysis provides a total particle carbonation of 30 %. 3.2 Porosimetry Changes in surface area (dS) after calcination for Katowice and Cadomin limestones are shown in Figs. 6 and 7, respectively. It is apparent that the material calcined with oxyfuel with recycled flue gas has a substantially lower surface area for both limestones than does the oxygen-enhanced case. The changes Table 3. EDX results (wt %) for sorbent particles taken from the calciner or carbonator with regions corresponding to those seen in Figs. 1 to 5. Region 1–1 1–2 2–1 2–2 3–1 3–2 4–1 4–2 4–3 4–4 5–1 5–2 Run ID 1a–p 1a–p 1a–p 1a–p 1a–p 1a–p 1o–c 1o–c 1o–c 1o–c 1o–c 1o–c C 1.84 0.15 16.22 2.81 1.36 5.36 4.23 20.2 24.12 4.33 1.07 2.81 O 23.24 5.5 10.7 15.91 10.35 26.79 21.71 10.54 11.83 21.87 5.69 17.03 Mg 0.19 0.18 0.28 0.41 0.06 0.75 0.69 0.27 0.21 0.43 0.11 0.09 Al 0 0 0 0.08 0 0.07 0.78 0.39 0.27 0.68 0.18 0 Si 0 0.2 0.13 0.22 0 0.19 1.47 0.7 0.52 1.42 0.39 0 Ca 26.95 15.04 12.51 20.96 20.21 21.4 22.06 11.46 9.76 22.51 16.33 23.46 Mn 0 0 0 0 0 0 0.22 0.11 0.08 0.19 0.09 0 Na 0 0.02 0.27 0 0 0 0 0.11 0.23 0 0 0 K 0 0.04 0.21 0 0 0 0 0.14 0.27 0 0 0 Fe 0 0.07 0 0.07 0 0.07 0.4 0.4 0.11 0.24 0.16 0 Sum 52.22 21.2 40.32 40.46 31.98 54.63 51.56 44.32 47.4 51.67 24.02 43.39 © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.cet-journal.com Chem. Eng. Technol. 2009, 32, No. 3, 425–434 Figure 2. SEM image of a Katowice limestone particle carbonated (600 °C) with a fluidizing gas of 8 % CO2, balance air, after having been calcined with a fluidizing gas of oxygen-enhanced air (53 % O2, balance N2), burning low-ash wood pellets (860 °C). Spectrum 1 – particle surface with high C content; Spectrum 2 – particle interior partially carbonated. CO2 looping combustion 429 hanced Cadomin calcine is lower than for the one that was heated rapidly (3.0 m2/g). For comparison, we have previously reported surface areas for this material, calcined in a TGA in a N2 atmosphere, of 2.2 m2/g (915 °C) to 3.2 m2/g (750 °C) [8]. Changes in pore volume (dV) after calcination of the two limestones are provided in Figs. 8 and 9. Here we see greater variation between each of the samples analyzed. There are differences in the way the two limestones have behaved, differences between heating rates, and differences between the calcination atmospheres. The pore volume is greatest for the Katowice limestone (Fig. 8) that has been calcined using oxygen-enhanced combustion, followed by oxyfuel with recycled flue gas using the fast heating method. The greatest loss in pore volume as we proceed to lower total pore volumes is for pores of radii of 30 to 40 nm. Note that the surface areas of the calcines from oxyfuel with recycled flue gas are similar, but the sample that has been heated rapidly has greater pore volume, especially in the 10–50-nm range. This should allow the rapidly-heated material to achieve greater conversion as pore pluggage will occur only at a higher level of carbonation. In the case of Cadomin limestone (Fig. 9), the pore volume is greatest for the samples that were calcined using oxygen-enhanced air combustion (cumulative totals: fast – 0.0085 mL/g; slow – 0.0070 mL/g). The sample heated using the fast method has similar pore volume in the 30- to 50-nm pore radius range, as does that heated by the slower method, but significantly more pore volume at lower pore sizes. The samples treated in oxyfuel with recycled flue gas both have similar pore volume distributions (cumulative totals: fast – 0.0039 mL/g, slow – 0.0044 mL/g). 3.3 CO2 Capture Initial fluidized bed carbonation resulted in carbonator flue gas outlet CO2 concentrations near equilibrium levels, based on the equilibrium vapor pressures reported by Baker [9] for the first 10 to 20 minutes of reaction. After this period, the CO2 concentration in the carbonator outlet rose rapidly to a value 1 to 4 percentage points below the CO2 inlet concentration. Fig. 10 shows the carbonator Figure 3. SEM image of a Katowice limestone particle carbonated (600 °C) with outlet flue gas concentration for Katowice limea fluidizing gas of 8 % CO2, balance air, after having been calcined with a fluidizstone carbonated (600 °C) with a fluidizing gas of ing gas of oxygen-enhanced air (53 % O2, balance N2), burning low-ash wood 8 % CO2, balance air, after having been calcined pellets (860 °C). Spectrum 1 – particle interior with light carbonation; Spectrum with a fluidizing gas of oxygen-enhanced air (53 % 2 – particle surface with more heavily carbonated grains. O2, balance N2), burning low-ash wood pellets (860 °C) for three cycles of calcination and carbonation. Fluctuations in CO2 concentration in the first four in the heating rate of the oxyfuel-with-recycled flue gas cases minutes of reaction are due to variations in carbonator temhave resulted in similar surface areas (1.3–1.4 m2/g), whereas perature as the reactor was being brought to steady state the surface area (1.9 m2/g) for the slowly heated oxygen-en- © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.cet-journal.com 430 Chem. Eng. Technol. 2009, 32, No. 3, 425–434 R. W. Hughes et al. Figure 4. SEM image of a Cadomin limestone particle calcined with a fluidizing gas of oxygen mixed with recycled flue gas (60 % O2, 36 % CO2, balance mainly N2), burning high-ash wood pellets (910 °C). Spectrum 1 – particle surface; Spectrum 2 and Spectrum 3 – sorbent surface coated with a blend of sorbent and ash constituents (elevated Na and K) with high C; Spectrum 4 – highly sintered sorbent grains at the particle surface. operation. The introduction of CO2 to the carbonator results in heat release due to the exothermic carbonation reaction. This heat release requires that most of the power to the electric heaters surrounding the fluid bed be shut off to maintain bed temperature. The pilot facility operators gained experience in predicting the change in power requirements, so this variation is less evident in the later tests. The lowest values of CO2 concentration shown here are approaching the minimum detection limits of the CO2 analyzer, so the minimum outlet concentration of CO2 should not be considered to be less than equilibrium. Taking the difference between inlet concentration and outlet concentration, and then integrating over the period of carbonation indicates that the amount of CO2 captured is the same as would be expected for the extent of carbonation measured from the thermogravimetric analysis. The carbonation conversion for the three cycles was nearly the same at about 16 %. Carbonator outlet flue gas concentration for Katowice limestone carbonated (∼ 600 °C) with a fluidizing gas of 8 % CO2, balance air, after having been calcined with a fluidizing gas of oxygen mixed with recycled flue gas (60 % O2, 36 % CO2, balance mainly N2), burning high-ash wood pellets (910 °C) can be seen in Fig. 11 for three cycles of calcination and carbonation. This sorbent also captured enough CO2 to achieve near-equilibrium levels; however, it did so for a shorter period. The carbonation conversion for the three cycles was similar at about 12 %. 4 Figure 5. SEM image of a Cadomin limestone particle carbonated (600 °C) with a fluidizing gas of 8 % CO2, balance air, after having been calcined with a fluidizing gas of oxygen mixed with recycled flue gas (60 % O2, 36 % CO2, balance mainly N2), burning high-ash wood pellets (910 °C). Spectrum 1 – particle surface; Spectrum 2 – particle interior. © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Discussion A very thin shell with low porosity was formed on the surface of the sorbent. There are a number of phenomena that may lead to this surface feature, including char and ash deposition, ash sintering, cyclic carbonation and calcination, and carbonate formation during sample extraction. Increased CaO sintering due to the high CO2 concentration is not believed to be the direct cause of this phenomenon, as the interior of the sorbent – where CO2 partial pressure is expected to be highest during calcination – has similar surface and grain structure to sorbent prepared in high-CO2 partial pressure calcination TGA experiments. Since the interior of the sorbent has similar morphology to the surface and interior of sorbents prepared via TGA, it would seem reasonable to look to exterior influences on the particle that are related to fluidized bed combustion for the cause of the shell formation. http://www.cet-journal.com Chem. Eng. Technol. 2009, 32, No. 3, 425–434 CO2 looping combustion 431 1 Oxy - Fuel Calcination - 1st Cycle - Slow Oxy - Fuel Calcination - 1st Cycle - Fast Air Calcination - 1st Cycle dS/dln(r) 0.8 0.6 0.4 0.2 0 1 10 100 1000 Ave rage Pore Radius (nm) Figure 6. dS/dln(r) vs. average pore radius for Katowice limestone calcined with low-ash wood pellets by (a) oxyfuel combustion with recycled flue gas using the slow heating method, (b) oxyfuel combustion with recycled flue gas using the fast heating method, and (c) oxygen-enhanced air combustion. Oxy - Fuel Calcination - 1st Cycle - Slow 2.5 Oxy - Fuel Calcination - 1st Cycle - Fast 2 dS/dln(r) Air Calcination - 1st Cycle Slow 1.5 Air Calcination - 1st Cycle Fast 1 0.5 0 1 10 100 1000 Average Pore Radius (nm) Figure 7. dS/dln(r) vs. average pore radius for Cadomin limestone calcined with high-ash wood pellets by (a) oxyfuel combustion with recycled flue gas using the slow-heating method, (b) oxyfuel combustion with recycled flue gas using the fast heating method, (c) oxygenenhanced air combustion using the slow heating method, and (d) oxygen-enhanced air combustion using the fast heating method. Char in fluid bed combustors may be at temperatures far exceeding the bulk bed temperature [10]. Chirone [11] has concluded that in combination with char fines and high particle temperatures caused by combustion, ash-layered bed material will be formed. In a CaO-based CO2 looping system, ash could build up on the surface of the sorbent in the calciner if the ash composition was appropriate. © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Joutsenoja et al. [12] showed that for a series of coals and a coal char, the average particle temperature was on average 87–191 °C above the average bed temperature. Peak particle temperatures ranged from 275–592 °C above the nominal bed temperature. In fluidized beds blown with high-oxygen-concentration fluidizing gas, peak particle temperatures can be expected to be still higher since the rate of reaction and, hence, http://www.cet-journal.com 432 Chem. Eng. Technol. 2009, 32, No. 3, 425–434 R. W. Hughes et al. Oxy - Fuel Calcination 1st Cycle - Slow 0.0018 0.0016 Oxy - Fuel Calcination 1st Cycle - Fast 0.0014 Air Calcination - 1st Cycle dV/dln(r) 0.0012 0.001 0.0008 0.0006 0.0004 0.0002 0 1 10 100 1000 Average Pore Radius (nm ) Figure 8. dV/dln(r) vs. average pore radius for Katowice limestone calcined with high-ash wood pellets by (a) oxyfuel combustion with recycled flue gas using the slow heating method, (b) oxyfuel combustion with recycled flue gas using the fast heating method, and (c) oxygen-enhanced air combustion using the slow heating method. 0.004 Oxy - Fuel Calcination 1st Cycle - Slow Oxy - Fuel Calcination 1st Cycle - Fast Air Calcination - 1st Cycle - Slow Air Calcination - 1st Cycle - Fast 0.0035 0.003 dV/dln(r) 0.0025 0.002 0.0015 0.001 0.0005 0 1 10 100 1000 Average Pore Radius (nm ) Figure 9. dV/dln(r) vs. average pore radius for Cadomin limestone calcined with high-ash wood pellets by (a) oxyfuel combustion with recycled flue gas using the slow-heating method, (b) oxyfuel combustion with recycled flue gas using the fast heating method, (c) oxygenenhanced air combustion using slow heating method, and (d) oxygen-enhanced air combustion using the fast heating method. heat release are proportional to the concentration of the oxygen. It is reasonable to expect that average fuel char temperatures in an oxygen-blown calciner operating at 900 °C may be in the area of 1000 to 1150 °C. Fuel ashes do not melt at an exact single temperature; instead they melt over a temperature range in which both solid and liquid phases are present. It has been found that in certain © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim types of FBC boilers, ashes containing 10–20 wt % liquid will be sticky and tend to adhere to surfaces within the combustor. FACTSage (Facility for Advanced Chemical Thermodynamics) software was used to predict the portion of the fuel ash that is liquid for both the high-ash and low-ash wood pellets used for the tests described in this paper. Carbon and CaO mixed with the fuel ash stream were varied to represent fuel char in various http://www.cet-journal.com Chem. Eng. Technol. 2009, 32, No. 3, 425–434 CO2 looping combustion 433 8 1st Cycle 7 CO2 Concentration (vol.-%) 2nd Cycle 3rd Cycle 6 5 4 3 2 1 0 0 250 500 750 1000 Tim e (s) Figure 10. Carbonator outlet flue gas concentration for Katowice limestone carbonated (∼ 600 °C) with a fluidizing gas of 8 % CO2, balance air, after having been calcined with a fluidizing gas of oxygen-enhanced air (53 % O2, balance N2), burning low-ash wood pellets (860 °C). CO2 Concentration (vol.-%) 9 8 1st Cycle 7 2nd Cycle 3rd Cycle 6 5 4 3 2 1 0 0 100 200 300 400 500 600 700 Tim e (s) Figure 11. Carbonator outlet flue gas concentration for Katowice limestone carbonated (∼ 600 °C) with a fluidizing gas of 8 % CO2, balance air, after having been calcined with a fluidizing gas of oxygen mixed with recycled flue gas (60 % O2, 36 % CO2, balance mainly N2), burning high-ash wood pellets (910 °C). © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.cet-journal.com 434 Chem. Eng. Technol. 2009, 32, No. 3, 425–434 R. W. Hughes et al. concentration [14] and, hence, will reduce char particle temperatures and the portion of the ash that is liquid, and should reduce ash deposition. However, increasing flue gas recirculation will increase parasitic energy losses for recycle blower operation, and increase the size of the calciner and ancillaries. Fuels [15] or blends of fuels, such as biomass with coals or petroleum coke, that do not form high liquidfraction eutectics at temperatures expected in the calciner, should be used. Such fuels include most coals and petroleum coke. Finally, the calciner should be operated at lower temperatures by using a stripping gas such as steam to reduce the partial pressure of CO2 and thereby allowing a reduced operating temperature. For some fuels, operating with a wet recycled flue gas may sufficiently reduce the required operating temperature of the calciner. 0.7 910 deg C 1000 deg C 0.6 1150 deg C Mass fraction liquid 0.5 0.4 0.3 0.2 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 5 Conclusions The sorbents in these tests were able to capture CO2 from a synthetic flue gas to achieve near-equiFigure 12. Mass fraction of the liquid phase vs. CaO added to ash from the highlibrium levels down to a CO2 concentration of ash wood pellets; from FACTSage analysis. around 0.4 vol.-%. Sorbent capacity was significantly lower than expected based on previous thermogravimetric analyses. This is believed to be at least partially due to the formation states of burnout and contact with the sorbent, respectively. of a thin, low-porosity shell around the sorbent, enhanced by Fig. 12 shows the weight percent of ash that is liquid vs. the the deposition of ash from the solid fuel under oxygen-fired CaO content for the high-ash wood pellets for the temperature conditions. There are a number of methods that could be used of the calciner operating with recycled flue gas (910 °C) and at to reduce or eliminate the shell formation and these will be two possible fuel char temperatures (1000 °C and 1150 °C). It explored in the near future at our laboratory. is apparent that even at the bulk bed temperature, the fuel ash contains more than 15 % liquid, which, depending on the viscosity of the slag, may or may not stick to the sorbent. The fuel References ash at 1150 °C is more than half liquid and has a significantly greater chance of depositing on the surface of the sorbent. The [1] P. Sun, C. J. Lim, J. R. Grace, AIChE J. 2008, 54, 1668. high-ash wood pellet ash has a lower initial deformation tem[2] D. Glasson, J. Appl. Chem. 1958, 8, 793. perature (1129 °C) than the low-ash wood pellet ash (1216 °C). [3] D. Glasson, J. Appl. Chem. 1961, 11, 201. The initial deformation temperatures indicate that the FACT[4] P. Anderson, R. Horlock, R. Avery, Proc. Brit. Ceram. Soc. Sage simulation is providing reasonable values for the temper1965, 3, 33. ature at which a substantial portion of the ash is liquid. [5] R. H. Borgwardt, Chem. Eng. Sci. 1989, 44 (1), 53. Skrifvars et al. [13] characterized the sintering tendency of [6] R. W. Hughes, D. Y. Lu, E. J. Anthony, A. Macchi, Fuel Proc. biomass ashes with compression-strength-based sintering analTech. 2005, 86 (14–15), 1523. ysis. The ash pellets were heated to various temperatures, [7] R. Symonds, M.Sc. Thesis, University of Ottawa 2008. cooled, and then their compressive strengths determined. If [8] R. W. Hughes, D. Lu, E. J. Anthony, Y. Wu, Ind. Eng. Chem. the compressive strength increased significantly, the ash was Res. 2004, 43, 5529. considered to have been heavily sintered. The compressive [9] E. Baker, J. Chem. Soc. 1962, 464. strength typically increased rapidly over a fairly narrow tem[10] F. Scala, R. Chirone, P. Salatino, Fuel Proc. Tech. 2003, 84, perature range at a liquid weight fraction of about 15 % for 7 229. out of the 10 samples. If the ashes examined here behave in a [11] R. Chirone, P. Salatino, F. Scala, Proc. Combust. Inst. 2000, similar manner to those tested by Skrifvars et al. [13], then we 28, 2279. can expect them to sinter heavily in areas where the ash con[12] T. Joutsenoja, P. Heino, R. Hernberg, B. Bonn, Combust. tent is greater than about 30 wt %. Severe sintering can result Flame 1999, 118, 707. in low-porosity structures, as we see on the sorbent surface. [13] B. Skrifvars, R. Backman, M. Hupa, Fuel Proc. Tech. 1998, 56, There have been a number of approaches that have been used 55. to reduce ash agglomeration that may be of use in improving [14] I. Obernberger, Biomass Bioenergy 1998, 14 (1), 33. sorbent performance if the shell formation is due to ash deposi[15] J. Werther et al., Prog. Energy Combust. Sci. 2000, 26, 1. tion. Increasing flue gas recirculation will reduce the oxygen Mass fraction CaO; balance ash © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.cet-journal.com