Article pubs.acs.org/IECR
In Situ CO2 Capture Using CaO/γ-Al2O3 Washcoated Monoliths for Sorption Enhanced Water Gas Shift Reaction Melis S. Duyar, Robert J. Farrauto, Marco J. Castaldi,† and Tuncel M. Yegulalp* Earth and Environmental Engineering Department, Columbia University in the City of New York, 500 West 120th Street, New York, New York 10027, United States ABSTRACT: In situ capture of CO2 allows the thermodynamically constrained water gas shift (WGS) process to operate at higher temperatures (i.e., 350 °C) where reaction kinetics are more favorable. Dispersed CaO/γ-Al2O3 was investigated as a sorbent for in situ CO2 capture for an enhanced water gas shift application. The CO2 adsorbent (CaO/γ-Al2O3) and WGS catalyst (Pt/γ-Al2O3) were integrated as multiple layers of washcoats on a monolith structure. CO2 capture experiments were performed using thermal gravimetric analysis (TGA) and a bench scale flow through reactor. Enhancement of the water gas shift (EWGS) reaction was demonstrated using monoliths (400 cells/in.2) washcoated with separate layers of dispersed CaO/γ-Al2O3 and Pt/γ-Al2O3 in a flow reactor. Capture experiments in a reactor using monoliths coated with CaO/γ-Al2O3 indicated that increased concentrations of steam in the reactant mixture increase the capture capacity of the CO2 adsorbent as well as the extent of regeneration. A maximum capture capacity of 0.63 mol of CO2/kg of sorbent (for 8.4% CaO on γ-Al2O3 washcoated with a loading of 3.45 g/in.3 on monolith) was observed at 350 °C for a reactant mixture consisting of 10% CO2, 28% steam, and balance N2. Hydrogen production was enhanced in the presence of monoliths coated with a layer of 1% Pt/γ-Al2O3 and a separate layer of 9.4% CaO/γ-Al2O3. A greater volume of hydrogen compared to the baseline WGS case was produced over a fixed amount of time for multiple cycles of EWGS. The CO conversion was enhanced beyond equilibrium during the period of rapid CO2 capture by the nanodispersed adsorbent. Following saturation of the adsorbent, the monoliths were regenerated (CO2 was released) in situ, at temperatures far below the temperature required for decomposition of bulk CaCO3. It was demonstrated that the water gas shift reaction could be enhanced for at least nine cycles with in situ regeneration of adsorbent between cycles. Isothermal regeneration with only steam was shown to be a feasible method for developing a process.
■
INTRODUCTION As the global energy demand is increasing, environmental and economic effects of using fossil fuels for energy are gaining more recognition. Hydrogen has the potential to become a widespread fuel that can help satisfy global energy demand while simultaneously reducing toxic combustion emissions when used in a fuel cell. Reforming of fossil fuels, especially natural gas, will be the feasible method of hydrogen production for the foreseeable future since it is conveniently available in pipelines in the current infrastructure. Carbon capture and storage can be combined with hydrogen production from fossil fuels to provide a means of power production with near zero emissions. Once CO2 is captured it can be utilized to produce other chemicals or fuels or be safely stored. Steam reforming of hydrocarbon fuels produces synthesis gas, a mixture of hydrogen and carbon monoxide, which can be used to synthesize H2 or other chemicals and fuels through the Fischer−Tropsch process. Equation 1 shows an example of the steam reforming of methane (SMR) yielding a 3:1 ratio of hydrogen to carbon monoxide. steam reforming of methane: CH4 + H 2O ↔ CO + 3H 2
the water gas shift reaction (WGS) shown in eq 2, thereby maximizing hydrogen output from fuel reforming. water gas shift: CO + H 2O ↔ CO2 + H 2
(2)
The water gas shift reaction is slightly exothermic and constrained by equilibrium at temperatures around 300 °C. Hence, the reaction as written is thermodynamically favored at lower temperatures. However, the kinetics of the reaction is slow at desired low temperatures (i.e., below 300 °C) to allow economical production. The thermodynamic and kinetic restrictions on the water gas shift reaction can be overcome by employing Le Chatelier’s principle by removing the CO2 as it is formed. If sufficient product removal is achieved, the reaction can be driven to completion at higher temperatures (i.e., 350−400 °C) where the kinetics are more favorable. The flexibility in operating conditions can be exploited to operate at higher temperatures where the kinetics of the reaction proceed with a much lower amount of catalyst (higher gas hourly space velocity (GHSV)). Product removal also eliminates the need to use large quantities of steam to force the reaction toward completion.
ΔH = 206.2 kJ/mol (1)
Processing of the synthesis gas via Fischer−Tropsch typically requires a 2:1 ratio of hydrogen to carbon monoxide for diesel fuel production. Additionally, if the end goal is hydrogen production, the carbon monoxide in synthesis gas can react via © 2013 American Chemical Society
ΔH = − 41.2 kJ/mol
Received: Revised: Accepted: Published: 1064
September 11, 2013 November 28, 2013 December 9, 2013 December 9, 2013 dx.doi.org/10.1021/ie402999k | Ind. Eng. Chem. Res. 2014, 53, 1064−1072
Industrial & Engineering Chemistry Research
Article
temperatures compared to the release of CO2 reacted with bulk CaO. The driving force for desorption of CO2 from the dispersed CaO adsorbent is the CO2 partial pressure difference between the adsorbent surface and the boundary gas layer. In contrast, to reverse the carbonation reaction (eq 3) for bulk CaO, the sorbent needs to be calcined at very high temperatures(∼800 °C). This regeneration mode causes CaO particles to lose surface area, resulting in a decrease in capture site utilization. It also requires significant energy to achieve the high temperatures. Gruene et al. have shown that dispersed CaO/γ-Al2O3 retains 90% of its initial CO2 capture capacity after 20 cycles of CO2 adsorption and desorption.9 Under the same conditions the CO2 capture capacity of bulk CaO during the 20th cycle is 60% of its initial capture capacity. Previous work on enhanced WGS and SMR reactions have only considered the use of sorbent and catalyst as particles in fixed or fluidized bed reactors. Balasubramanian et al. demonstrated that WGS and SMR could be performed in the same reactor using a fixed bed reactor packed with commercial Cu-based catalyst and CaO.10 In order to operate continuously, Balasubramanian et al.10 included an extra reactor in their system for sorbent regeneration. Hildenbrand et al.11 proposed a similar sorbent enhanced SMR system where nickel catalyst and dolomite particles are to be circulated between a fluidized bed reactor and a fluidized bed regenerator. Oxidation and coking of the catalyst are noted as possible problems associated with high temperature regeneration of a mixture of catalyst and sorbent.11 Maintaining good solid−gas contact and transporting the consumed sorbent to the regenerator result in large energy and attrition penalties when applied to real fluidized bed reactors. Separating catalyst from sorbent before regeneration is another potential problem if it is desired to prevent unnecessary thermal treatment of catalyst and reduce the mass of solids transported to the regenerator. Washcoated monolith reactors offer several advantages over fixed and fluidized bed reactors, which can be used to overcome some of the challenges associated with adding CO2 adsorbents to reforming or shift reactor systems. Monolith preparation procedures can be manipulated to obtain a washcoat containing a layer of catalyst on top of a layer of sorbent.12 Since both shift and capture reactions take place on the surface, this will allow for the CO2 molecules to be captured by the sorbent as they are formed. Having the reactions take place in thin washcoats eliminates internal diffusion effects common in larger particle fixed bed adsorbent reactors. The monolith channels offer the advantages of low pressure drop, attrition resistance, and higher mechanical strength when compared to packed or fluidized bed reactors.13 The weak interaction between CO2 and CaO/γ-Al2O3 as measured by Gruene et al. makes it possible to regenerate the sorbent in situ; thermal damage to the catalyst can be prevented because moderate conditions for regeneration are feasible. In situ regeneration of sorbent has the benefit of eliminating solids transport. Hence, large energy penalties and mechanical damage problems associated with solids handling are avoided. The focus of the present study is to combine nanodispersed CaO on γ-Al2O3 on a monolith structure with a separate washcoat layer of nanodispersed Pt, a known WGS catalyst, to enhance the kinetics of H2 production while capturing CO2. This reactor configuration is expected to eliminate mechanical attrition and prevent the formation of CaCO3 that requires 800 °C for decomposition. Methods of regenerating the CO2saturated adsorbent are also investigated.
Selective and efficient capture of CO2 will result in a reactor effluent stream that consists of nearly pure hydrogen gas (after water separation) which can be used directly to generate electricity from fuel cells. Such technology will provide a means of power generation without releasing any of the typical emissions associated with conventional combustion driven power plants. The enhancement of water gas shift (EWGS) in the presence of CO2 adsorbents has been demonstrated in a recent study by Jang et al. using a packed bed reactor containing an admixture of low temperature water gas shift catalyst and K2CO3 promoted hydrotalcite as the adsorbent.1 CO conversions and rates of production of hydrogen were both observed to have increased during CO2 capture, and a temperature swing method to release CO2 from the adsorbent was proposed. In a separate study by van Selow et al.,2 the stability of a hydrotalcite based sorbent material was examined over more than 300 cycles in an EWGS reactor setting. Using simulated syngas as a feed, van Selow et al.2 observed a stable rate of reaction and CO conversion enhancement for 300 cycles. Stevens et al. have also demonstrated enhanced conversion of CO during WGS in a fixed bed reactor containing Na-promoted CaO in the temperature range 400−600 °C.3 Maximum CO conversions of 100% were observed at 500−600 °C and the use of catalyst− sorbent physical mixtures resulted in the production of up to 45% more H2 compared to equilibrium conditions when only catalyst was used.3 Application of the sorption enhanced reaction concepts to the water gas shift reaction requires a sorbent capable of selective and reversible capture of CO2. Reviews of various sorbents previously compared in terms of their suitability for water gas shift and combined steam methane reforming (SMR) and WGS applications are available in the literature.4−6 The WGS reaction conditions dictate that, in order to capture CO2 in situ, the sorbent needs to have good capture capacity (>0.5 mol/kg) and fast sorption and desorption kinetics at high temperatures (∼350 °C for WGS) and low CO2 partial pressures (∼0.05−0.3 atm).4 Bulk calcium oxide offers good CO2 capture capacity and kinetics at high temperatures by undergoing the reversible carbonation reaction shown in eq 3. carbonation of CaO: CaO(s) + CO2 (g) ↔ CaCO3(s)
(3)
The carbonation reaction is initially fast, where it is controlled by chemical kinetics. As the reaction proceeds, the product layer (i.e., CaCO3) on the surface of CaO particles imposes a diffusion limitation which slows the rate of reaction.7 Thus the formation of CaCO3 prevents most of the CaO from contacting CO2 and lowers the CO2 capture efficiency. Belova et al. have demonstrated that dispersing nanosized particles of CaO on a γ-Al2O3 support can overcome the limitations associated with bulk CaO.8 The nano dispersed CaO sorbent behaves differently compared to bulk CaO. The high dispersion of CaO particles on the γ-Al2O3 support minimizes sintering of CaO, prevents the encapsulation of unreacted CaO sites by CaCO3, and allows more efficient use of CaO sites. Gruene et al. have determined that the dispersed CaO adsorbent bonds weakly with adsorbed CO2 without forming bulk CaCO3.9 Hence the interaction between CO2 and CaO/γAl2O3 is more similar to a chemisorption process. Desorption of CO2 from the CaO/γ-Al2O3 surface takes place at much lower 1065
dx.doi.org/10.1021/ie402999k | Ind. Eng. Chem. Res. 2014, 53, 1064−1072
Industrial & Engineering Chemistry Research
■
Article
EXPERIMENTAL SECTION Adsorbent and Catalyst Preparation. CaO/γ-Al2O3 Preparation. CaO was dispersed on a high surface area γAl2O3 support (SBA 150 from SASOL) using the incipient wetness technique. The γ-Al2O3 support was impregnated using Ca(NO3)2 as a precursor. The sample was then dried in air for 3 h at 200 °C and calcined in air for 4 h at 500 °C. A loading of 10.07% (by mass) CaO on γ-Al2O3 was used for the thermal gravimetric experiments. For flow through reactor experiments average CaO loadings of 8.4 and 9.4% (by mass) on γ-Al2O3 were used for CO2 adsorption and enhanced water gas shift experiments, respectively. Pt/γ-Al2O3 Preparation. A platinum salt precursor (aqueous solution of platinum amine) was impregnated on a γ-Al2O3 support (SBA 150 from SASOL) using the incipient wetness technique. Drying at 180 °C for 2 h followed by calcination at 500 °C for 3 h in air produced Pt dispersed on γ-Al2O3. A loading of 1% (by weight) Pt on γ-Al2O3 was used for the enhanced water gas shift reactor experiments. Washcoated Monolith Preparation. Corning ceramic monoliths 0.75 in. in diameter with cell densities of 400 cells/in.2 were used for all flow through reactor experiments. The monolith was washcoated with adsorbent and/or catalyst to achieve a thin, uniform coating on the walls of the monolith channels. This process involves dipping the monolith in an aqueous slurry, removal of excess slurry with a purge of air, and heat treatment to ensure adhesion. The adsorbent slurry was prepared using a 25% (by weight) solution of ethanol as a base and a solids (CaO/γ-Al2O3) content of 26.7%. The catalyst slurry was prepared by mixing the Pt/γ-Al2O3 powder with distilled water to achieve a mixture with a solids content of 20 wt %. Successful adhesion of the slurry onto the monolith requires a slurry particle size less than 10 μm.13 The catalyst and adsorbent slurries were ball milled separately. In each case the slurry was milled for 24 h and the particle size distribution was measured using laser diffraction to confirm that the mixtures comply with this criterion. For enhanced WGS catalysts, the monoliths were first coated with an aqueous slurry of Pt/γ-Al2O3. The monoliths were then tested in a flow through reactor to determine whether the loading of catalyst on monolith was sufficient for equilibrium product concentrations to be achieved for the space velocities and temperatures used in EWGS experiments. A 1% (by weight) loading of Pt on γ-Al2O3 was found to give satisfactory results. A washcoat of 9.4% (by weight) CaO/γ-Al2O3 was applied on top of the Pt/γ-Al2O3 layer. The concentration of CaO in the total washcoat was 7.9% by weight. The monoliths coated with the CaO/γ-Al2O3 slurry were dried in air at 180 °C for 3 h. If multiple washcoats were needed to achieve a higher loading of adsorbent on the monolith, the latter was dipped in slurry again after drying. Once all coatings had been applied, a final calcination was performed for 5 h in air at 550 °C to fix the washcoats on the monolith. The monoliths coated with Pt/γ-Al2O3 slurry were dried at 180 °C for 2 h. Calcination at 500 °C was performed to fix the washcoat on the monolith wall. Washcoat loading was determined by weight gain compared to the bare monolith. It was assumed that additional mass transfer limitations were not introduced to the EWGS system because the adsorbent and catalyst washcoats had the same porosity (resulting from use of the same γ-Al2O3 carrier). Since the focus of the present study was to demonstrate a novel EWGS system, monolith
optimization has not been discussed in detail. However, in order to allow for easier pore diffusion, a support consisting of larger pores can be used to disperse CaO, or a thinner layer can be deposited on the monolith. Thermal Gravimetric Analysis (TGA). Thermal gravimetric analysis was performed on powdered samples of CaO/γAl2O3 containing 10.07% CaO by mass using a Netsch STA449 F3 Jupiter type TGA. In all experiments, the sample was first heated to 350 °C in nitrogen. In an experiment aimed at determining the average particle size of CaO in the material, the sample was exposed to 10% CO2/N2 for 1 h at 350 °C. In the experiment aimed at determining the optimum temperature for regenerating the sorbent, CO2 capture was initiated at a temperature of 350 °C from a feed consisting of 10% CO2 and balance N2 by volume. After 30 min the feed was switched to 100% N2 under isothermal conditions. After being exposed to pure nitrogen at 350 °C for 30 min, the temperature was increased using a ramp of 5 °C/min up to 600 °C to determine the temperature at which all captured CO2 is released. In the experiments where the effect of exposure to steam on CO2 capture was investigated, the sample first received either a “steam treatment” in a flow of 10% steam (balance N2) or no treatment (100% N2), both at 350 °C for 30 min. This was followed by a period of CO2 capture from a 10% CO2/N2 mixture at 350 °C for 60 min. At the end of each test, the sample was cooled to room temperature in N2. Reactor Design. CO2 Partial Pressure Swing Reactor for Adsorbent Testing. A laboratory scale flow through reactor was used to test the adsorption/desorption behavior of CO2 on CaO/γ-Al2O3 monoliths. Monoliths were housed inside a quartz tube, and the temperature inside this reaction zone was controlled using a tube furnace (Thermocraft) connected to a variable transformer. Steam was generated by means of a syringe pump (KD Scientific) injecting water into a heated (∼130 °C) reactant stream. The high temperature of the reactant stream ensured that the mixture was not saturated with water vapor. Calibration of the syringe pump and leak testing of the reactor were performed to ensure accurate control of the steam flow rate. Product analysis was performed using a micro gas chromatograph (micro-GC; Agilent Quad) capable of detecting CO, CO2, H2, CH4, and N2 down to a concentration of 0.0001%. Steam in the product mixture was condensed prior to entering the micro-GC. Preheating of the reactants prior to entering the furnace and the products prior to entering the condenser was achieved using heat tape (Omega STH101-040) controlled by variable transformers. The reactor was operated in two cycles, which are referred to in this study as “capture” and “regeneration”. Only the reactor feed was changed when the cycle was switched; during “capture”, the monoliths were exposed to a mixture of carbon dioxide, nitrogen, and steam, whereas during “regeneration” the carbon dioxide flow was shut off while the nitrogen and steam flow rates remained unchanged. In all “capture” cycles, the CO2 concentration was maintained at 10% in the reactant mixture, whereas steam concentrations varied from 0 to 28%. In “regeneration” cycles, steam content in the feed varied from 0 to 48%. During the “capture” cycles, the adsorption capacities of CaO/γ-Al2O3 coated monoliths were tested under conditions with varying steam concentration at a temperature of 350 °C. This was a preliminary test to understand how water gas shift process conditions such as temperature and steam concentration would affect the performance of the adsorbent. 1066
dx.doi.org/10.1021/ie402999k | Ind. Eng. Chem. Res. 2014, 53, 1064−1072
Industrial & Engineering Chemistry Research
Article
“Regeneration” cycles enabled detection of the CO2 released in response to decreasing CO2 partial pressure to zero. Desorption of CO 2 at 350 °C under various steam concentrations was examined. Water Gas Shift (WGS) and Enhanced Water Gas Shift (EWGS) Reactor. The CO2 partial pressure swing reactor was modified and used for the water gas shift and enhanced water gas shift reactions. A temperature controller (Omega CN7800) was used to control the furnace. A gas mixture consisting of 4.35% CO, 8.7% H2O, and balance N2 was used for the reactions. The stability of the steam flow rate was verified by stable CO2 concentrations measured by the micro-GC and their accordance with thermodynamically determined values for a variety of steam flow rates. Carbon balances were calculated using N2 as internal standard for each test. Since the N2 flow rate is known and remains unchanged throughout all reactions, flow rates of all product species were calculated from the ratios of their concentrations to N2 as measured by the GC. The carbon balances took into account CO, CO2, and CH4 in the product stream and were between 99 and 101%. WGS was performed with a constant flow of reactants at 350 °C over monoliths coated with Pt/γ-Al2O3. The monolith loading of catalyst was determined via reactor testing of monoliths coated with Pt/γ-Al2O3. The catalyst loading was adjusted so that equilibrium was reached at the conditions used for EWGS tests. Thus WGS baselines were created for comparison with EWGS cycles. EWGS required cyclic operation consisting of “reaction” and “regeneration” cycles. During reaction cycles a mixture of CO, N2, and steam was introduced into the reactor. During regeneration, the CO was removed from the reactant stream. Two types of regeneration were investigated: During “temperature swing regeneration” the steam and N2 flow rates were maintained at levels used during EWGS while CO was absent from the mixture. The temperature was then increased to either 450 or 550 °C. This procedure was continued until CO2 could no longer be detected by the micro-GC. During “isothermal regeneration” the temperature was kept constant at 350 °C and the steam concentration was increased to 67% (with balance N2) in order to determine the feasibility of regenerating in high steam conditions. This was to provide some insight into the likely performance in a pure steam environment, which would simplify the product separation. Isothermal regeneration periods were kept fixed at 25 min.
Figure 1. Adsorption efficiency calculated from TGA data for 10.07% CaO/γ-Al2O3 exposed to 10% CO2 for 1 h.
every CaO site, the final adsorption efficiency reached in Figure 1 can be taken as a measure of dispersion. Thus, the adsorption efficiency θ can also be expressed as a ratio of the number of surface CaO sites (Ns) to the total amount of CaO (Nt) in the sample (eq 5). θ=
n(CaO)
·100%
SA t = Nt(4πrCaO2)
(6)
d particle =
(7)
(SA t/Ns)/π
The final adsorption efficiency attained in Figure 1 is 11.4%, and the weight of the CaO/γ-Al2O3 is 24.8 mg. Using this information to solve eqs 5−7, an average CaO particle diameter of 3.3 nm was calculated for the 10.07% CaO/γ-Al2O3 used in this study. Figure 2 displays the adsorption efficiency of washcoats of CaO/γ-Al2O3 coated on monoliths under conditions of varying steam concentration in the feed. Figure 2a shows that greater adsorption efficiency is observed when the feed consists of 28% steam. Since the partial pressure of CO2 is fixed for all experiments, this is interpreted as a beneficial effect of steam on CO2 uptake capacity of the adsorbent. In the study by Stevens et al.,3 steam treated bulk CaO showed increased CO2 capture capacity compared to untreated CaO, when used in a fixed bed water gas shift reactor. On the basis of X-ray diffraction (XRD) results, it was hypothesized that steam converts CaO to Ca(OH)2 and thus increases the CO2 capture capacity of the sorbent.3 In the present study, steam is thought to interact with CaO/γ-Al2O3 via the hydration reaction or weak chemisorption. Figure 2b shows that when CO2 is removed from the feed at 350 °C, CO2 desorption is enhanced by steam. Exposure of the CaO/γ-Al2O3 to steam is thought to result in the hydration of CaO sites (eq 8) or chemisorption of H2O molecules on CaO sites (eq 9). When CO2 is present in the feed stream, it replaces
RESULTS AND DISCUSSION When comparing the performances of different adsorbents, adsorption efficiency was used as parameter of merit. Adsorption efficiency θ, shown in eq 4, is defined as the ratio of the moles (n) of CO2 adsorbed to the moles of CaO that are present in the adsorbent sample, expressed as a percentage. Adsorption efficiency provides a measure of how much of the theoretical CO2 capture capacity of an adsorbent sample is utilized under a given set of experimental conditions. adsorption efficiency: n(CO2,ads )
(5)
In eq 5, m is the weight of the sample and MWCaO is the molecular weight of CaO. The surface area of each surface site can be calculated by dividing the total surface area (SAt) of CaO in the sample by the number of surface sites. Assuming a spherical geometry for CaO with a radius (rCaO) of 0.242 nm (sum of Ca and O ionic radii), the average diameter (d) for CaO particles in CaO/γ-Al2O3 can be calculated using eqs 6 and 7.
■
θ=
Ns Ns = Nt (m(0.1007)NAvogadro)/MWCaO
(4)
Adsorption efficiency was also used to estimate the average CaO particle size in a sample of 10.07 wt % CaO/γ-Al2O3. Figure 1 shows TGA data for a sample exposed to 10% CO2/ N2 for 1 h. By assuming that 1 molecule of CO2 adsorbs on 1067
dx.doi.org/10.1021/ie402999k | Ind. Eng. Chem. Res. 2014, 53, 1064−1072
Industrial & Engineering Chemistry Research
Article
Figure 2. (a) Adsorption efficiencies of CaO/γ-Al2O3 coated monoliths at 350 °C exposed to different steam concentrations. (b) Volume of CO2 released during regeneration at 350 °C as a percentage of initial volume adsorbed.
these surface species due to its higher affinity toward CaO. Likewise, when CO2 is removed from the feed and the steam concentration is increased, the chemisorbed CO2 molecules are displaced due to the interaction of CaO with H2O molecules (eqs 10a and 10b). The increase in steam concentration is thought to be responsible for a greater degree of replacement of adsorbed CO2 molecules, hence a greater extent of regeneration of the sorbent (eqs 11a and 11b). CaO hydration: CaO(s) + H 2O(g) → Ca(OH)2(s)
(8)
chemisorption of H2O on CaO: CaO(s) + H 2O(g) → CaO ··· H 2O(s)
(9)
possible CO2 adsorption mechanisms in the presence of steam: Ca(OH)2(s) + CO2(g) → CaO ··· CO2(s) + H 2O(g)
(10a)
Figure 3. TG signal during the period of exposure of CaO/γ-Al2O3 to steam.
(10b)
CO2 (balance N2) mixture, shows 0.12 mg of steam is adsorbed, which supports this hypothesis. Figure 4 shows
CaO ··· H 2O(s) + CO2(g) → CaO ··· CO2(s) + H 2O(g)
possible mechanisms for steam regeneration of saturated adsorbent: CaO ··· CO2(s) + H 2O(g) → CaO ··· H 2O(s) + CO2(g) (11a)
CaO ··· CO2(s) + H 2O(g) → Ca(OH)2(s) + CO2(g)
(11b)
Table 1 shows results from BET analysis performed on fresh and treated samples of CaO/γ-Al2O3 powder. Exposure to 350 Table 1. BET Analysis for CaO/γ-Al2O3 Samples Exposed to Different Conditions sample
BET surf. area (m2/g)
fresh 9.4% CaO/Al2O3 adsorbent 9.4% CaO/Al2O3, 2 h at 350 °C 9.4% CaO/Al2O3, 2 h at 350 °C in 66% H2O
69.24 64.41 76.72
Figure 4. Adsorption efficiencies for a steam treated sample and an untreated sample of CaO/γ-Al2O3 in the first minute of CO2 capture.
adsorption efficiencies for the first minute of CO2 capture for a sample that has previously been exposed to steam and a sample that has never been exposed to steam. The decreased weight gain for the steam treated sample agrees with the hypothesis that preadsorbed steam (or hydroxide group) is being replaced by CO2. This replacement results in a weight increase due to water molecules (or hydroxide groups) leaving the surface and a weight gain due to adsorbing CO2. After the first 30 s of capture, the steam-treated sample follows the same rate of CO2 uptake as the untreated sample.
°C causes a slight decrease in surface area, whereas exposure to the same temperature in the presence of steam causes a small increase. Steam appears to be chemisorbing on or hydrating the CaO crystallites (forming Ca(OH)2 or CaO···H2O) possibly rendering them mobile which opens pores that were previously inaccessible and hence causes a small increase in BET area. TGA data in Figure 3, where the CaO/γ-Al2O3 sample was exposed to steam at 350 °C, prior to CO2 capture from a 10% 1068
dx.doi.org/10.1021/ie402999k | Ind. Eng. Chem. Res. 2014, 53, 1064−1072
Industrial & Engineering Chemistry Research
Article
Figure 5. TGA for CaO/γ-Al2O3 sample during CO2 capture and regeneration.
Figure 6. Dry product concentrations for (a) water gas shift reaction on Pt/γ-Al2O3 coated monolith for feed with S/C = 2. (b) Enhanced water gas shift reaction on CaO/γ-Al2O3−Pt/γ-Al2O3 coated monolith for feed with S/C = 2.
saturated with CO2 at 350 °C. The sample was then regenerated in a flow of N2 while the temperature was increased. Figure 5 indicates that the saturated adsorbent can be fully regenerated at 390 °C in a flow of nitrogen. On the basis of previous reactor studies (shown in Figure 2b), it can be suggested that the optimum temperature for regeneration in the presence of steam will be even lower with a faster rate. The temperature of 390 °C is much lower than that required for decomposition of bulk CaCO3 (∼>800 °C). This result is consistent with the work by Gruene et al.9 which has shown that nano dispersed CaO interacts more weakly with CO2 compared to bulk CaO.9 Figure 6 compares results from reactor tests where monoliths coated with Pt/γ-Al2O3 (Figure 6a) and monoliths coated with both CaO/γ-Al2O3 and Pt/γ-Al2O3 (Figure 6b) were used to perform the water gas shift reaction at 350 °C and a steam to CO ratio (S/C) of 2. It should be noted that small amounts of CH4 were observed in the products for WGS as well as EWGS reactions (
