3544
Energy & Fuels 2008, 22, 3544–3550
CO2 Splitting via Two-Step Solar Thermochemical Cycles with Zn/ZnO and FeO/Fe3O4 Redox Reactions: Thermodynamic Analysis M. E. Ga´lvez,† P. G. Loutzenhiser,‡ I. Hischier,† and A. Steinfeld*,†,‡ Department of Mechanical and Process Engineering, ETH Zurich, 8092 Zurich, Switzerland, and Solar Technology Laboratory, Paul Scherrer Institute, 5232 Villigen PSI, Switzerland ReceiVed April 2, 2008. ReVised Manuscript ReceiVed June 5, 2008
Two-step thermochemical cycles for CO2 reduction via Zn/ZnO and FeO/Fe3O4 redox reactions are considered. The first, endothermic step is the thermal dissociation of the metal oxide into the metal or a reduced valence metal oxide and O2 using concentrated solar energy as the source of high-temperature process heat. The second, nonsolar, exothermic step is the reaction of the reduced metal/metal oxide with CO2, yielding CO and/or C, together with the initial form of the metal oxide that is recycled to the first step. Chemical equilibrium compositions of the pertinent reactions are computed as a function of temperature and pressure. A second-law thermodynamic analysis for the net reaction of CO2 ) CO + 0.5O2 indicates a maximal solar-chemical energy conversion efficiency of 39 and 29% for the Zn/ZnO and FeO/Fe3O4 cycle, respectively. Efficiencies are lower for both cycles yielding C. Major sources of irreversibility are associated with the re-radiation losses of the solar reactor operating at 2000 K and the quenching of its products to avoid recombination.
1. Introduction Stabilizing CO2 atmospheric concentrations is of major concern. Considerable efforts are currently underway to attain a zero-emission energy production scenario involving the development of more efficient energy systems, renewable energy use, as well as CO2 capture, sequestration, and/or use.1,2 CO2 capture, either by means of fuel decarbonization prior to combustion,3,4 separation from the combustion flue gas,5-7 or directly from air,8,9 produces a stream of pure CO2 that is stored long-term10-12 or used as feedstock for the synthesis of chemical commodities.1,13,14 A promising alternative to direct CO2 sequestration is to decompose CO2 into its base elements: C, CO, and O2. Solid carbon can be safely stored. Both C and CO can be used as combustion fuels or further processed to synthetic liquid fuels for transportation. O2 is needed for the oxycombustion and gasification technologies. Direct thermal de* To whom correspondence should be addressed. Fax: +41-44-6321065. E-mail:
[email protected]. † ETH Zurich. ‡ Paul Scherrer Institute. (1) Intergovernmental Panel on Climate Change (IPCC). Fourth Assessment Report. (2) Green, C.; Baksi, S. Energy Policy 2007, 35, 616–626. (3) Halmann, M.; Steinberg, M. Greenhouse Gas Carbon Dioxide Mitigation. Science and Technology; Lewis Publishing: Boca Raton, FL, 1999. (4) Herzog, H.; Golomb, D. Encyclopedia of Energy 2004, 277–287. (5) Abanades, J. C. Chem. Eng. J. 2002, 90, 303–306. (6) Gupta, H.; Fan, L. S. Ind. Eng. Chem. 2002, 41, 4035–4042. (7) de Diego, L. F.; Garcı´a-Labiano, F.; Ada´nez, J.; Gaya´n, P.; Abad, A.; Corbella, B. M.; Palacios, J. M. Fuel 2004, 83, 1749–1757. (8) Lackner, K. Science 2003, 300, 1677–1678. (9) Nikulshina, V.; Ga´lvez, M. E.; Steinfeld, A. Chem. Eng. J. 2007, 129, 75–83. (10) Haywood, H. M.; Eyre, J. M.; Scholes, H. EnViron. Geol. 2001, 41, 11–16. (11) Goel, N. J. Pet. Sci. Eng. 2006, 51, 169–184. (12) Golomb, D.; Barry, E.; Ryan, D.; Lawton, C.; Swett, P. EnViron. Sci. Technol. 2004, 38, 4445–4450. (13) Edwards, J. H. Catal. Today 1995, 23, 59–66. (14) Omae, I. Catal. Today 2006, 115, 33–52.
Figure 1. Variation of the equilibrium composition as a function of the temperature for 1 mol of CO2 at 1 bar.
composition of CO2 at atmospheric pressure occurs at ultrahigh temperatures; i.e., theoretically 30% dissociation is obtained at above 2700 K, as can be observed in Figure 1. It has been experimentally demonstrated with a prototype solar reactor,15-17 obtaining yields around 5%. Further complication arises from the need to separate the product gases at these high temperatures to avoid recombination or end up with an explosive mixture at ambient temperature. The operating temperature can be reduced and the separation problem bypassed by making use of thermochemical cycles. Of special interest is the two-step-cycle-based metal oxide redox reactions, represented by (15) Traynor, A. J.; Reed, J. J. Ind. Eng. Res. 2002, 41, 1935–1939. (16) Price, R. J.; Morse, D. A.; Hardy, S. L.; Fletcher, T. H.; Hill, S. C.; Reed, J. J. Ind. Eng. Chem. Res. 2004, 43, 2446–2453. (17) Price, R. J.; Fletcher, T. H.; Reed, J. J. Ind. Eng. Chem. Res. 2007, 46, 1956–1967.
10.1021/ef800230b CCC: $40.75 2008 American Chemical Society Published on Web 08/13/2008
Two-Step Solar Thermochemical Cycles
first endothermic step: MxOy ) xM + y/2O2
Energy & Fuels, Vol. 22, No. 5, 2008 3545
(1)
second endothermic step: xM + y/2CO2 ) MxOy + y/2C (2a) or xM + yCO2 ) MxOy + yCO
(2b)
Figure 2 shows schematically the proposed cycle. The two cycle steps encompass (1) an endothermic reaction to thermally dissociate the metal oxide into the metal or a reduced-valence metal oxide and O2 using concentrated solar energy as the source of process heat and (2) a nonsolar exothermic reaction with the reduced metal/metal oxide and CO2, which yields CO and/or C, together with the initial form of the metal oxide. The latter is recycled to the first step. The net reaction is CO2 ) C + O2 (eqs 1 + 2a) or CO2 ) CO + 0.5O2 (eqs 1 + 2b), but in either case, O2 is the only gas released in the first step, thereby eliminating the need for high-temperature gas separation. Similar thermochemical cycles have been proposed for splitting water and converting intermittent solar energy in the form of storable and transportable chemical fuels.18,19 Cd-, Sb-, Zn-, Fe-, and Ni-based cycles were proposed.20 A theoretical analysis was performed for the FeO/Fe3O4 redox pairs to ascertain the optimal temperatures of the two reactions.21 The chemical aspects of the dissociation of ZnO have been previously investigated, and activation energies ranging from 310-350 kJ mol-1 were determined by thermogravimetry.22-25 The reaction proceeds at reasonable rates above 2000 K, and the product gases, Zn(g) and O2, must either be quenched or separated at high temperatures to avoid recombination.26 The solar reactor technology, featuring a rotating cavity-received lined with ZnO particles and directly exposed to high-flux solar irradiation, was demonstrated in a solar furnace.27-29 Less recent experimental data are available for the reaction of Zn with CO2. Most of the previous studies deal with the kinetics of Zn(g) re-oxidation within the imperial smelting process.30-34 The thermal reduction of Fe3O4 to FeO was also thermodynamically (18) Steinfeld, A. Sol. Energy 2005, 78/5, 603–615. (19) Steinfeld, A. Int. J. Hydrogen Energy 2002, 27, 611–619. (20) Martin, L. R. Sol. Energy 1980, 24, 271–277. (21) Palumbo, R. D.; Larson, C. L. Energy 1990, 15, 479–487. (22) Palumbo, R.; Lede, J.; Boutin, O.; Elorza-Riccart, E.; Steinfeld, A.; Moeller, S.; Weidenkaff, A.; Fletcher, E. A.; Bielicki, J. Chem. Eng. Sci. 1998, 53, 2503–2518. (23) Weidenkaff, A.; Reller, A.; Wokaun, A.; Steinfeld, A. Thermochim. Acta 2000, 359, 69–75. (24) Weidenkaff, A.; Steinfeld, A.; Wokaun, A.; Eichler, B.; Reller, A. Sol. Energy 1999, 65, 59–69. (25) Weidenkaff, A.; Reller, A.; Sibieude, F.; Wokaun, A.; Steinfeld, A. Chem. Mater. 2000, 12, 2175–2181. (26) Gstoehl, D.; Brambilla, A.; Schunk, L.; Steinfeld, A. J. Mater. Sci., in press. (27) Haueter, P.; Mo¨ller, S.; Palumbo, R.; Steinfeld, A. Sol. Energy 1999, 67, 161–167. (28) Mu¨ller, R.; Haeberling, P.; Palumbo, R. D. Sol. Energy 2006, 80, 500–511. (29) Schunk, L.; Haeberling, P.; Wepf, S.; Wuillemin, D.; Meier, A.; Steinfeld, A. J. Sol. Energy Eng. 2008, 130, 021009. (30) Clarke, J. A.; Fray, D. J. The rate of deposition and the morphology of zinc oxide deposited from Zn(v)/CO/CO2/Ar gas mixtures. J. Mater. Sci. 1978, 13, 1921–1925. (31) Lewis, L. A.; Cameron, A. M. Metall. Mater. Trans. B 1995, 26, 911–918. (32) Lewis, L. A.; Cameron, A. M. Metall. Mater. Trans. B 1995, 26, 919–924. (33) Cox, A.; Fray, D. J. Trans. Inst. Min. Metall., Sect. C 2000, 109, 97–104. (34) Cox, A.; Fray, D. J. Trans. Inst. Min. Metall., Sect. C 2000, 109, 105–111.
Figure 2. Scheme of the two-step solar thermochemical cycle for CO2 reduction via M/MxOy redox reactions.
examined35 and experimentally demonstrated.36 Recently, stoichiometric and non-stoichiometric forms of wustite (Fe1-yO) from Fe2O3 were produced with 100% conversion at 1973 K using high concentrations of inert gas to dilute O2.37 CO2 reduction has been experimentally demonstrated in exploratory studies with FeO,38,39 oxygen-defective iron oxides,40-43 nanocrystalline Fe2O3,44 and Ni-, Co-, and Cu-doped ferrites.45-50 In the present work, the Zn/ZnO and FeO/Fe3O4 cycles are thermodynamically examined in detail and maximum theoretical energy conversion efficiencies are determined. In this analysis, a specific reactor design is not considered and several assumptions were made for the purpose of comparing the two cycles under the same operating conditions. 2. Chemical Thermodynamic Equilibrium 0 Reactions enthalpies ∆H298 K are listed in Table 1 for both cycles on a 1 mol of CO2 basis. Note that the reduction of CO2 to C(s) requires double amount of ZnO or Fe3O4, as compared to the reduction of CO2 to CO. Thermochemical equilibrium
(35) Steinfeld, A.; Sanders, S.; Palumbo, R. Sol. Energy 1999, 65, 43– 53. (36) Tofighi, A.; Sibieude, F. Int. J. Hydrogen Energy 1984, 9, 293– 296. (37) Charvin, P.; Abanades, S.; Flamant, G.; Lemort, F. Energy 2007, 32, 1124–1133. (38) Ehrensberger, K.; Palumbo, R.; Larson, C.; Steinfeld, A. Ind. Eng. Chem. Res. 1997, 36, 645–648. (39) Yamasue, E.; Yamaguchi, H.; Nakaoku, H.; Okumura, H.; Ischihara, K. N. J. Mater. Sci. 2007, 42, 5196–5202. (40) Tamaura, Y.; Tabata, M. Nature 1990, 346, 255–256. (41) Zhang, C. L.; Li, S.; Wu, T. H.; Peng, S. Y. Mater. Chem. Phys. 1999, 58, 139–145. (42) Zhang, C. L.; Li, S.; Wang, L. J.; Wu, T. H.; Peng, S. Y. Mater. Chem. Phys. 2000, 62, 44–51. (43) Zhang, C. L.; Liu, Z. Q.; Wu, T. H.; Yang, H. M.; Jiang, Y. Z.; Peng, S. Y. Mater. Chem. Phys. 1996, 44, 194–198. (44) Khedr, M. H.; Bahgat, M.; Nasr, M. I.; Sedeek, E. K. Colloids Surf., A 2007, 302, 517–524. (45) Khedr, M. H.; Farghali, A. A. Appl. Catal., B 2005, 61, 219–226. (46) Farghali, A. A.; Khedr, M. H.; Abdelkhalek, A. J. Mater. Process. Technol. 2007, 181, 81–87. (47) Kato, H.; Kodama, M.; Tsuji, M.; Tamaura, Y. J. Mater. Sci. 1994, 29, 5689–5692. (48) Ma, L. Y.; Chen, L. S.; Chen, S. Y. J. Phys. Chem. Solids 2007, 68, 1330–1335. (49) Ma, L. Y.; Chen, L. S.; Chen, S. Y. Mater. Chem. Phys. 2007, 105, 122–126. (50) Siegel, N. P. Proceedings of the 2007 Annual American Institute of Chemical Engineers (AIChE) Meeting, Salt Lake City, UT, Nov 3-9, 2007.
3546 Energy & Fuels, Vol. 22, No. 5, 2008
Ga´lVez et al.
Table 1. Reaction Equations and Enthalpy Change of the Two-Step Thermochemical Cycles for CO2 Reduction via Zn/ZnO (Cycle 1) and FeO/Fe3O4 (Cycle 2) Redox Reactions, Yielding Either C or CO 0 -1 ∆H298 K (kJ mol )
cycle 1
cycle 2
(1a′) 2ZnO ) 2Zn + O2 (2a′) 2Zn + CO2 ) 2ZnO + C (1b′) ZnO ) Zn + 0.5O2 (2b′) Zn + CO2 ) ZnO + CO (1a′′) 2Fe3O4 ) 6FeO + O2 (2a′′) 6FeO + CO2 ) 2Fe3O4 + C (1b′′) Fe3O4 ) 3FeO + 0.5O2 (2b′′) 3FeO + CO2 ) Fe3O4 + CO
701 -307.5 350.5 -67.5 633.1 -239.6 316.6 -33.6
equilibrium products and large amounts of FeO in stoichiometric and non-stoichiometric forms (i.e., FeO1.056 and Fe1-yO) and CO2. The changes in the slopes were far less pronounced than those of the Zn/CO2 reactions. The reaction extents with respect to CO and C are defined in eqs 3 and 4, respectively, as XCO )
XC )
code,51
calculations were performed using HSC Outokumpu selecting the elements involved in each system, assuming ideal mixtures, and omitting species with mole fractions of less than 10-5. Parts A and B of Figure 3 show the variation of the equilibrium composition with temperature at 1 bar for the systems: 2Zn + CO2, Zn + CO2, 6FeO + CO2, and 3FeO + CO2. For the 2Zn + CO2 system, CO2 is completely decomposed to C and Zn is fully converted to ZnO below 1000 K. Above 1000 K, ZnO is carbothermally reduced to produce Zn. At above 1200 K, Zn(g), CO, and CO2 are the only stable products. Hence, complete reduction of CO2 to C is thermodynamically possible only below 1000 K. For the Zn + CO2 system, Zn is completely converted to ZnO within the range of 570-1000 K. Below 570 K, the formation of ZnCO3 is thermodynamically favored. At above 1000 K, ZnO is reduced to Zn(g) in the presence of CO. C formation peaks at 700 K and decreases in accordance with the Boudouard equilibrium, while CO formation peaks at 1130 K. For the 6FeO + CO2 system, CO2 is completely reduced to C at below 550 K. FeO is oxidized primarily to Fe3O4 with residual amounts of Fe2O3. Increasing the temperature results in larger amounts of nonstoichimetric FeO. Also, at elevated temperatures, carbothermal reduction takes place, yielding pure Fe. CO is stable only above 800 K, and the formation of iron carbides is not seen at these low temperatures. For the 3FeO + CO2 system, C is produced by oxidization of FeO to Fe2O3 and Fe3O4 at low temperatures, while nearly 40% by volume of CO2 remains unreacted. Similar to the 6FeO + CO2 system, CO is not stable below 800 K, while at higher temperatures, the FeO remains unreacted and only negligible amounts of Fe are formed. Figure 4 shows the enthalpy change of these reactions as a function of the temperature assuming the reactants at 298 K and the corresponding equilibrium products at given temperatures. The 2Zn + CO2 system turns endothermic at 1210 K. The change in the slope occurring in the range of 1100-1400 K is due to the carbothermal reduction of ZnO, leading to the production of Zn(g), CO, and CO2. The Zn + CO2 system turns endothermic at about 1190 K. The change in the slope between 390 and 480 K is associated with the gradual decomposition of ZnCO3. A further change in the slope takes place in the range of 700-1100 K, mainly because of the conversion of CO2 and C to CO following the Boudouard equilibrium. Finally, at above 1100 K, a final change in the slope is related to the carbothermal reduction of ZnO in the presence of CO. Both FeO/CO2 reactions are thermodynamically favorable below 975 K and change from exo- to endothermic at just above 700 K. The slopes of the curves for both reactions increase at 600 K, which is indicative of less C being produced and more unreacted CO2. A final change in the slopes occurs at around 800 K, which corresponds to the appearance of small amounts of CO in the (51) Roine, A. Outokumpu HSC Chemistry for Windows 5.0. Outokumpu ResearchPori, Finland, 1997.
eq nCO i nCO 2
neq C i nCO 2
(3)
(4)
where neq and ni denote the number of moles in equilibrium and available for producing CO/C, respectively. Figure 5 shows the reaction extent with respect to CO and C as a function of the temperature at 0.1, 1, and 10 bar for 2Zn + CO2, Zn + CO2, 6FeO + CO2, and 3FeO + CO2 systems. For the Zn/CO2 reactions, an increase in the pressure shifts the reaction extent curves to higher temperatures. For a given temperature, this results in a XCO reduction followed by a XC increase. Similar trends are also seen for the FeO/CO2 reactions. A change in pressure forces shifts in the amounts of CO and C produced. At lower pressures, less C is produced at elevated temperature but is offset with increased CO. However, for the temperature ranges considered for the FeO/CO2 reactions, the amount of CO produced continues to increase and does not reach a maximum, as is the case for the Zn/CO2 reactions. For all reactions at higher pressures, the formation of C is favored according to Le Chatelier’s principle. 3. Second-Law Analysis This section presents a second-law (exergy) analysis that assesses the maximum possible energy conversion efficiency of the CO2-splitting solar thermochemical cycle using the proposed two-step Zn/ZnO and FeO/Fe3O4 redox reactions. A flow diagram for a general CO2-splitting cycle is shown schematically in Figure 6. It is comprised of a solar reactor, a quench unit, and a CO2-reducer. Readily available CO2 is assumed, after capture.1 The molar flow rate of CO2 to the CO2reducer is set to 1 mol/s to produce either CO or C, which implies different molar flow rates of the metal oxide to the solar reactor according to the stoichiometry provided in Table 1. The complete process is assumed to be carried out at steady state at a constant pressure of 1 bar. In practice, pressure drops will occur throughout the system, and pumping work will be required. Heat exchangers for recovering sensible latent heat are omitted from consideration. Additional assumptions include the following: the solar reactor is a blackbody absorber; all products separate naturally without expending work; kinetic and potential energies are neglected; and all reactions reach completion. The analysis follows the methodology and governing equations derived previously for H2O-splitting solar thermochemical cycles.19,35 The thermodynamic properties were all taken from HSC 5.0.51 The solar–chemical energy conversion efficiency is that defined as the portion of solar energy that is converted into chemical energy given by the Gibbs free energy of the products, i.e., the maximum possible amount of work that can be extracted from the products when transformed back to the reactants at 298 K in a reversible, ideal fuel cell ηsolar-chemical )
-n˙∆Gproducts|298 K WFC,ideal ) Qsolar Qsolar
(5)
Two-Step Solar Thermochemical Cycles
Energy & Fuels, Vol. 22, No. 5, 2008 3547
Figure 3. (A) Variation of the equilibrium composition with temperature at 1 bar of the carbon species for the chemical systems 2Zn + CO2, Zn + CO2, 6FeO + CO2, and 3FeO + CO2. (B) Variation of the equilibrium composition with temperature at 1 bar of the Zn or Fe species for the chemical systems 2Zn + CO2, Zn + CO2, 6FeO + CO2, and 3FeO + CO2.
The solar energy absorption efficiency is defined as the net rate at which energy is being absorbed divided by the solar power input through the aperture of the solar reactor. Assuming a perfectly insulated blackbody cavity-receiver (no convection or conduction heat losses; effective absorptivity and emissivity approaching 1), it is given by ηabsorption )
( )
Qreactor,net σTR4 )1Qsolar IC
(6)
where I is normal beam insolation, C is the solar flux concentration ratio of the solar concentrating system (The solar flux concentration ratio C is defined as the ratio of the solar flux intensity achieved after concentration to the normal beam insolation. It is a non-dimensional number, usually reported in units of “suns”), TR is the nominal reactor temperature, and σ is the Stefan-Boltzmann constant. ηabsorption expresses the capability of a solar reactor to absorb incoming concentrated
solar energy but does not include the losses incurred in collecting and concentrating solar energy [Losses in solar thermal power and solar flux concentration depend upon the power level and whether parabolic troughs, power towers, or distributed dishes are used and are due to geometrical imperfections (such as misalignment and segmented approximation to the exact reflector profile), optical imperfections (such as reflectivity and specularity of the mirrors, glass absorption, and shading effects), and tracking imperfections]. Assuming reactants enter the reactor at TL, are then heated to TR, and undergo full conversion, the net power absorbed in the solar reactor equals the rate of enthalpy change Qreaction,net ) n˙∆H|MxOy at TLfxM+y/2O2 at TR
(7)
Irreversibility in the solar reactor arising from the non-reversible chemical transformation and re-radiation losses to the surroundings at TL are
3548 Energy & Fuels, Vol. 22, No. 5, 2008
Irrreactor )
Ga´lVez et al.
-Qsolar Qre-radiation + + n˙∆S|MxOy at TLfxM+y/2O2 at TR TR TL (8)
assuming the heat source, obtained by absorbing concentrated solar radiation in the blackbody cavity-receiver, is at TR. The radiation heat losses from the reactor are Qre-radiation ) (1 - ηabsorption)Qsolar
(9)
To avoid recombination, the products M and O2 exit the reactor at TR and are rapidly cooled to TL in a quench unit. It is assumed that the chemical composition of the products remain unchanged upon cooling. The latent and sensible heat rejected to the surroundings and the irreversibility arising from this quench unit are given by Qquench ) -n˙∆H|xM+y/2O2 at TRfxM+y/2O2 at TL
(10)
Qquench Irrquench ) + n˙∆S|xM+y/2O2 at TRfxM+y/2O2 at TL TL
(11)
After quenching, the products are assumed to separate naturally without expending work. The reduced metal/metal oxide is then sent to the CO2-reducer to react with CO2 and form solid C or
CO, as represented by eq 2a or eq 2b, respectively. This second step is exothermal; thus, heat is released according to eqs 12a and 12b, for the production of C and CO, respectively QCO2-reducer ) -n˙∆H|xM+y/2CO2 at TLfMxOy+y/2C at TL
(12a)
QCO2-reducer ) -n˙∆H|xM+y/2CO2 at TLfMxOy+y/2CO at TL (12b) Work expenditure for product separation has been neglected. If CO is produced, the gaseous products separate naturally from the solid metal oxide. If C is produced, additional energy input would be necessary. The heat released in the CO2-reducer is assumed to be lost to the surroundings. Thus, the irreversibility associated with reducing CO2 to C and CO are given by eqs 13a and 13b, respectively, as IrrCO2-reducer )
QCO2-reducer TL
+ n˙∆S|xM+y/2CO2 at TLfMxOy+y/2C at TL (13a)
IrrCO2-reducer )
QCO2-reducer TL
+ n˙∆S|xM+y/2CO2 at TLfMxOy+y/2CO at TL (13b)
Figure 4. Standard enthalpy change as a function of the temperature for the chemical systems 2Zn + CO2, Zn + CO2, 6FeO + CO2, and 3FeO + CO2, assuming reactants at 298 K and the corresponding equilibrium products at a given temperature T.
Figure 5. Reaction extent as a function of the temperature at 0.1, 1, and 10 bar for the chemical systems 2Zn + CO2, Zn + CO2, 6FeO + CO2, and 3FeO + CO2. Equilibrium composition is assumed.
Two-Step Solar Thermochemical Cycles
Energy & Fuels, Vol. 22, No. 5, 2008 3549
Figure 6. Model flow diagram of the two-step solar thermochemical cycle for CO2 reduction applied for the second-law analysis. Table 2. Second-Law Analysis of the Two-Step Thermochemical Cycles for CO2 Reduction via Zn/ZnO (Cycle 1) and FeO/Fe3O4 (Cycle 2) Redox Reactions, Yielding Either C or CO, Based on the Process Modeling Depicted in Figure 6a CO2 ) C + O2
net reaction cycle
1
2
1333.7 1776.8 Qsolar (kW) Qre-radiation (kW) 242.0 322.4 Qreactor,net (kW) 1091.7 1454.4 -1 Irrreactor (kW K ) 0.73 1.26 Qquench (kW) 390.7 821.3 Irrquench (kW K-1) 0.93 1.94 QCO2 decomposer (kW) 307.5 239.6 -1 IrrCO2 decomposer (kW K ) 0.83 0.54 QFC,ideal (kW) -0.86 WFC,ideal (kW) 394.4 ηabsorption (%) 82 ηsolar-chemical (%) 30 22
CO2 ) CO + 0.5O2 1
2
666.9 121.0 545.9 0.36 195.4 0.46 67.5 0.21
888.4 161.2 727.2 0.63 410.6 0.97 33.6 0.07
25.8 257.2 82 39
The baseline parameters are nCO2 ) 1 mol/s, I ) 1 kW 5000 suns, TR ) 2000 K, and TL ) 298 K. a
29 m-2,
C)
According to the definition given in eq 5, the solar-chemical energy conversion efficiency is calculated for C and CO, respectively, as ηsolar-chemical ) ηsolar-chemical )
WFC,ideal -n˙∆G|C+O2 at TLfCO2 at TL ) (14a) Qsolar Qsolar
WFC,ideal -n˙∆G|CO+0.5O2 at TLfCO2 at TL ) Qsolar Qsolar (14b)
Besides work, the ideal fuel cell releases heat in an amount given by QFC,ideal ) -TLn˙∆S|C+O2 at TLfCO2 at TL
(15a)
QFC,ideal ) -TLn˙∆S|CO+0.5O2 at TLfCO2 at TL
(15b)
The results of the second-law analysis are summarized in Table 2 for the two different cycles 1 and 2 and for CO2 reduction yielding either C or CO. The baseline parameters are n ) 1 mol/s, I ) 1 kW m-2, C ) 5000 suns, TR ) 2000 K, and TL ) 298 K. For cycle 1 (Zn/ZnO), ηsolar-chemical reaches 30 and 39% for the case of C and CO production, respectively. For cycle 2 (FeO/Fe3O4), ηsolar-chemical reaches 22 and 29% for the case of
C and CO production, respectively. Higher efficiencies for cycle 1 than for cycle 2 are attributed to two factors: (1) the lower enthalpy change of ZnO dissociation, resulting in 25% lower Qsolar, and (2) the lower heat capacities (on a molar basis) for Zn and ZnO compared to FeO and Fe3O4, resulting in a reduction of Qquench by a factor of more than 2, as also indicated by the calculation of the irreversibility arising from the quench unit. In general, CO2 reduction to CO results in higher efficiencies than reduction to C as a consequence of the better use of the chemical energy within the cycle; the decomposition to C requires twice as much M compared to the reduction to CO, while the theoretical work output of a C-fuel cell (394.4 kW) is less than twice as much as that of a CO-fuel cell (257.0 kW). The major sources of irreversibility are associated with the re-radiation losses and the quenching of the products: 18% of Qsolar is lost by re-radiation, and 29 and 46% are lost in the quenching for cycles 1 and 2, respectively. A recent work introduces a novel reactor design for Fe-based thermochemical cycles that eliminates the need for quenching the products after the first solar-driven step.52 However, this design cannot be applied to Zn-based cycles because Zn is obtained in the gaseous phase and quenching is required for avoiding its re-oxidation. It is worth mentioning that the Zn/ZnO cycle for reducing CO2 to CO has a higher theoretical work potential than reducing H2O to H2 with the same quantity of Zn.19 In general, the second-law analysis indicates that a favorable aspect of using solar energy at high temperatures is the potential of achieving high solar-chemical conversion efficiencies. High efficiencies directly translate to lower solar collection area and associated costs of the heliostat field, which amount to 40-50% of the capital cost for the entire solar CO2-splitting plant. Numerous assumptions were made in these analyses that will need to be accounted for in future research and reactor development. Of particular interest is the separation of the products, especially in the cases where carbon is produced and mixed with MxOy. Although the main goal of C production is the definitive and stable sequestration of CO2, combustion of MxO/C delivers high-temperature heat that may be used for driving a heat engine, while avoiding the work required to extract carbon from the mixture. 4. Summary and Conclusions Two-step thermochemical cycles for CO2-splitting via Zn/ ZnO and FeO/Fe3O4 redox reactions have been thermodynamically examined. Thermochemical equilibrium composition calculations at 1 bar showed that CO2 can be reduced to C and CO with Zn(s) at below 1000 K and above 700 K, respectively, and with FeO at below 550 K and above 800 K, respectively. In the case of the stoichiometric Zn + CO2 and 3FeO + CO2 reactions, C(s) formation reaches maximums below 700 and 300 K, respectively. For all cycles, higher pressures favor the formation of C. The second-law analysis using a solar blackbody receiver-reactor at 2000 K subjected to a solar concentration of 5000 suns indicates a maximal solar-chemical energy conversion efficiency (given by ηsolar-chemical ) -n˙∆G|CO + 0.5O2 f CO2/ Qsolar) of 39% for the Zn/ZnO cycle and 29% for the FeO/Fe3O4 cycle. Higher efficiencies are obtained for the Zn/ZnO cycle because of the lower solar energy input required and lower heat loss by quenching per mole of CO2 reduced. Major sources of irreversibility are associated with the re-radiation losses of the solar reactor operating at 2000 K and the quenching of the (52) Diver, R. B.; Siegel, N. P. Moss, T. A.; Miller, J. E.; Stueker, J. N.; James, D. L. Proceedings from 14th Biennial Concentrating Solar Power (CSP) SolarPACES Symposium, Las Vegas, NV, March 4-7, 2008.
3550 Energy & Fuels, Vol. 22, No. 5, 2008
products exiting the solar reactor. These results provide a foundation for pursuing an experimental study for reducing CO2 with Zn and FeO. Additional measures could be applied in a real system to increase the overall efficiencies that were not considered in these analyses. For example, waste heat may be recovered from the quenching process and the exothermic xM + CO2 reaction. Nomenclature C ) solar flux concentration ratio, suns I ) normal beam solar insolation, W m-2 IrrCO2-reducer ) irreversibility associated with the CO2-reducer, kW K-1 Irrquench ) irreversibility from the quench unit, kW K-1 Irrreactor ) irreversibility from the reactor, kW K-1 MxOy ) metal oxide M ) metal or reduced-valence metal oxide XCO ) reaction extent with respect to CO XC ) reaction extent with respect to C eq ) number of moles of CO in equilibrium, moles nCO n˙ ) molar flow rate, mol/s
Ga´lVez et al. nCeq ) number of moles of C in equilibrium, moles i nCO2 ) number of moles of CO2 available for producing CO and C, moles QCO2-reducer ) heat released from the CO2-reducer, kW QFC,ideal ) heat rejected from the ideal fuel cell, kW Qquench ) heat reject to the surroundings during the quench process, kW Qreactor,net ) net energy absorbed in the reactor, kW Qre-radiation ) radiation heat loss in the reactor, kW Qsolar ) solar energy input, kW TR ) nominal reactor temperature, K TL ) surroundings temperature, K WFC,ideal ) work output of an ideal fuel cell, kW ∆G ) Gibb free-energy change, kJ mol-1 ∆H ) enthalpy change, kJ mol-1 ∆S ) entropy change, kJ mol-1 K-1 ηsolar-chemical ) solar-chemical energy conversion efficiency ηabsorption ) solar absorption efficiency σ ) Stefan-Boltzmann constant, 5.670 × 10-8 W m-2 K-4 EF800230B