Fundamental aspects of chemical-looping with oxygen uncoupling using copper oxide-based oxygen carriers

The chemical-looping technology is quickly becoming an attractive alternative for the combustion of fossil fuels in energy production. With the rapid growth in the anthropogenic production of carbon dioxide (CO2) due, in large part, to the combustion of fossil fuels, it is becoming increasingly important to identify technologies that are capable of producing energy, at the same rate as traditional fossil-fuels units, while serving in a secondary capacity of managing emissions. Among the materials suggested for use in chemical-looping, copper-oxide has emerged as a front runner. Materials such as copper oxide allow the direct combustion of solid fuels, without an intermediary gasification step, by spontaneously decomposing from cupric oxide (CuO) to cuprous oxide (Cu2O), liberating free oxygen in the process. The rate at which these particles decompose can be as much as 50 times faster than the rates of gasification. This work provides the academic community with results geared toward the production and implementation of copper-based oxygen carriers by: 1. Developing novel oxygen carrier materials and, 2.Developing models describing both the oxidation and decomposition of copper oxide-based materials. It has been determined that the decomposition of copper-oxide is well described by a global reaction rate for a wide range of copper oxide-based oxygen carrier materials. This global reaction rate suggests that the activation energy of decomposition that best predicts reaction rates of the different CuO-based materials is 62 kJ/mol and that the rate law has a zero order dependence on the concentration of the solid carrier. The modeled oxidation kinetics of Cu2O to CuO for two different oxygen carrier materials is presented. Unlike the simple case of decomposition of CuO, the oxidation of Cu2O is highly dependent on the solid concentration. During oxidation, the volume change of the solid is about a 5% increase. This causes a pore-blocking effect which is observed at low temperatures (below 700°C). However, at higher temperatures (above 800°C), this effect is not apparent. These two regimes are best described by two different kinetic models: pore-blocking kinetics is used for the lower temperature regime while nucleation/growth kinetics, given by Avrami-Erofeev, is used to model the higher temperature regimes.

FUNDAMENTAL ASPECTS OF CHEMICAL-LOOPING WITH OXYGEN UNCOUPLING USING COPPER OXIDE-BASED OXYGEN CARRIERS by Christopher K. Clayton A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemical Engineering The University of Utah May 2016 i Copyright © Christopher K. Clayton 2016 All Rights Reserved i The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL The dissertation of Christopher K. Clayton has been approved by the following supervisory committee members: Kevin J. Whitty , Chair 3/18/2016 Date Approved JoAnn S. Lighty , Member 8/12/2015 Date Approved Swomitra K. Mohanty , Member 2/26/2016 Date Approved Terry A. Ring , Member 8/12/2015 Date Approved Hong Y. Sohn , Member 8/12/2015 Date Approved and by Milind D. Deo , Chair of the Department of Chemical Engineering and by David B. Kieda, Dean of The Graduate School. ii ABSTRACT The chemical-looping technology is quickly becoming an attractive alternative for the combustion of fossil fuels in energy production. With the rapid growth in the anthropogenic production of carbon dioxide (CO2) due, in large part, to the combustion of fossil fuels, it is becoming increasingly important to identify technologies that are capable of producing energy, at the same rate as traditional fossil-fuels units, while serving in a secondary capacity of managing emissions. Among the materials suggested for use in chemical-looping, copper-oxide has emerged as a front runner. Materials such as copper oxide allow the direct combustion of solid fuels, without an intermediary gasification step, by spontaneously decomposing from cupric oxide (CuO) to cuprous oxide (Cu2O), liberating free oxygen in the process. The rate at which these particles decompose can be as much as 50 times faster than the rates of gasification. This work provides the academic community with results geared toward the production and implementation of copper-based oxygen carriers by: 1. Developing novel oxygen carrier materials and, 2. Developing models describing both the oxidation and decomposition of copper oxide-based materials. It has been determined that the decomposition of copper-oxide is well described by a global reaction rate for a wide range of copper oxide-based oxygen carrier materials. iii This global reaction rate suggests that the activation energy of decomposition that best predicts reaction rates of the different CuO-based materials is 62 kJ/mol and that the rate law has a zero order dependence on the concentration of the solid carrier. The modeled oxidation kinetics of Cu2O to CuO for two different oxygen carrier materials is presented. Unlike the simple case of decomposition of CuO, the oxidation of Cu2O is highly dependent on the solid concentration. During oxidation, the volume change of the solid is about a 5% increase. This causes a pore-blocking effect which is observed at low temperatures (below 700°C). However, at higher temperatures (above 800°C), this effect is not apparent. These two regimes are best described by two different kinetic models: pore-blocking kinetics is used for the lower temperature regime while nucleation/growth kinetics, given by Avrami-Erofeev, is used to model the higher temperature regimes. iv I would like to dedicate this work to my beautiful wife and loving children. Without your support this does not exist. My love, always. v Anybody who has been seriously engaged in scientific work of any kind realizes that over the entrance to the gates of the temple of science are written the words: "Ye must have faith." It is a quality which the scientist cannot dispense with. ~Max Planck: Where is Science Going? (1932) vi TABLE OF CONTENTS ABSTRACT....................................................................................................................... iii LIST OF TABLES ............................................................................................................. ix ACKNOWLEDGEMENTS ................................................................................................ x CHAPTERS 1. INTRODUCTION .......................................................................................................... 1 1.1 CO2 Capture Technologies ............................................................................... 3 1.2 Traditional Chemical-looping Combustion (CLC) ........................................... 5 1.3 Chemical-looping with Oxygen Uncoupling (CLOU) ..................................... 7 1.4 Selection of Material for Oxygen Carriers........................................................ 9 1.5 Objectives of This Work ................................................................................. 12 1.6 References ....................................................................................................... 15 2. REVIEW OF LITERATURE ....................................................................................... 18 2.1 CLC ................................................................................................................. 18 2.2 CLOU .............................................................................................................. 21 2.3 Kinetics of CLOU ........................................................................................... 22 2.4 Combustion of Solid Fuels ............................................................................. 27 2.5 References ....................................................................................................... 30 3. EXPERIMENTAL APPROACH ................................................................................. 34 3.1 Oxygen Carrier Selection................................................................................ 34 3.2 Physical Characterization................................................................................ 38 3.3 Chemical Testing ............................................................................................ 41 3.4 Solid Fuels Testing ......................................................................................... 49 3.5 References ....................................................................................................... 54 4. RESULTS ..................................................................................................................... 55 4.1 Preliminary and CLOU Operation of Carriers ................................................ 55 4.1.1 Ilmenite ................................................................................................ 55 4.1.2 Pure Copper ......................................................................................... 58 4.1.3 Supported Copper-based Oxygen Carriers .......................................... 60 vii 4.1.4 Oxidation and Decomposition Kinetics ............................................... 72 4.1.5 Production and Evaluation of Novel Materials .................................... 78 4.2 Conversion of Solid Fuels ............................................................................... 80 4.3 References ....................................................................................................... 91 5. CONCLUSIONS........................................................................................................... 92 APPENDICES A: MEASUREMENT AND MODELING OF DECOMPOSITION KINETICS FOR COPPER OXIDE-BASED CHEMICAL LOOPING WITH OXYGEN UNCOUPLING ........................................................................................................................................... 96 B: OXIDATION KINETICS OF CU2O IN OXYGEN CARRIERS FOR CHEMICAL LOOPING WITH OXYGEN UNCOUPLING............................................................... 105 C: CHARACTERISTICS AND CLOU PERFORMANCE OF A NOVEL SIO2SUPPORTED OXYGEN CARRIER PREPARED FROM CUO AND β-SIC .............. 117 D: ILMENITE AS AN INERT SUPPORT FOR COPPER-BASED OXYGEN CARRIER MATERIAL FOR USE IN CLOU ................................................................................. 126 ........................................................................................................................................................... viii LIST OF TABLES Tables 1: United States electricity production by source. Adapted from U.S. EIA, 2011. ............2 2: The melting points of different CLOU capable species. ...............................................10 3: List of CLOU reactions and the resultant heats of reaction. ..........................................12 4: Copper-based materials tested at the University of Utah for suitability as oxygen carriers within a fluidized-bed based on CLOU. ...............................................................36 5: Characteristic ranges used for the TGA and fluidized-bed reactors. ............................42 6: Ultimate and proximate analyses of solid fuels used in this study. ..............................50 7: Materials tested for suitability as oxygen carriers in chemical-looping. ......................56 8: SEM/EDS Micrographs of four copper oxide-based materials tested for oxygen carrier suitability in CLOU operation. Courtesy: Crystal Allen. ..................................................66 9: Surface area and pore size data for four copper based carriers studied as oxygen carriers in CLOU systems. Data courtesy of Crystal Allen. .............................................67 10: Results of crushing strength testing performed on four materials tested for suitability as oxygen carriers in CLOU systems. Courtesy: Crystal Allen ........................................67 11: Constants for kinetic rate expression modeling the decomposition of CuO................75 12: Results obtained from modeling of low temperature oxidation of two CLOU materials. ............................................................................................................................77 13: Model used to predict the conversion of Cu2O in two different oxygen carriers using the nucleation/growth kinetics expression Avrami-Erofeev. .............................................79 14: Measured agglomeration temperature in the fluidized-bed ........................................81 15. Summary of oxygen carrier materials prepared. ........................................................132 16: Illinois #6 coal proximate, ultimate and heating value analysis. ..............................136 ix ACKNOWLEDGEMENTS I would like to thank my wife and kids for the love and support they have given throughout the course of this work. I would also like to thank my parents for their support - thank you for instilling within me from such a young age the desire to achieve greatness. To my siblings and in-laws, there have been many a night that you have helped to inspire and support me. Thanks go to Crystal Allen, Blake Wilde, Gabor Konya, Sean Peterson and Edward Eyring for assisting with all the experimental work. To my committee, thank you for your time and your thoughts without which I would have been lost. To my advisor, Kevin, thank you for being patient with and pushing me when I didn't want to be pushed. Not to forget the money, this material was based upon work supported by the Department of Energy under Award Number DE-NT0005015. x 1 CHAPTER 1 INTRODUCTION According to the U.S. Department of Energy (DOE), of the world's total carbon dioxide (CO2) emissions, the United States is responsible for more than 18%, producing nearly 5.5 million metric tons in 2010 - second only to China. The greatest source for production of CO2 is the combustion of fossil fuels. From study of Figure 1, it may be understood that the United States is heavily dependent on fossil fuels. Table 1 tabulates the amount of electricity produced from different sources. Nearly 70% of electricity produced in the U.S. in 2010 was created from fossil fuel sources. This reliance on the combustion of fossil fuels has led to a near 2 ppm/year increase of CO2 globally according to the National Oceanic and Atmospheric Administration (NOAA) [1]. According to the United States Environmental Protection Agency (EPA) CO2 is the largest of the reported greenhouse gases emitted in 2011 at nearly 84% of the total reported greenhouse gases emitted [2]. The emission of greenhouse gases, specifically CO2, into the atmosphere along with the allegedly induced effects has been the subject of numerous studies. These studies have sparked a growing concern on the effects of these gases on the environment. This concern has, in turn, driven researchers to develop technologies directed toward the mitigation of these effects. These technologies are generally aimed at reducing emissions at the source as opposed to 2 Nuclear 19% Renewable 10% Misc. 1% Fossil Fuel 70% Figure 1: United States electricity production by source in 2010. Adapted from U.S. EIA 2011. Table 1: United States electricity production by source. Adapted from U.S. EIA, 2011. Year Fossil Fuels Nuclear Renewable Misc. Total 2011 2790.3 790.2 520.1 - - 2010 2883.4 807.0 427.4 33.3 4151.0 Proportion 2010 69.5% 19.4% 10.3% 0.8% 100.0% 2009 2726.5 798.9 417.7 41.4 3984.4 2008 2926.7 806.2 380.9 38.3 4152.2 3 environmental cleanup post emission. 1.1 CO2 Capture Technologies Current CO2 capture techniques are expensive and require large pieces of equipment. The removal of CO2 gas from a gaseous stream of mixed gaseous species can be performed either precombustion, postcombustion or in-situ. Although there are other greenhouse gases being emitted, due to the vast difference between emission rates (84% of GHG emissions are CO2), it seems prudent to focus a large portion of our resources toward the carbon emission mitigating technologies. The U.S. Department of Energy via the National Energy and Technology Laboratories (NETL) has selected various technologies in each of the categories while assigning priority to the research of these technologies for carbon capture and storage (CCS). For postcombustion solutions, they have listed various solvents, solid sorbents, and membranes. For precombustion solutions, they have identified physical solvents, solid sorbents, and membranes, including H2/CO2 and water-gas-shift membranes, as potentially suitable CCS technologies. For the oxy-combustion approach (also sometimes referred to as in-situ), the list of likely suitable candidate technologies includes oxy-fuel combustion and chemical-looping [3]. Although many challenges still exist in the development of the chemical-looping combustion technology, the payoff seems to be well worth it. In the same report, the NETL reported the results of their study to determine the cost reduction benefits versus an expected demonstration ready unit time frame. These results have been reproduced in Figure 2. Chemical-looping technology has been identified as having the greatest 4 Cost Reduction Benefits † - Post Combustion (existing, new PC) ‡ - Post Combustion (IGCC) • - Oxy-combustion (new PC) Chemical Looping, ‡ • H2 and CO2 Membranes, † ‡ Amine Solvents, † Physical Solvents, ‡ Cryogenic Oxygen, ‡ • Current 1st Gen Physical Solvents, ‡ Physical Solvents, ‡ 2015 2nd Gen Solvents, † ‡ 2nd Gen Oxyboiler, • Biological Processes, ‡ † Oxygen Membranes, ‡ Solid Sorbents, ‡ † 2020 2025 2030 Expected Demonstration Ready Figure 2: Expected CCS technologies demonstration date compared by cost reduction benefits according to the U.S. Department of Energy. 5 expected cost reduction benefit while also indicating the many challenges still existing in its development by assigning the demonstration ready time for some time after 2025 [4, 5]. 1.2 Traditional Chemical-looping Combustion (CLC) The name Chemical-looping combustion was first coined in 1987 by Ishida, Zheng and Akehata in their exergy analysis of a chemical-looping combustor [6]. From there, the Japanese group seemed to take the forefront of CLC studies and reported work completed from several other studies in the 1990s [7-11]. Mattisson [12] reports that by the new millennium, there were just a handful of institutions producing the bulk of the studies into chemical-looping, namely: Chalmers University of Technology in Sweden [13, 14, 27,16] , ICB-CSIC Spain [7, 15 , 22, 25], the Technical University of Vienna [16, 17, 18] and the Korea Institute of Energy Research [19, 22-25] with a few other institutions beginning studies [26, 27]. From then, several institutions have investigated various transition metals and their oxides for suitability as oxygen carriers (OC) in a few different reactor designs, including transport reactors, bubbling-fluidized-beds, and moving- and fixed-bed reactors. This technology has become increasingly popular among researchers as given by the number of literature sources reporting chemicallooping studies increasing, seemingly exponentially, over the past few years. Mattisson references over 130 sources in his latest review of the technology while focusing mainly on sources presenting results for the nontraditional CLC called Chemical-looping with Oxygen Uncoupling (CLOU) [12]. CLC takes advantage of a the redox characteristics of a material, called the 6 oxygen carrier (OC), which is typically a transition metal, by oxidizing the material at high-temperatures using an oxygen rich stream, typically air, in one reactor (the Air Reactor - AR) and reducing that material in a second reactor (Fuel Reactor - FR) using a fluid fuel [3, 20, 21, 22, 27]. The basic idea behind CLC may be visualized in Figure 3. The air for oxidation is introduced in a reactor separate from the fuel; in so doing, the product CO2 from combustion is inherently separated and may be purified further by condensing the product water from the effluent stream. This inherent separation of CO2 is what makes the CLC technology so attractive. The drawback of CLC lies in the challenge of utilizing solid fuels. For a traditional CLC oxygen carrier, the fuel directly reduces the carrier; therefore, a fluid fuel is necessary as the reaction rates of a solid-solid system are highly unfavorable. Therefore, in order to utilize a chemical-looping combustion system with solid fuels, the fuel must first be gasified. CO2 H2O Air Metal Oxide (MexOy) Air Reactor (AR) Fuel Reactor (FR) Metal (MexOy-1) N2 O2 Fuel Figure 3: Basic schematic of CLC technology. 7 1.3 Chemical-looping with Oxygen Uncoupling (CLOU) The use of traditional CLC materials has been successfully performed and reported in a multitude of studies. Moreover, these oxygen carriers have been tested with solid fuels and syngas mixtures. In an attempt to bypass the necessary gasification step, Mattisson et al. from Chalmers University of Sweden uncovered a technology they coined "Chemical-looping with Oxygen Uncoupling." They discovered that with certain materials, there exists a thermodynamic regime within which these materials may be oxidized in an oxygen rich atmosphere and may be reduced by subjecting the oxidized metal to an oxygen-depleted atmosphere [22]. This spontaneous reduction of the metal oxide may be more clearly understood when termed "decomposition" as opposed to "reduction." The mechanism for this reaction is altered from traditional CLC only within the FR. Although traditional CLC relies on direct reduction of the OC from the fuel, CLOU first decomposes liberating gaseous oxygen which is then available for combustion. This oxygen then reacts with the fuel, once again depleting the atmosphere of oxygen driving the decomposition. The generalized reactions occurring within the fuel reactor for both CLC and CLOU are given in equations (1) - (3). Equation (2) defines the decomposition of the metal oxide and equation (3) gives the combustion of the fuel using the oxygen liberated by reaction (2). It may be readily shown that the sum of equations (2) and (3) is equation (1). Although the name CLOU was not coined until 2009, this technology was actually tested and reported more than 50 years prior by Lewis, Gilliland and Sweeney [23]. Their results suggested that combustion of the solid fuel was roughly 50 times 8 faster, using the oxygen liberated from CuO decomposition, than the rate of gasification [24]. CLC (2𝑛 + 𝑚)𝑀𝑀𝑥 𝑂𝑦 + 𝐶𝑛 𝐻2𝑚 → (2𝑛 + 𝑚)𝑀𝑀𝑥 𝑂𝑦−1 + 𝑛𝐶𝐶2 + 𝑚𝐻2 𝑂 (1) CLOU 2𝑀𝑀𝑥 𝑂𝑦 → 2𝑀𝑀𝑥 𝑂𝑦−1 + 𝑂2 𝐶𝑛 𝐻2𝑚 + 𝑛 + 𝑚 𝑂 → 𝑛𝑛𝑛2 + 𝑚𝑚2 𝑂 2 2 (2) (3) The determination of a suitable oxygen carrier is pivotal to the success of CLC/CLOU research and implementation. These materials must exhibit suitable physical and chemical reactivity characteristics. Typically, oxygen carrier materials are put upon a supporting substrate to improve both physical durability and internal surface area of the particles. The metal/substrate combination must retain high reactivity rates. In order to reduce the total amount of solids circulating between the reactors, it is important to optimize the loading of the metal on the substrate, thereby maximizing the oxygen carrying capacity of the solids. 9 1.4 Selection of Material for Oxygen Carriers A few materials have been identified as having CLOU capabilities, including: CuO, Mn2O3, Co3O2 and the bimetallic compound (Mn0.8Fe0.2)2O3 [22, 25]. These metal oxides all undergo a change in the equilibrium partial pressure of oxygen between the oxidized and reduced states at elevated temperatures. These elevated temperatures are in the range of a typical chemical-looping system, which, for long chain hydrocarbons such as solid fuels or oils, is above 700 °C according to Fan [26]. Fan also gives an upper limit on the temperature by referencing the Boudouard reaction given in equation (4). The Boudouard reaction is favored at higher temperatures and will decrease the degree of fuel conversion if not fully considered. Also, when deciding on a temperature ceiling for chemical-looping, it is important to consider the melting point and softening point temperatures of the materials being considered. 𝐶 + 𝐶𝑂2 → 2𝐶𝐶 (4) The melting point of copper is relatively low compared to other CLOU carriers such as manganese and cobalt. The melting points of these metals along with some of the associated oxides are given in Table 2. Although the melting point of copper is considerably lower than the melting points of the other base metals, a comparison of the melting points of the relevant oxides associated with each base metal reveals the copperbased oxides have the highest melting points. This consideration of these melting points is critically important for the determination of a suitable carrier in a chemical-looping system for both traditional CLC and CLOU technologies - especially when applying the generally accepted concept of an interconnected dual fluidized-bed system for chemical- 10 Table 2: The melting points of different CLOU capable species. Species Cu Melting Point (°C) 1,084 Cu2O 1,232 CuO 1,201 Mn 1,246 Mn2O3 940 Mn3O4 1,567 Co Co3O4 CoO 1,495 895 1,933 looping given by Lyngfelt et al. in 2001 [27]. Although the melting points of these species are known, a more difficult measure to quantify is the softening point of these species at which point these species begin to agglomerate and eventually sinter. Agglomeration and sintering of fluidized-bed materials can prove catastrophic for operation and must be avoided. Along with other benefits, including additional internal surface area and improved structural durability, supporting substrates have been and are continually investigated to mitigate the risk of agglomeration within a fluidized-bed by raising the agglomeration temperature. Typical supports either being investigated or having been reported on include: TiO2, ZrO2, Al2O3, Ilmenite, Sand, MgAl2O4, ZrO2/MgO and others. These supports are reported as having varying degrees of success in chemical-looping environments, and are continually investigated as to their effectiveness. Mattisson compiled a list of literature sources reporting investigations of CLOU using many different carriers at various temperatures [12]. 11 Selection of materials for use in a CLOU system may be performed based upon reaction rates, carrying capacity, relevant reaction temperature regime, cost of the material (both for the initial charge and replacement of fragmented or deactivated material) and auto-thermal capabilities. The oxidation reaction of the relevant species in CLOU is always exothermic while the reverse reaction, decomposition, is always endothermic [28]. The endothermic nature of the decomposition reaction may present a challenge from the necessary additional heat needed to drive the reaction. For the reactions with lower heats of reaction, given in Table 3, the heat release from fuel combustion may overcome the endothermic nature of the oxygen carrier decompositions giving overall exothermic reactions in both the AR and FR. Of the four listed below, only the cobalt-based material has a heat of reaction too large to be overcome by the heat release from fuel combustion. Cobalt, therefore, creates challenges that almost certainly remove it from the list of suitable oxygen carriers for a chemical-looping system [22]. Further selection of a proper base for the oxygen carrier requires a closer thermodynamic analysis. Cao and Pan (2006) investigated the selection of both the proper carrier and reactor for a chemical-looping system. Their investigation revealed that manganese oxide-based carriers were not suitable for a chemical-looping reactor for a couple of different reasons. The lower oxidation state pair (Mn3O4/MnO) did not favor full conversion of the fuel to CO2. They also mentioned that for the higher oxidation state pair (Mn3O4/Mn2O3), the oxygen carrying capacity is too low to be economical [29]. Additionally, the equilibrium partial pressure versus temperature plot, reference Figure 4, identifies the higher temperature capabilities of copper versus manganese and cobalt. Copper-based oxygen carriers have been determined as having a high likelihood 12 Table 3: List of CLOU reactions and the resultant heats of reaction. Species Reaction Heat of Reaction 4𝐶𝐶𝐶 ↔ 2𝐶𝑢2 𝑂 + 𝑂2 ∆𝐻850 = 263.2 2𝐶𝐶3 𝑂4 ↔ 6𝐶𝐶𝐶 + 𝑂2 ∆𝐻850 = 408.2 6𝑀𝑀2 𝑂3 ↔ 4𝑀𝑀3 𝑂4 + 𝑂2 (𝑀𝑀0.8 𝐹𝐹0.2 )2 ↔ (𝑀𝑀0.8 𝐹𝐹0.2 )3 𝑂4 + 𝑂2 ∆𝐻850 = 193.9 ∆𝐻850 = 254 𝑘𝑘 𝑚𝑚𝑚 𝑂 2 𝑘𝑘 𝑚𝑚𝑚 𝑂 2 𝑘𝑘 𝑚𝑚𝑚 𝑂 2 𝑘𝑘 𝑚𝑚𝑚 𝑂 2 (5) (6) (7) (8) of success for a CLOU-operated system. Although there are more expensive alternatives to copper, there are also more economical ones. Copper is an expensive substance which happens to have a relatively low melting point. Therefore, it is critically important to create an oxygen carrier material with a suitable support and production method. These two factors will determine whether a suitable copper-based carrier may or may not be found. One challenge that exists for the production of a suitable carrier is to homogeneously disperse the copper (oxide) throughout the particle. This will maximize structural support from the substrate, decrease the risk of agglomeration by minimizing copper-copper contact between particles and increase internal surface area, thereby diffusing possible internal mass transfer limitations to redox rates. 1.5 Objectives of This Work Given the myriad of information available in the field of chemical-looping, it is essential to the further development of this technology that the fundamental kinetic 13 1 Partial Pressure O2 (bar) Co3O4/CoO 0.1 CuO/Cu2O Mn3O4/Mn2O3 0.01 700 800 900 1000 Temperature(°C) 1100 Figure 4: Equilibrium partial pressure of oxygen for three different metal oxide pairs. modeling of the suggested oxygen carrier materials be understood. In the case of copper oxide, the kinetics of decomposition and, more particularly, oxidation are not adequately understood. For the industrial and commercial implementation of this technology and these materials, the fundamental kinetics must be understood for proper modeling. Additionally, the development of novel carriers and production methods is important to the industrialization of this technology. The development of materials that are optimally suited for solid fuels combustion in a fluidized-bed environment requires further investigation. 14 Although gaining popularity among researchers, the investigation of directly injected solid fuels combusted using copper oxide-based oxygen carriers still needs additional attention. The effects fuel type, fuel particle size and temperature must be understood for successful design and operation of commercial/industrial-scale CLOU reactors. Hence, it is the purpose of this work to: 1. Improve understanding of and develop kinetic models for the oxidation and decomposition of copper oxide-based oxygen carriers, 2. Evaluate performance of promising novel carriers, and 3. Investigate conversion of solid fuels with regard to fuel type, particle size and temperature. 15 1.6 References [1] U.S. NOAA. The NOAA Annual Greenhouse Gas Index (AGGI). Boulder, CO. Web 18 June 2013. [2] United States Environmental Protection Agency. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990 - 2011. EPA 430-R-13-001. Washington D.C. Web 13 June 2013. [3] United States Department of Energy. DOE/NETL Carbon Dioxide Capture and Storage RD&D Roadmap. Web 17 June 2013. [4] Ekström, C., Schwendig, F., Biede, O., Franco, F., Haupt, G., De Koeijer, G., Papapavlou, C., Røkke, P.E., Techno-Economic Evaluations and Benchmarking of Precombustion CO2 Capture and Oxy-fuel Processes Developed in the European ENCAP Project. Energy Procedia 2009; 1 (1): 4233-4240. [5] Marion, J., Mohn, N., Liljedahl, G.N., Nsakala, N., Morin, J-X., Henriksen, P.-P., Technology options for controlling CO2 emissions from fossil-fueled power plants, Third Annual Conference on Carbon Capture and Sequestration, 3-6 May 2004, Alexandria, VA; 2004. [6] Ishida, M., Zheng, D., Akehata, T. Evaluation of a chemical-looping combustion power-generation system by graphic exergy analysis. Energy. 1987, 12 (2): 147154. [7] Ishida, M., Jin, H. A new advanced power-generation system using chemicallooping combustion. Energy. 1994, 19 (4): 415-422. [8] Ishida, M., Jin, H. A novel chemical-looping combustor without NOx formation. Ind Eng Chem Res. 1996, 35 (7): 2469-2472. [9] Ishida, M., Jin, H. CO2 recovery in a power plant with chmical looping combustion. Energ Conv Mange. 1997, 38 (1): 2469-2472. [10] Ishida, M., Jin, H., Okamoto, T. A fundamental study of a new kind of medium material for chemical-looping combustion. Energ Fuel. 1996, 10 (4): 958-963. [11] Ishida, M., Jin, H., Okamoto, T. Kinetic behavior of solid particle in chemicallooping combustion: suppressing carbon deposition in reduction. Energ Fuel. 1998, 12 (2): 223-229 [12] Mattisson, T. Materials for chemical-looping with oxygen uncoupling. ISRN Chem. Eng. 2013. http://dx.doi.org/10.1155/2013/526375. [13] Johansson, E., Lyngfelt, A., Mattisson, T., Johnsson, F. Gas leakage measurements in a cold model of an interconnected fluidized-bed for chemical-looping combustion. Powd. Tech. 2003, 134 (3): 210-217. 16 [14] Mattisson, T., Lyngfelt, A., Cho, P. The use of iron oxide as an oxygen carrier in chemical-looping combustion. Energ Fuel. 2003, 17 (3): 643-651. [15] Garcia-Labiano, F., de Diego, L., Adanez, J., Abad, A., Gayan, P. Temperature variations in the oxygen carrier particles during their reduction and oxidation in a chemical-looping combustion system. Chem Eng Sci. 2005, 60 (3): 851-862. [16] Kronberger, B., Johansson, E., Loffer, G., Mattisson, T., Lyngfelt, A., Hofbauer, H. A two-compartment fluidized-bed reactor for CO2 capture by chemical-looping combustion. Chem Eng Tech. 2004, 27 (12): 1318-1326. [17] Kronberger, B., Loffler, G., Hofbauer, H. Simulation of mass and energy balances of a chemical-looping system. Int J Energ Clean Env. 2005, 6 (1): 1-14. [18] Kronberger, B., Lyngfelt, A., Loffler, G., Hofbauer, H. Design and fluid dynamic analysis of a bench-scale combustion system with CO2 separatoin-chemicallooping combustion. Ind Eng Chem Res. 2005, 44: 546-556. [19] Ryu, H., Bai, D., Jin, G. Carbon deposition characteristics of NiO based oxygen carrier particles for chemical-looping combustor. Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, J. Gale and Y. Kaya, Eds., pp. 175-180, Elsevier, Oxford, UK, 2002. [20] Ishida, M., Jin, H. A Novel Combustor based on chemical-looping reactions and its reaction kinetics. J Chem Eng. 1994, 3: 296-301. [21] Lyon, R. K., Cole, J. A. Unmixed combustion: an alternative to fire. Combust Flame. 2000, 121 (1-2): 249-261. [22] Mattisson, T., Lyngfelt, A., Leion, H. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int J Greenh Gas Con. 2009, 3 (1): 11-19. [23] Lewis, W.K., Gilliland, E.R., Sweeney, M.P. Gasification of carbon: metal oxides in a fluidized powder bed. Chem Eng Pro. 1951, 47: 251-256. [24] Eyring, E., Konya, G., Lighty, J., Sahir, A., Sarofim, A., Whitty, K. Chemical looping with copper oxide as carrier and coal as fuel. Oil Gas Sci Tech. 2011, 66: 209-221 [25] Azimi, G., Rydén, M., Leion, H., Mattisson, T., Lyngfelt, A. (MnzFe1-z)yOx combined oxides as oxygen carrier for Chemical-Looping with Oxygen Uncoupling (CLOU). AIChE J. 2013, 59: 582-588. [26] Fan, L. Chemical-looping systems for fossil energy conversion. John Wiley and Sons, Hoboken, NJ, USA, 2010. [27] Lyngfelt, A., Leckner, B., Mattisson, T. A fluidized-bed combustion process with inherent CO2 separation; application of chemical-looping combustion. Chem Eng 17 Sci. 2001, 56 (10): 3101-3113. [28] Jerndal, E., Mattisson, T., Lyngfelt, A. Thermal analysis of chemical-looping combustion. Chem Eng Res Des. 2006, 9: 795-806. [29] Cao, Y., Pan, W.P. Investigation of chemical looping combustion by solid fuels. 1. process analysis. Energ. Fuel. 2006, 20: 1836-1844. 18 CHAPTER 2 REVIEW OF LITERATURE The term chemical-looping was first introduced into the literature in 1987 by Ishida et al. [1], but Lewis and Gilliland studied a copper-based chemical-looping system in the 1940s and 50s [1, 2]. About 30 years later, Richter and Knoche [3] proposed that combustion efficiency be improved through a CLC process in 1983. By 2008, there were nearly 100 published articles reporting CLC testing with different materials as potential oxygen carriers [4]. Among those tested up to that point, the most common materials tested were NiO and Fe2O3, although there were also a few studies involving other materials such as CuO, CoO and MnO. Only six of those articles reported using solid fuels as the reductant (i.e., coal, petroleum coke and lignite char). The remaining sources all reported results of CLC testing using gaseous fuels. It should be noted that a significant number of those reports discussed the use of syngas as a reductant, which may also be included as solid fuels testing due to syngas being the main product of gasification reactions [4]. 2.1 CLC Traditional CLC carriers refer to those unable to spontaneously uncouple from scavenged oxygen such as Ni-, Fe- and Ca-based materials. Fang et al. reported that 19 nickel-oxide-based carriers were believed to be the most promising oxygen carrier candidates [5]. Jin and Ishida [6] established that unsupported NiO lost reactivity over extended cycling times and attributed it to the agglomeration of NiO particles. Due to the low melting points of most transition metals, especially those most suitable for CLC, agglomeration becomes a significant design challenge and is generally circumvented by the use of an inert support material. These inert supports range from those commonly used in other applications such as Al2O3, bentonite, SiO2 and TiO2 to some less common supports such as ZrO2/MgO, MgAl2O4,YSZ and NiAl2O4. Judging from reactivity studies of different oxygen carriers with natural gas, NiO and CuO show the most intriguing rates [7-10]. The disadvantages of Ni include its toxicity and thermodynamic limitations in conversion of fuels. However, Ni remains attractive due to its higher melting point than Cu [11]. Mattisson et al. presented results of using NiO/NiAl2O4 in TGA, lab-scale fluidized-bed (FB) and a 10 kW continuous CLC reactor at temperatures from 750 °C to 900 °C. They concluded that the FB displayed faster reaction rates than those observed in TGA which translated to a solids inventory of 9-24 kg/MW assuming a strictly plug-flow model (for the gas) in the FB. This assumption proved insufficient for estimating scale-up parameters for the 10 kW system where an incomplete gas yield was observed [11]. Fe-based carriers are very attractive due to their low cost, availability and high strength, according to Lyngfelt et al. [4]. Fan et al. have done extensive work on Febased systems, including H2 production, coal and biomass studies [26,12]. Among the Fe-based carriers, iron-ores have received some attention due to the very low cost. An attractive iron-ore, ilmenite (generally accepted as FeTiO2), has been described as an 20 attractive material as an oxygen carrier [13, 14]. Leion et al. (2008) tested the reactivity and effectiveness of ilmenite as an oxygen carrier in a lab-scale fluidized-bed [14]. They concluded that ilmenite showed no decrease in reactivity when exposed to alternating oxidizing/reducing cycles over 37 cycles. They stated also that the reduction of ilmenite is endothermic, but less so than both NiO/Ni and Fe2O3/Fe3O4 on a per mole basis. Adanez et al. reported a study of manganese oxide on 5 different support materials in 2004 [15]. They reported testing on Al2O3, sepiolite, SiO2, TiO2 and ZrO2. Of these supports, ZrO2 was found to be the only material suitable for CLC based on mechanical strength and reactivity. Further exploration by Mattisson et al. [16] and Cho et al. [8] suggested that Mn supported by Al2O3 was unsuitable as an oxygen carrier due to particles sintering. Johansson et al. then investigated ZrO2-supported manganesebased carriers [17]. In all, four different carriers were prepared and tested which include: pure ZrO2, and then ZrO2 stabilized by CaO, MgO and CeO2. They reported that the carrier supported by ZrO2 stabilized by MgO displayed the highest reactivity. Although the advantages and disadvantages of oxygen carriers developed using single-metal oxides have been established repeatedly, some institutions thought it wise to develop an oxygen carrier that takes the advantages of different single-metal oxygen carriers and combine the metals in an effort to utilize the desirable characteristics of each while overcoming the undesirable characteristics [5]. Jin et al. suggested that the hybrid Ni-Co/YSZ carrier outperformed the single-metal counterparts (Ni/YSZ and Co/YSZ) [18]. While nickel normally performs well as a carbon deposition catalyst (which is undesirable in CLC), Adanez et al. [19] reported that when used in a mixed Ni-Cu-based oxygen carrier, 100% conversion of CH4 to CO2 and H2O was observed with no carbon 21 deposition. Additionally, due to the higher melting point of nickel, the low melting point agglomeration effects of copper were not observed. Johansson et al. discovered that a combination of 3% nickel oxides in 97% iron oxides nearly doubled the CO2 output per unit time when compared to the sum of the CO2 output by the two individual metals separately [9]. Adánez et al. recently presented a review of oxygen carriers proposed for CLC using gaseous fuels. They reviewed carriers based on copper, nickel, manganese, iron and cobalt. Through thermodynamic analysis, the report shows the 6 forms of copperbased oxygen carriers as among the most favorable toward full conversion of the fuel to CO2 and H2O [20]. 2.2 CLOU Lewis, Gilliland and Sweeney first reported the CLOU capabilities of copper in 1951 [2]. They were using the Cu/CuO loop as a means to generate CO2. Their findings showed that the rate of decomposition of CuO to Cu2O is many times faster than the rates of coal gasification. By utilizing the spontaneous reduction tendencies of CuO, they determined it was possible to bypass the gasification step altogether and directly combust the fuel using the oxygen given off during CLOU. Mattisson et al. demonstrated the CLOU process using both CuO/Cu and Mn2O3/Mn3O4 as oxygen carriers. They demonstrated that the reaction rate of petroleum coke could be increased by a factor of 50 when using a CLOU process compared to an iron-based carrier under the same conditions, but operating under a CLC process [21]. Copper oxide has exhibited the fastest reaction rates of the possible metal oxides 22 for CLC [22]. In addition to fast reaction rates, the typically overall endothermic nature of the FR in a CLC system is exothermic for CuO. Copper oxide displays the following positive characteristics: (1) lower solids inventory and circulation due to high reactivity and high oxygen carrying capacity [23]; (2) Cu/CuO exhibit an overall exothermic reaction system in both the AR and FR, eliminating the need to heat the FR [24]; (3) thermodynamically speaking, CuO is favored to completely convert gaseous hydrocarbons to CO2 and H2O [22]. Early testing indicated the need for an inert support because copper has a very low melting point and agglomerates readily when used in a CLC reactor. The reaction rate of CuO decreased very quickly with increasing cycle number according to de Diego et al. [25]. They also reported upon further investigation that any mass loading of CuO on and inert support material below 10 wt% would never agglomerate while anything over 20 wt% would always agglomerate. Furthermore, they recommended that impregnation methods may be the only effective method for carrier preparation. However, Chuang et al. in 2008 rejected the preparation methods of mechanical mixing as well as wetimpregnation and displayed positive results for carriers prepared using co-precipitated techniques [26]. These carriers did not show any signs of agglomeration and maintained high reactivity over the 18 test cycles. 2.3 Kinetics of CLOU Although previous studies have addressed many issues regarding the performance of a multitude of potential carriers, there is still much to be understood. One important gap in the understanding is the fundamental kinetics of the oxygen carriers. Although 23 Adánez-Rubio et al. began to discuss the specific kinetic parameters of copper-based oxygen carriers in 2012 [20], the level of scientific work is, as of yet, incomplete. These parameters include reaction rate as a function of: temperature, pressure, oxygen loading, support material, particle size, oxygen partial pressure and coal type. The oxidation kinetics of copper oxide, particularly, has yet to be adequately described within the literature. As temperature is increased, the reaction rates for oxidation begin to slow, thereby making the Arrhenius plot reverse direction and thereby generate a negative apparent activation energy. This is mainly due to a decrease in oxidation rate at high temperature which has been reported as a reversal of activation energy at elevated temperatures. Adánez-Rubio et al. ascribe the reversing activation energy strictly due to the decrease in the oxygen partial-pressure driving force as a function of temperature [20]. Wilkins and Rideal [27] suggested that the decreasing oxidation rate with increasing temperature was governed by the rate of diffusion of oxygen through the film of oxide. Although conflicting ideas are proposed here, it is possible that more than 1 of these mechanisms play a role in this phenomenon. What is certain is that for larger-scale copper oxide-based CLOU plants, a better understanding of these kinetics are necessary for design and scale-up. Mattisson provided a generalized reaction rate as follows: 𝑝𝑂 ,𝑒𝑒 − 𝑝𝑂2 𝑟𝑑 = 𝑘𝑟 2 𝑝𝑂2 ,𝑒𝑒 (9) where 𝑝𝑂2,𝑒𝑒 is the equilibrium partial pressure of oxygen, an exponential function of temperature and kr is a reaction rate constant. The rate of decomposition is challenging to 24 determine. Adánez-Rubio et al. reported an apparent decrese in the decomposition rate with an increase in the oxygen partial pressure by supplying different concentrations of oxygen during decomposition [29]. Because the rate of decomposition is a function of the local oxygen partial pressure, inversely related, it is necessary to either estimate the partial pressure at the particle surface or to design experiments to deplete the local partial pressure of oxygen enough to remove any influence generated. This may be accomplished in a few ways, all of which have their own challenges. In order to deplete the oxygen partial pressure at the surface of the particle, a sweeping diluent may be added in sufficient quantity as to remove the oxygen molecules quickly enough to mitigate any influence created. Alternatively, a fuel may be added in order to immediately react with the liberated oxygen. This second approach presents the challenge of fuel selection. Most fuels with directly reduce the oxygen carrier, which commonly has faster reaction rates than the rates of CuO decomposition. This direct reduction likely occurs with most gaseous reactants. Likewise, due to the thermal breakdown of solid fuels (devolatilization), the liberated volatile compounds will also directly reduce the oxygen carrier particles. For this method to be successfully performed, it is necessary to use a char produced at a temperature above the targeted operating temperature and at high heating rates in order to prevent the release of any volatile species after introduction to the fuel reactor. In an effort to decrease the local partial pressure of oxygen without directly reducing the oxygen carrier, Arjmand et al. [28] introduced an excess of devolatilized wood char to a bed of freeze-granulated CuO/MgAl2O4. The goal of this approach was to have the released oxygen consumed by the char as close to the instant of release as 25 possible. In order for this to be successful, the char must be completely devolatilized or the released volatiles may react directly with the oxygen carrier, thereby reducing the oxygen carrier in a path other than spontaneous decomposition and skewing the decomposition rate data. The tests were performed in the range of 850-900 °C and the decomposition reaction was modeled using the Avrami-Erofeev mechanism. The resulting Arrhenius expression produced an apparent activation energy of 139.3 kJ/mole. In a similar manner, Adańez-Rubio et al. [29] obtained a decomposition rate of 2.3×10-3 kgO2/kg carrier. Sahir et al. reported an activation energy of 281 kJ/mole [30], which was somewhat smaller to the work produced by Chadda et al. [31] where an activation energy of 313 kJ/mole was determined. The work by Eyring et al. [32] produced an activation energy very similar to Chadda et al. of 327 kJ/mole. Although only a small range of models and methods exist for the description of the kinetics of decomposition of CuO, there are relatively fewer descriptions for the oxidation of Cu2O to CuO. This particular reaction is challenging to adequately describe while developing a simple enough model to be used in large-scale modeling projects. It has been reported in several sources that the rate of oxidation reaches a maximum at elevated temperatures and then begins to decrease with increasing temperature [26, 32 -34]. This is due at least in part to a decrease in the difference between the equilibrium O2 partial pressure and air's partial pressure of 0.21 atm, at sea level. In addition to the change in the driving force of oxidation, the molar volume of CuO is 5% larger than that of Cu2O; therefore, during oxidation, the solid must "grow" in order to accommodate the added molar volume [35, 36]. This "growth" of the solid particle may cause collapsing of the smaller pores within the particle [37]. The 26 collapsing of these smaller pores may "block" access to the remaining unreacted solid. This effect may be recognized by a distinct change in the reaction rate, creating two different regions of the conversion vs. time curve: the first region shows high reaction rates, whereas the second shows a distinct change in the rate at a considerably slower value [38, 39, 40]. The mechanism used to describe this phenomenon is labeled "poreblocking kinetics" and uses a logarithmic rate law which includes an empirically determined pore-blocking constant. This pore-blocking constant may be a function of temperature [39, 40] as well as number of redox cycles. Therefore, if the pore-blocking constant is a function of temperature, then the full kinetic expression includes three terms, all dependent on temperature (activation energy, 𝑝𝑂2,𝑒𝑒 and the pore-blocking constant). This triple dependency on temperature makes it difficult to ascertain the truly inherent activation energy of this reaction, and has thus given rise to the very limited amount of literature on the subject. The most applicable source on this matter was produced by Zhu et al. [36], where they describe a logarithmic rate law to which they were able to adequately fit their data. They report two activation energies, one at low temperatures and one at the higher temperatures where the oxidation rates begin to slow. This work is not directly applicable due to the experimental differences. For chemical-looping, the generally accepted design is an interconnected fluidized-bed of porous particles. The experimental approach of Zhu et al. utilized the oxidation/decomposition of pure copper rods. The effects seen and described by Zhu et al. are almost certainly influenced by mass-transfer resistances due to the sample size and nonporous nature and are, therefore, inadequate for the description of CuO-based particles in a chemical-looping process. 27 2.4 Combustion of Solid Fuels The first demonstration of the CLOU concept by Mattisson et al. was performed in a batch fluidized-bed reactor by the oxidation of 0.1 g of petroleum coke at 985 °C using two different cupric oxide-based materials: CuO/Al2O3 and CuO/ZrO2. Their results showed that the rate of CuO decomposition was faster than the oxidation rate of the petcoke observed by a nonzero value of the O2 volume fraction during fuel combustion [22]. The individual reaction regimes for this test (1) OC oxidation, (2) CLOU decomposition, (3) fuel oxidation and (4) OC regeneration are all clearly identified by the changes in temperature. It is clear that during the reduction cycle, even though the decomposition reaction is endothermic, the combustion of the fuel is exothermic enough to overcome the temperature drop. Burnout of the petcoke is completed within 30 seconds, and the rate of conversion observed within the reaction is roughly a 45 fold increase to that observed using an iron-based carrier [12]. The conversion of 0.1 g of petcoke during the decomposition of a CuO-based oxygen carrier with a zirconia support may be seen in Figure 5. Three obvious regimes may be seen. The figure shows the end of the oxidation cycle with the oxygen volume fraction around 0.21. The air is turned off and the nitrogen purge is cycled on, leading to the oxygen fraction decrease seen around 30 seconds. The remaining excess oxygen is purged from the reactor before the fuel is introduced into the reactor. The petcoke is then combusted as evidenced by the spikes of the carbonaceous species. Upon fuel burnout, the oxygen volume fraction once again increases toward the equilibrium partial pressure. When the fuel is first introduced, a small methane peak may be seen, which represents devolatilization of the fuel. 28 Figure 5: Concentration profile during the conversion of 0.1 g of petroleum coke with 15 g CuO/ZrO2. The temperature was measured to 985 °C in the bed of material at the start of the experiment. The fluidizing gas is pure nitrogen. From Mattisson et al., Chemicallooping with oxygen uncoupling for combustion of solid fuels, Int J Green H Gas Con, 2009, 3: 11-19. Leion et al. performed a set of CLOU experiments in the range of 850 - 985 °C using 6 different solid fuels while testing the performance of a CuO/ZrO2 material. This set of experiments helped to solidify the advantage CLOU has in solid fuel conversion over traditional CLC. This advantage was especially realized in the low volatile fuels. Gayán et al. performed a set of experiments covering 27 different fuels and several different oxygen carriers [33]. Through a set of three different campaigns, the most promising materials tested were spray dried CuO/ZrO2 and CuO/MgAl2O4. They were able to obtain complete combustion using a solid inventory in the FR of 235 29 kg/MWth [33, 41, 42]. A large number of sources report the use of solid fuels for combustion in both CLC and CLOU. The fuels ranged from lignite [14, 42, 43], to petcoke [21, 22, 44-48], to biomass chars [44, 49, 50] to coals [29, 44, 46, 51]. 30 2.5 References [1] Lewis, W. K., & Gilliland, E. R. Patent No. 2,665,972. United States of America. 1954. [2] Lewis, W. K., Gilliland, E. R., & Sweeney, W. P. Gasification of carbon: metal oxides in a fluidized powder bed. Chem Eng Prog. 1951, 251-256. [3] Richter, H., & Knoche, K. Reversibility of combustion processes. ACS Symposium Series. 1983, 235: 415-422. [4] Lyngfelt, A., Johansson, M., & Mattisson, T. Chemical-looping combustion - status of development. 9th International Conference on Circulating Fluidized-beds. Hamburg. 2009. [5] Fang, H., Haibin, L., & Zengli, Z. Advancements in developmnet of chemicallooping combustion: a review. Int J Chem Eng. 2009, 16. [6] Jin, H., & Ishida, M. A new type of coal gas fueled chemical-looping combustion. Fuel. 2004, (83), 2411-2417. [7] Adanez, J., Adanez, J., de Diego, L. F., Garcia-Labiano, F., Gayan, P., Abad, A. Selection of oxygen carriers for chemical-looping combustion. Energ Fuel. 2004, 18 (2), 371-377. [8] Cho, P., Mattisson, T., & Lyngfelt, A. Comparison of iron-, nickel-, copper- and manganese-based oxygen carriers for chemical-looping combustion. Fuel. 2004, 83 (9), 1215-1225. [9] Johansson, M., Mattisson, T., & Lyngfelt, A. Comparison of oxygen carriers for chemical-looping combustion. Therm Sci. 2006 (10), 93-107. [10] Johansson, M., Mattisson, T., & Lyngfelt, A. Comparison of oxygen carriers for chemical-looping combustion of methane rich fuels. 19th International Conference on Fluidized-bed Combustion. Vienna. 2006. [11] Mattisson, T., Jerndal, E., Linderholm, C., & Lyngfelt, A. Reactivity of a spraydried NiO/NiAl2O4 oxygen carrier for chemical-looping combustion. Chem Eng Sci. 2011, 66 (20): 4636-4644. [12] Fan, L. S., & Li, F. Ind Eng Chem Res. 2010, (49): 10200-10211. [13] Bergurand, N., & Lyngfelt, A. Chemical-Looping Combustion of Petroleum Coke Using Ilmenite in a 10 kWth. Energ. Fuel. 2009: 5257-5268. [14] Leion, H., Lyngfelt, A., Johansson, M., Jerndal, E., & Mattisson, T. The use of ilmenite as an oxygen carrier in chemical-looping combustion. Chem Eng Res Des. 2008, (86): 1017-1026. 31 [15] Adanez, J., Adanez, J., de Diego, L. F., Garcia-Labiano, F., Gayan, P., Abad, A. Selection of Oxygen Carriers for Chemical-Looping Combustion. Energ Fuel. 2004, 18 (2): 371-377. [16] Mattisson, T., Jardnas, A., & Lyngfelt, A. Reactivity of some metal oxides supported on alumina with alternating methane and oxygen - application for chemical-looping combustion. Fuel. 2003, 17 (3): 1749-1757. [17] Johansson, M., Mattisson, T., Lyngfelt, A. Investigation of Mn3O4 with stabilized ZrO2 for chemical-looping combustion. Chem Eng Res Des. 2006, 84 (9): 807-818. [18] Jin, H., Okamoto, T., & Ishida, M. Development of a novelchemical-looping combustion: synthesis of a lopping material with a double metal oxide of CoO-NiO. Energ Fuel. 1998, 12 (6): 1272-1277. [19] Adanez, J., Garcia-Labiano, F., de Diego, L. F., Gayan, P., Celaya, J., & Abad, A. Nickel-copper oxygen carrier to reach zero CO and H2 emissions in chemicallooping combustion. Ind Eng Chem Res. 2006, 45 (8): 2617-2625. [20] Adanez, J., Abad, A., Garcia-Labiano, F., Gayan, P., & de Diego, L. F. Progress in Chemical-Looping Combustion and Reforming technologies. Prog Energ Comb. 2012, (38): 215-282. [21] Mattisson, T., Leion, H., & Lyngfelt, A. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int J Greenh Gas Cont. 2009, 3 (1): 11-19. [22] Garcia-Labiano, F., de Diego, L. F., Adanez, J., Abad, A., & Gayan, P. Reduction and oxidation kinetics of a copper-based oxygen carrier prepared by impregnation for chemical-looping combustion. Ind Eng Chem Res. 2004, 43 (26): 8168-8177. [23] Abad, A., Adanez, J., Garcia-Labiano, F., de Diego, L. F., Gayan, P., & Celaya, J. Mapping of the range of operational conditions for Cu-, Fe-, and Ni-based oxygen carriers in chemical-looping combustion. Chem Eng Sci. 2007, 62 (1-2), 533-549. [24] de Diego, L. F., Garcia-Labiano, F., Gayan, P., Celaya, J., Palacios, J. M., & Adanez, J. Operation of a 10 kWth chemical-looping combustor during 200 h with a CuO-Al2O3 oxygen carrier. Fuel. 2007, 86: 1036-1045. [25] de Diego, L. F., Garcia-Labiano, F., & Adanez, J. Development of cu-based oxygen carriers for chemical-looping combustion. Fuel. 2004, 83 (13): 643-651. [26] Chuang, S. Y., Dennis, J. S., Hayhurst, A. N., & Scott, S. A. Development and performance of Cu-based oxygen carriers for chemical-looping combustion. Comb Flame. 2009, 154 (1-2): 109-121. [27] Wilkins, F. J., & Rideal, E. K. The kinetics of the oxidation of copper. Part I. The initial Oxidation of Copper at Low Pressures. The Royal Society. 1980, 128: 394406. 32 [28] Arjmand, M., Keller, M., Leion, H., Mattisson, T., Lyngfelt, A. Oxygen release and oxidation rates of MnAl2O4-supported CuO oxygen carrier for chemical-looping with oxygen uncoupling (CLOU). Energ Fuel. 2012, 26: 6528-6539. [29] Adańez-Rubio, I., Abad, A., Gayan, P., de Diego, F., Garcia-Labiano, F., Adanez, J. Identification of operational regions in the chemical-looping with oxygen uncoupling (CLOU) process with a cu-based oxygen carrier. Fuel. 2012, 102: 634645. [30] Sahir, A., Sohn, H., Leion, H., Lighty, J. Rate analysis of chemical-looping with oxygen uncoupling (CLOU) for solid fuels. Energ Fuel. 2012, 26: 4395-4404. [31] Chadda, D., Ford, J., Fahim, M. Chemical energy storage by the reaction cycle of CuO/Cu2O. Int J Energ Res. 1989, 13 (1): 63-73. [32] Eyring, E., Konya, G., Lighty, J., Sahir, A., Sarofim, A., Whitty, K. Chemical looping with copper oxide as carrier and coal as fuel. Oil Gas Sci. Tech. 2011, 66: 209-221. [33] Gayan P, Adánez-Rubio I, Abad, A, de Diego LF, García-Labiano F, Adánez J. Development of Cu-based oxygen carriers for chemical-looping with oxygen uncoupling (CLOU) process. Fuel. 2012, 96:226-238. [34] Brandt B, Gruene P, Rosowski F, Farrauto RJ, Castaldi MJ. Oxygen uptake and release kinetics of Cu/Cu2O/CuO on Al2O3 and ZrO2-SiO2 as materials for chemical looping combustion, AIChE Annual Meeting. Minneapolis, MN, USA; 2011. [35] Prisedsky V, Vinogradov V. Fragmentation of diffusion zone in high-temperature oxidation of copper. J Solid State Chem 2004; 177: 4258-4268. [36] Zhu Y, Mimura K, Isshiki M. Oxidation mechanism of Cu2O to CuO at 600-1050 °C. Oxid Met. 2004, 62 (3): 207-222. [37] Evans, U.R. The Corrosion and Oxidation of Metals: Scientific Principles and Practical Applications. Edward Arnold Ltd.: London, 1960. [38] Sohn, H.Y., Byung, S.K. A novel cyclic reaction system involving CaS and CaSO2 for converting sulfur dioxide to elemental sulfur without generating secondary pollutants. 2. Kinetics of the reduction of sulfur dioxide by calcium sulfide powder. Ind Eng Chem Res. 2002, 41: 3087-3091 [39] A novel reaction system involving BaS and BaSOR for converting SO2 to elemental sulfur without generating pollutants: Part I. Feasibility and kinetics of SO2 reduction with BaS. Chem Eng Sci. 2006, 61: 5088-5093. [40] Won, S., Sohn, H.Y. Kinetics of the Reaction between Hydrogen Sulfide and Lime Particles. Metall Trans B. 1985, 16 (B): 163-168. 33 [41] Abad, A., Adánez-Rubio, I., Gayán, P., García-Labiano, F., de Diego, L. F., Adánez, J. Demonstration of chemical-looping with oxygen uncoupling (CLOU) process in a 1.5 kWth continuously operating units using a Cu-baed oxygen-carrier. Int J Greenh Gas Con. 2012, 6: 189-200. [42] Adanez-Rubio, I., Gayan, P., Abad, A., Garcia-Labiano, F., de Diego, L., Adanez, J. CO2 capture in coal combustion by chemical-looping with oxygen uncoupling (CLOU) with a cu-based oxygen carrier. Proceedings of the 5th International Conference on Clean Coal Technologies (CCT '11), Zaragoza, Spain, 2011. [43] Adanez-Rubio, I., Abad, A., Gayan, P., de Diego, L., Garcia-Labiano, F., Adanez, J. Performance of CLOU process in the combustion of different types of coal with CO2 capture. Int J Greenh Gas Con. 2013, 12: 430-440. [44] Leion, H., Mattisson, T., Lyngfelt, A. Using chemical-looping with oxygen uncoupling (CLOU) for combustion of six different solid fuels. Energ Proc. 2009, 1: 447-453. [45] Leion, H., Larring, Y., Bakken, E., Bredesen, R., Mattisson, T., Lyngfelt, A. Use of CaMn0.875Ti0.125O3 as oxygen carrier in chemical-looping with oxygen uncoupling. Energ. Fuel. 2009, 23: 5276-5283. [46] Azimi, G., Leion, M., Ryden, M., Mattisson, T., Lyngfelt, A. Investigation of different Mn-Fe oxides as oxygen carrier for chemical-looping with oxygen uncoupling (CLOU). Energ Fuel. 2013, 27: 367-377. [47] Arjmand, M., Leion, H., Lyngfelt, A., Mattisson, T. Use of manganese ore in chemical-looping combustion (CLC) - Effect on steam gasification. Int J Greenh Gas Con. 2012, 8: 56-60. [48] Arjmand, M., Leion, H., Mattisson, M., Lyngfelt, A. Evaluation of different manganese ores as oxygen carrier in chemical-looping combustion (CLC) for solid fuels. In Proceedings of the 2nd International Conference on Chemical Looping, Darmstadt, Germany, 2012. [49] Arjmand, M., Hedayati, A., Azad, A. M. Leion, H., Ryden, M., Mattisson, T. CaxLa1-yMyO3-σ (M = Fe, Ti, Mg, Cu) as oxygen carriers for chemical-looping with oxygen uncoupling (CLOU). Energ Fuel. 2013. [50] Azimi, G., Ryden, M., Leion, H., Mattisson, T., Lyngfelt, A. (MnzFe1-z)yOx combined oxides as oxygen carrier for chemical-looping with oxygen uncoupling (CLOU). AIChE J. 2013, 59: 582-588. [51] Abad, A., Adanez-Rubio, I., Gayan, P., Garcia-Labiano, F., de Diego, L.F., Adanez, J. Demonstration of chemical-looping with oxygen uncoupling (CLOU) process in a 1.5 kWth continuously operating unit using a Cu-based oxygen-carrier. Int J Greenh Gas Con. 2012, 6: 189-200 34 CHAPTER 3 EXPERIMENTAL APPROACH AND DATA ANALYSIS 3.1 Oxygen Carrier Selection For the selection of oxygen carriers, there are several factors needing consideration: (1) oxygen carrying capacity, (2) reactivity, (3) chemical durability (deactivation resistance), (4) physical durability (attrition resistance) and (5) total cost (base materials, production and replacement). Each of these attributes plays a significant role in OC selection and the pros and cons of each must be weighed for a proper selection to be made. For example, if a material has very low attrition resistance, but has a very low cost associated with procurement, then, it may be less expensive, both in money and time, to use it as opposed to a high-cost/low-attrition manufactured material. Alternatively, if the reactivity and oxygen carrying capacity are low, then the required solids inventory within the reactor at any given time will be larger, thereby requiring a higher capital investment. An additional consideration to be understood is the suitability for combusting, or oxidizing, the particular fuel in question. For example, while iron oxides are relatively inexpensive and perform well in chemical-looping systems involving gaseous reactants, they are ultimately inferior to the CLOU-type materials of Cu or Mn in the combustion of solid fuels [1]. In their study of the use of copper oxide to produce CO2, Lewis, Gilliland 35 and Sweeney were able to show reaction rates roughly 50 times faster than that of gasification of solid fuel [2]. For solid fuel combustion, using an iron-based OC gasification is necessary to avoid the very slow solid-solid reaction rates between the OC particles and the solid fuel. Therefore, likely the first consideration to be made in the selection of an OC requires a decision to use a CLC material or a CLOU material. At the University of Utah's Institute for Clean and Secure Energy, the decision was made to investigate chemical-looping on the premise for combustion of solid fuels. Therefore, the CLOU materials were looked at very closely to understand the application of CLOU-based materials in the combustion of solid fuels in a chemical-looping system. As outlined within the previous chapter, of the three main CLOU capable materials (Cu, Mn, Co), cobalt showed the least amount of promise [3] and was therefore discarded immediately. The choice between Mn and Cu came down to an understanding of the region. The University of Utah is nestled up against the Wasatch Mountain Range and overlooks the Salt Lake Valley. Across the valley, on a clear day, a smoke stack is visible. This smoke stack belongs to a smelter plant of Kennecott's Bingham Canyon Mine - one of the largest open pit copper mines in the world. Along with the availability of copper, several studies had been performed where cu-based materials were reported as having very high potential as suitable OCs in a fluidized-bed CLOU system [4, 5, 6]. Therefore, several different copper-based materials were either acquired or produced and, subsequently, tested in an attempt to find a carrier suitable for large-scale testing. A list of these materials is provided in Table 4. The carrier name designations are assigned based upon, first, either CuO wt% loading or 36 Table 4: Copper-based materials tested at the University of Utah for suitability as oxygen carriers within a fluidized-bed based on CLOU. Source Preparation Method CuO Loading (wt%) Ilmen Mountains USA NA NA 3N Cu Atlantic Engineers, USA NA 99.9 12_Al2O3_IW Sigma Aldrich, USA Incipient Wetness 12 50_TiO2_MM ICPC, Poland Mechanical mixing 50 45_ZrO2_FG Chalmers, Sweden Freeze granulation 45 16_SiO2_IW University of Utah Incipient Wetness 16 64_SiO2_IW University of Utah Incipient Wetness 64 20_FeTiO2_IW University of Utah Incipient Wetness 20 Carrier Non Cu-Based FeTiO2 (Ilmenite) Cu-Based purity (3N is 99.9%), second, support material, and, lastly, production method. As a base case, a 3N (99.9%) pure copper powder was acquired from Atlantic Engineers and tested. All additional materials were subsequently tested to find a more suitable candidate. The substrates were selected based on a variety of reasons. There are a handful of supporting materials which appear repeatedly within the literature such as: ZrO2, Al2O3, TiO2, Sepiolite and SiO2 [7]. The materials selected for this study were among the list of the most common support used in the literature providing additional characteristics such 37 as durability, additional internal surface area and increased melting point. Several production methods were used during this work. This was mostly out of a desire to understand variability of materials based upon production method. The methods of production used and discussed in the following sections are: freeze granulation, mechanical mixing and wet impregnation (also referred to as incipient wetness). Freeze granulation is frequently used because of its ease of operation and ability to produce oxygen carrier particles with a favorable porosity and specific surface. In this method of production, chosen proportions of metal oxide, inert support and a dispersant are mixed with water and ball milled. A binder is added and then the slurry is pumped through a nozzle into liquid nitrogen. Moisture is removed from the spherical particles during the freeze drying process. The particles are then calcined at 950 and 1050 ⁰C. The mechanical mixing procedure uses a powdered mixture of metal oxide and inert combined in desired concentrations. A concentration of 10 wt. % graphite is included as a pore forming additive, increasing reactivity. Water is added to achieve a suitable viscosity and the resulting paste is extruded through a syringe. The material is allowed to soft dry at 80 ⁰C overnight and then cut to a desired length. The particles are then sintered at various temperatures ranging from 950 to 1300 ⁰C in a muffle oven. This method generally produces irregularly shaped particles compared to the sphericity of particles prepared by other methods. Spray drying method begins with metal oxides and inert materials in a solution or slurry. The solution is atomized or sprayed through a nozzle, dispersing the liquid in a controlled drop size. The dispersed solution comes into contact with a cool gas (usually air or nitrogen) where it is rapidly hardened. This method generally provides the most 38 consistent particle size distribution. Wet impregnation (or Incipient Wetness) is carried about by dropping an inert support particle into a saturated copper nitrate aqueous solution. The particles are then dried overnight and then calcined at 500 ⁰C in order to decompose the copper nitrate to a copper oxide. This method provides a preferred particle porosity and surface. For comparison, an iron-based material was selected to stand as a base material to be compared against with the copper-based materials. In the literature, the material ilmenite has received quite a bit of attention [8, 9]. Ilmenite is attractive due to its strong support material, TiO2, and because it is a naturally occurring iron-ore, meaning it is unprocessed and therefore relatively inexpensive. Ilmenite is the largest source of titanium in the world, accounting for roughly 90% of the world's consumption of titanium materials [10]. Being both a naturally occurring ore and abundant, this material is an inexpensive alternative for CLC oxygen carriers. 3.2 Physical Characterization Certain physical characteristics are important considerations for a suitable oxygen carrier material in a chemical-looping system. These characteristics include: 1. Particle size distribution 2. Surface area 3. Mechanical strength 4. Particle shape (Sphericity) Each of these characteristics was determined for each material tested at the University of Utah. 39 Particle size distribution may affect both the reaction rates and fluidizability of the selected material. If particles are too large, the material may exhibit mass transfer resistances during reaction. If the particle size distribution is too large, it is difficult to determine a proper gas velocity to maintain proper fluidization. According to Kunii and Levenspiel [11], the minimum fluidization velocity is a function of particle density, bed voidage and the sphericity of the particle along with the fluidizing fluid properties. Figure 6 gives a qualitative representation of fluidizability as a function of the solid material characteristics. The physical characteristics of the oxygen carrier materials tested were ascertained by a few different methods. To determine the particle size distribution of each of the powders, both a sieve method and a particle size analyzer were employed. The sieve trays were used to cut the powder to different size ranges. The ranges employed in this work were: 5. Dp < 45 μm - TGA Kinetic Studies 6. 45 μm < Dp < 75 μm - TGA Kinetic Studies 7. 75 μm < Dp < 105 μm - TGA/FzB (Fluidized-bed) Classification Studies 8. 105 μm < Dp 150 μm - FzB Performance and Attrition Studies 9. 150 μm < Dp 250 μm - FzB Attrition Studies Surface area, particle morphology and attrition studies were determined using BET, SEM and a fines collector system. The BET employed is a Micromeritics Tristar II. SEM/EDS micrographs were used to understand particle sphericity as well as visual understanding of surface area and porosity. Particle attritted fines were collected in a dual fine filter apparatus designed at the University of Utah. 40 Sphericity Bed Voidage 1/ρ Fluidizability Figure 6: Qualitative representation of fluidizability as a function of sphericity, bed voidage and density. The mechanical strength of the individual particles may serve as an indication of attrition resistance; therefore, the mechanical strength of the particles was determined using a Shimpo FGE-5X mounted on an aluminum frame for stability. The influences of gas velocity, particle size, production method, operating temperature and support material on the rate of particle attrition were determined using a fluidized-bed. The attritted materials were collected using two in-line parallel PTFE (Teflon) filters. These filters were isolated from each other using ball valves at both the inlet and outlet of the filters. By so doing, the attritted materials were allowed to collect on the surface of one filter while the isolated filter could be removed and weighed to determine the amount of collected material. These collected materials were then viewed using SEM/EDS micrographs to determine surface morphology and to identify the elemental makeup of the fines collected. 41 3.3 Chemical Testing The characteristics designated as "chemical" include oxygen carrying capacity, reactivity and deactivation resistance. These characteristics were determined using two different reactors - thermogravimetric analyzer (TGA) and a bubbling fluidized-bed reactor (FzB). Typical reaction conditions are given for each apparatus in Table 5. Each oxygen carrier material was analyzed by a TA Instruments Q500 thermogravimetric analyzer (TGA) to determine its reactivity without fuel in multicycle tests. The TGA experiments cannot predict oxygen carrier performance under CLOU conditions when fuel is present, but were used to observe complete oxidation and reduction reactions over multiple cycles. Reaction gases used during redox cycling ranged from pure nitrogen to 21% oxygen (Air) with the balance as N2. A schematic of the Q500 with EGA (Evolved Gas Analysis) furnace used in these tests is provided as Figure 7. The quartz liner of the EGA furnace decreases the internal volume to around 15mL. The sample is suspended from the balance into the center of the furnace chamber and reaction gases are fed to the sample perpendicular to the sample cup suspension. A small purge flow is provided to the balance chamber to reduce the temperatures observed there. This purge gas mixes with the exhaust gases and proceeds out of the reactor chamber. The reaction gas flow (80 - 100mL) is typically four to five times higher than the balance purge gas and therefore ensures that the reaction gas is not diluted prior to sample contact. To analyze the TGA data, the conversion of the oxygen carriers for each oxidation and reduction reaction was determined using the equations (9) & (10). Xox and Xred represent fractional conversion for the oxidation and reduction reactions, mt is the sample 42 Table 5: Characteristic ranges used for the TGA and fluidized-bed reactors. Sample Size (g) Particle Size (μm) Temperature (°C) Flow Rate (L/min) Oxidation Time (min) TGA Q500 0.010 - 0.050 < 45 - 105 650 - 1000 0.1 - 0.12 5-120 10-120 Fluidizedbed 10 - 50 75 - 250 650 - 1000 0.5 - 2 5 - 60 5 - 60 Decomp Time (min) mass at a given time, mred is the mass of fully reduced sample with all copper as Cu2O, and mox is the mass of fully oxidized sample with all copper as CuO. (10) (11) The rates of the reduction portions of the experiments were found by treating the conversion data with a linear model. The calculated regression slope of the line approximates the rate constant, dXred/dt, in units of percent-conversion/second. The data were trimmed from 1% to 90% to capture the linear portion of the data. Each rate constant was calculated by trimming the data so that initial reaction completion (Xi) was 5% and final reaction completion (Xf) was 100% to capture the portion of data that fit the pseudo first order model. The FzB system (Figure 8) is very similar to that developed at Chalmers University. The reactor is made of quartz and is housed within a Carbolite VST 12/600 43 Figure 7: Picture of TGA Q500 with reactor schematic (bottom). The Q500 shown gives the EGA furnace which has quartz lining the furnace walls. Images courtesy of TA Instruments 2009, http://files.instrument.com.cn/bbs/upfile/200959204245.pdf. 44 2.5 cm 3.75 cm 2.5 cm Particle Bed Quartz Frit or Distributor Figure 8: Schematic of quartz reactor used for fluidized-bed studies. The particle bed is held and the reactant gas is distributed by a sintered quartz frit. The particle bed is 2.5 cm in diameter. Clamshell furnace (shown in Figure 9) with a maximum operating temperature of 1200°C. The reactor has four zones: (1) inlet, (2) sintered quartz distributor supporting the particle bed, (3) freeboard expansion zone and (4) outlet. Gas flow is controlled by a series of mass flow controllers and is introduced to the reactor housed within the furnace. The exit gas is filtered to remove fines and water is condensed from the gas before it is sent to a 4-channel California Analytical Instruments ZRE NDIR/fuel call analyzer. The analyzer measures the concentration of CO, CO2, CH4 and O2 (see Figure 10). The ID of the reaction tube is 2.5 cm. In order to maintain a well-fluidized system, the typical resting bed height was about 2.5 cm. Additionally, tests were conducted at about 5 cm and 1 cm bed height. In order to reduce the heating and cooling effects of the oxidation and reduction reactions, some samples were diluted with 45 Figure 9: Chemical-looping fluidized-bed reactor system at the University of Utah. zirconium silicate beads with an approximate particle diameter of 100 µm. Reactivity data are determined by the signals received from the gas analyzer. The gas analyzer data do not directly indicate what is going on within the reactor. The data are convoluted due to the reaction gas residence time distribution, gas dispersion in the gas lines after the reactor and analyzer time delay. To account for this convolution of the actual data, a deconvolution procedure was developed. Several approaches for deconvolution of data are available. If the data set is a discrete set, then the set may be fit to a polynomial expression, which can then be subjected to a Laplace transform. The continuous Laplace transform method obeys the following relationship [11]: 46 Figure 10: Fluidized-bed reactor system employed at the University of Utah for the testing of CLC/CLOU materials. ℒ[𝐹(𝑡)] = ℒ[𝐴(𝑡)] × ℒ[𝐶(𝑡)] = 𝐹(𝑠) (12) where F denotes the collected data, A denotes the actual data and C represents the convolution of the data. The function C may be determined from residence time distribution tests, with data then fit to a polynomial in similar fashion to the transformation of F from a discrete array to a continuous function. The resulting transformed equations may be rearranged as follows: 𝐹(𝑠) = 𝐴(𝑠) 𝐶(𝑠) The inverse Laplace transform generates the final result of actual data as a function of time, or: (13) 47 𝐹(𝑠) 𝐴(𝑡) = ℒ −1 [𝐴(𝑠)] = ℒ −1 𝐶(𝑠) (14) This process has proven useful and may be utilized, but the accuracy depends on the accuracy of the polynomial fit. Another method used to deconvolve a data set is much simpler and more quickly employed. In this method, the measured signal is subtracted from a second signal obtained by looping reaction gases over a bed of inert material (90 micron ceramic beads). The observed difference between the signal with inert material and a perfect step change represents the residence time distribution (RTD signal), or degree of data convolution. This method is displayed in Figure 11, which shows the result obtained when the measured signal is subtracted from the RTD signal. The reasonableness of using this method was evaluated by comparing the results obtained in the fluidized-bed and deconvolved using this simple method against results obtained in a TGA using the same material. Figure 12 shows a comparison between these methods. These tests were conducted at 800°C using air as the oxidizer and ilmenite as the oxygen carrier. While the signal lines do not line up exactly on top of each other, the two results agree very well. Due to the simplicity and reasonably good accuracy associated with this method, data analysis was conducted in this manner. 48 Figure 11: Method of signal deconvolution used in this study. Figure 12: Comparison between deconvolved rates obtained within a fluidized-bed and TGA. The error appears small enough that the method of deconvolution is justified. 49 3.4 Solid Fuels Testing For the testing of solid fuels in a CLOU reactor, three fuels were chosen and were tested while using 3 different oxygen carriers. The fuels selected along the ultimate and proximate analyses are listed in Table 6. For further testing, the solid fuels listed were each used to produce a char resulting from the low heating rate release of volatile compounds at 900°C under nitrogen for 8 hours. Prior to testing, each of the fuels was heated to 110°C for 2 hours in an effort to remove accumulated moisture. An additional unit was added to the quartz reactor scheme in order to provide a means of solid fuel delivery to the oxygen carrier particle bed. This additional unit was made of quartz as well and included a 0.3175 cm delivery port (Figure 13). For testing with solid fuels, a brief purge cycle of N2 at 1 L/min was used to clear the excess oxygen from the reactor to a level below the equilibrium partial pressure of oxygen. At this time, the purge gas was shut off to reduce blow-out of the fuel while it was added from the top of the reactor (Figure 13). After the addition of the solid fuels, the purge N2 (1 SLPM) was once again switched on. The volume percentages of CO2, CO, CH4 and O2 were measured and recorded continuously during the combustion of the fuel. Figure 14 is a plot showing the evolution of these carbonaceous gases during the combustion of char made from Illinois#6 coal at 900°C. The oxygen carrier used to produce Figure 14 consisted of 30 wt% CuO loaded onto an ilmenite support. The oxygen trace quickly drops below 2 vol% before the fuel is added. Once the fuel is added, the oxygen volume percent drops to near zero, but is not completely consumed and the production of CO2 dominates the evolved gases. For the measured signals, the carbonaceous gases are very near real-time whereas 50 Table 6: Ultimate and proximate analyses of solid fuels used in this study. Black Thunder Illinois #6 (PRB) Bituminous Sub-bituminous Green Coke Petroleum Coke Moisture (wt% as received fuel) 2.54 21.30 0.4 Ash (wt% Dry) 12.33 6.46 0.39 Volatile matter (wt% dry) 39.40 54.26 11.03 Fixed carbon (wt% dry) 48.28 39.28 88.01 Carbon 78.91 74.73 89.21 Hydrogen 5.50 5.40 3.78 Nitrogen 1.38 1.00 1.73 Sulfur 4.00 0.51 5.82 Oxygen 10.09 18.27 4.41 Chlorine 0.11 0.08 - 12,233 12,815 15,622 Coal - Type Proximate Analysis Ultimate Analysis (wt% dry, ash-free) Heating Value HHV, dry (Btu/lb) 51 Solid Fuel Feeder 0.3175 cm 2.5 cm 3.75 cm 2.5 cm Particle Bed Quartz Frit or Distributor Figure 13: Quartz bubbling-fluidized-bed reactor showing additional solid fuel feeder. CO CH4 and CO Vol % 0.7 CH4 0.6 25 20 CO2 0.5 O2 0.4 15 10 0.3 0.2 CO2 and O2 Vol% 0.8 5 0.1 0 0 0 200 400 Time (Seconds) Figure 14: Evolution of gases during combustion of Illinois #6 Char at 900°C while using 52 in order to understand the true evolution of oxygen, the method of deconvolution previously discussed must be employed. For the combustion of solid fuels, it is requisite to perform a carbon balance analysis to evaluate conversion of the fuel and to understand effective combustion rates. The carbon balance is performed using the flow rate of the purge gas as a baseline. It is assumed that the gases recorded by the analyzer, along with the inert purge, combine to make the majority of the gases evolved from fuel combustion. By assuming that any other species evolved from the combustion are in quantities small enough to have a negligible effect on the total flow rate of the effluent gases, the molar flow rate of the analyzed species may be determined. The total volumetric flow rate may be determined by calculating the volume fraction of the inert purge gas (𝜙𝑝𝑝𝑝𝑝𝑝 ) assuming that is the difference between one and the sum the volume fractions (𝜙𝑖 ) of CO2, CO, CH4 and O2. 𝜙𝑝𝑝𝑝𝑝𝑝 = 1 − 𝜙𝑖 ; 𝑖 = 𝐶𝐶2 , 𝐶𝐶, 𝐶𝐶4 , 𝑂2 (15) The volume fraction of the purge gas (N2) is then used to calculate the total effluent volumetric flow rate of gases (𝑄𝑡𝑡𝑡 ). 𝑄𝑡𝑡𝑡 = 𝜙𝑁2 𝑄𝑁2 (16) The total volumetric flow rate multiplied by the individual specie volume fractions gives the volumetric flow rate of each of the individual gaseous species. The 53 ideal gas equation is then employed to determine the molar flow rate of each component. 𝑄𝑖 = 𝜙𝑖 × 𝑄𝑡𝑡𝑡 ; 𝑖 = 𝐶𝐶2 , 𝐶𝐶, 𝐶𝐶4 , 𝑂2 𝑛𝑖 = 𝑄𝑖 𝑃 𝑅𝑅 (17) (18) The carbon balance is then performed by adding the total number of moles of carbonaceous gases and then multiplying by the molecular weight of carbon - 12 g/mole. The resulting mass of carbon is compared to the mass of the sample introduced to the reactor, generating a conversion profile based on the conversion of carbon. 54 3.5 References [1] Eyring, E., Konya, G., Lighty, J., Sahir, A., Sarofim, A., Whitty, K. Chemical looping with copper oxide as carrier and coal as fuel. Oil Gas Sci. Tech. 2011, 66: 209-221. [2] Lewis, W.K., Gilliland, E.R., Sweeney, M.P. Gasification of carbon: metal oxides in a fluidized powder bed. Chem Eng Pro. 1951, 47: 251-256. [3] Mattisson, T., Lyngfelt, A., Leion, H. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int J Greenh Gas Con. 2009, 3 (1): 11-19. [4] Mattisson, T., Lyngfelt, A., Leion, H. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int J Greenh Gas Con. 2009, 3 (1): 11-19. [5] Lewis, W.K., Gilliland, E.R., Sweeney, M.P. Gasification of carbon: metal oxides in a fluidized powder bed. Chem Eng Pro. 1951, 47: 251-256. [6] Adanez, J., Adanez, J., de Diego, L. F., Garcia-Labiano, F., Gayan, P., Abad, A. Selection of oxygen carriers for chemical-looping combustion. Energ Fuel. 2004, 18 (2), 371-377. [7] Mattisson, T. Materials for Chemical-looping with oxygen uncoupling. ISRN Chem. Eng. 2013. http://dx.doi.org/10.1155/2013/526375. [8] Leion, H., Lyngfelt, A., Johasson, M., Jerndal, E., Mattisson, T. The use of ilmenite as an oxygen carrier in chemical-looping combustion. Chem Eng Res. Des. 2008, 86 (9): 1017-1026. [9] Berguerand, N., Lyngfelt, A. The use of petroleum coke as fuel in a 10 kWth chemical-looping combustor. Int J Greenh Gas Con. 2008, 2 (2): 169-179. [10] U.S. Geological Survey. Mineral Commodity Summaries. Web: Minerals.usgs.gov/minerals/pubs/commodity/titanium/mcs-2012-timin.pdf [11] Kunii, D., Levenspiel, O. Fluidization Engineering. Butterworth-Heinemann, Newton, MA, 1991. [12] Blair, D. W.; Wendt, J. O.; Bartok, W. Evolution of nitrogen and other species during controlled pyrolysis of coal. International Symposium on Combustion (pp. 475-489). Elsevier 1977. 55 CHAPTER 4 RESULTS 4.1 Preliminary and CLOU Operation of Carriers Several materials, listed in Table 7, were tested to determine their suitability as oxygen carriers. It was established early on during this work that copper was to be the material of choice for the reasons previously discussed. Because CuO is readily decomposed to Cu2O at a rate greater than the rate of gasification of solid fuels, it was seen as the obvious choice for this work. 4.1.1 Ilmenite While CuO was selected as the oxygen carrier of choice, it was necessary to test other materials in order to effectively compare the performance of the copper-based oxygen carriers. A commonly tested and inexpensive material was selected to serve as this comparison. Ilmenite is a naturally occurring iron-ore used most often in the manufacture of rutile and titanium metal. In its fully reduced state, ilmenite is comprised mostly of FeTiO3, but admittedly, because it is a naturally occurring iron-ore, the composition of ilmenite has some variability. At temperatures below 800°C, not all of the ilmenite is oxidized obeying the following reaction: 56 Table 7: Materials tested for suitability as oxygen carriers in chemical-looping. Support Production Method Acquired from Ilmenite FeTiO3 - - - Atlantic Engineers, USA Cu 99.999 100 - - Atlantic Engineers, USA 13_Al2O3_IW 13 Al2O3 IW Sigma-Aldrich 50_TiO2_MM 50 TiO2 MM ICPC Poland 45_ZrO2_FG 45 ZrO2/MgO FG Chalmers University Sweden 16_SIO2_IW 16 SiO2 IW University of Utah 42_SIO2_IW 42 SiO2 IW University of Utah 68_SIO2_IW 68 SiO2 IW University of Utah OC Name CuO Wt% 57 3 6𝐹𝐹𝐹𝐹𝑂3 + 𝑂2 ↔ 2𝐹𝐹2 𝑇𝑇3 𝑂9 + 𝐹𝐹2 𝑂3 2 (19) At temperatures above 800°C, two reactions occur: 4𝐹𝐹𝐹𝐹𝑂3 + 𝑂2 ↔ 2𝐹𝐹2 𝑇𝑇𝑂5 + 2𝑇𝑇𝑂2 (20) 𝐹𝐹2 𝑇𝑖3 𝑂9 ↔ 𝐹𝐹2 𝑇𝑇𝑂5 + 2𝑇𝑇𝑂2 (21) where TiO2 is rutile and pseudobrookite (Fe2TiO5) is the most stable phase [1]. Therefore, theoretically, the oxygen transfer capacity is 5 wt% between the states FeTiO3 and Fe2TiO5. Ilmenite has a melting point of 1470°C [2]. Ilmenite was tested in a number of different campaigns involving the FzB and TGA. Ilmenite has also been studied as a potential substrate for a copper oxide-based oxygen carrier. Due to the CLC capabilities and enhanced physical durability of iron oxide and the CLOU capabilities of copper, a material was created in an attempt to create an inexpensive copper-based oxygen carrier. Ilmenite performed well within the fluidized-bed. It is important to note that ilmenite does not spontaneously decompose and, therefore, must be reduced by a gaseous fuel (i.e., methane, natural gas, hydrogen, synthesis gas etc.). For the redox reactions of ilmenite, air was used as the oxidizer whereas 5% CH4 in N2 was used as the fuel. At lower temperatures, the oxidation of ilmenite proceeds very slowly, but can reach completion within 10 minutes at temperatures above 850°C. Figure 15 shows the completion profile for the oxidation of ilmenite in fluidized-bed whereas Figure 16 shows 58 1 0.9 % Completion 0.8 0.7 0.6 0.5 700 C 750 C 800 C 850 C 0.4 0.3 0.2 0.1 0 0 200 Time (Sec) 400 Figure 15: Oxidation of ilmenite at various temperatures in FzB using air. the rates of oxidation at the corresponding temperatures. With high physical durability evidenced by low attrition rates and high resistance to agglomeration, ilmenite proved a suitable candidate for a CLC unit employing a fluidized-bed design. It has been reported that ilmenite displays slower redox rates than other oxygen carrier candidates, especially during initial cycles, and displays an activation period after which it shows more attractive reaction rates and has even been shown to increase the rate of gasification of a bituminous coal [2]. 4.1.2 Pure Copper With ilmenite as a comparison, the study of copper-based oxygen carriers commenced. To form a baseline, pure copper was first tested in both TGA and FzB 59 Oxidation Rate (g/s/g_ilmenite) 4.0E-04 3.5E-04 700 C 750 C 800 C 850 C 3.0E-04 2.5E-04 2.0E-04 1.5E-04 1.0E-04 5.0E-05 0.0E+00 0 30 60 Time (sec) 90 120 Figure 16: Oxidation rate for ilmenite in TGA under air at several temperatures. studies. It was quickly understood that the use of copper as an oxygen carrier would be impractical for a fluidized-bed apparatus. Copper has a melting point below 1000°C and becomes soft at temperatures much lower than that. For CLOU operation of copper, the fuel reactor must be above 700°C. This may be understood by inspection of Figure 17. For the spontaneous decomposition of CuO, the kinetic driving force is the difference between the equilibrium partial pressure of oxygen and the actual partial pressure of oxygen in the reaction zone. Therefore, at temperatures below 850°C, reaction rates begin to slow significantly. Because of the need for higher temperature operation, the physical durability of an oxygen carrier must not allow for agglomeration of particles at these temperatures. Agglomeration of oxygen carrier, even prior to sintering, may cause catastrophic failure of the reactor system. Particulate agglomeration of copper particles was observed at temperatures near Partial Pressure of O2 (atm) 60 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 CuO Cu2O 850 900 950 Temperature (°C) 1000 Figure 17: Equilibrium partial pressure of oxygen over the CuO/Cu2O redox pair. 600°C during TGA testing, but that temperature was expected to increase with the harsh agitation of a fluidized-bed environment, allowing for operation of copper particles at temperatures near 900°C. Figure 18 shows the agglomeration of 99.9% copper powder during chemical-looping cycling above 900°C. The agglomeration tendencies of copper force the use of support materials in a fluidized-bed design. 4.1.3 Supported Copper-based Oxygen Carriers 4.1.3.1 Physical Performance and Characterization The literature is replete with reports of various materials used as supports for a variety of oxygen carrier materials. These supports include materials such as: Al2O3, bentonite, SiO2 and TiO2, ZrO2/MgO, MgAl2O4,YSZ, NiAl2O4, etc. [2,3]. Each of these supporting materials is reported as having varying levels of success and suitability as 61 Figure 18: Agglomeration of 99.9% pure copper at 900°C during chemical-looping cycling. oxygen carriers. For monetary and availability incentives, it is generally beneficial to find a material that is already in production - either as a primary product or secondary byproduct of another process. A material was obtained from Sigma-Aldrich sold as 13 wt% CuO on Al2O3 support (Figure 19). The material is sold as solid spheres in a range of 14-20 mesh (approximately 1,400 to 840 μm). The material was purchased and crushed using a mortar and pestle then sieved to a size range of 75 μm < Dp < 105 μm. The powdered material was then tested at various temperatures under CLOU conditions within the fluidized-bed and TGA. Since the particle size of the material as received was 14-20 mesh, it was necessary to crush and sieve the material to a more suitable size range (75 - 150 μm). From a physical durability standpoint, this material performed rather well. The 62 Figure 19: SEM micrograph of 13 wt% CuO on Al2O3. material exhibited no agglomeration at the tested temperatures and the observed attrition rates were comparable to the other materials tested. However, this material performed poorly from a chemical standpoint. This will be discussed in a subsequent section. As commercially available copper-based powders suitable for the chemicallooping process are not overly abundant, the Institute for the Chemical Processing of Coal (ICPC) in Poland was contracted to prepare a 50 wt% CuO material on a titania (TiO2) substrate. The material was produced using a mechanical mixing technique at a cost of $5,000 USD/kg. The material was observed using SEM. Figure 20 shows the SEM micrograph of this material as received. The surface of the material appears very rough when compared to the alumina material shown in Figure 19. The titania material was produced using a mechanical mixing technique. This technique employs the use of a machine such as a ball mill or other apparatus that physically forces one material to meld 63 Figure 20: SEM micrograph of 50 wt% CuO on TiO2. with another through sheer force of impact. Figure 20 provides the SEM micrograph images of the titania material. The roughly spherical balls appear to be conglomerates of much smaller particles. This fragmented topography is likely due to the method of production. The titania material did not perform as well in the fluidized-bed as was hoped. The material was subject to agglomeration at temperatures greater than 850°C and displayed the largest attrition values recorded compared to the other materials tested. To combat the poor physical performance of the carrier, a diluent powder was added to help curb the agglomeration of the material. The material was diluted by 50 wt% zirconia 64 silicate beads in the size range of 125 microns in diameter. The addition of these beads allowed the material to be tested at temperatures up to 1000°C with minimal agglomeration of particles. The effective CuO loading of the diluted material went from 50 wt% to 25 wt%. The diluted material was used in the testing of solid fuels to be discussed in a subsequent section. Another acquired material was donated by the Chalmers University of Technology in Sweden. This material (Figure 21) was prepared by a freeze granulation method that involves mixing and binding of the metal oxide and support material and then spraying into a cold bath to create well-mixed spherical particles. Cupric oxide was loaded onto a zirconia-based support which was stabilized by magnesia at a tested loading of 45 wt% CuO using this method. Of all materials tested, this material displayed the greatest physical attributes and was tested at temperatures up to 1000°C without any sign of agglomeration without using diluents for added support. SEM/EDS micrographs of four of the copper oxide-based materials tested are presented in Table 8. For comparison, rows 1 and 2 show the titania material as a fresh sample as well as after reaction. There appears to be no large qualitative difference between the two. This would seem to suggest that there is no migration of the copper compounds throughout the particle, but, instead, that the copper oxide stays near the surface of the particle. Table 9 displays the corresponding porosity measurements. Additional characterizations are provided in materials. Some materials show higher attrition at the lower temperature while others show higher attrition at the higher temperature. Gas velocity, expectedly, does appear to have a large influence on the rates of attrition. 65 Figure 21: SEM micrograph of 45_ZrO2_FG particles as received. Attrition results from the collisions between the oxygen carrier particles within the fluidized-bed. In order to compare collisions of equal intensity, the gas velocities used were determined from the minimum fluidization velocity. While a similar particle size was used between the materials, each of the materials tested has a different density and, therefore, a different minimum fluidization velocity. The ratio of the gas velocity (U) to the minimum fluidization velocity (Umf) was varied between 5 and 10. A greater attrition rate was observed for all materials at the higher ratio. Table 10 includes the crushing strength, BET surface area and pore size data. It is interesting to note that the zirconia material exhibited the smallest crushing strength. It has been suggested in the literature that crushing strength may be an indication of attrition rates. Results of an investigation of this claim are presented in Figure 22 - Figure 25. 66 Table 8: SEM/EDS Micrographs of four copper oxide-based materials tested for oxygen carrier suitability in CLOU operation. Courtesy: Crystal Allen. Oxygen Carrier Un-reacted 50_TiO2_MM Reacted 50_TiO2_MM Reacted 45_ZrO2/Mg O_FG Reacted 50_SiO2_IW Reacted 70_SiO2_IW 100×Magnification Metal Support 67 Table 9: Surface area and pore size data for four copper based carriers studied as oxygen carriers in CLOU systems. Data courtesy of Crystal Allen. Oxygen Carrier Un-reacted 50_TiO2_MM Reacted 50_TiO2_MM Reacted 45_ZrO2/MgO_ FG Reacted 50_SiO2_IW Reacted 70_SiO2_IW Particle Diameter (μm) BET Surface Area (m2/g) Total Volume of Pores 1.7-300 nm Width (cm3/g) Average Pore Width (nm) 75 - 105 0.6178 0.001392 10.2463 75 - 105 0.1656 0.000286 27.9278 75 - 105 0.8963 0.001926 8.3488 106 - 250 0.0823 0.000375 40.7755 106 - 250 0.0726 0.000883 109.2892 Table 10: Results of crushing strength testing performed on four materials tested for suitability as oxygen carriers in CLOU systems. Courtesy: Crystal Allen Oxygen Particle Diameter Crushing Strength Standard Carrier (μm) (N) Deviation 180 - 250 4.45 1.14 180 - 250 3.94 1.28 180 - 250 2.21 0.88 180 - 250 4.10 1.26 180 - 250 4.48 1.02 Un-reacted 50_TiO2_MM Reacted 50_TiO2_MM Reacted 45_ZrO2/MgO_FG Reacted 50_SiO2_IW Reacted 70_SiO2_IW 68 0.0160% 50_TiO2_MM 0.0140% 45_ZrO2/MgO_FG 0.0120% 50_SiO2_IW 0.0100% 70_SiO2_IW 0.0080% 0.0060% 0.0040% 0.0020% 0.0000% Figure 22: Observed attrition rates during redox cycling of four different materials tested for suitability as oxygen carriers in CLOU systems. Figure courtesy of Crystal Allen. 0.0600% 50_TiO2_MM 0.0500% 45_ZrO2/MgO_FG 0.0400% 50_SiO2_IW 0.0300% 70_SiO2_IW 0.0200% 0.0100% 0.0000% 1 2 3 4 5 6 7 8 Figure 23: Observed bed weight percent loss per hour of testing for four materials tested for oxygen carrier suitability in CLOU systems. Figure courtesy of Crystal Allen. 69 0.0300% 0.0250% 0.0200% 50_TiO2_MM 0.0150% 45_ZrO2/MgO_FG 0.0100% 50_SiO2_IW 70_SiO2_IW 0.0050% 0.0000% 800 900 Temperature, Celsius Figure 24: Influence of temperature on attrition of copper oxide-based materials in a fluidized-bed. 0.3500% 0.3000% 0.2500% 50_TiO2_MM 0.2000% 45_ZrO2/MgO_FG 0.1500% 50_SiO2_IW 0.1000% 70_SiO2_IW 0.0500% 0.0000% 5 10 Figure 25: Influence of gas velocity on attrition rates. Gas velocity is given as U/Umf calculated using correlations given by Kunii and Levenspiel. Figure courtesy of Crystal Allen. 70 Figure 22 displays an overall view of weight loss from the fluidized-bed during CLOU cycling. It is clear that the titania material displayed the lowest attrition resistance while the zirconia material exhibited the highest attrition resistance showing the least amount of bed loss over all conditions. This is especially interesting when considering the crushing strength information given in Table 10. While the zirconia material had the lowest crushing strength, it performed the best when considering attrition. However, demonstrating a much higher crushing strength, the titania material displayed much greater attrition. Therefore, crushing strength is a poor indicator of attrition resistance; at least when considering materials prepared using different methods. Interestingly, when comparing materials prepared using the same method, crushing strength appears to effectively predict attrition characteristics. The 70_SiO2_IW material has a higher crushing strength than its lower CuO loading counterpart 50_SiO2_IW and shows greater attrition resistance. Another interesting note is gleaned from Figure 23 which shows the observed weight loss per hour for each of the materials investigated. The results show a very high attrition for the titania material for the first 2 hours followed by a significantly decreased rate. In fact, all four materials showed a decrease in the amount of material leaving the bed over time. Comparing the topographies of the titania material (Figure 20) and the zirconia material (Figure 21), it may be understood that the difference in early stage attrition is due to the very rough nature of the titania material, where the zirconia material is much smoother and exhibited much lower attrition. Figure 24 displays the influence of temperature on the attrition rates of the materials. It does not appear that there is any effect which is observed by each of the 71 materials. Some materials show higher attrition at the lower temperature while others show higher attrition at the higher temperature. Gas velocity, however, does appear to have a large influence on the rates of attrition, as may be seen in Figure 25. Attrition results from the collisions between the oxygen carrier particles within the fluidized-bed. In order to compare collisions of equal intensity, the gas velocities used were determined from the minimum fluidization velocity. While a similar particle size was used between the materials, each of the materials tested has a different density and, therefore, a different minimum fluidization velocity. The ratio of the gas velocity (U) to the minimum fluidization velocity (Umf) was varied between 5 and 10. A greater attrition rate was observed for all materials at the higher ratio. 4.1.3.2 Chemical Characterization Under CLOU Conditions The failed operation of the pure copper oxide powder due to its poor physical performance merits not going into detail into its chemical performance; therefore, it is not treated here. However, all other carriers performed successfully enough physically to merit some discussion into their respective chemical performances. While no agglomeration was observed at any temperature during testing of the alumina material, the overall performance of the carrier was very poor as a CLOU material. The % mass change decreased after each redox cycle under CLOU conditions. This is likely due to the formation of Cu-aluminate spinal compounds. These compounds, while still reducible, are not CLOU capable and, therefore, do not spontaneously uncouple oxygen at elevated temperatures and was deemed not optimal for the combustion of solid fuels in a fluidized-bed design of a chemical-looping process. 72 The titania material far outperformed the alumina material when considering CLOU operation. There was no decrease in reactivity after 20 cycles during cycling within TGA which may be seen from Figure 26. The material achieved full conversion under a variety of conditions and temperatures ranging from 600°C to 1000°C. The trace shown in Figure 27 shows the zirconia-supported material tested for 20 cycles between air and nitrogen at 925°C in TGA. The trace shows a decrease in the mass of the sample over time. While the sample mass slowly decreases, the change in mass between the fully oxidized and fully reduced remains nearly constant. Therefore, this drift is likely just that, a drift in the TGA balance, and does not likely represent a mass loss of the sample. The ZrO2/MgO material was tested under many different conditions. The material was tested for use in solid fuels combustion as well as normal CLOU cycling. The results of TGA testing for both the oxidation and decomposition reactions were used to obtain the relevant reaction rate constants and expressions. This information was used in conjunction with the same gathered from TGA results from cycling of the TiO2 supported material. These rate expressions were used to develop a model that may be applied to predict the characteristic redox patterns of the CuO/Cu2O pair. 4.1.4 Oxidation and Decomposition Kinetics The rate expression obtained for the decomposition of CuO using the titania- and zirconia- supported materials was developed using the methods described in Chapter 6. The developed rate expression was then used to predict reaction rates of other materials. These predictions were then compared to rate data obtained during CLOU testing of two Mass (mg) 73 58.5 58 57.5 57 56.5 56 55.5 55 54.5 54 0 500 1000 Time (minutes) Figure 26: Cycling of 50_TiO2_MM in TGA for 20 loops using air for oxidation and nitrogen during decomposition at 925°C. 23.2 Sample Mass (mg) 23 22.8 22.6 22.4 22.2 22 0 500 1000 Time (Minutes) Figure 27: Multiple redox cycling for the 45_ZrO2_FG material at 925°C in TGA under air and nitrogen. 74 other materials. These other materials differed from the previous two in CuO wt% loading (16 wt% and 64 wt%), production method (Incipient Wetness using a Rotary Evaporator), support type (SiO2) and area produced (in house at the University of Utah). The CuO/SiO2 materials exhibited rates and characteristics similar to those observed with the titania and zirconia materials. The obtained model (Table 11) generated Figure 28 which compares the actual extracted decomposition rates of the oxygen carriers and the rates predicted by the model. More information on this study is provided in the manuscript titled Measurement and Modeling of Decomposition Kinetics for CopperBased Chemical-Looping with Oxygen Uncoupling in Appendix A. For the modeling and scale-up of copper oxide-based CLOU reactors, it is important to understand the kinetics of both decomposition and oxidation of the oxygen carrier. While it appears that the decomposition of cupric oxide to cuprous oxide may have a universal set of kinetic parameters that may be applied globally for all copperbased oxygen carriers, the same may not be true for the reverse reaction (oxidation of Cu2O). Details into this investigation are provided in Appendix B. This section will only serve to highlight the results discussed therein. Similar to the decomposition reaction of CuO, the oxidation of Cu2O is largely dependent on a driving force that is a function of the partial pressure of oxygen around the sample and the equilibrium partial pressure of oxygen for the CuO/Cu2O redox pair at each temperature. For the decomposition of CuO, the reaction rate is accelerated as the temperature is increased by both a direct and indirect affect. The direct affect is explained by use of the activation energy for the reaction, while the indirect affect is 75 Table 11: Constants for kinetic rate expression modeling the decomposition of CuO. 50_TiO2_MM 45_ZrO2_FG Value in Rate Expression Activation Energy, Ea (kJ/mol) 67 58 62 Frequency Factor, A (atm-1×s-) 4.15 × 10-4 3.64 × 10-4 3.90 × 10-4 Order of PO2, α (atm) 1.0 1.0 1.0 Order of CCuO, β (-) 0 0 0 Parameter 0.01 50_TiO2_MM Series3 16_SiO2_IW 64_SiO2_IW Corrected Ea Apparent Ea 0.009 Rate (gO2/gCu/s) 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0 700 750 800 850 Temp (°C) 900 950 1000 Figure 28: Rates of decomposition of four different copper-based oxygen carriers at several temperatures using nitrogen as inert purge during decomposition compared with the rates predicted by the developed kinetic rate expression. 76 caused by the increased equilibrium partial pressure of oxygen with temperature. Unlike decomposition, the reaction rate of oxidation is slowed by the increase of the equilibrium partial pressure of oxygen. Therefore, as the temperature is increased above 850°C, the observed reaction rate begins to slow. This results in a seemingly negative activation energy at high temperatures, but, in reality, is explained within the driving force term. Figure 29 shows the rate of oxidation at several temperatures for the 45_ZrO2_FG material using air as the oxidizer. The decreased reaction rate is readily observed above about 850°C. While this phenomenon is easily explained by the decreased driving force with increased temperature, there are other mechanisms at work during the oxidation of Cu2O that further complicate determination of an applicable kinetic expression. In fact, while one mechanism appears to be at work during lower temperature oxidation, another mechanism seems to be more representative of the observed kinetics. For low temperature oxidation of Cu2O, the pore-blocking model (a logarithmic rate expression) best describes the observed kinetics. This is likely due to an increase in the molar volume of the solid during oxidation from Cu2O to CuO (approximately 5 vol%). The extracted values describing the pore-blocking kinetic expression are provided in Table 12. While useful for understanding the oxidation of cuprous oxide, the low temperature kinetic expression is not very applicable to CLOU operation. Generally speaking, the temperatures at which the air reactor will be operated is likely over 900°C. With this in mind, it is obvious that a kinetic expression should be developed closer to the likely temperature range of operation. For the oxidation of Cu2O at temperatures above 850°C, the applicable kinetic 77 0.009 0.008 rate (g/s/gCu) 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0 600 700 800 900 1000 Temperature (C) 1100 Figure 29: Oxidation rate extracted at 50% conversion for the 45_ZrO2_FG material at temperatures from 750 - 1000°C. Table 12: Results obtained from modeling of low temperature oxidation of two CLOU materials. Influence 50_TiO2_MM X 𝑓(𝑋): 1/exp 𝑝𝑂𝛼2,𝑒𝑒 λ 𝑓𝑝𝑂2 : 𝑝𝑂𝛼2 − 𝑝𝑂𝛼2 ,𝑒𝑒 𝐸 𝑓(𝑇) = 𝐴 × 𝑒𝑒𝑒 − 𝑅𝑅𝑎 45_ZrO2_FG 𝜆 = −2.95 × 10−4 × 𝑇 + 0.34 𝜆 = −4.16 × 10−4 × 𝑇 + 0.48 𝛼 = 1.3 𝛼 = 1.3 𝑃𝑂2 (𝑎𝑎𝑎) = 6.057 × 10−11 𝑒 0.02146 𝑇(°𝐶) A =6.2 × 10−13 (min-1); Ea =172 (kJ/mole) A = 3.4× 1012 (min-1); Ea = 165 (kJ/mole) 78 expression was developed assuming the Avrami-Erofeev nucleation/growth kinetics equation. The true scientific nature describing the transition from the pore-blocking kinetic expression to the nucleation/growth kinetic expression is beyond the scope of this work, but is possibly a result of increased diffusion rates to the point at which pores blocked by the expansion of the molar volume do not truly hinder the diffusion of reactive species (likely the diffusion of Cu through CuO). This requires a closer investigation for further understanding. The kinetic constants used for the completed model are provided in Table 13. The developed model failed to effectively capture the conversion profile of the tested materials at low temperatures, but seemed to capture the higher temperature oxidation of Cu2O. Figure 30 displays the model compared to the obtained data for both cases. 4.1.5 Production and Evaluation of Novel Materials Along with the materials obtained from other institutions, a few materials were produced and tested in house at the University of Utah. First among these materials were produced from a high surface area SiC. Due to the oxidation of SiC at CLOU temperatures, the material was baked in air within a muffle furnace for roughly a week. The oxidation of SiC, even at high temperatures, is very slow and required the whole week to form 99% SiO2. Various productions methods were employed during the creation of this material in an effort to find the optimal loading and distribution of CuO. In the end, the material of choice was produced by first loading CuO onto the surface of the SiC powder using a rotary evaporator. The CuO-coated SiC material was then placed within the muffle furnace to convert the SiC to SiO2. The method of coating the particle 79 Table 13: Model used to predict the conversion of Cu2O in two different oxygen carriers using the nucleation/growth kinetics expression Avrami-Erofeev. Influence Model Values 𝑓(𝑋) 1−𝑋 𝑃𝑂2 (𝑎𝑎𝑎) = 6.057 × 10−11 𝑒 0.02146 𝑇(°𝐶) 𝑝𝑂2,𝑒𝑒 𝑓𝑝𝑂2 : 𝑝𝑂𝛼2 − 𝑝𝑂𝛼2 ,𝑒𝑒 𝛼 = 1.3 𝐸 A = 7.90× 104 (min-1); Ea = 69 (kJ/mole) 𝑓(𝑇) = 𝐴 × 𝑒𝑒𝑒 − 𝑅𝑅𝑎 1 0.9 0.8 Conversion 0.7 0.6 0.5 45_ZrO2_FG 0.4 50_TiO2_MM 0.3 Prediction 0.2 0.1 0 0 0.2 0.4 0.6 Time (minutes) 0.8 1 Figure 30: Conversion profiles for the oxidation of Cu2O at 925°C for 45_ZrO2_FG (□) and 50_TiO2_MM (Δ) are compared against the model prediction (- - -). 80 first, then baking within the furnace provided a means whereby the CuO was thermally driven beneath the surface of the particles. This thermally driven migration is evidenced in Figure 31 which shows SEM micrographs of the material before and after baking with the corresponding EDS images. The copper seen in the first row seems to disappear in the figure below. This corresponds to the increased level of Si seen in the second row compared to the first. By utilizing the coat-then-bake method (CTB) as opposed to the bake-then-coat (BTC) method, the observed agglomeration temperature ceiling was increased from 850°C to over 1000°C for two of the materials. The measured agglomeration temperatures are given in Table 14. The performances of these materials are more adequately presented in Appendix D. 4.2 Conversion of Solid Fuels The results from the combustion of solid fuels using different oxygen carriers are presented below. Three different fuels (PRB, Illinois #6 and Petroleum Coke) were used for this study and were tested using two different oxygen carriers (50_TiO2_MM and 45_ZrO2_MM). The proximate and ultimate analyses of the tested fuels are provided in Table 6. For the combustion of solid fuels, many factors play major roles in the design of industrial-scale combustors. For example, the selection of an appropriate fuel will have a major impact on the size and design of the boilers. A higher reactivity fuel will require a shorter residence time, allowing for a smaller boiler design. Another factor is the water and/or ash content of the fuel. The higher the water content, the less efficient the fuel, thus requiring more fuel to achieve the same thermal output. A higher ash content can 81 C C-Cu C-Si 50μm Figure 31: SEM and corresponding EDS images for elemental Cu and Si of RV-CTB-15 (a) before baking, (b) after baking, and (c) after fluidization. Table 14: Measured agglomeration temperature in the fluidized-bed Oxygen Carrier CuO loading (wt%) Agglomeration temp. (°C) RV-BTC-20 19 850 RV-CTB-15 15 > 1000 RV-CTB-40 40 > 1000 RV-CTB-60 61 950 82 also affect the thermal efficiency of the boiler, and can also cause great concern due to slagging (collection and cooling of molten bottom ash within the boiler) which reduces heat transfer efficiency to water pipes. In the case of a chemical-looping combustor, these factors must all be considered. For example, the collection of ash within the fluidized-bed may increase solids inventory and decrease OC circulation rates, thereby decreasing the overall efficiency of the process. Therefore, proper selection of a fuel is critically important to the overall design of a chemical-looping combustor. The following results will help to highlight some of these concerns and potential design considerations to overcome possible challenges. Petcoke is stripped of most of the volatile content from the creation process, hence, it is less reactive than the coals studied. The Powder River Basin (PRB) coal is a sub-bituminous and thus has a lower rank than the bituminous Illinois #6 coal. The lower rank coals generally have higher reactivities. This ordering may be observed in Figure 32 where the conversion of carbon versus time is displayed for the three fuels during combustion using 50_TiO2_MM material at 930°C. The PRB reacts very quickly, reaching full conversion in just a few seconds. The Illinois #6 coal and petcoke, however, do not reach full conversion. From a stoichiometric standpoint, the oxygen supplied by the metal oxide carrier should be more than sufficient to fully combust all of the fuels. The amount of fuel supplied to the reactor was evaluated by determining the amount of oxygen to be released from the total sample carrier between 80% and 20% conversion of the carrier. Therefore, the amount of oxygen released from the carrier should be 40% more than is necessary for complete carbon burnout of the fuel. A simple explanation could be that the unreacted fuel was simply blown out of the reactor before it Carbon Conversion 83 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Illinois #6 PRB 0 100 200 300 Time (Seconds) 400 Figure 32: Combustion of three different fuels using 50_TiO2_MM material at 930°C. The conversion is calculated from the mass of carbon (as gas) analyzed and collected by infrared analyzer. could be consumed. If that is the case, the ensuing oxidation cycle would yield no additional carbonaceous products - or, at least, it would not account for all of the missing carbon. The full conversion profile of the petcoke fuel is given in Figure 33. Clearly, as the conversion approaches 50%, the rate slows considerably and is near zero around 700 seconds of reaction. Air is once again cycled to the reactor around 700 seconds and the missing carbon is burned off. It is clear from this figure that the fuel is not being ejected from the reactor, but is simply not completely reacted within the fluidized-bed. The full reaction profile obtained during the combustion of Illinois #6 coal is very similar to the petcoke, in that it is not fully combusted until the air is once again cycled through the reactor. Once again, the incomplete combustion of these fuels is not a result of insufficient 84 1 0.9 Carbon Conversion 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Air Flow Rate (SLPM) 0 1 0.5 0 0 200 400 600 Time (Seconds) 800 Figure 33: Conversion of carbon during combustion of green petcoke at 930°C using 50_TiO2_MM in a fluidized-bed. oxygen supply. Certainly, the fuel is oxygen starved, but it is due to the slow reactivity of the fuel and not due to an excessive charge of fuel to the reactor. This characteristic is accentuated and more easily understood by looking at the effects of increasing the fuel particle size. The combustion of PRB particles ranging from 150 μm to 6,000 μm (0.3 cm) at 930°C using the 45_ZrO2_FG material is given in Figure 34. The three tests were performed by selecting particles that fit the desired size range while having a very similar combined mass. For example, the sample tested in the 6,000 μm range consisted of only one particle. This single particle had a mass very similar to the combined masses of the other two samples individually. While, once again, the combustion of the small particle size (150 μm) was completed very quickly (roughly one minute), the other two samples did not reach full conversion of the carbon even though the masses of each of the three Fuel Conversion 85 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 150 Microns 3000 Microns 6000 Microns 0 100 200 300 Time (Seconds) 400 500 Figure 34: Carbon conversion is tracked during the combustion of PRB at 930°C using 45_ZrO2_FG in a fluidized-bed. samples was very nearly equal. Admittedly, it may be possible to have complete burnout of the larger fuel particles if the amount of oxygen carrier is increased. However, it is likely the case that this will not be due to the increased amount of available oxygen, but instead, due to an increase in the reaction time due to a deeper particle bed. The slower decomposition rates may be achieved by the oxygen carrier simply due to an increased oxygen partial pressure. The increase in the oxygen partial pressure effectively decreases the driving force for decomposition. This phenomenon is discussed further in Appendix A and Appendix B where the kinetic characteristics of both oxidation and decomposition of 86 the CuO/Cu2O redox pair are discussed in detail, including the effect of the driving force on the reaction. The reactivity of PRB was previously mentioned as being higher than the other two fuels tested. It may be assumed, however, that even for a more reactive fuel, there exists a particle size threshold, above which a particular oxygen carrier may be ineffective, not because it decomposes too slowly, but to the contrary, because it decomposes too quickly. The higher rates of decomposition then produce oxygen that is unreacted and dilutes the fuel reactor product stream. It should be noted that the driving force fueling the research of technologies such as chemical-looping is the production of a nearly pure CO2 stream which may easily be sequestered. Excess air in the fuel reactor will dilute the CO2, stream making purification of the CO2 more challenging and decreasing the effectiveness of the technology. Figure 35 shows the conversion of the solid oxygen carrier from CuO to Cu2O during the combustion of large petcoke particles (roughly 3,000 μm) using 45_ZrO2_FG at 930°C and 960°C. The conversion of the solid oxygen carrier material at 960°C, is shown as well. At roughly 200 seconds the oxygen carrier is nearly completely spent and the rate of decomposition slows, essentially to zero. In both cases, the fuel is incompletely combusted and stops as the oxygen is no longer devolved. Due to the increased reactivity of the petcoke at 960°C, the conversion profile for that temperature is extended. Once again, while the amount of oxygen is sufficient for the combustion of this fuel, it is simply released and swept away too quickly to react with the remaining petcoke. Figure 36 suggests that the oxygen contained within carbonaceous gas evolved 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Carbon Conversion 0.8 0.7 930 C 0.6 960 C 0.5 0.4 0.3 0.2 0.1 0 0 200 Time (Seconds) Conversion of CuO to Cu2O 87 400 Figure 35: Conversion of carbon during the combustion of petcoke particles with a diameter of roughly 3,000 μm at two temperatures using the 45_ZrO2_FG material as oxygen carrier is compared against the conversion of the oxygen carrier (---), 1 % Of Oxygen Released 0.9 0.8 0.7 % of Liberated Oxygen Reacted % of Liberated Oxygen Unreacted 0.6 0.5 0.4 0.3 0.2 0.1 0 0 50 100 150 Time (Seconds) 200 Figure 36: Depiction showing the amount of oxygen liberated from the reactor unreacted compared to the oxygen reacted during combustion of petcoke at 960°C in a fludized-bed with 45_ZrO2_FG as oxygen carrier. 88 during the combustion of the large petcoke particles accounted for only about 30% of the total evolved oxygen. The remaining near 70% of the liberated oxygen was swept through the outlet as unreacted oxygen. This number actually speaks well to the performance of the oxygen carrier. The challenge of having too much oxygen in the effluent of the fuel reactor may be overcome by simply adding a simple recycle loop. This will deliver unreacted oxygen back to the reactor which will also serve to effectively slow the rate of oxygen carrier decomposition. The slower rates of decomposition would prove more effective for the combustion of less reactive fuels such as petcoke or larger fuel particles; both cases would require a more substantial residence time for complete combustion. One of the very large challenges in the combustion of solid fuels is understanding how best to deal with the resulting ash. In the case of chemical-looping, the general desire is to handle the separation of oxygen carrier from the ash simply by the differences in their respective fludization characteristics. Generally speaking, the lower the density and smaller the particle, the easier it is to fluidize. In the case of ash associated with solid fuels, that particle sizes are very small and the density is much lower than the solid fuel. Since most oxygen carriers are made from metals (or metal oxides), it can generally be assumed that it is more difficult to fluidize the oxygen carrier than it is to fluidize the resulting ash. By creating an environment where the fluidization properties encourage the expulsion of the smaller ash particles, while still not quite large enough to eject the oxygen carrier, this separation may be effectively and efficiently mitigated. Unfortunately, the bubbling fluidized-bed did not exhibit such behavior. A set of in-line filters has been installed into the fluidized-bed reactor system with the hope of 89 catching any elutriated fines (one may hope that coal ash is on that list of elutriated fines). Unfortunately, as one may see in Figure 37, that hardly anything is collected on the filter, while the product oxygen carrier is riddled with coal ash. 90 Figure 37: Coal ash that has collected within the particle bed and is not elutriated during the combustion of different solid fuels above 900°C using the 45_ZrO2_FG material shown here. 91 4.3 References [1] Chen, Y. (1997). Low-temperature oxidation of ilmenite (FeTiO3) induced by high energy ball milling at room temperature. J Alloy Compd. 257 (1-2):156 - 160. [2] Wang, C. Y., & Zhou, M.-F. (2004). Mineral Chemistry of Fe-Ti oxides from Xinjie PGE-bearing layered mafic ultramafic intrusion in Sichuan, SW China. Mineral Deposit Research, meeting the global challenge (pp. 481-485). Beijing: Springer. [3] Cuadrat, A. Chemical-looping Combustion of Coal Using Ilmenite as Oxygen Carrier. Ph.D. Dissertation, Instituto de Carboquimica (ICB-CSIC). Zaragosa, Spain, 2012. 92 CHAPTER 5 CONCLUSIONS Chemical-looping is quickly becoming a suitable alternative for the cleaner combustion of fossil fuels. While there is still much that needs to be understood before this technology may be commercialized, the promise that this technology holds gives hope to those involved in its development. Several materials have been suggested as possible suitable candidates for use in a chemical-looping system. Among these, copper oxide stands out toward the front. Copper oxide has a relatively high oxygen carrying capacity, high reactivity and creates an overall endothermic reaction in the fuel reactor when oxidizing a number of fuel sources. The most notable characteristic of copper oxide is that it is CLOU capable. This means that at certain temperatures, both an oxidized state (CuO) and a reduced state (Cu2O) are thermodynamically favored, depending on the partial pressure of oxygen within the system. This characteristic enables the cycling between CuO and Cu2O without a fluid fuel. Much improved reaction rates are observed during the combustion of solid fuels without the need for initial gasification of the fuel. A seemingly plethora of potential copper oxide-based material have been suggested as suitable materials for use in a chemical-looping system. Among them, the use of high surface area SiC and ilmenite have proven effective due, in large part, to their 93 ease of production and relatively low cost. Other prepared materials may possibly exhibit slightly better physical durability, but it is challenging to compete with a cost-effective support such as the abundantly available ilmenite. The ilmenite-supported oxygen carrier performed very well both physically and chemically. There was no long-term decrease in reactivity and it has the added benefit of being available for CLC combustion. The future of chemical-looping will be shaped using metal blended oxygen carriers. These type of carriers can be highly cost effective while providing the positive benefits from both metal oxide sources. Because copper oxide is such an attractive material for a chemical-looping system, it becomes necessary to more adequately understand the mechanisms responsible for both the oxidation and decomposition of the CuO/Cu2O redox pair. An effective model describing the kinetics of oxidation and another describing the kinetics of decomposition have been developed. These models adequately describe the observed behavior of the relevant copper oxide states while remaining simple enough to be effectively utilized in larger-scale modeling of a copper oxide-based chemical-looping system. Moreover, a single equation which adequately describes the decomposition of CuO over a wide range of supports and production methods has been developed. This, in particular, is useful for larger-scale modeling applications to effectively compare the theoretical performance of a host of copper-based carriers. The decomposition of CuO under CLOU conditions is adequately described with an activation energy 62 kJ/mol and a pre-exponential constant of 3.9 ×10-4. The order of reaction with respect to the partial pressure of oxygen is 1 and the reaction is independent of the solid conversion. 94 The oxidation of Cu2O, on the other hand, is not so easily modeled. For low temperatures, the oxidation mechanism appears to obey a pore-blocking rate law while there appears to be a transition around 700 - 800°C where the conversion profile begins to more accurately follow a nucleation/growth rate model. The Avrami-Erofeev model adequately described the conversion profile with n = 1. The influence of the solid, unlike the reverse reaction, appears to be pseudo-first order. This pseudo-first order influence incorporates potential underlying mechanisms such as solid state Cu+ diffusion. Copper oxide-based materials have been proven suitable as oxygen carriers for the combustion of solid fuels. The successful combustion of solid fuels was shown in this work while using two different oxygen carrier materials. The fuels with lower reactivities create challenges that are easily overcome with simple engineering strategies, whereas the higher reactivity fuels may successfully employ the use of larger particles. The combustion of PRB at 930°C is completed very rapidly - around 30 seconds for 0.1 g. This equates to a consumption rate of 0.003 gfuel/s. The developed kinetic equation predicts an oxygen release rate of 0.0036 gO2/s-gCu. For the lower reactivity fuels, while the oxygen release rate remains essentially the same, the combustion rate drops thereby releasing unspent oxygen into the flue gas. There are still plenty of opportunities for the continued research of chemicallooping technologies. A great many topics relating to this technology still have yet to be adequately researched. Among those topics needing the most attention: 1. Determining the most cost-effective oxygen carrier production method. 2. At what rate are the oxygen carrier particles "poisoned" by different solid fuels? 95 3. How does one go about separating the ash produced from the combustion of solid fuels from the oxygen carrier materials? 96 APPENDIX A MEASUREMENT AND MODELING OF DECOMPOSITION KINETICS FOR COPPER OXIDE-BASED CHEMICAL LOOPING WITH OXYGEN UNCOUPLING Reprinted with permission from Elsevier Limited. Clayton, C.K., Whitty, K.J. Measurement and modeling of decomposition kinetics for copper oxide-based chemical looping with oxygen uncoupling. Applied Energy 2013, 116: 116-423 97 98 99 100 101 102 103 104 105 APPENDIX B OXIDATION KINETICS OF CU2O IN OXYGEN CARRIERS FOR CHEMICAL LOOPING WITH OXYGEN UNCOUPLING Reprinted with permission from ACS Publications. Clayton, C.K., Sohn, H.Y., Whitty, K.J. Oxidation Kinetics of Cu2O in Oxygen Carriers for Chemical Looping with Oxygen Uncoupling. Industrial and Engineering Chemistry Research 2014, 53 (80): 2976-2986. 106 107 108 109 110 111 112 113 114 115 116 117 APPENDIX C CHARACTERISTICS AND CLOU PERFORMANCE OF A NOVEL SIO2-SUPPORTED OXYGEN CARRIER PREPARED FROM CUO AND β-SIC Reprinted with permission from ACS Publications. Peterson, S.B., Konya, G., Clayton, C.K., Lewis, R.J., Wilde, B.R., Eyring, E.M., Whitty, K.J. Characteristics and CLOU Performance of a Novel SiO2-Supported Oxygen Carrier Prepared from CuO and β-SiC. Energy and Fuels 2013, 27 (10): 6040-6047. 118 119 120 121 122 123 124 125 126 APPENDIX D ILMENITE AS AN INERT SUPPORT FOR COPPER-BASED OXYGEN CARRIER MATERIAL FOR USE IN CLOU 127 D.1 Abstract Among the critical aspects needing investigation for the implementation of chemical-looping into an industrial-scale reactor, oxygen carrier development remains at the forefront. A successful oxygen carrier will have a high carrying capacity, high redox rates and an optimal relationship between cost and physical durability. Both iron and copper have displayed desirable characteristics for a chemical-looping reactor. The high reactivity, oxygen carrying capacity and CLOU capabilities make copper a potentially suitable oxygen carrying material. When coupled with the superior physical durability of iron, a copper/iron bimetallic carrier proves very effective as an oxygen carrier material. Ilmenite is a low-cost iron-ore that has been proven an effective and inexpensive oxygen carrier material. The performance of ilmenite/copper bimetallic carrier in both TGA and fluidized-bed has been investigated in this work. To test the reactivity with solid fuels, the char of a bituminous coal (Illinois #6) has been tested along with the Cu-FeTiO2 carrier in a fluidized-bed. D.2 Introduction Chemical-looping with oxygen uncoupling (CLOU) is an energy generation technology suitable for use with solid fuels such as coal, petroleum coke and biomass, that offers inherent separation of CO2 with comparatively little energy penalty. CLOU involves cycling metal-based solid "oxygen carrier" particles between two fluidized-bed reactors. In the air reactor, the metal is oxidized by oxygen, resulting in an O2-depleted effluent stream. The oxidized carrier is transported to a fuel reactor fluidized by steam and/or CO2, where the thermodynamics of CLOU carriers are such that gaseous oxygen is 128 favored. The released O2 combusts the fuel, resulting in a product gas of mostly CO2 and steam. Condensation of the steam results in a nearly pure stream of CO2 suitable for sequestration with little additional processing required. CLOU is recognized as one of the most promising CO2 capture-ready technologies for energy production from coal, and research in this area has accelerated rapidly in recent years. A key consideration of CLOU technology is the oxygen carrier material. Several pure and mixed metals have been identified as having suitable thermodynamics for oxygen release in the fuel reactor [1,2,3,4]. Many investigations have focused on copper, cycling between oxidized cupric (CuO) and cuprous (Cu2O) oxide: Cu2O(s) + ½ O2(g) ↔ 2 CuO(s) (22) In thermogravimetric (TGA) experiments, CuO/Cu2O powder, on its own, can be oxidized to CuO in an air environment and reduced to Cu2O in an inert environment over many cycles [100]. However, in fluidized-bed conditions similar to those that would be used in a commercial CLOU system, pure CuO agglomerates, which makes it unsuitable for use in a fluidized system. Support materials, such as alumina, titania, zirconia, ceria and sepiolite, have been investigated to act as inert binders that allow the use of CuO in the fluidized conditions required for CLOU [1, 4, 6-11]. While the main concern that justifies the need of a support material is agglomeration, the support material also improves the kinetics of the oxidation and reduction reactions of the copper oxides [7]. A simple method for introducing CuO to a support material to prepare oxygen carriers for CLOU experiments is incipient wetness impregnation [8]. The incipient wetness method has provided supported CuO that performed well in TGA experiments. 129 However, de Diego et al. has reported that CuO on the external surface of the support material can be worn off in a fluidized environment, contributing to agglomeration of the material. This means that the CuO content is limited by the pore size of the support material for oxygen carriers prepared by incipient wetness. Compounds that are stable in both oxidizing and reducing environments at temperatures up to 1000°C and have a significant pore size are not readily available. Other methods for preparing supported copper particles, such as mechanical mixing [12], spray drying [13] and freeze granulation [1] have shown promise. But in some cases these methods have produced materials that lack the mechanical strength needed to be viable candidates for use in CLOU [12, 13]. A few studies have been performed which look into bimetallic oxygen carriers. A bimetallic carrier presents the developer with the opportunity to customize an oxygen carrier based upon the strengths of the individual components. For example, by creating an oxygen carrier based upon iron and copper, a material may be engineered with the structural integrity of iron while maintaining the CLOU capabilities of copper. Siriwardane et al. reported that the bimetallic combination of iron and copper as an oxygen carrier performed well while using a synthesis gas for reduction up to 900 °C [14]. 130 D.3 Experimental D.3.1 Preparation and Characterization of Oxygen Carriers D.3.1.1 Ilmenite Natural ilmenite powder (Atlantic Equipment Engineers) was sieved to a range of 150-250 microns and used as a support for the CuO. Adánez et al. [15] observed an increase in the rates of both oxidation and reduction of ilmenite after multiple cycles of oxidation in air and reduction in gas mixtures containing H2, CO or CH4. They observed that the increase in reaction rates occurs in only a few cycles, and referred to this phenomenon as an "activation period." Although the oxidation and reduction rates increased, a decrease in oxygen carrying capacity from 4 wt% to 2 wt% was observed after approximately 100 cycles. It is not clear, however, what mechanism(s) may be responsible for these changes. It is possible that during the first few cycles as the reaction rates increase that the redox conditions may be increasing a transport resistance by, for example, altering the micropore volume. Because the properties of ilmenite change during activation, it is important to investigate the deposition of CuO on both activated and nonactivated ilmenite. Some of the ilmenite material was thus activated by processing the raw material in a lab-scale fluidized-bed at 900°C while cycling between air (oxidation) and 5% CH4 in N2 (reduction) for 30 minutes each, for five complete cycles. D.3.1.2 Preparation Technique CuO was loaded onto the ilmenite by impregnation via rotary evaporation. A copper nitrate solution of about 2 M was prepared by dissolving copper(II) nitrate tri- 131 hydrate (Sigma-Aldrich) in acetone. The ilmenite powder and copper nitrate solution were mixed in a flask using a rotary evaporator. The materials were mixed and the acetone evaporated at an external temperature of approximately 50 °C to promote the quick evaporation of the solvent while avoiding bumping. Once the acetone had evaporated, a heating mantle was used to increase the temperature at a rate of approximately 25°C/min to a final external temperature of about 375°C. Once all of the material in the flask had turned black and the decomposition of copper(II) nitrate to CuO was complete, the flask was removed and allowed to cool. After the flask had cooled, the next addition of CuO took place. For each sample, copper(II) nitrate was added to the ilmenite in multiple additions to reach the target CuO content. A summary of the oxygen carriers produced in this study is reported in Table 15. D.3.1.3 Thermogravimetric Analysis Each oxygen carrier material was analyzed in a thermogravimetric analyzer (TGA) to determine its reactivity in multicycle tests. The TGA experiments were conducted using a TA Q500 instrument. For each experiment, the sample with a mass between 30-40 mg was heated to the target temperature in 100 mL/min of air. Nitrogen gas, at a rate of 100 mL/min, was then used to promote spontaneous reduction of the sample. The cycle was then completed as the sample was re-oxidized in air. Fractional completion, X, describes reaction progression. Fractional completion is defined in terms of mass at a given time (mt), mass of the reduced (Cu2O) sample (mred) and mass of fully oxidized (CuO) material (mox). Fractional completion for oxidation, Xox, and reduction, Xred, are defined as follows: 132 Table 15. Summary of oxygen carrier materials prepared. The name indicates if the ilmenite was activated prior to CuO deposition (ACT) or if nonactivated ilmenite was used (NA) and the number indicated the % CuO by weight of the material. CuO Loading Number of CuO BET Surface Area Particle Size (wt%) Additions (m2/g) (µm) ACT-20 20 6 0.299 150 - 250 NA-20 20 6 4.2 150 - 250 NA-30 30 9 2.3 150 - 250 Code 1 All agglomeration was soft - particles were easily separated with mild physical agitation. mt − mred mox − mred m − mox = t mred − mox X ox = (23) X red (24) D.3.2 Fluidized-Bed Performance Performance of the carriers in a fluidized-bed environment was evaluated in a labscale fluidized-bed system (FZB, Figure 38), which is modeled on a system developed at Chalmers University. The reactor is made of quartz and is housed within a Carbolite VST 12/600 Clamshell furnace with a maximum operating temperature of 1200°C. The reactor has four zones (Figure 39): (1) inlet, (2) sintered quartz distributor supporting the particle bed, (3) freeboard expansion zone and (4) outlet. The ID of the reaction tube is 2.5 cm. In order to maintain a well-fluidized system, the typical resting bed height was about 2.5 cm (or 1 diameter tall). Gas flow is controlled by a series of mass flow controllers and is introduced to the reactor housed within the furnace. The exit gas is filtered to remove fines and water is condensed from the gas before it is sent to a 4channel California Analytical Instruments ZRE NDIR/fuel call analyzer. The analyzer 133 Figure 38: Schematic of the bench-scale fluidized-bed reactor system used at the University of Utah. measures the volume percents of CO, CO2, CH4 and O2. A dry gas meter measures the total volume of gas fed through the system. For typical operation, the reactor was charged with 20-30 g of oxygen carrier particles and heated at 7 °C/min to 850°C. During heating, the bed was fluidized with nitrogen to mitigate agglomeration. Upon reaching 850°C, the fluidizing gas was cycled between nitrogen and air for 2 hours to encourage thermally driven migration of the copper oxide below the surface of the carrier. During cycling, each gas was fed for 30 minutes before switching to the other. The reactor was then heated further to 900°C and gases cycled again for 2 hours. Finally, the reactor temperature was raised to 950°C and 3 air/nitrogen cycles were performed. Performance of the CuO/ilmenite carrier during CLOU combustion of coal was tested by introducing small batches of coal into the reactor. After the carrier underwent the initial cycling at progressively higher temperatures as described above, the temperature was adjusted to that for the coal experiments. The carrier was oxidized in air, after which excess air was removed by passing a very small flow of nitrogen (1 ml/min) 134 Solid Fuel Feeder 2.5 cm 15 cm 18 cm 6 cm 3.8 cm 18 cm 6 cm 2.5 cm Particle Bed 28 cm Quartz Frit or Distributor 2.5 cm 20 cm Figure 39: Schematic of quartz fluidized-bed reactor. The particle bed sets upon a porous quartz distributor with a nominal maximum pore size of 16 - 140 microns. through the reactor for approximately 2 minutes. The gas flow was then shut off to reduce blow-out of the fuel while it was added from the top of the reactor (Figure 39). After the addition of the solid fuels the flow of nitrogen was increased to the standard flow of 1 SLPM and the coal was allowed to react with oxygen released from the carrier particles. The evolution of gases was monitored by the analyzer and gas meter. The fuel was allowed to burn out evidenced by the increase in oxygen partial pressure and 135 decrease in CO2 partial pressure. Upon completion, the reacting gas was switched back to air to regenerate the carrier. Coal processing was performed at two temperatures, 900 and 950°C. The fuel used for testing was char made from Illinois #6, an eastern U.S. bituminous coal. The ultimate and proximate analyses of the coal are provided in Table 16. The coal was dried at 125°C for 3 hours to remove the moisture and then heated to 1000°C under nitrogen for 8 hours for devolatilization. The fuel was sieved to a particle size range of 150-250 microns. The amount of fuel used in the tests was calculated so that the carbon in the fuel could be completely oxidized by the theoretical amount of oxygen released from the oxygen carrier sample in the reactor, while allowing for 10% excess oxygen. Conversion of the fuel is calculated based on the conversion of carbon to carbonaceous gases - assuming the char has been stripped of all volatile matter. For the combustion of solid fuels, it is requisite to perform a carbon balance analysis to evaluate conversion of the fuel and to understand effective combustion rates. The carbon balance is performed using the flow rate of the purge gas as a baseline. It is assumed that the gases recorded by the analyzer, along with the inert purge, combine to make the majority of the gases evolved from fuel combustion. By assuming that any other species evolved from the combustion are in quantities small enough to have a negligible effect on the total flow rate of the effluent gases, the molar flow rate of the analyzed species may be determined. The total volumetric flow rate may be determined by calculating the volume fraction of the inert purge gas (ϕpurge ), assuming that is the difference between one and the sum the volume fractions (ϕi ) of CO2, CO, CH4 and O2. 136 Table 16: Illinois #6 coal proximate, ultimate and heating value analysis. Proximate Analysis Moisture (wt% as received fuel) 2.54 Ash (wt% Dry) 12.33 Volatile matter (wt% dry) 39.40 Fixed carbon (wt% dry) 48.28 Ultimate Analysis (wt% dry ash-free) Carbon 78.91 Hydrogen 5.50 Nitrogen 1.38 Sulfur 4.00 Oxygen 10.09 Chlorine 0.11 Heating Value HHV, dry (Btu/lb) 12,233 137 ϕpurge = 1 − ϕi ; i = CO2 , CO, CH4 , O2 (25) The volume fraction of the purge gas (N2) is then used to calculate the total effluent volumetric flow rate of gases (Qtot ). Qtot = ϕN 2 Q N2 (26) The total volumetric flow rate multiplied by the individual specie volume fractions gives the volumetric flow rate of each of the individual gaseous species. The ideal gas equation is then employed to determine the molar flow rate of each component (equation 28). 𝑄𝑖 = 𝜙𝑖 × 𝑄𝑡𝑡𝑡 ; 𝑖 = 𝐶𝐶2 , 𝐶𝐶, 𝐶𝐶4 , 𝑂2 𝑛𝑖 = 𝑄𝑖 𝑃 𝑅𝑅 (27) (28) The carbon balance is then performed by adding the total number of moles of carbonaceous gases and then multiplying by the molecular weight of carbon - 12 g/mole. The resulting mass of carbon is compared to the mass of the sample introduced to the reactor generating a conversion profile based on the conversion of carbon. 138 D.4 Results and Discussion D.4.1 Characteristics of the Oxygen Carrier Particles D.4.1.1 Microscopic Imaging The investigated materials were imaged by a Hitachi S-300N scanning electron microscope (SEM) to permit visualization of the surface on the scale of 10 μm to 1 mm. Energy dispersive X-ray spectroscopy (EDS) was carried out with an EDAX HIT S3000N 132-10 alongside the SEM images to confirm the appearance of and locate elemental titanium, iron, and copper. SEM images of ilmenite in both the fresh and activated states are included. Carrier NA-20 was observed by SEM immediately after being produced and after undergoing five oxidation and reduction cycles in TGA experiments. As seen in Figure 40, the surface of the nonactivated ilmenite is smooth while the surface of the activated ilmenite has acquired a rough surface. The surface roughness may be a consequence of carbon deposition during the reducing phase of the activation process, since methane may have been catalytically cracked by the ilmenite. EDS elemental mapping was used to identify the distribution of copper on the samples and to determine how TGA experiments affected carrier structure. Figure 40 images C and D show that the TGA experiments with air/N2 cycling change the makeup of the particles. After TGA experiments, the EDS shows that little Cu is present on the surface in contrast to those particles which had not been used in TGA experiments. The particles that were exposed to several redox cycles in TGA experiments display a sponge-like surface. This could be the result of CuO migration due to the multiple reduction and oxidation cycles, but the EDS maps suggest that that is not the case; instead, the shown that activation of the ilmenite can alter the surface properties of ilmenite, it is possible 139 A A-Cu A-Ti B-Cu B-Ti C-Cu C-Ti D-Cu D-Ti 25μm B 10μm C 200μm D 200μm Figure 40: SEM images of A) activated ilmenite, B) nonactivated ilmenite, C) ACT20 before TGA experiment, and D) ACT20 after TGA experiment. C and D include EDS elemental mapping for elemental Cu and Ti. that the addition of CuO and the multiple oxidation and reduction cycle have caused the surface of the ilmenite to be altered and engulf the CuO. The sponge-like surface could be the channels through which the oxygen travels to and from the particle. D.4.1.2 Distribution of Cu, Fe and Ti Within Single Particles Freshly prepared carriers NA-20 and ACT-20 were analyzed by laser ablation ICP-MS to determine composition across a cross-section of the material. Particles were fixed in an epoxy and allowed to set. The epoxy was then sliced and polished to expose cross-sections of the particles. As the 15μm wide laser scanned in a line across the material, the material was vaporized and the gas was analyzed by mass spectroscopy. The analytical technique allows determination of what elements are present to 140 within15μm. For carrier ACT-20, Cu was present only very near the surface of the particle (Figure 41). In the inside of the particle, essentially no Cu was detected. In contrast, the results for carrier NA-20 (Figure 42) show a distribution of Cu much less concentrated at the particle surface and appears to have penetrated to the center of the particle, albeit at low concentrations. This indicates that the activation of the ilmenite with the procedure used in this study does not improve penetration or evenness of copper distribution in the particle. Thus, preprocessing of the ilmenite can affect final carrier properties. Further investigation would be needed to determine the ideal method of preparation of the oxygen carrier material and the ideal particle size to maximize the amount of CuO present and the reactivity of the material. ACT-20 700 Ti Fe Cu 600 Concentration (ppt) 500 400 300 200 100 0 0 50 100 150 200 Distance (µm) Figure 41: Laser ablation results for carrier ACT-20, produced using activated ilmenite. The elemental concentrations are shown in parts per thousand (ppt). 141 NA-20 Concentration (ppt) 300 Ti Fe Cu 200 100 0 0 50 100 150 Distance (µm) Figure 42: Laser ablation results for carrier NA-20. The elemental concentrations are shown in parts per thousand (ppt). D.4.1.3 Oxygen Carrying Capacity Multicycle TGA experiments were performed in which the oxygen carriers were cycled between the reduced and oxidized forms for over 10 cycles (Figure 43). The measured mass increase during oxidation indicated that the copper loading was very close to the target loading. No significant change in mass between cycles was observed, indicating that oxygen carrying capacity of the copper could be sustained, and that no interaction between the ilmenite and copper reduced carrying capacity. It was observed that completion of the reduction cycle slowed somewhat after several cycles, such that complete reduction was not achieved (as evidenced by a slight downwards slope of the mass curve) even after 30 minutes. 142 Figure 43: Mass of carrier NA-30 over 10 oxidation/reduction cycles in the TGA. D.4.1.4 Rates of Reduction and Oxidation Figure 44 shows the rates of carrier reduction (O2 release) in nitrogen at 850, 900 and 950°C. At 900 and 950°C, more than 75% conversion was achieved in less than 2 minutes, but the rate slowed significantly after that, so that 90% conversion was not achieved until approximately 10 minutes. This may be a result of difficulty accessing copper in the core of the particle (Figure 42). The rate at 50% conversion was determined from the slope of the mass loss curve and the activation energy for reduction was calculated to be 240 kJ/mol (Figure 45). The time required for oxidation in the TGA experiments was found to decrease as the temperature was increased from 850 to 950°C (Figure 46). This is consistent with what has been observed in other studies of oxidation of Cu2O to CuO, and is due primarily to a decrease in driving force of the oxidation reaction as the equilibrium O2 partial pressure approaches the partial pressure of the fed gas environment at higher 143 Reduction Completion (%) 100 50 850°C 900°C 950°C 0 0 8 16 Time (min) Figure 44: Reduction in TGA of NA-30 measured at 850, 900 and 950°C. copper-based CLOU carriers have found. -4 ln (k) (s-1) -5 -6 -7 8.2 8.4 8.6 4 8.8 -1 10 /T (K ) Figure 45: Arrhenius plot for reduction of NA-30 carrier. 9.0 144 Oxidation Completion (%) 100 50 850°C 900°C 950°C 0 0 5 10 Time (min) Figure 46: Oxidation in TGA of NA-30 measured at 850, 900, and 950°C. temperatures. D.4.2 Fluidized-bed Performance D.4.2.1 Cycling Between Air and N2 Each of the materials was cycled between air and nitrogen at temperatures from 850 to 950°C. The clamshell design of the furnace housing allowed observation of the bed, so agglomeration could be observed visually. The carriers were able to be successfully heated to 850°C without significant agglomeration. At temperatures above 850°C, the materials began to show signs of agglomeration. At temperatures above 950°C, catastrophic agglomeration was observed. In order to mitigate agglomeration, the temperature ramp during heating was 145 slowed from 10°C per minute to 5°C per minute. This change raised the agglomeration temperature above 850°C, suggesting that more gentle heating allows the material to stabilize, possibly avoiding high concentrations of agglomeration-prone copper at the particle surface. During heating, it was also observed that the bed must be maintained under full fluidization conditions or agglomeration may begin to occur. Regions with limited gas flow are most prone to agglomeration in fluidized-bed systems. While good fluidization was maintained throughout testing, it was observed that stagnant zones would begin to collect defluidized particles. These zones included the sloped region of the free board expansion zone as well as zones near the wall at the base of the particle bed closest to the gas distributor plate. The particles were not strongly stuck together; simple agitation of the particle bed would be enough to break it free. Expected operating temperatures of a copper-based chemical-looping system are between 900 and 1000°C, so it is important to test oxygen carrier materials above 900 °C. Because of the agglomeration propensity observed at temperatures above 850°C, testing of the CuO-ilmenite carriers above that temperature was conducted carefully, with slow heating and periodic redox cycling. The material was heated slowly to 850 °C and then cycled between CuO and Cu2O under alternating atmospheres of air and nitrogen for 1 hour each cycle for 4 cycles. After the 4 cycles were completed, the temperature was increased another 50°C and the cycling was repeated. The temperature was then set to the desired operating temperature. This process allowed the agglomeration temperature for the NA-30 material to be increased from 850°C to greater than 950°C. This increase in agglomeration temperature may be explained using the laser ablation ICP findings displayed in Figure 41. For the materials that were exposed to elevated temperatures, the 146 copper oxide on the particle surface began to migrate from the surface toward the inside of the particles. A similar technique is used within the semiconductor industry, where thermally driven diffusion of impregnating dopants is performed regularly in order to alter the conductive properties of the silicon. This thermal treatment of the material was most effective during gas cycling. To verify this, a sample of the NA-30 material was subjected to baking at 950°C under air for 36 hours in a muffle furnace without being subjected to fluidization conditions. After baking, the material was agglomerated, but was easily broken apart by simple agitation, which indicates sintering was not significant and that fluidization conditions may keep the particles from agglomerating. This product did not perform any better than nonbaked material under fluidization and agglomerated in the fluidized-bed before reaching 950°C. Therefore, it is believed that cycling between oxidizing and inert conditions may provide a better means for the thermally driven diffusion of the copper oxide molecules. D.4.2.2 With Solid Fuel Addition The evolution of gases during the combustion of the char prepared from the coal represented in Table 16 is presented in Figure 47. The figure shows three regions for the reaction: initially, air is used to oxidize the oxygen carrier. The gas is switched to nitrogen at about 50 seconds and after a period of about 50 more seconds, the oxygen is depleted and levels off around 2% by volume, which corresponds to the equilibrium concentration at the reactor temperature. Once the oxygen vol% levels off, the purge gas is turned off and the fuel is introduced, which may be seen in Figure 47 as the spike in the carbonaceous gases and the drop in the oxygen level. Interestingly, the oxygen does not 147 CO CH4 and CO Vol % 0.7 CH4 0.6 25 20 CO2 0.5 O2 0.4 15 10 0.3 0.2 CO2 and O2 Vol% 0.8 5 0.1 0 0 0 200 400 Time (Seconds) Figure 47: Evolution of effluent gases during combustion of Illinois #6 Char using the NA30 material at 900 °C. drop to zero and continues to evolve from the surface of the oxygen carrier particles faster than it can be consumed. Figure 48 and Figure 49 display the conversion of the fuel versus time. As may be seen in Figure 49, the rate of consumption of the fuel is much faster at 950°C than it is at 900°C, and the conversion of the solid fuel would be attained in approximately 1 hour. This, however, is unlikely. Due to the operating temperature, the decomposition of the CuO should occur in less than a few minutes. It is quickly seen that the oxygen carrier has completely released all of its oxygen from Figure 49. The rate of conversion of the fuel changes drastically around 30 seconds. Initially, the fuel burns quickly, but seems to slow around 30 seconds. This is most likely due to the particles being completely converted from CuO to Cu2O long 148 0.6 Mass Carbon Fuel Conversion 0.002 0.5 0.4 0.0015 0.3 0.001 0.2 0.0005 Fuel Conversion Mass Carbon Gasses (g/s) 0.0025 0.1 0 0 0 100 Time (Seconds) 200 Figure 48: Conversion of the fuel vs. time with the evolution of carbon (as gas) from the combustion of Illinois #6 Char at 900 °C. 0.45 0.4 Fuel Conversion 0.35 0.3 0.25 0.2 0.15 0.1 900 C 0.05 950 C 0 0 200 Time (Seconds) 400 Figure 49: Conversion of Illinois #6 char at two temperatures using NA30 material under CLOU conditions. 149 before the fuel is converted. This may be caused from poor mixing in the fluidized-bed between the oxygen carriers and the fuel particles or simply the fuel not being very highly reactive. This may be accounted for by using a more reactive fuel, decreasing fuel particle size, increasing mixing from fluidization or even increasing the oxygen carrier reactor charge (thereby increasing the bed depth). By increasing the bed depth, the release of oxygen from the particles in the bottom of the bed essentially inhibits the decomposition of the particles downstream by decreasing the oxygen partial pressure driving force. This inhibition would slow the release of oxygen allowing for a slower burning fuel. In Figure 50, it may be seen that the production ratio CO2/CO is roughly twice as large at 950°C than at 900°C. The largest difference between these two temperatures is seen around 25 seconds and is possibly an effect of residual volatiles not removed during the charring process. The char is produced at relatively low heating rates; therefore, the volatile compounds evolve from the fuel. Therefore, the volatiles are better consumed at the higher temperatures likely due to the increased equilibrium partial pressure of oxygen at the higher temperatures which increases the rate at which the oxygen is liberated (Figure 44). The faster CuO decomposition rate provides more oxygen in the chamber to react with the low residence time volatiles. D.5 Conclusions The development of a novel oxygen carrier by the mixture of ilmenite and copper oxide is presented. The performance of the oxygen carrier in both a fluidized-bed and TGA is also presented. The bimetallic Cu-Fe carrier made from ilmenite and copper 150 60 950 C 900 C Ratio of moles CO2/CO 50 40 30 20 10 0 0 50 Time (Seconds) 100 Figure 50: Ratio of moles of CO2 to moles of CO evolved during the combustion of Illinois #6 char at two temperatures using NA30 material as oxygen carrier. nitrate trihydrate performed well a nitrogen atmosphere within TGA and while using a char made from the devolatilization of a bituminous coal. The increase in temperature from 900 °C to 950 °C significantly increased the ratio of CO2/CO production during combustion of the char. The material developed successfully operated at 950 °C without significant agglomeration using the described preparation method. D.6 Acknowledgements This material is based upon work supported by the Department of Energy under Award Number DE-NT0005015. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. The authors express thanks to Dana Overacker and Crystal Allen of the Department of Chemical Engineering at the University of Utah for their assistance with the laboratory experiments. 151 D.7 References [1] Mattisson, T., Lyngfelt, A., Leion, H. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int J Greenh Gas Con. 2009, 3 (1): 11-19. [2] Azimi, G., Rydén, M., Leion, H., Mattisson, T., Lyngfelt, A. (Mnz Fe1-z)yOx combined oxides as oxygen carrier for chemical-looping with oxygen uncoupling. AIChE J. 2013, 59(2): 582-588. doi:10.1002/aic.13847. [3] Arjmand, M., Hedayati, A., Azad, A.-M., Leion, H., Rydén, M., & Mattisson, T. Ca(x)La(1-x)Mn(1-y)M(y)O(3−δ) (M=Mg, Ti, Fe, or Cu) as Oxygen Carriers for Chemical-Looping with Oxygen Uncoupling (CLOU). Energ & Fuel. 2013, 27 (8): 4907 - 4107. doi:10.1021/ef3020102 [4] Mattisson, T. Materials for Chemical-Looping with Oxygen Uncoupling. ISRN Chem Eng. 2013(1): 1-19. doi:10.1155/2013/526375 [5] Baraki, R., Konya, G., Eyring, E. M. Investigation of the Reaction Kinetics of Oxygen Carriers in Chemical Looping Combustion. Twenty-Sixth Annual International Pittsburgh Coal Conference, Pittsburgh, PA, 2009. [6] Arjmand, M., Azad, A.-M., Leion, H., Lyngfelt, A., Mattisson, T. Prospects of Al2O3 and MgAl2O4-Supported CuO Oxygen Carriers in Chemical-Looping Combustion (CLC) and Chemical-Looping with Oxygen Uncoupling (CLOU). Energ Fuel. 2011, 25 (11): 5493-5502. [7] Wang, S., Luo, M., Wang, G., Wang, L., Lv, M. Analysis of Reactivity of a CuOBased Oxygen Carrier for Chemical Looping Combustion of Coal. Energ Fuel. 2012, (26): 3275-3283. [8] de Diego, L. F.; Garcia-Labiano, F.; Gayan, P.; Celaya, J.; Palacios, J. M.; Adanez, Operation of a 10-kW Chemical-Looping Combustor during 200 h with a CuOAl2O3 Oxygen carrier. Fuel. 2007, 86 (7-8): 1036-1045. [9] Johansson, M.; Mattisson, T.; Lyngfelt, A. Investigation of Mn3O4 With Stabilized ZrO2 for Chemical-Looping Combustion. Chem Eng Res Des. 2006, 84 (9): 807818. [10] Adanez-Rubio, I.; Gayan, P.; Garcia-Labiano, F.; de Diego, L. F.; Adanez, J.; Abad, A. Development of CuO-based oxygen-carrier materials suitable for ChemicalLooping with Oxygen Uncoupling (CLOU). Energ Proc. 2011, (4): 417-424. [11] Adanez-Rubio, I., Arjmand, M., Leion, H., Gayan, P., Abad, A., Mattisson, T. P., & Lyngfelt, A. (2013). Investigation of Combined Supports for Cu-based Oxygen Carriers for Chemical-Looping with Oxygen Uncoupling (CLOU). Energ Fuel. 2013, 27 (7): 3918 - 3927. doi:10.1021/ef401161s. [12] Gayan, P.; Adanez-Rubio, I.; Abad, A.; de Diego, L. F.; Garcia-Labiano, F.; 152 Adanez, J. Fuel 2012, 96, 226-238. [13] Adanez-Rubio, I.; Gayan, P.; Abad, A.; de Diego, L. F.; Garcia-Labiano, F.; Adanez, J. Energy Fuels 2012, 26 (5), 3069-3081. [14] Siriwardane, R. V., Ksepko, E., Tian, H., Poston, J., Simonyi, T., Sciazko, M. Interaction of Iron-Copper Mixed Metal Oxide Oxygen Carriers with Simulated Synthesis Gas Derived from Steam Gasification of Coal. Applied Energy, 2013, 107 (C): 111-123. [15] Adánez, J., Cuadrat, A., Abad, A., Gayán, P., de Diego, L. F., Garcia-Labiano, F. Ilmenite during Consecutive Redox Cycles in Chemical-Looping Combustion. , 2010 , 24 (2): 1402-1413.