ABSTRACT. The chemical looping combustion (CLC) process uses metal oxides, also referred to as oxygen

Transcription

1 ABSTRACT GALINSKY, NATHAN LEE. Investigation of Redox Metal Oxides for Carbonaceous Fuel Conversion and CO2 Capture. (Under the direction of Dr. Fanxing Li.) The chemical looping combustion (CLC) process uses metal oxides, also referred to as oxygen carriers, in a redox scheme for conversion of carbonaceous fuels into a concentrated stream of CO2 and steam while also producing heat and electricity. The unique redox scheme of CLC allows CO2 capture with minimal energy penalty. The CLC process performance greatly depends on the oxygen carrier that is chosen. To date, more than 1000 oxygen carriers have been developed for chemical-looping processes using metal oxides containing first-row transition metals. Oxygen carriers are typically mixed with an inert ceramic support to improve their overall mechanical stability and recyclability. This study focuses on design of (i) iron oxide oxygen carriers for conversion of gaseous carbonaceous fuels and (ii) development of perovskite CaMnO3-δ with improved stability and redox properties for conversion of solid fuels. Iron oxide is cheap and environmentally benign. However, it suffers from low activity with carbonaceous fuels due partially to the low ionic conductivity of iron oxides. In order to address the low activity of iron-oxide-based oxygen carriers, support addition has been shown to lower the energy barrier of oxygen anion transport within the oxygen carrier. This work adds a mixedionic-and-electronic-conductor (MIEC) support to iron oxide to help facilitate O 2- transport inside the lattice of iron oxide. The MIEC-supported iron oxide is compared to commonly used supports including TiO2 and Al2O3 and the pure ionic conductor support yttria-stabilized zirconia (YSZ) for conversion of different carbonaceous fuels and hydrogen. Results show that the MIEC-supported iron oxide exhibits up to 70 times higher activity than non-miec-

2 supported iron oxides for methane conversion. The MIEC supported iron oxide also shows good recyclability with only minor agglomeration and carbon formation observed. The effect of support-iron oxide synergies is further investigated to understand other physical and chemical properties that lead to highly active and recyclable oxygen carriers. Perovskite and fluorite-structured MIEC supports are tested for conversion of methane. The perovskite supported iron oxides exhibit higher activity and stability resulting from the high mixed conductivity of the support. Fluorite-structured CeO2 oxygen carriers deactivated by 75% after 10 redox cycles. This deactivation was attributed to agglomeration of iron oxide. The agglomeration was determined to occur due to Fe x+ transport during the oxidation step leading to high content of Fe on the surface of the oxygen carrier. Besides the MIEC supports, inert MgAl2O4 supported iron oxide is observed to activate in methane. The activation is attributed to carbon formation causing physical degradation of the oxygen carrier and leading to higher surface area and porosity. To achieve high activity with solid fuels, chemical looping with oxygen uncoupling (CLOU) is commonly used. This process uses oxygen carriers with high PO2 that allows the oxygen carrier to release a portion of their lattice oxygen as gaseous oxygen. In turn, the gaseous oxygen can react with solid fuel particles at a higher rate than the lattice oxygen. CaMnO3 perovskite oxygen carriers offer high potential for CLOU. However, pure CaMnO3 suffers from long-term recyclability and sulfur poisoning. Addition of A-site (Ba and Sr) and B-site (Fe, Ni, Co, Al, and V) dopants are used to improve the performance of the base CaMnO3 oxygen carrier. Sr (A-site) and Fe (B-site) exhibit high compatibility with the base perovskite structure. Both dopants observe oxygen uncoupling properties up to 200 C below that of pure CaMnO3. Additionally, the doped structures also exhibit higher stability at high temperatures

3 (>1000 C) and during redox cycles. The doped oxygen carriers also demonstrate significantly improved activity for coal char conversion.

4 Copyright 2016 by Nathan Lee Galinsky All Rights Reserved

5 Investigation of Redox Metal Oxides for Carbonaceous Fuel Conversion and CO2 Capture by Nathan Lee Galinsky A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy Chemical Engineering Raleigh, North Carolina 2016 APPROVED BY: Dr. Fanxing Li Committee Chair Dr. Phillip Westmoreland Dr. Erik Santiso Dr. Sunkyu Park

6 DEDICATION To my parents, Martin and Robin Galinsky, for your love and sacrifices, to my friends who supported me through tough times, and to my professors who inspire me continuously. ii

7 BIOGRAPHY Nathan Galinsky was born in Charleston, West Virginia on July 28, Upon graduation from high school in May 2007, he accepted to attend college at West Virginia University Institute of Technology for Chemical Engineering. Upon entering the end of his freshman year, he was offered a research position to work on various renewable energy technologies including biofuels and solar panels. Following that research experience, he joined a collaboration project between the chemical and electrical engineering departments research on direct carbon fuel cells. During his research experiences in undergrad, Nathan gained invaluable experience and passion for research pushing him towards a career in graduate school. Nathan was accepted to attend North Carolina State University in the department of chemical and biomolecular engineering in 2011 in pursuit of his PhD. In January 2012, Nathan joined the research group of Dr. Fanxing Li. During his PhD, Nathan grew a passion for teaching and took the opportunity to assist in teaching classes of thermodynamics and reactor design. Following the completion of his PhD, Nathan will be working through the Postgraduate Research Program at National Energy Technology Laboratory in Morgantown, West Virginia. iii

8 ACKNOWLEDGEMENTS First, I want to express my sincere gratitude to my advisor Dr. Li. His unsurpassed knowledge and encouragement during my graduate school career is greatly appreciative and inspirational. His willingness to support me in my personal endeavors as well as my research did not go unnoticed. His desire to see his students succeed was demonstrated by his actions, and I am thankful to be in the position I am through his help. In addition to Dr. Li, I want to thank my committee Dr. Phillip Westmoreland, Dr. Erik Santiso, Dr. Wesley Henderson, and Dr. Sunkyu Park for your suggestions, patience, and inspiration. I want to thank Dr. Santiso for being able to participate in my oral final defense on short notice. I want to thank my family for their truly sacrificing love and support during my life. My parents, Robin and Martin Galinsky, sacrificed a lot, so that I may succeed. I appreciate your sacrifices and am the man I am through a lot of your actions of support. My sister, Shaina, has helped me tremendously through transitioning from undergrad to graduate school. I am extremely proud of her. I have several individuals who have impacted my life through one way or another. My closest friend during my undergrad, Rebecca Minnick, shared the struggles, laughs, and good times of undergraduate chemical engineering. The late night talks during difficult times, long all night homework sessions, and fun outings we had made undergraduate a blast. The moment in meeting you, it was apparent that you were someone to have great impact on my life, and I will be eternally grateful for the time and hardships we shared. I want to thank Jared Wright, who was my roommate during undergraduate school. Jared always brought a light to the room and smile to my face. He was a good support and friend that I cherished during my time as his roommate. Michelle Casto was an amazing person to talk to during all times. She knew a way iv

9 to put a smile on my face despite any situation. All the people I got to meet through CSU, you all were the best support group and helped me grow in my faith that could be sent out into the world. Following my acceptance into NCSU and joining Dr. Li s group, I have got to meet and work with some of the best, most intelligent people. Feng He was the funniest guy I have had the pleasure to meet. He also laid the ground work to getting out lab set-up and being able to get into the lab right away to perform experiments. His time was appreciated, and our conversations are not something I will ever forget. Arya Shafiefarhood joined the group at the same time, so we got to share a lot of the struggles and difficulties of joining a group at the same time. Arya and I worked on similar projects to get accustomed to the lab, so we spent a lot of time in discussions, and I always appreciate his insight into our research. I also appreciate your help and support whenever I needed something from a ride to help in a lab, you truly are inspiring through your sacrifices to help others. Besides Arya, Donny Bhuana joined our group at the same time which allowed for a group of us to experience the challenges together. Amit Mishra was the next student to join our group. His project overlapped with a portion of research I was working on conducting, so we started by working together in the lab. Amit is extremely intelligent, and his help has been invaluable through the time he has been in the group. Yunfei Gao and Vasudev Haribal have been a pleasure to work with and I wish them the best in their research pursuits. I want to thank Seif Yusuf for his work and help in managing the labs after joining the group. I wish the best to newer members Ryan Dudek and Michael Gerber, I wish I had more time to get know you guys better. I also want to thank Dr. Luke Neal, Dr. Junshe Zhang, Dr. Yanguang Chen, and Dr. Yan Huang for their help, suggestions, and guidance during my research. v

10 I have gotten to also meet some incredible people here at state. First I want to mention my roommate, David Rutkowski, for his effort and support during my time. David was always willing to help me out if I needed a ride or had an emergency. Besides David, I want to thank Mark Schulte, Steve Benner, and Kenny Mineart for our homework struggles during the first year as well as our random food outings. They were some of the most fun moments during that year. I also want to thank the rest of the incoming class of 2011 that I got to meet some incredible people. I want to thank my friends Jackson Wooten, Austin Ocheltree, Alex Mull, Christopher Nichols, Elizabeth Oliveros, and Cameron Howell for their support, care, and faith. You are all inspirations and took the time to make me feel at home in a new city. You guys were selfless in your support to me over my time here. I appreciate you all very much for your encouragement and help through my faith. I want to give a big thank you to Amy Stewart. She was initially an undergraduate student who I met through as a mentor for her research endeavors. Amy was incredible in her time she spent in the lab. She was easy to work with and mentor. But, her biggest contribution to me was through her friendship. Amy was a highlight of my time at state. I can t express enough sincere gratitude through our long talks that you helped me with so much personal matter. I know that Amy is one of the few people I can count on in any situation. I also want to thank Marwan Sendi. He was an incredible undergraduate student whose work ethics are second to none. He was always on time and was able to grasp different aspects of the laboratory very quickly. He was able to contribute unbelievable amounts to my work. Additionally, Evan Scott, Ziren Wang, Shivalik Daga, Lindsey Bowers, Tabitha Cabral, and Clay Hamill contributed to building the lab and working with me. vi

11 In addition to my research duties, I have had the pleasure of teaching undergraduate students. I am thankful of the opportunities of working with Dr. Carol Hall, Dr. Balaji Rao, and Dr. Li as teaching assistants. The invaluable lessons I have learned through teaching and the passion I have seen from being able to teach to undergraduates bring an extreme delight. Besides the professors for their inspiration and encouragement for pursuing this passion, I want to thank the students I have gotten to interact with over the time I have been at NCSU. I hope I was able to do this school justice and provide you educational opportunities that I also had. I hope that my passion was able to shine through. And finally, I hope that you all know that through your hard work and dedication that you inspired me to keep teaching. I want to thank every individual I have had to meet or interact in some way. Every situation and person helps define me to this day and I could not be where I am without all the people in my life. I want to end with saying that without my faith in Jesus, I would be far from where I am today. I thank Him for supporting me in all my endeavors good or bad. vii

12 TABLE OF CONTENTS LIST OF TABLES... xiii LIST OF FIGURES... xv CHAPTER 1 Introduction Carbon Emissions Carbon Capture Chemical Looping Processes Chemical Looping Combustion (CLC) Chemical Looping with In-Situ Gasification Chemical Looping with Oxygen Uncoupling (CLOU) Oxygen Carriers Fe-Based Oxygen Carriers Ni-Based Oxygen Carriers Cu-Based Oxygen Carriers Mn-Based Oxygen Carriers Co-Based Oxygen Carriers Mixed-Metal Based Oxygen Carriers Summary REFERENCES CHAPTER 2 Iron Oxide with Facilitated O 2- Transport for Facile Fuel Oxidation and CO2 Capture in a Chemical Looping Scheme viii

13 Abstract Introduction Experimental Oxygen Carrier Synthesis Reactivity Studies Sample Characterizations Results and Discussion Reactivity with Hydrogen Long Term Stability of the LSF Supported Oxygen Carrier Reaction with Gaseous Carbonaceous Fuels Morphology Studies and Reaction Scheme Conclusions REFERENCES CHAPTER 3 Effect of Support on Redox Stability of Iron Oxide for Chemical Looping Conversion of Methane Abstract Introduction Experimental Methods Oxygen Carrier Synthesis Reactivity Studies Sample Characterizations ix

14 3.3 Results Activity Comparisons Ca0.8Sr0.2Ti0.8Ni0.2O3 Supported Fe2O CeO2 Supported Fr2O MgAl2O4 Supported Fe2O Conclusion REFERENCES CHAPTER 4 Ca1-xAxMnO3 (A= Sr and Ba) Perovskite Based Oxygen Carriers for Chemical Looping with Oxygen Uncoupling (CLOU) Abstract Introduction Experimental Oxygen Carrier Synthesis Oxygen Uncoupling Experiments Fluidized Bed Experiments Sample Characterizations DFT Calculations Results and Discussions Motivation for Dopant Addition Thermal Stability of CaMnO Effect of A-Site Dopants on Oxygen Uncoupling Properties x

15 4.3.4 Structural and Stability Studies Redox Testing Conclusions REFERENCES CHAPTER 5 CaMn1-xBxO3-δ (B= Al, V, Fe, Co, and Ni) Perovskite Based Oxygen Carriers for Chemical Looping with Oxygen Uncoupling (CLOU) Abstract Introduction Experimental Oxygen Carrier Synthesis Oxygen Uncoupling Experiments Fluidized Bed Experiments Sample Characterizations Results and Discussion Effect of Dopant into the B-site Effects of Iron Dopant Concentrations Redox Testing Conclusion REFERENCES CHAPTER 6 Conclusions and Future Work REFERENCES xi

16 APPENDICES Appendix A Supplementary Information for Chapter Appendix B Supplementary Information for Chapter Appendix C Supplementary Information for Chapter Appendix D Supplementary Information for Chapter xii

17 LIST OF TABLES Table 1.1- Carbon capture energy penalty to gas-fired and coal-fired power plants[10,11,23 25] Table 2.1. Experimental conditions for 1-cycle, 5-cycle, and 50-cycle redox experiments using hydrogen as the reducing gas Table 2.2. Crystalline size change during multi-cyclic redox reactions of the LSF supported Fe2O Table 2.3. Carbon Deposition (% wt.) results when using CO and CH4 as reducing gases Table 2.4. EDX results of the LSF supported oxygen carriers at the points represented in Figure 2.8 (b) taken at 10.0kV for enhanced special resolution Table 2.5. Elemental Composition comparison of the Fe2O3 supported on LSF oxygen carrier of the bulk compared to the surface. Carbon content in the XPS result is removed since it is introduced for binding energy calibration Table 3.1. List of oxygen carriers synthesized Table 3.2. Carbon Formation (within 10 mins. of reduction in CH4) and BET surface areas of the CSTN supported iron oxide oxygen carrier Table 3.3. Summary of BET, crystallite sizes, and carbon formation study on the Fe2O3:CeO2 (4:6) CP oxygen carrier Table 3.4. Summary of carbon formation, crystallite sizes, and BET surface areas of the MgAl2O4:Fe2O3 (4:6) oxygen carrier Table 4.1. Oxygen vacancy (OV) formation energy of BaMnO3 and CaMnO3 under FM (ferromagnetic) and AFM (antiferromagetic) orders Table 5.1- Gas conditions for redox cycles used during fluidized bed experiments xiii

18 Table 5.2-Lattice parameters for the CaMn0.8B0.2O3 oxygen carriers Table 6.1- Apparent Arrhenius activation energies for surface exchange for the synthesized oxygen carriers xiv

19 LIST OF FIGURES Figure 1.1- World carbon dioxide emissions (by fuel) yearly projections from the U.S. Energy Information Administration Figure 1.2- Simplified plant schematics for the three different types of carbon capture techniques Figure 1.3- Schematic of the chemical looping combustion process Figure 1.4- CLOU process schematic (only the fuel reactor is illustrated) Figure 2.1. Reduction of as-prepared LSF and TiO2 supported iron oxide particles in a TGA apparatus (10% H2 balanced with N2/He; 900 C; Reference oxygen carrier refers to TiO2 supported iron oxide) Figure 2.2. Reactivity comparisons of as-prepared TiO2 supported Fe2O3 (reference oxygen carrier) and LSF supported Fe2O3 (oxygen carrier with enhanced mixed-conductivity); reducing gas: 10.5% hydrogen, temperature: 900 C Figure 2.3. Initial Temperature for Reduction of Fe2O3 supported on LSF and TiO2 in TGA (a) and H2-TPR (b) Figure 2.4. Time to achieve 60% conversion over the first 5 cycles using 10% H Figure 2.5. Cyclic redox experiment of (a) La0.8Sr0.2FeO3-δ supported Fe2O3 in reduction cycles; (b) La0.8Sr0.2FeO3-δ supported Fe2O3 in oxidation cycles. Reducing gas: 10% hydrogen, temperature: 900 C; oxidizing gas: 10% oxygen, temperature: 900 C Figure 2.6. XRD spectra of LSF supported Fe2O3 in oxidized (a) and reduced (b) forms over 51 reduction-oxidation cycles. (the small unlabelled peaks in 1 st cycle samples can be attributed to the insignificant amount of unconverted strontium ferrite from solid state reactions) xv

20 Figure 2.7. Reactivity comparisons of the reference and LSF supported oxygen carriers in 10% CO at 900 C Figure 2.8. Reactivity comparisons of synthesized oxygen carriers in 10% CH4 showing (a) time to achieve 11% conversion and (b) time to achieve 33% conversion Figure 2.9. SEM Image (20 kv) of LSF supported Fe2O3 of (a) fresh particles and (b) after 51 oxidations and two locations where EDX point analyses are performed Figure Surface composition analysis using XPS of the Fe2O3 supported on La0.8Sr0.2FeO Figure Proposed reduction reaction scheme of the LSF supported Fe2O3 oxygen carrier Figure 3.1- Comparison of the 2 nd cycle reduction of the various oxygen carriers by comparison of average conversion rate (Xave) to achieve 11% and 33% conversion in (a) 10% H2 and (b) 10% CH4 at 900 C Figure 3.2- Average conversion rate (Xave) to 33% conversion comparison between 2 nd and 10 th cycles of the various oxygen carriers at 900 C in 10% CH Figure 3.3- Average conversion rate ( Xave ) to achieve 33% conversion of Ca0.8Sr0.2Ti0.8Ni0.2O3 supported Fe2O3 via CA method with (a) 10 cycles with H2 and (b) 50- cycles with CH4 at 900 C Figure 3.4- XRD analyses of the (a) fresh and (b) 10-cycled CSTN supported iron oxide samples Figure 3.5- Oxygen carrier weight loss (TGA) curves for cycle 2 and cycle 10 of the Fe2O3:CSTN (CA) in CH4 at 900 C. Vertical lines represent the estimated coke formation for the reaction with methane. Lower horizontal line represents the demarcation between xvi

21 Fe2O3/Fe3O4 and Fe3O4/FeO; Upper horizontal line represents the demarcation between Fe3O4/FeO and FeO/Fe (calculated based on average oxidation state) Figure 3.6- Average reduction rate (Xave) to achieve 33% conversion of Fe2O3:CeO2 (4:6) prepared via a co-precipitation route over the first 10 redox cycles during a 50 cycle experiment. Reducing gas: methane; Oxidizing gas: oxygen; Temperature: 900 C Figure 3.7- XRD of the (a) fresh and 50th cycle oxidized (b) 1 st reduced and 51 st reduced during the redox in methane and oxygen of the (4:6) co-precipitation Fe2O3:CeO2 sample Figure 3.8- XPS of the (a) fresh and (b) 51 st oxidized ceria supported oxygen carrier Figure 3.9- Average conversion rate (Xave) to (a) 11% and (b) 33% conversion comparison of the 800 C 1,100 C sintered CeO2 supported Fe2O3 (6:4) in CH4 at 900 C synthesized by a co-precipitation method. The lower sintered oxygen carrier does not achieve 33% conversion in the 25 th and 50 th cycles Figure Comparison of the activity of CeO2 supported iron oxides versus the Fe/Ce ratio on the surface. Cycle 5 and 50 data are used for the 1,100 C sample. Cycle 1 and 50 data are used for the other samples Figure Average conversion rate (Xave) (%/min) to 33% conversion of the MgAl2O4 supported Fe2O3 (7:3) via SSR tested in (a) 10 cycles with H2 and (b) 5-cycles with CH4 and (c) 5 cycles in CH4 of 1200 C sintered 4:6 Fe2O3:MgAl2O4 at a reaction temperature of 900 C Figure XRD of the (a) fresh and 5th oxidized samples of Fe2O3 MgAl2O4 SSR sample and (b) of the 1 st and 6 th reduced oxygen carrier Figure TEM images of the 6 th reduced MgAl2O4 supported iron oxide xvii

22 Figure 4.1- (a) TPD of the pure CaMnO3 at 5 C/min to 1000 C in pure helium environment. (b) In-situ XRD of the same oxygen carrier at 5 C/min to 1200 C in argon Figure 4.2- Inert temperature programmable desorption (TPD) profiles of (a) Ba-doped and (b) Sr-doped CaMnO3 ramped to 1000 C at a rate of 20 C/min in a pure helium atmosphere. 117 Figure 4.3- Total oxygen release and initial decomposition temperature comparisons between CaMnO3 and Sr- and Ba-doped synthesized oxygen carriers Figure 4.4- XRD of various doped CaMnO3 with (a) Sr doping and (b) Ba doping. Phases: ( ) SrMnO3, (x) CaMnO3, ( ) BaMnO3, and ( ) Mn3O Figure 4.5- In-situ XRD of the (a) Ca0.75Sr0.25MnO3 and (b) Ca0.95Ba0.05MnO3 in argon at a ramping rate of 5 C/min to 1200 C Figure 4.6- Isothermal oxygen donation at temperatures between C. Reduction step: 5.0 grade helium. Oxidation step: 10% O Figure cycles at 850 C Ca0.75Sr0.25MnO3 oxygen carrier Figure 4.8- Char combustion using TGA of pure and Sr-doped CaMnO3. Heating in inert to 250 C for drying, then heated at a rate of 20 C/min. in helium to 950 C Figure 4.9- Char conversion and CO2 selectivity of the Ca0.75Sr0.25MnO3 oxygen carrier in a fluidized bed reactor. Temperature: 850 C Gas Flow Rates: nitrogen: 800mL/min and helium: 280mL/min (Char Reduction) and oxygen: 120mL/min (during oxidation only) Figure 5.1- XRD spectra of the CaMn0.8B0.2O3 oxygen carriers. ( )CaMnO3, ( ) Ca3Co1.5Mn0.5O6, ( ) V2O5 and ( ) NiO phases Figure 5.2- DTG/TPD curves of CaMn0.8B0.2O3 oxygen carriers (B=Fe, V, Co, Ni, and Al) to 1000 C at 10 C/min. in a pure inert atmosphere after treatment in 25% O2 at 850 C. Sample xviii

23 weight: 30-35mg Gas flow rate: 20mL/min He (grade 5.0) and 100mL/min Ar (grade 5.0) Figure 5.3- XRD spectra of the CaMn1-xFexO3 spectra Figure 5.4- DTG/TPD curves of CaMn1-xFexO3 oxygen carriers (x=0, 0.05, 0.2, and 0.5) to 1000 C at 10 C/min. in a pure argon atmosphere after treatment in 25% O2 at 850 C. Sample weight: 30-35mg Gas flow rate: 20mL/min He (grade 5.0) and 100mL/min Ar (grade 5.0) Figure 5.5- (a) and (c) In-situ XRD spectra of the CaMn0.95Mn0.05O3 and CaMnO3 perovskite oxygen carrier and (b) and (d) a better resolution of the (110) plane for both oxygen carriers Figure 5.6- Isothermal weight loss at temperatures between C. Sample weight: 30mg. Flow Rates: Reduction step: 30mL/min He (5.0 grade) and 60mL/min Ar (5.0 grade). Oxidation step: additional 10mL/min O2 (Extra Dry grade) added to aforementioned gas stream Figure cycle tests at 850 C for CaMn0.95Fe0.05O3 oxygen carrier. Sample weight: 50mg. Gas flow rates: Reduction step: 30mL/min He (5.0 grade) and 60mL/min Ar (5.0 grade). Oxidation step: additional 10mL/min O2 (Extra Dry grade) added to aforementioned inert gas. The XRD pattern of the fresh and post 100 CLOU cycles XRD of the oxygen carrier is embedded Figure 5.8- Char conversion and CO2 selectivity of the CaMn0.95Fe0.05O3 oxygen carrier in a fluidized bed reactor. Sample weight: 15 gms. Oxygen carrier and 10mg of char Temperature: 850 C Gas Flow Rates: nitrogen: 800mL/min (Grade 5.0) and helium (Grade 5.0): 280mL/min (Char Reduction) and oxygen (Extra dry grade): 120mL/min (during oxidation only) xix

24 Figure 6.1- (a) CaMnO3, (b) CaMn0.75Fe0.25O3, and (c) Ca0.75Sr0.25MnO3 fractions of 16 O2, 16 O 18 O, 18 O2 from pulse isotope exchange experiments at PO2= 0.1 atm. Sample: 50 mg. Gas Flow: 50 ml/min Figure 6.2- R0, Ra, and Ri for (a) CaMnO3, (b) CaMn0.75Fe0.25O3, and (c) Ca0.75Sr0.25MnO3 oxygen carriers at a PO2=0.1 atm (d) Compares the R0 for the three oxygen carriers at 500 C and 725 C xx

25 CHAPTER 1 Introduction 1.1 Carbon Emissions Anthropogenic carbon emissions from fossil fuel conversion have led to unintended consequences of climate change. Figure 1.1 shows the 2015 and 2030 (projected) world CO2 emissions predicted by U.S. Energy Information Administration (EIA) in their international energy outlook reports from 2006 and 2016, respectively [1,2]. Figure 1.1- World carbon dioxide emissions (by fuel) yearly projections from the U.S. Energy Information Administration. The EIA made an accurate prediction of 2015 CO2 emissions in their 2006 report. With respect to the 2030 projections, the 2016 EIA report offers a more optimistic picture for CO2 emission reductions. This can be attributed to the better than anticipated progress in terms of the development and adoption of renewable energy technologies such as those using solar and wind energies. However, EIA does still predict a steady increase in CO2 emissions over time. The EIA reported in 2009 that 98% of the CO2 emissions in the U.S. come from energy related sources with the primary source being fossil fuel combustion. At a world scale, average CO2 1

26 concentrations in the atmosphere have risen from ~310ppm in 1959 to ~400ppm today. This rise of CO2 in the atmosphere correlates with a nearly 0.8 C rise in the global surface temperature. The continued reliance of fossil fuels prompts the needs to develop cost-effective carbon capture and storage techniques (CCS). 1.2 Carbon Capture CO2 is the largest contributor to greenhouse-gases-induced climate change by human activity. Therefore, development of reliable, efficient, and low-cost carbon-capture technologies is essential to mitigate the issues related to CO2 emissions [3]. The three most developed carboncapture techniques for capturing CO2 emissions from carbonaceous-fueled power plants are pre-combustion, post-combustion, and oxy-fuel combustion techniques. Figure 1.2 indicates the typical schematic for how the carbon capture occurs for each technique. Figure 1.2- Simplified plant schematics for the three different types of carbon capture techniques. Post-combustion CO2-capture technologies separate CO2 in flue gas produced from combustion of carbonaceous fuels. Post-combustion separation can be retrofitted to be used on 2

27 current and newly developed coal-fired power plants. Typically, concentration of CO2 in flue gas in coal combustion plants is approximately 15% [4]. Post-combustion techniques mainly are derived from chemical absorption using amines in particular monoethanolamine (MEA) or a sterically hindered amine [5 14]. The amine-based systems use the reaction of CO2 with the amine to form water-soluble compounds. However, regeneration of the amine solvent is very energy-intensive requiring high temperature steam. Because of the large steam usage, increased electricity costs and lower plant efficiencies are observed [15]. Besides amine absorption, carbonate, aqueous ammonia, membranes, metal organic frameworks, enzyme, and ionic liquid technologies have shown promise at the laboratory scale for post-combustion capture [16]. Pre-combustion carbon capture technologies remove CO2 from the process prior to burning of the fuel. The most developed pre-combustion capture technology is associated with the Integrated Gasification Combined Cycle (IGCC). The process involves gasification of coal into syngas containing CO, H2, and CO2 followed by using the water-gas shift reaction to take CO and convert it into CO2. From here, H2 can be separated out and fed to a gas turbine. IGCC systems suffer from very high capital costs and cannot be retrofitted to existing coal-fired power plants [7,11,16,17]. Oxy-fuel combustion techniques work by removing nitrogen from flue gases, leading to primarily CO2 and H2O. This is commonly achieved through the use of a cryogenic airseparation unit (ASU) to obtain a pure (>95%) oxygen stream for combustion of the fuel. Early techno-economic assessments suggested that oxy-fuel combustion would be the most energy and cost efficient of carbon capture techniques [7,18 21]. This conclusion was drawn from the assumptions from higher boiler efficiency from smaller flue gas volume and lack of need to 3

28 separate out NOx and SOx from the flue gas. However, the energy intensive oxygen separation step and the needs to recirculate portion of the flue gas lead to significant energy penalty [16,22]. Moreover, although oxy-fuel combustion was proposed to be retrofitted to existing coal-fired boilers, recent demonstrations indicates that existing boilers tend to leak significant amount of air into the system, leading to a flue gas stream that requires further separation. By utilizing these carbon capture techniques, overall plant efficiencies are seen to drop significantly. Table 1.1 reports average thermal efficiencies with and without carbon capture for both gas-fired and coal-fired plants. Table 1.1- Carbon capture energy penalty to gas-fired and coal-fired power plants[10,11,23 25]. Carbon Capture Technology Thermal Efficiency (% LHV) Overall Efficiency Drop (%) Gas-Fired No Capture Post-Combustion Pre-Combustion Oxy-fuel Combustion Coal-Fired No Capture Post-Combustion Pre-Combustion Oxy-fuel Combustion As can be seen from Table 1.1, thermal efficiencies decrease between 11-29%. As a result of the energy penalties and sometimes difficulty in retrofitting these technologies to existing plants, more cost- and energy-efficient carbon-capture technologies need to be developed. Of these technologies, the chemical looping process has garnered significant attention in the last 15 years. 4

29 1.3 Chemical Looping Processes Chemical looping processes offer an alternative and potentially efficient strategy for conversion of fossil fuels for power generation with in-situ CO2 capture. Issues commonly faced in the abovementioned carbon capture techniques are mitigated in chemical looping. Metal oxide particles, which are commonly referred to as oxygen carrier particles, are used in a cyclic redox scheme for conversion of solid and gaseous carbonaceous fuels. The redox cycle involves initially a reaction between oxygen carrier s lattice oxygen and carbonaceous fuel to form a concentrated stream of CO2 and steam. The CO2 separation can take place prior to the combustion of the oxygen depleted metal oxide in air. As a result, chemical looping is considered as both pre-combustion and oxy-combustion capture techniques Chemical Looping Combustion (CLC) The CLC scheme, as shown in Figure 1.3, uses the lattice oxygen stored in the oxygen carriers to combust carbonaceous fuels such as natural gas, syngas, refinery gas, etc. Oxygen carriers, once depleted, can recover their oxygen through combustion with air or reaction with steam. Figure Schematic of the chemical looping combustion process. 5

30 The CLC process uses two interconnected reactors termed the fuel reactor (reducer) and air reactor (oxidizer) respectively. The oxidized oxygen carrier in the fuel reactor is able to oxidize the carbonaceous fuel into a highly concentrated stream of CO2 and steam. The resulting flue gas can be separated through steam condensation allowing for a concentrated stream of CO2. After the lattice oxygen is depleted from the oxygen carrier, it can be transferred to the air reactor where lattice oxygen can be restored to the particle through combustion with air. The combustion process in air is highly exothermic and produces a N2 enriched flue gas stream and heat, which can be used for power generation. Moreover, the oxygen carriers also act as heat carriers allowing the transfer of heat from the air reactor to the fuel reactor which it typically an endothermic reaction. By transferring the heat through the use of the oxygen carriers, this can lead to improved thermal efficiencies [26]. As can be seen, the CLC process avoids the cryogenic ASU required by typical oxy-fuel combustion processes. Compared to postcombustion CO2-capture processes, the highly energy-intensive CO2 separation step is also eliminated. For the gas-fired CLC scheme, excellent contact is required between the gas phase and solid phase materials, as well as capability of transferring the oxygen carriers from the air reactor to fuel reactor and vice versa. As a result, circulating fluidized bed (CFB) reactors are almost exclusively used for CLC [26 28]. In CFB reactors, the fuel reactor is operated in the bubbling fluidized bed regime at low superficial gas velocity. The solid oxygen carriers are carried from the fuel reactor to the air reactor in which regeneration of depleted oxygen with air can be completed before recirculation to the fuel reactor to close the loop. The CFB reactor has been demonstrated at pilot plant scale up to 3 MWth [26,29 31]. Additionally, process design and development began on a 1000 MWth chemical looping CFB reactor [32]. Besides the CFB 6

31 reactor design, moving bed reactors have been proposed due to their ability to achieve higher metal oxide and fuel conversions [26,33,34]. Other alternative reactor schemes tested in laboratory scale include spout-fluid bed reactor[35,36], double loop circulating fluidized bed reactors[37], rotating reactor[38], and packed bed reactor systems [39 41]. Early CLC studies focused on converting gaseous carbonaceous fuels such as natural gas/methane[42 45] and coal derived syngas [46,47]. However, solid fuels such as coal offer a lower cost on same energy content basis. As a result, it is advantageous to develop chemical looping processes for direct conversion of solid fuels. However, low solid-solid reaction kinetics between solid fuel and oxygen carrier particles hinders the direct use of solid fuels. One potential design for implementing solid fuels in chemical looping is through gasification into syngas using oxygen before sending the fuel into the system. However, gasification increases process costs and complexity due to the need of a gasifier and ASU. Design of chemical looping process schemes for more direct conversion of the solid fuels is desired Chemical Looping with In-Situ Gasification One method developed for solid fuel CLC is through in-situ gasification [48 53]. This approach requires a gasification enhancer, typically CO2 or steam. The enhancer s role is to convert the solid carbon content in the fuel into H2 and CO through a combination of steamcarbon and reverse Boudouard reactions. Steam-Carbon Reaction: C + H2O CO + H2 Reaction 1.1 Reverse Boudouard Reaction: C + CO2 2CO Reaction 1.2 The resulting syngas will then be allowed to react with the oxygen carrier particles at higher rate than solid fuel-oxygen carrier particle reactions. However, the steam-carbon and reverse 7

32 Boudouard reactions are slow at the temperatures typically operated in the fuel reactor, leading to high residence times for both the oxygen carrier and solid fuel particles to achieve high conversions [26]. The resulting higher residence times lead to a larger reactor volume for the fuel reactor and higher solids inventory causing higher capital and operational costs Chemical Looping with Oxygen Uncoupling (CLOU) An alternative to the in-situ gasification approach is the chemical looping with oxygen uncoupling method (CLOU). The CLOU strategy was first suggested in the 1950 s by Lewis and Gilliland [54,55]. The CLOU process scheme is shown in Figure 1.4. Figure CLOU process schematic (only the fuel reactor is illustrated). The basic principle behind the CLOU scheme is the use of oxygen carriers that take advantage of the difference of PO2 between inside the lattice of the metal oxide and the atmosphere inside the fuel reactor. The difference in PO2 allows the oxygen carrier to release or uncouple its oxygen directly into gaseous oxygen. The gaseous oxygen is able to react at a much higher rate with the solid char particles than the solid-solid interactions between fuel and oxygen carrier 8

33 particles. Additionally, the use of a gasifier enhancer such as CO2 can improve the fuel conversion through gasification of the solid fuels to volatiles that additionally will react with the oxygen carriers leading to chain reactions. 1.4 Oxygen Carriers Oxygen carrier design is of upmost importance for design and operation of chemical looping processes. Design parameters of oxygen carriers typically include high oxygen transport capacity, reaction kinetics, long-term recyclability, cost, mechanical strength, high melting points, resistance to contaminants of different fuels, resistance to carbon formation, and minimal effects to health and the environment [56 58]. Active oxygen carriers are often based on Fe, Ni, Mn, Cu, and Co oxides. However, most pure single metal oxides suffer from several disadvantages such as low reaction kinetics and particle agglomeration [56,59]. As a result, a secondary metal oxide termed a support is added to the oxygen carrier to improve its properties. Typically, supports consist of inert ceramics such as Al2O3 and TiO2. Additionally, mixed metal oxides and perovskites have been investigated as oxygen carriers. Based on literature, approximately 1000 different oxygen carriers have been developed and tested for chemical looping processes [26,56 58,60 65]. Certain oxygen carrier properties are based on the metal oxide redox system chosen. These properties include cost, environmental impact, oxygen transport capacity, and thermodynamic favorability to convert fuel into CO2 and H2O. Mechanical properties such as attrition and agglomeration resistance as well as activity for different fuels must be performed experimentally on oxygen carriers. For example, attrition can be measured in a number of ways, including using crushing strength or testing in a fluidized bed reactor for attrition. 9

34 Attrition affects the lifetime of oxygen carriers, which is defined as the amount of time under reaction (reduction or oxidation) without decrease in reactivity or elutriation of oxygen carrier particles out of the reactor [56]. Particles sizes of <45μm are considered as fines that would not contribute to the reactivity of oxygen carrier particles due to their low residence time in the reactors [56,66,67]. Additionally, smaller particle sizes have varying health and environmental impacts that must be considered when chemical looping processes are being developed Fe-Based Oxygen Carriers Iron oxides are very attractive to the CLC application due to their low cost and environmental impacts. However, iron oxides suffer from poor activity with carbonaceous fuels and poor oxygen transport rate. Iron oxides consist of several transition oxidation states when reduced from Fe2O3 including Fe3O4, FeO, and Fe. For CLC applications in CFB reactors, the transition from hematite (Fe2O3) to magnetite (Fe3O4) is primarily considered. This is due to the thermodynamic limitation that further reduction to wustite (FeO) and metallic iron (Fe) would decrease CO2 purity. This is attributed to increase production in CO and H2 as hematite is reduced through its oxidation states to metallic iron [68]. The use of countercurrent moving bed reactors can help improve CO2 selectivity in iron oxide reactions. Additionally, reduction to wustite and metallic iron has shown agglomeration issues that must be addressed in design of iron oxide oxygen carriers. The addition of supports has been shown to help address these issues with Fe-based oxygen carriers. Common supports used with iron oxide include TiO2, Al2O3, SiO2, MgAl2O4, and Zrbased supports [61,64,69 75]. The primary role of these supports is to improve the mechanical properties of pure iron oxide. Additionally, supports such as Al2O3 and TiO2 form compounds of FeAl2O4 and FeTiO3, which are of the oxidation state of Fe 2+ allowing for full conversion 10

35 of the fuel to CO2 and H2O [69,76,77]. Activity changes, i.e. deactivation or activation, have been reported for supported iron oxide over multiple redox cycles [70,78 80]. While pure iron oxide is shown to deactivate due to reduced surface area and porosity, iron oxide ores such as ilmenite and synthetic supported iron oxides are shown to activate during the chemical looping combustion (CLC) processes. This could be attributed to an increase in porosity over cyclic redox reactions [78,80]. Li et al. investigated TiO2 supported iron oxides [81,82]. Slight increase in activity of the oxygen carrier was observed with decreasing surface area and porosity through redox cycles. Based on inert marker experiments and density function theory (DFT) calculations, the authors proposed that higher conductivities of electrons and lattice oxygen (O 2- ) of the supported sample are responsible for the improved redox activity of the supported iron oxide. Iron oxides also have an advantage as they can be oxidized by either air or steam to produce heat or H2, respectively. With regards to Fe oxidation, steam can be used to generate H2 from the water-splitting reaction while regenerating the depleted oxygen as shown in reaction 1.3. Water-Splitting Reaction with Fe: 3Fe +4H2O Fe3O4 + 4H2 Reaction 1.3 Additionally, a second regeneration step using air can be used to combust magnetite to hematite. Supports used in water splitting of iron oxides include CeO2-ZrO2 mixed oxides [83,84], Al2O3 [85,86], and MgAl2O4 [87]. More recently, He et al. have investigated La1- xsrxfeo3-δ (LSF) supported Fe2O3 for water splitting in a novel solar-redox scheme [88,89]. In this system, water-splitting conversions reached above 77% conversion, which surpasses the theoretical thermodynamic conversion of unsupported iron oxide. The excess in conversion contributed additionally to the LSF being able to convert steam. 11

36 1.4.2 Ni-Based Oxygen Carriers Nickel-based oxygen carriers are the most widely studied oxygen carrier materials. Nickel oxides have very good activity for gaseous fuels such as methane but suffer from high costs, low resistance to sulfur and carbon formation, and environmental and health concerns. Pure NiO, despite having high activity for fuel conversion, suffers from poor recyclability due to low re-oxidation rates from Ni agglomeration [90]. Al2O3 is the most commonly used support for Ni-based oxygen carriers [61,75,91 94]. However, Al2O3 and NiO tend to form spinel NiAl2O4, which is inactive at typical reaction temperatures of the fuel reactor of CLC operations ( C) [95,96]. To address this issue, researchers use excess NiO with Al2O3, so free NiO content is high enough to achieve high activity [97 99]. Alternatively, addition of secondary metals such as Mg, Ca, and La can form spinel with the Al2O3 and improve reactivity of the Ni oxygen carrier [ ]. Other supports have also been tested for Ni oxygen carriers. ZrO2 or YSZ have shown good activity for fuel conversion but suffer from undesirable mechanical properties such as attrition resistance [61,104,105]. Low activity has been reported for bentonite, TiO2, and MgO as a consequence of formation of stable, unreactive complex compounds under reaction conditions [100, ]. SiO2 and sepiolite supported NiO suffer from deactivation over cycles and poor mechanical properties [61,108]. Unlike Fe-based oxygen carriers, Ni is susceptible to sulfur poisoning. Presence of H2S, which is common in both gaseous and solid carbonaceous fuels, reacts with Ni to form nickel sulfide. Through formation of the nickel-sulfide phase, deactivation of the oxygen carrier was observed [109,110]. From tests conducted in a 500Wth fluidized bed rector, if the H2S content was greater than 100ppmv. then significant nickel sulfide would form, and deactivation would 12

37 occur [109]. However, NiO s high activity and good mechanical properties make NiO highly promising as an oxygen carrier. The toxicity of the oxygen carrier remains its biggest obstacle Cu-Based Oxygen Carriers Copper oxygen carriers offer the advantage of high reactivity with slightly lower costs and health concerns than nickel oxide oxygen carriers. Copper oxide s high reactivity with carbonaceous fuels stems from the ability of it to uncouple its oxygen from the CuO to Cu2O oxidation state. Copper oxides are also thermodynamically favorable for the full conversion of fuel into CO2 and H2O. However, Cu faces mechanical issues. Agglomeration is prominent as CuO reduces to Cu metal due to the low melting point of Cu (1085 C) which will lead to defluidization in chemical looping reactors [56]. Support additions include those of Al2O3, SiO2, MgAl2O4, TiO2, and ZrO2 [41,60,61,75,108, ]. While CuO also forms complex oxides with some supports such as Al2O3, the complex oxides are also active for CLC applications. However, CLOU properties of these carriers may be altered leading to slower reactions with fuels. Supports such as SiO2 and MgAl2O4 have shown to improve sintering resistance and recyclability of the oxygen carriers while maintaining the oxygen-uncoupling properties of CuO [ ] Mn-Based Oxygen Carriers Manganese oxygen carriers offer the interesting combination of iron oxide s low cost and health risks with copper oxides ability to uncouple oxygen. Mn oxides have several transition states similarly to iron oxide. However, most commonly only Mn3O4 to MnO is considered for CLC applications as Mn2O3 is not observed above 800 C in air [118]. Pure manganese-oxide oxygen carriers show low activity with gaseous and solid fuels despite its oxygen uncoupling 13

38 properties [119,120]. Additionally, inert supports TiO2, Al2O3, SiO2, and MgAl2O4 are ineffective supports for manganese oxygen carriers due to the formation of unreactive and irreversible complex oxides [61,75,93,108,121]. The most effective Mn oxygen carriers utilize ZrO2 as a support. However, ZrO2 supported manganese oxides suffer from cracking and agglomeration [122]. MgO, CaO, and CeO2 have all been used with ZrO2 to help stabilize the oxygen carrier s mechanical properties [122]. While all additions showed improved activity, MgO-stabilized ZrO2 demonstrated the most stability and recyclability. Despite the improved recyclability, these oxygen carriers are not as active for methane as other metal oxides like Cu and Ni Co-Based Oxygen Carriers Cobalt oxides face severe challenges despite its favorable oxygen carrying capacity. These include high costs and toxicity. While cobalt oxides have several transition states, only CoO to Co is considered for CLC applications due to decomposition of Co3O4 above 900 C. Due to this consideration, the thermodynamics are less favorable for complete conversion into CO2 and H2O [56]. Similar to Mn oxygen carriers, inert supports face the challenge of formation of unreactive oxide phases such as CoAl2O4 and CoTiO3 [93,123]. YSZ supports exhibit good activity with low carbon formation Mixed-Metal Based Oxygen Carriers As has been discussed above, the addition of supports improves the oxygen carrier properties of metal oxides. Alternatively, the formation of complex, mixed metal oxides as oxygen carriers has been investigated. The combination of metals used is chosen to improve properties of the single metal oxides. Fe is commonly used to combine with other primary metals Ni, Cu, 14

39 Mn, and Co as it offers a low-cost, low-toxicity, high-oxygen-carrying-capacity option. Fe- Mn, Fe-Cu, and Fe-Co oxides can be used to exploit the oxygen uncoupling properties of Cu, Mn, and Co while obtaining the properties of lower cost, less toxicity, and better mechanical stability of iron oxides [ ]. Fe-Ni oxides have also been investigated as alternatives to pure NiO to help minimize cost and toxicity issues plagued by NiO oxygen carriers. Johannson et al. observed significant improvement in activity with CH4 of 3% NiO combined with Fe2O3 as compared to pure Fe2O3[128]. More recently, perovskite structured oxygen carriers have received growing attention as oxygen carriers. Perovskites are based on the formula ABO3-δ where the A-site represents a large cation, usually from the alkali earth or rare earth metals, and B-site is a transition metal. Stability of perovskites is based on the ionic radii of the A- and B-metal ions chosen. The Goldschmidt tolerance factor is a factor that can determine the stability of the perovskite semiempirically and is shown in Eq. 1.1[129,130]. t = r A + r O 2(r B + r O ) (Eq. 1.1) Based on equation 1.1, ra, rb, and ro are the ionic radii of the A- and B- site and oxygen ions respectively. Ideal cubic perovskites have a tolerance factor value of 1. Perovskite oxygen carriers commonly use La, Ca, Sr, and Ba as potential A site candidates while the B-site typically consists of Ti, Ni, Fe, Co, Mn, and Mg [ ]. For CLC operations using methane, La1-xSrxFe1-yCoyO3-δ (LSFC) oxygen carriers have shown the best performance [131,133]. When LSF is used, partial oxidation is favored [141]. Besides Labased perovskites, CaMnO3-δ offers potential as both an oxygen carrier in CLC and CLOU applications [ , ]. CaMnO3-δ faces challenges such as irreversible and 15

40 unreactive phase changes into spinel (CaMn2O4) and Ruddlesden-Popper (Ca2MnO4) phases and is susceptible to sulfur poisoning [136, ]. In summary, oxygen-carrier design is of upmost importance for chemical looping to be successful. The commonly used trial-and-error approach gives insight into information regarding specific oxygen carriers and their properties. However, general design principles for oxygen carriers have not been extensively studied. The ideology of previous research results in (i) highly active, recyclable, costlier, and often-toxic oxygen carriers and (ii) cheap, low activity and recyclability, and less toxic oxygen carrier design. 1.5 Summary Design of highly active and recyclable oxygen carriers is desirable for CLC. Moreover, understanding mechanisms for oxygen carrier redox reactions is important in order to design effective oxygen carriers for different chemical looping processes. The work investigated covers two topics: (i) the role of oxygen transport in Fe-based oxygen carriers and development of support-metal oxide synergies allowing for high reactivity with gaseous carbonaceous fuels and (ii) development and understanding the effects of doping the A- and B-site of perovskitebased CaMnO3 oxygen carriers to enhance the activity for coal char combustion. Chapters 2 and 3 discuss the relationship of support s role in the oxygen transport of Fe-based oxygen carriers for chemical looping conversion of carbonaceous fuels and hydrogen. Chapter 2 introduces the concept of using a support with high oxygen anion and electron transport termed a mixed ionic-electronic conductor (MIEC) support. The support chosen was La0.8Sr0.2FeO3-δ as it is a well-known MIEC material. LSF is commonly used in oxygen membrane separations. The MIEC support is compared to commonly used iron-oxide supports 16

41 such as inert TiO2 and Al2O3 and pure ionic conductor YSZ. The oxygen carriers are tested in a thermogravimetric analyzer for conversion of H2, CO, and CH4. Experiments show the MIEC-supported iron oxide has up to 70 times higher activity with the fuels compared to the other used supports. Additionally, the MIEC-supported oxide showed nearly no carbon formation and only slight agglomeration was observed over 50 redox cycles. Chapter 3 investigates further the understanding of support and iron oxide synergies. Various MIEC supports with different physical and chemical properties and inert supports are tested with redox cycles involving methane and oxygen. Perovskite MIEC-supported iron oxides exhibit both high activity and stability. However, MIEC, fluorite-structured CeO2-supported iron oxide deactivated due to agglomeration. The agglomeration resulted from Fe x+ transport through the oxygen carrier to the surface during the oxidation step resulting in segregation of iron oxide on the surface. Inert MgAl2O4-supported oxygen carriers were observed to activate during methane redox cycles. The resulting activation was caused by carbon formation resulting in physical disintegration and higher surface area and porosity of the oxygen carrier. Chapters 4 and 5 focus on the effects of A- and B-site dopants, respectively, on the CaMnO3- δ oxygen carriers for chemical looping with oxygen uncoupling for conversion of solid fuels. From the A-site perspective, Sr and Ba was considered as possible dopants, however, Ba doped CaMnO3-δ formed segregated phases even at low concentrations. Sr dopant shows compatibility with the base CaMnO3-δ. Additionally, the Sr-doped oxygen carrier has high recyclability and offered a lower uncoupling temperature compared to undoped CaMnO3-δ. B- site dopants were Fe, Ni, Co, Al, and V. Of these, Fe-doped CaMnO3-δ offered the most promising oxygen uncoupling properties. At lower temperatures (<850 C), the uncoupling properties were even greater than the Sr-doped oxygen carrier. Both Sr and Fe doped CaMnO3-17

42 δ offer greater conversion of solid coal char in a fluidized bed reactor compared to undoped CaMnO3-δ. Chapter 6 focuses on summarizing the abovementioned works and discusses potential future work for understanding oxygen transport in oxygen carriers. 18

43 REFERENCES [1] U.S. Energy Information Administration. World Energy Outlook [2] U.S. Energy Information Administration. World Energy Outlook [3] Advancing the Science of Climate Change. Washington, D.C.: National Academies Press; [4] Hu Y, Naito S, Kobayashi N, Hasatani M. CO2, NOx and SO2 emissions from the combustion of coal with high oxygen concentration gases. Fuel 2000;79: doi: /s (00) [5] Boot-Handford ME, Abanades JC, Anthony EJ, Blunt MJ, Brandani S, Mac Dowell N, et al. Carbon capture and storage update. Energ Environ Sci 2014;7: doi: /c3ee42350f. [6] Markewitz P, Kuckshinrichs W, Leitner W, Linssen J, Zapp P, Bongartz R, et al. Worldwide innovations in the development of carbon capture technologies and the utilization of CO2. Energ Environ Sci 2012;5:7281. doi: /c2ee03403d. [7] Toftegaard MB, Brix J, Jensen PA, Glarborg P, Jensen AD. Oxy-fuel combustion of solid fuels. Prog Energy Combust 2010;36: doi: /j.pecs [8] Davison J. Performance and costs of power plants with capture and storage of CO2. Energy 2007;32: doi: /j.energy [9] Descamps C, Bouallou C, Kanniche M. Efficiency of an Integrated Gasification Combined Cycle (IGCC) power plant including CO2 removal. Energy 2008;33: doi: /j.energy [10] Gibbins J, Chalmers H. Carbon capture and storage. Energ Policy 2008;36: doi: /j.enpol [11] Kanniche M, Gros-Bonnivard R, Jaud P, Valle-Marcos J, Amann J-M, Bouallou C. Precombustion, post-combustion and oxy-combustion in thermal power plant for CO2 capture. Appl Therm Eng 2010;30: doi: /j.applthermaleng

44 [12] Sander MT, Mariz CL. The Fluor Daniel econamine FG process: Past experience and present day focus. Energ Convers Manage 1992;33: doi: / (92)90029-v. [13] Wall TF. Combustion processes for carbon capture. P Combust Inst 2007;31: doi: /j.proci [14] Stiegel GJ, Ramezan M. Hydrogen from coal gasification: An economical pathway to a sustainable energy future. Int J Coal Geol 2006;65: doi: /j.coal [15] Ramezan M, Skone. Carbon dioxide capture from existing coal-fired power plants. US DOE/ NETL; [16] Figueroa JD, Fout T, Plasynski S, McIlvried H, Srivastava RD. Advances in CO2 capture technology The U.S. Department of Energy s Carbon Sequestration Program. Int J Greenh Gas Con 2008;2:9 20. doi: /s (07) [17] Pehnt M, Henkel J. Life cycle assessment of carbon dioxide capture and storage from lignite power plants. Int J Greenh Gas Con 2009;3: doi: /j.ijggc [18] Herzog H, Golomb D, Zemba S. Feasibility, modeling and economics of sequestering power plant CO2 emissions in the deep ocean. Environ Prog 1991;10: doi: /ep [19] Horn FL, Steinberg M. Control of carbon dioxide emissions from a power plant (and use in enhanced oil recovery). Fuel 1982;61: doi: / (82) [20] Nakayama S, Noguchi Y, Kiga T, Miyamae S, Maeda U, Kawai M, et al. Pulverized coal combustion in O2/CO2 mixtures on a power plant for CO2 recovery. Energ Convers Manage 1992;33: doi: / (92)90034-t. [21] Rubin ES, Mantripragada H, Marks A, Versteeg P, Kitchin J. The outlook for improved carbon capture technology. Prog Energ Combust 2012;38: doi: /j.pecs

45 [22] Anheden M, Yan J, De Smedt G. Denitrogenation (or Oxyfuel Concepts). Oil Gas Sci Technol 2005;60: doi: /ogst: [23] Leung DYC, Caramanna G, Maroto-Valer MM. An overview of current status of carbon dioxide capture and storage technologies. Renew Sust Energ Rev 2014;39: doi: /j.rser [24] Finkenrath M. Cost and performance of carbon dioxide capture from power generation. IEA; [25] Nord LO, Anantharaman R, Bolland O. Design and off-design analyses of a precombustion CO2 capture process in a natural gas combined cycle power plant. Int J Greenh Gas Con 2009;3: doi: /j.ijggc [26] Fan L-S. Chemical looping systems for fossil energy conversions. Wiley, [27] Kronberger B, Johansson E, Löffler G, Mattisson T, Lyngfelt A, Hofbauer H. A Twocompartment fluidized bed reactor for CO2 capture by chemical-looping combustion. Chem Eng Technol 2004;27: doi: /ceat [28] Lyngfelt A, Leckner B, Mattisson T. A fluidized-bed combustion process with inherent CO2 separation; application of chemical-looping combustion. Chem Eng Sci 2001;56: doi: /s (01) [29] Abdully I, Beal C, Andrus H, Epple B, Lyngfelt A, Lani B. Alstom s chemical looping prototypes, program update. 37th Int. Tech. Conf. Clean Co Al Fuel Syst., Clearwater, FL: [30] Ströhle J, Orth M, Epple B. Design and operation of a 1MWth chemical looping plant. Appl Energ 2014;113: doi: /j.apenergy [31] Lyngfelt A, Thunman H. Chapter 36 - Construction and 100 h of operational experience of a 10-kWth chemical - looping combustor. Carbon Dioxide Capture Storage Deep Geol. Form., Amsterdam: Elsevier Science; 2005, p

46 [32] Lyngfelt A, Leckner B. A 1000MWth boiler for chemical-looping combustion of solid fuels Discussion of design and costs. Appl Energ 2015;157: doi: /j.apenergy [33] Li F, Fan L-S. Clean coal conversion processes progress and challenges. Energ Environ Sci 2008;1:248. doi: /b809218b. [34] Li F, Zeng L, Velazquez-Vargas LG, Yoscovits Z, Fan L-S. Syngas chemical looping gasification process: Bench-scale studies and reactor simulations. AIChE J 2010;56: doi: /aic [35] Shen L, Wu J, Xiao J, Song Q, Xiao R. Chemical-looping combustion of biomass in a 10 kwth reactor with iron oxide as an oxygen carrier. Energ Fuel 2009;23: doi: /ef900033n. [36] Gu H, Shen L, Xiao J, Zhang S, Song T. Chemical looping combustion of biomass/coal with natural iron ore as oxygen carrier in a continuous reactor. Energ Fuel 2011;25: doi: /ef101318b. [37] Bischi A, Langørgen ø., Saanum I, Bakken J, Seljeskog M, Bysveen M, et al. Design study of a 150kWth double loop circulating fluidized bed reactor system for chemical looping combustion with focus on industrial applicability and pressurization. Int J Greenh Gas Con 2011;5: doi: /j.ijggc [38] Dahl IM, Bakken E, Larring Y, Spjelkavik AI, Håkonsen SF, Blom R. On the development of novel reactor concepts for chemical looping combustion. Energ Procedia 2009;1: doi: /j.egypro [39] Noorman S, van Sint Annaland M, Kuipers JAM. Experimental validation of packed bed chemical-looping combustion. Chem Eng Sci 2010;65:92 7. doi: /j.ces [40] Noorman S, van Sint Annaland M, Kuipers. Packed bed reactor technology for chemical-looping combustion. Ind Eng Chem Res 2007;46: doi: /ie061178i. [41] Kooiman RF, Hamers HP, Gallucci F, van Sint Annaland M. Experimental demonstration of two-stage packed bed chemical-looping combustion using syngas with 22

47 CuO/Al2O3 and NiO/CaAl2O4 as Oxygen Carriers. Ind Eng Chem Res 2015;54: doi: /ie [42] Consonni S, Lozza G, Pelliccia G, Rossini S, Saviano F. Chemical-looping combustion for combined cycles with CO2 capture. J Eng Gas Turb Power 2006;128:525. doi: / [43] Johansson E, Mattisson T, Lyngfelt A, Thunman H. Combustion of syngas and natural gas in a 300 W chemical-looping combustor. Chem Eng Res Des 2006;84: doi: /cherd [44] Kronberger B, Lyngfelt A, Löffler G, Hofbauer H. Design and Fluid dynamic analysis of a bench-scale combustion system with CO2 separation chemical-looping combustion. Ind Eng Chem Res 2005;44: doi: /ie049670u. [45] Mattisson T, Lyngfelt A, Cho P. The use of iron oxide as an oxygen carrier in chemicallooping combustion of methane with inherent separation of CO2. Fuel 2001;80: doi: /s (01) [46] Abad A, Mattisson T, Lyngfelt A, Ryden M. Chemical-looping combustion in a 300W continuously operating reactor system using a manganese-based oxygen carrier. Fuel 2006;85: doi: /j.fuel [47] Mattisson T, García-Labiano F, Kronberger B, Lyngfelt A, Adánez J, Hofbauer H. Chemical-looping combustion using syngas as fuel. Int J Greenh Gas Con 2007;1: doi: /s (07) [48] Lyngfelt A. Chemical-looping combustion of solid fuels Status of development. Appl Energ 2014;113: doi: /j.apenergy [49] Sorgenfrei M, Tsatsaronis G. Design and evaluation of an IGCC power plant using ironbased syngas chemical-looping (SCL) combustion. Appl Energ 2014;113: doi: /j.apenergy [50] Tong A, Bayham S, Kathe MV, Zeng L, Luo S, Fan L-S. Iron-based syngas chemical looping process and coal-direct chemical looping process development at Ohio State University. Appl Energ 2014;113: doi: /j.apenergy

48 [51] Scott SA, Dennis JS, Hayhurst AN, Brown T. In situ gasification of a solid fuel and CO2 separation using chemical looping. AIChE J 2006;52: doi: /aic [52] Leion H, Mattisson T, Lyngfelt A. The use of petroleum coke as fuel in chemicallooping combustion. Fuel 2007;86: doi: /j.fuel [53] Luo S, Bayham S, Zeng L, McGiveron O, Chung E, Majumder A, et al. Conversion of metallurgical coke and coal using a Coal Direct Chemical Looping (CDCL) moving bed reactor. Appl Energ 2014;118: doi: /j.apenergy [54] Lewis WK, Gilliland ER, Sweeney WP. Gasification of carbon metal oxides in a fluidized power bed. Chem Eng Prog 1951;47: [55] Lewis W and Gillialand E. Production of pure carbon dioxide , [56] Adanez J, Abad A, Garcia-Labiano F, Gayan P, de Diego LF. Progress in chemicallooping combustion and reforming technologies. Prog Energy Combust 2012;38: [57] Nandy A, Loha C, Gu S, Sarkar P, Karmakar MK, Chatterjee PK. Present status and overview of chemical looping combustion technology. Renew Sust Energ Rev 2016;59: doi: /j.rser [58] Tang M, Xu L, Fan M. Progress in oxygen carrier development of methane-based chemical-looping reforming: A review. Appl Energ 2015;151: doi: /j.apenergy [59] Roux S, Bensakhria A, Antonini G. Study and improvement of the regeneration of metallic oxides used as oxygen carriers for a new combustion process. Int J Chem React Eng 2006;4. [60] Imtiaz Q, Hosseini D, Müller CR. Review of oxygen carriers for chemical looping with oxygen uncoupling (CLOU): Thermodynamics, material development, and synthesis. Energy Technol 2013;1: doi: /ente [61] Adánez J, de Diego LF, García-Labiano F, Gayán P, Abad A, Palacios JM. Selection of oxygen carriers for chemical-looping combustion. Energ Fuels 2004;18: doi: /ef

49 [62] Adanez J, García-Labiano F, de Diego LF, Gayán P, Celaya J, Abad A. Characterization of oxygen carriers for chemical-looping combustion. Greenh Gas Con Technol 7, Oxford: Elsevier Science Ltd; 2005, p [63] Kang K-S, Kim C-H, Bae K-K, Cho W-C, Kim S-H, Park C-S. Oxygen-carrier selection and thermal analysis of the chemical-looping process for hydrogen production. Int J Hydrogen Energ 2010;35: doi: /j.ijhydene [64] Lyngfelt A. Oxygen carriers for chemical looping combustion h of operational experience. Oil Gas Sci Technol 2011;66: doi: /ogst/ [65] Lyngfelt A, Johansson M, Mattisson T. Chemical-looping combustion - status of development, [66] Zhou Z, Han L, Bollas G. Overview of chemical-looping reduction in fixed bed and fluidized bed reactors focused on oxygen carrier utilization and reactor efficiency. Aerosol Air Qual Res doi: /aaqr [67] Adánez J, Gayán P, Celaya J, de Diego LF, García-Labiano F, Abad A. Chemical looping combustion in a 10 kwth prototype using a CuO/Al2O3 Oxygen carrier: Effect of operating conditions on methane combustion. Ind Eng Chem Res 2006;45: doi: /ie060364l. [68] Jerndal E, Mattisson T, Lyngfelt A. Thermal analysis of chemical-looping combustion. Chem Eng Res Des 2006;84: doi: /cherd [69] Corbella BM, Palacios JM. Titania-supported iron oxide as oxygen carrier for chemicallooping combustion of methane. Fuel 2007;86: doi: /j.fuel [70] Mattisson T, Johansson M, Lyngfelt A. Multicycle reduction and oxidation of different types of iron oxide particles application to chemical-looping combustion. Energ Fuel 2004;18: doi: /ef [71] Fan L-S, Li F. Chemical looping technology and its fossil energy conversion applications. Ind Eng Chem Res 2010;49: doi: /ie

50 [72] Hossain MM, de Lasa HI. Chemical-looping combustion (CLC) for inherent separations a review. Chem Eng Sci 2008;63: doi: /j.ces [73] Jerndal E, Leion H, Axelsson L, Ekvall T, Hedberg M, Johansson K, et al. Using lowcost iron-based materials as oxygen carriers for chemical looping combustion. Oil Gas Sci Technol 2011;66: doi: /ogst/ [74] Mattisson T, Lyngfelt A, Cho P. The use of iron oxide as an oxygen carrier in chemicallooping combustion of methane with inherent separation of CO2. Fuel 2001;80: doi: /s (01) [75] Paul Cho TM. Comparison of iron-, nickel-, copper- and manganese-based oxygen carriers for chemical-looping combustion. Fuel n.d.: doi: /j.fuel [76] Abad A, García-Labiano F, de Diego LF, Gayán P, Adánez J. Reduction kinetics of Cu-, Ni-, and Fe-based oxygen carriers using syngas (CO + H2) for chemical-looping combustion. Energ Fuel 2007;21: doi: /ef070025k. [77] 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: doi: /j.cherd [78] Adánez J, Cuadrat A, Abad A, Gayán P, de Diego LF, García-Labiano F. Ilmenite activation during consecutive redox cycles in chemical-looping combustion. Energ Fuel 2010;24: doi: /ef900856d. [79] Bleeker MF, Veringa HJ, Kersten SRA. Deactivation of iron oxide used in the steamiron process to produce hydrogen. Appl Catal A-Gen 2009;357:5 17. doi: /j.apcata [80] Cuadrat A, Abad A, Adánez J, de Diego LF, García-Labiano F, Gayán P. Behavior of ilmenite as oxygen carrier in chemical-looping combustion. Fuel Process Technol 2012;94: doi: /j.fuproc [81] Li F, Sun Z, Luo S, Fan L-S. Ionic diffusion in the oxidation of iron effect of support and its implications to chemical looping applications. Energ Environ Sci 2011;4:876. doi: /c0ee00589d. 26

51 [82] Li F, Luo S, Sun Z, Bao X, Fan L-S. Role of metal oxide support in redox reactions of iron oxide for chemical looping applications: experiments and density functional theory calculations. Energ Environ Sci 2011;4:3661. doi: /c1ee01325d. [83] Galvita V, Hempel T, Lorenz H, Rihko-Struckmann LK, Sundmacher K. Deactivation of modified iron oxide materials in the cyclic water gas shift process for CO-free hydrogen production. Ind Eng Chem Res 2008;47: doi: /ie [84] Yamaguchi D, Tang L, Wong L, Burke N, Trimm D, Nguyen K, et al. Hydrogen production through methane steam cyclic redox processes with iron-based metal oxides. Int J Hydrogen Energ 2011;36: doi: /j.ijhydene [85] Kidambi PR, Cleeton JPE, Scott SA, Dennis JS, Bohn CD. Interaction of iron oxide with alumina in a composite oxygen carrier during the production of hydrogen by chemical Looping. Energ Fuel 2012;26: doi: /ef200859d. [86] Kierzkowska AM, Bohn CD, Scott SA, Cleeton JP, Dennis JS, Müller CR. Development of iron oxide carriers for chemical looping combustion using sol gel. Ind Eng Chem Res 2010;49: doi: /ie100046f. [87] Rydén M, Arjmand M. Continuous hydrogen production via the steam iron reaction by chemical looping in a circulating fluidized-bed reactor. Int J Hydrogen Energ 2012;37: doi: /j.ijhydene [88] He F, Li F. Perovskite promoted iron oxide for hybrid water-splitting and syngas generation with exceptional conversion. Energ Environ Sci 2015;8: doi: /c4ee03431g. [89] He F, Trainham J, Parsons G, Newman JS, Li F. A hybrid solar-redox scheme for liquid fuel and hydrogen coproduction. Energ Environ Sci 2014;7:2033. doi: /c4ee00038b. [90] Jin H, Ishida M. A new type of coal gas fueled chemical-looping combustion. Fuel 2004;83: doi: /j.fuel [91] Adánez J, García-Labiano F, de Diego LF, Gayán P, Celaya J, Abad A. Nickel Copper oxygen carriers to reach zero CO and H2 emissions in chemical-looping combustion. Ind Eng Chem Res 2006;45: doi: /ie050558l. 27

52 [92] Gayán P, Dueso C, Abad A, Adanez J, de Diego LF, García-Labiano F. NiO/Al2O3 oxygen carriers for chemical-looping combustion prepared by impregnation and deposition precipitation methods. Fuel 2009;88: doi: /j.fuel [93] Mattisson T, Järdnäs A, Lyngfelt A. Reactivity of some metal oxides supported on alumina with alternating methane and oxygen application for chemical-looping combustion. Energ Fuel 2003;17: doi: /ef020151i. [94] Moghtaderi B, Song H. Reduction properties of physically mixed metallic oxide oxygen carriers in chemical looping combustion. Energ Fuel 2010;24: doi: /ef [95] Ryu H-J, Lim N-Y, Bae D-H, Jin G-T. Carbon deposition characteristics and regenerative ability of oxygen carrier particles for chemical-looping combustion. Korean J Chem Eng 2003;20: doi: /bf [96] Villa R. Ni based mixed oxide materials for CH4 oxidation under redox cycle conditions. J Mol Catal A-Chem 2003; : doi: /s (03) [97] Shulman A, Linderholm C, Mattisson T, Lyngfelt A. High reactivity and mechanical durability of NiO/NiAl2O4 and NiO/NiAl2O4 /MgAl2O4 oxygen carrier particles used for more than 1000 h in a 10 kw CLC reactor. Ind Eng Chem Res 2009;48: doi: /ie900342f. [98] Erri P, Varma A. Diffusional effects in nickel oxide reduction kinetics. Ind Eng Chem Res 2009;48:4 6. doi: /ie071588m. [99] Blas L, Dorge S, Michelin L, Dutournié P, Lambert A, Chiche D, et al. Influence of the regeneration conditions on the performances and the microstructure modifications of NiO/NiAl2O4 for chemical looping combustion. Fuel 2015;153: doi: /j.fuel [100] Baek J-I, Yang S-R, Eom TH, Lee JB, Ryu CK. Effect of MgO addition on the physical properties and reactivity of the spray-dried oxygen carriers prepared with a high content of NiO and Al2O3. Fuel 2015;144: doi: /j.fuel

53 [101] Hamers HP, Gallucci F, Williams G, van Sint Annaland M. Experimental demonstration of CLC and the pressure effect in packed bed reactors using NiO/CaAl2O4 as oxygen carrier. Fuel 2015;159: doi: /j.fuel [102] Kooiman RF, Hamers HP, Gallucci F, van Sint Annaland M. Experimental demonstration of two-stage packed bed chemical-looping combustion using syngas with CuO/Al2O3 and NiO/CaAl2O4 as oxygen carriers. Ind Eng Chem Res 2015;54: doi: /ie [103] Hossain MM, Lopez D, Herrera J, de Lasa HI. Nickel on lanthanum-modified γ-al2o3 oxygen carrier for CLC: Reactivity and stability. Catal Today 2009;143: doi: /j.cattod [104] Ishida M, Jin H, Okamoto T. A Fundamental study of a new kind of medium material for chemical-looping combustion. Energ Fuel 1996;10: doi: /ef950173n. [105] Mattisson T, Johansson M, Lyngfelt A. The use of NiO as an oxygen carrier in chemicallooping combustion. Fuel 2006;85: doi: /j.fuel [106] Corbella BM, de Diego LF, García-Labiano F, Adánez J, Palacios JM. Performance in a fixed-bed reactor of titania-supported nickel oxide as oxygen carriers for the chemicallooping combustion of methane in multicycle tests. Ind Eng Chem Res 2006;45: doi: /ie050756c. [107] Siriwardane R, Poston J, Chaudhari K, Zinn A, Simonyi T, Robinson C. Chemicallooping combustion of simulated synthesis gas using nickel oxide oxygen carrier supported on bentonite. Energ Fuel 2007;21: doi: /ef [108] Zafar Q, Mattisson T, Gevert B. Redox investigation of some oxides of transition-state metals Ni, Cu, Fe, and Mn Supported on SiO2 and MgAl2O4. Energ Fuel 2006;20: doi: /ef [109] García-Labiano F, de Diego LF, Gayán P, Adánez J, Abad A, Dueso C. Effect of fuel gas composition in chemical-looping combustion with ni-based oxygen carriers. 1. Fate of sulfur. Ind Eng Chem Res 2009;48: doi: /ie801332z. [110] Tian H, Simonyi T, Poston J, Siriwardane R. Effect of hydrogen sulfide on chemical looping combustion of coal-derived synthesis gas over bentonite-supported metal oxide oxygen carriers. Ind Eng Chem Res 2009;48: doi: /ie900638p. 29

54 [111] Chuang S, Dennis J, Hayhurst A, Scott S. Development and performance of Cu-based oxygen carriers for chemical-looping combustion. Combust Flame 2008;154: doi: /j.combustflame [112] de Diego LF, Garcıá-Labiano F, Adánez J, Gayán P, Abad A, Corbella BM, et al. Development of Cu-based oxygen carriers for chemical-looping combustion. Fuel 2004;83: doi: /j.fuel [113] Chuang SY, Dennis JS, Hayhurst AN, Scott SA. Kinetics of the chemical looping oxidation of CO by a co-precipitated mixture of CuO and Al2O3. P Combust Inst 2009;32: doi: /j.proci [114] Gayán P, Forero CR, Abad A, de Diego LF, García-Labiano F, Adánez J. Effect of support on the behavior of Cu-based oxygen carriers during long-term CLC operation at temperatures above 1073 K. Energ Fuel 2011;25: doi: /ef101583w. [115] Arjmand M, Keller M, Leion H, Mattisson T, Lyngfelt A. Oxygen release and oxidation rates of MgAl2O4-supported CuO oxygen carrier for chemical-looping combustion with oxygen uncoupling (CLOU). Energ Fuel 2012;26: doi: /ef [116] 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: doi: /ef201329x. [117] Gayán 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: doi: /j.fuel [118] Zafar Q, Abad A, Mattisson T, Gevert B, Strand M. Reduction and oxidation kinetics of Mn3O4/Mg ZrO2 oxygen carrier particles for chemical-looping combustion. Chem Eng Sci 2007;62: doi: /j.ces [119] Roux S, Bensakhria A, Antonini G. Study and improvement of the regeneration of metallic oxides used as oxygen carriers for a new combustion process. Int J Chem React Eng 2006;4. doi: /

55 [120] Siriwardane R, Tian H, Richards G, Simonyi T, Poston J. Chemical-looping combustion of coal with metal oxide oxygen carriers. Energ Fuel 2009;23: doi: /ef [121] Zafar Q, Mattisson T, Gevert B. Integrated hydrogen and power production with CO2 capture using chemical-looping reformingredox reactivity of particles of CuO, Mn2O3, NiO, and Fe2O3 using SiO2 as a support. Ind Eng Chem Res 2005;44: doi: /ie048978i. [122] Johansson M, Mattisson T, Lyngfelt A. Investigation of Mn3O4 with stabilized ZrO2 for chemical-looping combustion. Chem Eng Res Des 2006;84: doi: /cherd [123] Jin H, Okamoto T, Ishida M. Development of a novel chemical-looping combustion: Synthesis of a looping material with a double metal oxide of CoO NiO. Energ Fuel 1998;12: doi: /ef980080g. [124] He F, Galinsky N, Li F. Chemical looping gasification of solid fuels using bimetallic oxygen carrier particles Feasibility assessment and process simulations. Int J Hydrogen Energ 2013;38: doi: /j.ijhydene [125] Shafiefarhood A, Stewart A, Li F. Iron-containing mixed-oxide composites as oxygen carriers for chemical looping with oxygen uncoupling (CLOU). Fuel 2015;139:1 10. doi: /j.fuel [126] Adánez-Rubio I, Arjmand M, Leion H, Gayán P, Abad A, Mattisson T, et al. Investigation of combined supports for Cu-based oxygen carriers for chemical-looping with oxygen uncoupling (CLOU). Energ Fuel 2013;27: doi: /ef401161s. [127] Moghtaderi B, Song H. Reduction properties of physically mixed metallic oxide oxygen carriers in chemical looping combustion. Energ Fuel 2010;24: doi: /ef [128] Johansson M, Mattisson T, Lyngfelt A. Creating a synergy effect by using mixed oxides of iron- and nickel oxides in the combustion of methane in a chemical-looping combustion reactor. Energ Fuel 2006;20: doi: /ef060068l. 31

56 [129] Bhalla AS, Guo R, Roy R. The perovskite structure - a review of its role in ceramic science and technology. Mater Res Innov 2000;4:3 26. doi: /s [130] Zhang K, Sunarso J, Shao Z, Zhou W, Sun C, Wang S, et al. Research progress and materials selection guidelines on mixed conducting perovskite-type ceramic membranes for oxygen production. RSC Adv 2011;1: doi: /c1ra00419k. [131] Readman JE, Olafsen A, Larring Y, Blom R. La0.8Sr0.2Co0.2Fe0.8O3 δ as a potential oxygen carrier in a chemical looping type reactor, an in-situ powder X-ray diffraction study. J Mater Chem 2005;15:1931. doi: /b416526h. [132] Dai XP, Li J, Fan JT, Wei WS, Xu J. Synthesis gas generation by chemical-looping reforming in a circulating fluidized bed reactor using perovskite LaFeO3-based oxygen carriers. Ind Eng Chem Res 2012;51: doi: /ie300033e. [133] Rydén M, Lyngfelt A, Mattisson T, Chen D, Holmen A, Bjørgum E. Novel oxygencarrier materials for chemical-looping combustion and chemical-looping reforming; LaxSr1 xfeyco1 yo3 δ perovskites and mixed-metal oxides of NiO, Fe2O3 and Mn3O4. Int J Greenh Gas Con 2008;2: doi: /s (07) [134] Zhao K, He F, Huang Z, Zheng A, Li H, Zhao Z. Three-dimensionally ordered macroporous LaFeO3 perovskites for chemical-looping steam reforming of methane. Int J Hydrogen Energ 2014;39: doi: /j.ijhydene [135] Arjmand M, Hedayati A, Azad A-M, Leion H, Rydén M, Mattisson T. CaxLa1 xmn1 ymyo3 δ (M = Mg, Ti, Fe, or Cu) as oxygen carriers for chemical-looping with oxygen uncoupling (CLOU). Energ Fuel doi: /ef [136] Arjmand M, Kooiman RF, Rydén M, Leion H, Mattisson T, Lyngfelt A. Sulfur tolerance of CaxMn1 ymyo3 δ (M = Mg, Ti) perovskite-type oxygen carriers in chemical-looping with oxygen uncoupling (CLOU). Energ Fuel 2014;28: doi: /ef402383v. [137] Cabello A, Abad A, Gayán P, de Diego LF, García-Labiano F, Adánez J. Effect of Operating conditions and H2S presence on the performance of CaMg0.1Mn0.9O3 δ perovskite material in chemical looping combustion (CLC). Energ Fuel 2014;28: doi: /ef [138] Hallberg P, Jing D, Rydén M, Mattisson T, Lyngfelt A. Chemical looping combustion and chemical looping with oxygen uncoupling experiments in a batch reactor using 32

57 spray-dried CaMn1 xmxo3 δ (M = Ti, Fe, Mg) particles as oxygen carriers. Energ Fuel 2013;27: doi: /ef [139] Hallberg P, Rydén M, Mattisson T, Lyngfelt A. CaMnO3-δ made from low cost material examined as oxygen carrier in chemical-looping combustion. Energ Procedia 2014;63:80 6. doi: /j.egypro [140] 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: doi: /ef900444d. [141] He F, Li X, Zhao K, Huang Z, Wei G, Li H. The use of La1 xsrxfeo3 perovskite-type oxides as oxygen carriers in chemical-looping reforming of methane. Fuel 2013;108: doi: /j.fuel [142] Källén M, Rydén M, Dueso C, Mattisson T, Lyngfelt A. CaMn0.9Mg0.1O3-δ as oxygen carrier in a gas-fired 10 kwth chemical-looping combustion unit. Ind Eng Chem Res 2013;52: doi: /ie303070h. [143] Pishahang M, Larring Y, McCann M, Bredesen R. Ca0.9Mn0.5Ti0.5O3 δ : A suitable oxygen carrier material for fixed-bed chemical looping combustion under syngas conditions. Ind Eng Chem Res 2014;53: doi: /ie500928m. [144] Pour NM, Azimi G, Leion H, Rydén M, Lyngfelt A. Production and examination of oxygen-carrier materials based on manganese ores and Ca(OH)2 in chemical looping with oxygen uncoupling. AIChE J 2014;60: doi: /aic [145] Rydén M, Lyngfelt A, Mattisson T. CaMn0.875Ti0.125O3 as oxygen carrier for chemicallooping combustion with oxygen uncoupling (CLOU) Experiments in a continuously operating fluidized-bed reactor system. Int J Greenh Gas Con 2011;5: doi: /j.ijggc [146] Sundqvist S, Leion H, Rydén M, Lyngfelt A, Mattisson T. CaMn0.875Ti0.125O3 δ as an oxygen carrier for chemical-looping with oxygen uncoupling (CLOU)-solid-fuel testing and sulfur interaction. Energ Technol 2013;1: doi: /ente [147] Imtiaz Q, Hosseini D, Müller CR. review of oxygen carriers for chemical looping with oxygen uncoupling (CLOU): Thermodynamics, material development, and synthesis. Energ Technol 2013;1: doi: /ente

58 [148] Bakken E, Norby T, Stolen S. Nonstoichiometry and reductive decomposition of CaMnO3. Solid State Ionics 2005;176: doi: /j.ssi

59 CHAPTER 2 Iron Oxide with Facilitated O 2- Transport for Facile Fuel Oxidation and CO2 Capture in a Chemical Looping Scheme Nathan L. Galinsky, Yan Huang, Arya Shafiefarhood, and Fanxing Li* Department of Chemical and Biomolecular Engineering, North Carolina State University, 911 Partners Way, Raleigh, NC , USA. Chapter 2 is a reprint of a manuscript published in ACS Sustainable Chemistry & Engineering, 2013, 1 (3), The supplementary information is in Appendix A. *To whom the correspondence should be addressed. Telephone: +1 (919) Fax: +1(919) fli5@ncsu.edu. 35

60 Abstract The chemical looping strategy offers a potentially viable option for efficient carbonaceous fuel conversion with a reduced carbon footprint. In the chemical looping process, an oxygen carrier is reduced and oxidized in a cyclic manner to convert a carbonaceous fuel into separate streams of concentrated carbon dioxide and carbon-free products such as electricity and/or hydrogen. The reactivity and chemical and physical stability of the oxygen carrier are of pivotal importance to chemical looping processes. A typical oxygen carrier is composed of a multivalence transition metal oxide supported on an inert support. Although the support does not get reduced or oxidized at any significant extent, numerous studies have indicated that certain supports such as TiO2 and Al2O3 can improve oxygen carrier stability and/or reactivity. This study reports the use of mixed ionic-electronic conductive support in iron-based oxygen carriers. By incorporating a perovskite based mixed-conductive support such as lanthanum strontium ferrite (LSF), the reactivity of the oxygen carrier is enhanced by 5-70 times when compared to oxygen carriers with conventional TiO2, Al2O3, or yttria stabilized zirconia (YSZ) support. The mixed-conductivity enhanced oxygen carrier also shows good stability and coke resistance. Characterization studies indicate that the enhanced oxygen carrier is composed of intermixed nano-scale (<100 nm) crystallites of iron oxide and support. The mixed-conductive support enables facile O 2- transport to and from the iron oxide nano-crystallites to participate in the surface redox reactions. The support also allows counter-current or concurrent diffusion of electrons or holes to maintain charge balance within the oxygen carrier. With iron oxide as the nano-scale oxygen source and mixed-conductive support as the oxygen/electron conductor, the mixed-conductivity enhanced oxygen carrier particle can be considered as an ensemble of nano-scale, mixed-conductive membrane reactors that possess excellent redox activity. 36

61 2.1 Introduction The chemical looping strategy represents a unique alternative to conventional combustion or gasification based approaches for power and/or hydrogen generation from carbonaceous fuels. 1 6 In chemical looping processes, carbonaceous fuels are converted by an oxygen carrier particle in two or more reduction and oxidation (redox) steps. These redox steps allow the oxygen carrier to shuttle oxygen from oxidants such as air or steam to the fuel in a cyclic manner. During the reduction step, the oxygen carrier donates oxygen to the fuel, resulting in a gaseous mixture with concentrated carbon dioxide and water vapor. The oxygen deprived oxygen carrier is then replenished in the regeneration or oxidation step(s) in which heat and/or hydrogen is produced. 1 3,7 11 Simplified reactions in the reduction and regenerations steps are given below: Reduction Step: CxHy (Fuel) + MeOm (oxygen carrier) MeOm-n + y/2h2o + xco2 Regeneration Step: MeOm-n (oxygen carrier) + H2O/O2 MeOm + H2/Heat (products) Since the oxygen carrier is essential to key chemical looping reactions, its selection and optimization is of utmost importance to the novel chemical looping process. An oxygen carrier is typically composed of a primary metal oxide and a support. 7,12 14 The primary metal oxide, often composed of oxides of iron, nickel, copper, manganese, and/or cobalt, functions as the active redox material for oxygen storage and donation purposes. 7,12,15 The support, on the other hand, usually does not directly participate in chemical looping reactions. The presence of support is found to improve the reactivity and/or long term stability of the oxygen carrier. 1 Besides supported transition metal oxides, mixed metal oxide based oxygen carriers such as iron containing perovskites and mixed Mn and Fe oxides are also 37

62 investigated as potential oxygen carrier candidates for chemical looping applications Although mixed metal oxides of Mn and Fe showed promising properties for chemical looping with oxygen uncoupling, they do not possess the thermodynamic properties required for hydrogen generation. 10,20 Perovskite materials, although potentially feasible for hydrogen generation 17,21, generally exhibit a low oxygen transport capacity. 4,16,17 Due to the lack of oxygen storage and donation capabilities, fuel conversion using perovskite based oxygen carriers is often limited. 4,16,17,22 The selection criteria for oxygen carriers have been discussed in a number of previous publications. 7,12,15 A rational strategy for oxygen carrier design and optimization, however, has yet to be developed because: (i) oxygen carrier design can be a challenging task considering the many desired physical and chemical properties, and (ii) there is a lack of fundamental understanding of the roles of the primary metal oxide, support, and their interactions. To date, more than 700 oxygen carriers obtained from various metal oxidesupport combinations and synthesis techniques have been reported. 1 A trial and error type of approach is often adopted in these studies. In an attempt to further understand the redox scheme of iron-based oxygen carriers, Li et al. recently investigated the effect of TiO2 support on the reactivity of iron oxide based oxygen carriers. Using both experiment and ab-initio calculation tools, they demonstrated that TiO2 support lowers the activation energy for oxygen anion (O 2- ) transfer through the metal oxide lattice. 23,24 The lowered activation energy for O 2- conduction and the substitution defects of O 2- created by support addition are considered to be the key contributors to the enhanced oxygen carrier performance after support addition. In this article, we attempt to rationalize the design strategy of iron-based oxygen carriers by incorporating mixed ionic-electronic conductive (MIEC) support. It is proposed that the MIEC support facilitates countercurrent conduction of O 2- and electrons, thereby allowing facile O 2-38

63 transport to and from the iron oxide irrespective to the porosity of the oxygen carrier particle. Although TiO2 supported Fe2O3 is considered as an active oxygen carrier in previous research, 1,15 the present study indicates that more than one order of magnitude reactivity improvement for methane conversion can be achieved by replacing the TiO2 with an MIEC support such as La0.8Sr0.2FeO3-δ (LSF). The use of LSF support instead of conventional, TiO2 based support has also led to 5-70 fold reactivity enhancements in oxygen carrier reactivity for converting a variety of fuels including H2, CO, and CH4. The LSF supported iron oxide exhibits excellent activity at a low temperature of 450 C, which is 250 C lower than the initial reduction temperature of TiO2 supported oxygen carrier. The stability of the LSF supported oxygen carrier is confirmed by a 50-cycle continuous redox experiment, which indicates stable phase composition, near constant crystalline sizes, and minimal decrease in reactivity. Coke formation is found to be minimal in all cases. Compared to oxygen carriers with other commonly used supports such as Al2O3 and YSZ, LSF supported Fe2O3 is also significantly more active. Scanning electron microscope (SEM), energy dispersive X-ray spectroscopy (EDX), and X-ray photoelectron spectroscopy (XPS) studies reveal that the oxygen carrier particles are composed of micron-sized dense grains. Each grain can be considered as an interpenetrated matrix of nano-scale LSF support and iron oxide crystallites. Although oxygen conductivity in iron oxides tends to be low, the nano-scale interaction between the MIEC support and iron oxide creates a solid-state O 2- and electron/hole pathway that allows full accessibility of active oxygen in iron oxide, even at a low surface area. The present study validates that MIEC support can significantly enhance the reactivity of iron oxide based oxygen carriers. Since a key limitation of iron oxide based oxygen carriers is their low 39

64 reactivity, 7,12,13,25 the MIEC supported oxygen carrier developed in the present study can potentially be effective for chemical looping combustion and gasification applications. 2.2 Experimental Oxygen Carrier Synthesis Fe2O3 supported with TiO2 and La0.8Sr0.2FeO3, both with 60 w.t.% Fe2O3 loading, are prepared. TiO2 supported Fe2O3 is used as a reference oxygen carrier since TiO2 supported iron oxide and ilmenite are frequently studied oxygen carriers with satisfactory reactivity A simple, one pot solid state reaction (SSR) method is used for oxygen carrier synthesis. The general procedure involves preparation of powder mixture followed by pelletization and annealing/solid state reaction. To prepare Fe2O3 supported TiO2 particles, calculated amounts of TiO2 (99.9%, Noah Chemicals) and Fe2O3 (99.9%, Noah chemicals) precursors are mixed for 1h in a planetary ball mill (XBM4X, Columbia International) at a rotation speed of 250 rpm. The mixture is then pressed into pellets by a hydraulic press (YLJ-15T, MTI Corporation) at the pressure of 6 MPa and sintered in a high temperature tube Furnace (GSL-1500-X50, MTI Corporation) at 1,200 C for 8h. Samples of Al2O3 (99% pure, Noah Chemical) and YSZ (99% pure, Noah Chemical) supported Fe2O3 are also prepared using the method identical to the TiO2 supported Fe2O3. Sintering time of 28 hours at 1,200 C was used to ensure comparable surface areas among the various oxygen carrier samples. Pure iron oxide, sintered for 8 hours, is also tested. The Fe2O3/La0.8Sr0.2FeO3 particles are prepared by mixing calculated amounts of La2O3 (99.9%, Aldrich), SrCO3 (99.9%, Noah Chemical) and Fe2O3 precursors for 3h in the planetary ball mill at a rotation speed of 250 rpm. The mixed powder is then pressed and sintered at 1,200 C. Two samples, one sintered for 8 hours and the other for 28 hours, were 40

65 prepared for comparison purpose. High pelletization pressure and sintering temperature are used to ensure sufficient solid state reaction and to obtain particles with comparable (low) porosity. The sintered pellets are finally fractured and sieved to arrive at particulates of various size ranges μm particles are used in all the reactivity studies since such a size range is suitable for chemical looping operations in a circulating fluidized bed reactor. 32, Reactivity Studies All the reactivity studies are conducted in a SETARAM SETSYS Evolution Thermal Gravimetric Analyzer (TGA). Initial studies are performed to determine the flow rate at which external mass transfer resistances are eliminated. This flow rate is determined by performing reduction experiments with varying reducing gas flow rate while keeping reaction temperature (900 C) and reducing gas composition (10% H2 balanced with He/N2) constant. The relationship between the reducing gas flow rate and the oxygen carrier reduction rate is recorded. At low gas flow rates, an increase in total gas flow rate generally leads to a higher oxygen carrier-h2 reaction rate. The external mass transfer limitations are considered to be minimal when no further rate enhancement is observed with increase in total gas velocity. Our experiments indicate that a total flow rate of 250 ml/min is adequate for eliminating the external mass transfer resistance. This corresponds to a linear gas velocity of approximately 1.2 m/s. Unless specified otherwise, the experiments are conducted at 900 C with a sample weight of approximately 20 mg. The total gas flow rate is maintained at around 300 ml/min. Three sets of redox experiments are conducted for both TiO2 supported and LSF supported oxygen carriers using hydrogen as the reducing gas. These experiments include one cycle reduction-oxidation (redox), five cycle redox, and fifty cycle redox. The one cycle test provides 41

66 information on initial redox activity of the oxygen carrier whereas the five and fifty cycle experiments determine short and long term recyclability of the oxygen carriers respectively. The experimental conditions for these experiments are listed in Table 2.1. Table 2.1. Experimental conditions for 1-cycle, 5-cycle, and 50-cycle redox experiments using hydrogen as the reducing gas. Test Sample Weight (mg) Reducing Gas Composition Oxidizing Gas Composition 1-cycle % H2 4.1% O2 5-cycle % H2 4.1% O2 50-cycle* 30 10% H2 10% O2 * 50-cycle experiment is carried out to determine particle stability under redox conditions, 30 mg of sample is used to obtain adequate sample for further characterizations. A higher O2 concentration is used for 50-cycle test to reduce the cycle time. To determine the redox contribution from the LSF support, pure La0.8Sr0.2FeO3 is tested for 5 hours in 10% H2 at 900 C. TiO2 support can also be reduced by H2 at high temperature; previous studies have indicated that the redox contribution from TiO2 tends to be low, 23,34 especially at the initial stage of the reduction. Besides hydrogen, CO and CH4 are also used as the reducing gases to determine the activity of the oxygen carriers and their resistance towards carbon deposition. 10% CH4 or CO balanced with inert gases are used in all the reducing experiments. Carbon analysis is performed on the solid samples using a CHN Elemental Analyzer (Perkin Elmer-2400). To determine the initial reduction temperature of the oxygen carriers, experiments are performed in the TGA on TiO2 and LSF supported oxygen carriers under a temperature ramping mode with 10% H2. The starting and ending points of the temperature ramp are 100 and 900 C respectively and the 42

67 temperature ramping rate is 5 C/min. A Quantachrome ChemBET Pulsar temperature programmed reduction/desorption (TPR/TPD) instrument is also used to determine the sample reducibility at various temperatures. 5% hydrogen mixed with argon is used as the reducing gas. Temperature ramping during the TPR experiment is 2 C/min Sample Characterizations The surface area, crystal structure, and surface and bulk elemental compositions of the oxygen carriers are analyzed using a variety of characterization tools. X-ray powder diffraction (XRD) test is carried out to analyze the crystal phase composition before and after reduction cycles. XRD patterns are recorded using a Rigaku SmartLab X-ray diffractometer with Cu-Kα (λ=0.1542) radiation operating at 40 kv and 44 ma. Scans are performed stepwise with 0.1 step size holding for 5 seconds at each step in angle range (2θ). The specific surface areas of the samples are measured with a BET surface analyzer (Quantachrome QuadraSorb Station 1) using krypton or nitrogen physisorption at 77.3 K. Scanning electron microscopy (SEM, Hitachi S3200) is used to observe the surface morphology and structure of the samples. An accelerating voltage of 5-20 kv is used for image capture. The same instrument is used for elemental analysis of the samples using energy dispersive X-ray spectroscopy (EDX) technique. Both the surface and the cross-section of the oxygen carriers are examined using both EDX line scan and spot analysis (10 kv). The surface compositions of the oxygen carrier are also analyzed using X-ray photoelectron spectroscopy (XPS, SPECS-XPS) with an Mg-Kα excitation (1254 ev) and a PHOIBIS 150 hemispherical analyzer at mbar pressure. Energy calibration for the XPS is established by referencing to adventitious Carbon (C1s line at ev binding energy). 43

68 2.3 Results and Discussion In order to achieve smooth operation of chemical looping processes, oxygen carriers are processed into particulates of 100 μm or larger. 10 The interactions between oxygen carrier particles and gaseous reactants typically proceed via a number of steps including diffusion of reactants from bulk phase to sample surface, intra-particle gaseous diffusion, surface reactions, and desorption and outwards diffusion of the reaction products, etc. 1 Recent studies indicated that active oxygen in the oxygen carrier may also diffuse to its surface, via solid-state ionicconduction, to participate in the redox reactions. 15,23 In order to accurately evaluate the effect of mixed conductivity on the redox reactivity of the oxygen carrier, the contributions from external mass transfer resistances and intra-particle gaseous diffusion need to be isolated. As discussed in the experimental section, external mass transfer resistance can be eliminated by performing the experiments at a high linear gas velocity. The effect of intra-particle gaseous diffusion can be offset by comparing oxygen carriers with similar porous structure but different mixed-conductivities. As mentioned earlier, TiO2 supported Fe2O3 is used as the reference oxygen carrier whereas LSF supported Fe2O3 is the oxygen carrier with improved mixedconductivity. In order to obtain oxygen carrier samples with comparable porous structure, a high sintering temperature of 1,200 C is used to eliminate most of the micro-pores and mesopores within the oxygen carrier particles. Surface analysis conducted using BET showed comparable low surface areas after sintering (See Supporting Information Table 2.S1) Reactivity with Hydrogen The reaction between the oxygen carrier and fuels represents a key step in all chemical looping processes. Hydrogen is selected as the fuel for initial assessment of the oxygen carrier reactivity since no side reactions are likely to take place between hydrogen and the oxygen 44

69 carrier samples. Therefore, reactivity with hydrogen can represent the intrinsic redox activity of the oxygen carrier. In addition, hydrogen is a commonly encountered fuel that is present in synthesis gas derived from various carbonaceous feedstocks. The TGA curves for oxygen carrier reduction with 10.5% H2 are illustrated in Figure 2.1. The TGA records the weight loss of the oxygen carrier during its reduction reactions. Since the weight loss of the solid sample directly corresponds to its extent of reduction, the points on the TGA curves in Figure 2.1 can be converted into oxygen carrier conversion via Equation 1: % conversion = (m initial m i ) m initial x O2 100% (1) where minitial is the initial mass of the oxygen carrier, mi is the mass at the time of interest during the reduction step, xo2 is the weight percent of active oxygen in the unreacted oxygen carrier. In the current study, the support materials, i.e. TiO2 or LSF, are considered as inert whereas all the oxygen atoms in Fe2O3 are considered active. Since 60% Fe2O3 is used in all the oxygen carriers, xo2 is ~18% in all cases. It should be noted that many previous studies only consider 11% of the oxygen in Fe2O3 as active oxygen (Fe2O3 Fe3O4). This is different from that in the present study which considers full reduction of Fe2O3 to Fe. It is also noted that both TiO2 and LSF can be reduced at a high temperature under a hydrogen environment. The contribution of TiO2 reduction to the overall reduction kinetics of the oxygen carrier; however, has been determined to be minimal. 23,34 Reduction experiments using pure LSF indicates that the contribution of LSF support to the oxygen carrier reduction kinetics is within 2% of the total conversion (see Figure 2.S1 of Supporting Information). Therefore, LSF mainly functions as a redox reaction promoter as opposed to a reactant. 45

70 Figure 2.1. Reduction of as-prepared LSF and TiO2 supported iron oxide particles in a TGA apparatus (10% H2 balanced with N2/He; 900 C; Reference oxygen carrier refers to TiO2 supported iron oxide) The reaction rates of the TiO2 supported and LSF supported iron oxide samples are compared using three quantitative parameters including initial reaction rate, the time required to reach 11% conversion, and the time required to achieve 33% conversion. Initial reaction rate (%conversion/sec) reflects the intrinsic reactivity between oxygen carrier surface region and reducing gas. The time required (min) to reach 11% and 33% conversions provides indication on the overall activity of the oxygen carrier. 11% and 33% conversions correspond to iron oxide reduction to Fe3O4 and FeO, respectively. The rate comparisons for the oxygen carriers with the two different supports are given in Figure 2.2. As can be seen, the initial rate of the LSF supported oxygen carrier is approximately 5 times higher than that of the oxygen carrier with TiO2 support. To reach conversions of 11% and 33%, the TiO2 supported oxygen carrier requires approximately 6 times longer reaction time. Beyond 33% conversion, the reduction rate of TiO2 supported oxygen carrier slows down significantly faster than LSF supported oxygen carrier. As indicated in Figure 2.1, LSF supported oxygen carrier shows complete 46

71 oxygen donation from iron oxide within a much shorter time-frame, indicating that the active oxygen atoms in iron oxide can effectively react with hydrogen even at a very low surface area. The continued weight loss beyond 18 w.t.% oxygen donation mainly results from the slow reduction of the LSF support. To determine the temperature dependence of the oxygen carrier redox reactivity, the two aforementioned oxygen carriers are reduced in hydrogen under a temperature ramp in both TGA and TPR. As illustrated in Figure 2.3, the initial reduction temperature of the LSF supported Fe2O3 is roughly 250 C lower than TiO2 supported Fe2O3. The significant decrease in initial reduction temperature further confirms the superior reactivity of the oxygen carrier enhanced by the mixed-conductive support. The TPR results of TiO2- supported Fe2O3 shown in Figure 2.3(b) are qualitatively consistent with those obtained by previous researchers on impregnated TiO2-Fe2O3 oxygen carriers using 10% H2 as the reducing gas. 26 In contrast, the LSF supported Fe2O3 exhibits its major reduction peak at a temperature much lower than the typical reduction temperature of β-oxygen in LSF. 17 This is likely to result from the synergistic effect of LSF and Fe2O3 coupled with the relatively low reducible oxygen content in the LSF support. 47

72 Figure 2.2. Reactivity comparisons of as-prepared TiO2 supported Fe2O3 (reference oxygen carrier) and LSF supported Fe2O3 (oxygen carrier with enhanced mixedconductivity); reducing gas: 10.5% hydrogen, temperature: 900 C. Figure 2.3. Initial Temperature for Reduction of Fe2O3 supported on LSF and TiO2 in TGA (a) and H2-TPR (b) Previous studies have indicated that ilmenite and TiO2-supported iron oxide particles can be activated during the first few redox cycles due to its morphology and porosity change. 24,25 Figure 2.4 compares the reactivity of the TiO2 and LSF supported iron oxide over five redox 48

73 cycles. Consistent with previous studies, the reactivity of the TiO2 supported oxygen carrier significantly increases over the first 3 redox cycles and then stabilizes. To compare, reactivity of the LSF supported oxygen carrier is much more stable. Even at the 5 th cycle, the LSF supported oxygen carrier is still five times more reactive than the TiO2 supported oxygen carrier for hydrogen conversion. Reactivity studies on porous oxygen carriers obtained using starch template also indicate that LSF supported oxygen carrier is far more reactive than the oxygen carrier with the TiO2 support (see Figure 2.S4 in Supporting Information). Figure 2.4. Time to achieve 60% conversion over the first 5 cycles using 10% H Long Term Stability of the LSF Supported Oxygen Carrier The ability of an oxygen carrier to maintain a high reactivity and phase stability over multiple redox cycles at high temperatures is highly desirable for chemical looping applications. With LSF supported oxygen carrier exhibiting promising reactivity, studies are carried out to investigate its long term stability. The TGA curves of the LSF supported oxygen carrier over 50 redox cycles are provided in Figure 2.5. As indicated by Figure 2.5, the LSF supported oxygen carrier shows less than 4% decay in reactivity for both reduction and oxidation 49

74 reactions over the 50 redox cycles. During the experiment, iron oxide is reduced to iron phase in all reduction cycles. An insignificant extent of sintering is observed on the sample, as manifested by loose attachment of particulates. Sintering may account for the small loss in reactivity over multiple redox cycles. It is likely to result from the complete reduction of Fe2O3 to iron and the subsequent combustion reaction. Combustion of metallic iron with oxygen is highly exothermic and can lead to melting of iron, which has a much lower melting point than iron oxides. Sintering can be completely avoidable in actual chemical looping combustion or gasification operations since the oxygen carrier will be in the form of Fe3O4/FeO prior to combustion. 10 Figure 2.5. Cyclic redox experiment of (a) La0.8Sr0.2FeO3-δ supported Fe2O3 in reduction cycles; (b) La0.8Sr0.2FeO3-δ supported Fe2O3 in oxidation cycles. Reducing gas: 10% hydrogen, temperature: 900 C; oxidizing gas: 10% oxygen, temperature: 900 C. XRD analysis is performed to investigate the phase stability of the oxygen carrier. Figure 2.6 compares the dominant phases at the 1 st reduction and oxidation as well as the 51 st reduction and oxidation. The sizes of the key crystalline phases are given in Table 2.2. The crystalline sizes are estimated using the Williamson-Hall method. A significant change in 50

75 crystalline size is observed during the first redox cycle, which can be attributed to the solid state reactions promoted by the redox reaction, leading to the stabilization of the crystalline phases. In fact, a noticeable amount of strontium ferrite phase co-exists with perovskite and iron oxide phases in the fresh oxygen carrier, resulting from the incomplete solid state reaction during the sample preparation/annealing step. This phase impurity is almost completely incorporated to the LSF structure after the first redox cycle (Figure 2.6). Table 2.2 also indicates that the crystalline sizes within the oxygen carrier remained relatively stable from the first redox cycle onwards. This confirms the observations of insignificant sintering and oxygen carrier deactivation. XRD analysis on the Al2O3 and YSZ supported samples show two primary phases: Fe2O3 and the support phase (see ESI Figure 2.S7). Figure 2.6. XRD spectra of LSF supported Fe2O3 in oxidized (a) and reduced (b) forms over 51 reduction-oxidation cycles. (the small unlabelled peaks in 1 st cycle samples can be attributed to the insignificant amount of unconverted strontium ferrite from solid state reactions) 51

76 Table 2.2. Crystalline size change during multi-cyclic redox reactions of the LSF supported Fe2O3 Crystalline Size in Fresh Oxygen Carrier (Å) Crystalline Size after 1 st oxidation(å) Crystalline Size after 50 th oxidation(å) Fe2O La0.8Sr0.2FeO Reaction with Gaseous Carbonaceous Fuels An important feature of the chemical looping strategy is its ability for carbonaceous fuel conversion with in-situ CO2 capture. Syngas and natural gas are among the most investigated gaseous carbonaceous fuels in chemical looping applications. This section reports the conversion of CO and CH4. CO is a major component in syngas besides H2 whereas CH4 is a key component of natural gas. The reactivity comparisons of the LSF supported and TiO2 supported oxygen carriers are given in Figure 2.7. In terms of initial reaction rate with CO, Figure 2.7 indicates that LSF supported oxygen carrier is an order of magnitude more reactive than the reference, TiO2 supported oxygen carrier. The LSF enhanced oxygen carrier is approximately 40 times faster for achieving 33% conversion when CO is used as the fuel. An even more significant contrast is observed in terms of the oxygen carrier reactivity for CH4 conversion, as shown in figure 2.8. For comparison purpose, the CH4 oxidation activities of YSZ and Al2O3 supported oxygen carriers are also included in Figure 2.8. Although the formation of mixed Fe-Al oxides were reported to reduce the activity of Al2O3 supported Fe2O3 in some previous studies, XRD analysis in the current study indicates distinct support and iron oxide phases for both Al2O3 and YSZ supported Fe2O3 (see Supporting Information Figure 2.S7). Thus, the formation of mixed metal oxide is insignificant. In terms of the initial rate for CH4 oxidation, the LSF supported oxygen carrier approximately 8 times higher than the 52

77 reference material, 2 times higher than the Al2O3 supported oxygen carrier, and 3.5 times higher than the YSZ supported oxygen carrier. The LSF enhanced oxygen carrier is 64 times faster to achieve 11% conversion than the reference material, 9 times faster than the Al2O3 supported oxygen carrier, 54 times that of the YSZ supported oxygen carrier, and 79 times faster than pure Fe2O3. A more pronounced reactivity disparity is observed for oxygen carrier conversion at above 11%. In fact, the TiO2 and YSZ supported iron oxide as well as pure iron oxide did not reach 33% conversion during the 45 minutes (shown in grey in Figure 2.8) testing period. The Al2O3 supported sample achieved only 45% conversion within the 45 minutes reduction period. In contrast, LSF supported oxygen carrier is able to reach 100% conversion within a short duration. Although the excellent CH4 conversion activity of the LSF supported oxygen carrier may partially result from the surface catalytic activity of the perovskite (LSF) support, the main contributor should be the mixed-conductivity of LSF considering the relatively large crystallite sizes of both iron oxide and LSF support (Table 2.2). Unlike typical methane combustion/gasification reactions, gaseous oxygen is not used in chemical looping combustion of methane. Therefore, transfer of active O 2- from the solid primary oxide lattice to the oxygen carrier surface is likely to be the rate limiting step for methane oxidation. The ability of LSF to shuttle O 2- and electron at a high conductivity allows effective combustion of methane and other fuels. The importance of mixed-conductivity is further confirmed by the high activity of LSF supported oxygen carrier at low surface area. For instance, LSF supported oxygen carrier sintered for 28 hours shows slightly higher activity when compared to the oxygen carrier sintered for 8 hours (see Supporting Information Figure 2.S3). Therefore, a low surface area did not adversely affect the activity of the oxygen carrier. The likely reason for a slightly higher activity is the more complete solid state reaction for the formation of more 53

78 homogeneous LSF phase. It is also worthwhile to point out that, although YSZ is a good O 2- conductor, it is a poor electronic conductor and thus cannot shuttle O 2- to and from the primary iron oxide. This explains the poor activity of YSZ when compared to LSF supported oxygen carrier even though both YSZ and LSF are O 2- -conductive materials. Figure 2.7. Reactivity comparisons of the reference and LSF supported oxygen carriers in 10% CO at 900 C Figure 2.8. Reactivity comparisons of synthesized oxygen carriers in 10% CH4 showing (a) time to achieve 11% conversion and (b) time to achieve 33% conversion. 54

79 Besides being active for carbonaceous fuel conversion, an effective oxygen carrier should also be resistant towards coke formation, which may occur from a number of side reactions including the Boudouard reaction, methane decomposition, etc. 1,10,35 Coke formation data after 2 hours of reaction between CO and oxygen carriers is given in Table 2.3. As can be seen, the LSF supported oxygen carrier is highly resistant towards coking formation through the Boudouard reaction. Merely 600 ppm (by weight) of carbon is observed after the 2 hour reaction with the completely reduced oxygen carrier. In many previous studies, significant carbon deposition tends to occur upon near complete reduction of the oxygen carrier, since reduced metal tends to catalyze the Boudouard reaction. 10,35 A lower carbon formation is observed for TiO2 supported oxygen carrier since it is only reduced by 45% after the two hour duration. The good resistance towards carbon formation is likely to result from the mixed conductive support which provides high mobility lattice oxygen for in-situ coke removal. In contrast to the insignificant coke formation observed when reacting with CO, significant coke formation (nearly 20% by weight) is identified on LSF supported oxygen carrier when CH4 is used as the fuel in a 2-hour reduction experiment. In order to investigate carbon deposition during the reduction stage, the LSF supported oxygen carrier is reacted with CH4 for 20 minutes prior to carbon analysis. The 20 minutes reduction time allows 100% oxygen carrier reduction while avoiding overexposure of the fully reduced oxygen carrier to CH4. The coke content in the resultant solid sample is merely 400 ppm by weight. This indicates that significant coke formation in CH4 conversion only takes place after complete depletion of active oxygen in the oxygen carrier. From a thermodynamics standpoint, the active oxygen in the oxygen carrier retards coke formation. Therefore, depletion of the same will favor coke 55

80 formation on the oxygen carrier. In addition, transition metals including iron, which is formed upon reduction, can catalyze methane decomposition reactions. The fact that insignificant coke deposition occurred before 100% oxygen carrier reduction indicates that the LSF supported oxygen carrier can be suitable for chemical looping applications. Prior studies have indicated that the maximum oxygen carrier conversion for iron oxide based oxygen carrier in a practical chemical looping process is limited to around 50%. 36 In fact, a low conversion of less than 11% (based on the current definition) is often maintained in chemical looping combustion processes. 1 The low oxygen carrier conversion in actual chemical looping processes is likely to completely retard coke formation on the LSF supported oxygen carrier. It is noted that TiO2 supported oxygen carrier exhibit low (0.3% by weight) coke formation after 2 hour reaction with CH4; this is attributed to the slow reduction kinetics. In fact, the TiO2 supported oxygen carrier is only reduced by 25% after a 2-hour duration. Regeneration of the reduced oxygen carrier is also performed. Coke formed on the LSF supported oxygen carriers is found to be completely removed from the regeneration step. Table 2.3. Carbon Deposition (% wt.) results when using CO and CH4 as reducing gases. Material Reduced for 2 hours in 10% CO Reduced for 2 hours in 10% CH4 Reduced for 20 minutes in 10% CH4 Fe2O3 supported on La0.8Sr0.2FeO3 Fe2O3 supported on TiO

81 2.3.4 Morphology Studies and Reaction Scheme To obtain further understanding of the potential effect of the LSF supported oxygen carrier, the oxygen carrier samples are characterized using a number of approaches including SEM/EDX, XRD, and XPS. As revealed by XRD analysis (Table 2.2 and Figure 2.6), the primary phases in oxidized oxygen carriers are Fe2O3 and LSF with crystalline sizes of around 50 nm. SEM images of the oxygen carriers as well as EDX spectra of several sample locations are illustrated in Figure 2.9 and Table 2.4. Aluminum and carbon in the EDX spectra results from impurities introduced during the sample preparation stage in which a powder form of the LSF supported Fe2O3 is spread across a piece of carbon tape placed on an aluminum sample holder. The SEM images indicate that the oxygen carrier is composed of dense grains of roughly 1-2 μm. At accelerating voltage of 10 kv (special resolutions of ~0.5 μm), the elemental compositions obtained by EDX analysis at the various sample locations do not exhibit the presence of either pure LSF or Fe2O3 even though local enrichments of either LSF or Fe2O3 are identified (See Table 2.4). EDX line scan of La, Sr, and Fe reveal continuous distribution of these elements (See Figure 2.S6 of Supporting Information). We also observed some deviation from the stoichiometric La: Sr ratio at selected locations in the fresh sample. This is in accordance with the presence strontium ferrite observed under XRD, which leads to local enrichment of Sr compared to La. The strontium ferrite phase is fractionated into perovskite and ferrite phases after the first redox cycle. Based on the XRD and SEM/EDX results, it can be concluded that the Fe2O3 crystallites are finely divided by LSF at nano-scale. XPS analysis is carried out to characterize the surface composition of the oxygen carrier. The XPS spectra of the LSF supported oxygen carrier surface is shown in Figure The corresponding elemental composition of the surface is provided in Table 2.5. As indicated in 57

82 Table 2.5, compared to the bulk LSF and Fe2O3 compositions, the oxygen carrier surface is enriched with LSF by approximately 45%. The surface enrichment of LSF is desired since it reduces the surface exposure of Fe2O3, which undergoes structural and phase changes over the redox cycles. Encapsulation of Fe2O3 can potentially lead to enhanced structural stability of the oxygen carrier. Although the mechanism for surface enrichment of LSF is subject to further studies, the high surface LSF concentration indicates that the redox reactions are likely to occur on the LSF surface, which provides active lattice oxygen conducted from the bulk phase. It is also noted that the oxygen concentration on the surface is slightly higher than that in the bulk. This indicates that the preferred termination surface is likely to contain higher oxygen concentration. The result is consistent with Density Functional Theory (DFT) calculations performed by Mastrikov et al on relevant perovskites. 37 Based on the morphology of the oxygen carrier, a reaction scheme for the redox reaction of the oxygen carrier is illustrated in Figure As can be seen, the unique oxygen carrier is composed of nano-scale crystallites of (i) iron oxide as primary oxygen carrying material for lattice oxygen (O 2- ) storage; (ii) LSF as MIEC support which facilities O 2- and electron/hole conduction. To participate in the fuel oxidation reactions, active O 2- is shuttled from the primary oxide to the oxygen carrier surface through the support. In the meantime, the electrons released by O 2- on the surface are conducted back to the primary oxide to maintain charge balance and to facilitate primary oxide reduction. With primary oxide providing O 2- and driving force (μo2) for oxygen conduction in nano-crystallites of MIEC, each redox catalyst particle can be considered as a large collection of MIEC membrane reactors at nano-scale. 4 Although Fe2O3 is not a good conductor for both electrons and oxygen anion, the very small length scale for O 2- diffusion in the Fe2O3 nano-crystallites ensures easy accessibility of 58

83 oxygen. As confirmed by the experimental results, nearly 100% of the O 2- stored in the primary oxide is accessible to the redox reactions. Compared to conventional MIEC membrane reactors, the redox catalyst can be more robust and flexible since a dense membrane layer is unnecessary. The abovementioned mechanism is consistent with the authors previous DFT and experimental studies on supported oxygen carriers. 23,24 Figure 2.9. SEM Image (20 kv) of LSF supported Fe2O3 of (a) fresh particles and (b) after 51 oxidations and two locations where EDX point analyses are performed. Table 2.4. EDX results of the LSF supported oxygen carriers at the points represented in Figure 2.8 (b) taken at 10.0kV for enhanced special resolution. Point 1 (atomic percentage) Point 2 (atomic percentage) O 64.3% 54.52% Fe 27.2% 41.14% La 6.1% 1.71% Sr 2.3% 2.63% 59

84 Figure Surface composition analysis using XPS of the Fe2O3 supported on La0.8Sr0.2FeO3 Table 2.5. Elemental Composition comparison of the Fe2O3 supported on LSF oxygen carrier of the bulk compared to the surface. Carbon content in the XPS result is removed since it is introduced for binding energy calibration. Bulk atomic composition Surface atomic composition O 60.00% 72.83% Fe 33.72% 19.35% La 5.03% 6.39% Sr 1.25% 1.42% 60

85 Figure Proposed reduction reaction scheme of the LSF supported Fe2O3 oxygen carrier. 2.4 Conclusions The present study investigates the use of mixed-conductive support such as lanthanum strontium ferrite (LSF) to enhance the performance of iron-based oxygen carriers. Using a simple solid-state synthesis method, oxygen carriers composed of interpenetrated nanocrystallites of iron oxide and LSF support are obtained. The LSF enhanced oxygen carrier exhibits more than an order of magnitude higher reactivity when compared to a reference, TiO2 supported oxygen carrier as well as Al2O3 and YSZ supported oxygen carriers. The mixconductive support enhanced oxygen carrier also becomes active for redox reactions at a much lower temperature. In addition, the improved oxygen carrier shows excellent recyclability and good resistance towards coke formation for CO and CH4 conversion. The significantly enhanced redox performance of the LSF-supported oxygen carrier is likely to result from the facile lattice oxygen and electron/hole pathways created by the mixed conductive support. To be more specific, the mixed-conductive support enables fast solid state O 2- and electron/hole exchange between the iron oxide nano-crystallites and the oxygen carrier particle surface, 61

86 thereby allowing effective oxygen removal and restoration of the oxygen carrier during the redox reactions while maintaining charge balance within the oxygen carrier. By embedding and stabilizing iron oxide nano-crystallites of less than 100 μm in a matrix of nano-sized mixed-conductive support, all active lattice oxygen within the iron oxide are made accessible for fast redox reactions via solid state conduction. The findings from the present research can lead to a potentially effective approach for rational optimization of oxygen carriers for chemical looping applications. These findings indicate that the incorporation of mixedconductive and structurally compatible support with an oxygen carrying primary metal oxide can lead to high performance oxygen carriers with good reactivity, robustness, and sintering resistance. The resulting oxygen carrier particle can be considered as an ensemble of nanoscale mixed-conductive membrane reactors in which primary metal oxide is used as oxygen source or sink in the redox reactions of the chemical looping process. Acknowledgment This work was supported by the North Carolina State University Start-Up Fund. We would like to acknowledge Feng He and Kit Yeung in the set-up of the experimental apparatus. Supporting Information Additional information regarding synthesis and characterization procedures and results are available. This information is available free of charge via the Internet at 62

87 REFERENCES (1) Adanez, J.; Abad, A.; Garcia-Labiano, F.; Gayan, P.; de Diego, L. F. Progress in Chemical-Looping Combustion and Reforming technologies. Prog. Energ. Combust. 2012, 38, (2) Shah, K.; Moghtaderi, B.; Wall, T. Selection of Suitable Oxygen Carriers for Chemical Looping Air Separation: A Thermodynamic Approach. Energ. Fuel. 2012, 26, (3) Hossain, M. M.; de Lasa, H. I. Chemical-looping combustion (CLC) for inherent separations a review. Chem. Eng. Sci. 2008, 63, (4) Thursfield, A.; Murugan, A.; Franca, R.; Metcalfe, I. S. Chemical looping and oxygen permeable ceramic membranes for hydrogen production a review. Energ. Environ. Sci. 2012, 5, (5) Li, F.; Fan, L.-S. Clean coal conversion processes progress and challenges. Energ. Environ. Sci. 2008, 1, 248. (6) Ishida, M.; Jin, H. A new advanced power-generation system using chemical-looping combustion. Energy 1994, 19, (7) Kang, K.-S.; Kim, C.-H.; Bae, K.-K.; Cho, W.-C.; Kim, S.-H.; Park, C.-S. Oxygencarrier selection and thermal analysis of the chemical-looping process for hydrogen production. Int. J. Hydrogen Energ. 2010, 35, (8) Ishida, M.; Jin, H. CO2 recovery in a power plant with chemical looping combustion. Energ. Convers. and Manage. 1997, 38, Supplement, S187 S192. (9) Lyngfelt, A. Oxygen Carriers for Chemical Looping Combustion h of Operational Experience. Oil Gas Sci. Technol.2011, 66, (10) Fan, L.-S. Chemical Looping Systems for Fossil Energy Conversions;

88 (11) Svoboda, K.; Slowinski, G.; Rogut, J.; Baxter, D. Thermodynamic possibilities and constraints for pure hydrogen production by iron based chemical looping process at lower temperatures. Energ. Convers. and Manage.2007, 48, (12) Adánez, J.; de Diego, L. F.; García-Labiano, F.; Gayán, P.; Abad, A.; Palacios, J. M. Selection of Oxygen Carriers for Chemical-Looping Combustion. Energ. Fuel. 2004, 18, (13) Lyngfelt, A.; Johansson, M.; Mattisson, T. Chemical-looping combustion - status of development. In; Proceedings of the 9 th International Conference on Circulating Fluidized Beds, Hamburg, Germany (14) Rydén, M.; Johansson, M.; Lyngfelt, A.; Mattisson, T. NiO supported on Mg ZrO2 as oxygen carrier for chemical-looping combustion and chemical-looping reforming. Energ. Environ. Sci.2009, 2, 970. (15) Li, F.; Kim, H. R.; Sridhar, D.; Wang, F.; Zeng, L.; Chen, J.; Fan, L.-S. Syngas Chemical Looping Gasification Process: Oxygen Carrier Particle Selection and Performance. Energ. Fuel. 2009, 23, (16) Rydén, M.; Lyngfelt, A.; Mattisson, T.; Chen, D.; Holmen, A.; Bjørgum, E. Novel oxygen-carrier materials for chemical-looping combustion and chemical-looping reforming; LaxSr1-xFeyCo1-yO3 δ perovskites and mixed-metal oxides of NiO, Fe2O3 and Mn3O4. Int. J. Greenh. Gas Con.2008, 2, (17) Murugan, A.; Thursfield, A.; Metcalfe, I. S. A chemical looping process for hydrogen production using iron-containing perovskites. Energ. Environ. Sci. 2011, 4, (18) Sarshar, Z.; Kleitz, F.; Kaliaguine, S. Novel oxygen carriers for chemical looping combustion: La1 xcexbo3 (B = Co, Mn) perovskites synthesized by reactive grinding and nanocasting. Energ. Environ. Sci.2011, 4, (19)Azimi, G.; Rydén, M.; Leion, H.; Mattisson, T.; Lyngfelt, A. (MnzFe1-z)yOx combined oxides as oxygen carrier for chemical-looping with oxygen uncoupling. AIChE J. 2012, n/a n/a. (20) Fan, L.-S.; Zeng, L.; Wang, W.; Luo, S. Chemical looping processes for CO2 capture and carbonaceous fuel conversion prospect and opportunity. Energ. Environ. Sci. 2012, 5,

89 (21) Bouwmeester, H. J. M. Dense ceramic membranes for methane conversion. Catal. Today 2003, 82, (22) Zhang, K.; Sunarso, J.; Shao, Z.; Zhou, W.; Sun, C.; Wang, S.; Liu, S. Research progress and materials selection guidelines on mixed conducting perovskite-type ceramic membranes for oxygen production. RSC Adv. 2011, 1, (23) Li, F.; Luo, S.; Sun, Z.; Bao, X.; Fan, L.-S. Role of metal oxide support in redox reactions of iron oxide for chemical looping applications: experiments and density functional theory calculations. Energ. Environ. Sci. 2011, 4, (24) Li, F.; Sun, Z.; Luo, S.; Fan, L.-S. Ionic diffusion in the oxidation of iron effect of support and its implications to chemical looping applications. Energ. Environ. Sci.2011, 4, 876. (25) Mattisson, T.; Lyngfelt, A.; Cho, P. The use of iron oxide as an oxygen carrier in chemicallooping combustion of methane with inherent separation of CO2. Fuel 2001, 80, (26) Corbella, B. M.; Palacios, J. M. Titania-supported iron oxide as oxygen carrier for chemical-looping combustion of methane. Fuel 2007, 86, (27) Jerndal, E.; Leion, H.; Axelsson, L.; Ekvall, T.; Hedberg, M.; Johansson, K.; Källén, M.; Svensson, R.; Mattisson, T.; Lyngfelt, A. Using Low-Cost Iron-Based Materials as Oxygen Carriers for Chemical Looping Combustion. Oil Gas Sci. Technol. 2011, 66, (28) Leion, H.; Mattisson, T.; Lyngfelt, A. Use of Ores and Industrial Products As Oxygen Carriers in Chemical-Looping Combustion. Energ. Fuel. 2009, 23, (29) Mattisson, T.; Johansson, M.; Lyngfelt, A. Multicycle Reduction and Oxidation of Different Types of Iron Oxide Particles Application to Chemical-Looping Combustion. Energ. Fuel. 2004, 18, (30) Roux, S.; Bensakhria, A.; Antonini, G. Study and Improvement of the Regeneration of Metallic Oxides Used as Oxygen Carriers for a New Combustion Process. Int. J. Chem. React. Eng.2006, 4. 65

90 (31) Adanez, J.; García-Labiano, F.; de Diego, L. F.; Gayán, P.; Celaya, J.; Abad, A. - Characterization of oxygen carriers for chemical-looping combustion. In Greenhouse Gas Control Technologies 7; Elsevier Science Ltd: Oxford, 2005; pp (32) Kolbitsch, P.; Pröll, T.; Hofbauer, H. Modeling of a 120 kw chemical looping combustion reactor system using a Ni-based oxygen carrier. Chem. Eng. Sci. 2009, 64, (33) Pröll, T.; Mayer, K.; Bolhàr-Nordenkampf, J.; Kolbitsch, P.; Mattisson, T.; Lyngfelt, A.; Hofbauer, H. Natural minerals as oxygen carriers for chemical looping combustion in a dual circulating fluidized bed system. Energy Procedia 2009, 1, (34) Zhao, Y.; Shadman, F. Reduction of ilmenite with hydrogen. Ind. Eng. Chem. Res. 1991, 30, (35) Rohmund, F.; Falk, L. K. L.; Campbell, E. E. B. Iron catalyzed growth of carbon nanotubes. AIP Conference Proceedings 2000, 544, (36) Li, F.; Zeng, L.; Velazquez-Vargas, L. G.; Yoscovits, Z.; Fan, L.-S. Syngas chemical looping gasification process: Bench-scale studies and reactor simulations. AIChE J. 2010, 56, (37) Mastrikov, Y. A.; Merkle, R.; Heifets, E.; Kotomin, E. A.; Maier, J. Pathways for Oxygen Incorporation in Mixed Conducting Perovskites: A DFT-Based Mechanistic Analysis for (La, Sr)MnO3 δ. J. Phys. Chem. C 2010, 114,

91 CHAPTER 3 Effect of Support on Redox Stability of Iron Oxide for Chemical Looping Conversion of Methane Nathan L. Galinsky, Arya Shafiefarhood, Yanguang Chen, Luke Neal, and Fanxing Li* Department of Chemical and Biomolecular Engineering, North Carolina State University, 911 Partners Way, Raleigh, NC , USA. Chapter 3 is a reprint of a manuscript published in Applied Catalysis B: Environmental, 2015, 164, The supplementary information is in Appendix B. *To whom the correspondence should be addressed. Telephone: +1 (919) Fax: +1(919) fli5@ncsu.edu. 67

92 Abstract The chemical looping processes utilize lattice oxygen in oxygen carriers to convert carbonaceous fuels in a cyclic redox mode while capturing CO2. Typical oxygen carriers are composed of a primary oxide for active lattice oxygen storage and a ceramic support for enhanced redox stability and activity. Among the various primary oxides reported to date, iron oxide represents a promising option due to its low cost and natural abundance. The current work investigates the effect of support on the cyclic redox performance of iron oxides as well as the underlying mechanisms. Three ceramic supports with varying physical and chemical properties, i.e. perovskite-structured Ca0.8Sr0.2Ti0.8Ni0.2O3, fluorite-structured CeO2, and spinel-structured MgAl2O4, are investigated. The results indicate that the redox properties of the oxygen carrier, e.g. activity and long-term stability, are significantly affected by support and iron oxide interactions. The perovskite supported oxygen carrier exhibits high activity and stability compared to oxygen carriers with ceria support, which deactivate by as much as 75% within 10 redox cycles. The high stability of perovskite supported oxygen carrier is attributable to its high mixed ionic-electronic conductivity. Deactivation of ceria supported samples results from solid-state migration of iron cations and subsequent enrichment on the oxygen carrier surface. This leads to agglomeration and lowered lattice oxygen accessibility. Activity of MgAl2O4 supported oxygen carrier is found to increase during redox cycles in methane. The activity increase is a consequence of surface area increase caused by filamentous carbon formation and oxygen carrier fragmentation. While higher redox activity is desired for chemical looping processes, physical degradation of oxygen carriers can be detrimental. 68

93 3.1 Introduction CO2 emissions from the oxidation of fossil fuels can lead to unintended consequences of climate change.[1,2] Thus, novel approaches for CO2 capture, utilization, and sequestration (CCUS) are highly desired. The chemical looping strategy offers a potentially viable option for efficient carbonaceous fuel conversion with reduced carbon footprints. In the chemical looping process, an oxygen carrier is reduced and oxidized in a cyclic manner to convert carbonaceous fuels into separate streams of concentrated carbon dioxide and carbon-free products such as heat/electricity or hydrogen.[3 9] During the reduction step, the oxygen carrier, a.k.a. redox material, donates its lattice oxygen to the fuel while getting reduced, thereby converting the carbonaceous fuels to a gaseous mixture of carbon dioxide and water vapor. In the subsequent oxidation step(s), the oxygen deprived oxygen carrier is regenerated with gaseous oxygen and/or steam, resulting in heat and/or hydrogen products.[4,10] Since the effectiveness of chemical looping processes is largely affected by the performance of the oxygen carrier, i.e. its physical robustness, redox activity, stability, cost, etc., oxygen carrier development and optimization is a key area for chemical looping research. A key component of an oxygen carrier is its active material or primary oxide, which directly participates in the abovementioned cyclic redox operations. Resulting from their suitable thermodynamic properties, oxides of first row transition metals such as iron, nickel, copper, manganese, and/or cobalt are among the most frequently used primary oxides for chemical looping applications.[3,11 16] With most unsupported metal oxides showing diminishing redox activities over multiple redox cycles[3,17], a support, which is typically selected from ceramic materials with relatively stable structures, is usually added to the primary oxide to improve its long-term redox performance. Among the various primary oxides, iron oxide(s) is 69

94 particularly attractive due to its low cost, high oxygen capacity, and abundance.[3,18] In addition, reduced iron and wustite can be effective for hydrogen production through watersplitting.[5,12,19] A key challenge for iron oxide based oxygen carrier development is to obtain oxygen carriers with improved activity and recyclability. To address such a challenge, addition of supports including Al2O3, MgAl2O4, CeO2, TiO2, and YSZ have been investigated.[11,15,20 25] While supports can improve the activity and stability of iron based oxygen carriers [3,11,20,24,26,27], relatively few studies have investigated the roles of support for overall oxygen carrier performance. Activity changes, i.e. deactivation or activation, for supported iron oxide over multiple redox cycles have been reported. [28 31] Bleeker et. al reported deactivation of iron oxide in a steamiron process.[29] The authors attributed deactivation to the decrease in surface area by iron (oxide) sintering. In contrast, iron oxide ores such as ilmenite are shown to activate during the chemical looping combustion (CLC) processes, possibly due to an increase in porosity over cyclic redox reactions.[28,30] Li et al. investigated TiO2 supported iron oxides.[32,33] Slight increase in activity of the oxygen carrier was observed with decreasing surface area and porosity through redox cycles. Based on inert marker experiments and density function theory (DFT) calculations, the authors proposed that high conductivities of electrons and lattice oxygen (O 2- ) of the supported sample are responsible for improved redox activity of the supported iron oxide. A recent study by Galinsky et. al. have indicated that a mixed ionicelectronic conductive (MIEC) lanthanum strontium ferrite (LSF) support significantly improves the redox activity and stability of iron oxide based oxygen carriers.[34] The activity increase is believed to result from the excellent mixed conductivity of the LSF perovskite support. Shafiefarhood et. al. examined a novel strategy of covering iron oxide with an LSF 70

95 shell for methane partial oxidation. [35] The LSF shell helped to achieve high activity, stability, and syngas selectivity over duration of 100 redox cycles. Similar to LSF, ceria is a frequently studied mixed-conductive support.[36 47] Ceria supported iron oxide is reported to possess good redox stability for cyclic reactions with CH4 (reduction) and steam (regeneration).[23] Bhavsar and Veser deposited Fe2O3 onto synthesized ceria via an incipient wetness method.[40] The authors report stable activity of the Fe-Ce-O oxygen carrier in hydrogen at 800 C with H2 and CH4 for 10 cycles. He et al. showed, at 850 C in CH4, that Ce-Fe-O with a Ce/Fe of 7/3 had stable activity between Fe2O3 and FeO phases over 20 redox cycles.[42] Contrast to the mixed-conductive perovskite and ceria, MgAl2O4 is a commonly encountered inert support.[22, 24, 31, 48-52] Johansson et al. concluded that sintering temperature has a significant effect to the activity of Fe2O3 supported with MgAl2O4.[22] They observed a slight increase in redox activity over 9-17 cycles in a 1:1 ratio of CH4 and H2O. The authors postulated that the activation is resulted from cracks and pore size change. To summarize, although a large number of supported iron oxides are tested, their deactivation/activation mechanisms have not been thoroughly investigated. This article aims to investigate the roles of support on the redox activity and long-term stability of iron oxide based oxygen carriers. Three ceramic supports, i.e. perovskite-structured Ca0.8Sr0.2Ti0.8Ni0.2O3, fluorite-structured CeO2, and spinel-structured MgAl2O4 are investigated. Of these supports, Ca0.8Sr0.2Ti0.8Ni0.2O3 and CeO2 represent mixed-conductors with distinct crystal structures. MgAl2O4, on the other hand, is a frequently encountered inert support. Through detailed investigation of the relationship between oxygen carrier activation/deactivation and its support conductivity, surface and phase compositions, 71

96 morphology, and surface area, general design principles for highly active and recyclable oxygen carriers are obtained. 3.2 Experimental Methods Oxygen Carrier Synthesis Several oxygen carriers are synthesized using three approaches, i.e. solid state reaction (SSR), co-precipitation (CP), and citric acid (CA) methods. The oxygen carriers tested are shown in Table 3.1. Oxygen carrier samples prepared via SSR reactions are sintered at 1,000-1,050 C while those prepared through CP and CA methods are sintered at 800-1,000 C. X-ray powder diffraction (XRD) is used to confirm the formation of the desired phases. A brief summary of these synthesis methods is given in the following section. Further details can be found in several references. [34,53,54] To ensure generality of the mechanistic findings, synthesis method and compositions of ceria and MgAl2O4 supported oxygen carriers are consistent with those reported in literature [22,24-26,35,36,39-47]. Activity and recyclability of the ceria based oxygen carriers in the current study are comparable, if not higher, than those reported in previous literatures under similar operating conditions (see Supplemental Figure 3.S3). 72

97 Table 3.1. List of oxygen carriers synthesized. Material Method and composition 1) Fe2O3:CeO2 SSR (3:7 molar ratio) 2) Fe2O3:CeO2 Co-precipitation (3:7 molar ratio) 3) Fe2O3:CeO2 Co-precipitation (4:6 molar ratio) 4) Fe2O3:Ca0.8Sr0.2Ti0.8Ni0.2O3 citric acid (3:7 molar ratio) 5) Fe2O3:MgAl2O4 SSR (3:7 weight ratio) SSR samples are prepared by measuring calculated amounts of precursors Fe2O3 (99.9%, Noah Chemicals), CeO2 (99.9%, Noah Chemicals), and MgAl2O4 (99.9%, Noah Chemicals) and placing them for 6 hours in a planetary ball mill (XBM4X, Columbia International) at a rotation of 250 rpm. Following the milling step, the resulting mixture is pressed into pellets by a hydraulic press (YLJ-15T, MTI Corporation) at a pressure of 20MPa. The pellets are then annealed, fractured, and sieved to powder (<150 microns) for testing. The co-precipitation method involves taking calculated precursor nitrates Fe(NO3)3 9H2O (98+%, Sigma-Aldrich) and Ce(NO3)3 6H2O (99%, Sigma-Aldrich) into solution with deionized water. The solution is then mixed at 70 C where a 10% ammonia solution is introduced drop wise to increase the ph of the solution to 9. The solution is allowed to precipitate at 70 C for 1 hour. It is then aged for 2 hours. The resulting solids are subsequently filtered and washed thoroughly with deionized water and ethanol. They are then dried at 110 C for 24 hours. Calcination is performed at 300 C to convert the hydroxide groups into oxides, and then the particles are crushed into powder and sintered for 6 hours at 800 C. A sintering 73

98 temperature of 800 C is used for CP sample to reproduce oxygen carriers reported in literature. [53] The citric acid method is used to prepare the Fe2O3:Ca0.8Sr0.2Ti0.8Ni0.2O3 oxygen carrier from Fe2O3 nanoparticles (<50nm, Sigma-Aldrich), nitrate salts, and alkoxides. First, Fe2O3 nanoparticles are dispersed in ethanol via sonication. The solution is settled for 6 hours to deposit the nanoparticles and to remove the upper liquid of the solution. The nanoparticles are re-dispersed in deionized water and sonicated. A second solution comprised of the precursors Ca(NO3)2 4H2O ( 99.9%, Sigma-Aldrich), Sr(NO3)2 (ACS Reagent, Noah Chemicals), and Ni(NO3)2 6H2O (Sigma-Aldrich) are placed into deionized water. Citric acid and EDTA are used as chelating agents. An ammonia solution is used to change the ph of the solution to 8. A third solution containing ethanol and citric acid is used to dissolve Ti(OC4H9)4 (97%, Sigma- Aldrich). Solutions containing the metal ions are mixed thoroughly. The dispersed nanoparticles are introduced to the solution and heated at 80 C under mixing conditions until gel formation. The gel is then dried at 120 C for 6 hours followed by calcination at 1,000 C for 6 hours Reactivity Studies Reactivity studies are conducted in a SETARAM SETSYS Evolution Thermal Gravimetric Analyzer (TGA). Up to 75 mg powdery samples are loaded into the TGA. Both H2 and CH4 are used as the reducing gas and O2 is used as the oxidizing gas. Gas flow rates are adjusted so that there is 10% reducing (H2 or CH4) or oxidizing (O2) gas balanced with inert (He). Unless otherwise specified, the temperature inside the TGA is held at 900 C with a total flow rate of 300mL/min. Carbon formations are quantified on samples reduced in methane. Carbon content is estimated using the TGA curves by examining the weight gain after near complete reduction 74

99 of the oxygen carrier samples. Such an analysis method, which was reported in prior literatures [55,56], assumes that sample weight loss from further reduction is minimal compared to the weight gain from coke formation Sample Characterizations A number of characterization tools are utilized to characterize surface areas, pore size distributions, bulk compositions, crystal structure, and morphology of the oxygen carriers before and after redox reactions. X-ray powder diffraction (XRD) is carried out to analyze crystal phase compositions and crystallite sizes before and after redox cycles. XRD patterns are recorded using a Rigaku SmartLab X-ray diffractometer with Cu-Kα (λ=0.1542) radiation operating at 40 kv and 44 ma. Scans are performed stepwise with a 0.1 step size holding for 3.5 seconds at each step between a angle range (2θ). Surface areas and pore size distributions are measured using a BET surface analyzer (Quantachrome QuadraSorb Station 1) using nitrogen physisorption at 77.3 K. For image capturing, a scanning electron microscope (SEM, Hitachi S3200) is used to observe surface morphology and structure. The same instrument is used to perform X-ray spectroscopy (EDX) scans on the particles to obtain bulk and area compositions at a wide range of accelerating voltages (10-30 kev). Transmission Electron Microscopy (TEM, JEOL JEM 2010F) with 200 kev accelerating voltage are used to characterize carbon formation on the MgAl2O4 sample. X-ray photoelectron spectroscopy (XPS) was used to probe the near-surface composition of the fresh and cycled ceria supported oxygen carrier. The custom built system was comprised of a Thermo-Fisher Alpha 110 hemispherical energy analyzer, a Thermo-Fisher XR3, 300 watt duel anode x-ray source, and a chamber with a base pressure 1*10-9 Torr. The Mg anode was used. Survey spectra were 75

100 taken with a pass energy of 100 ev, and narrow scans were taken with a pass energy of 20 ev. The spectra were processed in CasaXPS. 3.3 Results Activity Comparisons Redox activities of oxygen carriers under H2 and CH4 environments are summarized in Figure 3.1. Second cycle activity data is used for comparison purpose since they tend to be more stable than first cycle results. Average conversion rates (conversion %/min.) to achieve 11% conversion and 33% conversion, which correspond to Fe2O3 to Fe3O4 and Fe2O3 to FeO respectively, are used to characterize the activities of the oxygen carrier. The conversion profiles are provided in Figure 3.S1 of the supplemental document. Oxygen carrier conversion is defined by Equation 1: x OC = (m int. m i ) 100% m int x O2 (1) where mint. is the initial weight of the oxygen carrier, mi is mass at time i during the reduction, and xo2 is the amount of active lattice oxygen in the oxygen carrier. Average conversion rates are determined by taking the conversion of 11% or 33% and dividing it with the time to achieve the corresponding conversion. As can be seen from Figure 3.1, the perovskite supported oxygen carrier has a higher activity than those of CeO2 and MgAl2O4 supported oxygen carriers in both hydrogen and methane. Compared to the activity to achieve 11% conversion, oxygen carriers exhibit higher activity to achieve 33% when H2 is the reducing gas. This indicates that, between 0 and 33% conversion of the oxygen carrier, reaction rates tend to increase with deeper reduction of the iron oxide. 76

101 Such a sigmoidal behavior has been reported in previous literatures for ilmenite reduction.[57] An activation period, which is associated with ion migration, is proposed to be responsible for the relatively low activity at the initial stage of the reduction. When CH4 is used as the reducing gas; however, oxygen carrier activity tends to decrease with increasing degree of reductions. The difference in reduction behavior of iron oxide in H2 and CH4 can be partially attributed to the difference in diffusivities of H2 and CH4 molecules as well as the surface activation scheme. All the oxygen carriers investigated exhibit lower activities in CH4 than those in H2. This can be explained by the high stability C-H bond in methane. Figure 3.2 compares the activity change for the oxygen carriers between the 2 nd and 10 th reduction cycles. As can be seen, the perovskite supported iron oxide deactivates by approximately 16% from cycle 2 to 10. While all the ceria samples exhibit deactivation, 4:6 (Fe:Ce) co-precipitated ceria supported oxygen carrier deactivates to the largest extent (~75%). In contrast, MgAl2O4 supported oxygen carrier showed activity increase by almost 60% over the 9 redox cycles. The potential causes of activity changes are discussed in the following sections. It is noted that the current study focuses on reduction cycles since oxidation rates tend to be similar to, if not faster than, the reduction steps when CH4 is used as the fuel (See Figure 3.S2 in Supplemental document). 77

102 Figure 3.1- Comparison of the 2 nd cycle reduction of the various oxygen carriers by comparison of average conversion rate (X ave) to achieve 11% and 33% conversion in (a) 10% H2 and (b) 10% CH4 at 900 C. Figure 3.2- Average conversion rate (X ave) to 33% conversion comparison between 2 nd and 10 th cycles of the various oxygen carriers at 900 C in 10% CH4. 78

103 3.3.2 Ca0.8Sr0.2Ti0.8Ni0.2O3 Supported Fe2O3 The multi-cyclic redox activities of perovskite supported iron oxide with H2 and CH4 are shown in Figure 3.3. Excellent stability of the oxygen carrier in H2 is observed within the second to tenth redox cycles. The activation of the perovskite from the first to second cycles in H2 may be due to incomplete perovskite phase formation of the oxygen carrier. In contrast, the same oxygen carrier deactivates by 15% within the first 5 redox cycles when methane is used as the fuel. When cycled further, the perovskite supported oxygen carrier is more stable, showing only slight deactivation (<1%) between cycle 5 and 50. Deactivation in the short term is likely to be caused by phase instability of the support. While the surface area also decreases, our previous works show very weak correlation between activity and surface area with regards to perovskite supported iron oxide.[34,35] XRD spectra of the fresh and ten cycle (ending on oxidation) perovskite supported oxygen carrier are shown in Figure 3.4. As indicated by the absence of Ni oxides, Ni is incorporated in the perovskite B-site in the fresh sample. A slight addition of Ni to the B-site did not change the diffraction pattern of the parent perovskite (Orthorhombic Ca0.8Sr0.2TiO3) to any significant extent. After 10 cycles, a distinct Ni-Fe spinel structure is observed. This indicates that Ni is likely to be partially substituted by Fe since substitution of Ni 2+ with Fe 3+ is likely to result in a more stable CaxSr1-xTiyM1-yO3-δ structure from a charge balance standpoint. This is consistent with refinement results which indicate a lower lattice constant in the b direction, causing unit cell volume to decrease by nearly 0.6% (See SI Table 3.S1). Such a decrease can be explained by the addition of a secondary B-site dopant, as Fe 3+ is notably smaller than Ni 2+. The activity of the oxygen carrier becomes stable once the phase transformation is completed. 79

104 From TGA curves shown in Figure 3.5, notable coke formation starts to occur when the oxygen carrier is close to being fully reduced. This is expected due to the high activity of reduced Ni (and Fe) for methane dissociation.[58 60] Coke formation is calculated from the TGA curves as shown in Figure 3.5. From the TGA, it can be observed that carbon formation does not begin until near complete reduction of the iron oxide to metallic iron (~9% weight loss). This agrees with our previous work on perovskite supported iron oxides.[34] It is noted that oxygen carrier conversion is typically limited to less than 50% for iron oxide based oxygen carriers. Therefore, carbon formation in actual process operations can be low. BET surface area results (see Table 3.2) indicate decreasing surface area with increase redox cycles. The surface area is decreased by almost a factor of 4 over the first 10 cycles. The corresponding deactivation; however, is less than 15%. No volume expansion is observed for this oxygen carrier over 50 redox cycles. To summarize, perovskite supported iron oxide, although deactivates slightly in the first few redox cycles, exhibits high redox activity and stability over multiple redox cycles. The high electronic and ionic conductivities are likely to be responsible for the high activity of the oxygen carrier. Figure 3.3- Average conversion rate ( X ave ) to achieve 33% conversion of Ca0.8Sr0.2Ti0.8Ni0.2O3 supported Fe2O3 via CA method with (a) 10 cycles with H2 and (b) 50- cycles with CH4 at 900 C. 80

105 Figure 3.4- XRD analyses of the (a) fresh and (b) 10-cycled CSTN supported iron oxide samples. Figure 3.5- Oxygen carrier weight loss (TGA) curves for cycle 2 and cycle 10 of the Fe2O3:CSTN (CA) in CH4 at 900 C. Vertical lines represent the estimated coke formation for the reaction with methane. Lower horizontal line represents the demarcation between Fe2O3/Fe3O4 and Fe3O4/FeO; Upper horizontal line represents the demarcation between Fe3O4/FeO and FeO/Fe (calculated based on average oxidation state). 81

106 Table 3.2. Carbon Formation (within 10 mins. of reduction in CH4) and BET surface areas of the CSTN supported iron oxide oxygen carrier 2 nd reduced 10 th cycle reduced Carbon Formation (%) (10 minute reduction) BET surface areas (oxidized samples) (m 2 /gm) CeO2 Supported Fe2O3 As shown in Figures 3.2 and 3.6, CeO2 supported iron oxides exhibit high initial activity with CH4. However, their activities decrease over multiple redox cycles. Change in activity for a co-precipitated Fe2O3:CeO2 (4:6) sample is further illustrated in Figure 3.6. As can be seen, activity of the oxygen carrier decreases from cycle 1 onwards. By cycle 10, the sample reduction rate has decreased by 75%. When the oxygen carrier is cycled further, it is unable to achieve 33% conversion during the 10-minute reduction period. The decreasing reduction rate also affects the onset of coke formation. While carbon formation is observed during the first 5 redox cycles, coke formation is insignificant in the subsequent cycles. This can be explained by incomplete reduction of the oxygen carrier: The cycled oxygen carriers, being less active, are not fully reduced during the 10-minutes reduction period and contain active lattice oxygen. The lack of metallic iron on oxygen carrier surface and the high lattice oxygen conductivity of the CeO2 support lead to reduced coke formation by (i) decreasing methane dissociation rate; (ii) oxidation of surface-carbon with active lattice oxygen. 82

107 Figure 3.6- Average reduction rate (X ave) to achieve 33% conversion of Fe2O3:CeO2 (4:6) prepared via a co-precipitation route over the first 10 redox cycles during a 50 cycle experiment. Reducing gas: methane; Oxidizing gas: oxygen; Temperature: 900 C. XRD spectra of the fresh, 1 st cycle reduced, 50 th cycle oxidized, and 51 st cycle reduced samples are summarized in Figure 3.7. As can be seen, oxidized oxygen carriers (fresh and 50 th cycled) indicate only the primary metal oxide (Fe2O3) and support (CeO2) phases. Fe2O3 peaks are less intense after 50 cycles, which corresponds to smaller crystallite sizes (see Table 3.3). With respect to the reduced oxygen carriers, 1 st cycle reduced oxygen carrier is composed of iron and CeO2 phases. The results are consistent with TGA weight loss as well as observations reported by Galvita et. al.[36] The 51 st reduced oxygen carrier; however, consists of a ternary phase of metallic iron, CeO2, and CeFeO3 perovskite. The formation of CeFeO3 has also been reported in previous literatures. [23] As can be seen in Table 3.3, surface area of the sample is reduced by nearly an order of magnitude during 50 cycles. PSDs indicate a smaller amount of meso-pores in the cycled sample, likely due to sintering (see Figure 3.S5 in Supplemental Document). 83

108 Figure 3.7- XRD of the (a) fresh and 50th cycle oxidized (b) 1 st reduced and 51 st reduced during the redox in methane and oxygen of the (4:6) co-precipitation Fe2O3:CeO2 sample. Table 3.3. Summary of BET, crystallite sizes, and carbon formation study on the Fe2O3:CeO2 (4:6) CP oxygen carrier Fresh (2 nd reduced) 50 th cycle Carbon Formation (%) (10 minute reduction) BET Surface Area (oxidized samples) (m 2 /gm) Crystallite Size CeO2 (Å) Crystallite Size Fe2O3 (Å) To further investigate the sintering behavior, morphology and compositions of the oxygen carriers are examined using SEM and EDX (see supplemental document figure 3.S6 and tables 3.S2 and 3.S3). As the material is cycled, sintering of these particles take place, leading to larger particles with reduced porosity. This is consistent with the changes in surface area and PSD. EDX analyses are performed to obtain the average (near surface) composition of the oxygen carrier. Fe:Ce ratio for as-prepared sample increases from 2.1 to 6.5 when acceleration voltage decreases from 30 kev to 10 kev (Figure 3.S6). Since EDX detects elemental 84

109 compositions within a few micrometers from the sample surface, iron oxide is likely to have been enriched on the oxygen carrier surface. To determine the surface composition changes over redox cycles, EDX area scans on the fresh, 51 st oxidation, and 52 nd reduced samples are conducted (See Supplemental Table 3.S2). Redox reactions appear to enhance the surface enrichment of iron oxides. For instance, the Fe:Ce ratio of 51 st oxidized sample is 2.4 times higher than that of the as-prepared sample. Large agglomerates of iron oxide are also observed. The reduced oxygen carriers tend to be more homogeneous although isolated regions enriched with iron are nevertheless identified. Increased homogeneity of the reduced sample can be explained by CeFeO3 phase formation through solid state reactions between Fe and Ce oxides, as confirmed by XRD. X-ray photoelectron spectroscopy (XPS) is performed on fresh and 51 st oxidized oxygen carrier as shown in Figure 3.8. XPS results indicate that nearly 40 at.% of the initial oxygen carrier surface is iron. Upon examining the 51 st oxidized oxygen carrier, enrichment of iron to almost 80 at.% was observed. Since XPS is sensitive to atoms within a few nm below the sample surface, these results, along with the EDX data, further confirm iron cation diffusion and enrichment on the sample surface. Figure 3.8- XPS of the (a) fresh and (b) 51 st oxidized ceria supported oxygen carrier 85

110 While surface iron enrichment on CeO2 supported iron oxide has not been reported, outward growth of iron in unsupported iron oxides during redox cycles has been reported in literature.[29,33] This can be explained by lower ionic diffusivity of O 2- compared to Fe x+ (x = 2 or 3) in both magnetite and wustite. As a result, cations of iron need to migrate out from the oxide lattice to react with oxygen during the oxidation step. Since the iron cations may not migrate back during the reduction cycles, continued redox cycles leads to continued surface enrichment of iron oxide layers. A similar phenomenon of iron migration is also reported by Zhao et al.[57] for the reduction of ilmenite (FeTiO3) to Fe and TiO2 in H2. Although CeO2 possesses good mixed-conductivity for electron and lattice oxygen transfer, migration of Fe and Ce cations are likely to occur throughout the redox cycles, as evidenced by mixed Fe-Ce oxide formation. To compare, the energy barrier for Fe 3+ diffusion in Ca0.8Sr0.2Ti0.8Ni0.2O3 perovskite is likely to be high: typical energy barrier for the diffusion of B-site cation in acceptor-doped perovskites is ~3 ev compared to <1 ev for O 2- diffusion.[63,64] Therefore, Ca0.8Sr0.2Ti0.8Ni0.2O3 supported iron oxide possesses a more stable structure by allowing facile lattice oxygen exchange. In contrast, Fe cations may be easier to migrate than oxygen anions in CeFeO3, since oxygen vacancy within CeFeO3 is likely to be very low. Based on the above findings, three possible causes for deactivation of ceria supported oxygen carriers can be identified: (i) formation of a stable CeFeO3 perovskite phase, (ii) decrease in surface area, and (iii) the enrichment and segregation of iron oxides on the oxygen carrier surface. While mechanism (i) may partially account for the decrease in activity, it is not the primary cause since one would expect the redox activity of the oxygen carrier to stabilize once CeFeO3 is formed. This; however, is not observed (Figure 3.S8). In order to determine whether surface area loss (mechanism ii) is the main deactivation factor, a low surface area Fe2O3:CeO2 86

111 is prepared by sintering an identical sample at 1100 C for 8 hours. The 300 C increase in sintering temperature compared to the reference sample leads to an oxygen carrier sample with lower initial surface area. Activity comparisons of the low sintering temperature (800 C) and high sintering temperature (1,100 C) samples are shown in Figure 3.9. With respect to the first reduction cycle, the 1,100 C sintered oxygen carrier has almost 6 times lower activity (for 33% conversion) compared to the 800 C sintered sample, indicating that surface area plays a role in sample activity. While the 800 C sintered sample deactivates from the 5 th cycle onwards, the 1,100 C sintered sample activates over the first 10 cycles by approximately a factor of 3 before deactivating at a slower pace. The initial activation is resulted from the formation of cracks and pores from the thoroughly sintered oxygen carrier. Such activation-slow deactivation phenomena clearly indicate that loss of surface area during high temperature redox reactions is not the primary cause for deactivation. Figure 3.9- Average conversion rate (X ave) to (a) 11% and (b) 33% conversion comparison of the 800 C 1,100 C sintered CeO2 supported Fe2O3 (6:4) in CH4 at 900 C synthesized by a co-precipitation method. The lower sintered oxygen carrier does not achieve 33% conversion in the 25 th and 50 th cycles. 87

112 To further validate the migration of Fe x+, Fe/Ce ratios are obtained for both the 1100 C 4:6 Fe:Ce oxygen carrier and 800 C 3:7 Fe:Ce oxygen carrier (see Figure 3.S7 and Table 3.S3). SEM and EDX scans of the 1,100 C sintered oxygen carrier indicate a more homogeneous starting material compared to the 800 C sintered sample. EDX scans on the 50 th cycle oxidized 1,100 C sample also show an enrichment of iron. The lower rate of iron phase segregation for the 1,100 C sample is likely to be responsible for the slower deactivation of the high temperature sintered sample. Figure 3.10 summarizes the relationship between oxygen carrier activity and near surface Fe/Ce ratio for all three ceria supported iron oxides. As can be seen, surface enrichment of Fe occurs in all three ceria supported samples, leading to a general trend of lowering activity with increasing near surface Fe/Ce ratio. Based on the above findings, it is concluded that iron oxide aggregation on the surface, through Fe x+ migration, is the primary cause for the deactivation of ceria supported oxygen carriers. Aggregation of iron oxide causes deactivation through increased gaseous and solid-state diffusion resistances. Decrease in sample surface area, as observed in 800 C sintered sample, is a consequence of iron migration/agglomeration as opposed to the primary cause for deactivation. 88

113 Figure Comparison of the activity of CeO2 supported iron oxides versus the Fe/Ce ratio on the surface. Cycle 5 and 50 data are used for the 1,100 C sample. Cycle 1 and 50 data are used for the other samples MgAl2O4 Supported Fe2O3 Unlike CeO2, MgAl2O4 supported oxygen carrier becomes more active for methane conversion as the number of cycles increase. Figure 3.11(a) and (b) illustrate the conversion rate of MgAl2O4 supported oxygen carrier in H2 and CH4, respectively. Interestingly, activity of the sample in H2 is found to decrease by 30% from cycle 2 to 10 while CH4 activity increases by 30% between cycle 2 and 5. To further investigate this fuel-dependent activation phenomenon, a second MgAl2O4 supported oxygen carrier with increased amount of iron oxide (6:4 instead of 3:7 weight ratio between Fe2O3 and MgAl2O4) sintered at 1,200 C is prepared. The sample is found to activate in methane at a faster rate, as shown in Figure 3.11 (c). Carbon formation for this sample based on sample weight gain, which tends to underestimate actual carbon formation, is much more severe than the CeO2 supported material. 89

114 XRD spectra of the oxygen carrier developed in Figure 3.11 (c) are shown in Figure The phases are fairly consistent in the oxidized samples along with iron oxide crystallite size. The phase analysis of the reduced samples; however, presents significant difference over multiple redox cycles. Initially, the iron oxide is only reduced to FeO within the 10 minute reduction cycle; this is consistent with the low activity of the sample. It is also observed that Fe is partially substituted into the spinel structure of the support. As the activity of the oxygen carrier increases, a new iron carbide phase is observed. This indicates coke formation and increased degree of iron oxide reduction. In the meantime, BET surface area of the sample is found to increase by 15% over the 5 redox cycles. A summary of the surface area, crystallite sizes, and carbon formation are shown in Table 3.4. The PSDs of the MgAl2O4 supported oxygen carrier indicate an increase in meso-pore volumes with increasing number of redox cycles (See Figure 3.S10). Further characterization of the sample under TEM indicates notable carbon filaments in reduced oxygen carrier samples after multiple redox cycles (See Figure 3.13). While TEM images of the first cycle reduced sample display little carbon formation (See Figure 3.S11), TEM images of 6 th cycle reduced sample illustrates grapheme layers covering the surface of the oxygen carrier and growing out in thread like filaments. Most of the hollow fiber strands can be categorized as filamentous carbon or multi-walled carbon nanotubes. These observations are in agreement with what was observed by Ermakova et. al.[65] The nanotube growth on the sample has led to the breakage of active catalyst grains, i.e. iron, from the oxygen carrier surface, thereby compromising the structural integrity of the oxygen carrier. The deformation of the particle is due to growth of the filamentous carbon causing interstitial gaps within the oxygen carrier. These effects correspond well to surface area 90

115 increase of the sample. Visual inspection of the sample also indicates significant expansion of the oxygen carrier sample over multiple redox cycles. In contrast, deactivation of the same sample is observed in H2 due to the lack of carbon formation and hence continued sintering of the oxygen carrier. It should be noted, while high redox activity of oxygen carriers is a sought after feature in chemical looping processes. Physical disintegration of the oxygen carrier can be detrimental to CLC, which is operated in fluidized bed and/or moving bed reactors. Based on the above discussions, formation of filamentous carbon on the MgAl2O4 supported oxygen carrier is the primary cause for its activation phenomena in CH4-O2 redox cycles and the corresponding surface area increase through structural disintegration. 91

116 Figure Average conversion rate (X ave) (%/min) to 33% conversion of the MgAl2O4 supported Fe2O3 (7:3) via SSR tested in (a) 10 cycles with H2 and (b) 5-cycles with CH4 and (c) 5 cycles in CH4 of 1200 C sintered 4:6 Fe2O3:MgAl2O4 at a reaction temperature of 900 C. Figure XRD of the (a) fresh and 5th oxidized samples of Fe2O3 MgAl2O4 SSR sample and (b) of the 1 st and 6 th reduced oxygen carrier. 92

117 Table 3.4. Summary of carbon formation, crystallite sizes, and BET surface areas of the MgAl2O4:Fe2O3 (4:6) oxygen carrier Fresh (2 nd reduced) 5 th Carbon Formation (%) (60 minute reduction) BET Surface Area (oxidized samples) (m 2 /gm) Crystallite Size MgAl2O4 (Å) Crystallite Size Fe2O3 (Å) Figure TEM images of the 6 th reduced MgAl2O4 supported iron oxide. 3.4 Conclusion The present study investigates supported iron oxides with distinct structural and conductive properties for chemical looping applications. While oxygen carriers supported by mixedconductive perovskite (CaxSr1-xTiyNi1-yO3-δ) and fluorite (CeO2) both exhibited high initial 93

118 activity, CeO2 supported iron oxide notably deactivates with increasing number of redox cycles. To compare, perovskite supported iron oxide only slightly deactivates within the first 5 redox cycles, resulting from substitution of Ni 2+ on perovskite B-site by Fe 3+. Complete substitution of Ni 2+ leads to stable redox activity. XRD, SEM, and EDX analyses indicate that continued deactivation of ceria supported iron oxide in methane, under deep redox cycles, can be explained by iron migration, enrichment, and agglomeration on the sample surface over multiple redox cycles. The inability for Fe to stabilize inside the CeO2 supported particle has led to large, segregated iron oxide on the surface. This causes decreased accessibility of O 2- to gaseous fuels. Unlike the mixed-conductor supported oxygen carriers, MgAl2O4 supported iron oxide is found to activate over CH4-O2 redox cycles but deactivate in H2-O2 redox cycles. Further investigation by TEM confirms significant carbon formation during CH4-O2 cycles. The graphitic carbon appears to cover the oxygen carrier where it threads out into filamentous carbon or multi-walled carbon nanotubes. The nanotube growth on the sample has led to breakage of active catalyst grains, i.e. iron, from the oxygen carrier surface, thereby compromising the structural integrity of the oxygen carrier while increasing its surface area and meso-pore volume. While the deformation of the oxygen carrier can significantly enhance the activity of the oxygen carriers, it leads to undesired structural degradation of the oxygen carrier over multiple redox cycles. From an oxygen carrier design and optimization standpoint, supports that stabilize iron (oxide) grains in a mixed-conductive matrix can result in highly active oxygen carriers that exhibit stable activity and structural integrity. One should avoid supports that promote outwards 94

119 diffusion of Fe x+ or carbon filament formation since such supports could lead to oxygen carrier deactivation or apparent activation through structural expansion/disintegration. Acknowledgements This work is supported by the U.S. National Science Foundation under Award (CBET ) and the U.S. Department of Energy under Award (FE ). The authors acknowledge the use of the Analytical Instrumentation Facility (AIF) at North Carolina State University, which is supported by the State of North Carolina and the National Science Foundation. We would also like to acknowledge Amy Stewart for her help in conducting some of the reactivity experiments. 95

120 REFERENCES [1] C. Bessou, F. Ferchaud, B. Gabrielle, B. Mary, Biofuels, Greenhouse Gases and Climate Change, in: E. Lichtfouse, M. Hamelin, M. Navarrete, P. Debaeke (Eds.), Sustain. Agric. Vol. 2, Springer Netherlands, Dordrecht, 2011: pp [2] D.A. Lashof, D.R. Ahuja, Relative Contributions of Greenhouse Gas Emissions to Global Warming, Nature. 344 (1990) [3] J. Adanez, A. Abad, F. Garcia-Labiano, P. Gayan, L.F. de Diego, Progress in Chemical- Looping Combustion and Reforming Technologies, Prog. Energ. Combust. 38 (2012) [4] L.-S. Fan, Chemical Looping Systems for Fossil Energy Conversions, [5] L.-S. Fan, F. Li, Chemical Looping Technology and Its Fossil Energy Conversion Applications, Ind. Eng. Chem. Res. 49 (2010) [6] L.-S. Fan, F. Li, Conversion of Carbonaceous Fuels Into Carbon Free Energy Carriers, EP A2, [7] L.-S. Fan, L. Zeng, W. Wang, S. Luo, Chemical Looping Processes for CO2 Capture and Carbonaceous Fuel Conversion prospect and opportunity, Energ. Environ. Sci. 5 (2012) [8] M.M. Hossain, H.I. de Lasa, Chemical-Looping Combustion (CLC) for Inherent Separations A Review, Chem. Eng. Sci. 63 (2008) [9] A. Lyngfelt, M. Johansson, T. Mattisson, Chemical-Looping Combustion - Status Of Development, In: Proceedings of the 9 th International Conference on Circulating Fluidized Beds [10] F. Li, L.-S. Fan, Clean Coal Conversion Processes Progress and Challenges, Energ. Environ. Sci. 1 (2008) 248. [11] J. Adánez, L.F. de Diego, F. García-Labiano, P. Gayán, A. Abad, J.M. Palacios, Selection of Oxygen Carriers for Chemical-Looping Combustion, Energ. Fuel. 18 (2004)

121 [12] K.-S. Kang, C.-H. Kim, K.-K. Bae, W.-C. Cho, S.-H. Kim, C.-S. Park, Oxygen-Carrier Selection and Thermal Analysis of the Chemical-Looping Process for Hydrogen Production, Int. J. Hydrogen Energ. 35 (2010) [13] A. Lyngfelt, Oxygen Carriers for Chemical Looping Combustion - 4,000 h of Operational Experience, Oil Gas Sci. Technol. 66 (2011) [14] T.M. Paul Cho, Comparison of Iron-, Nickel-, Copper- and Manganese-Based Oxygen Carriers for Chemical-Looping Combustion, Fuel. (2004) [15] K. Shah, B. Moghtaderi, T. Wall, Selection of Suitable Oxygen Carriers for Chemical Looping Air Separation: A Thermodynamic Approach, Energ. Fuel. 26 (2012) [16] F. He, N. Galinsky, F. Li, Chemical Looping Gasification of Solid Fuels Using Bimetallic Oxygen Carrier Particles Feasibility Assessment and Process Simulations, Int. J. Hydrogen Energ. 38 (2013) [17] S. Roux, A. Bensakhria, G. Antonini, Study and Improvement of the Regeneration of Metallic Oxides Used as Oxygen Carriers for a New Combustion Process, Int. J. Chem. React. Eng. 4 (2006). [18] E. Jerndal, H. Leion, L. Axelsson, T. Ekvall, M. Hedberg, K. Johansson, et al., Using Low-Cost Iron-Based Materials as Oxygen Carriers for Chemical Looping Combustion, Oil Gas Sci. Technol. 66 (2011) [19] K. Svoboda, G. Slowinski, J. Rogut, D. Baxter, Thermodynamic Possibilities and Constraints for Pure Hydrogen Production by Iron Based Chemical Looping Process at Lower Temperatures, Energ. Convers. Manage. 48 (2007) [20] J. Adanez, F. García-Labiano, L.F. de Diego, P. Gayán, J. Celaya, A. Abad, - Characterization of Oxygen Carriers for Chemical-Looping Combustion, In: Greenhouse Gas Control Technologies 7, Elsevier Science Ltd, Oxford, 2005: pp [21] B.M. Corbella, J.M. Palacios, Titania-Supported Iron Oxide as Oxygen Carrier for Chemical-Looping Combustion of Methane, Fuel. 86 (2007)

122 [22] M. Johansson, T. Mattisson, A. Lyngfelt, Investigation of Fe2O3 with MgAl2O4 for Chemical-Looping Combustion, Ind. Eng. Chem. Res. 43 (2004) [23] X. Zhu, Y. Wei, H. Wang, K. Li, Ce Fe Oxygen Carriers for Chemical-Looping Steam Methane Reforming, Int. J. Hydrogen Energ. 38 (2013) [24] Q. Zafar, T. Mattisson, B. Gevert, Redox Investigation of Some Oxides of Transition- State Metals Ni, Cu, Fe, and Mn Supported on SiO2 and MgAl2O4, Energ. Fuel. 20 (2006) [25] K. Li, H. Wang, Y. Wei, D. Yan, Transformation of Methane into Synthesis Gas Using the Redox Property of Ce Fe Mixed Oxides: Effect of Calcination Temperature, Int. J. Hydrogen Energ. 36 (2011) [26] A. Hedayati, A.-M. Azad, M. Rydén, H. Leion, T. Mattisson, Evaluation of Novel Ceria- Supported Metal Oxides As Oxygen Carriers for Chemical-Looping Combustion, Ind. Eng. Chem. Res. 51 (2012) [27] F. Li, H.R. Kim, D. Sridhar, F. Wang, L. Zeng, J. Chen, et al., Syngas Chemical Looping Gasification Process: Oxygen Carrier Particle Selection and Performance, Energ. Fuel. 23 (2009) [28] J. Ada nez, A. Cuadrat, A. Abad, P. Gaya n, L.F. de Diego, F. Garcıá-Labiano, Ilmenite Activation during Consecutive Redox Cycles in Chemical-Looping Combustion, Energ. Fuel. 24 (2010) [29] M.F. Bleeker, H.J. Veringa, S.R.A. Kersten, Deactivation of Iron Oxide Used in the Steam-Iron Process to Produce Hydrogen, Appl. Catal. A-Gen. 357 (2009) [30] A. Cuadrat, A. Abad, J. Adánez, L.F. de Diego, F. García-Labiano, P. Gayán, Behavior of Ilmenite as Oxygen Carrier in Chemical-Looping Combustion, Fuel Process. Technol. 94 (2012) [31] T. Mattisson, M. Johansson, A. Lyngfelt, Multicycle Reduction and Oxidation of Different Types of Iron Oxide Particles Application to Chemical-Looping Combustion, Energ Fuel. 18 (2004)

123 [32] F. Li, S. Luo, Z. Sun, X. Bao, L.-S. Fan, Role of Metal Oxide Support in Redox Reactions of Iron Oxide for Chemical Looping Applications: Experiments and Density Functional Theory Calculations, Energ Environ. Sci. 4 (2011) [33] F. Li, Z. Sun, S. Luo, L.-S. Fan, Ionic Diffusion in the Oxidation of Iron Effect of Support and Its Implications to Chemical Looping Applications, Energ. Environ. Sci. 4 (2011) 876. [34] N.L. Galinsky, Y. Huang, A. Shafiefarhood, F. Li, Iron Oxide with Facilitated O 2- Transport for Facile Fuel Oxidation and CO2 Capture in a Chemical Looping Scheme, ACS Sustain. Chem. Eng. 1 (2013) [35] A. Shafiefarhood, N. Galinsky, Y. Huang, Y. Chen, F. Li, Fe2O3@LaxSr1-xFeO3 Core- Shell Redox Catalyst for Methane Partial Oxidation, ChemCatChem. 6 (2014) 790. [36] V.V. Galvita, H. Poelman, V. Bliznuk, C. Detavernier, G.B. Marin, CeO2-Modified Fe2O3 for CO2 Utilization via Chemical Looping, Ind. Eng. Chem. Res. 52 (2013) [37] X. Zhu, H. Wang, Y. Wei, K. Li, X. Cheng, Hydrogen and Syngas Production from Two-Step Steam Reforming of Methane Using CeO2 as Oxygen Carrier, J. Nat. Gas Chem. 20 (2011) [38] X. Zhu, H. Wang, Y. Wei, K. Li, X. Cheng, Reaction Characteristics of Chemical- Looping Steam Methane Reforming over a Ce ZrO2 Solid Solution Oxygen Carrier, Mendeleev Commun. 21 (2011) [39] Galvita, V., Hempel, T., Lorenz, H., Rihko-Struckmann, L.K., Sundmacher, K., Deactivation of Modified Iron Oxide Materials in the Cyclic Water Gas Shift Process for CO-Free Hydrogen Production, Ind. Eng. Chem. Res. 47 (2008) [40] S. Bhavsar, G. Veser, Bimetallic Fe Ni Oxygen Carriers for Chemical Looping Combustion, Ind. Eng. Chem. Res. 52 (2013) doi: /ie400612g. [41] Z. Gu, K. Li, H. Wang, Y. Wei, D. Yan, T. Qiao, Syngas production from methane over CeO2-Fe2O3 mixed oxides using a chemical-looping method, Kinet. Catal. 54 (2013) doi: /s x. 99

124 [42] F. He, Y. Wei, H. Li, H. Wang, Synthesis Gas Generation by Chemical-Looping Reforming Using Ce-Based Oxygen Carriers Modified with Fe, Cu, and Mn Oxides, Energy Fuels. 23 (2009) doi: /ef800922m. [43] K. Li, H. Wang, Y. Wei, M. Liu, Catalytic performance of cerium iron complex oxides for partial oxidation of methane to synthesis gas, J. Rare Earths. 26 (2008) doi: /s (08) [44] K. Li, H. Wang, Y. Wei, D. Yan, Direct conversion of methane to synthesis gas using lattice oxygen of CeO2 Fe2O3 complex oxides, Chem. Eng. J. 156 (2010) doi: /j.cej [45] K. Li, H. Wang, Y. Wei, D. Yan, Syngas production from methane and air via a redox process using Ce Fe mixed oxides as oxygen carriers, Appl. Catal. B Environ. 97 (2010) doi: /j.apcatb [46] Y. Wei, H. Wang, K. Li, Ce-Fe-O mixed oxide as oxygen carrier for the direct partial oxidation of methane to syngas, J. Rare Earths. 28 (2010) doi: /s (09)60154-x. [47] X. Zhu, K. Li, Y. Wei, H. Wang, L. Sun, Chemical-Looping Steam Methane Reforming over a CeO2 Fe2O3 Oxygen Carrier: Evolution of Its Structure and Reducibility, Energy Fuels. 28 (2014) doi: /ef402203a. [48] M. Arjmand, A.-M. Azad, H. Leion, A. Lyngfelt, T. Mattisson, Prospects of Al2O3 and MgAl2O4-Supported CuO Oxygen Carriers in Chemical-Looping Combustion (CLC) and Chemical-Looping with Oxygen Uncoupling (CLOU), Energ. Fuel. 25 (2011) [49] J.-I. Baek, J.-W. Kim, J.B. Lee, T.H. Eom, J. Ryu, C.K. Ryu, et al., Effects of Support on the Performance of NiO-Based Oxygen Carriers, Oil Gas Sci. Technol. 66 (2011) [50] S.-C. Baek, J.-W. Bae, J.Y. Cheon, K.-W. Jun, K.-Y. Lee, Combined Steam and Carbon Dioxide Reforming of Methane on Ni/MgAl2O4: Effect of CeO2 Promoter to Catalytic Performance, Catal. Lett. 141 (2010) [51] Q. Zafar, A. Abad, T. Mattisson, B. Gevert, Reaction Kinetics of Freeze-Granulated NiO/MgAl2O4 Oxygen Carrier Particles for Chemical-Looping Combustion, Energ. Fuel. 21 (2007)

125 [52] S. Wang, G. Wang, F. Jiang, M. Luo, H. Li, Chemical Looping Combustion of Coke Oven Gas by using Fe2O3/CuO with MgAl2O4 as Oxygen Carrier, Energ. Environ. Sci. 3 (2010) [53] K. Li, H. Wang, Y. Wei, M. Liu, Preparation and Characterization of Ce1-xFexO2 Complex Oxides and Its Catalytic Activity for Methane Selective Oxidation, J. Rare Earth. 26 (2008) [54] T. Hayakawa, H. Harihara, A.G. Andersen, A.P.E. York, K. Suzuki, H. Yasuda, et al., A Sustainable Catalyst for the Partial Oxidation of Methane to Syngas: Ni/Ca1-xSrxTiO3, Prepared In Situ from Perovskite Precursors, Angew. Chem. Int. Edit. 35 (1996) [55] T. Mendiara, J.M. Johansen, R. Utrilla, P. Geraldo, A.D. Jensen, P. Glarborg, Evaluation of Different Oxygen Carriers for Biomass Tar Reforming (I): Carbon deposition in Experiments with Toluene, Fuel. 90 (2011) [56] T. Mendiara, J.M. Johansen, R. Utrilla, A.D. Jensen, P. Glarborg, Evaluation of Different Oxygen Carriers for Biomass Tar Reforming (II): Carbon Deposition in Experiments with Methane and Other Gases, Fuel. 90 (2011) [57] Y. Zhao, F. Shadman, Reduction of Ilmenite with Hydrogen, Ind. Eng. Chem. Res. 30 (1991) [58] P. Cho, T. Mattisson, A. Lyngfelt, Carbon Formation on Nickel and Iron Oxide- Containing Oxygen Carriers for Chemical-Looping Combustion, Ind. Eng. Chem. Res. 44 (2005) [59] P. Debokx, The Formation of Filamentous Carbon on Iron and Nickel Catalysts I. Thermodynamics, J. Catal. 96 (1985) [60] A. Kock, The Formation of Filamentous Carbon on Iron and Nickel Catalysts II. Mechanism, J. Catal. 96 (1985) [61] E.T. Turkdogan, J.V. Vinters, Gaseous reduction of iron oxides: Part I. Reduction of hematite in hydrogen, Metall. Mater. Trans. B. 2 (1971)

126 [62] J. Smith, T. Norby, Cation Self-Diffusion in LaFeO3 Measured by the Solid State Reaction Method, Solid State Ionics 177 (2006) [63] A. Chroneos, R.V. Vovk, I.L. Goulatis, L.I. Goulatis, Oxygen Transport in Perovskite and Related Oxides: A Brief Review, J. Alloy. Compd. 494 (2010) [64] J.L. Routbort, S.J. Rothman, B.K. Flandermeyer, L.J. Nowicki, J.E. Baker, Oxygen Diffusion in La2-xSrxCuO4-y, J. Mater. Res. 3 (2011) [65] M. Ermakova, Decomposition of Methane over Iron Catalysts at the Range of Moderate Temperatures: The Influence of Structure of the Catalytic Systems and the Reaction Conditions on the Yield of Carbon and Morphology of Carbon Filaments, J. Catal. 201 (2001)

127 CHAPTER 4 Ca1-xAxMnO3 (A= Sr and Ba) Perovskite Based Oxygen Carriers for Chemical Looping with Oxygen Uncoupling (CLOU) Nathan L. Galinsky, Amit Mishra, Jia Zhang, and Fanxing Li* Department of Chemical and Biomolecular Engineering, North Carolina State University, 911 Partners Way, Raleigh, NC , USA. Chapter 4 is a reprint of a manuscript published in Applied Energy, 2015, 157, The supplementary information is in Appendix C. *To whom the correspondence should be addressed. Telephone: +1 (919) Fax: +1(919) fli5@ncsu.edu. 103

128 Abstract Operated under a cyclic redox mode with an oxygen carrier, the chemical looping with oxygen uncoupling (CLOU) process offers the potential to effectively combust solid fuels while capturing CO2. Development of oxygen carriers capable of reversibly exchanging their active lattice oxygen (O 2- ) with gaseous oxygen (O2) under varying external oxygen partial pressure (PO2) is of key importance to CLOU process performance. This article investigates the effect of A-site dopants on CaMnO3 based oxygen carriers for CLOU. Both Sr and Ba are explored as potential dopants at various concentrations. Phase segregations are observed with the addition of Ba dopant even at relatively low concentrations (5% A-site doping). In contrast, stable solid solutions are formed with Sr dopant at a wide range of doping level. While CaMnO3 perovskite suffers from irreversible change into Ruddlesden-Popper (Ca2MnO4) and spinel (CaMn2O4) phases under cyclic redox conditions, Sr doping is found to effectively stabilize the perovskite structure. In-situ XRD studies indicate that the Sr doped CaMnO3 maintains a stable orthorhombic perovskite structure under an inert environment (tested up to 1200 C). The same oxygen carrier sample exhibited high recyclability over 100 redox cycles at 850 C. Besides being highly recyclable, Sr doped CaMnO3 is found to be capable of releasing its lattice oxygen at a temperature significantly lower than that for CaMnO3, rendering it a potentially effective oxygen carrier for solid fuel combustion and carbon dioxide capture. 4.1 Introduction Climate change caused by anthropogenic carbon dioxide emissions along with the ever increasing demand for energy requires new power generation processes with lower carbon footprints. Of the various carbon capture technologies under development, chemical looping 104

129 combustion (CLC) has emerged as a potentially promising option.[1,2] The CLC processes utilize a solid oxygen carrier to transfer oxygen from the air to the fuel. This is achieved using interconnected fluidized bed or moving bed reactors.[3 8] In the first reactor, commonly termed the reducer/fuel reactor, the oxidized form of the oxygen carrier is reacted with the fuel to form carbon dioxide and water vapor. The oxygen depleted carrier is then transported to a second reactor referred to as the oxidizer/air reactor. Here, air is used to regenerate the oxygen depleted carrier while producing heat for power generation. While most oxygen carriers are effective for gaseous fuel combustion, they tend to be less active in converting solid fuels such as biomass or coal due to mass transfer and kinetic limitations. To enhance the combustion rate for solid fuels, two methods are typically used. In the first method, solid fuel is gasified into syngas by steam, carbon dioxide, or oxygen. The resulting syngas is then reacted with the oxygen carrier at a faster rate than direct solid-solid interactions. [6 11] A variety of oxygen carriers including Fe [8,10,12 14], Mn [13,15], and Ni [16,17] based oxides have been used in the conversion of solid fuels via such an in-situ gasification approach. An even faster approach to combust solids fuel is the so-called chemical looping with oxygen uncoupling (CLOU). [3,18,19] The CLOU process takes the advantage of oxygen carriers with high equilibrium oxygen partial pressure (PO2). As a result, solid fuels are combusted with gaseous oxygen released from the oxygen carrier, resulting in significantly improved reaction kinetics. Typical oxygen carriers that exhibit CLOU properties include oxides and mixed oxides containing Cu, Mn, and/or Co. [18 30] Despite possessing excellent oxygen uncoupling properties, copper oxides face challenges including a low melting point of the reduced Cu phase which can lead to agglomeration and defluidization. Supports, such as MgAl2O4, SiO2, and Al2O3, have been shown to help prevent 105

130 sintering and defluidization while maintaining the oxygen uncoupling properties of the copper oxide. [19,26 28] Nevertheless, the relatively high cost of copper oxide may limit its applications. Cobalt oxides, while potentially feasible, are typically avoided due to their low decomposition temperature, toxicity, and negative environmental impact. [18,23] Manganese oxides hold advantages over copper oxides in terms of high availability and low cost. However, pure manganese oxide is difficult to regenerate with air due to its high equilibrium PO2. Formation of mixed oxide phases through the addition of secondary metal oxides such as Fe, Ni, Mg, and Ca to manganese oxides has been shown as an effective approach to improve the regenerability and structural stability of Mn based oxygen carriers. [15,19,21 25,31 36] More recently, perovskite structured oxygen carriers have received increasing attentions as oxygen carriers for redox applications. [23,31,33 54] Typical perovskite takes the form of ABO3-δ, where A is a large cation of either the alkali earth or rare earth metal and B is a smaller transition metal cation.[55] As a support, mixed ionic and electronic conductive (MIEC) perovskites such as La1-xSrxFeO3 (LSF) have shown to enhance the redox activity of iron oxides by nearly two orders of magnitude.[51 54] Perovskite and perovskite supported iron oxide have also been explored as redox catalysts for syngas generation and watersplitting.[56,57] In terms of CLOU applications, La containing perovskite supports are reported to be effective to enhance the oxygen donation properties of mixed Mn-Fe and Co-Fe oxides.[23] Up to 8.8% decrease in initial decomposition temperature was reported for perovskite supported mixed metal oxides. As a standalone oxygen carrier, perovskites of the CaMnO3 family are extensively studied for CLOU and CLC applications.[20,29 39,49,58 62] While its oxygen uncoupling capacity is smaller than copper oxides, CaMnO3, which can be synthesized from abundantly available manganese ores and Ca precursors [29,61], has been 106

131 reported to be active for solid fuel conversions.[29,31,34,36] A challenge to CaMnO3 is its lack of long term stability and activity with fuels. Undoped CaMnO3 has been reported to undergo irreversible phase change to CaMn2O4 and Ca2MnO4 phases, leading to eventual deactivation of the oxygen carrier.[19,20,32,59,60] In addition, CaMnO3 has been shown to deactivate in the presence of sulfur, a common constituent of coal, through formation of CaSO4.[32,36,60] Prevention of irreversible phase transition has been investigated by doping secondary metals into the A and B-sites of the CaMnO3 parent material. The dopants can also be used to increase the activity of the oxygen carrier. The most common elements used to stabilize the structure are Ti and Mg. Pishahang et. al investigated Ca1-yMn1-xTixO3-δ oxygen carriers under fixed bed conditions using syngas as the fuel. [38] The authors reported that adding Ti and decreasing the Ca content increased stability of the oxygen carrier. 95% conversion of the syngas was achieved with no observable coke formation. Rydèn et. al studied a CaMn0.875Ti0.125O3-δ oxygen carrier for CLOU.[35] The oxygen carrier exhibited oxygen uncoupling at temperature above 720 C in a fluidized bed with a maximum oxygen uncoupling rate of 0.03 LN/min at 950 C. The maximum oxygen concentration was 4 vol. %. Besides the good uncoupling properties, the oxygen carrier exhibited good redox activity and stability for methane conversion. Kallen et al. used Mg doped CaMnO3 and tested the oxygen carrier in a 10kWth gas-fired CLC unit. [37] The synthesized oxygen carrier released oxygen at above 700 C. The oxygen carrier did not agglomerate for 120 hrs. at hot conditions (including 55hrs. of reaction) and formation of fines (<45 μm.) was below 0.01%, indicating possible longevity in the particle s lifetime. Another study using Mg, Ti, and Fe dopants conducted by Hallberg et. al. shows good activity with methane and syngas while releasing % by weight gaseous oxygen at 900 C. [33] A 107

132 maximum oxygen release of ~1 w.t.% was reported at 1000 C. The authors noted CaMn2O4 spinel formation after testing the oxygen carrier s CLOU properties in three cycles and activity for methane conversion in a single cycle. Hallberg et. al. also investigated the addition of both Ti and Mg in a 300W fluidized bed system using natural gas as the fuel. [62] The designed oxygen carriers showed higher conversion of the fuel than Ni oxides while showing little attrition, however, it was determined that the spent particles contained a noticeable amount of the spinel CaMn2O4 phase. Pour et. al. examined different oxygen carriers created by combining cheap manganese ores with Ca(OH)2. [29] While the primary phase was CaMnO3- δ, the manganese ores also contained Al, Fe, K, Ti, Mg, etc. in small quantities. In reaction with methane, ores containing high concentrations of Al showed poor performance due to spinel formation between Al and Mn. The best performing oxygen carrier made from a South African manganese ore, which was mainly composed of Fe, Si, Ca, Mg, and Mn, had an oxygen capacity of 1.5 w.t. % in methane. Oxygen uncoupling capacity was between w.t.% for the various oxygen carriers investigated. Although extensive studies on B-site dopants were performed, limited studies using A-site dopants have been conducted. Arjmand et. al. investigated CaxLa1-xMnyM1-yO3-δ (M=Mg, Ti, Fe, or Cu) oxygen carriers for CLOU properties between C in a laboratory scale fluidized bed.[31] Oxygen release for pure CaMnO3 had the highest oxygen concentration of 0.59% and 0.96% at 900 and 1000 C respectively. When doped with La, the oxygen concentration decreased to 0.3% and 0.77%. Doping of both A and B-sites caused even further decrease in oxygen release except for copper doping which exhibited higher oxygen concentration than CaMnO3. However, copper caused the perovskite materials to defluidize when fuel was added. The authors did detect small amounts of spinel CaMn2O4 although its 108

133 effect on the oxygen carrier performance was not documented. To summarize, further studies on the effect of A-site substitution using low cost alkali earth metals are desired in order to obtain a more informed strategy to develop oxygen carriers with lower cost and improved CLOU performance. This article investigates the effect of A-site dopants on CaMnO3 based oxygen carriers for CLOU applications. Both Sr and Ba are explored as potential dopants at various concentrations. The impacts of A-site dopants on phase stability, oxygen release properties, redox stability, and reactivity of the Ca1-xAxMnO3 (A= Sr and Ba) based oxygen carriers are investigated using an in-situ X-ray diffraction (XRD) instrument, a thermogravimetric analyzer (TGA), and a fluidized bed reactor. It is determined that Sr dopant can enhance the stability of the perovskite structure while providing facile oxygen release for solid fuel conversions. 4.2 Experimental Oxygen Carrier Synthesis Mixed oxides with a general formula of Ca1-xAxMnO3 (A= Ba or Sr and x=0, 0.05, 0.25, 0.5, 0.75, and 1) are prepared using a citric acid sol-gel method. General procedure for the sol-gel samples includes dispersion of cation precursors Ca(NO3)2 4H2O (Sigma-Aldrich), Ba(NO3)2 (Sigma Aldrich), Sr(NO3)2 (ACS Reagent, Noah), and Mn(NO3)2 4H2O (Sigma Aldrich) in distilled water followed by addition of citric acid. In the synthesis procedure used, citric acid (Sigma Aldrich) is added to the solution in a molar ratio of 2.5:1 (citric acid: total cations) as a chelating agent. After a short time, ethylene glycol (Sigma Aldrich) is used as a gel promoter in a molar ratio of 1.5:1 (ethylene glycol: citric acid) at a temperature of 80 C. After most of 109

134 the water is boiled out, the sol-gel sample is dried in a vacuum oven for 12 hours to remove any extra moisture. The dried sample is then annealed in air at 1200 C for 12 hours. The resulting powder is crushed and sieved to under 75 micron particle size. A solid state reaction (SSR) method is used to scale up the synthesis process. The SSR method procedure requires powder preparation, followed by pelletization, and annealing/solid state reaction. Preparation of a Ca0.75Sr0.25MnO3 oxygen carrier was conducted by mixing calculated amounts of CaCO3 ( 99.0%, Sigma-Aldrich), SrCO3 (99.0%, Noah), and Mn3O4 (99.0%, Noah) precursors for 12 hours in a planetary ball mill (XBM4X, Columbia International) at a rotation speed of 250 rpm. The mixture is then pressed into ¾ pellets by a hydraulic press (YLJ-15T, MTI Corporation) at a pressure of 20MPa and sintered in a high temperature furnace (GSL-1500-X50, MTI Corporation) at 1200 C for 12 hours. After annealing, the particles are fractured and sieved into a range of microns Oxygen Uncoupling Experiments Initial oxygen uncoupling properties of the oxygen carriers are characterized in a TGA. Two primary experiments are conducted in the TGA: (1) temperature programmable desorption TPD experiments using inert helium (5.0, Airgas) up to 1000 C at a ramp of 20 C/min and (2) isothermal experiments between C. Isothermal experiments are conducted at various temperatures to determine the oxygen uncoupling capabilities at low to high temperatures. At each temperature, 5 cycles of switching between pure inert and 10% O2 (Extra Dry, Airgas) balanced with a mixture of argon (5.0, Airgas) and helium are conducted for short term recyclability of the oxygen carrier. Best performing oxygen carriers are also tested for 100 cycles at 850 C to demonstrate the oxygen uncoupling performance of the perovskite oxygen carriers over the long term. 110

135 4.2.3 Fluidized Bed Experiments A laboratory-scale fluidized-bed reactor is used for solid fuel conversion experiment (Figure 4.S1). The quartz fluidized-bed reactor has an outer diameter of 25.4 mm. The reactor is heated through external heating from a tube furnace (MTI OTF-1200X-S-VT). Temperature is measured inside the reactor with an Inconel (Type K) thermocouple to set the temperature of the reactor to 850 C. A 3.75mm ID stainless steel tube is used for solid fuel injection into the reactor. A gas mixing panel is used to send the desired ratio of gases from both the bottom of the reactor for fluidization and through the smaller stainless steel tube for solid fuel injection. For maintaining fluidization, flow rates of 1080 and 1200 ml/min (at S.T.P) are used for solid fuel conversion (inert environment) and oxidation cycles, respectively. The flow rates correspond to 6-7 times the calculated minimum fluidization velocity, Umf, of the synthesized oxygen carrier. Prior to experiments, the reactor bed is loaded with 16 mesh silicon carbide (Kramer Industries) at the bottom of the reactor. The SiC particles act as a gas preheater and distributor. The SiC layer also supports the oxygen carrier bed. After the SiC layer is placed, approximately 15 grams of the oxygen carrier is placed into the reactor. For redox reactions of solid fuel, bituminous coal (Asbury Inc.) is converted into devolatized coal char through heating in the reactor at 800 C in a N2 environment. Injection of the coal is conducted in a pulse mode by loading from the top of the reactor through the small stainless steel tube and is pulsed into the bottom of the oxygen carrier layer of the reactor bed using N2. 10 redox cycles are tested using the fluidized bed reactor conducting two experiments in those 10 cycles. During cycles 1, 3, 5, and 10 solid fuel conversion is tested. Cycles 2, 4, and 6-9 are examining the oxygen uncoupling properties. Exhaust gas concentrations are measured using 111

136 a quadrupole mass spectrometer (MKS Cirrus 2) and a near-ir based gas analyzer (Emerson X-Stream) Sample Characterizations Phase identification of the oxygen carriers is conducted using X-ray powder diffraction. Powdered samples are prepared and tested using a Rigaku SmartLab X-ray diffractometer with Cu-Kα (λ=0.1542nm) radiation operating at 40kV and 44mA. A scanning range of (2θ) with a step size of 0.1 holding for 3.5 seconds at each step is used to generate the XRD patterns. To examine phase properties during TPD experiments, in-situ X-ray diffraction was conducted using an Empyrean PANalytical XRD using a similar Cu-Kα radiation operating at 45kV and 40mA. A 2θ range of is used at a ramp rate of 0.1 holding each step for 0.1s. A TPD experiment is conducted in the in-situ XRD by heating the sample at 5 C/min in pure argon. XRD scans are conducted the entire length of the ramp to 1200 C DFT Calculations Density functional theory (DFT) calculations are conducted to interpret the oxygen uncoupling behavior of cubic CaMnO3 and hexagonal BaMnO3. The calculations are performed using the Vienna ab initio simulation package (VASP) [63,64] in which a plane-wave basis set is used. The electron-ion interaction is modeled by the projector-augmented wave (PAW) method.[65,66] The Perdew-Burke-Enzerhoff (PBE) form of the generalized-gradient approximation (GGA) is used for the exchange and correlation functional.[67] The plane-wave cutoff energy is set to 425 ev. To calculate the O vacancy (OV) formation energy, 2x2x2 supercells are chosen for both BaMnO3 and CaMnO3 systems. A 2x2x2 Γ-centered k-mesh and a 4x4x4 Monkhost-Pack k-mesh are used to the BaMnO3 and CaMnO3 supercells, respectively. In our calculations, both the ferromagnetic (FM) and the antiferromagnetic (AFM) order are 112

137 used for considered perovskite structures. Here, A-type AFM and G-type AFM spin configurations are adopted by BaMnO3 and CaMnO3, respectively. In this work, removal of one O atom from the perfect structure is done to create a defect. The formation energy of the O vacancy is defined as E OV = E AMnO3 δ E O 2 E AMnO3 where E AMnO3 δ is the total energy of the perovskite supercells with one O vacancy (A= Ba, Ca and δ= 0.125), E O2 is the energy of the isolated O2 molecules in the spin-polarized triplet state, and E AMnO3 corresponds to the energy of the perfect AMnO Results and Discussions Motivation for Dopant Addition Previous studies indicate that CaMnO3 undergoes an irreversible phase transition into CaMn2O4-δ and Ca2MnO4 under redox conditions.[59] Since perovskites can accommodate significant amount of oxygen vacancies, the decomposition would occur in two steps. In the first step, manganese undergoes reduction from a 4+ valence state (CaMnO3) to 3+ (CaMnO2.5) with the perovskite structure maintained. Further reduction of Mn will result in the formation of new phases: 6CaMnO2.5 2CaMn2O4-δ + 2Ca2MnO4 + (δ-0.5)o2 Reaction 1 One potential approach to avoid such an undesired transition is to add dopants that enhance the stability of the perovskite phase. By inspection of the Goldschmidt or bond-valence model 113

138 (BVM) tolerance factor, it is possible to project dopant effect on the structural stability of perovskites semi-empirically. The Goldschmidt tolerance factor is defined in equation 1 [68]: t = r A + r O 2(r B + r O ) (Eq. 1) Where ra, rb, and ro are the ionic radii of the A-site ion, B-site ion, and oxygen ion. BVM tolerance factor replaces the summations of effective ionic radii in Equation 1 with ideal A-O and B-O bond distances.[69] Based on the report by Dabrowski et. al.,[70] CaMnO3 has a tolerance factor of While this value is within the suitable range to form stable cubic perovskites (0.9 t<1), further increase of the tolerance factor towards an ideal value of 1 for cubic perovskite can potentially enhance the stability of the oxygen vacant Ca1-xAxMnO2.5 perovskite. It is apparent from Equation 1 that larger A-site dopants such as Ba and Sr should render such an effect. It is also noted that the ionic radius of Ba 2+ (1.61 angstrom) is significantly larger than Mn 3+ (0.58 angstrom under low spin) and Ca 2+ (1.34 angstrom). Therefore, significant doping of Ba will lead to a tolerance factor larger than 1, resulting in a hexagonal perovskite structure. Effect of A-site dopant on stability of CaMnO3 perovskite family can also preliminarily be understood from DFT calculations. The BaMnO3/CaMnO3 structures have different magnetic ordering patterns over different temperature ranges. Table 4.1 shows the oxygen vacancy formation energies for BaMnO3 and CaMnO3 in both FM and AFM states. As can be seen, the formation energy is sensitive to the magnetic state adopted by the structure. For the AFM ordering, the CaMnO3 needs more energy for vacancy formation (2.28 ev) than BaMnO3 (1.75 ev), however, the order is reversed when the FM state is taken, i.e. the OV formation energy of CaMnO3 (1.20 ev) is much lower than that of BaMnO3 (2.63 ev). Indeed, the AFM ordered 114

139 state can only be remained at a relatively low temperature (below the Néel temperature). The computational results imply that the structural stability coupled with the oxygen donation capability can potentially be changed by substituting Ca in the A-site with larger cations such as Sr or Ba. In addition, the doping effect is likely to be correlated to the experimental temperature. Since Sr has an ionic radius slightly larger than Ca but smaller than Ba, one could anticipate Sr addition to CaxA1-xMnO3 causing similar effects as Ba addition, but to a lesser extent. Although stability of the perovskite can only be accurately estimated by comparing the relative stability of original and decomposed phases, vacancy formation energy calculations together with semi-empirical analysis give important clues to the selection of potentially effective dopants (like Sr and Ba) for the CaMnO3 parent structure. Table 4.1. Oxygen vacancy (OV) formation energy of BaMnO3 and CaMnO3 under FM (ferromagnetic) and AFM (antiferromagetic) orders E OV, FM (ev) E OV, AFM (ev) BaMnO CaMnO Thermal Stability of CaMnO3 Although previous studies revealed irreversible decomposition of CaMnO3 to Ruddlesden- Popper and spinel phases, the decomposition was observed under extended redox cycles at relatively low temperatures (<1000 C).[31] An alternative approach that can conveniently quantify phase stability of perovskites without undergoing extended redox cycles would be 115

140 desirable for efficient screening of perovskite based oxygen carriers. In the current study, temperature programmed desorption (TPD) and in-situ XRD are used to determine the stability of perovskites as well as their initial oxygen uncoupling temperature and rate of oxygen uncoupling. Figure 4.1 shows the TPD and in-situ XRD results. The sample weight loss profile in TPD indicates an initial decomposition temperature at approximately 800 C followed with significant release of lattice oxygen. This is corroborated by the in-situ XRD spectra, which exhibit significant increase in lattice parameter (decreasing 2θ angles) at above 800 C. When heated above 1000 C, the CaMnO3 phase decomposes into undesired Ca2MnO4 and CaMn2O4 structures as evidenced by the peaks formed at or near 29.5, 33, 34.5, 37.7, 39, 40.4, 43.2, and etc. 2θ angles. 116

141 Figure 4.1- (a) TPD of the pure CaMnO3 at 5 C/min to 1000 C in pure helium environment. (b) In-situ XRD of the same oxygen carrier at 5 C/min to 1200 C in argon Effect of A-Site Dopants on Oxygen Uncoupling Properties Figure 4.2- Inert temperature programmable desorption (TPD) profiles of (a) Ba-doped and (b) Sr-doped CaMnO3 ramped to 1000 C at a rate of 20 C/min in a pure helium atmosphere. 117

142 Perovskite samples with various dopant amounts are tested using TPD to observe initial oxygen release temperature and total amount of oxygen released. Differential thermal gravimetry (DTG) curves of selected samples are shown in Figure 4.2. As can be seen, CaMnO3 exhibits a significant decomposition peak at around 800 C. A similar feature is also observed for Ca0.95Ba0.05MnO3. Both Sr and Ba dopant affect the initial decomposition temperature, oxygen release rate, and total oxygen release of Ca1-xAxMnO3. Both pure SrMnO3 and BaMnO3 are more stable than CaMnO3 and, as a result, release less total oxygen. BaMnO3 offers about 7 times less total oxygen release and SrMnO3 offers about 3.5 times less oxygen compared to pure CaMnO3. When Sr and Ba are doped to the A-site, properties of both oxides can be observed. Ba and Sr doped samples tend to decompose at significantly lower temperatures (<700 C) than Ca based perovskites. When Ba is doped, the trend is apparent to shift between pure CaMnO3 and pure BaMnO3. While 0.05 Ba doping does not affect the properties of oxygen coupling, increasing the dopant amount to 0.25 can shift the initial reduction temperature by approximately 200 C. Total oxygen release for Ca0.75Ba0.25MnO3 is 3 w.t.% during the TPD experiment. This is only slightly lower than the 3.7 w.t.% oxygen release by pure CaMnO3. The rate of oxygen release in the high temperature ranges (>750 C), however, is significantly slower for Ca0.75Ba0.25MnO3. Another intriguing property is the significant weight loss observed at low temperatures (<600 C) in which almost 0.5 w.t.% oxygen is released. Further doping of Ba onto the perovskite follows a trend to reduce total oxygen release and oxygen donation rate in the high temperature range. Sr-doped samples follow a trend similar to that exhibited by the Ba-doped samples. Rates of oxygen release at high temperatures are slightly faster than the Ba-doped samples, although still slower than pure CaMnO3. Doping up to 0.25 Sr does not significantly affect the total oxygen released. Initial 118

143 temperature for oxygen release; on the other hand, is lowered. For instance, Ca0.75Sr0.25MnO3 exhibits significant weight loss in two distinct temperature regions, i.e C and > 600 C. The low temperature oxygen release could be associated with chemisorbed oxygen in the lattice structure. Figure 4.3 summarizes the total oxygen release and initial decomposition temperature for several representative perovskite samples. Figure 4.3- Total oxygen release and initial decomposition temperature comparisons between CaMnO3 and Sr- and Ba-doped synthesized oxygen carriers Structural and Stability Studies Effectiveness of dopant addition largely relies on the compatibility of the dopant and its parent structure. Ideally, dopant should be fully incorporated into the host structure to form a homogeneous solid state solution. Figure 4.4 illustrates the XRD patterns for the oxygen carriers for Ba and Sr doped CaMnO3 at a range of x=0 to x=1. As can be seen, Ba cannot be fully incorporated into the CaMnO3 structure in most cases. Even at small dopant levels (x<0.1), a secondary hexagonal BaMnO3 phase can be identified. When doped further, ternary phases, such as Mn oxides, are observed until A-site is fully occupied by Ba. Sr-doped CaMnO3, on the other hand, forms a relatively homogenous solid-solution. Sr doping up to 119

144 x=0.25 exhibits complete incorporation of Sr into CaMnO3 while shifting the structure from cubic to orthorhombic. At x=0.5, co-existence of SrMnO3 and CaMnO3 structures are observed. At higher doping levels (>0.5), a single orthorhombic SrMnO3 phase is observed. Figure 4.4- XRD of various doped CaMnO3 with (a) Sr doping and (b) Ba doping. Phases: ( ) SrMnO3, (x) CaMnO3, ( ) BaMnO3, and ( ) Mn3O4. The primary motivation for dopant addition is to stabilize the CaMnO3 structure and prevent the formation of stable spinel phases at high temperatures. Figure 4.5 investigates the in-situ XRD of the Sr- and Ba-doped CaMnO3. For the Sr-doped sample, the structure of the perovskite phase is maintained even at a high temperature of 1200 C. To compare, CaMnO3 is decomposed at 1000 C. Significant shift of 2θ angles, which are indicative of oxygen release, begins at approximately 600 C. This is consistent with TPD observations. Ca1- xbaxmno3 sample with a low dopant level (x=0.05) was chosen for in-situ XRD due to the difficulty to incorporate Ba into the CaMnO3 host structure. As shown in Figure 4.5, the Badoped sample, similar to CaMnO3, exhibited phase decomposition at 1000 C. It is therefore concluded that Sr is likely to be a more suitable dopant for CaMnO3, since it readily incorporates into the host structure and helps to enhance the stability of the perovskite phase. 120

145 Ba-doped samples, on the other hand, do not exhibit the desired CLOU properties. This may have resulted from the significant differences in ionic sizes between Ba 2+ and Ca 2+, which can lead to large lattice distortions and stresses. Figure 4.5- In-situ XRD of the (a) Ca0.75Sr0.25MnO3 and (b) Ca0.95Ba0.05MnO3 in argon at a ramping rate of 5 C/min to 1200 C Redox Testing CLOU performance of an oxygen carrier can be characterized by its oxygen donation properties under cyclic redox conditions. In order to determine the oxygen donation properties of the reference (CaMnO3) and Sr doped (Ca0.75Sr0.25MnO3) oxygen carrier samples, shortterm isothermal redox experiments are carried out at 650 C, 750 C, 850 C, and 950 C. The results are summarized in Figure

146 Figure 4.6- Isothermal oxygen donation at temperatures between C. Reduction step: 5.0 grade helium. Oxidation step: 10% O2. As can be seen, undoped CaMnO3 does not observe any noticeable weight loss at 650 C. Once the temperature reaches 850 C, significant oxygen release is observed. When the temperature is increased even further, the oxygen carrier exhibits even higher oxygen donation. When doping Sr and Ba into CaMnO3, different phenomena are observed. When Ba is doped, there is no indication that the material is being reoxidized to any extent at any temperature using 10% O2. From XRD of the reduced sample, the formation of spinel CaMn2O4 phase is observed. This may indicate that Ba could be distorting the structure to promote the decomposition of the cubic CaMnO3 phase. In-situ XRD (Figure 4.5b) also corroborates the formation of spinel and reduced perovskite phases at a lower temperature than pure CaMnO3. When Sr is doped into the CaMnO3 structure, it promotes oxygen donation at a much lower temperature while still being able to reoxidize. While at 650 C, ~0.1 w.t.% oxygen release is observed. Once the material is heated to 850 and 950 C, total oxygen release becomes more significant. In pure CaMnO3, near complete oxidation using 10% oxygen is achieved. 122

147 However, Sr-doped CaMnO3 is only able to restore roughly 99% of the initial weight. Each sample was cycled 5 times at each temperature and no change in performance was observed. In chemical looping schemes, long term cycling is important due to the need to replenish the solid oxygen carriers over time. Figure 4.7 examines the Sr-doped perovskite at 850 C for 100 uncoupling cycles. Over 100 cycles, the material has no observable deactivation and is very stable throughout all 100 cycles. After 100 cycles are performed, oxygen release is still observed at lower temperatures 650 C and 750 C. The oxygen capacity and rate of oxygen release at these lower temperatures are comparable to what was observed in Figure 4.6. XRD analysis shows a small amount of SrMnO3 appearing after the 100 cycles (See Figure 4.S2). However, no spinel or Ruddlesdon-Popper phases were identified. Figure cycles at 850 C Ca0.75Sr0.25MnO3 oxygen carrier. Initial coal char reactions were conducted in a TGA. Coal char was well mixed with the oxygen carrier and placed in the TGA and heated to 250 C to remove any moisture. The TGA is then heated to 950 C in helium and the reaction is allowed to proceed to completion. Figure 4.8 examines the results of pure CaMnO3 versus a Sr-doped sample. The Sr-doped oxygen carrier 123

148 has an initial reaction temperature of approximately 100 C lower compared to the undoped CaMnO3 oxygen carrier. Besides the lower reaction temperature, the Sr-doped has a higher activity in reaching full conversion of solid char. The results indicate that Sr-doped CaMnO3, not only being more stable, but also more active for solid fuel combustion. Figure 4.8- Char combustion using TGA of pure and Sr-doped CaMnO3. Heating in inert to 250 C for drying, then heated at a rate of 20 C/min. in helium to 950 C. Fluidized bed results are shown in Figure 4.9. Examination of the solid char cycles of the Ca0.75Sr0.25MnO3 reveals nearly 100% conversion of the coal char in every cycle. This is confirmed through no observable CO2 during the oxidation step. Further comparison of the CO2 to CO selectivity reveals around 90% selectivity during every cycle. The fluidized bed results confirm the viability of the Ca0.75Sr0.25MnO3 oxygen carrier in CLOU applications. 124

149 Figure 4.9- Char conversion and CO2 selectivity of the Ca0.75Sr0.25MnO3 oxygen carrier in a fluidized bed reactor. Temperature: 850 C Gas Flow Rates: nitrogen: 800mL/min and helium: 280mL/min (Char Reduction) and oxygen: 120mL/min (during oxidation only). 4.4 Conclusions The present study investigates the redox properties of doped CaMnO3 as oxygen carriers for CLOU processes. The effects of A-site substitution with Sr and Ba dopants on structural stability, oxygen carrying capacity, and oxygen donation temperature of the oxygen carriers are investigated. The A-site dopants are chosen through both a semi-empirical approach and DFT calculations for oxygen vacancy formation energy. Ba addition to the CaMnO3 structure is found to be ineffective due to the inability of Ba to incorporate into the CaMnO3 structure. Sr, on the other hand, is found to form homogeneous solid solutions with the CaMnO3 host structure. Sr-doped samples are also found to be effective in preventing irreversible decomposition of the parent perovskite structure to the Ruddlesden-Popper (Ca2MnO4) and spinel (CaMn2O4) phases. Excellent stability is observed for Sr doped sample (Ca0.75Sr0.25MnO3) through various conditions including heating to 1200 C, 100 isothermal redox cycles, and reaction with solid fuels. Besides the increased stability of the structure, Srdoped oxygen carriers exhibit noticeable oxygen release at significantly lower temperature 125

150 than that observed for pure CaMnO3. Due to its various advantages compared to undoped CaMnO3, Sr-doped CaMnO3 oxygen carrier offers a potentially viable option for solid fuel conversion in CLC schemes. Acknowledgements Funding supports from the U.S. Department of Energy (Award Number FE001247) and Illinois Clean Coal Institute (Award Number 14/5C-2) are greatly appreciated. We acknowledge the use of the Analytical Instrumentation Facility (AIF) at North Carolina State University, which is supported by the State of North Carolina and the National Science Foundation. We would also like to acknowledge Feng He for his help in setting up some of the experiments. 126

151 REFERENCES [1] Figueroa JD, Fout T, Plasynski S, McIlvried H, Srivastava RD. Advances in CO2 capture technology The U.S. Department of Energy s Carbon Sequestration Program. Int. J. Greenh. Gas Con. 2008;2:9 20. doi: /s (07) [2] Hansen J, Sato M. Greenhouse gas growth rates. Proc. Natl. Acad. Sci. USA 2004;101: [3] Adanez J, Abad A, Garcia-Labiano F, Gayan P, de Diego LF. Progress in chemicallooping combustion and reforming technologies. Prog. Energ. Combust. 2012;38: doi: /j.pecs [4] Fan L-S, Li F. Chemical looping technology and its fossil energy conversion applications. Ind. Eng. Chem. Res. 2010;49: doi: /ie [5] Fan L-S. Chemical looping systems for fossil energy conversions [6] Lyngfelt A. Chemical-looping combustion of solid fuels status of development. Appl. Energ. 2014;113: doi: /j.apenergy [7] Sorgenfrei M, Tsatsaronis G. Design and evaluation of an IGCC power plant using ironbased syngas chemical-looping (SCL) combustion. Appl. Energ. 2014;113: doi: /j.apenergy [8] Tong A, Bayham S, Kathe MV, Zeng L, Luo S, Fan L-S. Iron-based syngas chemical looping process and coal-direct chemical looping process development at Ohio State University. Appl. Energ. 2014;113: doi: /j.apenergy [9] Scott SA, Dennis JS, Hayhurst AN, Brown T. In situ gasification of a solid fuel and CO2 separation using chemical looping. AIChE J 2006;52: doi: /aic [10] Leion H, Mattisson T, Lyngfelt A. The use of petroleum coke as fuel in chemicallooping combustion. Fuel 2007;86: doi: /j.fuel [11] Luo S, Bayham S, Zeng L, McGiveron O, Chung E, Majumder A, et al. Conversion of metallurgical coke and coal using a Coal Direct Chemical Looping (CDCL) moving bed reactor. Appl. Energ. 2014;118: doi: /j.apenergy

152 [12] Cuadrat A, Linderholm C, Abad A, Lyngfelt A, Adánez J. Influence of limestone addition in a 10 kwth chemical-looping combustion unit operated with petcoke. Energ. Fuel. 2011;25: doi: /ef200806q. [13] Linderholm C, Lyngfelt A, Cuadrat A, Jerndal E. Chemical-looping combustion of solid fuels operation in a 10kW unit with two fuels, above-bed and in-bed fuel feed and two oxygen carriers, manganese ore and ilmenite. Fuel 2012;102: doi: /j.fuel [14] Leion H, Mattisson T, Lyngfelt A. Solid fuels in chemical-looping combustion. Int. J. Greenh. Gas Con. 2008;2: doi: /s (07)00117-x. [15] Arjmand M, Leion H, Mattisson T, Lyngfelt A. Investigation of different manganese ores as oxygen carriers in chemical-looping combustion (CLC) for solid fuels. Appl. Energ. 2014;113: doi: /j.apenergy [16] Shen L, Wu J, Xiao J. Experiments on chemical looping combustion of coal with a NiO based oxygen carrier. Combust. Flame 2009;156: doi: /j.combustflame [17] Shen L, Wu J, Xiao J, Song Q, Xiao R. Chemical-looping combustion of biomass in a 10 kwth reactor with iron oxide as an oxygen carrier. Energ. Fuel. 2009;23: doi: /ef900033n. [18] Mattisson T, Lyngfelt A, Leion H. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int. J. Greenh. Gas Con. 2009;3:11 9. doi: /j.ijggc [19] Mattisson T. Materials for chemical-looping with oxygen uncoupling. ISRN Chem. Eng. 2013;2013:1 19. doi: /2013/ [20] Imtiaz Q, Hosseini D, Müller CR. Review of oxygen carriers for chemical looping with oxygen uncoupling (CLOU): thermodynamics, material development, and synthesis. Energy Technol. 2013;1: doi: /ente [21] Azimi G, Leion H, Mattisson T, Lyngfelt A. Chemical-looping with oxygen uncoupling using combined Mn-Fe oxides, testing in batch fluidized bed. Energy Procedia 2011;4: doi: /j.egypro

153 [22] Azimi G, Leion H, Rydén 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: doi: /ef301120r. [23] Shafiefarhood A, Stewart A, Li F. Iron-containing mixed-oxide composites as oxygen carriers for chemical looping with oxygen uncoupling (CLOU). Fuel 2015;139:1 10. doi: /j.fuel [24] Shulman A, Cleverstam E, Mattisson T, Lyngfelt A. Manganese/iron, manganese/nickel, and manganese/silicon oxides used in chemical-looping with oxygen uncoupling (CLOU) for combustion of methane. Energ. Fuel. 2009;23: doi: /ef [25] Shulman A, Cleverstam E, Mattisson T, Lyngfelt A. Chemical Looping with oxygen uncoupling using Mn/Mg-based oxygen carriers oxygen release and reactivity with methane. Fuel 2011;90: doi: /j.fuel [26] 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: doi: /ef201329x. [27] Arjmand M, Keller M, Leion H, Mattisson T, Lyngfelt A. Oxygen release and oxidation rates of MgAl2O4-supported CuO oxygen carrier for chemical-looping combustion with oxygen uncoupling (CLOU). Energ. Fuel. 2012;26: doi: /ef [28] Gayán 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: doi: /j.fuel [29] Pour NM, Azimi G, Leion H, Rydén M, Lyngfelt A. Production and examination of oxygen-carrier materials based on manganese ores and Ca(OH)2 in chemical looping with oxygen uncoupling. AIChE J. 2014;60: doi: /aic [30] Rydén M, Leion H, Mattisson T, Lyngfelt A. Combined oxides as oxygen-carrier material for chemical-looping with oxygen uncoupling. Appl. Energ. 2014;113: doi: /j.apenergy

154 [31] Arjmand M, Hedayati A, Azad A-M, Leion H, Rydén M, Mattisson T. CaxLa1- xmn1-ymyo3-δ (M = Mg, Ti, Fe, or Cu) as oxygen carriers for chemical-looping with oxygen uncoupling (CLOU). Energ. Fuel doi: /ef [32] Arjmand M, Kooiman RF, Rydén M, Leion H, Mattisson T, Lyngfelt A. Sulfur tolerance of CaxMn1 ymyo3 δ (M = Mg, Ti) perovskite-type oxygen carriers in chemical-looping with oxygen uncoupling (CLOU). Energ. Fuel. 2014;28: doi: /ef402383v. [33] Hallberg P, Jing D, Rydén M, Mattisson T, Lyngfelt A. Chemical looping combustion and chemical looping with oxygen uncoupling experiments in a batch reactor using spray-dried CaMn1 xmxo3 δ (M = Ti, Fe, Mg) particles as oxygen carriers. Energ. Fuel. 2013;27: doi: /ef [34] 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: doi: /ef900444d. [35] Rydén M, Lyngfelt A, Mattisson T. CaMn0.875Ti0.125O3 as oxygen carrier for chemicallooping combustion with oxygen uncoupling (CLOU) experiments in a continuously operating fluidized-bed reactor system. Int. J. Greenh. Gas Con. 2011;5: doi: /j.ijggc [36] Sundqvist S, Leion H, Rydén M, Lyngfelt A, Mattisson T. CaMn0.875Ti0.125O3 δ as an oxygen carrier for chemical-looping with oxygen uncoupling (CLOU)-solid-fuel testing and sulfur interaction. Energ. Technol. 2013;1: doi: /ente [37] Källén M, Rydén M, Dueso C, Mattisson T, Lyngfelt A. CaMn0.9Mg0.1O3-δ as oxygen carrier in a gas-fired 10 kwth chemical-looping combustion unit. Ind. Eng. Chem. Res. 2013;52: doi: /ie303070h. [38] Pishahang M, Larring Y, McCann M, Bredesen R. Ca0.9Mn0.5Ti0.5O3 δ : A suitable oxygen carrier material for fixed-bed chemical looping combustion under syngas conditions. Ind. Eng. Chem. Res. 2014;53: doi: /ie500928m. [39] Moldenhauer P, Rydén M, Mattisson T, Hoteit A, Jamal A, Lyngfelt A. Chemicallooping combustion with fuel oil in a 10 kw pilot plant. Energ. Fuel. 2014;28: doi: /ef

155 [40] Dai XP, Li J, Fan JT, Wei WS, Xu J. Synthesis gas generation by chemical-looping reforming in a circulating fluidized bed reactor using perovskite LaFeO3-based oxygen carriers. Ind. Eng. Chem. Res. 2012;51: doi: /ie300033e. [41] He F, Li X, Zhao K, Huang Z, Wei G, Li H. The use of La1-xSrxFeO3 perovskite-type oxides as oxygen carriers in chemical-looping reforming of methane. Fuel 2013;108: doi: /j.fuel [42] Mihai O, Chen D, Holmen A. Chemical looping methane partial oxidation: the effect of the crystal size and O content of LaFeO3. J Catal. 2012;293: doi: /j.jcat [43] Murugan A, Thursfield A, Metcalfe IS. A chemical looping process for hydrogen production using iron-containing perovskites. Energ. Environ. Sci. 2011;4:4639. doi: /c1ee02142g. [44] Readman JE, Olafsen A, Larring Y, Blom R. La0.8Sr0.2Co0.2Fe0.8O3 δ as a potential oxygen carrier in a chemical looping type reactor, an in-situ powder X-ray diffraction study. J. Mater. Chem. 2005;15:1931. doi: /b416526h. [45] Rydén M, Lyngfelt A, Mattisson T, Chen D, Holmen A, Bjørgum E. Novel oxygencarrier materials for chemical-looping combustion and chemical-looping reforming; LaxSr1 xfeyco1 yo3 δ perovskites and mixed-metal oxides of NiO, Fe2O3 and Mn3O4. Int. J. Greenh. Gas Con. 2008;2: doi: /s (07) [46] Sarshar Z, Sun Z, Zhao D, Kaliaguine S. Development of sinter-resistant core shell LaMnxFe1 xo3@msio2 oxygen carriers for chemical looping combustion. Energ. Fuel. 2012;26: doi: /ef [47] Zhao K, He F, Huang Z, Zheng A, Li H, Zhao Z. Three-dimensionally ordered macroporous LaFeO3 perovskites for chemical-looping steam reforming of methane. Int. J. Hydrogen Energ. 2014;39: doi: /j.ijhydene [48] Sarshar Z, Kleitz F, Kaliaguine S. Novel oxygen carriers for chemical looping combustion: La1 xcexbo3 (B = Co, Mn) perovskites synthesized by reactive grinding and nanocasting. Energ. Environ. Sci. 2011;4: doi: /c1ee01716k. [49] Noorman S, Gallucci F, van Sint Annaland M, Kuipers JAM. Experimental investigation of chemical-looping combustion in packed beds: a parametric study. Ind. Eng. Chem. Res. 2011;50: doi: /ie

156 [50] Chen Y, Galinsky N, Wang Z, Li F. Investigation of perovskite supported composite oxides for chemical looping conversion of syngas. Fuel 2014;134: doi: /j.fuel [51] Galinsky NL, Huang Y, Shafiefarhood A, Li F. Iron oxide with facilitated O 2- transport for facile fuel oxidation and CO2 capture in a chemical looping scheme. ACS Sustain. Chem. Eng. 2013;1: doi: /sc300177j. [52] Shafiefarhood A, Galinsky N, Huang Y, Chen Y, Li F. Fe2O3@LaxSr1 xfeo3 core-shell redox catalyst for methane partial oxidation. ChemCatChem 2014;6: doi: /cctc [53] Neal LM, Shafiefarhood A, Li F. Dynamic methane partial oxidation using a Fe2O3@La0.8Sr0.2FeO3-δ core shell redox catalyst in the absence of gaseous oxygen. ACS Catal. 2014;4: doi: /cs [54] Galinsky NL, Shafiefarhood A, Chen Y, Neal L, Li F. Effect of support on redox stability of iron oxide for chemical looping conversion of methane. Appl. Catal. B- Environ. 2015;164: doi: /j.apcatb [55] Peña MA, Fierro JLG. Chemical structures and performance of perovskite oxides. Chem. Rev. 2001;101: doi: /cr980129f. [56] He F, Trainham J, Parsons G, Newman JS, Li F. A hybrid solar-redox scheme for liquid fuel and hydrogen coproduction. Energ. Environ. Sci. 2014;7:2033. doi: /c4ee00038b. [57] Mihai O, Chen D, Holmen A. Chemical looping methane partial oxidation: the effect of the crystal size and O content of LaFeO3. J Catal. 2012;293: doi: /j.jcat [58] Jing D, Mattisson T, Leion H, Rydén M, Lyngfelt A. Examination of perovskite structure CaMnO3-δ with MgO addition as oxygen carrier for chemical looping with oxygen uncoupling using methane and syngas. Int. J. Chem. Eng. 2013;2013:1 16. doi: /2013/ [59] Bakken E, Norby T, Stolen S. Nonstoichiometry and reductive decomposition of CaMnO3-δ. Solid State Ionics 2005;176: doi: /j.ssi

157 [60] Cabello A, Abad A, Gayán P, de Diego LF, García-Labiano F, Adánez J. Effect of operating conditions and H2S presence on the performance of CaMg0.1Mn0.9O3 δ perovskite material in chemical looping combustion (CLC). Energ. Fuel. 2014;28: doi: /ef [61] Fossdal A, Bakken E, Øye BA, Schøning C, Kaus I, Mokkelbost T, et al. Study of inexpensive oxygen carriers for chemical looping combustion. Int. J. Greenh. Gas Con. 2011;5: doi: /j.ijggc [62] Hallberg P, Källén M, Jing D, Snijkers F, van Noyen J, Rydén M, et al. Experimental investigation of CaMnO3-δ based oxygen carriers used in continuous chemical-looping combustion. Int. J. Chem. Eng. 2014;2014:1 9. doi: /2014/ [63] Kresse G, Hafner J. Ab initio molecular dynamics for open-shell transition metals. Phys. Rev. B. 1993;48: doi: /physrevb [64] Kresse G, Hafner J. Ab initio molecular-dynamics simulation of the liquid-metal amorphous-semiconductor transition in germanium. Phys. Rev. B. 1994;49: doi: /physrevb [65] Blöchl PE. Projector augmented-wave method. Phys. Rev. B. 1994;50: doi: /physrevb [66] Kresse G. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B. 1999;59: doi: /physrevb [67] Perdew JP, Burke K, Ernzerhof M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1996;77: doi: /physrevlett [68] Bhalla AS, Guo R, Roy R. The perovskite structure - a review of its role in ceramic science and technology. Mater. Res. Innov. 2000;4:3 26. doi: /s [69] Lufaso MW, Woodward PM. Prediction of the crystal structures of perovskites using the software program SPuDS. Acta Crystallogr. B. 2001;57: doi: /s

158 [70] Dabrowski B, Chmaissem O, Mais J, Kolesnik S, Jorgensen JD, Short S. Tolerance factor rules for Sr1 x ycaxbaymno3 perovskites. J. Solid State Chem. 2003;170: doi: /s (02)

159 CHAPTER 5 CaMn1-xBxO3-δ (B= Al, V, Fe, Co, and Ni) perovskite based oxygen carriers for chemical looping with oxygen uncoupling (CLOU) Nathan L. Galinsky, Marwan Sendi, Lindsey Bowers, and Fanxing Li* Department of Chemical and Biomolecular Engineering, North Carolina State University, 911 Partners Way, Raleigh, NC , USA. Chapter 5 is a reprint of a manuscript published in Applied Energy, 2016, 174, The supplementary information is in Appendix C. *To whom the correspondence should be addressed. Telephone: +1 (919) Fax: +1(919) fli5@ncsu.edu. 135

160 Abstract Operated under a cyclic redox mode in the presence of an oxygen carrier, the chemical looping with oxygen uncoupling (CLOU) process has the potential to effectively combust solid carbonaceous fuels while capturing CO2. The overall process is enabled by an oxygen carrier that is capable of reversibly exchanging its lattice oxygen (O 2- ) with gaseous oxygen (O2) under varying external oxygen partial pressures (PO2). As such, further improvements of the CLOU process relies largely on the identification of oxygen carriers with higher activity, better recyclability, and improved resistance towards physical degradation. This article investigates dopant effects on CLOU properties of oxygen carriers with a CaMnO3 parent structure. Various B-site compatible metal cations including Fe, Ni, Co, V, and Al are incorporated into the perovskite. While CaMnO3 suffers from stability issues resulting from irreversible transitions to spinel (CaMn2O4) and Ruddlesden-Popper (Ca2MnO4) structures under typical CLOU redox conditions, a number of B-site doped perovskites exhibited promising phase stability and redox activity. Of the oxygen carriers investigated, Fe-doped CaMnO3 exhibits the most promising CLOU properties while showing high compatibility with the CaMnO3 parent structure. In terms of redox performance, CaMn1-xFexO3-δ exhibit notable redox activity at temperatures as low as 600 C. No deactivation was observed over 100 redox cycles. The doped perovskite structure was also significantly more stable than undoped CaMnO3, showing no signs of decomposition at 1200 C. When operated under identical conditions, the Fe-doped oxygen carrier is observed to achieve significantly higher conversion of Pittsburgh #8 coal char compared to undoped CaMnO3 oxygen carrier, when operated at 850 C. 136

161 5.1 Introduction Concerns over global climate change resulting from anthropogenic CO2 emissions necessitate development of new power generation technologies with lower carbon footprints. Of the various carbon capture technologies, chemical looping combustion (CLC) has emerged as a promising option. [1,2] The CLC processes aims at lowering the energy penalty for CO2 separation in combustion processes. This is achieved though cyclic redox reactions using interconnected fluidized bed or moving bed reactors. [3 8] Solid oxygen carrier particles are used to carry oxygen and heat from air to the fuel without mixing these two components. As a result, a concentrated CO2 stream can be generated in the fuel reactor. The oxygen depleted carrier is then transported to a second reactor referred to as the air reactor. Here, air is used to regenerate the oxygen carrier while producing heat for power generation. Development of low cost oxygen carriers with high activity for carbonaceous fuel conversions is highly desirable for chemical looping processes. While many oxygen carriers are effective for gaseous fuel combustion, they tend to be less active in converting solid fuels such as biomass or coal due to mass transfer and kinetic limitations. Enhancement of oxygen carrier activity towards solid carbon-based fuels is typically achieved through two methods. First, insitu gasification uses CO2 or steam to gasify the solid fuel into syngas for oxygen carrier conversion. [6 11] However, gasification of solid fuels with steam or CO2 is often slow and becomes a rate limiting step. A variety of oxygen carriers including oxides of Fe [8,10,12 14], Mn [13,15], and Ni [16,17] have been used in the conversion of solid fuels through the in-situ gasification approach. The second method is known as chemical looping with oxygen uncoupling (CLOU). [3,18,19] In CLOU, the oxygen carrier is selected from metal oxides with significant equilibrium partial pressure of oxygen (PO2) at high temperature. As a result, the 137

162 oxygen carrier is able to release a portion of its lattice oxygen into the gas phase. The gaseous oxygen then reacts with the solid fuel, leading to improved combustion kinetics. Typical oxygen carriers that exhibit CLOU properties include oxides and mixed oxides containing Cu, Mn, and/or Co. [18 30] However, each oxide exhibits challenges, such as high cost and agglomeration of Cu oxides [19,26 28], low decomposition temperature and toxicity of Co oxides [18,23], and difficulty of regeneration of Mn oxides [15,19,21 25,31 36]. Recently, perovskite structured oxides have received increasing attention as oxygen carriers for redox applications. [23,31,33 56] Typically, perovskites take the form of ABO3-δ, where A is a large cation of either the alkali or rare earth metal, and B is a smaller cation of the transition metal group. [57] Perovskite structured oxides have been used both as supports and primary oxygen carriers. As a support, mixed ionic and electronic conductive (MIEC) perovskites such as La1-xSrxFeO3 (LSF) have shown to enhance the redox activity of iron oxides by nearly two orders of magnitude. [51 54] Perovskite and perovskite supported iron oxides have also been explored as redox catalysts for syngas generation and water-splitting. [42,56,55] La containing perovskite supports have been shown to enhance the oxygen donation properties of Mn-Fe and Co-Fe oxides. [23] Up to 8.8% decrease in the initial decomposition temperature was reported for perovskite supported mixed metal oxides. Besides its role as a support, perovskite oxygen carriers have been shown to be effective as standalone oxygen carriers in CLOU and CLC applications. CaMnO3 perovskites are of interest due to its low cost, availability, and CLOU properties. [20,29 39,49,58 62] CaMnO3 based oxygen carriers have lower oxygen uncoupling capacities than copper oxides but can be synthesized from cheap manganese ores and Ca precursors [29,61], while obtaining high solid fuel conversions. [29,31,34,36] CaMnO3 faces the challenge of long term stability. CaMnO3 138

163 oxygen carriers undergo an irreversible phase change to spinel (CaMn2O4) and Ruddlesden- Popper (Ca2MnO4) phases. [19,20,32,59,60] In addition, CaMnO3 has also been shown to be susceptible to sulfur poisoning, forming CaSO4. [32,36,60] Stabilization of the perovskite structure has been investigated through addition of secondary metal atoms into the A- and B- sites of the parent CaMnO3. [29,31 38,60,61,63 65] Arjmand et. al. investigated Ca1-xLaxMn1-yMyO3 (where x=0 or 0.1 and M= Mg, Ti, Fe, or Cu) oxygen carriers for CLOU properties between C in a laboratory scale fluidized bed reactor. [31] It was determined that undoped CaMnO3 had the highest oxygen release capability (~0.7% after 360s). Doping of the A- and B-site was shown to decrease the oxygen release capability except when Cu was used. However, Cu-doped samples were prone to defluidization. Rydèn et. al. studied a Ti-doped CaMnO3 oxygen carrier for CLOU. [35] The oxygen carrier exhibited oxygen uncoupling properties above 720 C in a fluidized bed reactor. The maximum rate for oxygen uncoupling was observed to be 0.03LN/min at 950 C with a maximum concentration at 4 volume% when a mass of 200 g of the oxygen carrier is used. The oxygen carrier was also tested in a redox mode using methane as a fuel. Slight deactivation was observed. The deactivation was attributed to CaMn2O4 formation. Kallèn et. al. used CaMn1-xMgxO3 oxygen carrier in a 10kWth gas-fired CLC unit. [37] The oxygen carrier was observed to release oxygen above 700 C. No agglomeration was observed for 120 hr of high temperature operations. The total attrition of fines (<45 μm) was determined to be less than 0.01%. Hallberg et. al. investigated Mg, Ti, and Fe doped CaMnO3 oxygen carriers prepared by spray drying for CLOU properties and redox activity with methane. [33] The oxygen carriers had oxygen release of % by weight at 900 C with a maximum release of nearly 1% by weight at 1000 C. After 3 cycles of oxygen uncoupling and one methane cycle, the 139

164 authors reported the formation of CaMn2O4 spinel phase in all tested oxygen carriers. CaMnO3 oxygen carriers have also been synthesized from combinations of manganese ores and Ca(OH)2.[29] In reaction with methane, ores containing high concentrations of Al showed poor performance due to spinel formation between Al and Mn. The best performing oxygen carrier had an oxygen uncoupling capacity of 0.68 w.t.%. Although a number of doped CaMnO3 oxygen carriers have shown promise for the CLOU process, detailed characterization of dopant effects on B-site doped CaMnO3 oxygen carriers and their long term performance has not been extensively studied. In this study, we investigate the effect of B-site doping for CaMnO3 based oxygen carriers. Fe, Co, V, Ni, and Al are chosen as potential dopants. It was determined that oxygen release properties of the oxygen carriers vary with dopant type and concentrations. Of the various dopants chosen, Fe doped oxygen carrier exhibits long term stability and has low temperature CLOU properties around 650 C. Fe dopants also induced noticeable α-oxygen release between C. The Fe-doped samples exhibit long term stability (100 cycles) of its uncoupling properties. As a result, Fe-doped samples observe more facile oxygen release for solid fuel conversions relative to undoped CaMnO3 oxygen carriers. 5.2 Experimental Oxygen Carrier Synthesis Mixed oxides with a general formula of CaMn1-xBxO3 (B= V, Fe, Co, Ni, and Al and x 0.5) are prepared using a citric acid sol-gel method. General procedure for the sol-gel samples includes dispersion of cation precursors Ca(NO3)2 4H2O (Sigma Aldrich), Fe(NO3)3 9H2O (98+%, Sigma Aldrich), Co(NO3)2 6H2O (ACS Reagent, Noah), Ni(NO3)2 6H2O (Sigma 140

165 Aldrich), VCl3 (97%, Sigma Aldrich), Al(NO3)3 9H2O (ACS Reagent, Sigma Aldrich) and Mn(NO3)2 4H2O (Sigma Aldrich) in distilled water followed by addition of citric acid. In the synthesis procedure used, citric acid (Sigma Aldrich) is added to the solution in a molar ratio of 2.5:1 (citric acid: total cations) as a chelating agent. The mixture is heated to 80 C, and ethylene glycol (Sigma Aldrich) is added as a gel promoter in a molar ratio of 1.5:1 (ethylene glycol: citric acid). After most of the water is boiled out, the sol-gel sample is dried in a vacuum oven for 12 hours to remove any extra moisture. The dried sample is then annealed in air at 1200 C for 12 hours. The resulting powder is crushed and sieved to under 90 micron particle size. A solid state reaction (SSR) method is used to scale up the synthesis process. The SSR method procedure requires powder preparation, followed by pelletization, and annealing/solid state reaction. Preparation of CaMnO3 and CaMn0.75Fe0.25O3 oxygen carriers are conducted by mixing calculated amounts of CaCO3 ( 99.0%, Sigma-Aldrich), Fe2O3 (99.9%, Noah), and Mn3O4 (99.0%, Noah) precursors for 12 hours in a planetary ball mill (XBM4X, Columbia International) at a rotation speed of 250 rpm. The mixture is then pressed into ¾ pellets by a hydraulic press (YLJ-15T, MTI Corporation) at a pressure of 20MPa and sintered in a high temperature furnace (GSL-1500-X50, MTI Corporation) at 1200 C for 24 hours. After annealing, the particles are fractured and sieved into a range of microns Oxygen Uncoupling Experiments Initial oxygen uncoupling properties of the oxygen carriers are characterized in a TGA (Q600 TA Instruments). Two primary experiments are conducted in the TGA: (1) temperature programmed desorption TPD experiments using inert helium (5.0, Airgas) and argon (5.0, Airgas) up to 1000 C at a ramp of 10 C/min and (2) isothermal experiments between

166 950 C. During the TPD experiments, the oxygen carriers are pre-treated in 25% O2 (Extra Dry, Airgas) balanced with Ar at 850 C for one hour and cooled down at a rate of 10 C/min to 100 C where the system is purged with a mixture of helium and argon to remove any excess oxygen gas. A TPD experiment is then performed on the oxygen carriers using a 10 C/min ramp to 1000 C in a helium and argon mixture. Isothermal experiments are conducted at various temperatures to determine the oxygen uncoupling capabilities at different temperatures. At each temperature, 5 cycles of switching between pure inert and 10% O2 balanced with a mixture of argon and helium are conducted for short term recyclability of the oxygen carrier. Best performing oxygen carriers are also tested for 100 cycles at 850 C to demonstrate the oxygen uncoupling performance of the perovskite oxygen carriers over the long term Fluidized Bed Experiments A laboratory-scale fluidized-bed reactor is used for solid fuel conversion experiment (see Figure 5.S1). The quartz fluidized-bed reactor has an outer diameter of 25.4mm. The reactor is heated through external heating from a tube furnace (MTI OTF-1200X-S-VT). Temperature is measured inside the reactor with an Inconel (Type K) thermocouple to set the temperature of the reactor to 850 C. A 3.75mm ID stainless steel tube is used for solid fuel injection into the reactor. A gas mixing panel is used to send the desired ratio of gases from both the bottom of the reactor for fluidization and through the smaller stainless steel tube for solid fuel injection. For maintaining fluidization, flow rates of 1080 and 1200mL/min (at S.T.P) are used for solid fuel conversion and oxygen carrier regeneration (in 10% O2), respectively. The flow rates correspond to 6-7 times of the calculated minimum fluidization velocity (Umf) for the synthesized oxygen carriers. Prior to experiments, the reactor is loaded with 16 mesh silicon 142

167 carbide (Kramer Industries) at the bottom of the reactor. The SiC particles act as a gas preheater and distributor. Approximately 15 grams of the oxygen carrier is placed into the reactor for the experiments. For redox reactions of solid fuel, Pittsburgh #8 coal (-60 mesh) is converted into coal char through heating in the reactor at 950 C in a N2 environment. The resulting char is sieved into a range of microns. Injection of the coal is conducted in a pulse mode by loading from the top of the reactor through the small stainless steel tube and is pulsed into the bottom of the oxygen carrier layer of the reactor bed using N2. 20 redox cycles are tested using the fluidized bed reactor conducting two experiments in those 20 cycles. The conditions for each cycle are given in Table 5.1. Exhaust gas concentrations are measured using a quadrupole mass spectrometer (MKS Cirrus 2) and a near-ir based gas analyzer (Emerson X-Stream). Table 5.1- Gas conditions for redox cycles used during fluidized bed experiments. Cycle Number Gas conditions 10% O2 with coal injection for reference (balance grade 5.0 N2) 100% N2 with char injection (reduction step)/10% O2 balance N2 (oxidation step) 100% N2 without coal injection (reduction step)/10% O2 balanced with N2 for oxidation (oxidation step) 100% N2 with char injection (reduction step)/10% O2 balanced with N2 for oxidation (oxidation step) Sample Characterizations Phase identification of the oxygen carriers is conducted using X-ray powder diffraction. Powdered samples are prepared and tested using a Rigaku SmartLab X-ray diffractometer with 143

168 Cu-Kα (λ=0.1542nm) radiation operating at 40kV and 44mA. A scanning range of (2θ) with a step size of 0.1 holding for 3.5 seconds at each step is used to generate the XRD patterns. To examine phase properties during TPD experiments, in-situ X-ray diffraction was conducted using an Empyrean PANalytical XRD using a similar Cu-Kα radiation operating at 45kV and 40mA. A 2θ range of is used at a ramp rate of 0.1 holding each step for 0.1s. A TPD experiment is conducted in the in-situ XRD by heating the sample at 5 C/min in pure argon (Grade 5.0). XRD scans are conducted the entire length of the ramp to 1200 C. 5.3 Results and Discussion Effect of Dopant into the B-site Effectiveness of dopant addition largely relies on the compatibility of the dopant and its parent structure. Ideally, dopants should be fully incorporated into the host structure to form a homogeneous solid solution. Figure 5.1 illustrates the XRD spectra of the various B-site doped CaMnO3 oxygen carriers. Figure 5.1- XRD spectra of the CaMn0.8B0.2O3 oxygen carriers. ( )CaMnO3, ( )Ca3Co1.5Mn0.5O6, ( ) V2O5 and ( ) NiO phases 144

169 Of the dopants, Fe is effectively incorporated into the structure without any secondary phases. Ni, V, and Co doped CaMnO3 oxygen carriers are present with separate identified phases of NiO, V2O5, and mixed oxide phases respectively. Al has an insignificant amount of unidentified phase or phases, this resulted from the low intensity of these peaks and limited database of the HighScore Plus software. Lattice parameter refinement of the base orthorhombic CaMnO3 crystal structure with and without dopants from Figure 5.1 can be seen in Table 5.2. The change of the lattice parameters and volume of the unit cell stems from the ionic radii and oxygen vacancies formed during preparation. From table 5.2, it is observed that the volume of the unit cell decreases with all dopants. The decrease in volume is likely to stem from the stabilization of the crystal structure. As a result, less oxygen vacancy formation at higher temperature is observed despite a change in ionic radii when these dopants replace the Mn 4+ ion in the CaMnO3 structure. This is confirmed through isothermal oxygen uncoupling studies of the undoped and Fe-doped oxygen carriers (see Figure 5.6) which we will discuss later. Additionally, shifts in peak locations can indicate substitution into the A- or B-site of the perovskite. For example, the diffraction peak for the (110) plane of CaMnO3 shifts from to 33.8 corresponding to shift in the lattice from Fe incorporation. Table 5.2-Lattice parameters for the CaMn0.8B0.2O3 oxygen carriers Dopant a (Å) b (Å) c (Å) Volume (Å 3 ) No Dopant Fe Co V Ni Al

170 Through addition of secondary metals into the A- or B-site of the perovskite, it is possible to improve the stability of the parent CaMnO3 structure while obtaining more favorable oxygen uncoupling properties. [63] Figure 5.2 examines the addition of the dopants on the oxygen uncoupling properties of the oxygen carriers using temperature programmed desorption (TPD) and differential thermal gravimetry (DTG) curves. Figure 5.2- DTG/TPD curves of CaMn0.8B0.2O3 oxygen carriers (B=Fe, V, Co, Ni, and Al) to 1000 C at 10 C/min. in a pure inert atmosphere after treatment in 25% O2 at 850 C. Sample weight: 30-35mg Gas flow rate: 20mL/min He (grade 5.0) and 100mL/min Ar (grade 5.0). From Figure 5.2, it can be observed that dopants have a noticeable impact on the oxygen uncoupling properties. Undoped CaMnO3 sees most of its oxygen released between C. Cobalt and vanadium doped perovskites observe a very sharp peak at 775 and 800 C respectively. The peaks correspond to approximately 0.08 and 0.12mg of oxygen released for the Co-doped and V-doped oxygen carriers. The sharp peaks observed can be explained by the oxidation state change of Co and V cations. The Al-doped oxygen carrier exhibits a small peak around 700 C and at higher temperatures follows a trend similar to pure CaMnO3. The Nidoped oxygen carrier is observed to follow the pure CaMnO3 behavior at a slightly lower oxygen uncoupling capacity. As the XRD profile (see Figure 5.1) shows, NiO is observed to 146

171 have formed, which does not exhibit CLOU properties under the conditions tested and is likely to contribute to the slightly lower oxygen uncoupling capacity compared to pure CaMnO3. The Fe-doped oxygen carrier is the only sample that observes a noticeable amount of oxygen at temperatures lower than 700 C. The Fe-doped oxygen carrier peak around 400 C is most likely α-oxygen which is attributed to chemisorbed oxygen in oxygen vacancies of the perovskite structure. [66 68] A previous study by Fino et. al showed that perovskites with high α-oxygen content correlated to high activity in catalytic combustion of soot. [67] As a result, it is considered advantageous of oxygen carriers for CLOU to have properties that include having suitable amounts of α-oxygen to achieve high activity with solid fuels. Coupling of the low temperature α-oxygen, incorporation of the metal into the perovskite, and the comparable high temperature CLOU properties, the Fe-doping effect on the CaMnO3 parent structure is further investigated for CLOU applications Effects of Iron Dopant Concentrations It is expected that the amount of dopant could have significant impact on the oxygen uncoupling properties. Figure 5.3 shows the XRD patterns for various dopant amounts of Fe into the CaMnO3 structure. Fe dopant concentrations at or below 20% observe only the CaMnO3 phase with no secondary phase present. However, a high dopant concentration (50%) induces a Ruddlesden-Popper structured Ca2(Mn,Fe)O4 phase. 147

172 Figure 5.3- XRD spectra of the CaMn1-xFexO3 spectra. Figure 5.4 examines the effect of Fe dopant amount into the parent CaMnO3 oxygen carrier for the CLOU capabilities of the oxygen carriers. Figure 5.4- DTG/TPD curves of CaMn1-xFexO3 oxygen carriers (x=0, 0.05, 0.2, and 0.5) to 1000 C at 10 C/min. in a pure argon atmosphere after treatment in 25% O2 at 850 C. Sample weight: 30-35mg Gas flow rate: 20mL/min He (grade 5.0) and 100mL/min Ar (grade 5.0). Doping of iron in the structure noticeably lowers the oxygen uncoupling properties compared to the undoped CaMnO3. Addition of dopants creates an increase in the amount of α-oxygen that is observed. At 5% dopant, a peak around 500 C is observed, while increasing the dopant 148

173 to 20% shows the peak shift to a lower temperature around 400 C. The amount of α-oxygen released by the 5% and 20% Fe doped oxygen carriers is 0.04 and 0.08 mg respectively. While the 50% Fe-doped CaMnO3 loses noticeable α-oxygen around 500 C, the higher temperature properties of the oxygen carrier is minimal. The undesirable CLOU properties of the 50% dopant oxygen carrier can be explained through the formation of the Ruddlesden-Popper phase which is known to be very stable. [59] The motivation behind B-site doping is to stabilize the perovskite oxygen carrier s stability. Undoped CaMnO3 is observed to decompose from the CaMnO3 structure into the Ruddlesden- Popper and spinel structures above 1050 C in an inert atmosphere (See Figure 5.5 (c) and (d)) Reduction to these phases have been observed at lower temperatures ( C) in highly reductive environments that are typically observed in chemical looping schemes. [33,35] These phases are undesirable due to their difficulty to regenerate back to the CaMnO3 structure in air under temperatures typical of chemical looping systems. Figure 5.5 (a) and (b) shows the insitu XRD spectra of the 5% Fe doped CaMnO3. 149

174 Figure 5.5- (a) and (c) In-situ XRD spectra of the CaMn0.95Mn0.05O3 and CaMnO3 perovskite oxygen carrier and (b) and (d) a better resolution of the (110) plane for both oxygen carriers. As can be seen, the perovskite consists of a single CaMn1-xFexO3 phase at room temperature. During the ramping, no phase transformation is seen as the 5 peaks observed at 2θ s of 24, 34, 48, 60, and 71 are present at all temperatures tested (up to 1200 C). The observation of a sharp peak shift around 650 C agrees with the oxygen release temperature in TPD experiments. In contrast, undoped CaMnO3 sample undergoes decomposition to CaMn2O4 and Ca2MnO4 phases at around 1050 C. Therefore, Fe doping is shown to be effective to enhance the phase stability of CaMnO3. 150

175 5.3.3 Redox Testing CLOU performance can be characterized by testing oxygen carriers under redox cycles. For CLOU, cycles of inert (mixture of helium and argon) and air are used to typically test for oxygen carrier s recyclability and stability. Our tests use both undoped CaMnO3 and the Fe doped sample under a small number of redox cycles at various temperatures to investigate the ability of these oxygen carriers to perform under numerous conditions. Figure 5.6 shows the oxygen uncoupling capacity under isothermal conditions between C Figure 5.6- Isothermal weight loss at temperatures between C. Sample weight: 30mg. Flow Rates: Reduction step: 30mL/min He (5.0 grade) and 60mL/min Ar (5.0 grade). Oxidation step: additional 10mL/min O2 (Extra Dry grade) added to aforementioned gas stream. Consistent with TPD results, CaMnO3 has no noticeable oxygen release at 650 C, while the incorporation of iron allows for noticeable amounts of oxygen to be released. Once the temperature reaches a temperature high enough to drive oxygen release in undoped CaMnO3 (>850 C), CaMnO3 releases approximately 20% more oxygen than the Fe doped samples. This result confirms that the Fe doped sample can accommodate less oxygen vacancy than undoped 151

176 CaMnO3 under high temperature, low PO2 environments. This is consistent with the oberseved phase stability of the doped sample under in-situ XRD. It also explains the smaller unit cell sizes for the doped samples (see Table 5.2). The spontaneous oxygen release at low temperature coupled with the improved phase stability make the Fe doped oxygen carriers promising for CLOU. The oxygen carriers were cycled 5 times at each temperature with no noticeable change in performance. Figure cycle tests at 850 C for CaMn0.95Fe0.05O3 oxygen carrier. Sample weight: 50mg. Gas flow rates: Reduction step: 30mL/min He (5.0 grade) and 60mL/min Ar (5.0 grade). Oxidation step: additional 10mL/min O2 (Extra Dry grade) added to aforementioned inert gas. The XRD pattern of the fresh and post 100 CLOU cycles XRD of the oxygen carrier is embedded. In chemical looping schemes, long term cycling is important due to the need to replenish the solid oxygen carriers over time. Figure 5.7 investigates a 100 redox cycles of Fe-doped CaMnO3 at 850 C. From the long term cycling, the Fe-doped oxygen carrier exhibits little to no deactivation. After the 100 cycles, the oxygen carrier was tested for redox reactions at lower temperatures, and comparable results of the oxygen uncoupling properties shown in Figure 5.6 were observed. Previous studies on B-site doped perovskites have shown the formation of 152

177 undesired CaMn2O4 and Ca2MnO4 phases after as few as 3 cycles. [31,33,35] After 100 cycles, the 5% Fe doped oxygen carrier was not observed to undergo any undesired phase transitions. Oxygen carriers for CLOU are highly effective in converting solid fuels due to their ability to release gaseous O2. Solid fuel conversion is then of upmost importance of these oxygen carriers. Fluidized bed experiments are conducted to test the undoped CaMnO3 and Fe-doped oxygen carriers ability to convert char prepared from Pittsburgh # 8 coal. The results are shown in Figure 5.8. Figure 5.8- Char conversion and CO2 selectivity of the CaMn0.95Fe0.05O3 oxygen carrier in a fluidized bed reactor. Sample weight: 15 gms. Oxygen carrier and 10mg of char Temperature: 850 C Gas Flow Rates: nitrogen: 800mL/min (Grade 5.0) and helium (Grade 5.0): 280mL/min (Char Reduction) and oxygen (Extra dry grade): 120mL/min (during oxidation only). The Fe-doped oxygen carrier was able to convert, on average, 90% char during each cycle with close to 99% selectivity to CO2 conversion at 850 C. During the first 3 cycles, coal char is injected while 10% O2 is present resulting in complete combustion of the char. The less than 100% conversion of coal char is likely to be due to the limitation of solids inventory in the tubular reactor, which results in relatively low oxygen generation rate (and concentration) 153

178 when compared to the flow rate of the inert fluidization gas. The Fe-doped oxygen carrier shows an attrition rate of 0.003w.t.% per hour. For comparison purpose, undoped CaMnO3 oxygen carrier was also tested for coal char conversion under identical conditions. The char conversion of the undoped CaMnO3 oxygen carrier was approximately 50%. This is a significant drop in performance compared to when Fe was doped into the B-site. A higher temperature experiment at 950 C was also tested for the undoped CaMnO3 oxygen carrier. At this temperature, a conversion of % is obtained within 5 redox cycles (see Figure 5.S3). This correlates with the superior low temperature CLOU activity for Fe doped CaMnO3. Therefore, the Fe-doped oxygen carrier is observed to have facile oxygen release for effective conversion of solid fuels. It is also capable of maintaining long term stability compared to undoped CaMnO Conclusion The present study investigates the effect of B-site doping on oxygen carriers with a general composition of CaMnxM1-xO3 (M = Fe, V, CO, Ni, Al). The phase compatibility with CaMnO3 parent structure, oxygen carrying capacity, oxygen release temperature, phase stability, and CLOU performance of these B-site doped oxygen carriers are investigated. Secondary phases are formed for all dopants except for Fe. For V and Ni doped oxygen carriers, pure oxides of the dopant metals are formed along with a doped CaMnO3 structure. Fe dopants are found to be fully compatible with the CaMnO3 parent structure for up to 20% B-site doping. These Fe doped oxygen carriers also exhibit excellent CLOU property, especially at lower temperatures (~600 C) when compared to the undoped CaMnO3 oxygen carrier. Fe dopant is also found to significantly improve the phase stability of CaMnO3 based oxygen carriers. No phase transformation was observed in both high temperature operation (up to 1200 C) and

179 oxygen uncoupling cycles (at 850 C). The Fe doped oxygen carrier also demonstrated significantly higher activity for coal char conversion when compared to the undoped CaMnO3 oxygen carrier at 850 C. Acknowledgements Funding supports from the U.S. Department of Energy (Award Number FE001247) and Kenan Institute are greatly appreciated. We acknowledge the use of the Analytical Instrumentation Facility (AIF) at North Carolina State University, which is supported by the State of North Carolina and the National Science Foundation. 155

180 REFERENCES [1] Figueroa JD, Fout T, Plasynski S, McIlvried H, Srivastava RD. Advances in CO2 capture technology The U.S. Department of Energy s Carbon Sequestration Program. Int J Greenh Gas Con. 2008;2:9 20. doi: /s (07) [2] Hansen J, Sato M. Greenhouse gas growth rates. Proc. Natl. Acad. Sci. USA 2004;101: [3] Adanez J, Abad A, Garcia-Labiano F, Gayan P, de Diego LF. Progress in chemicallooping combustion and reforming technologies. Prog Energ Combust 2012;38: doi: /j.pecs [4] Fan L-S, Li F. Chemical looping technology and its fossil energy conversion applications. Ind Eng Chem Res 2010;49: doi: /ie [5] Fan L-S. Chemical Looping Systems for Fossil Energy Conversions [6] Lyngfelt A. Chemical-looping combustion of solid fuels status of development. Appl Energ 2014;113: doi: /j.apenergy [7] Sorgenfrei M, Tsatsaronis G. Design and evaluation of an IGCC power plant using ironbased syngas chemical-looping (SCL) combustion. Appl Energ 2014;113: doi: /j.apenergy [8] Tong A, Bayham S, Kathe MV, Zeng L, Luo S, Fan L-S. Iron-based syngas chemical looping process and coal-direct chemical looping process development at Ohio State University. Appl Energ 2014;113: doi: /j.apenergy [9] Scott SA, Dennis JS, Hayhurst AN, Brown T. In situ gasification of a solid fuel and CO2 separation using chemical looping. AIChE J 2006;52: doi: /aic [10] Leion H, Mattisson T, Lyngfelt A. The use of petroleum coke as fuel in chemicallooping combustion. Fuel 2007;86: doi: /j.fuel [11] Luo S, Bayham S, Zeng L, McGiveron O, Chung E, Majumder A, et al. Conversion of metallurgical coke and coal using a Coal Direct Chemical Looping (CDCL) moving bed reactor. Appl Energ 2014;118: doi: /j.apenergy

181 [12] Cuadrat A, Linderholm C, Abad A, Lyngfelt A, Adánez J. Influence of limestone addition in a 10 kwth chemical-looping combustion unit operated with petcoke. Energ Fuel 2011;25: doi: /ef200806q. [13] Linderholm C, Lyngfelt A, Cuadrat A, Jerndal E. Chemical-looping combustion of solid fuels operation in a 10kW unit with two fuels, above-bed and in-bed fuel feed and two oxygen carriers, manganese ore and ilmenite. Fuel 2012;102: doi: /j.fuel [14] Leion H, Mattisson T, Lyngfelt A. Solid fuels in chemical-looping combustion. Int J Greenh Gas Con 2008;2: doi: /s (07)00117-x. [15] Arjmand M, Leion H, Mattisson T, Lyngfelt A. Investigation of different manganese ores as oxygen carriers in chemical-looping combustion (CLC) for solid fuels. Appl Energ 2014;113: doi: /j.apenergy [16] Shen L, Wu J, Xiao J. Experiments on chemical looping combustion of coal with a NiO based oxygen carrier. Combust Flame 2009;156: doi: /j.combustflame [17] Shen L, Wu J, Xiao J, Song Q, Xiao R. Chemical-Looping Combustion of biomass in a 10 kwth reactor with iron oxide as an oxygen carrier. Energ Fuel 2009;23: doi: /ef900033n. [18] Mattisson T, Lyngfelt A, Leion H. Chemical-looping with oxygen uncoupling for combustion of solid fuels. Int J Greenh Gas Con 2009;3:11 9. doi: /j.ijggc [19] Mattisson T. Materials for chemical-looping with oxygen uncoupling. ISRN Chem Eng 2013;2013:1 19. doi: /2013/ [20] Imtiaz Q, Hosseini D, Müller CR. Review of oxygen carriers for chemical looping with oxygen uncoupling (CLOU): thermodynamics, material development, and synthesis. Energy Technol 2013;1: doi: /ente [21] Azimi G, Leion H, Mattisson T, Lyngfelt A. Chemical-looping with oxygen uncoupling using combined Mn-Fe oxides, testing in batch fluidized bed. Energy Procedia 2011;4: doi: /j.egypro

182 [22] Azimi G, Leion H, Rydén 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: doi: /ef301120r. [23] Shafiefarhood A, Stewart A, Li F. Iron-containing mixed-oxide composites as oxygen carriers for chemical looping with oxygen uncoupling (CLOU). Fuel 2015;139:1 10. doi: /j.fuel [24] Shulman A, Cleverstam E, Mattisson T, Lyngfelt A. Manganese/iron, manganese/nickel, and manganese/silicon oxides used in chemical-looping with oxygen uncoupling (CLOU) for combustion of methane. Energ Fuel 2009;23: doi: /ef [25] Shulman A, Cleverstam E, Mattisson T, Lyngfelt A. Chemical looping with oxygen uncoupling using Mn/Mg-based oxygen carriers oxygen release and reactivity with methane. Fuel 2011;90: doi: /j.fuel [26] 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: doi: /ef201329x. [27] Arjmand M, Keller M, Leion H, Mattisson T, Lyngfelt A. Oxygen release and oxidation rates of MgAl2O4-supported CuO oxygen carrier for chemical-looping combustion with oxygen uncoupling (CLOU). Energ Fuel 2012;26: doi: /ef [28] Gayán 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: doi: /j.fuel [29] Pour NM, Azimi G, Leion H, Rydén M, Lyngfelt A. Production and examination of oxygen-carrier materials based on manganese ores and Ca(OH)2 in chemical looping with oxygen uncoupling. AIChE J 2014;60: doi: /aic [30] Rydén M, Leion H, Mattisson T, Lyngfelt A. Combined oxides as oxygen-carrier material for chemical-looping with oxygen uncoupling. Appl Energ 2014;113: doi: /j.apenergy

183 [31] Arjmand M, Hedayati A, Azad A-M, Leion H, Rydén M, Mattisson T. CaxLa1 xmn1 ymyo3 δ (M = Mg, Ti, Fe, or Cu) as oxygen carriers for chemical-looping with oxygen uncoupling (CLOU). Energ Fuel doi: /ef [32] Arjmand M, Kooiman RF, Rydén M, Leion H, Mattisson T, Lyngfelt A. Sulfur tolerance of CaxMn1 ymyo3 δ (M = Mg, Ti) perovskite-type oxygen carriers in chemical-looping with oxygen uncoupling (CLOU). Energ Fuel 2014;28: doi: /ef402383v. [33] Hallberg P, Jing D, Rydén M, Mattisson T, Lyngfelt A. Chemical looping combustion and chemical looping with oxygen uncoupling experiments in a batch reactor using spray-dried CaMn1 xmxo3 δ (M = Ti, Fe, Mg) particles as oxygen carriers. Energ Fuel 2013;27: doi: /ef [34] 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: doi: /ef900444d. [35] Rydén M, Lyngfelt A, Mattisson T. CaMn0.875Ti0.125O3 as oxygen carrier for chemicallooping combustion with oxygen uncoupling (CLOU) Experiments in a continuously operating fluidized-bed reactor system. Int J Greenh Gas Con 2011;5: doi: /j.ijggc [36] Sundqvist S, Leion H, Rydén M, Lyngfelt A, Mattisson T. CaMn0.875Ti0.125O3 δ as an oxygen carrier for chemical-looping with oxygen uncoupling (CLOU)-solid-fuel testing and sulfur interaction. Energy Technol 2013;1: doi: /ente [37] Källén M, Rydén M, Dueso C, Mattisson T, Lyngfelt A. CaMn0.9Mg0.1O3-δ as oxygen carrier in a gas-fired 10 kwth chemical-looping combustion unit. Ind Eng Chem Res 2013;52: doi: /ie303070h. [38] Pishahang M, Larring Y, McCann M, Bredesen R. Ca0.9Mn0.5Ti0.5O3 δ : A suitable oxygen carrier material for fixed-bed chemical looping combustion under syngas conditions. Ind Eng Chem Res 2014;53: doi: /ie500928m. [39] Moldenhauer P, Rydén M, Mattisson T, Hoteit A, Jamal A, Lyngfelt A. Chemicallooping combustion with fuel oil in a 10 kw pilot plant. Energ Fuel 2014;28: doi: /ef

184 [40] Dai XP, Li J, Fan JT, Wei WS, Xu J. Synthesis gas generation by chemical-looping reforming in a circulating fluidized bed reactor using perovskite LaFeO3-based oxygen carriers. Ind Eng Chem Res 2012;51: doi: /ie300033e. [41] He F, Li X, Zhao K, Huang Z, Wei G, Li H. The use of La1 xsrxfeo3 perovskite-type oxides as oxygen carriers in chemical-looping reforming of methane. Fuel 2013;108: doi: /j.fuel [42] Mihai O, Chen D, Holmen A. Chemical looping methane partial oxidation: the effect of the crystal size and O content of LaFeO3. J Catal 2012;293: doi: /j.jcat [43] Murugan A, Thursfield A, Metcalfe IS. A chemical looping process for hydrogen production using iron-containing perovskites. Energ Environ Sci 2011;4:4639. doi: /c1ee02142g. [44] Readman JE, Olafsen A, Larring Y, Blom R. La0.8Sr0.2Co0.2Fe0.8O3 δ as a potential oxygen carrier in a chemical looping type reactor, an in-situ powder X-ray diffraction study. J Mater Chem 2005;15:1931. doi: /b416526h. [45] Rydén M, Lyngfelt A, Mattisson T, Chen D, Holmen A, Bjørgum E. Novel oxygencarrier materials for chemical-looping combustion and chemical-looping reforming; LaxSr1 xfeyco1 yo3 δ perovskites and mixed-metal oxides of NiO, Fe2O3 and Mn3O4. Int J Greenh Gas Con 2008;2: doi: /s (07) [46] Sarshar Z, Sun Z, Zhao D, Kaliaguine S. Development of sinter-resistant core shell LaMnxFe1 xo3@msio2 oxygen carriers for chemical looping combustion. Energ Fuel 2012;26: doi: /ef [47] Zhao K, He F, Huang Z, Zheng A, Li H, Zhao Z. Three-dimensionally ordered macroporous LaFeO3 perovskites for chemical-looping steam reforming of methane. Int J Hydrog Energ 2014;39: doi: /j.ijhydene [48] Sarshar Z, Kleitz F, Kaliaguine S. Novel oxygen carriers for chemical looping combustion: La1 xcexbo3 (B = Co, Mn) perovskites synthesized by reactive grinding and nanocasting. Energ Environ Sci 2011;4: doi: /c1ee01716k. [49] Noorman S, Gallucci F, van Sint Annaland M, Kuipers JAM. Experimental Investigation of chemical-looping combustion in packed beds: a parametric study. Ind Eng Chem Res 2011;50: doi: /ie

185 [50] Chen Y, Galinsky N, Wang Z, Li F. Investigation of perovskite supported composite oxides for chemical looping conversion of syngas. Fuel 2014;134: doi: /j.fuel [51] Galinsky NL, Huang Y, Shafiefarhood A, Li F. Iron oxide with facilitated O 2 transport for facile fuel oxidation and CO2 Capture in a Chemical Looping Scheme. ACS Sustain Chem Eng 2013;1: doi: /sc300177j. [52] Shafiefarhood A, Galinsky N, Huang Y, Chen Y, Li F. Fe2O3@LaxSr1-xFeO3 core-shell redox catalyst for methane partial oxidation. ChemCatChem 2014;6: doi: /cctc [53] Neal LM, Shafiefarhood A, Li F. Dynamic methane partial oxidation using a Fe2O3@La0.8Sr0.2FeO3-δ core shell redox catalyst in the absence of gaseous oxygen. ACS Catal 2014;4: doi: /cs [54] Galinsky NL, Shafiefarhood A, Chen Y, Neal L, Li F. Effect of support on redox stability of iron oxide for chemical looping conversion of methane. Appl Catal B- Environ 2015;164: doi: /j.apcatb [55] Peña MA, Fierro JLG. Chemical structures and performance of perovskite oxides. Chem Rev 2001;101: doi: /cr980129f. [56] He F, Trainham J, Parsons G, Newman JS, Li F. A hybrid solar-redox scheme for liquid fuel and hydrogen coproduction. Energ Environ Sci 2014;7:2033. doi: /c4ee00038b. [57] He F, Li F. Perovskite promoted iron oxide for hybrid water-splitting and syngas generation with exceptional conversion. Energ Env Sci 2015;8: doi: /c4ee03431g. [58] Jing D, Mattisson T, Leion H, Rydén M, Lyngfelt A. Examination of perovskite structure CaMnO3-δ with MgO addition as oxygen carrier for chemical looping with oxygen uncoupling using methane and syngas. Int J Chem Eng 2013;2013:1 16. doi: /2013/ [59] Bakken E, Norby T, Stolen S. Nonstoichiometry and reductive decomposition of CaMnO3-δ. Solid State Ionics 2005;176: doi: /j.ssi

186 [60] Cabello A, Abad A, Gayán P, de Diego LF, García-Labiano F, Adánez J. Effect of operating conditions and H2S presence on the performance of CaMg0.1Mn0.9O3 δ perovskite material in chemical looping combustion (CLC). Energ Fuel 2014;28: doi: /ef [61] Fossdal A, Bakken E, Øye BA, Schøning C, Kaus I, Mokkelbost T, et al. Study of inexpensive oxygen carriers for chemical looping combustion. Int J Greenh Gas Con 2011;5: doi: /j.ijggc [62] Hallberg P, Källén M, Jing D, Snijkers F, van Noyen J, Rydén M, et al. Experimental investigation of based oxygen carriers used in continuous chemical-looping combustion. Int J Chem Eng 2014;2014:1 9. doi: /2014/ [63] Galinsky N, Mishra A, Zhang J, Li F. Ca1 xaxmno3 (A=Sr and Ba) perovskite based oxygen carriers for chemical looping with oxygen uncoupling (CLOU). Appl Energ doi: /j.apenergy [64] Hallberg P, Rydén M, Mattisson T, Lyngfelt A. CaMnO3-δ made from low cost material examined as oxygen carrier in chemical-looping combustion. Energy Procedia 2014;63:80 6. doi: /j.egypro [65] Schmitz M, Linderholm C, Lyngfelt A. Chemical looping combustion of sulphurous solid fuels using spray-dried calcium manganate particles as oxygen carrier. Energy Procedia 2014;63: doi: /j.egypro [66] Ciambelli P, Cimino S, De Rossi S, Faticanti M, Lisi L, Minelli G, et al. AMnO3 (A=La, Nd, Sm) and Sm1 xsrxmno3 perovskites as combustion catalysts: structural, redox and catalytic properties. Appl Catal B-Environ 2000;24: doi: /s (99) [67] Fino D, Russo N, Saracco G, Specchia V. The role of suprafacial oxygen in some perovskites for the catalytic combustion of soot. J Catal 2003;217: doi: /s (03)00143-x. [68] Zhu J, Li H, Zhong L, Xiao P, Xu X, Yang X, et al. Perovskite oxides: preparation, characterizations, and applications in heterogeneous catalysis. ACS Catal 2014;4: doi: /cs500606g. 162

187 CHAPTER 6 Conclusions and Future Work Chemical looping combustion (CLC) uses a novel, energy efficient redox scheme for in-situ carbon dioxide capture. The performance of CLC relies on the oxygen carrier used in the redox scheme. In CLC, the oxygen carrier undergoes a redox reaction where it provides its lattice oxygen to carbonaceous fuels before being regenerated through combustion in air. Research in oxygen carriers have resulted in more than 1000 different oxygen carriers tested. Among the various transition metal oxides considered as oxygen carrying materials, iron oxide is more promising for CLC due to its low cost, good mechanical properties, and minimal health and environmental impacts. However, iron oxides suffer from low reactivity and agglomeration from reduction to metallic Fe. Support additions have offered improved activity and recyclability, but the activity is still not as high as other metal oxides such a CuO and NiO. It has been observed that oxygen anion transport in the lattice is related to iron oxide s activity. However, pure iron oxide and inert supported iron oxides have low oxygen anion conduction. To help improve the activity of iron oxide oxygen carriers, improvement of the oxygen anion and electron conduction within the lattice is attempted. The developed support La1-xSrxFeO3 (LSF) is commonly used in oxygen separation membranes and has been utilized as an oxygen carrier. The LSF supported iron oxide is compared to traditionally used inert supports, TiO2 and Al2O3, and pure ionic conductor support, yttria-stabilized zirconia (YSZ). The LSF supported iron oxide achieves activity up to 70 times higher than that of the other supported iron oxides in methane. In addition to the higher activity, the LSF supported iron oxide exhibited excellent recyclability over 50 cycles in hydrogen. Coke formation with iron oxide oxygen carriers usually results from either the reverse Boudouard reaction or hydrocarbon decomposition reaction. The coke formation is minimal on the mixed conductor supported 163

188 oxygen carrier for methane conversion, and no coke formation is observed until the theoretical conversion of Fe2O3 to Fe occurs in the reactor. This leads to the fact that oxygen stored in iron oxides is easily transported through the oxygen carrier. The LSF-Fe2O3 oxygen carrier can be considered as an ensemble of nanoscale mixed conductive membrane reactors where the Fe2O3 acts as an oxygen source or sink in the redox reactions of the chemical looping process. To further probe the effects of support and iron oxide interactions in chemical looping redox reactions, supports with varying physical and chemical properties are chosen for iron oxide based oxygen carriers. These supports include mixed conductor perovskite (Ca1-xSrxTi1- yniyo3) and fluorite (CeO2) and inert spinel (MgAl2O4). These oxygen carriers are investigated for performance in redox reactions containing hydrogen and methane as fuels. The perovskite oxygen carrier exhibits the highest activity of the three oxygen carriers. The same oxygen carrier deactivates slightly over 5 cycles before the activity stabilizes. The deactivation is attributed to Fe 3+ substituting into the perovskite phase in favor of the Ni 2+. Once the phases are observed to stabilize, the oxygen carrier s activity is also stabilized. CeO2 supported iron oxide is observed to deactivate by up to 75% in methane over 10 cycles. During the oxidation step, iron oxide is migrated and segregated on the surface which leads to agglomeration. The inability of the Fe to stabilize within the CeO2 support leads to large segregated amounts of iron oxide on the surface leading to decreased accessibility of O 2- to gaseous fuels. The MgAl2O4 supported iron oxide offered contrasting behavior in hydrogen and methane. In hydrogen, the oxygen carrier is observed to deactivate over 10 cycles. However, methane causes the oxygen carrier to activate, i.e. redox kinetics improved from cycle to cycle. The resulting activation is attributed to the increased in porosity and surface area caused through coke formation. The carbon filaments degrade the surface through breakage of the iron oxide 164

189 located at the surface of the oxygen carrier. Despite the activity increase of the oxygen carrier, the disintegration is considered detrimental in fluidized bed reactors. These indicate important key design parameters in choosing a support-primary oxide in which stable phase compatible materials with improved mixed conduction can lead to highly active, recyclable oxygen carriers. Solid fuels offer additional challenges for chemical looping combustion due to slow reaction kinetics between solid oxygen carrier and fuel particles. To improve the ability for solid conversion in chemical looping combustion, a process known as chemical looping with oxygen uncoupling (CLOU) has been developed. CLOU utilizes oxygen carriers with high PO2 that are able to release or uncouple their lattice oxygen as gaseous oxygen. This gaseous oxygen can react with the solid fuels at a much higher rate than the lattice oxygen. CaMnO3 offers high potential as an oxygen carrier for CLOU as it is abundant, cheap, and less toxic material than CLOU oxygen carriers based on Cu or Co. CaMnO3 suffers from irreversible phase transitions and sulfur poisoning, so development into stabilizing the structure and improving the oxygen uncoupling properties of CaMnO3 can lead to highly active and recyclable oxygen carriers for solid fuel conversion. To address the concerns of pure CaMnO3, addition of A- and B-site dopants are chosen to investigate their compatibility with the parent structure. Additionally, their effect on the oxygen uncoupling properties of the oxygen carriers are investigated. A-site dopants Sr and Ba are chosen to investigate the effect of A-site dopants on the base CaMnO3 oxygen carrier. Ba has poor phase compatibility with the CaMnO3. Even at low concentrations of Ba, separate phases of BaMnO3 and CaMnO3 are observed. However, addition of Sr shows full incorporation of Sr into the CaMnO3 structure until concentrations reach 50% Ca and 50% Sr. 165

190 In-situ X-ray diffraction (XRD) indicates that the Sr-doped oxygen carrier is stable in inert atmospheres up to 1200 C whereas pure CaMnO3 decomposes at 1000 C into stable Ruddlesden-Popper and spinel structures. Besides the improved stability of the oxygen carrier, the Sr-doped oxygen carrier exhibits oxygen release at temperatures of 200 C lower than CaMnO3. This improved oxygen uncoupling property coupled with enhanced stability leads to improved recyclability. The Sr-doped oxygen carrier is tested in a fluidized bed reactor under batch redox cycles with solid coal char particles as the fuel. It is observed to be stable over 5 cycles leading to minimal attrition and averages >94% conversion of the solid char particles injected. B-site dopants include Fe, Ni, Co, V, and Al. The incorporation of most B-site dopants results in small quantities of secondary phases. However, Fe exhibited excellent incorporation into the parent structure. Similar to the Sr-doped oxygen carrier, Fe-doped CaMnO3 also saw stability up to 1200 C and observed oxygen uncoupling properties at low temperatures. The Fe-doped oxygen carrier exhibits slightly higher oxygen uncoupling properties than Sr-doped at low temperature, however at high temperatures, it has a lower amount of oxygen uncoupled. The Fe-doped sample is highly recyclable, and no other phases are observed after 100 cycles. The Fe-doped oxygen carriers observes excellent conversion of Pittsburgh #8 coal char at 850 C whereas pure CaMnO3 does not observe much activity. The doped CaMnO3 oxygen carriers offer potentially attractive alternatives for oxygen carriers utilized for chemical looping with oxygen uncoupling of solid fuels. 166

191 FUTURE WORK- Quantification of Oxygen Transport Redox reactions involving oxygen carriers are highly complex. These reactions involve gaseous dissociation on the surface, surface exchange, and bulk conduction. Bulk diffusion can incorporate transport through a single phase or multiple phases meaning that transport across grain boundaries can occur. Moreover, the reaction can be controlled through one or multiple of these steps under different operating conditions inside the reactor. As a result, understanding and quantifying oxygen transport during the redox process can lead to optimization and design specifications for oxygen carriers in CLC. The importance of oxygen anion transport in iron oxide oxygen carrier s design has been discussed. It has been shown that supports exhibiting high mixed-conduction offer highly active oxygen carriers for a variety of fuels. To date, there has been minimal research involving correlation of oxygen transport in the oxygen carrier to redox activity. Imtiaz et al. investigated the effect of Na + on Al2O3-CuO in reaction for its activity with methane [1]. It was determined that Na + doping had a higher activity compared to undoped oxygen carriers. The increase in activity was contributed to both the prevention of Cu-Al spinel oxides and the effect of the alkali on activation energy on charge transfer. The doped samples had approximately 0.05eV lower activation barrier on charge transfer than undoped samples. This enhanced charge transfers also correlated to a total conduction increase of 2 times that of the undoped oxygen carriers. Surface exchange and bulk oxygen conduction measurements of MIEC materials are often conducted through techniques such as secondary ion mass spectroscopy (SIMS), conductivity relaxation, or impedance spectroscopy techniques [2]. For SIMS, the common technique of isotope exchange depth profiling (IEDP) is used, in which a dense pellet of material is annealed 167

192 in an enriched 18 O stream for a fixed time t. To determine the oxygen diffusion and surface exchange coefficient of the material, it is required to fit theoretical diffusion equations to the experimental data obtained from SIMS depth profiling [3,4]. Conductivity relaxation techniques can be used to measure both ionic and electronic conduction by choosing electrodes that only allow either electronic conduction or ionic conduction. By holding current fixed between two electrodes, voltage measurements can be conducted. As current increases, voltage increases in a linear regression allowing for a slope to be calculated which is directly related to the electrical or ionic conductivity of material [5 7]. Impedance spectroscopy utilizes fitting equivalent circuits to experimental data which can be correlated through different rate limiting steps such as electronic and ionic conduction steps. Impedance spectroscopy is commonly used in fuel cell materials [8,9]. Due to the complexity of redox reactions involving oxygen carriers, probing of oxygen transport through the oxidation reaction in single phase oxygen carriers will allow for deep understanding of the overall redox process. Discussing of individual transport process such as surface exchange and bulk conduction can lead to important mechanistic understanding of the oxygen transport. Recently, a novel, alternative approach to surface exchange measurements have been developed by Bouwmeester et al.[10] This procedure utilizes oxygen pulse isotopic experiments (PIE) to monitor the exchange of oxygen from the gaseous phase to adsorbed species on the surface of metal oxide particles and eventual incorporation into the lattice. The procedure uses a known amount of sample in a U-tube reactor with a fixed PO2 16 O2 and temperature. A pulse of known volume 18 O2 is injected into the stream. Oxygen fractions of 16 O2, 16 O 18 O, and 18 O2 are measured at the exit of the reactor using a mass spectrometer. The advantage of this technique over isotopic exchange depth profiling using SIMS is that key 168

193 mechanistic information can also be ascertained from experimental data. Under the assumption that the reactor behaves using ideal plug-flow behavior, the mass balance yields a simplified equation for relationship for surface exchange rate (R0) shown in Eq. 1 [10]: R 0 = 18 n τ r S ln (f g,i 18 ) Eq. 1 f g,e where f g,i and f g,e are the isotopic fractions of 18 O at the inlet and exit of the reactor, respectively. Additionally, n is the total amount of oxygen atoms in the gas phase while S is the total surface area available of the particles in the reactor. Average residence time of the reactor is described in equation 1 as τr. The isotopic fractions of 18 O can be calculated from equations 2 and 3: f g 18 = f g f g 34 Eq. 2 f g 34 = 1 f g 32 f g 36 Eq. 3 Where f g 32, f g 34, and f g 36 are isotopic fractions of 16 O2, 16 O 18 O, and 18 O2. Preliminary probing of CaMnO3 oxygen carriers are examined using the technique described by Bouwmeester et al. Effectiveness of incorporation of the dopants are discussed further in chapters 4 and 5. Oxygen carriers are investigated at a PO2 of 0.10atm and over a temperature range of C. Experimental data of the oxygen isotope fraction (f 32, f 34, and f 36 ) are shown in Figure 6.1 for the oxygen carriers. The undoped oxygen carrier sees more than 50% 18 O2 uptake around 575 C. The Fe-doped oxygen carrier doesn t observe this uptake until ~650 C while the Sr-doped oxygen carriers observe this at 525 C. Moreover, this comparison of isotope fractions acts as a fingerprint of O2 dissociation. 169

194 Figure 6.1- (a) CaMnO3, (b) CaMn0.75Fe0.25O3, and (c) Ca0.75Sr0.25MnO3 fractions of 16 O2, 16 O 18 O, 18 O2 from pulse isotope exchange experiments at PO2= 0.1 atm. Sample: 50 mg. Gas Flow: 50 ml/min From Figure 6.1, surface exchange rates can be calculated using Eq. 1. However, R0 is a lumped parameter consisting of several possible sequential exchange steps including adsorption, dissociation, and incorporation of oxygen atoms into the oxide lattice. The transient distribution of 18 O in the gas phase, as shown in Figure 6.1, offer insight into important kinetic information including rate limiting steps and possible mechanisms of oxygen surface exchange. Bouwmeester et al. discuss different possible mechanisms for oxygen incorporation in their development of the PIE technique [10]. By assuming that the adatoms intermediates of the oxygen species are minimal (steady state), using a simple 2-step mechanism shown below, one is able to show that R0 can be treated as a resistance comprised of the summation of 2 170

195 serial resistances Ra and Ri for the adsorption and incorporation, respectively. This is shown in Eq. 4. Step 1: O Ra 2 2O ad Step 2: O ad Ri O lattice 1 R 0 = 1 R a + 1 R i Eq. 4 Plots of R0, Ra, and Ri vs 1/T are shown in Figure 6.2. All oxygen carriers are rate limited through the dissociation of oxygen from the gaseous phase to the adsorbed species on the surface on the gas phase at low temperatures. As temperature increases, a more competing effect of adsorption and incorporation is observed. Additionally, the Fe-doped oxygen carrier sees the biggest difference in adsorption and incorporation rates at low T (<700 C). 171

196 Figure 6.2- R0, Ra, and Ri for (a) CaMnO3, (b) CaMn0.75Fe0.25O3, and (c) Ca0.75Sr0.25MnO3 oxygen carriers at a PO2=0.1 atm (d) Compares the R0 for the three oxygen carriers at 500 C and 725 C. Additionally, apparent activation energy for the overall surface exchange, dissociative adsorption, and incorporation steps can all be calculated. Generally, the dopants have all lowered activation energies except for the dissociative adsorption of Fe-doped oxygen carriers. Table 6.1- Apparent Arrhenius activation energies for surface exchange for the synthesized oxygen carriers. Oxygen Carrier Activation Energy (kj/mol) R0 Ra Ri CaMnO CaMn0.75Fe0.25O Ca0.75Sr0.25MnO

197 The initial results indicate that a correlation between the exchange and uncoupling properties is observed. The undoped CaMnO3 offers the lowest surface exchange rates and highest activation energies which correlate with the lack of uncoupled oxygen observed with the oxygen carrier at temperatures lower than 800 C. The Sr-doped sample has the lowest activation energy with Fe-doped samples between the undoped and Sr-doped. This correlates well with the oxygen uncoupling results presented in chapters 4 and 5 where Sr-doped uncouples more oxygen and at faster rates than the Fe-doped oxygen carrier at higher temperatures. In addition to the surface exchange experiments presented here, the use of IEDP can be coupled with TOF-SIMS and conductivity relaxation methods to probe the bulk oxygen conduction in these materials. Further investigation of the relationship between oxygen transport and redox activity of oxygen carriers can lead to fundamental design parameters in future development of oxygen carriers. Besides the direct measurement of correlation between oxygen transport and redox activity, kinetic modeling can give insight into possible reaction mechanisms and rate limiting steps. Common kinetic models operate under the assumption that the reaction is either surface or diffusion limited. However, several models including ionic diffusion can be incorporated into future studies to offer potential mechanistic insights into reaction mechanisms in oxygen carrier redox systems. 173

198 REFERENCES [1] Imtiaz Q, Abdala PM, Kierzkowska AM, van Beek W, Schweiger S, Rupp JLM, et al. Na + doping induced changes in the reduction and charge transport characteristics of Al2O3-stabilized, CuO-based materials for CO2 capture. Phys Chem Chem Phys 2016;18: doi: /c6cp00257a. [2] Sunarso J, Baumann S, Serra JM, Meulenberg WA, Liu S, Lin YS, et al. Mixed ionic electronic conducting (MIEC) ceramic-based membranes for oxygen separation. J Membrane Sci 2008;320: doi: /j.memsci [3] Boehm E, Bassat J-M, Steil MC, Dordor P, Mauvy F, Grenier J-C. Oxygen transport properties of La2Ni1-xCuxO4+δ mixed conducting oxides. Solid State Sci 2003;5: doi: /s (03) [4] De Souza R., Kilner J., Walker J. A SIMS study of oxygen tracer diffusion and surface exchange in La0.8Sr0.2MnO3+δ. Mater Lett 2000;43: doi: /s x(99) [5] Chen D, Shao Z. Surface exchange and bulk diffusion properties of Ba0.5Sr0.5Co0.8Fe0.2O3 δ mixed conductor. Int J Hydrogen Energ 2011;36: doi: /j.ijhydene [6] Lane J. Measuring oxygen diffusion and oxygen surface exchange by conductivity relaxation. Solid State Ionics 2000; : doi: /s (00) [7] Niedrig C, Wagner SF, Menesklou W, Baumann S, Ivers-Tiffée E. Oxygen equilibration kinetics of mixed-conducting perovskites BSCF, LSCF, and PSCF at 900 C determined by electrical conductivity relaxation. Solid State Ionics 2015;283:30 7. doi: /j.ssi [8] de Larramendi IR, Ortiz N, López-Antón R, de Larramendi JIR, Rojo T. Structure and impedance spectroscopy of La0.6Ca0.4Fe0.8Ni0.2O3 δ thin films grown by pulsed laser deposition. J Power Sources 2007;171: doi: /j.jpowsour [9] Wagner SF, Menesklou W, Ivers-Tiffée E. Kinetics of oxygen exchange in strontium titanate. Ionics 2003;9: doi: /bf

199 [10] Bouwmeester HJM, Song C, Zhu J, Yi J, van Sint Annaland M, Boukamp BA. A novel pulse isotopic exchange technique for rapid determination of the oxygen surface exchange rate of oxide ion conductors. Phys Chem Chem Phys 2009;11:9640. doi: /b912712g. 175

200 APPENDICES 176

201 Appendix A Supplementary Information for Chapter 2 This supporting information contains the following sections: S1 Experimental S2 Results 177

202 S1 Experimental S1.1 Oxygen Carrier and Mixed-Conductive Support Synthesis As described in the manuscript, a one pot solid state reaction (SSR) method is used for synthesizing most of the oxygen carriers. The general procedure involves preparation of powder mixture followed with pelletization and annealing/solid state reaction. For instance, the Fe2O3/La0.8Sr0.2FeO3-δ particles are prepared by mixing calculated amounts of La2O3 (99.9%, Aldrich), SrCO3 (99.9%, Noah Chemical) and Fe2O3 (99.9%, Noah) precursors for 3h in a planetary ball mill (XBM4X, Columbia International) at a rotation speed of 250 rpm. This is followed with annealing at 1200 C for 8 hours. A sequential synthesis method which involves formation of mixed-conductive support followed with primary oxide (Fe2O3) addition is also used to synthesize LSF and La0.8Sr0.2Co0.5Fe0.5O3-δ (LSCF) supported oxygen carriers. Single phase LSF and LSCF is synthesized via the one pot SSR approach described above. Co3O4 is used as the precursor to provide Co for the B site of the LSCF perovskite. The phase purity of the support is confirmed by XRD. Once the mixed-conductive support is formed, it is ground into powders and then mixed with additional iron oxide in the ball mill. The mixed powder is annealed again at 1200 C. In order to evaluate the reactivity of porous oxygen carrier particles, a starch template technique is used. Pores are created in the TiO2 and LSF supported oxygen carrier by mixing the sintered oxygen carrier powders with starch at a 9:1 weight ratio in the plenary ball mill for 5 hours at a rotation speed of 250 rpm. The starch is subsequently burnt-off and leaves pores behind. 178

203 S1.2 Reactivity Studies All the reactivity studies are conducted in a SETARAM SETSYS Evolution Thermal Gravimetric Analyzer (TGA). Unless specified otherwise, the experiments are conducted at 900 C with a total gas flow rate of 300 ml/min and a sample weight of approximately 20 mg. 10.5% (by volume) H2 balanced with He/N2 is used to test the redox activity of pure LSF and one pot and sequentially produced oxygen carriers. In order to test the reactivity of the porous oxygen carriers with methane, 10% CH4 is used as the reducing gas. 10% oxygen balanced with inert gas is used as the regeneration/oxidation gas in all cases. S1.3 Morphology Characterizations The morphology and elemental compositions of the oxygen carriers are analyzed using scanning electron microscopy (SEM, Hitachi S3200). An accelerating voltage of 5 20 kv is used for image capture. The same instrument is used for elemental analysis of the samples using energy dispersive X-ray spectroscopy (EDS) technique. Both the surface and the crosssection of the oxygen carriers are examined using both EDS line scan (20 kv) and spot analysis (10 kv). A lower acceleration voltage is used to obtain sub-micro special resolution of the elemental compositions. S2 Results S2.1 Reduction of LSF Support Surface and lattice oxygen in perovskite can be partially removed under a reducing environment. S1 Our current definition of oxygen carrier conversion does not account for the removable lattice oxygen in the LSF based perovskite support. In order to quantify the redox contribution from LSF support, pure peroskite is reduced in 10% H2. As shown in Figure 2.S1, 179

204 pure LSF donated less than 0.8% (by weight) oxygen in a 5-mins period. Since the LSF supported oxygen carrier is composed of 40% (by weight) support, the oxygen donation from support will be roughly 0.32% by weight within 5 minutes. Based on our definition of oxygen carrier conversion in Equation 1 (see manuscript), support is likely to account for 1.8% conversion out of the 97.5% total conversion achieved within the first 5 minutes of the reduction. This confirms that the contribution of the LSF support to the overall oxygen carrier conversion is negligible. Figure 2.S1. Weight loss of LSF as a function of time in the presence of 10% H2. Total gas flow rate: 300 ml/min, reduction temperature: 900 C. S2.2 Effect of Oxygen Carrier Compositions and Synthesis Techniques In order to determine the effect of oxygen carrier composition and synthesis methods, a numbers of oxygen carriers are synthesized and tested. The results are presented in Figure 2.S2. As can be seen, all the mixed-conductor supported oxygen carriers are much more reactive 180

205 than TiO2 based oxygen carriers. In addition, the simpler, one-pot technique produces more reactive oxygen carrier than the sequential technique. This may result from the more homogeneous mixing of LSF and Fe2O3 nano-crystallites in the oxygen carriers produced from the one-pot method. Among the oxygen carriers synthesized using sequential method, LSCF supported oxygen carrier is much more reactive than LSF supported oxygen carrier. This is likely to result from the higher O 2- conductivity of LSCF compared with LSF. S2-S4 It is noted that the one-pot method, although simple and effective, may not be suitable for some oxygen carrier compositions. For instance, one-pot mixing of La/Sr/Co/Fe precursors may lead to many possible product combinations containing LSF, LSCF, and/or oxides of Co/Fe. Further studies on phase compatibilities of various support and primary oxide combinations are being conducted. The effect of LSF support in enhancing the shuttling of O 2- and electrons is substantiated by the experimental finding shown in Figure 2.S3: increased sintering time leads to slightly increased reactivity with methane. This can be attributed to more complete SSR that allows the formation of more homogeneous LSF phase. Figure 2.S2. Reactivity comparisons of oxygen carriers obtained with different compositions and synthesis techniques. Total gas flow rate: 290 ml/min, gas compositions: 10.5% H2, reduction temperature: 900 C. 181

206 Figure 2.S3. Reactivity comparison in methane of LSF supported oxygen carrier between 8 hours of sintering and 28 hours of sintering. Total gas flow rate: 300 ml/min, gas compositions: 10% CH4, reduction temperature: 900 C. S2.3 Effect of Oxygen Carrier Porosity Intra-particle diffusion can be important for the overall reactivity of the oxygen carrier. In order to isolate the effect of mixed conductivity to oxygen carrier reactivity, we intentionally reduced the porosity of the oxygen carriers in most of our studies via high temperature annealing. For practical applications, it may be desirable to maximize the oxygen carrier porosity so that the reaction rate can be enhanced. In order to compare the effect of mixed-conductive support in porous oxygen carriers, oxygen carriers synthesized using the starch-template technique are tested using methane as the reducing gas. As can be seen from Figure 2.S4, the porous LSF supported oxygen carrier is still far more reactive than the reference, TiO2 supported oxygen carrier. Figure 2.S5 indicates slight reactivity improvement of the TiO2 supported, porous oxygen carrier overall the first few redox cycles. The reactivity of the LSF supported oxygen carrier is still far more superior from a reactivity standpoint. 182

207 Figure 2.S4. Reactivity comparisons of porous oxygen carriers obtained using starch-template method. Total gas flow rate: 300 ml/min, gas compositions: 10% CH4, reduction temperature: 900 C. Figure 2.S5. Reactivity comparisons of porous oxygen carriers within the first 5 redox cycles. Total gas flow rate: 300 ml/min, gas compositions: 10% CH4, reduction temperature: 900 C. 183

208 S2.4 Morphology and Elemental Compositions of the Oxygen Carrier Surface area analysis done using BET is shown in Table 2.S1. Comparable surface areas are obtained for the oxygen carriers after sintering. The amounts of YSZ and Al2O3 samples have led to slightly larger uncertainty of the measurement. Table 2.S1. Surface areas of the oxygen carriers before and after sintering Material Surface Area of asprepared samples (m 2 /g) Surface Area of samples after sintering (m 2 /g) Fe2O3 supported on La0.8Sr0.2FeO Fe2O3 supported on TiO Fe2O3 supported on Al2O3 n/a 0.54 Fe2O3 supported on YSZ n/a Figure 2.S5 shows the morphology of the LSF supported oxygen carrier at various redox cycles and oxidation states. As can be seen, the morphological change over 50 redox cycles is insignificant. EDS line scan illustrated in Figure 2.S6 exhibits continuous distribution of La, Sr, and Fe. Based on their relative content, it is apparent that the perovskite phase and iron oxide phase cannot be distinguished based on the spatial resolution of the EDS. Point EDS analysis of the sample further validate the co-existence of two phases at the current resolution. Since the spatial resolution of the EDS is below 1 micron at the current acceleration voltage, we can conclude that the oxygen carrier grain is composed of interpenetrated nano-crystallites of Fe2O3 and LSF. It is noted that the La and Sr ratios at different locations vary in a wide range. This may result from inhomogeneous mixing during solid-state synthesis. Since the LSF structure allows essentially any La and Sr ratio on the A site, local enrichment of La or Sr is possible. XRD analysis confirms that both La and Sr are fully incorporated in the LSF lattice. 184

209 These results further confirm the intimate interaction between the primary oxide and support. Such an intimate interaction ensures high oxygen carrier activity and long term recyclability. Figure 2.S5. SEM images at 10000x (5.0kV) LSF supported Fe2O3 sample (a) fresh, (b) 1 st cycle reduced (c) 1 st cycle oxidized (d) 51 st cycle oxidized 185

210 Figure 2.S6. EDS line scans at 5000x and 10 kv of the (a) fresh LSF supported oxygen carrier and (b) the 51 st cycle oxidized LSF supported oxygen carrier With regards to the Al2O3 and YSZ supported oxygen carriers, XRD analysis was performed to ensure the absence of solid state reactions between iron oxide and supports since the formation of new phases will affect the reactivity measurements. XRD results shown in Figure 2.S7 show no evidence of new phase formation. The slight differences in noise and peak intensity levels between the fresh sample and after 5 cycles for Al2O3 supported iron oxide is due to the sample size and amount for the fresh sample compared to that of the after 5-cycles. 186

211 Figure 2.S7. XRD spectra of (a) fresh and after 5-cycles Al2O3 supported iron oxide and (b) after 5-cycles YSZ supported iron oxide. REFERENCES S1. A. Murugan, A. Thursfield, and I. S. Metcalfe, Energ. Environ. Sci., 2011, 4, S2. T. Ishihara, T. Yamada, H. Arikawa, H. Nishiguchi, and Y. Takita. Solid State Ionics. 2000, 135, S3. S.B. Ha, Y.H. Cho, Y.C. Kang, and J.H. Lee. J Eur Ceram Soc. 2010, 30, , S4. J. Sunarso, S. Baumann, J.M. Serra, W.A. Meulenberg, S. Liu, Y.S. Lin, and J.C. Diniz da Costa.J Membrane Sci. 2008, 320,

212 Appendix B Supplementary Information for Chapter 3 This supporting information contains the following sections: S1 Experimental S2 Results 188

213 S1 Experimental S1.1 Oxygen Carrier Synthesis Oxygen carrier synthesis procedures are described in the main body of text in the present work. S1.2 Reactivity Studies Reactivity studies are conducted in a SETARAM SETSYS Evolution Thermal Gravimetric Analyzer (TGA). Up to 75 mg powdery samples are loaded into the TGA. Both H2 and CH4 are used as the reducing gas and O2 is used as the oxidizing gas. Gas flow rates are adjusted so that there is 10% reducing (H2 or CH4) or oxidizing (O2) gas balanced with inert (He). Unless otherwise specified, the temperature inside the TGA is held at 900 C with a total flow rate of 300mL/min to eliminate external mass transfer limitation. S1.3 Morphological Characterizations A number of characterization tools are utilized to characterize surface areas, pore size distributions, bulk compositions, and morphological properties of the oxygen carriers before and after redox reactions. Surface areas and pore size distributions are measured using a BET surface analyzer (Quantachrome QuadraSorb Station 1) using nitrogen or krypton physisorption at 77.3 K. For image capturing, a scanning electron microscope (SEM, Hitachi S3200) is used to observe surface morphology and structure. The same instrument is used to perform X-ray spectroscopy (EDX) scans on the particles to obtain point and area compositions at a range of accelerating voltages (10-30 kev). Transmission Electron Microscopy (TEM, JEOL JEM 2010F) with 200 kev accelerating voltage is used to characterize carbon formation on MgAl2O4 sample. 189

214 S2 Results S2.1 Oxidation and Reduction Rates Figure 3.S1 illustrates typical TGA weight loss curves for oxygen carriers in the main paper. The 2 nd cycle reduction with H2 and CH4 are shown for the perovskite, ceria, MgAl2O4 supported iron oxide samples. It is observed that in all cases CH4 reduction is slower than when reduced with H2. For the perovskite supported iron oxide (Figure 3.S1 (a)), the conversion is seen to be larger than 100%. This is due to the oxygen contribution from the perovskite support. Figure 3.S1- Typical TGA conversion profiles of the Ca0.8Sr0.2Ti0.8Ni0.2O3, CeO2, and MgAl2O4 supported iron oxides at 900 C during the 2 nd reduction in (a) 10% H2 and (b) 10% CH4 Figure 3.S2 examines the reduction and oxidation times in the redox cycles using H2 and O2. CH4 and O2 redox cycles are not examined since it is difficult to determine the initial onset of oxidation due to coke burn-off. Oxidation rates are as fast as or faster in all cases than the reduction step. There is also much less observed deactivation during the oxidations steps. Figure 3.S2 compares the time until the oxidation curve flattens out (having no more weight gained) and the time to achieve that same conversion during the reduction step. 190

215 Figure 3.S2- Time to achieve full oxidation/reduction at 900 C (a) Ca0.8Sr0.2Ti0.8Ni0.2O3, (b) CeO2 and (c) MgAl2O4 supported iron oxides during the 2 nd redox cycle in 10% H2/O2. S2.2 Ceria Support Ce-Fe oxygen carriers have been extensively studied in previous literatures. [1 11] Change in the redox activity has not been specifically reported. Most of the articles did not provide adequate multi-cyclic data for comparison purpose except for references 1 and 4. To confirm the generality of the oxygen carriers synthesized in the current study, their performances are compared with those reported in literature under similar conditions. Results shown in Figure 3.S3 are completed using the 7/3 Ce-Fe co-precipitated oxygen carrier. Figure 3.S3 (a) compares results of the 5 th cycle of the reduction in H2 at 800 C with results given by Bhasvar et al.[1] In that article, the authors synthesized the oxygen carrier by incipient wet method of 191

216 iron onto ceria. It should be noted that concentrations of methane was not specified in the article. Our work consistently used 10% volume concentration of the redox gases. Figure 3.S3 (b) compares activity with methane. He et al. use a similar oxygen carrier to the one synthesized in this work for their methane tests.[4] Under their conditions, the authors looked at the reduction from Fe2O3 to FeO. The authors used the degree of oxidation, X, shown in equation S1 to quantify oxygen carrier activity. X = m m red m ox m red (eq. S1) where m is the instantaneous mass at time t, mred is the mass in the fully reduced state, and mox is the mass in the fully oxidized form. The authors also contributed the reduction of CeO2 to Ce2O3 in their activity comparison. Our work did not show any conclusive evidence that CeO2 is being reduced in any distinguishable amount by XRD. The authors did not specify CH4 concentration. In our study, 10% methane was used. As can be seen, the oxygen carrier synthesized in the current article exhibits comparable or faster reduction rate compared to the literatures. The faster rates of the current study may have stemmed from differences in reducing gas concentrations since they are not reported in available literature. 192

217 Figure 3.S3- (a) Comparison of the 7:3 co-precipitated Ce:Fe oxygen carrier with Bhavsar et al. in H2 at 800 C.[1] (b) Comparison of the 7:3 co-precipitated oxygen carrier with work done by He et. al in CH4 at 850 C. [4] To examine whether kinetic or diffusion limitations are seen, particle size effects are determined at 900 C in 10% methane. The results are shown in Figure 3.S4. As can be seen, initially in cycle 1 the reaction rate up to approximately 40% conversion is close between the and micron sized particles. This indicates that ionic conduction or surface reaction controls the reaction during this step of the reaction. Past 40% conversion, the reaction rate for the larger particle slows down significantly, indicating potential diffusion limitation due to the formation of small pores. [12] Diffusion limitation is further confirmed with 5 th cycle data: far more significant deactivation is observed for larger oxygen carrier than its smaller counterpart. This confirms that reaction rate is limited by intra-particle gaseous diffusion due to loss of porous structure over the redox cycles (as confirmed by Table 3.3 in the main article). As confirmed by experiments, porosity loss has more significant for larger oxygen carriers, indicating intra-particle diffusion control. Figure 3.S4-1 st and 5 th cycle comparisons of micron and micron Ce:Fe (7:3) CP oxygen carrier in methane at 900 C. 193

218 Table 3.S1 indicates the lattice parameters of the perovskite phase before and after the first 10 redox cycles in methane/oxygen. As seen, the b lattice parameter decreases indicating that likely Ni 2+ is being replaced with smaller Fe 3+ cations. Fresh 10 th cycled oxidized a (Å) b (Å) c (Å) V (10 6 pm 3 ) Figure 3.S5 indicates the pore-size distributions (PSD) of the ceria supported oxygen carrier before and after methane redox cycling. As can be seen, initially there is a large amount of meso-pores in the 10-50nm range. However, after 51 cycles most of the pores are found in the 1-4 nm scale. Figure 3.S5- Pore size distribution of the (a) fresh and (b) 51 st oxidized co-precipitated CeO2:Fe2O3 (6:4) Morphological and surface analysis are performed using SEM/EDX analysis at accelerating voltages between 10-30keV. The penetration depth, which is on the order of a few microns, 194

219 changes with accelerating voltages. This procedure leads to an estimate on how relative metal ions are distributed towards the surface of the oxygen carrier. Figure 3.S6 (a)-(c) shows SEM images under various accelerating voltages and example areas in which EDX scans were performed. Figure 3.S6 (d) illustrates the Fe/Ce ratio found at the various accelerating voltages. It gives a clear indication at the lowest accelerating voltage (10keV) there is a higher value of Fe/Ce ratio indicating that Fe is concentrated at the sample surface. Figure 3.S6- SEM images of the fresh CeO2 supported Fe2O3 oxygen carrier at (a) 10 kev (b) 20 kev and (c) 30 kev. EDX scans were taken of 5-7 randomly selected particles and averaged to obtain an average Fe/Ce ratio which is shown in (d) along with the standard deviation error bars for each accelerating voltage. 195

220 To further confirm the enrichment of Fe on the surface of the cycled materials, we investigated the EDX spectras of the fresh and cycled 6:4 and 7:3 Ce:Fe CP oxygen carriers. These are shown in Tables 3.S2 and 3.S3. Table 3.S2- EDX spectra of the 6:4 Ce:Fe co-precipitated oxygen carrier in the fresh, 51 st oxidized, and 52 nd reduced states. Standard deviations and 95% confidence intervals are also given. Metal Ions (Atomic %) Bulk Fresh 51 st oxidized 52 nd reduced Average Fe 57.1% 71.7% 86.1% 66.7% Average Ce 42.9% 28.3% 13.9% 33.3% Average Fe/Ce Ratio Sample Size n/a % Confidence Interval for Fe/Ce Ratio ± ±4.4* 2.0± % CI for surface Fe percentage (%) 57.1% 71.7± ± ±7.6 *large confidence interval exhibited by 51 st oxidized sample is stemmed from significant enrichment of iron oxide on large portion of the sample surface (>95% in some cases), significantly skewing the Fe/Ce ratio. Fe enrichment on the surface is also confirmed by surface Fe percentage, defined by Fe/(Fe+Ce) 100% While the 7:3 co-precipitated oxygen carrier deactivated at a slower rate, we still observed the enrichment of Fe on the surface of the oxygen carrier. Figure 3.S7 shows the activity of the 7:3 oxygen carrier over the 50 cycles. 196

221 Figure 3.S7- Average conversion rate to 33% conversion of the 7:3 Ce:Fe co-precipitated oxygen carrier during the redox in methane-oxygen at 900 C. Table 3.S3- EDX spectra of the 7:3 Ce:Fe co-precipitated oxygen carrier in the fresh and 52nd oxidized. Standard deviations and 95% confidence intervals are also given. Metal Ions (Atomic %) Bulk Fresh 52 nd Oxidized Average Fe 46.2% 57.6% 74.4% Average Ce 53.8% 42.4% 25.6% Average Fe/Ce Ratio Sample Size n/a % Confidence Interval n/a 1.36± ±1.32 By plotting the TGA curves (Figure 3.S8), we observe continual deactivation with respect to cycle number. If the formation of CeFeO3 phase causes the deactivation, it would be expected that the activity would stabilize upon formation of the CeFeO3. This, however, was not observed. As can be seen, the initial activity of the sample is high and the reduction curve follows an inverse sigmoidal shape. With the formation of the ternary perovskite phase, CeFeO3, a longer reduction period is observed (cycles 5 and 10). While the formation of perovskite phase decelerates, it is nonetheless reducible to metallic iron. With increasing 197

222 number of cycles, it is observed that the reduction curve begins to exhibit a sigmoidal shape. To investigate the effect of fuel type, Figure 3.S9 (a) illustrates the activity of the Ce-Fe (6:4) oxygen carrier in hydrogen for 10 cycles at 900 C. The oxygen carrier deactivates by only 20% in hydrogen whereas under similar conditions it deactivated by 75% in methane. To elucidate whether the migration may occur retrospectively of the fuel used, the oxygen carrier was reduced in methane after already being exposed to 10 cycles in hydrogen/oxygen redox. It is observant that the activities are very similar. It is likely that enrichment of iron is thus occurring despite the type of the fuel. The main contributing factor between the large differences in deactivation is that methane has a lower diffusivity through iron oxide agglomerates than hydrogen. Figure 3.S8- TGA curve of 800 C (6:4) co-precipitated ceria supported oxygen carrier over 50 cycle redox at 900 C in methane. 198

223 Figure 3.S9- (a) Average conversion rate to 33% conversion in H2 at 900 C and (b) comparison of fuel effect on activity of the ceria based oxygen carrier at 900 C. S2.3 MgAl2O4 Support Figure 3.S10 indicates the pore size distribution of the MgAl2O4 supported iron oxide before and after 5 redox cycles in methane. Initially, it is observed that most pores are the micro pores in the range of 0.4-2nm and meso-pores in the range of 2-50 nm. After only 5 redox cycles, macro pores are observed to form up to 400nm. Figure 3.S10- Pore size distribution of the (a) fresh and (b) 5 th oxidized SSR MgAl2O4:Fe2O3 (4:6) 199

224 We observed carbon filament and nanotube formation on the 6 th reduced inert supported oxygen carrier indicating that carbon formation was the driving force of removal of small metal nanoparticles causing interstitial gaps and pores. This increase in surface area is believed to cause the activation observed in the inert supported oxygen carrier. Figure 3.S11 indicates that carbon formation was not significant during the first reduction. This is consistent with the XRD results in which no carbon or iron carbide phases was identified. Figure 3.S11- TEM image of the 1 st cycle reduced MgAl2O4 supported iron oxide in methane. REFERENCES [1] S. Bhavsar, G. Veser, Ind. Eng. Chem. Res. 52 (2013) [2] V. Galvita, T. Hempel, H. Lorenz, L.K. Rihko-Struckmann, K. Sundmacher, Ind. Eng. Chem. Res. 47 (2008) 303. [3] Z. Gu, K. Li, H. Wang, Y. Wei, D. Yan, T. Qiao, Kinet. Catal. 54 (2013) 326. [4] F. He, Y. Wei, H. Li, H. Wang, Energy Fuels 23 (2009) [5] A. Hedayati, A.-M. Azad, M. Rydén, H. Leion, T. Mattisson, Ind. Eng. Chem. Res. 51 (2012)

225 [6] K. Li, H. Wang, Y. Wei, D. Yan, Chem. Eng. J. 156 (2010) 512. [7] K. Li, H. Wang, Y. Wei, D. Yan, Appl. Catal. B Environ. 97 (2010) 361. [8] K. Li, H. Wang, Y. Wei, M. Liu, J. Rare Earths 26 (2008) 705. [9] Y. Wei, H. Wang, K. Li, J. Rare Earths 28 (2010) 560. [10] X. Zhu, H. Wang, Y. Wei, K. Li, X. Cheng, J. Rare Earths 28 (2010) 907. [11] X. Zhu, K. Li, Y. Wei, H. Wang, L. Sun, Energy Fuels 28 (2014) 754. [12] K. Piotrowski, K. Mondal, H. Lorethova, L. Stonawski, T. Szymanski, T. Wiltowski, Int. J. Hydrog. Energy 30 (2005)

226 Appendix C Supplementary Information for Chapter 4 This supporting information contains the following sections: S1 Experimental S2 Results 202

227 S1 Experimental S1.1 Oxygen Carrier Synthesis Oxygen carrier synthesis procedures are discussed in the main document of the present work. S1.2 Fluidized Bed Studies A laboratory-scale fluidized-bed reactor is used for solid fuel conversion experiment (Figure 4.S1). The quartz reactor has an inner diameter of 2.54cm. The reactor is heated through external heating from a tube furnace (MTI OTF-1200X-S-VT). Temperature is measured inside the reactor initially with an Inconel (Type K) thermocouple to set the temperature of the reactor to 850 C. The thermocouple is replaced with a 31.75mm ID stainless steel tube that is used for solid fuel injection into the reactor. A gas mixing panel is used to send the desired ratio of gases from both the bottom of the reactor for fluidization and through the smaller stainless steel tube for solid fuel injection. For maintaining fluidization, flow rates of 1080 and 1200 ml/min are used for solid fuel conversion (inert environment) and oxidation cycles respectively. Corresponding gas flow rates give a residence time of 0.13 and 0.14 s. The flow rates correspond to 6-7 times the calculated minimum fluidization velocity, Umf, of the synthesized oxygen carrier. Prior to experiments, the reactor bed is loaded with 16 mesh silicon carbide (Kramer Industries) at the bottom of the reactor. The SiC particles act as a gas preheater and distributer. The SiC layer helps supports the oxygen carrier layer as well. After the SiC layer is placed, approximately 15 grams of the oxygen carrier is placed into the reactor. 203

228 Figure 4.S1- Fluidized bed reactor design that is used for solid fuel redox cycles in the present work. For the redox reactions of solid fuel, a bituminous coal (Asbury Inc.) is converted into devolatized coal char through heating in the reactor at 800 C in a N2 environment. Injection of the coal is conducted in a pulse mode by loading from the top of the reactor through the small stainless steel tube and pulsing it into the bottom of the oxygen carrier layer of the reactor bed using N2. 10 redox cycles are tested using the fluidized bed reactor conducting two experiments in those 10 cycles. During cycles 1, 3, 5, and 10 solid fuel conversion is tested. Cycles 2, 4, and 6-9 are examining the oxygen uncoupling properties. Exit gas concentrations are measured using a quadrapole mass spectrometer (MKS Cirrus 2) and a near-ir based gas analyzer (Emerson X-Stream). S1.3 Oxygen Carrier Characterizations 204

229 Phase identification of the oxygen carriers are conducted using X-ray powder diffraction. Powdered samples are prepared and tested using a Rigaku SmartLab X-ray diffractometer with Cu-Kα (λ=0.1542nm) radiation operating at 40kV and 44mA. A scanning range of (2θ) with a step size of 0.1 holding for 3.5 seconds at each step is used to generate the XRD patterns. S2 Results S2.1 Phase stability in CLOU redox cycles Figure 4.S2- XRD pattern of the Ca0.75Sr0.25MnO3 as prepared and after 100 oxygen uncoupling cycles. Figure 4.S2 examines the phase stability of the Sr-doped oxygen carrier during the 100 cycle oxygen uncoupling experiments at 850 C. As can be seen, the as prepared sample has a single distinct orthorhombic perovskite phase. After cycling, it is observed some of the Sr and Mn come out of the structure to form a SrMnO3 phase. It should also be noted no Ca2MnO4 or CaMn2O4 phases observed. The resulting phase segregation does not observable effect the performance of the oxygen carrier. 205

230 S2.2 Fludized Bed Experiments Figure 4.S3- (a) CO2 and (b) CO concentration profiles during the fluidized bed experiments. Temperature: 850 C Flow rates: 1080 ml/min during char injection and 1200mL/min during oxidation (10% O2 balanced with inert) Figure 4.S3 examines the concentration profiles of CO2 and CO during the char oxidation. Integration of the concentration profiles will lead to overall char conversion and selectivity of the combustion process. Calculated results are presented in the main text in which the oxygen carrier converted nearly 100% of the char during every cycle with nearly 90% selectivity towards CO2. 206

231 Appendix D Supplementary Information for Chapter 5 This supporting information contains the following sections: S1 Experimental S2 Results 207

232 S1 Experimental S1.1 Oxygen Carrier Synthesis Oxygen carrier synthesis procedures are discussed in the main document of the present work. S1.2 Fluidized Bed Studies A laboratory-scale fluidized-bed reactor is used for solid fuel conversion experiment (Figure 5.S1). The quartz fluidized-bed reactor has an outer diameter of 25.4 mm. The reactor is heated through external heating from a tube furnace (MTI OTF-1200X-S-VT). Temperature is measured inside the reactor with an Inconel (Type K) thermocouple to set the temperature of the reactor to 850 C. A 3.75mm ID stainless steel tube is used for solid fuel injection into the reactor. A gas mixing panel is used to send the desired ratio of gases from both the bottom of the reactor for fluidization and through the smaller stainless steel tube for solid fuel injection. For maintaining fluidization, flow rates of 1,080 and 1,200 ml/min (at S.T.P) are used for solid fuel conversion in inert environment and 10% O2 environments, respectively. The flow rates correspond to 6-7 times the calculated minimum fluidization velocity, Umf, of the synthesized oxygen carrier. Prior to experiments, the reactor bed is loaded with 16 mesh silicon carbide (Kramer Industries) at the bottom of the reactor. The SiC particles act as a gas preheater and distributor. The SiC layer also supports the oxygen carrier bed. After the SiC layer is placed, approximately 15 grams of the oxygen carrier is placed into the reactor. 208

233 Figure 5.S1- Fluidized bed reactor design that is used for solid fuel redox cycles in the present work. For redox reactions of solid fuel, Pittsburgh #8 coal (-60 mesh) is converted into devolatized coal char through heating in the reactor at 950 C in a N2 environment. The resulting char is sieved into a range of microns. Injection of the coal is conducted in a pulse mode by loading from the top of the reactor through the small stainless steel tube and is pulsed into the bottom of the oxygen carrier layer of the reactor bed using N2. 20 redox cycles are tested using the fluidized bed reactor conducting two experiments in those 20 cycles. The exact conditions for the redox experiments with solid fuel can be found in Table 4.1 in the main text. S2 Results S2.1 Fludized Bed Experiments 209

234 Figure 5.S2- (a) CO and (b) CO2 concentration profiles during the fluidized bed experiments with the CaMn0.95Fe0.05O3 oxygen carrier. Sample Weight: 15 gms ( microns) Temperature: 850 C Flow rates: 1080 ml/min during char injection and 1200mL/min during oxidation (10% O2 balanced with inert) Figure 5.S2 examines the concentration profiles of CO2 and CO during the char oxidation of the Fe-doped oxygen carrier. Integration of the concentration profiles will lead to overall char conversion and selectivity of the combustion process. Calculated results are presented in the main text in which the oxygen carrier converted approximately 90% of the char during every cycle with nearly 99% selectivity towards CO2. In Figure 5.S3, an example concentration profile for one injection cycle on CaMnO3 for solid injection at 850 and 950 C is shown. At a lower temperature of 850 C, undoped CaMnO3 only observed a 50% conversion of the char while the increase in temperature increases to the conversion to 95%. The carbon balance for all cycles was within 20% of the theoretical estimate. 210