Calcium and Iron Oxide Reactivity Studies for Chemical Looping Applications of Clean Energy Conversion

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Niranjani Deshpande

Graduate Program in Chemical Engineering

The Ohio State University

2015

Dissertation Committee:

Professor Liang-Shih Fan, Advisor

Professor Bhavik R. Bakshi

Professor David L. Tomasko

Niranjani Deshpande

2015

Abstract

The following study entails independent investigations carried out on the reactivity of metal oxides involved in the calcium and chemical looping applications. The Chapters 2 through 5 involve studies on the various applications and aspects of the calcium looping process, and Chapter 6 and 7 discuss two independent investigations of chemical looping oxygen carrier particles.

The hydration of calcium oxide (CaO) sorbent has been investigated as a reactivation method in the three step calcium looping process for pre and post combustion carbon dioxide (CO2) capture. The feasibility of the process concept was established using lab scale fixed bed reactor setup, and reactivation of sorbent was achieved with high temperature steam at 500°C over multiple cycles. Further development of the design and operation of a fluidized bed hydrator is reported upon, and fast fluidization regime was identified as the most suitable for a scalable steam hydrator design. Further, a screening study was conducted on multiple egg and sea shells as a renewable source of the CaO sorbent, and excellent reactivity towards CO2 is reported. A novel method for the simultaneous cleanup of CO2, SOx and NOx impurities from coal combustion flue gas is proposed based on the calcium looping process. Proof of concept experiments were performed and 90% CO2 and NO and 100% SO2 removal was demonstrated at 1 atm,

650°C fixed bed experiments, using the calcium sorbent and coal char. For pre- ii combustion application of the calcium looping process (CLP), the fate of sulfurous species is explored, which are formed as a byproduct of the coal to H2 plant with the

CLP. The CaS formed in the carbonator at the operating conditions of about 600°C and

23 bar is found to be oxidized to CaSO4 at the calciner operating conditions of the CLP.

Treatment options for the purge stream are discussed for the oxidation of unreacted CaS for the safe disposal and integration with the cement industry.

In the latter half of the present study, the iron-based metal oxide oxygen carriers are investigated for the chemical looping partial oxidation (CLPO) of CH4 for the production of syngas at elevated pressures. The favorable impact of increased pressure on the redox reaction rates is illustrated through experiments conducted on the iron-titanium complex metal oxide (ITCMO) particles between 1 and 10 atm at 900-950°C in a thermogravimetric setup. The observed change in morphology through SEM and BET analysis at increased pressures is related to the change in reactivity obtained. Lastly, an application of chemical looping gasification (CLG) for the coproduction of H2 and electricity is explored. Specifically, the recyclability of iron based oxygen carriers is investigated in steam redox environments using a specialized thermogravimetric setup.

Isothermal tests are conducted for 20 redox cycles using steam as the oxidizing agent for iron and cobalt based metal oxide oxygen carriers at 900°C. MgAl2O4 is used as an inert support. While cobalt-based samples exhibited a loss in reactivity, the excellent recyclability of iron-based oxygen carriers has thus been established.

iii

This document is dedicated to my little brother, Sukumar. Your memory always inspires

me to be a better person, and gives me warmth in my darkest hour.

Acknowledgments

I would like to express my deepest gratitude towards my advisor, Professor Liang-Shih

Fan, for offering me the opportunity to work on such a fast and exciting field as chemical looping. I am thankful to the Ohio State University and particularly the William G.

Lowrie Department of Chemical and Biomolecular Engineering (CBE) for all the excellent resources and infrastructure that was provided for my easy use that made my graduate research experience such a joy. The state-of-the-art facilities available for students gave me a one of kind experience and a unique flavor of academic research. The

CBE family at large, and the Fan group in particular, has helped shape my keen research acumen over the last five years, and instilled in me a deep appreciation of the role of scientific investigation in the overall human development. I would also like to thank Prof.

James Rathman, Prof. Bhavik Bakshi, Prof. David Tomasko and Prof. Lisa Hall for serving on my qualifier, candidacy, and dissertation committees. Their discussions always provided me with new ideas to further my research objectives.

Dr. Fan has been a constant source of inspiration to me, not only in his role as a direct advisor for my research progress, but also leading by example a life of dedication and discipline. His endless enthusiasm and optimism towards researching solutions for various technological challenges is something I will always aspire to imbibe in myself.

Dr. Nihar Phalak, who was a colleague and a senior in the calcium looping sub-group, shared the better part of his graduate career with mine. Nihar has been invaluable to me in his steadfast friendship, guidance, mentorship and support. I am forever indebted to him for his close involvement and interest in my graduate career development, as well as the close personal friendship, which made my doctoral study a fulfilling and enriching experience. In addition, Dr. Shwetha Ramkumar played a key role in my initial mentoring in the Fan lab. William Wang was always available to discuss any doubts I may have had about the calcium looping system. His in-depth understanding of the process and the power generation process always provided me fodder for new ideas. My other seniors

Deepak Sridhar, Ray Kim, Liang Zeng, and Andrew Tong guided me in various capacities, and I am thankful for their guidance and support. I also enjoyed working with

Alan Wang for the steam hydration process for reactivation of the calcium sorbent. Dr.

Lang Qin provided me her immense expertise in the FIB, EDS and various microscopy techniques. Discussions with her were invaluable to the development of oxygen carrier studies. Ankita Majumder, Elena Chung, and Mandar Kathe were a delight to work with, and transcended the boundaries of co-workers and formed close personal friendships with me. To all three of them, I am forever indebted. Other members of my Fan group family include Dr. Samuel Bayham, Dr. Qiang Zhou, Dr. Dawei Wang, Omar McGiveron,

Aining Wang, Cheng Chung, Dikai Xu, and Tien-Lin Hseh; who were always available to discuss any concepts, and lend support in the best team spirit and great rapport.

I would also like to mention Nicholas Blum, who I had the pleasure to mentor in his undergraduate research efforts. His insightful questions and keen interest in research were

vi extremely helpful for my development as well. I must also mention Brian Yuh, who I mentored for his high school summer internship. The Yuh family’s kindness and warmth will stay with me forever.

Dr. Robert Statnick (ClearSkies Consulting) and Mr. Dan Connell (CONSOL Energy

R&D) provided great insight from their vast industrial experience, through our many discussions, conference calls, and collaborative efforts that I was a part of. I am thankful for the contributions of Mr. Bob Brown in an advisory capacity for the development of the calcium study funded the Ohio Coal Development Office (OCDO). I would like to gratefully acknowledge Mr. Joe Eutizi (San Miguel Electric Cooperative Inc.) and Mr.

David Martin (Walnut Creek Mining Company) for providing the lignite coal samples, and Mr. Daniel Wilson (CONSOL Energy R&D) for help in char production from the coal samples. I am grateful for the financial support provided by projects funded through

OCDO as well as the United States Department of Energy (USDOE).

Special thanks to Mr. Paul Green and Mr. Michael Wilson. Their skills in the machine shop and their willingness to always help were key to the successful and timely completion of many of my lab scale studies. I must mention Dr. Carlo Scaccia, under whose guidance I completed my teaching assignments as a part of my doctoral studies. I learnt a great deal about professional ethics and maintaining good professional relationships from him. Angela Bennett, Lynn Flanagan, and Susan Tesfai of the CBE department always provided professional and timely assistance in all my administrative tasks. I greatly appreciate all their help.

vii

Special thanks to my extended family, your love keeps me strong even from halfway across the world. My friends near and far, who are always ready to lend me a patient hearing, or provide good counsel when I need it. Last but most importantly, I want to mention my family, my parents Ashwini and Rajendra, who have supported me unconditionally in every endeavor, who are my biggest pillars of strength and support.

They are shiny examples of a purposeful life well-lived, I aspire to be like them every day. Also, my fiancé, Harshavardhan, who is my rock. Your love uplifts and inspires me.

viii

Vita

June 2003 ...... S.S.C., Balmohan Vidyamandir June 2005 ...... H.S.C., Mumbai University June 2009 ...... B. Chem., Mumbai University, Institute of Chemical Technology Sept 2009 to present ...... Graduate Research Associate, Department of Chemical and Biomolecular Engineering, The Ohio State University

Publications

Deshpande, N.; Majumder, A.; Qin, L.; Fan, L.-S. High-Pressure Redox Behavior of

Iron-Oxide-Based Oxygen Carriers for Syngas Generation from Methane. Energy Fuels

2015.

Deshpande, N.; Phalak, N.; Fan, L.-S.; Sankaran, S. Carbon Dioxide (CO2) Capture from

Coal-Fired Power Plants Using Calcium Looping. Chem. Eng. Educ. 2015.

Fan, L.-S.; Deshpande, N.; Phalek, N. United States Patent: 8877150 - Single-Step

Process for the Simultaneous Removal of CO2, SOx and NOx from a Gas Mixture.

8877150, November 4, 2014.

Wang, A.; Deshpande, N.; Fan, L.-S. Steam Hydration of Calcium Oxide for Solid

Sorbent Based CO2 Capture: Effects of Sintering and Fluidized Bed Reactor Behavior.

Energy Fuels 2014, 29, 321–330 ix

Luo, S.; Zeng, L.; Xu, D.; Kathe, M.; Chung, E.; Deshpande, N.; Qin, L.; Majumder, A.;

Hsieh, T.-L.; Tong, A.; Sun, Z.; Fan, L.-S. Shale Gas-to-Syngas Chemical Looping

Process for Stable Shale Gas Conversion to High Purity Syngas with a H2:CO Ratio of

2:1. Energy Environ. Sci. 2014, 7, 4104–4117.

Deshpande, N.; Yuh, B. Screening of Multiple Waste Animal Shells as a Source of

Calcium Sorbent for High Temperature CO2 Capture. Sustain. Eng. Res. 2013, 23, 227–

232.

Wang, A.; Wang, D.; Deshpande, N.; Phalak, N.; Wang, W.; Fan, L.-S. Design and

Operation of a Fluidized Bed Hydrator for Steam Reactivation of Calcium Sorbent. Ind.

Eng. Chem. Res. 2013, 52, 2793–2802.

Phalak, N.; Deshpande, N.; Fan, L.-S. Investigation of High-Temperature Steam

Hydration of Naturally Derived Calcium Oxide for Improved Carbon Dioxide Capture

Capacity over Multiple Cycles. Energy Fuels 2012, 26, 3903–3909.

Phalak, N.; Ramkumar, S.; Deshpande, N.; Wang, A.; Wang, W.; Statnick, R. M.; Fan,

L.-S. Calcium Looping Process for Clean Coal Conversion: Design and Operation of the

Subpilot-Scale Carbonator. Ind. Eng. Chem. Res. 2012, 51, 9938–9944.

Fields of Study

Major Field: Chemical Engineering

Table of Contents

Abstract ...... ii Acknowledgments...... v Vita ...... ix Table of Contents ...... xi List of Tables ...... xv List of Figures ...... xvi CHAPTER 1: Introduction ...... 1 1.1 Type I and type II chemical looping systems ...... 1 1.2 Major demonstration plants ...... 3 1.3 Outline of chapters ...... 5 CHAPTER 2: Steam Hydration as Reactivation for the Calcium Sorbent ...... 11 2.1 Introduction ...... 11 2.2 Feasibility of steam hydration ...... 17 2.2.1 Experimental materials and methods ...... 17 2.2.2 Recyclablility of the sorbents over steam hydration ...... 25 2.2.3 Change in sorbent morphology ...... 28 2.2.4 Extended number of cycles ...... 32 2.3 Bench scale semi-batch fluidized bed hydrator ...... 37 2.4 Continuous hydration on entrained bed reactor ...... 39 2.4.1 Experimental setup...... 39 2.4.2 Cold/dry flow tests ...... 42 2.4.3 High temperature experiments ...... 44 2.5 Conclusions ...... 52 CHAPTER 3: Screening of Multiple Waste Animal Shells as a Source of Calcium Sorbent ...... 53 xi

3.1 Introduction ...... 53 3.2 Experimental ...... 55 3.2.1 Materials and methods ...... 55 3.3 Observations and discussion ...... 61 3.4 Conclusions ...... 71 CHAPTER 4: A Novel Calcium-Char (Cal-C) Process for the Simultaneous Removal of NOx, SOx and CO2 from Combustion Flue Gas ...... 72 4.1 Introduction ...... 72 4.1.1 Process overview ...... 75 4.2 Experimental section ...... 80 4.3 Results and discussion ...... 83 4.3.1 Effect of temperature on NO reduction in presence of CaO: ...... 83 4.3.2 Effect of addition of CaO (presence and absence of calcium) ...... 86

4.3.3 Effect of O2 concentration ...... 90 4.3.4 Effect of inlet NO concentration ...... 93

4.3.5 Simultaneous capture of NO, SO2, and CO2 ...... 95 4.4 ASPEN simulations ...... 97 4.5 Conclusion ...... 105

CHAPTER 5: Calcium Looping Process for Coal-to-H2 Production: Fate of Sulfur ..... 107 5.1 Introduction ...... 107 5.2 Motivation/Problem statement ...... 110 5.3 Background and literature review ...... 115 5.4 Conditions tested ...... 117 5.5 Materials ...... 119 5.5.1 Experimental setup and procedure ...... 119 5.6 Results and discussion ...... 123

5.6.1 Reaction of CaS with CO2 as the oxidizing agent ...... 123

5.6.2 Reaction of CaS with H2O as the oxidizing agent ...... 125

5.6.3 Reaction of CaS with O2 as the oxidizing agent ...... 128

5.6.4 Reaction of CaO with SO2 released from oxycombustion of coal ...... 131

5.6.5 Reaction of CaSO4 with H2 ...... 135 xii

5.6.6 Reaction of CaSO4 with CO ...... 139

5.6.7 Reaction of CaSO4 with H2 and CO ...... 144 5.6.8 Treatment of purge stream ...... 149 5.7 Commercial implications and conclusions ...... 156 CHAPTER 6: Chemical Looping Applications: High Pressure Redox Behavior of Iron- Oxide Based Oxygen Carriers ...... 162 6.1 Introduction ...... 162 6.2 Thermodynamic analysis...... 168 6.3 Experimental setup, materials and procedure: ...... 178 6.4 Results and discussion ...... 180

6.4.1 Reduction in H2 ...... 181

6.4.2 Reduction in CH4 ...... 190 6.4.3 Pressure correction ...... 198 6.4.4 Air oxidation ...... 200 6.4.5 XRD, SEM, EDS, and BET analysis ...... 202 6.5 Conclusions ...... 208 CHAPTER 7: Chemical Looping Applications: Redox Reactivity of Steam Oxidation for Chemical Looping Particles ...... 211 7.1 Introduction ...... 211 7.2 Thermodynamic analysis...... 218 7.3 Materials and methods ...... 221 7.4 Results and discussion ...... 222 7.4.1 Fe-based oxygen carriers ...... 222 7.4.2 Co-based oxygen carriers ...... 226 7.5 Conclusions ...... 231 FUTURE DIRECTIONS ...... 233 APPENDIX: Supplemental Data ...... 236 A.1 Calcium sorbent reactivation by hydration ...... 236 A.1.1 Steam hydration TGA experiments ...... 236 A.1.2 Decay in reactivity of CaO sorbent over continuous TGA testing ...... 238 A.2 Fate of Sulfur ...... 242

xiii

A.2.1 Soot formation ...... 242 A.2.2 The Buoyancy change ...... 242

A.2.3 Investigation of CaSO4 reaction ...... 244 A.3 Reduction of ITCMO particles under pressure ...... 245

A.3.1 CH4 reduction at 900°C ...... 245

A.3.2 H2 reduction at 900°C ...... 246 A.4 Steam oxidation of reduced oxygen carrier samples ...... 247 A.4.1 Sample calculation of extent of oxidation for Fe-based oxygen carriers ...... 247 A.4.2 Raw data of 20 redox cycles of Fe-based oxygen carriers ...... 250 A.4.3 Co-based oxygen carriers upon re-oxidation ...... 252 REFERENCES ...... 254

xiv

List of Tables

Table 1: Properties for sorbents generated from different limestones ...... 19

Table 2: Different animal shells tested and their sources, with initial CO2 capture capacities ...... 58

Table 3: Calcined and hydrated CO2 capture capacities of sorbents after 30 min, tested in TGA ...... 62 Table 4. Bulk pore volume and surface area analysis using BET theory ...... 64 Table 5: Modeling parameters for CCR Process...... 98

Table 6: 550 MWe coal-fired power plant – process conditions for Case 1 ...... 99 Table 7: Modeling parameters for NO removal for the Cal-C process ...... 100 Table 8: Summary of simulation results for the four cases ...... 101 Table 9: Calciner equilibrium gas concentrations from Aspen Plus simulations ...... 118 Table 10: Carbonator equilibrium gas concentrations derived from Aspen Plus simulations ...... 118

Table 11: Cost comparison for H2 and electricity generation for coal to H2 plant, base case and CLP plant...... 160 Table 12: Methane optimum equilibrium conversion results for partial oxidation using ITCMO particles ...... 175

Table 13: YH2 as a function of total system pressure and PPH2 for section 6.4.1.1 ...... 184 Table 14: The thermodynamic properties for reactions of the various oxidation states of Fe and Co with steam, per mole of H2 produced, at 900°C ...... 220 Table 15: Calculation of extent of steam oxidation of the Fe-based oxygen carriers by the three methods...... 249 Table 16: Raw weight data for the 20 redox cycles for Fe-based oxygen carriers ...... 251

List of Figures

Figure 1: Concept of type I and type II chemical looping, for post combustion scenario .. 2 Figure 2: The steam pressure as a function of temperature for the hydration reaction. ... 14 Figure 3: Exponential increase in the equilibrium reaction pressure with temperature, in the temperature range of interest...... 16 Figure 4: Fixed bed reactor setup for the high temperature steam hydration. Figure from 39 ...... 21 Figure 5: The post-hydration capture capacity of the CaO sorbents derived from the six different limestones over five hydration cycles ...... 26

Figure 6: Comparison of the post-hydration CO2 capture capacities of the calcitic sorbents, on calcium and hydrate basis...... 27 Figure 7: Particle size distribution post calcination and post hydration for (a) MV and (b) sorbent ...... 29 Figure 8: Calcined MV sorbent (a) cycle 1 and (b) cycle 5 ...... 30 Figure 9: Calcined BR sorbent, (1) cycle 1 and (2) cycle 5 ...... 30 Figure 10: BR sorbent cycle 5 (a)calcined and (b) steam hydrated followed by dehydration in TGA ...... 31 Figure 11: MV sorbent cycle 5 (a) calcined and (b) steam hydrated followed by dehydration in TGA ...... 31 Figure 12: Extended number of cycles showing the effectiveness of hydration as reactivation mechanism...... 34 Figure 13: Hydration conversion as a function of number of cycles...... 36 Figure 14: Hydration performance of the semi-batch fluidized bed hydrator40 ...... 38 Figure 15: Schematic of the sub-pilot scale riser reactor used for continuous steam hydration experiments...... 41 Figure 16: Degree of entrainment obtained as function of gas velocity, cold flow experiments ...... 43 Figure 17: Maximum steam partial pressures achievable as a function of temperature, at different gas velocities...... 46 Figure 18: Wall coating observed in the entrained bed reactor ...... 47 Figure 19: Hydration conversions obtained in the entrained and bed samples as a function of steam partial pressures used ...... 48 xvi

Figure 20: Comparison of entrained bed performance and conversions predicted by Schaube et.al...... 50 Figure 21: Predictions of hydration conversion at OSU’s test conditions using equation 1...... 51 Figure 22: Original shells from which samples were prepared...... 56

Figure 23: CO2 capture capacity (wt%) for (a) calcined sorbents and (b) hydrated sorbents ...... 63 Figure 24: X-ray diffraction patterns obtained from CH-E, OS-E and OY-S samples: determining crystallographic differences...... 68 Figure 25: SEM images of egg and sea-shell derived sorbents. (a) CH-E (b) OS-E and (c) OY-S...... 69 Figure 26: The Cal-C process – conceptual schematic...... 76 Figure 27: Experimental setup of reactor and gas analysis system...... 81 Figure 28: TPR of NO reduction using bituminous and lignite chars. Total flow rate =

210 ml/min, inlet NO concentration = 924 ppm, inlet O2 concentration = 1.5%, Ca:char loading = 10:1 (by wt.) ...... 85 Figure 29: Effect of addition of calcium sorbent on the NO-char reduction reaction using (a) BC, (b) LC-1 and (c) LC-2. Gas flow rate = 210 ml/min, inlet NO concentration = 924 ppm, O2 concentration = 1.5%, temperature = 650 ̊C. Ca:char loading = 10:1 by wt. Hollow symbols indicate absence of calcium sorbent and solid symbols indicate presence of calcium...... 88

Figure 30: Effect of different Ca(OH)2:char loading (by wt) on the selectivity of NO-char reaction, LC-2. Total flow rate = 210 ml/min, inlet NO concentration = 600 ppm, inlet O2 concentration = 3%, temperature = 650 ̊C...... 89

Figure 31: Effect of inlet O2 concentration on the selectivity of char-NO reduction reaction in presence of calcium sorbent and LC-2. Total gas flow rate = 210 ml/min, inlet NO concentration = 600 ppm, temperature = 650 ̊C, Ca:char loading = 10:1 by wt...... 91

Figure 32: Effect of O2 concentration on NO-char isothermal reduction reaction in presence of calcium sorbent and LC-2, pre-breakthrough periods. Inlet NO concentration = 1800 ppm, Ca:char loading = 10:1 by wt, total gas flow rate = 210 ml/min, temperature = 650 ̊C...... 92 Figure 33: Effect on inlet NO concentration on the selectivity of char-NO reduction reaction in presence of calcium sorbent and LC-2. Total gas flow rate = 210 ml/min, inlet

O2 concentration = 1.5%, temperature = 650 ̊C, Ca:char loading = 10:1 by wt...... 94 Figure 34: The simultaneous removal of NO, SO2 and CO2 from a simulated gas mixture, in presence of calcium sorbent and lignite coal char LC-2. Inlet CO2 concentration = 13%, Inlet NO = 1800 ppm, Inlet SO2 = 3050 ppm, Inlet O2 = 1.5%, Ca:char loading = 10:1 by wt. Temperature = 650 ̊C...... 96

xvii

Figure 35: CCR process block flow diagram - heat exchangers not shown for simplicity...... 102 Figure 36: The block flow diagram of CCR process with NOx control – Cal-C process...... 103

Figure 37: Schematic of the three-step CLP process for high purity H2 production from coal-derived syngas39 ...... 110

Figure 38: The Aspen Plus flowsheet showing the CLP system applied to a coal-to-H2 plant. The red dotted arrows indicate the locations of high quality heat recovery to run an auxiliary steam turbine cycle for the coproduction of electricity...... 112 Figure 39: Schematic of the Rubotherm MSB setup used for performing thermogravimetric experiments...... 122

Figure 40: CaS conversion to CaSO4 as a function of time, at different isothermal temperatures.CO2 concentration fixed at 80%, total pressure = 1 atm...... 124 Figure 41: CaS conversion to CaSO4 as a function of time, at different H2O concentrations. Isothermal experiments at (a) 875°C, (b) 900°C, and (c) 925°C...... 127

Figure 42: Oxidation of CaS with oxygen at calciner operating conditions. Varying the O2 concentration at isothermal conditions, product is CaSO4 at (a) 875°C, (b) 900°C, and (c) 925°C ...... 129

Figure 43: Oxidation of CaS to CaSO4 with different oxidizing agents at concentrations relevant to calciner operating conditions, H2O = 24%, CO2 = 80% and O2 = 5%, always balance N2. Total gas flow rate was maintained at ~600 ml/min (at room T) for all experiments and Texperiment = 900°C...... 131 Figure 44: Typical TGA graph for reaction between CaO and SO2/O2 mixture, starting from CaCO3 decomposition in inert N2 ...... 132 Figure 45: CaSO4 formation from CaO at calciner operating conditions at various temperatures. 1% O2, 2000 ppm SO2 ...... 134 Figure 46: Equilibrium constant of CaSO4 decomposition reactions as a function of temperature ...... 136 Figure 47: XRD analysis of (a) reactant and (b) product solid samples - Matched against CaSO4 and CaS...... 138 Figure 48: Carbon deposition indicated by weight increase upon injection of CO. 650°C, 5 atm, 30% CO...... 140

Figure 49: Equilibrium constants of CaSO4 reduction, Boudouard and WGS reactions as a function of temperature ...... 141 Figure 50: 30% CO experiment with prior steam injection, 650°C and 8 atm (a) outlet gas concentrations, and (b) sample weight change measured by the MSB...... 143

xviii

Figure 51: Typical experiment showing weight change (TGA) and outlet gas composition

(micro-GC) as a function of time, upon injection of reactive gases. 30% CO, 30% H2 and steam injection before the experiment. T = 650°C and P = 10 atm...... 145

Figure 52: Comparison of weight changes, when using (a) CaSO4 and (b) CaO solids in the TGA at 650°C, 10 atm, 30% H2 and CO...... 147 Figure 53: Process flow diagram of the CLP unit as applied to the coal-to-H2 process. Dotted line indicates the solid purge stream and possible treatment locations for the sulfurous species (original figure from reference 93)...... 150

Figure 54: CaS oxidation between 700 and 900°C at (a) 5% O2, (b) 10% O2 and (c) 21% O2...... 153 Figure 55: reaction rate data as a function of reaction conversion X, for (a) 5% O2, (b) 10% O2 and (c) 21% O2. The graph clearly shows an initial fast reaction rate followed by a slow diffusion controlled regime of reaction...... 154

Figure 56: Base case coal-to-H2 plant with 2-stage Selexol and Claus plant for sulfur removal93 ...... 159 Figure 57: Schematic of Fe-oxide based system for syn-gas generation from partial oxidation of CH4 ...... 167 Figure 58: Effect of temperature on the methane conversion and syngas purity for different metal oxide systems. (a) CeO2, (b) NiO, (c) Fe2O3 and (d) ITCMO system .... 170 Figure 59: Simulated equilibrium iron oxide phases and fractional carbon deposition as a function of inlet gas:solid ratios at elevated pressure. T = 950 ˚C, P = 5 atm...... 173 Figure 60: Gas composition profiles and methane conversion at 900C, 1 atm with respect to loading of Fe2O3...... 176 Figure 61: Gas composition profiles and methane conversion at 900C, 10 atm with respect to loading of Fe2O3 ...... 176 Figure 62: The effect of total system pressure on rates of reduction at X = 0.75, and constant partial pressure of H2. T = 900 ˚C ...... 183 Figure 63: The effect of mole fraction of reducing gas (YH2) on rates of reduction at X = 0.75, and constant partial pressures. T = 900°C ...... 186 Figure 64: The effect of sysyem pressure on rates of reduction at X = 0.5 and X = 0.75, and constant mole fraction of reducing gas YH2 = 50%, T = 900 ˚C ...... 188 Figure 65: The effect of partial pressure of reducing gas PPH2 on (a) conversion curves obtained and (b) rates of reduction at constant system pressure P = 5 atm. T = 900°C . 189

Figure 66: Reduction conversion curves obtained using CH4 from the thermogravimetric analysis between 1 and 10 atm at constant mole fraction of reducing gas YCH4 = 50%. T = 950 ˚C...... 191 Figure 67: The effect of system pressure on reaction rate for the three-step reduction with

CH4 as the reducing gas. YCH4 = 50%, T = 950°C, P = 1 to 10 atm ...... 194

xix

Figure 68: Effect of gas dispersion in the reduction kinetics of ITCMO particles in presence of CH4...... 197 Figure 69: Pressure data and Reduction conversion obtained based on the original data and the data obtained by applying pressure correction. Reducing gas = CH4 with YCH4 = 50%, T = 950°C, and P = 8 atm...... 199 Figure 70: Oxidation conversion curves obtained from the thermogravimetric analysis between 1 and 10 atm at YO2 = 0.1, T = 900 ˚C ...... 201 Figure 71: XRD analysis of (a) reduced and (b) oxidized samples at 1 and 10 atm, 900°C.

Reducing environment is under H2 with YH2 = 0.5, and oxidizing environment is air... 203 Figure 72: SEM and EDS elemental mapping of cross sections of reduced particles.

Samples reduced under H2, YH2 = 0.5 and T= 900 ˚C. (a) 1 atm and (b) 10 atm ...... 205 Figure 73: Surface grains in samples reduced under H2, YH2 = 50% and T= 900 ˚C. (a) 1 atm and (b) 10 atm ...... 206 Figure 74: Simplified block flow diagram of chemical looping combustion...... 216 Figure 75: Simplified process schematic for the chemical looping gasification process using steam to produce H2 ...... 217 Figure 76: Thermodynamic equilibrium constants as a function of temperature for various oxidation states of Fe and Co based materials ...... 219 Figure 77: Effect of pressure on reaction rates for Fe samples, using at 900°C for (a) reduction with 50% H2 as reducing agent, and (b) oxidation with steam...... 223 Figure 78: The 20 cycle redox reactivity of Fe-based oxygen carriers at 900°C...... 225 Figure 79: Effect of pressure on reaction rates for Co samples, using at 900°C for (a) reduction with 50% H2 as reducing agent, and (b) oxidation with steam...... 228 Figure 80: The 20 cycle redox reactivity of Co-based oxygen carriers at 900°C ...... 229 Figure 81: SEM of the cross section of reduced Co sample at 5 atm, and 900°C. The EDS mapping clearly shows the presence of Co, Mg, Al, and O phases in a single microparticle tested here...... 230 Figure 82: Hydration of CaO sorbent, 50% steam, 3 atm, 400°C ...... 237 Figure 83: Hydration of CaO sorbent, 50% steam, 1.5 atm, 400°C ...... 237 Figure 84: Carbonation-calcination cycles performed in the TGA, effect of calcination environment on sintering...... 238

Figure 85: Effect of CO2 concentration on sintering of the limestone and dolomite sorbent...... 239 Figure 86: Modeling the decay in reactivity by the equation given by equation A1 for MV sorbent ...... 241 Figure 87: Carbon deposition observed at the coupling zone of the Rubotherm MSB .. 242

Figure 88: Simultaneous addition of ~30% CO and ~20% H2 to the TGA in presence of CaSO4 solids. Temperature maintained at 650 C and pressure of 5 atm. The outlet gas

xx concentrations are shown on primary y-axis and sample weight recorded by the MSB is shown on the secondary y-axis...... 243

Figure 89: Product XRD analysis after experiment with high purity (95%) H2, 650°C, 10 atm...... 244

Figure 90: ITCMO particles reduced under 50% CH4 at 900 °C, effect of pressure on reaction rates ...... 245

Figure 91: ITCMO particles reduced under 50% H2 at 900 °C, effect of pressure on reaction rates ...... 246

Figure 92: ITCMO particles reduced under constant partial pressure of H2 of 1.5 atm at 900 °C, effect of pressure on reaction rates ...... 246

Figure 93: The 20 redox cycles of Fe-based oxygen carriers supported on MgAl2O4, each series corresponds to a new day of testing. Compiled data ...... 250 Figure 94: Reoxidized Co-based sample, SEM image of the surface. Sample reoxidized at 5 atm using steam, 900°C ...... 252

Figure 95: EDS mapping of MgAl2O4 support used in steam oxidation experiments for chemical looping ...... 253

xxi

CHAPTER 1: Introduction

The field of chemical looping research has made great strides in the recent years, and emerged as one of the foremost leading technologies for the efficient fossil fuel conversion in a carbon constrained world energy scenario1. This research work has been conducted at the Ohio State University (OSU) in one of the topmost research centers, recognized as an authority on various aspects and applications of the chemical looping processes. In that capacity, this work deals with reactivity studies for the solid metal oxides for various applications of the type I and type II chemical looping systems.

1.1 Type I and type II chemical looping systems

The chemical looping systems can be broadly classified as type I and type II looping, according to the applications and the type of solids circulating in the system.2,3 The schematics of the two process concepts are seen in Figure 1. The type I chemical looping consists of a metal-metal oxide solid loop circulating between two (or more) reactors, and this involves the transport of oxygen between reactor blocks via the solid metal oxides.

Type I chemical looping is an example of oxycombustion of fuel, i.e. a controlled formation of carbon dioxide (CO2) and other combustion products via fuel oxidation by the solid metal oxides. This results in the formation of an inherently separate and pure

CO2 stream, in applications involving complete combustion. On the other hand, the type

II chemical looping consists of a solid loop of metal oxides-metal carbonate circulating

1 between the two or more reactors, and this involves the transport of CO2 between the reactor blocks via solid metal carbonates. Thus, the Type II chemical looping is an example of pre or post-combustion CO2 capture, i.e., the separation of formed CO2.

Figure 1: Concept of type I and type II chemical looping, for post combustion scenario

1.2 Major demonstration plants

The individual chapters of the present study pertain to specific applications of the type I and type II chemical looping, and each chapter contains a review of the literature relevant to the application being discussed. This section outlines the major demonstration efforts of these two chemical looping concepts throughout the world.

Both type I and type II chemical looping systems have made significant technological advances in the recent years. The type I technology was developed for various applications such as chemical looping gasification, chemical looping combustion, and partial oxidation. These processes were developed at OSU and demonstrated on various lab scale, bench scale and subpilot scale units.4,5 The 25 kWth subpilot demonstration units were constructed and tested for the syngas chemical looping (SCL) for testing gaseous fuels6,7 and coal direct chemical looping (CDCL) for conversion of solid fuels8,9,10. One particular application of the SCL process is currently under construction for the 250 kWth pilot scale demonstration at the National Carbon Capture Center

(NCCC) in Wilsonville, Alabama, to be operated under pressurized syngas combustion, using OSU’s iron-titanium complex metal oxide (ITCMO) particles as the oxygen carriers.5 The lab scale studies reported on in chapters 6 and 7 are in support of this effort.

Elsewhere in the world also, the chemical looping processes are being demonstrated over large scales. Recently, a 10 MW pilot plant has been designed by the Cenovus Energy

Inc., in association with Vienna University of Technology (TUV), and is proposed to be constructed in Canada by the year 2020 for CLC application for the production of

3 electricity.11 The proposed plant is a CLC steam generator, for combustion of natural gas that is used for Bitumen recovery from oil sands. Coupling of CLC and biomass gasification plants has also been tested at the TUV for the co-production of syngas and electricity in 120 and 100 kWth dual fluidized bed systems.12 The Technical University of

Darmstadt (TUD) has a 1 MWth unit operating with an interconnected circulating fluidized bed CFB) system using the iron-titanium ore ilmenite as the oxygen carrier materials for coal conversion.13,14 Alstom, U.S. has demonstrated the type I chemical looping concept on a 3 MWth pilot plant for the conversion of coal using CaSO4 oxygen carriers.15

Type II chemical looping uses solid metal oxides that can easily react reversibly with

CO2 to form metal carbonates. Calcium oxide (CaO) is the main metal oxide that is researched in type II chemical looping, and it is also renamed calcium looping. Since the concept was first proposed using twin fluidized bed reactors by Shimizu et al,16 it has been widely researched for both pre and post combustion applications.17,18 The majority of the present study is focused on various applications of this system.

The three-step calcium looping technology was pioneered at the OSU, through

19 20 21 synthetically derived and limestone-based calcium sorbents. >90% CO2 capture and

100% SO2 capture from coal-combustion flue gas has been demonstrated by the highly

22 reactive calcium hydroxide (Ca(OH)2) sorbent at a 120 kWth sub-pilot scale. This technology has been licensed to Industrial Technology Research Institute (ITRI), and will be demonstrated on a 1.9 MWth scale in Taiwan for a plant integrated with the cement industry.23 The pre-combustion application of the calcium looping technology was also

4 demonstrated at OSU for H2 production from simulated coal-derived syngas on a 25 kWth scale.24 The calcium looping system has also been demonstrated on large scales worldwide. The La-Pereda power plant by HUNOSA company in Spain has been supplemented with a 1.7 MWth pilot plant designed to remove 70-90% CO2 under the

CaOling project using the calcium looping process.25 At TUD at Darmstadt, Germany, a dual circulating fluidized bed reactor system is used for a 1 MWth demonstration of the calcium looping process.26

1.3 Outline of chapters

Due to the rising greenhouse gas concentrations, such as CO2, in the atmosphere; major research activities have been currently focused in the field of carbon capture from point sources. Calcium looping, which uses calcium oxide (CaO) sorbent derived from naturally occurring sources such as limestone, is a viable high temperature technology that is in advanced stages of development. The CO2 capture from power generation is an energy intensive step and therefore the competent CO2 capture technology is the one which is economical, and incurs minimum energy production penalty. One way of achieving this is a method which uses inexpensive and abundant natural resources.

The CaO sorbent derived from naturally occurring limestone loses its reactivity towards

CO2 over multiple high temperature cycles through a process called sintering. This decay in reactivity of the sorbent has been a major focus area in this field of research in recent years. The three step calcium looping process developed at OSU proposes hydration of sintered CaO as the third step in the solids loop to reactivate the sorbent to overcome this decay. Chapter 2 dwells on the significant development in this process block achieved

5 over the course of 3-4 years, from proof of concept experiments to bench scale demonstration. Lab scale fixed bed experiments showed the feasibility of high temperature steam hydration to effectively reactivate the CaO sorbent. The process concept was further investigated in semi-batch and continuous modes in fluidized bed reactors. The fluidization behavior of steam and CaO system was investigated, and fast fluidized bed was identified as the best design for operation of the hydrator reactor.

In Chapter 3, the CaO sorbent obtained from various waste egg and sea shells is demonstrated for its successful CO2 capture properties. Five different animal shells were selected, which are common waste of the food industry. All shells were also sintered at high temperature (950 °C) and subsequently regenerated using hydration. The inherent differences in the capture capacities of the different shells are further investigated using different analysis techniques such as bulk surface property analysis, X-ray diffraction and scanning electron microscopy. The successful achievement of CaO sorbent from all shells gives us a viable method of converting the waste material such as animal shells, which is a nuisance to the environment, into a useful resource as a sorbent for CO2 capture.

A novel process – termed as the Cal-C process – is conceived in Chapter 4 for the simultaneous removal of NO, SO2 and CO2 from coal-combustion flue gas. The Cal-C process consists of three reactors, while the gas cleanup reactions occur in a single gas- solid reactor with CaO sorbent and coal chars. The char is consumed in the process of

NO reduction and CaO is converted into calcium carbonate (CaCO3) and calcium sulfate

(CaSO4). The calciner regenerates the calcium sorbent and consumes any unreacted char.

The third reactor, hydrator, reactivates the calcium sorbent to restore its reactivity.

Results from laboratory-scale tests are presented which verify the process concept. A detailed study of the NO-char reduction reaction was performed isothermally at 650 ̊C, and in presence of the calcium sorbent. Complete NO reduction was achieved for various process conditions tested. The process analysis was conducted using the ASPEN simulation software. This new process is capable of 90% CO2 and NO removal and 100%

SO2 removal in a single process step, thereby reducing post-combustion processing steps for coal-combustion power plants.

The work in Chapter 5 was undertaken to investigate the side-reactions involving sulfurous species formed in the application of the calcium looping process (CLP) to a coal-to-H2 plant with CO2 capture. The sulfur present in the coal is converted to gaseous species such as hydrogen sulfide (H2S) and carbonyl sulfide (COS) in the gasifier, where coal is converted to syngas (CO and H2). High purity H2 can be produced from this syngas by subjecting it to the water gas shift reaction (WGS). The CaO sorbent aids this equilibrium limited reaction by consuming the CO2 reaction product. The WGS and carbonation reaction is carried out in a single reactor in the CLP system. In addition to

CO2, the CaO sorbent also removes the sulfurous (and halide) species by fixing them in the solid form of calcium sulfide (CaS) in the same reactor. Thus, the CLP system causes the multi-gas removal, enabling the production of high purity H2 and results in process intensification by combining several unit operations in a single reactor.

The specific work in this chapter was undertaken to explore the fate of the solid sulfurous species formed in the carbonator reactor. The CaS is converted to calcium sulfate

(CaSO4) under oxidizing conditions in the calciner. The CaSO4 so formed is recirculated

7 to the carbonator with the solid sorbent. The possible reductive decomposition of CaSO4 under the carbonator operating conditions was probed and it was revealed to be stable and unreactive at reducing conditions of the carbonator tested, at 650°C and 10 atm. Thus, it is concluded that CaSO4, once formed, is the stable species at the CLP operating conditions, and the sulfur will exit the solids loop as such, along with the unreacted CaS, via the purge stream of solids. Possible recommendations are made towards the end of this chapter for the treatment of this unreacted CaS in the purge stream, including possible treatment locations; gas concentrations; operating temperatures; and modifications to existing process scheme.

In Chapters 6 and 7 of this study, experimental investigations of two specific applications of type I chemical looping processes are illustrated. For gas to liquid (GTL) type of applications, partial oxidation of methane (CH4) is a viable route for conversion of natural gas to valuable chemicals. The oxygen for this partial oxidation process can be supplied using solid oxygen carriers, which can be single or mixed metal oxides. A particular partial oxidation scheme for CH4 conversion is called chemical looping partial oxidation (CLPO). The CLPO reaction scheme consists of two reactors, using Fe-based oxygen carrier particles which circulate within the two units and undergo cyclic reduction-oxidation (redox) reactions. The solid carriers therefore serve as a vehicle for oxygen between the units, enabling clean conversion of the fossil fuel with high purity product streams generated. Unlike the conventional combustion and/or gasification, the gaseous products of the two reactors are inherently separated. This allows minimization

8 of downstream processing and gas separation, making it a highly efficient energy conversion process.

For applications involving high pressure downstream processing (such as producing syngas as intermediate feedstock for liquid fuels and chemical synthesis), it may be advantageous to operate this gas-solid CLPO system at elevated pressures. Thus, it is desirable to study the effect of pressure on the reaction kinetics of the various reactions involved. Therefore in Chapter 6, the high pressure experiments were conducted for reduction and oxidation of oxygen carrier particles in a specialized thermogravimetric analyzer (TGA). The relative reactions rates were computed for all experiments conducted in the range of 1-10 atm. Specifically, the rate of reduction under H2 was observed to double when pressure was increased from 1 to 10 atm, compared to a fivefold increase in reduction rate under CH4. By comparison, oxidation reaction rate under air was observed to increase by ~50%. The reduced and oxidized samples were analyzed using SEM, XRD and BET techniques to determine the possible role of pressure in producing a more reactive particle, which explains the superior reaction rates observed at elevated pressures.

Similar to CLPO, which produces syngas, the versatile Fe-based oxygen carrier materials can be used in another application called chemical looping gasification (CLG) to co- produce high purity H2 and electricity. In this configuration, the chemical looping reaction scheme is split over three reactors. The reducer is used to completely oxidize a carbonaceous fuel, and the reduced metal oxides produced as a result are oxidized sequentially in two reactors. In the first reactor, called the oxidizer, steam (H2O) is used

9 to partially re-oxidize the reduced metal oxides and produce H2 as a result. The solids are further fully oxidized in the combustor, where the high exothermic heat of the oxidation reactions is harnessed to produce electricity. This process concept has been previously verified as the syngas chemical looping (SCL) process at OSU at the 25kWth scale, with

>99.99% purity H2 produced during steady state operation.

The Fe-based oxygen carriers used in the SCL process are typically thermogravimetrically tested for their recyclability under redox conditions, under conventional air oxidation. Chapter 7 reports on TGA tests conducted to verify the reactivity and recyclability of the Fe-based oxygen carriers under redox conditions involving steam. The reactive Fe2O3 supported on MgAl2O4 was subjected to up to 20 redox cycles in the TGA. Identical test conditions were also repeated on Co3O4 supported on MgAl2O4 for comparison. Unlike Co-based samples, which showed sluggish reaction rates and loss in reactivity over cycles, the Fe-based oxygen carriers exhibited excellent reactivity and recyclability over the 20 cycles tested. The steam oxidation resulted in

Fe3O4 formation after every cycle, verifying the stability and suitability of Fe-based materials for CLG applications.

CHAPTER 2: Steam Hydration as Reactivation for the Calcium

Sorbent

2.1 Introduction

The calcium oxide (CaO) sorbent is a high temperature sorbent for the capture of carbon dioxide (CO2) for pre-and post-combustion applications. Several unique characteristics of the calcium sorbent make it highly amenable to CO2 capture from large point sources such as coal combustion power plants.

1. It is derived from naturally occurring limestone, making it a highly inexpensive

sorbent.

2. It is an environmentally benign sorbent which is available in large quantities.

3. The sorbent material is robust towards other impurities that may be present in the

syngas or flue gas, such as fly-ash, halide, sulfur, and heavy metal impurities.

4. Unlike other physical and chemical CO2 scrubbing techniques which use lower

temperature solvents/sorbents, the calcium looping process is a high temperature

process, which allows high temperature heat recovery and therefore reduces the

energy penalty of the CO2 separation step.

Traditionally, the CO2 removal by calcium sorbent is achieved by employing the reversible reaction between CaO and CO2 in a closed solids loop, in an aptly named

11 calcium-looping technology. CaO reacts with the dilute CO2 present in the gas mixture

(~10-15% in coal combustion flue gas) and thereby fixes the CO2 in the solid calcium carbonate (CaCO3) form in a high temperature reactor (carbonator, 600-700°C). This

CaCO3 is separated from the gas mixture by any high temperature gas-solid separation technique, and transported to second reactor (calciner, 900°C). Here, the high temperature endothermic decomposition of CaCO3 takes place to release CO2 in a concentrated form. This concentrated stream of CO2 may then be used for other applications or sequestered. The regenerated CaO sorbent is then recycled back to the carbonator, closing the solids loop.

In this two-step process, the CaO sorbent loses its reactivity towards CO2 over multiple cycles. The high temperature thermal decomposition (calcination) reaction causes sintering of the CaO particles, resulting in loss of surface morphological characteristics.

Thus, over an extended number of cycles, a large excess of CaO sorbent is required to achieve the same amount of CO2 removal. This results in large solid circulation rates and/or large purge and makeup rates to achieve the desired level of CO2 removal.

The traditional calcium loop requires repeated cycling of the CaO sorbent through high temperature calcination, resulting in rapid degeneration of the sorbent reactivity towards

CO2 due to sintering. Over the last decade or so, various methods have been researched, developed and proposed to overcome this decay in sorbent reactivity.27,28,29 Some of these methods include structurally engineered sorbents to improve their mechanical strength30, doping the sorbents with chemical additives to resist sintering31,32, use of dolomitic sorbents (which results in more robust sorbents which resist sintering, but at the expense

33 34 35 of lower CO2 carrying capacity) , thermal pre-treatment , recarbonation etc. One such method of overcoming the loss in reactivity is the regeneration of the calcined sorbent through its conversion to Ca(OH)2. It is a well-established fact that the conversion of

22 CaO to Ca(OH)2 leads to increased reactivity of the sorbent towards CO2. Thus, various options exist for applying hydration as a sorbent reactivation method, such as: sorbent pre-treatment with water/steam, and a regeneration step of hydration after a specific number of cycles, or in-line partial/complete hydration of spent sorbent stream every cycle.36,37 The last option necessitates the inclusion of hydration as a separate unit of the calcium looping process. Not only that, in order for hydration to become a viable third step of this high temperature cyclic process, it is imperative to operate this step at appropriately high temperatures. This concept forms the basis of the development of

Ohio State’s 3-step calcium looping technology.

In this envisioned process, the hydration is a separate step between the high temperature thermal calcination and carbonation. In this step, the CaO is hydrated to calcium hydroxide (Ca(OH)2), which shows superior reactivity towards CO2 as compared to CaO sorbent. Due to this high reactivity of Ca(OH)2 towards CO2, the solid to gas ratios as well as the overall solid circulation rates, and makeup and purge rates of the system are reduced greatly. This fact was first demonstrated at the 120KWth scale using an actual

22 coal combustion flue gas for post combustion CO2 capture. Here, using Ca(OH)2 instead of CaO, near stoichiometric gas: solid ratios were found to be sufficient to achieve ~90% CO2 removal and also 100% SO2 removal in a single reactor in a matter of seconds. Also, on this scale, the effectivity of hydration as a reactivation treatment for the

13 multicyclic use of CaO sorbent was tested for the first time. Offline hydration was performed after each cycle on a commercial lime hydration unit and the reactivity of this re-activated solid was found to be maintained over the 5 cycles tested.22

However, this offline hydration was performed at ambient conditions using liquid water as hydration medium. In order to successfully incorporate hydration as a third step in the solids loop in the calcium looping technology, it must necessarily be conducted at appropriately high temperatures, in order to minimize temperature swing and associated energy losses. Therefore, high temperature steam hydration was identified as a key area of research for the feasibility of this process.

0.1 0.01 0.001 0.0001 0.00001 0.000001 0.0000001

Steam partial pressure, atm partialpressure, Steam 1E-08 100 200 300 400 500 Temperature, C

Figure 2: The steam pressure as a function of temperature for the hydration reaction.

The steam hydration reaction is given below:

CaO + H2O(g) → Ca(OH)2 Rxn 1

The equilibrium relationship between the partial pressure of steam and temperature is shown in Figure 2. At a constant temperature, for steam pressures above the equilibrium curve, the forward reaction is favored and the solid product is Ca(OH)2. For steam pressures below the equilibrium curve at the same temperature, the backward reaction is favored and the stable product is CaO. In pure steam environments, at 512°C, the equilibrium steam pressure is 1 atm. Thus, this places an upper boundary for the operating temperature of the steam hydration reactor, as operation above 512°C will require elevated pressures to carry out the hydration reaction. According to process integration simulations performed by Wang et.al., a minimum hydrator operating temperature of 350°C is required for the effective heat recovery of the exothermic heat of hydration.38 Also, with the hydrator operation at 500°C, the drop in electricity generation efficiency is predicted to be around 20-22%, as compared to 25% using oxycombustion and 27% using amine scrubbing. Therefore temperature range of 350-512°C is thus chosen for the hydrator reactor to successfully integrate into the solid sorbent loop of the calcium looping technology.

The Figure 3 shows the equilibrium pressure curve in the temperature range of interest. It can be appreciated that the equilibrium pressure rises exponentially in the narrow operating window of 350 to 500°C. Further, the hydration reaction rate is proportional to the difference between operating steam pressure and the equilibrium pressure, i.e. R α

* n (PH2O – PH2O ) . Therefore, the challenge of the successful operation of hydration reactor is to achieve reaction rates comparable to that of carbonation and calcination, while attaining as high temperatures of operation as possible.

0.9 0.8 0.7 0.6 0.5 Ca(OH) 0.4 2 0.3 0.2 CaO

0.1 Steam partial pressure, atm partialpressure, Steam 0 300 350 400 450 500 Temperature, C

Figure 3: Exponential increase in the equilibrium reaction pressure with temperature, in the temperature range of interest.

Thus, the feasibility of high temperature steam hydration was studied in the present work, using lab scale fixed bed testing. Hydration was carried out using steam at 500°C on different sorbents. The sorbents were derived from various limestones as precursors, and were subjected to up to 5 cycles of carbonation, calcination and hydration. The reactivity of the sorbents upon hydration was found to be restored after every cycle, with negligible decay in reactivity observed over the five cycles tested. One of the limestones was further selected to carry out extended number of cycles (over 15 cycles) and its reactivity was compared to the traditional two step cycles which were also conducted on the same sorbent. The hydration resulted in superior sorbent morphology, with the pore size distribution shifting towards mesopores formation upon hydration.

2.2 Feasibility of steam hydration Reproduced in part with permission from Phalak, N.; Deshpande, N.; Fan, L.-S.

Investigation of High-Temperature Steam Hydration of Naturally Derived Calcium Oxide for Improved Carbon Dioxide Capture Capacity over Multiple Cycles. Energy Fuels

2012, 26 (6), 3903–3909. Copyright [2012] American Chemical Society.

2.2.1 Experimental materials and methods

Six different limestones were procured from various regions of the US Midwest. The three limestones obtained from Graymont Inc. were Pleasant Gap, PA (PG), Genoa, OH

(GN) and Townsend, MT (TS). The three limestones obtained from Carmeuse Lime and

Stone were Maysville, KY (MV), Blackriver, KY (BR) and Toledo, OH (TL). The GN and TL samples were dolomitic in origin, while the rest were calcitic. The composition and initial CO2 capture capacity of these limestone sorbents is shown in Table 1. The

17 limestones were crushed and sieved to the desired particle size of 250-300μ. The CaO sorbent was then obtained from the crushed limestone sample by calcining in a muffle furnace at 950°C for 2 hours in air environment. The CaO sorbent thus produced was hydrated using a fixed bed reactor setup at 500°C for 30 minutes, using 90% steam. After hydration, the sample particle size was found to be reduced to less than 20μ, and no further effort was expended on controlling particle size of the samples. The fixed bed reactor setup is shown in Figure 4. The carbonation was conducted in the same fixed bed reactor at 650°C using 12% CO2 for 2 hours.

Table 1: Properties for sorbents generated from different limestones

Limestone Initial Wt % CO2 capacity Inerts (% by wt) Toledo 37.93 15.49 Genoa 40.27 12.25 Maysville 47.98 4.07 Blackriver 52.63 4.26 Montana 54.08 5.28 Pleasant Gap 56.46 3.44

The desired gas flowrates were obtained by using a battery of mass flow controllers and a gas mixing panel. The gas was preheated in an electrically heated zone. Water was injected in the preheater section using a high precision syringe pump (ISCO 100DM).

This preheater zone was filled with quartz-wool to increase the contact area between the water and the heated gas, thereby producing steam in situ. The steam-gas mixture was injected from the top of the fixed bed reactor, which was encased in an electrically heated furnace. The solids bed was contained in the fixed bed reactor. Downstream of the fixed bed reactor, the gas passed through a condenser before being vented to the atmosphere.

Figure 4: Fixed bed reactor setup for the high temperature steam hydration. Figure from 39

The samples were tested for their carbonate, hydrate, and oxide content after each step in the cycle using a thermogravimetric analyzer (TGA) to decompose the sample under inert

N2 environment. After each hydration step, the sorbents’ reactivity towards CO2 was also tested using the same. The sorbent reactivity is quantified by means of “wt % CO2 capture”, which is defined as

푤푒𝑖푔ℎ푡 표푓 퐶푂 푐푎푝푡푢푟푒푑 푊푒𝑖푔ℎ푡 % 퐶푂 푐푎푝푡푢푟푒 = 2 ∗ 100 2 푤푒𝑖푔ℎ푡 표푓 퐶푎푂 푠표푟푏푒푛푡

The physical morphological changes in the sample were studied using qualifying and quantifying techniques such as scanning electron microscopy (SEM) and N2 physisorption analysis respectively.

The weight of inerts in the original sample from the TGA analysis is defined as weight of all species excluding CaCO3.This weight is assumed to be constant, i.e. unreactive and inert with respect to temperature. This assumption is an approximation, since the limestone contains other species which, though being non-reactive toward CO2 capture at high temperatures of operation, yet undergo decomposition to release gaseous species during the initial TGA analysis.

For the calcitic stones, this assumption is valid since the fraction of inerts present in the original sorbent is very low, and the weight loss upon heating is not appreciable with the analysis instrument available, viz. TGA. However, the dolomitic sorbents contain appreciable amounts of other species which are not fully inert (they undergo perceivable weight loss due to decomposition upon being heated in an inert atmosphere, which can be distinguished from the weight loss of CO2 from CaCO3). Therefore, this assumption

22 cannot be extended to the dolomitic sorbents to calculate the active fraction of dehydrated sorbent in TGA, for comparing the sorbents on calcium and hydrate basis.

Thus, the dolomitic sorbents are excluded from the following analysis.

Calculation of calcium basis:

퐶푂 푐푎푝푡푢푟푒푑 (푔푚) 푤푡% 퐶푂 푐푎푝푡푢푟푒 표푛 푐푎푙푐𝑖푢푚 푏푎푠𝑖푠 = 2 × 100 2 푤푒𝑖푔ℎ푡 표푓 푎푐푡𝑖푣푒 푠표푟푏푒푛푡 (푔푚)

Active weight is defined as weight of the sample present in the form of CaO and therefore capable of capturing CO2. It is calculated as follows:

푊푎푐푡𝑖푣푒 = 푊푐푎푙푐𝑖푛푒푑 × (1 – 푓𝑖)

푤𝑖 푓𝑖 = 푤퐶푎푂 + 푤𝑖

56 where, 푤 = 푓 × 푊 × 퐶푎푂 퐶푎퐶푂3 표 100

and 푤𝑖 = (1 − 푓퐶푎퐶푂3) × 푊표

Wactive = weight of active sorbent (gm)

WCalcinced = weight of sample in TGA after dehydration and before carbonation, in CaO

fi = fraction of inerts in weight of dehydrated sample in TGA

wi = weight of inerts in original sample Wo, in TGA

wCaO = weight of CaO in dehydrated sample in TGA

Calculation of hydrate basis:

Not all of the sorbent is hydrated during the steam hydration step. Also, the degree of hydration is believed to be different for different sorbents for fixed duration of steam hydration. Therefore, the weight % CO2 capture of all sorbents is compared on the basis of the actual hydrate present in each sorbent during each cycle.

To this end, estimation of the degree of hydration was required. For the purpose of the following calculations, it is assumed that weight of H2O evolved during dehydration prior to carbonation in TGA is indicative of the degree of hydration of sorbent.

푊퐶푂2,ℎ푦푑푟푎푡푒푑 푠표푟푏푒푛푡 푥ℎ푦푑푟푎푡푒 푏푎푠𝑖푠 = × 100 푊ℎ푦푑푟푎푡푒푑

where 푊퐶푂2,ℎ푦푑푟푎푡푒푑 푠표푟푏푒푛푡 = 푊퐶푂2,푡표푡푎푙– 푊퐶푂2,푢푛ℎ푦푑푟푎푡푒푑 푠표푟푏푒푛푡

푊 푊 = 퐻2푂 × 56 ℎ푦푑푟푎푡푒푑 18

푊푢푛ℎ푦푑푟푎푡푒푑 = 푊푎푐푡𝑖푣푒– 푊ℎ푦푑푟푎푡푒푑

푥 푊 = 푐푎푙푐푖푛푒푑 × 푊 퐶푂2,푢푛ℎℎ푦푑푟푎푡푒푑 푠표푟푏푒푛푡 100 푢푛ℎ푦푑푟푎푡푒푑

푊퐶푂2,푡표푡푎푙= total weight of CO2 captured

푊퐶푂2,푢푛ℎ푦푑푟푎푡푒푑 푠표푟푏푒푛푡= weight of CO2 captured by unhydrated sorbent

푊ℎ푦푑푟푎푡푒푑= weight of hydrated sorbent, in the form of CaO

푊푎푐푡𝑖푣푒= weight of active sorbent, in the form of CaO

푊푢푛ℎ푦푑푟푎푡푒푑= weight of unhydrated sorbent, in the form of CaO

푥푐푎푙푐𝑖푛푒푑= weight % capture for calcined sorbent in the corresponding cycle

2.2.2 Recyclablility of the sorbents over steam hydration

All the sorbents tested showed excellent recyclability upon being subjected to the five cycles of carbonation, calcination and hydration. The sorbents showed negligible loss in reactivity over five cycles, thus showing the effectiveness of steam hydration as a reactivation treatment of CaO sorbent. The following Figure 5 shows the five-cycle performance of the six sorbents tested here. It was also observed that the sorbents derived from different limestone precursors exhibited different reactivity towards CO2 despite being treated at identical reaction conditions. However, this disparity in the reactivities may be explained in part by the different amounts of inerts present in the sorbent. When the weight % CO2 capture is considered on the basis of the active calcium content of the sorbent, the difference in reactivity is reduced. Further, the sorbents are also observed to hydrate to dissimilar extents despite being subjected to identical hydration reaction conditions. The gap in the measured reactivity is further diminished when the CO2 capture is considered only on the fraction of the sorbent that is hydrated (hydrate basis in

Figure 6).

40 Cycle 1 Cycle 2 30 Cycle 3

% Capture % Cycle 4 20 Cycle 5 10

0 TL GN MV BR TS PG

Figure 5: The post-hydration capture capacity of the CaO sorbents derived from the six different limestones over five hydration cycles

average capture capacity hydrate basis calcium basis 70

30 % Capture % 20

0 MV BR TS PG Sample

Figure 6: Comparison of the post-hydration CO2 capture capacities of the calcitic sorbents, on calcium and hydrate basis.

Additionally, two dolomite limestone samples (TL and GN) were also procured and the

CaO/MgO sorbent generated from these precursors was also subjected to hydration reaction. The steam hydration conditions tested here are amenable to hydrate only CaO sorbent, and the MgO part of the sorbent remains in oxide form throughout the hydration step. The reactivity of these sorbents was also observed to be maintained over five cycles tested here, as seen in Figure 5.

2.2.3 Change in sorbent morphology

Of particular interest were the sorbents obtained from MV and BR limestones, since these samples exhibited markedly different reactivity towards CO2 despite having almost identical active calcium content (CaCO3 content of ~ 95.93% and ~95.74% for MV and

BR samples respectively). Therefore, the difference in their reactivity towards CO2 was further investigated by observing the changes in surface morphology of the two samples.

It was observed that the extent of hydration in the two samples was different in spite of the similar calcium content, and on the hydrate basis alone their CO2 capture capacity was almost identical. Furthermore, the N2 physisorption analysis was used to compute the total pore volume of the samples, before and after hydration. It is a well-established fact that hydration causes an increase in surface area and pore volume of the CaO sorbent.

The increase in total pore volume post hydration was 47% in the BR sample, as opposed to a 10% increase in MV sample. Furthermore, pore size distribution is seen to shift towards larger sized pores after hydration. This shift is also more prominent in the case of

BR sample (see Figure 7). The sorbents after calcination are studied using SEM analysis in Figure 8 and Figure 9. The calcined MV sample clearly shows more sintering than the

28 corresponding BR sample, after 1 as well as 5 cycles. This indicates towards the lower reactivity of the MV sample observed. The MV and BR samples are compared for their morphological changes before and after hydration. After hydration, the ‘popcorn effect’ is also observed in both samples (Figure 10 and Figure 11). All of these observations serve to explain the difference in reactivity of the two samples towards hydration, and therefore subsequently, carbonation.

Figure 7: Particle size distribution post calcination and post hydration for (a) MV and (b) sorbent

Figure 8: Calcined MV sorbent (a) cycle 1 and (b) cycle 5

Figure 9: Calcined BR sorbent, (1) cycle 1 and (2) cycle 5

Figure 10: BR sorbent cycle 5 (a)calcined and (b) steam hydrated followed by dehydration in TGA

Figure 11: MV sorbent cycle 5 (a) calcined and (b) steam hydrated followed by dehydration in TGA

2.2.4 Extended number of cycles

One of the limestones, Pleasant Gap (PG) was identified as producing the best performing CaO sorbent among the several samples tested. Therefore, it was chosen for testing its reactivity over an extended number of carbonation-calcination-hydration cycles. The PG sorbent was subjected to two types of cycles:

 Three-step cycles: these cycles consist of carbonation at 650°C followed by

calcination at 950°C in the muffle furnace, and hydration at 500°C in the fixed

bed reactor. The samples were tested for their reactivity towards CO2 after every

calcination and hydration.

 Conventional, two-step cycles: the two step cycles consist of alternating

carbonation at 650°C in the fixed bed reactor and calcination at 950°C in the

muffle furnace. The CO2 capture capacity of the sorbent was measured after

every calcination.

The three step process was carried out for 15 cycles, while the traditional two-step process was carried out for 10 cycles. The results of these extended cycles are shown in

Figure 12.

It can be appreciated that for the conventional two-step cycles, the reactivity of the sorbent drops drastically over the 10 cycles conducted (red bars), whereas, the post- hydration reactivity of the sorbent undergoing the three-step cycle suffers negligible loss in reactivity over the 15 cycles tested. Moreover, the samples from the three-step cycle were also tested for their post-calcination capture capacity in each cycle, and it was

32 observed that even this capture capacity, though less than half that of post-hydration, remains constant over the 15 cycles tested here.

CaCO3-CaO cycle

Capture

2 30 CaCO3-CaO-Ca(OH)2 cycle, calcined 20 Wt% CO Wt% CaCO3-CaO-Ca(OH)2 cycle, hydrated 10

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 cycle

Figure 12: Extended number of cycles showing the effectiveness of hydration as reactivation mechanism.

The extent of hydration, or hydration conversion, in each cycle was ascertained through

TGA decomposition of the sample under inert N2 conditions. Under 90% steam conditions at 500°C in a fixed bed reactor setup, the hydration conversion is found to be varied between 80-90% for the 15 cycles tested, as seen in Figure 13. Thus, the reactivity of the sorbent is observed to decay over multiple cycles (from 90% at cycle 1 to 80% at cycle 15), however, the effect is not as pronounced as that of carbonation, thus enabling the sorbent’s reactivity towards CO2 to be maintained above 45% for the 15 cycles tested.

0.8

0.6

0.4

0.2 Hydration conversion Hydration 0 0 5 10 15 cycle number

Figure 13: Hydration conversion as a function of number of cycles.

Thus, the feasibility of the steam hydration of CaO sorbent was established as a reactivation mechanism. It was also concluded that the reactivity of said reactivated sorbent towards CO2 may be maintained over multiple cycles.

Next, the need for a functional hydrator reactor design was identified for the process integration of the hydration reaction. This hydrator reactor must achieve high conversions of solids while minimizing the reaction time and excess steam requirements. On the commercial scale, CaO hydration is carried out at ambient conditions using water in the cement industry. High pressure steam hydrators designs are patented, primarily for the steam hydration of dolomite samples. However, operational data on said hydrators is rare.

Therefore, through these feasibility tests, the need for design and operational results of a high temperature steam hydrator was identified.

2.3 Bench scale semi-batch fluidized bed hydrator

The fact that hydration maintains reactivity of sorbent towards CO2 for multiple cycles has thus been established. This highly reactive Ca(OH)2 sorbent enables the use of near

22 stoichiometric Ca/C ratios for ~90% CO2 capture, as indicated by reference .

Accordingly, a semi-batch fluidized bed reactor (FBR) was designed for the high temperature steam hydration reaction.40 Due to the cohesive nature of the particles, several static mixing elements were installed, such as wall and embedded baffles, as well as a rotating impeller at the center of the solid bed. The fluidization behavior of the sorbent was methodically investigated. The calcined CaO solids were fed to this reactor in a batch manner and steam-air mixture was passed through this solid bed and the solids were sampled from the reactor at specific intervals to determine their hydration conversion. This semi-batch FBR has resulted in about ~70% hydration conversion using a steam to calcium ratio of 1.3 achieved in 30 minutes. The results are shown in Figure

14.

Figure 14: Hydration performance of the semi-batch fluidized bed hydrator40

One of the key outcomes of these FBR experiments was the recognition of the need for effective heat removal from the reactor. As can be seen from Figure 14, due to the highly exothermic nature of the CaO hydration reaction, isothermal experimentation is not feasible in experiments dealing with a large batch of solids. The semi-batch nature of the operation of this FBR resulted in a considerable increase of the temperature of solids as the reaction progressed. Operating this FBR as the hydrator in a fully integrated solid loop with a calciner and carbonator will enable isothermal operation by continuous gas- solid feeding. A continuous system will enable further study of parameters such as steam to calcium ratio, residence time of solids, and design of efficient heat recovery system.

The hydrator fluidization behavior may thus be improved further; to result in higher hydration conversion, reduced residence time of the solids in the reactor, and reduced excess steam requirement.

2.4 Continuous hydration on entrained bed reactor

2.4.1 Experimental setup

The sub-pilot scale fluidized bed reactor situated at west campus high-bay lab facilities was used to carry out the continuous high temperature steam hydration experiments. The schematic of this reactor setup as modified for steam injection is shown in Figure 15.

This reactor is a riser, which consists of four flanged sections of 2.25 ft. height. The ID of this riser is 4 in. Each flanged section is supplied with two temperature ports for thermocouples and one pressure port. An inclined solids feeding section is attached to the bottom flanged section. The solids are fed through a solids hopper by means of a screw

39 feeder and two gate valves which function as double-dump valves by alternate opening and closing mechanisms. At the base of the flanged section, a porous sintered plate functions as the gas distributor for the fluidization gas and also supports the bed of solids that may accumulate through the solid feeding mechanism. Sufficiently large velocities may be chosen such that majority of the solids fed to the reactor are entrained out through the top of the riser section. The solids exiting the top of the riser can be captured and tested for the solids hydration conversion. More details of this riser reactor are given in reference 24.

The reactor is heated by means of electrical heaters maintained at isothermal conditions through a temperature control program using DaQ-Factory software. Steam is generated by means of two stainless steel coils which are electrically heated and water is fed to these coils via metered pumps calibrated for specific volumetric flow. Air is used as a carrier gas to provide motive force for the steam so produced.

Figure 15: Schematic of the sub-pilot scale riser reactor used for continuous steam hydration experiments.

2.4.2 Cold/dry flow tests

Before conducting actual hydration experiments, certain cold and/or dry experiments were performed for determination of the operating velocities to achieve > 80% entrainment. These experiments were performed using air as the fluidization medium using mixtures of CaO and Ca(OH)2. Due to the highly cohesive nature of the Geldart C type particles of the sorbent, extremely high velocities of gas were needed to achieve appreciable entrainment.

The water pumps for steam generation, as well as the screw feeder for solids feeding, were calibrated at the time of the experiments. The screw feeder is a volumetric feeder and therefore, the actual feeding rate may vary depending on the packing density of the solids in the hopper. Thus, the solids fed to the hopper were measured prior to the experiment, and the solids remaining in the hopper after the experiment were weighed, to calculate the total solids fed and thus the solid flow rate during experiment. Very high velocities of up to 2 m/s were used to achieve >80% solid entrainment. The entrainment as a function of gas velocity is seen in Figure 16.

90 80 70 60 50 40 30

Entrainment, by wt% by Entrainment, 20 10 0 0 0.5 1 1.5 2 Velocity, m/s

Figure 16: Degree of entrainment obtained as function of gas velocity, cold flow experiments

2.4.3 High temperature experiments

During each experiment, the reactor was preheated using air to the operating temperature of 350°C. Isothermal experiments were carried out at 1 atm. Upon reaching the desired steady temperatures, steam and solids were injected into the reactor. The flowrates of both were calibrated before experiment and also quantified post-experiment. The entrained solids were sampled via a solid sampling port located at the top of the topmost flanged section. These entrained samples were analyzed and their hydrate content was determined using a TGA via thermal decomposition. In addition, after each experiment, solids samples were collected from the solids bed accumulated on the porous gas distributor plate, as witnessed by opening the reactor after cooling down. Thus for each test, hydrate content of two sets of samples was measured, one from the entrained solids, and the other from the bed collected on the gas distributor plate. The solid flow rates were maintained at 2.5 to 3 mols per min. Entrained solids were sampled near the top of the reactor, at the last heated section. The solid sampling probe (1/4” SS tube) contained a 1” slot along the length for capturing solid particles. Solids were sampled at 4 minute intervals and sealed for composition testing.

Due to the highly cohesive nature of sorbent, very high velocities were required to achieve appreciable degree of entrainment. The degree of entrainment was calculated by measuring the solids remaining over the porous plate after the experiment. The equivalent weight of reactant solids was then calculated based on the composition.

푒푞푢𝑖푣푎푙푒푛푡 푤푒𝑖푔ℎ푡 표푓 (푟푒푎푐푡푎푛푡) 푠표푙𝑖푑푠 표푣푒푟 푝표푟표푢푠 푝푙푎푡푒 Degree of entrainment = 푡표푡푎푙 푤푒𝑖푔ℎ푡 표푓 푟푒푎푐푡푎푛푡 푠표푙𝑖푑푠 푓푒푑

The superficial velocities of 1 to 1.9 m/s were required for achieving >80% entrainment of the fine cohesive solids. The excellent temperature control was achieved by operating dry flow experiments using air before testing with actual steam conditions. During steam experiments, the flowrates were limited by the maximum feed rate allowable by the water pumps. As a result, to obtain higher gas velocities, the supplemental air had to be increased, resulting in lowering of the steam partial pressure. The achievable steam partial pressures as a function of temperature are given in Figure 17. These cohesive solids also resulted in severe wall coating of the sorbent in absence of any internals in the reactor such as wall baffles. This is seen in Figure 18. These high velocities gave rise to a short residence time for the solids of the order of 2-3 seconds in the riser section. In this short contact time between steam and CaO sorbent, the entrained solids showed conversions between 5-15% for steam partial pressures between 0.2 to 0.77 at 350°C.

This can be seen in Figure 19.

1 0.9 0.8

0.7 thermodynamic 0.6 limit 0.5 1.5 m/s 0.4 1.2 m/s Steam PP, SteamPP, atm 0.3 0.2 1 m/s 0.1 1.8 m/s 0 350 400 450 500 temperature, C

Figure 17: Maximum steam partial pressures achievable as a function of temperature, at different gas velocities.

Scale: 1”

Figure 18: Wall coating observed in the entrained bed reactor

0.4 0.35

0.3 0.25 0.2 Entrained sample 0.15 bed sample

hydrationconversion 0.1 0.05 0 0 0.2 0.4 0.6 0.8 1 Partial pressure of steam, atm

Figure 19: Hydration conversions obtained in the entrained and bed samples as a function of steam partial pressures used

The conversion in the solid bed samples corresponding to the same experiments were proportionally higher, with up to 35% solids conversion to hydrate observed on the gas distributor plate. However, due to the continuous feeding and entrainment of the solids, it is difficult to state the exact reaction time for which the bed solids were contacted with steam.

Schaube et. al.41 have proposed an empirical reaction rate expression for the hydration of

CaO for two cases, case 1) T – Teq > 50 K (away from equilibrium) and case 2) T – Teq <

50 K (close to equilibrium). For all the conditions tested here, case 1 is valid. Therefore, the kinetic expression proposed by Schaube et. al. is:

푑푋 89465 푃 0.83 = 13495 ∗ exp (− ) ∗ ( – 1) × 3(1 − 푋) × ln(1 − 푋)0.66 푑푡 푅푇 푃푒푞

--- Eqn 1

Thus, this expression can be used to predict the conversions at the test conditions employed in this continuous steam hydration experiments. Such a comparison is shown in the following Figure 20. It can be appreciated that the trend in the predicted conversions is identical that observed in the experimental conversion values. Thus, the conversions obtained in entrained samples achieve performance close to that predicted by kinetic rate expressions.

Furthermore, Eqn 1 was used to predict the total reaction time required to achieve complete or near-complete conversions at the experimental conditions tested here (350°C,

0.2-0.8 atm of steam partial pressure). The predicted reaction conversion curves are given

49 below in Figure 21. The curves predict up to 3-4 minutes of residence time required for near complete reaction conversions of hydration.

0.3

0.25

0.2

0.15 Schaube et al. 0.1 experimental

hydration conversion hydration 0.05

0 1 2 3 4 5 6 Experiment number

Figure 20: Comparison of entrained bed performance and conversions predicted by Schaube et.al.

2 0.8 0.27 atm 0.6 0.45 atm 0.4 0.4 atm

0.2 0.54 atm conversion to Ca(OH) to conversion 0.77 atm 0 0 1 2 3 4 5 time, min

Figure 21: Predictions of hydration conversion at OSU’s test conditions using equation 1.

2.5 Conclusions

Thus, these experiments conclusively proved that using a fast fluidized bed, the mass transfer limitations for the high temperature steam hydration reaction may be easily overcome and near-kinetic performance may be achieved. However, the short residence times of 2-3 seconds provided by the entrained bed riser reactor employed here is not sufficient to achieve appreciable solids conversion for hydration reaction. Nevertheless, kinetic expressions proposed in the literature predict up to 3-4 minutes residence time required for near-complete hydration conversion. Thus, the hydrator reactor may be envisioned as a fast/turbulent fluidized bed reactor with a solids residence time on the order of a few minutes. Additionally, the intense solids wall-coating observed in these experiments confirmed the need for baffles and other internals for the uniform fluidization of these cohesive Ca(OH)2 powders.

CHAPTER 3: Screening of Multiple Waste Animal Shells as a Source of

Calcium Sorbent

3.1 Introduction

The rise in atmospheric carbon dioxide (CO2) levels in the last few centuries which was sparked by the industrial revolution, has spurred a plethora of research activities on the potential carbon capture and sequestration (CCS) technologies. The world energy market continues to grow due to the ever increasing demand, with the energy consumption estimated to grow from 5.3 x 1017 in 2008 to 8.1 x 1017 kJ by 2035. Fossil fuels continue to supply majority of the total energy required in the projected near future.42 The unmistakable rising trend in atmospheric CO2 levels is therefore a concern that demands urgent attention. Many CCS technologies are currently in various stages of development,

43,44 with the focus being on the separation of CO2 from the emission gases , and recent efforts are also focused on reutilization of the captured CO2 to convert it to energy products 45,46. The various capture techniques can be categorized as absorption and adsorption techniques, with either physical or reactive separation; membrane separation, oxycombustion, etc. Commercial or near-commercial technologies include using solvents such as amines (viz. monoethanol amine)47 and other specialty solvents such as Selexol, or the chilled ammonia process.48 The molecular sieves, physical adsorbents, low temperature chemisorbents, membrane separation technologies are some of the other

53 leading areas of research in the field of CO2 separation and capture. The high temperature reactive chemisorption using solid sorbents is of particular interest due to its relatively high CO2 carrying capacity, greater possibility of effective heat integration, comparatively lower energy penalty, etc1,49,17,50

The use of metal oxides, particularly calcium oxide (CaO), for dry CO2 removal has been theoretically proven for a few decades, and the first successful demonstration of calcium

51 oxide for CO2 capture was in the 1970s with the CO2 Acceptor Process. The underlying reactions for this process are

CaO + CO2 → CaCO3 Rxn 2

CaCO3 → CaO + CO2 Rxn 3

The exothermic forward reaction, carbonation, captures CO2 from the gas stream and converts the CaO sorbent into solid calcium carbonate (CaCO3). The subsequent endothermic high temperature thermal decomposition of CaCO3 releases the captured

CO2 and regenerates the CaO sorbent via the backward reaction. Thus, the CaO sorbent can be used over multiple cycles to achieve high temperature CO2 removal from point sources such as combustion flue gases emitted from power plants. The process has also been successfully demonstrated at lab and bench scale level for pre-combustion CO2 capture from synthesis gas for various applications.52,53,24

The availability of limestone as an inexpensive, abundant, and environmentally benign source of the sorbent makes this technology viable and economically attractive. Current research efforts in this area are focused on improving the degradation of the reactivity of

CaO over multiple cycles.1,50,29 Intermediate hydration of deactivated calcium sorbent has been shown to effectively restore the capture capacity of CaO sorbent.

CaCO3 is also an abundant biomaterial which is present in most animal shells. Poultry eggs and sea shells are a common feature of many cuisines around the globe, and therefore egg-shells and sea-shell form a major part of food industry waste. In many shore lined countries, sea-shell pollution is a grave environmental problem. According to the International Egg Commission, China is world’s largest egg producer, and about 53.4

Mt eggs were produced worldwide in 2002.54 In the United States alone, the poultry industry reports production of 6.54 x 109 eggs in April 2012.55 Thus these organic shells which are waste material can also be a potential source of CO2 capture, and research effort has also been focused on the use of these animal shell waste materials as a source

56,57 of calcium sorbent for high temperature CO2 capture. In this study, several different egg and sea shell waste materials were screened for their CO2 capture capacity at high temperatures.58 The effective regeneration of the sintered CaO sorbents was demonstrated using hydration. A few samples were selected to investigate the difference in inherent reactivity of the shells.

3.2 Experimental

3.2.1 Materials and methods

Five different animal shells were obtained from various sources. The shells are broadly classified as egg and sea-shells. Chicken eggs (CH-E) were obtained from conventional farm-raised eggs from grocery stores. The duck eggs (DU-E) were farm raised in the U.S. and purchased from food stores in the Columbus, Ohio. The ostrich egg sample (OS-E)

55 was obtained from a farm in Tehachapi, CA.59 The clam (CL-S) and oyster (OY-S) shells were similarly obtained from grocery stores (Figure 22).

Figure 22: Original shells from which samples were prepared.

The powdered sorbent was obtained from the shells by the following procedure: The samples from egg shells were prepared by pretreating the egg shells to a solution of 0.2

M acetic acid for 30 min. This step was carried out to separate the collagen and other proteinaceous material from the shell to obtain dry sorbent. After the wash with acetic acid, the samples were filtered and further treated with a water wash. The sea shell samples were cleaned of any meat by scraping and washing in soap solution followed by deionized water. The shells were then dried overnight and crushed and sieved until a required particle size was achieved (< 250 μm). Samples were stored in air-tight containers for testing. The samples were screened for their initial capture capacity, given in Table 2.

Table 2: Different animal shells tested and their sources, with initial CO2 capture capacities

Sorbent source Initial CO2 capture Acronym (type of shell) capacity (%) Chicken egg CH-E 60 Ostrich egg OS-E 45 Duck egg DU-E 61 Oyster shell OY-S 34 Clam shell CL-S 21

The wt% CO2 capture capacity of the sorbent, also referred to as sorbent reactivity, is

-1 calculated as weight of CO2 captured in g at the end of 30 min g sorbent, expressed in percentage value.

푤푒𝑖푔ℎ푡 표푓 퐶푂 푐푎푝푡푢푟푒푑 푤푡% 푐푎푝푡푢푟푒 푐푎푝푎푐𝑖푡푦 = 2 100% 푤푒𝑖푔ℎ푡 표푓 퐶푎푂 푠표푟푏푒푛푡

The characterization of all sorbent samples for their CO2 capture capacities was carried out using thermogravimetric analysis (TGA) equipment (Perkin Elmer) with carbonation carried out using 10% concentration of CO2 (balance N2) so as to emulate realistic flue gas CO2 concentrations. The details of reactivity testing using the TGA have been published elsewhere60.

The CaO sorbent needs to be subjected to high temperature calcination reaction to recover the CO2 in a pure form. This high temperature calcination causes the CaO sorbent to lose its reactivity toward CO2, and thus this deactivated sorbent is subjected to hydration to recover the reactivity. The samples used in the current study were accordingly tested for the calcination and hydration steps, with reactivity testing carried out after each step to monitor changes in reactivity. Thus the samples were first subjected to high temperature thermal sintering at 950 °C for 2 h. This step was carried out in an

Isotemp Muffle Furnace oven (Fisher Scientific). Following the sintering the samples are termed as calcined. Further, the samples were hydrated with excess water and dried overnight by placing the wet samples in controlled temperature fume hood.

The reactivity of the samples was tested after each step using TGA equipment. The morphological properties of the samples such as surface area and pore volume were

59 analyzed using BET techniques using a surface morphology analyzer (Qauntachrome).

Three samples were selected for further testing using X-ray diffraction (XRD) analysis and Scanning electron microscopy (SEM).

3.3 Observations and discussion

The following key observations are made: one, the sorbents obtained from various sources of waste animal shells exhibit inherently different CO2 sorption capacities. Two, all sorbents responded well to the regeneration by water hydration. Three, calcined samples exhibited a loss in reactivity, and the subsequently hydrated samples showed improved reactivity (Table 3, Figure 23) which is in agreement with previous studies.57,60

Note that evidence exists in published literature supporting regenerability of sorbent derived from oyster shell using hydration techniques.56 The BET analysis confirms the regeneration by increase in pore volume of the sorbent samples, as seen in Table 4.

Table 3: Calcined and hydrated CO2 capture capacities of sorbents after 30 min, tested in TGA

Calcined CO capture Hydrated CO capture Sample 2 2 capacity (wt%) capacity (wt%) CH-E 25 49 OS-E 21 53 DU-E 23 49 OY-S 13 37 CL-S 10 35

Figure 23: CO2 capture capacity (wt%) for (a) calcined sorbents and (b) hydrated sorbents

Table 4. Bulk pore volume and surface area analysis using BET theory

Pore volume Surface area 3 -1 2 -1 Sample (cm g ) (m g ) Calcined Hydrated Calcined Hydrated CH-E 0.024 0.033 7.3 5.5 DU-E 0.004 0.025 1.7 6.2 OS-E 0.005 0.014 1.9 4.1 OY-S 0.002 0.035 1.0 6.8 CL-S 0.045 0.068 2.7 3.2

It must be noted that the initial sample capture capacity for all samples is higher than that of the same after hydration. This is because, the solid sorbent initial capture capacity, shown in Table 2, is evaluated by first calcining the original samples (CaCO3) in pure N2 environment at lower temperatures of 700 °C, followed by immediate reactivity testing by exposure to 10% CO2 stream at 650 °C. Due to this low temperature ideal condition calcination, the original sample does not undergo any process of sintering and therefore exhibits higher reactivity. However, such ‘ideal conditions’ of calcination in pure N2 is not practical for actual cyclic process considerations, and therefore some sintering is inevitable in the actual process which is reflected in the lower reactivity even upon hydration.

The fourth key observation is that the samples derived from egg-shells consistently showed better reactivity toward CO2 than their sea-shell derived counterparts, as seen in

Figure 23. OY-S and CL-S samples exhibit lower reactivity toward CO2 than their egg- shell counterparts in all forms. The difference in reactivity of the samples can be due to various factors; such as inherently different chemical compositions, differences in the reaction surfaces of the different samples, different crystal structures or dominance of certain crystallographic planes, etc. Each of these factors was investigated using different analysis techniques, which are briefly discussed below.

The decomposition curve of the samples obtained via TGA was considered for estimating the active calcium component of each of the samples. This analysis confirmed that all samples contained similar fractions of active calcium component (~96 wt% CaCO3).

Therefore, this phenomenon was thought to be an attribute of the difference in surface

65 morphology of the sorbents and therefore their reactivity toward CO2. From the BET analysis carried out, the bulk surface properties such as total pore volume and surface area values (Table 4) do not indicate significant differences either. However, microscopic morphological differences play an important role in the overall reactivity of solid reactants. To better understand these microscopic morphological differences between the samples, two egg-shell derived samples and one sea-shell derived sample were chosen for

XRD and SEM analysis. The results obtained can be seen in Figure 24 and Figure 25.

From the Figure 24, all samples are calcite in the original crystal form, and were matched to JCPD 86-2335 C as the reference pattern. Only minute differences in peak heights of certain crystal planes are observed. The number associated with each peak represents a particular crystallographic plane, as matched to the reference calcite crystal form. The diffraction patterns of CH-E and OS-E samples match almost completely. The pattern of

OY-S sample shows higher concentration of planes 104, 006 and 018, and lower concentration than the egg-shell samples in a few minor peaks. However, these differences between the egg and shell derived samples from XRD analysis were insignificant to account for the large difference in reactivities observed. Finally, SEM analysis was carried out for the three selected samples. The major microscopic morphological differences are clearly evident in the SEM images, seen in Figure 25. CH-

E and OS-E samples (Figure 25a and b) have a porous surface structure exhibiting the so- called ‘popcorn effect’, whereas OY-S (Figure 25c) samples are seen to have much smoother surface morphology. Thus the higher reactivity of the egg-shell derived samples can be attributed to their enhanced microscopic surface characteristics, which lend them advantage over the sea-shell derived samples which do not exhibit such microscopic

66 surface enhancement. This advantage is apparent in the higher reactivity of these samples despite having the same bulk surface properties and chemical composition as well as crystal structure.

CH-E

OS-E 104

OY-S

018

116

113

202

012

110

122

006

024

214

300

211 ArbitraryIntensity

20 25 30 35 40 45 50 55 60 65 70 2θ

Figure 24: X-ray diffraction patterns obtained from CH-E, OS-E and OY-S samples: determining crystallographic differences.

Figure 25: SEM images of egg and sea-shell derived sorbents. (a) CH-E (b) OS-E and (c) OY-S.

The analysis techniques employed here reveal that the sea-shell derived samples exhibit lower affinities toward CO2 for carbonation reaction due to poor reaction surface area available as compared to the egg-shell derived samples. It is well documented that the acetic acid pretreatment of calcium sorbents increases the reactivity of the sorbents

61,62 toward CO2. From our results, the acetic acid wash given to egg-shell samples to separate the collagen membrane from the calcitic shell is believed to have enhanced the surface properties of the egg shell samples, resulting in higher reactivity.

3.4 Conclusions

In this study, it was successfully demonstrated that the calcium sorbent derived from different waste animal shells are cost-effective sorbents for CO2 capture from effluent gas streams of coal fired power plants. Three bird egg shells and two sea shells were tested.

Further, all the sorbents showed near complete re-activation or recovery of the CO2 sorption capacity upon hydration. The reactivity was confirmed using TGA as well as surface morphology measurements such as pore volume and surface area. The calcium sorbent obtained from all shells was confirmed to be predominantly in the carbonate form

(> 95%). Two egg-shell and one sea-shell sample were selected for studying the difference in reactivity. XRD analysis revealed the presence of calcite as the dominant crystal structure in all three samples. SEM revealed clear differences in surface morphology, indicating that the sea-shell sample has lower surface area for reaction, with appearance of smoother particle surfaces, which correlates well with the lower reactivity observed. The difference in the pretreatment of the egg and sea shells, viz. the acetic acid wash given to the egg shell samples to remove collagen membrane, is offered as an explanation to the general trend of lower reactivity of sea-shell samples towards CO2. As a result, acetic acid pretreatment is offered as a viable method of pretreatment of sorbent derived from waste animal shells.

CHAPTER 4: A Novel Calcium-Char (Cal-C) Process for the

Simultaneous Removal of NOx, SOx and CO2 from Combustion Flue

Gas

4.1 Introduction

The increase in anthropogenic gaseous emissions due to fossil fuel combustion has been accompanied in recent years with an increase in research efforts to curb them. Oxides of nitrogen (NOx) and sulfur (SOx) are common fossil fuel combustion pollutants that have been classified as criteria pollutants, with their ambient air quality standards defined from as early as 1971.63 Since then, numerous techniques have been developed for continuously improving the efficiency and/or extent of removal of these gases from

64,65 varied combustion sources. NOx, chiefly nitric oxide (NO), is produced mainly through high-temperature oxidation of atmospheric nitrogen (N2) that occurs during fuel combustion (thermal NOx) or from oxidation of nitrogen present in the fuel (fuel NOx).

The measures of reducing NO emissions include high-temperature techniques such as combustion modifications which entail producing environments that abate the formation

66 of NO by staged combustion, low NOx burners, etc. The medium- and low-temperature post-combustion techniques include selective catalytic or non-catalytic reduction (SCR and SNCR) of NO by a reducing agent (commonly ammonia),67, 68 reburning of fuel,69 or

70 ozone oxidation with subsequent removal of NO2, etc. For sulfur dioxide (SO2)

72 removal, the flue gas desulfurization (FGD) unit is an integral part of post-combustion flue gas processing in the majority of coal-fired power plants. The most common method of removing SO2 is using lime-based slurry, with the occasional dry sorbent injection also involving predominantly calcium oxide (CaO), and less commonly other alkali metal oxides such as sodium (Na) or magnesium (Mg).65

The worldwide increasing energy demand is resulting in ever-increasing amount of carbon dioxide (CO2) emission, particularly from large point sources such as coal-fired

42 power plants . Due to the greenhouse effect caused by CO2, there are several

71 technologies being considered for addressing CO2 reduction ; among which the limestone-based calcium looping technology is one of the most promising, having been demonstrated at various pilot-scale facilities around the world.17 The reversible reaction between CO2 and CaO forms the basis of this technology. This is a high-temperature technique in which the CO2 capture reaction occurs at 550-650°C. Based on this concept, the Carbonation Calcination Reaction (CCR) process was developed at The Ohio State

University (OSU) for the simultaneous removal of CO2 and SO2 from coal-combustion flue gas.20,2,22,38 The CCR process comprises a three-reactor system, with hydration as the preferred method of reactivating the CaO that otherwise loses its reactivity over multiple

60 cycles. The CCR process has successfully demonstrated >90% CO2 and 100% SO2 removal from flue gas generated from actual coal combustion in a 120 kWth unit at

22 OSU. The reaction of SO2 with CaO is virtually complete, with 100% capture achieved at 600-650 ̊C, which is lower than typical furnace dry sorbent injection temperatures.

This high extent of capture is feasible at lower temperatures owing to the extremely high

73 calcium-to-sulfur mol ratio – approximately 50:1 for a flue gas generated from a typical medium (4%) sulfur coal.22

Similar to SO2, NO is also present in the flue gas at very low concentration of a few hundred to thousand ppm. This NO can be reduced to N2 in the presence of a

72 carbonaceous material such as coal char. Oxygen (O2) present in the flue gas also catalyzes NO reduction using char at small concentrations.73 It is also well established that the presence of certain alkali and alkaline earth metals catalyze the NO reduction reaction.74,75 Various studies have quantified the effect of these metals, either present inherently in the char or added externally through experimental catalytic loading techniques.76,77,78,79 The effect of the presence of ash in the char has also been quantified on the basis of reaction temperature, total NO conversion, etc. Chang et al. noted that the activity of raw char is higher than that of de-ashed char.80 They also corroborated the known fact that the char reactivity toward NO decreases with increasing rank of the coal by testing anthracite, lignite, and bituminous coal chars. Activated lignite coal chars with a high Na content have been demonstrated to successfully remove NO from simulated flue gas at a pilot scale at Energy and Environment Research Center (North Dakota) in a demonstration of the CARBONOX process, which was developed at OSU.72 Synthetics chars with artificial CaO catalytic loading have also been reported to exhibit enhanced reactivity toward NO for the reduction reaction.81 If the char is chosen such that it possesses appropriate kinetics for NO reduction at the CCR process conditions (reaction time, temperature, etc.), then the simultaneous removal of NO, SO2, and CO2 may be

74 achieved in a single reactor by using a mixture of calcium sorbent and char. This forms the basis of the conceived Calcium-Char (Cal-C) process.82

This study provides an overview of the Cal-C process. The process concept has been verified by performing experiments in a fluidized bed reactor. Process analysis using

Aspen Plus simulation software has been performed and is also presented.

4.1.1 Process overview

A simplified block flow diagram of the process is shown in Figure 26. The process consists of a three-reactor loop; the simultaneous removal of the NO, SO2, and CO2 takes place in a single reactor, i.e. the Cal-C reactor.

Figure 26: The Cal-C process – conceptual schematic.

4.1.1.1 The Cal-C reactor

The flue gas containing NOx, SOx and CO2 enters the Cal-C reactor operating at 550-

650°C. Calcium hydroxide (Ca(OH)2) sorbent and char are also added to this reactor. The

Cal-C reactor is envisioned to be a fluidized-bed reactor with possibly entrained mode of operation. The Ca(OH)2 undergoes instantaneous dehydration to form CaO and the following reactions occur in the Cal-C reactor:

CaO + H2O(g) → Ca(OH)2 Rxn 1

CO2 and SO2 removal using calcium sorbent:

CaO + CO2 → CaCO3 Rxn 2

SO2 (g) + CaO (s) + ½ O2 (g) → CaSO4 (s) Rxn 4

NO removal by reduction on char:

2NO (g) + C (s) → CO2 (g) + N2 (g) Rxn 5

2NO (g) + 2C (s) → 2CO (g) + N2 (g) Rxn 6

Thus, by injecting the two solid reactants, viz. Ca(OH)2 and char, in a single reactor, it is possible to simultaneously extract NO, SO2, and CO2 from flue gas. The Ca(OH)2 is converted to CaO, CaCO3, and CaSO4, and the char is consumed in the process, leaving behind only inert compounds. The reaction kinetics are rapid and therefore the short residence times provided by an entrained mode operation are envisioned to be sufficient.

At the exit of the Cal-C reactor, the solids are separated from the gas stream by a high- temperature particulate capture device (PCD), such as a cyclone, and the clean gas is returned to the boiler for heat recovery, after which it is sent to the stack.

4.1.1.2 The calciner

The solid stream entering the calciner contains CaCO3, CaSO4, and may or may not contain unreacted CaO, and, depending upon the char combustion kinetics and amount of loading, unreacted char. The calciner operates at >900 ̊C and requires external energy input to both heat the reactants and to sustain the endothermic calcination reaction, which is backward reaction of

CaO + CO2 → CaCO3 Rxn 2

Typically, the external energy is provided by direct combustion of a fuel (e.g. coal, natural gas, etc.), or by indirect means through heat exchange. In addition, the unreacted char in the solid stream (if any) will also undergo combustion in the calciner and supply additional energy:

C(s) + O2 (g) → CO2 (g) ↑ Rxn 7

The direct combustion calciner is operated under an enriched O2 environment, which can be provided by the dilution of O2 with steam or CO2 (typically calciner tail gas). The decomposition of CaCO3 produces CO2, and the gaseous atmosphere in the calciner is required to be controlled in such a manner as to produce a concentrated stream of CO2 that can be easily sequestered with minimum downstream processing.

The gaseous products of the calciner are separated from the solid products at the exit of the calciner using a high-temperature PCD. The gas stream, mainly containing CO2, can be sequestered or utilized after appropriate processing and heat recovery.

4.1.1.3 The hydrator

High-temperature calcination resulting in a decrease in the reactivity of CaO toward CO2 is a well-known phenomenon. This occurs due to the thermal deactivation process known as sintering.29 Thus, the solid stream now containing sintered CaO is fed to the third and final reactor in this three-step process, viz. the hydrator. In the hydrator, the CaO undergoes reactivation by the hydration reaction:

CaO + H2O(g) → Ca(OH)2 Rxn 1

The hydration reaction may be carried out using steam in a reactor operating in the 300-

500 ̊C range, thus minimizing the temperature swing of the overall process and recuperating high-quality heat from the exothermic hydration reaction40. The hydrated solid sorbent is fed back into the Cal-C reactor along with a fresh stream of coal char, thus completing the three-step loop. A fraction of the total solids circulating in this loop consists of inerts (CaSO4, ash, etc.) that need to be removed from the system. Thus, a purge stream for the solids is warranted. The location of this purge stream is dependent upon several process considerations including solids throughput, reactor sizing, energy requirements, solids composition, and end value. For example, for the potential synergistic integration with the cement industry, purging solids after the calciner is advantageous. On the other hand, it would mean a higher solids loading through the calciner, resulting in increased calciner energy requirements etc.

The Cal-C process is a direct adaptation of the three-step calcium looping process for

20 CO2 and SO2 capture, viz. the CCR process, developed at OSU. This process has been

79 successfully tested at 120 kWth subpilot scale. Therefore, the focus of the present work is to determine the feasibility of the NO-char reduction reaction at the process conditions of the well-established CCR process.

4.2 Experimental section

Experimental setup: The experimental setup is shown in Figure 27. The gas mixtures were created using various mass flow controllers (MFCs) to obtain the desired inlet compositions of the reactant gases. The gas was then pre-heated in an electrically heated section before being introduced into the fluidized-bed reactor. The reactor is housed in an electrically heated tube furnace. The reactor has an inner diameter of 1”, and is loaded with reactive solids (either char or a mixture of char and CaO, depending upon the nature of the experiment). The solid bed is supported on a sintered plate, which acts as the gas distributor where gas enters the bed. Upon exiting the top of the reactor, the gas is cooled and dried. The gas exiting the reactor is continuously monitored for the NO, CO2, and

SO2 content, using chemiluminescence and gas chromatography techniques.

Figure 27: Experimental setup of reactor and gas analysis system.

Materials and methods: Different concentrations of inlet gases were achieved by creating gas mixtures of relevant compositions. The different gases used were O2, CO2, SO2, and

NO. The balance was always N2. Lignite coal samples were obtained from San Miguel

Electric Cooperative Inc., Texas and Walnut Creek Mining Company, Texas. Chars were prepared by thermal pretreatment (devolatilization) in N2 at more than 400 ̊C until tar ceased to evolve. The resulting char samples are hereafter termed LC-1 and LC-2, respectively. Bituminous coal char (BC) was obtained from Asbury Carbons. Laboratory grade (98%) Ca(OH)2 (ACROS Organics) was used as precursor for CaO. Both char and

CaO were taken in a fine powder form (<50 μm) to match the particle size used in the entrained flow operation of CCR process carbonator. However, in the lab-scale reactor, gas velocities were chosen so as to obtain a semi-batch operation, with solids being maintained in a fixed-fluidized bed mode.

Procedure: For the temperature programmed reduction (TPR), the reactor was loaded with a mixture of CaO and char and brought to the starting temperature (150°C) in N2.

Upon reaching the reaction temperature, the gases were switched to a mixture of ~900 ppm NO, 1.5% O2, and N2 and the TPR was performed at a ramp rate of 10 ̊C/min. The temperature was ramped until complete NO reduction was confirmed by the outlet NO concentration reaching zero. For the isothermal tests, the reactor was loaded with char or a mixture of CaO and char and brought to the reaction temperature (~650°C) in N2. Upon reaching this temperature, the inlet gas was switched to the reactant gas. The gas flow rate was always maintained at 210 ml/min (NTP). The progress of reactions was

82 monitored by measuring outlet gas concentrations. Breakthrough curves were obtained for each experiment.

4.3 Results and discussion

4.3.1 Effect of temperature on NO reduction in presence of CaO:

Based on the published literature, NO dissociatively chemisorbs onto the surface of coal char and is reduced at temperatures as low as 300°C.83 Nitrogen escapes in gaseous molecular form while the O is bound to the surface in various C(O) and C(O2) complexes.84 The energy required for desorption of these complexes is higher than that of the desorption of gaseous N2. However, at sufficiently high temperatures, these complexes can also be desorbed from the surface of coal char. When the C(O) complexes are released from the surface of the char (in the form of CO or CO2), they leave behind active sites for fresh NO molecules to chemisorb.85 Therefore, sustained NO removal necessitates that the rate of desorption of C(O) complexes be equal to or higher than the rate of chemisorption of NO molecules.

The presence of CaO on the surface of the coal char aids in the contact between the (O) and C on the char surface, thereby facilitating the regeneration of the active sites for

81 further NO chemisorption. Also, the presence of O2 in small concentrations aids the formation of active sites by gasification and aforesaid desorption of complexes. This results in lowering of the temperature of NO reduction. The O2 present in the flue gas and

CaO present for the removal of CO2 and SO2 thereby facilitate the char-NO reaction further. Therefore, TPR tests were carried out on the chars in presence of the CaO sorbent and O2 in the gas stream. The TPR showed a decrease in the outlet NO concentration

83 with an increase in temperature, as seen in Figure 28. For bituminous char, higher temperature may be necessary to achieve complete NO reduction, and the outlet NO concentration trend agrees with that reported by Gupta et.al73 The peak in NO concentration and subsequent rapid decrease around 550-600°C corresponds to the change in the availability of active sites on the char surface due to desorption of C(O) complexes. The temperature was not ramped above 700°C as the thermodynamics of the carbonation reaction prevent CO2 removal to the desired degree above this temperature, thus bituminous char seems incompatible with the desired process requirements.

However, in the presence of CaO, lignite chars exhibited complete NO removal at 650°C.

The feasibility of NO reduction at this temperature makes this process perfectly amenable with the CO2 and SO2 removal using calcium sorbent.

Figure 28: TPR of NO reduction using bituminous and lignite chars. Total flow rate =

210 ml/min, inlet NO concentration = 924 ppm, inlet O2 concentration = 1.5%, Ca:char loading = 10:1 (by wt.)

4.3.2 Effect of addition of CaO (presence and absence of calcium)

The presence of CaO is documented to have a catalytic effect on the NO reduction by char. This is due to metallic oxides facilitating the transfer of gaseous oxygen to the active sites on char by readily providing temporary binding sites for the (O) radical.81 For oxides having multiple oxidation states, this effect is more pronounced since there is a greater affinity towards oxygen by the metal oxide.81,86 The catalytic effect of CaO on the

NO-char reduction reaction has been quantified in earlier studies, where the CaO was loaded onto the char surface in small quantities by using methods such as solution impregnation.76,81 However, the samples prepared for the present study involved only physical mixing of Ca(OH)2 and char in pre-determined weight proportions. During the pre-heating of the reactor bed to the reaction temperature, Ca(OH)2 decomposes to produce the desired CaO sorbent by the reverse of reaction 1, as confirmed by the presence of moisture in the exit of the reactor.

It was observed that the addition of calcium sorbent to the char bed resulted in the enhancement of the NO reduction reaction. Using BC, sustained NO reduction was not observed in isothermal experiments carried out at 650 ̊C, as higher temperatures are known to be required for the reaction of NO with bituminous char.73 However, the breakthrough curve for outlet NO concentration exhibited a definite shift towards lower concentrations (Figure 29a) indicating increased reactivity of char. The same trend was also observed in case of the lignite chars, shown in Figure 29b for LC-1, in which a clear pre-breakthrough period was observed indicating complete and sustained NO reduction

86 during the same. Similar trends were also obtained for tests carried out with LC-2 (Figure

29c).

Since the char is also consumed by the competing reaction with O2 at a much higher rate

(reaction 7 and 8), the total amount of char used for NO reduction is calculated using selectivity, defined as total weight of NO reduced per unit weight of char consumed. The total amount of NO reduced can be obtained by calculating the cumulative area under the plot, in these isothermal experiments. Figure 30 shows the effect of various Ca(OH)2 to char ratios (referred to as Ca:char loading by wt) on the selectivity of NO reduction reaction for LC-2. The selectivity of NO reduction increased with an increase in the calcium loading. In the actual process, the solid loading ratios will be dictated by the relative concentrations of the corresponding reacting gases. Since the CO2 concentration in coal-combustion flue gas is orders of magnitude higher (12-15%) than the concentration of NO (in ppm), the calcium sorbent is expected to be in large excess as compared to the char in the Cal-C reactor.

Figure 29: Effect of addition of calcium sorbent on the NO-char reduction reaction using (a) BC, (b) LC-1 and (c) LC-2. Gas flow rate = 210 ml/min, inlet NO concentration = 924 ppm, O2 concentration = 1.5%, temperature = 650 ̊C. Ca:char loading = 10:1 by wt. Hollow symbols indicate absence of calcium sorbent and solid symbols indicate presence of calcium.

0.016

0.012

0.008

0.004

Selectivity Selectivity (mgNO/mg char) 0.000 0 5 10 15 Ca:char loading (by wt)

4.3.3 Effect of O2 concentration

The presence of O2 in the reacting gas mixture increases the rate of NO-char reduction due to the reaction between the oxygen and char. Char consumption by O2 results in the oxidation of the carbon to produce CO and CO2, thereby producing more active sites for the incoming NO molecules for dissociative adsorption. Therefore, tests were carried out to elucidate the effect of oxygen on NO-char reaction in presence of calcium sorbent. The

O2 concentrations were typical of combustion flue gas, varying between 1-5%. The combustion flue gas contains higher concentrations of O2 as compared to NO and the char-O2 reaction is thermodynamically favored over the char-NO reaction. This results in a lower selectivity of char for NO reduction at higher O2 concentrations, as shown in

Figure 31.

The oxygen reacts with char mainly via the following reactions:

C(s) + O2 (g) → CO2 (g) ↑ Rxn 7

2C (s) + O2 (g) → 2CO (g) Rxn 8

Thus, the char is consumed by O2, and higher O2 concentration results in faster consumption of char; therefore, shorter pre-breakthrough periods are observed with NO outlet concentration curves signifying a lower extent of reduction of NO in the reactor bed. These trends can be seen Figure 32. The results are in agreement with previous studies carried out on this reduction reaction in the absence of calcium sorbent.

Therefore, it can be concluded that the presence of calcium sorbent has no effect on the parameter of O2 concentration for the NO reduction reaction.

0.030

0.020

0.010

0.000

Selectivity Selectivity (mgNO/mg char) 0% 1% 2% 3% 4%

Inlet O2 concentration

1000 1% O2 3% O2 800 5% O2 600

400

200

0 OutletNO concentration (ppm) 0 10 20 Time (min)

4.3.4 Effect of inlet NO concentration

Different inlet NO concentrations were tested for the NO-char reduction reaction in the presence of calcium sorbent. Three different concentrations –1800, 900, and 600 ppm – were chosen to represent the presence or absence of various degrees of upstream NO abatement in an actual combustion system. At all the concentrations tested, complete reduction of NO was observed for a significant period at the start of the reaction, indicated by the flat pre-breakthrough region of the outlet NO concentration curves obtained. The selectivity of the char toward NO increased with an increase in the inlet

NO concentration (Figure 33). This observation makes the Cal-C process very appropriate for systems having high NO concentrations. However, the total amount of char consumed in the system will be dictated by the amount of O2 present in the gas, as it is the limiting reactant due to its higher concentration and higher reactivity toward char.

0.030

0.020

Selectivity Selectivity 0.010 (mgNO/mg char)

0.000 0 1000 2000 Inlet NO concentration (ppm)

Figure 33: Effect on inlet NO concentration on the selectivity of char-NO reduction reaction in presence of calcium sorbent and LC-2. Total gas flow rate = 210 ml/min, inlet

O2 concentration = 1.5%, temperature = 650 ̊C, Ca:char loading = 10:1 by wt.

4.3.5 Simultaneous capture of NO, SO2, and CO2

So far, the focus of this study has been the effect of several parameters on the NO-char reduction reaction, owing to a multitude of research material that has already extensively covered the topic of CO2 and SO2 removal using calcium sorbent. Previous experiments carried out at OSU have successfully demonstrated >90% CO2 removal and 100% SO2

22 removal using Ca(OH)2 sorbent, in the demonstration of the CCR process. The scope of the present study was therefore limited to ascertaining the suitability of the CCR process conditions for the reduction of NO by using inexpensive carbonaceous material such as char in conjunction with the calcium sorbent. Nevertheless, the simultaneous removal of all the species under consideration was also investigated, and the results are shown in

Figure 34. The duration of this test was 60 minutes, inlet gas stream contained 1800 ppm

NO, 3050 ppm SO2, 13% CO2, 1.5% O2, and the balance was N2. The temperature of the reactor was maintained at 650 ̊C. The Ca:char loading was set at 10. The reactor was heated to the reaction temperature by the electric furnace, while passing a stream of N2.

It was observed that outlet CO2 was maintained below 10% of the inlet value for a little over the first half of the experiment, after which it gradually increased and reached its inlet value by the end of the experiment. The pre-breakthrough period of the NO concentration curve lasted longer (~40 minutes), after which the char was completely consumed with unreacted NO evolving from the reactor, indicated by the gradual rise in outlet NO concentration in the latter half of the experiment. Finally, the SO2 was observed to react completely for the entire duration of the experiment, with the outlet concentrations at near zero values. Since SO2 can also react with the reaction product of

CaO and CO2, i.e. CaCO3, complete SO2 removal is possible well after the CaO reactant bed is exhausted by CO2 capture.

16 3000

SO2 inlet 14

2500 12 CO2 inlet

2000 10

concentration

2 8

1500 NO inlet (ppm)

6 concentration (%)

1000

NO out (ppm) 2 SO2 out (ppm) 4 500 CO

CO2 out (%) 2

OutletNO, SO

0 0 0 20 40 60 Time (min)

Figure 34: The simultaneous removal of NO, SO2 and CO2 from a simulated gas mixture, in presence of calcium sorbent and lignite coal char LC-2. Inlet CO2 concentration = 13%, Inlet NO = 1800 ppm, Inlet SO2 = 3050 ppm, Inlet O2 = 1.5%, Ca:char loading = 10:1 by wt. Temperature = 650 ̊C.

4.4 ASPEN simulations

Aspen Plus v 7.3.2 was used to evaluate the Cal-C Process. For comparison, four simulations were conducted: two simulations provide the base results and the remaining two simulations provide two possibilities for NOx integration into the CCR Process. In general, the parameters and guidelines provided by the US NETL-DOE were followed.

Case 1 is the base case. It models a 550 MWe net electricity generation coal-fired power plant combusting Illinois #6 coal. The material inputs are exactly identical to case 11 of reference 33. The boiler is modeled as a Gibbs reactor, which takes into account both physical and chemical equilibrium but does not account for reactor designs such as low-

NOx burners and overfire air. As a result, the NO concentration at the outlet of the Gibbs reactor is over 5200 ppm, which is artificially reduced to 545 ppm NO to better approximate the NO concentration in a coal-fired boiler with NOx control.87 Table 5 provides the inlet materials and major outlet flue gas components.

Case 2 models the CCR Process that simultaneously removes both CO2 and SO2 produced by flue gas by reactively fixating the gases using solid Ca(OH)2 to form CaCO3 and

CaSO4, respectively. Both experimental and process simulation results have been published elsewhere.20,22,40,88,89,90 The modeled CCR Process makes use of experimental results, published guidelines, and conservative estimates for any remaining unknown variables. Table 6 provides the important modeling parameters and Figure 35 shows a block flow diagram of the CCR Process.

Table 5: Modeling parameters for CCR Process.

Carbonator T 625 °C

CO2 Removal 90% of total CO2

SO2 Removal 100% Ca:C mol ratio 1.75:1 PCD-1 Efficiency 98% Purge Stream 1.5% of total solids Calciner T 1000 °C

O2 Purity 94.98%

Excess O2 7.71%

O2 concentration 25%

Fresh CaCO3 127 tons/hour Coal 142 tons/hour Oxygen 308 tons/hour PCD-2 99.5% Hydrator T 500 °C CaO Conversion 70%

H2O:Ca mol ratio 1.3:1 Steam Consumption 638 tons/hour

Table 6: 550 MWe coal-fired power plant – process conditions for Case 1

Input: Coal 204.75 tons/hour

Air 2040 tons/hour

Output: N2 72.5%

CO2 13.9%

H2O 8.9%

O2 2.0%

Ar 0.86%

Table 7: Modeling parameters for NO removal for the Cal-C process

Reactor T = (Carbonator T) 625 °C

NO removal 90%

Char Carbon (solid)

Reaction Route C + 2 NO → CO2 + N2

Case 3: Excess char combustion in calciner

Case 4: Excess char combustion in Cal-C reactor

100

Table 8: Summary of simulation results for the four cases

Thermal Energy (MWth) Case 1 Case 2 Case 3 Case 4

Input 1400.5 2371 2324 2420

Output 1232.5 2073 2045 2160

Coal Consumption (tph) 204.75 346.75 339.75 353.75

Steam Consumption (tph) 0 638 642 667

Fresh CaCO3 0 127 127 149

CO2 purity (dry) 96% 96% 96%

101

Figure 35: CCR process block flow diagram - heat exchangers not shown for simplicity.

102

Figure 36: The block flow diagram of CCR process with NOx control – Cal-C process.

103

Cases 3 and 4 consider NOx removal occurring with the CCR Process, i.e. the Cal-C process. This is achieved through char injection into the Cal-C reactor (Carbonator).

Although char can react with NOx through more than one reaction routes, only the reaction producing CO2 during NO destruction is considered, since formation of other species such as CO will not have any implication on the calcium sorbent. Another potential side reaction occurs with the char combusting with the O2 in the flue gas to form

CO2. Maximum CO2 production from the NOx removal process occurs when the char reacts with NO to form N2 and CO2 and the char completely combusts with the O2 in the

Cal-C reactor. While complete char combustion is thermodynamically favorable, the possibility exists that it may not due to kinetic limitations, given the short residence time in the entrained bed Cal-C reactor and 625°C reactor temperature. Case 3 considers 90%

NO reduction with char without char combustion while Case 4 is identical to Case 3 except considering complete char combustion in the Cal-C reactor. Table 7 provides the additional modeling parameters necessary to complete the NOx removal process and

Figure 36 shows the block diagram of the simulation.

From an energetics view, the CCR Process is a favorable method for carbon capture from coal-fired power plants. It also has the distinct advantage of removal of additional acid gases and maintaining the electric output of the process. Additional process intensification can occur by including a NOx removal step with the CCR Process. In one reactor, it is possible to achieve DOE target goals for carbon capture, remove additional unwanted acid gases, and also remove NO to below current environmental limits. Table 8 provides a summary of key results.

104

4.5 Conclusion

The Calcium-Char (Cal-C) process has been described for the single-stage removal of

NOx, SOx and CO2 from combustion flue gases from point sources and has been verified at laboratory-scale. Specifically, the NO-char reduction reaction has been studied in some detail and at process conditions suitable for CO2 and SO2 capture using a calcium sorbent. It is found that the presence of the calcium sorbent enhances the NO reduction reaction, evidenced by the breakthrough curves obtained in the presence and absence of

CaO. The well-known phenomenon of CaO catalyzing the char-NO reaction is observed by the simple physical mixing of the two solids employed here. Near complete reduction of NO over lignite chars is obtained at temperatures suitable for CO2 and SO2 removal using CaO. The O2 present in the flue gas is known to improve the extent of NO reduction. However, the higher concentration of O2 over NO, and the higher reaction rate between char and O2, results in parasitic consumption of char via combustion. The selectivity of char for the NO reduction is therefore higher at lower concentrations of O2.

This selectivity is also found to be higher at higher NO concentrations, making this process highly suitable for applications with low O2 and high NO emissions. The process simulation results corroborate the experimental findings and additionally highlight the opportunity for significant heat recovery in the two cases considered. In the case of char combustion in the calciner, the coal requirement of the calciner is reduced due to the heat produced by excess char combustion. In the case of char combustion in the Cal-C reactor, the exothermicity of char combustion enables lower inlet flue gas temperature, resulting in greater heat recovery in the steam turbine cycle prior to the Cal-C reactor. The Cal-C process uses bulk consumable raw materials such as coal char and limestone; and the

105 exothermic high-temperature reactions involved offer the opportunity for efficient heat recovery, offsetting some of the energy penalty for the removal of these species, and making the Cal-C process economically attractive.

106

CHAPTER 5: Calcium Looping Process for Coal-to-H2 Production:

Fate of Sulfur

5.1 Introduction The three stage Calcium Looping Process (CLP) has been proposed for the production of

91,92 high purity hydrogen (H2) from syngas derived from coal and methane (CH4). The process primarily consists of three reactors, much like the post-combustion calcium looping CO2 capture process. The schematic of the process is given in Figure 37.

In a conventional coal-to-H2 process, coal is gasified to produce syngas, which is the mixture of carbon monoxide (CO) and H2. This syngas is subjected to the water gas shift

(WGS) reaction in presence of steam (H2O) to produce carbon dioxide (CO2) and more

H2 in presence of high and low temperature WGS catalysts. A large amount of excess steam is used to overcome the equilibrium limitation of the WGS reaction. The acid gas removal is necessitated due to the formation of considerable amount of CO2 in this setup.

Additional steps of purification such as pressure swing adsorption (PSA) are also required downstream to enhance the purity of the H2 thus produced.

In the three step CLP shown in Figure 37, in the carbonator reactor the calcium oxide

(CaO) sorbent is generated from the decomposition of calcium hydroxide (Ca(OH)2).

This CaO then reacts with CO2 generated from the WGS reaction, by fixing it in solid

107 calcium carbonate (CaCO3) format and thereby pushing the equilibrium limited WGS reaction towards the right hand side. The use of solid sorbent reduces the excess steam requirement of the process, with the decomposition of Ca(OH)2 sufficiently supplying all the steam necessary to carry out the WGS reaction. The carbonator operates at ~600-

650°C.

The spent sorbent, in the form of CaCO3, is then transported to the calciner reactor, which operates at ~900°C. In this reactor, the CaO sorbent is regenerated by the endothermic decomposition of CaCO3. This high temperature regeneration step, however, causes deactivation of the sorbent through sintering. Therefore, the sorbent is ‘reactivated’ through steam hydration in a separate reactor, termed as the hydrator, at ~500°C. The

Ca(OH)2 thus formed is recirculated to the carbonator. The reactions are shown below.

Carbonator:

Ca(OH)2 → CaO + H2O Rxn 9

CO + H2O → CO2 + H2 Rxn 10

CaO + CO2 → CaCO3 Rxn 2

Calciner:

CaCO3 → CaO + CO2 Rxn 3

108

Hydrator:

CaO + H2O(g) → Ca(OH)2 Rxn 1

In addition to these main reactions (9), (10) and (2) from the carbonator reactor, the CaO sorbent also results in the removal of additional pollutants such as gaseous sulfur and halides.

Sulfur removal:

CaO + H2S → CaS + H2O Rxn 11

CaO + COS → CaS + CO2 Rxn 12

Halide removal:

CaO + 2HCl → CaCl2 + H2O Rxn 13

The use of the three step CLP system therefore results in process intensification through multipollutant removal. Therefore, several process units such as syngas scrubber, high and low temperature WGS reactors, two-stage Selexol unit, can all be replaced by the

93 single CLP system to produce high purity H2.

109

Figure 37: Schematic of the three-step CLP process for high purity H2 production from coal-derived syngas39

5.2 Motivation/Problem statement

The coal-to-H2 process using CLP was simulated using Aspen Plus software and the economic analysis was also performed by comparing the CLP plant with the base coal-to-

93 H2 plant with CO2 capture, using the same gasifier capacity as the baseline. The baseline case was adapted from the U.S. Department of Energy’s (DOE) report on the baseline state of the art plants for H2 production with CO2 capture with Illinois #6 coal.

As stated before, the use of CLP system obviates the need of several unit operations such as the syngas scrubber, the WGS reactors, syngas coolers, as well as the two stage

Selexol unit. In this process, the carbonator reactor operates at a pressure of 23 bar and a

110 temperature of 650°C. The calcium sorbent to carbon mole ratio of 1.3, which has been earlier proven sufficient to achieve 90% CO2 capture in a post combustion scenario, is employed. The steam required for the WGS reaction is provided by the thermal decomposition of Ca(OH)2 as well as steam already present in the syngas, and no additional steam is therefore required to be injected into the carbonator. In keeping with previous lab-scale experimental findings, very high purity H2 (~95% on a dry basis) is generated from the carbonator operated at the given temperature and pressure. It is further purified to 99.9% pure after a PSA unit.

The application of CLP to the coal-to-H2 process necessarily results in the significant co- production of electricity, due to the several opportunities of heat integration afforded by the high temperature process. The block diagram in Figure 38 of the simulated case shows the heat recovery opportunities for heat integration as indicated by the red dotted arrows.

111

112

The CLP plant requires substantially more fuel (coal) than the base plant, due to the high temperature exothermic reaction of calcination, but it also automatically provides substantially towards a significant co-production of electricity through heat recovery steam generation. While the base case 26 tph H2 power plant generates 7.2 MWe of net electric power, the CLP system applied to the identical coal-to-H2 process results in the generation of 320 MWe of net electric power, according to the process analysis conducted by CONSOL Energy.93

One of the uncertainties pointed out at the time of this techno-economic analysis was that of fate of sulfur. The sulfur that is removed from the syngas stream in the carbonator is fixed in the form of calcium sulfide (CaS). This CaS may undergo additional oxidation reactions at the calciner operating conditions. The sulfur enters the calciner via two routes, 1) in the form of CaS solids, and 2) in the form of sulfur dioxide (SO2) from the coal combustion. This CaS can undergo several reactions with the oxidizing agents in the calciner such as oxygen (O2), CO2, H2O and result in the formation of calcium sulfate

(CaSO4).

CaS + 4CO2 → CaSO4 + 4CO Rxn 14

CaS + 2O2 → CaSO4 Rxn 15

CaS + 4H2O → CaSO4 + 4H2 Rxn 16

The CaSO4 thus formed will circulate back to the carbonator with the active solids stream and may react with H2 or CO component of the syngas and reduce the H2 yield of the process by parasitically consuming the gases.

113

CaSO4 + 4CO → CaS + 4CO2 Rxn 17

CaSO4 + 4H2 → CaS + 4H2O Rxn 18

Alternatively, the CaSO4 formed in the calciner may also participate in a solid-solid reaction with CaS and thus produce CaO and release SO2.

CaS + CaSO4 → 2CaO + 2SO2 Rxn 19

Thermodynamically, all these reactions (14), (15), (16), and (19) are feasible at the calciner operating conditions.

From the standpoint of the techno-economic evaluation, the release of sulfur in the gaseous phase was desirable in the calciner in the form of gaseous SO2. There are twofold reasons for this. One, the sulfur would then exit the loop via the CO2-rich gaseous stream, where it would be treated with a dry CaO sorbent at lower temperature conditions.

Theoretically, if feasible, this would prevent buildup of the sulfurous species in the solid sorbent loop. Two, and more importantly, if the sulfur remains fixed in the solid loop in the form of CaSO4, there exists a possibility of it being to the detriment of the overall H2 yield of the carbonator, by parasitic consumption of CO and H2 at the carbonator operating conditions via reactions (17) and (18).

The CaS remaining unreacted in the calciner is the least desirable outcome. This unreacted CaS would then be present in the purge stream, which is located after the calciner, and this would pose waste disposal issues.

114

With this background, the techno-economic analysis was performed with the process design that assumed the release of sulfur in the form of gaseous SO2. The current work was therefore proposed as experimental substantiation/rebuttal of this design.

5.3 Background and literature review

The different routes of oxidation of CaS have been subject of a plethora of research for specific application in PFBC-CC (Pressurized Fluidized Bed Combustion-Combined

Cycle) and IGCC (Integrated Gasification Combined Cycle). In both processes, the sulfur in the coal is released and is mainly converted to hydrogen sulfide (H2S). Addition of calcium sorbents like dolomite and limestone is a popular way of capturing H2S, but this generates large quantities of hazardous CaS. Hence researchers have suggested the oxidation of CaS by O2, CO2 and H2O as a means to safely dispose the waste. Oxidation of CaS mainly produces CaSO4 which is benign and stable at ambient conditions.

94 Qiu et.al. have previously investigated reaction 10 at 500-1000 °C with 1-40% O2 in N2.

They reported the formation of both CaSO4 and CaO via reaction 10 and 15, respectively.

The kinetic data in that study revealed reaction 10 to be a first order reaction with respect to O2. Song et.al. reported that the molar contents of CaSO4 and CaO increased with temperature substantially and CaS content decreases accordingly, between 700-900°C.95

96 The oxidation of CaS by reaction with CO2 and H2O was studied by Anthony et.al.

They concluded that CO2 is the more important oxidant as compared to H2O. According to results of Marbán et. al.97, reaction 11 is the prominent reaction below <890°C, and this temperature is identified as the upper limit for suitability of CaS conversion to

CaSO4. In these studies, the aim was to ‘destroy’ CaS and the formation of CaSO4 was

115 favored because of ease of disposal. Though prior work performed by other researchers provides a strong indication that reactions 14-12, and 19 will occur in the calciner, the existing data does not allow us to predict the extents of these reactions in the residence time of interest. The CLP-calciner is expected to operate at a solids residence time which is of the order of seconds. Hence, the goal of the work is to establish if these reactions occur to a significant extent in such a short time, or if they are insignificant enough to be neglected. Finally, the competitive kinetics of CaS oxidation in the presence of different oxidants also needs to be investigated to draw accurate conclusions.

The reductive decomposition of CaSO4 by CO was investigated by Wheelock and Boylan at temperatures above 1100 °C.98 Their findings indicated that at higher temperatures, the formation of CaO was favored over CaS. Diaz-Bossio et.al. investigated the same using both CO and H2 between 900-1180°C and found the reactions to be of first order with

99 respect to CO and H2. Thus, the reductive decomposition of CaSO4 has been extensively investigated, however, at much higher temperatures than the 550-700°C range expected in the CLP carbonator. In general, the gas compositions, particle sizes and composition, temperature and pressure range of investigation in the proposed work were quite different from the existing literature. Therefore, in spite of a large body of work already existing on the solid compounds of interest involved in the sulfur chemistry in the

CLP system, a substantial knowledge gap from the standpoint of CLP conditions existed at the time that this work was undertaken. Thus, the present work was conducted to study the kinetics of reactions of the various calcium-sulfur species (reactions 14 to 19) as applicable to the CLP system.

116

5.4 Conditions tested

The calciner equilibrium concentrations are given in the following Table 9. These gas concentrations have been obtained from the Aspen Plus simulations performed on the three cases of H2 production with CLP system. For the present work, only the concentrations from the first column were selected (highlighted). In the ASPEN simulations, the calciner operates at 875°C and 1 atm.

Laboratory grade CaSO4 and CaS samples were used for the experiments conducted. Pure gas bottles were used to create the appropriate gas concentrations using a battery of mass flow controllers for simulating the carbonator and calciner operating conditions, as explained further in section 5.4.2. The

Table 10 shows the carbonator concentrations on dry basis derived from Aspen Plus simulations.

117

Table 9: Calciner equilibrium gas concentrations from Aspen Plus simulations

Calciner, gas concentration v/v%

Gas Coal-to-H2 IGCC SMR

CO2 58-70 73-80 69-78 CO 0-0.5 0-0.5 0-0.2

H2 0-8 0-0.1 0

H2O 24-26 17-23 20-24

O2 0-5 0-0.5 0-3

SO2 0.2-0.5 0.3-0.5 0.2-0.3

Table 10: Carbonator equilibrium gas concentrations derived from Aspen Plus simulations

Carbonator, gas concentration v/v%

Gas Coal-to-H2 IGCC SMR

CO2 - - - CO 0-36 0-36 0-7

H2 0-35 0-35 0-44

H2O - - -

O2 - - -

SO2 - - -

118

5.4.1 Materials

Calcium sulfide (CaS, 99% purity, <1% Mg) was obtained from Fisher Scientific. Pure

CO2, N2, and air gases were used to create gas mixtures of required concentrations.

Various temperatures within the range of 800-950°C were chosen and CO2 concentrations were chosen between 60-80% (total pressure = 1 atm). A total flowrate of 430 ml/min was maintained and a sample size of 0.12-0.14 g was used for each experiment. H2O concentrations of 17, 20 and 24% were used to match the calciner conditions, and 3 different temperatures of 875, 900 and 925°C were tested. For oxidation by O2, concentrations of 1,3, and 5% were used according to Table 9. In addition, ppm quantities of SO2 gas were introduced in the thermogravimetric (TGA) reactor to observe the reaction kinetics of CaO conversion to CaSO4.

5.4.2 Experimental setup and procedure

Isothermal experiments were carried out in the Rubotherm thermogravimetric analyzer

(TGA) employing a magnetic suspension balance (MSB). This MSB equipped TGA would be used for all future tests as it is capable of handling high pressure experiments proposed in this work, which are not achievable using typical TGA apparatus. Also, due to the unique weight measurement technique of this MSB, it is possible to carry out experiments involving high concentrations of steam, a feature absent from traditional

TGAs. The schematic of the Rubotherm MSB is shown in Figure 39. A battery of mass flow controllers (MFC) was used to control the flowrates of individual gases, in order to simulate various required concentrations of reacting species. Valves 1 through 7 were used to control flow towards reactor. The MSB is coupled to the sample weight through

119 magnetic coupling; therefore the balance chamber is decoupled from the sample cell and does not require a separate purge gas. Thus, higher concentrations of reactant gases are achievable in the sample cell. The gas mixture is heated prior to the inlet of the sample cell by means of electrical heating tapes, resulting in gas temperatures of 250-300°C at the sample cell inlet. The sample cell is maintained at reaction temperature by means of electrical heaters. The section between the sample cell inlet and electrical heaters is heated by a heated oil jacket connected to a circulating oil bath (Corning 550). The temperature and pressure of the sample cell is measured by the thermocouple (TC) and a pressure transducer respectively. The desired pressure in the system is achieved by the help of a back pressure regulator (BPR) situated downstream of the sample cell. Steam was generated by injecting metered water through a precision syringe pump (ISCO series

100DM). The water was injected into the heated gas inlet line, which was filled with quartz wool to increase contact area between the heated gas and water, to convert it to steam. For experiments with O2 reaction, the desired concentration was obtained by mixing air and N2 in appropriate ratios. The carrier gas (usually N2) passing through this preheater section aids the vaporization of this water and carries the produced steam to the reactor (sample cell). Due to this method, it is not possible to maintain precise control over the concentration of steam experienced by the sample. However, the total amount of water injected is found to be sufficient to maintain a steam environment throughout the course of 60-80 minutes duration of experiment. In addition, the outlet gases were analyzed using a micro-GC (CP-4900, Varian) to identify the species present in the gaseous product stream of certain experiments. In addition, the on-campus center for

120 microscopy and analysis (CEMAS) facility was used to carry out the XRD analysis

(Rigaku SmartLab XRD) of the solid product.

The sample weight and temperature was continuously recorded by a computer as the reaction progressed. Conversion was estimated from the change in weight by assuming that increase in weight is only caused by the reaction (10).

Molar conversion (X) of CaS into CaSO4 is calculated as:

푚표푙푒푠 표푓 퐶푎푆 푟푒푎푐푡푒푑 푀푊 ∗ ∆푊 푋 = = 퐶푎푆 푚표푙푒푠 표푓 퐶푎푆 푝푟푒푠푒푛푡 (푀푊퐶푎푆푂4 − 푀푊퐶푎푆) ∗ 푊𝑖 Where,

MWm is the molecular weight of species ‘m’,

ΔW = Total increase in weight, mg

Wi = initial weight of CaS sample, mg

121

Figure 39: Schematic of the Rubotherm MSB setup used for performing thermogravimetric experiments.

122

5.5 Results and discussion

5.5.1 Reaction of CaS with CO2 as the oxidizing agent

The isothermal experiments of CaS and CO2 were conducted in the range of concentration of 65%-80% CO2, and the temperatures of 825 to 900°C.

Figure 40 shows graphs of conversion at various CO2 concentrations, namely, 65, 70, 75 and 80% at 875°C. Every experimental condition was repeated 2 or 3 times and the average readings are reported, with the error bars indicating the standard deviation. As the concentration of CO2 is increased, the conversion increases, indicating a higher reaction rate. Very low conversion was achieved for 65% CO2 in nitrogen. At 80% CO2 concentration, the nature of the curve changed, indicating a fast, reaction controlled regime followed by a slow diffusion controlled regime. Nevertheless, the conversion remained extremely low within the first 20 minutes of isothermal experiments in all the conditions tested.

The effect of temperature was also verified at various CO2 concentrations. Figure 40 shows one such trend at a fixed CO2 concentration of 80% at four different temperatures,

825, 850, 875 and 900°C. As the temperature was increased, the reaction rate was seen to decrease slightly. Independent studies show a weak dependence of this CaS oxidation on temperature. Furthermore, this dependence is reported to grow weaker at higher pressures of CO2. Also, the two-regime reaction rate behavior mentioned above is also seen at lower temperatures. At the highest temperature tested here, the variance increased between replicates.

123

0.12

0.1

0.08

825 C 0.06 850 C

conversion 875 C 0.04 900 C

0.02

0 0 5 10 15 20 time, min

Figure 40: CaS conversion to CaSO4 as a function of time, at different isothermal temperatures.CO2 concentration fixed at 80%, total pressure = 1 atm.

124

At all conditions tested, the CaS conversion to CaSO4 is very low, and never exceeded

10% over a period of 20 minutes of isothermal reaction. The product samples were stored in airtight containers from tests conducted, and these samples were analyzed using XRD techniques. Also, the spectra obtained were compared with that of the lab grade reactant

CaS sample. The product spectrum was not significantly different from the reactant, as predicted by the very low conversions obtained in the TGA over the duration of each test.

Although not a quantifying technique, the areas of each peak were used to obtain qualitative information about the relative amounts of various species present in the product sample. It was evident that majority of the product was CaS (>90%). Other species identified were CaCO3, CaO (trace) and CaSO4 (trace). Therefore it was concluded that CaSO4 can react with CaS further in presence of CO2 to give CaCO3 if the conditions are amenable to carbonation – however, it is not of consequence in the real system, as temperatures of operation of the calciner will be maintained above the maximum temperature of carbonation at given CO2 concentration levels.

5.5.2 Reaction of CaS with H2O as the oxidizing agent

Similar to CO2, full parametric tests were performed for the oxidation of CaS with H2O.

Very small weight changes were observed with the addition of steam at high temperatures of calcination. The results are shown in Figure 41.

Following reaction is considered:

CaS + 4H2O → CaSO4 + 4H2 Rxn 16

125

CaS seems to be very mildly reactive towards steam at the calciner operating conditions.

At all concentrations and temperatures tested, the conversion did not rise above ~5% in

10 minutes. Also, in the narrow range of conditions tested, no significant change was seen in the reactivity with temperature or partial pressure of steam.

126

Figure 41: CaS conversion to CaSO4 as a function of time, at different H2O concentrations. Isothermal experiments at (a) 875°C, (b) 900°C, and (c) 925°C

127

Steam addition was conducted by injecting water using a micro-syringe pump as noted in section 5.4.2. Due to this technique, the determination of exact concentration of steam achieved during reaction is impossible. Nevertheless, a very narrow range of 17-24% steam concentration was studied here in accordance with the values given in Table 9. In the range of temperature and concentrations tested, no significant change in reaction rates was observed.

Similar to CO2 (as exhibited in the last section), oxidation with steam seems to be a two- stage reaction. Initially the reaction is kinetically controlled, with rapid increase in weight corresponding to formation of CaSO4 product. However, this phase is very short and after about 2 minutes of fast reaction, the TGA curves change to indicate a slower, diffusion- limited rate. Similar phenomenon was observed under CO2 studied in the previous section.

5.5.3 Reaction of CaS with O2 as the oxidizing agent

Full parametric tests were performed for the oxidation of CaS with O2, with the temperature varied between 875-925°C, and the concentration of O2 between 1-5%, in accordance with values from Table 9. The results are shown in Figure 42.

128

Figure 42: Oxidation of CaS with oxygen at calciner operating conditions. Varying the O2 concentration at isothermal conditions, product is CaSO4 at (a) 875°C, (b) 900°C, and (c) 925°C

129

As expected, the reactivity of CaS increases with increase in O2 concentration at all temperatures tested. However, as temperature is increased, any increase in the O2 concentration starts having a more pronounced increase in the CaS conversion or reactivity. The highest reactivity is obtained at 925°C with 5% O2, which is the upper boundary of O2 concentrations expected at the calciner conditions. Also, an interesting phenomenon of two-step reaction is observed only at this reaction condition. This indicates that a combination of high temperature and O2 concentration results in a fast kinetic controlled reaction followed by product layer diffusion resistance.

To compare these rates with that obtained due to oxidation by steam and CO2, the three oxidation curves at 900°C are given on the same plot in Figure 43. O2 is observed to be far more effective in converting CaS to CaSO4 even at the low concentrations it is expected to be present at in the calciner.

130

5.5.4 Reaction of CaO with SO2 released from oxycombustion of coal

Another important side reaction of sulfurous compounds that may occur in the calciner, is that between the calcined CaO sorbent, and SO2 released by the oxycombustion of coal in the calciner. This reaction is feasible in the presence of the small amounts of O2 present in the calciner (reaction 21).

To study this reaction for the formation of CaSO4 from calcined sorbent (CaO), pure

CaCO3 was calcined in inert atmosphere of N2. The temperature was ramped at a fixed rate, and complete decomposition of CaCO3 was achieved before isothermal conditions were reached. A reaction mixture of 1% O2 and 2000 ppm SO2 was created by mixing

131 appropriate amounts of standard gas mixtures. Figure 44 shows a typical TGA graph obtained.

Figure 44: Typical TGA graph for reaction between CaO and SO2/O2 mixture, starting from CaCO3 decomposition in inert N2

132

The following reaction takes place:

CaO + 0.5 O2 + SO2 → CaSO4 Rxn 20

Figure 45 shows the effect of temperature on CaSO4 formation at a fixed O2 and SO2 concentration. Substantial CaSO4 formation is observed at higher temperatures, however the reaction rate is low at T ≤ 900°C.

The conversion is calculated using the following formula:

∆푤 푚표푙푒푠 표푓 퐶푎푂 푟푒푎푐푡푒푑 ⁄80 푐표푛푣푒푟푠𝑖표푛 = = 푤표 푚표푙푒푠 표푓 퐶푎푂 푎푣푎𝑖푙푎푏푙푒 ⁄56

Where Δw = weight increase upon injection of reactive gas mixture wo = calcined sample weight

133

Figure 45: CaSO4 formation from CaO at calciner operating conditions at various temperatures. 1% O2, 2000 ppm SO2

134

Thus, the investigation of the reactions of sulfurous species at calciner conditions revealed that CaS formed in the carbonator is mostly likely to undergo oxidation to

CaSO4 in presence of the various oxidizing species, with O2 being the strongest oxidizing reactant despite the small equilibrium concentrations predicted by Aspen Plus simulations. In addition, CaSO4 may be formed by the reaction between CaO and SO2 released from coal combustion in presence of O2 in the calciner. However, the small residence times envisioned for a fast-fluidized or entrained flow calciner are not expected to be sufficient to result in complete oxidation of CaS to CaSO4. Therefore, the sulfurous solids recirculating towards the carbonator must be a mixture of unreacted CaS and

CaSO4. The solid stream recirculating back to the carbonator would consist Ca(OH)2, unhydrated CaO, and among the sulfur compounds, unreacted CaS and CaSO4. Hence, it is essential to study the reactivity of CaSO4 at carbonator conditions.

5.5.5 Reaction of CaSO4 with H2

CaSO4 can undergo the following reactions in presence of H2:

CaSO4 + 4H2 → CaS + 4H2O Rxn 18

CaSO4 + H2 → CaO + H2O + SO2 Rxn 21

CaSO4 + 4H2 → CaO + H2S + 3H2O Rxn 22

The above three reactions were analyzed using HSC chemistry software. The equilibrium constants of the three reactions as a function of temperature are given in Figure 46.

135

rxn. 18 rxn 21 rxn. 22

Figure 46: Equilibrium constant of CaSO4 decomposition reactions as a function of temperature

136

The pressure of the system was successively ramped with every test. Since the sample weight is highly sensitive to the pressure, as the pressure increases the weight decreases of the system. Therefore the pressure was ramped after the reaction temperature had been reached, in presence of N2. Once pressure stabilized, the H2 gas was introduced. The samples were collected after the experiment and sealed in an airtight container. They were analyzed using X-ray diffraction (XRD) to identify the product phases formed. The following Figure 47 shows the results. The XRD analysis did not reveal significant new insights, that is, the product and reactant spectra don’t look much different with product sample also predominantly CaSO4 after 40 minutes of reaction with 30% H2. The sample was collected at the end of several runs carried out at 8-10 atm.

137

Figure 47: XRD analysis of (a) reactant and (b) product solid samples - Matched against CaSO4 and CaS

138

5.5.6 Reaction of CaSO4 with CO

Pressure was ramped prior to injecting CO into the reactor. However, as soon as CO was injected into the reactor cell, the sample weight was observed to increase immediately.

The reaction of interest (reaction 17) must result in weight loss in accordance with the conversion of CaSO4 to CaS and therefore this weight gain was unexplained till the reactor chamber was opened after the apparatus was cooled. The weight gain was found to be due to the Boudouard reaction (see below) and soot deposition was visible on various interior parts of the TGA. Thus, this carbon deposition was observed immediately at all pressures at which the experiments were carried out up to 10 atm. For example, see

Figure 48.

139

1% samplewt change,% 0%

-1% 0 20 40 60 80 100 time, minutes

Figure 48: Carbon deposition indicated by weight increase upon injection of CO. 650°C, 5 atm, 30% CO.

Explanation of steam addition:

The reaction between CaSO4 and CO which may result in parasitic consumption of CO by

CaSO4 + 4CO → CaS + 4CO2 Rxn 17

Boudouard reaction, resulting in carbon deposition

2CO(g) = C + CO2(g) Rxn 23

And the desired water gas shift (WGS) reaction

CO + H2O → CO2 + H2 Rxn 10

140

10 reactionReduction 1

(17) Log(K) 5 reactionBouduoard 2 (23) reactionWGS 3 0 (10)

-5 0 200 400 600 800 Temperature, ̊ C

Figure 49: Equilibrium constants of CaSO4 reduction, Boudouard and WGS reactions as a function of temperature

It was identified that the addition of steam can be used to suppress the Boudouard reaction. This technique was used as an attempt to isolate reaction 13 and 2. Although the soot deposition was successfully eliminated using this technique, it should be noted that the addition of steam may affect the equilibrium of reaction (10) as well, however at this point it still remains the best method of suppressing any weight changes that might occur due to reaction (23). The thermodynamic equilibrium constants of the three reactions under consideration were calculated using the HSC chemistry software, and are shown in Figure 49 as a function of temperature. Although the soot deposition was successfully eliminated using this technique, it should be noted that the addition of steam may affect the equilibrium of reaction (17) as well, however at this point it still remains the best method of suppressing any weight changes that might occur due to reaction (23).

141

Steam was injected for up to 10 minutes prior to the injection of CO. Under these conditions, different pressures were tested for CO reaction and outlet gas was analyzed.

Due to the occurrence of WGS reaction in these experiments, the micro-GC detected presence of H2 in the outlet gas. The Boudouard reaction is known to occur at atmospheric pressure too as evidenced in the sub-pilot testing of the CLP unit, and it is suppressed through the addition of external steam injection.24 However, from these experiments it is evident that parasitic CO consumption due to reaction 23 is more significant than that due to reaction 17 (the original reaction of interest). Secondly, during the experiments with steam injection, reaction 23 was found to be suppressed (absence of/reduced degree of soot deposition). Further, even if WGS reaction occurs, the sample weight change would only be due to the reaction 17 in the present system. No significant weight loss is observed, as evidenced in Figure 50. This indicates that reaction 17 does not occur at a rate significant to the CLP system, at the conditions employed here.

142

Figure 50: 30% CO experiment with prior steam injection, 650°C and 8 atm (a) outlet gas concentrations, and (b) sample weight change measured by the MSB.

143

5.5.7 Reaction of CaSO4 with H2 and CO

The carbon deposition phenomenon was completely eliminated, by perfecting the prior steam injection method through a process of trial and error. It was observed that negligible weight change occurred during all the experiments at carbonator conditions.

Further, the slight weight “gain” observed in certain experiments at the carbonator conditions was determined as caused only due to buoyancy changes in the reactor.

Several identical tests were repeated to generate enough solid product for the XRD analysis. One such typical run is shown in Figure 51. The lack of weight loss indicated that CaSO4 decomposition did not occur during the testing, which was further confirmed upon XRD analysis. The product solids only identify as CaSO4, which confirms that

CaSO4 is unreactive to reductive decomposition at the carbonator conditions tested here.

The CO2 concentration trend in the outlet gives an indication of the presence of steam in the reactor (and its gradual decrease as reaction time progresses). Further evidence of steam during the reaction is the condensation observed on some of the internals, after cooling down to room temperature.

144

35% 1.4%

30% 1.2%

25% 1.0%

20% 0.8% CO, %

15% 0.6% H2, %

concentration

2 TGA wt% change

10% 0.4% CO CO2, % 5% 0.2%

0% 0.0% %weight change,and concentration gas 0 20 40 60 80 100 time, minutes

Figure 51: Typical experiment showing weight change (TGA) and outlet gas composition

(micro-GC) as a function of time, upon injection of reactive gases. 30% CO, 30% H2 and steam injection before the experiment. T = 650°C and P = 10 atm.

145

At the same operating conditions, the main reaction occurring in the carbonator was also tested, which is the water gas shift reaction in presence of CaO sorbent. In this case, the exact same experimental conditions were employed except that CaO solids were used instead of CaSO4. This was accomplished by generating the CaO in-situ by heating limestone sample to decomposition temperatures prior to the injection of CO/H2 mixture.

The comparison is shown in Figure 52.

It can be seen that the CaO sorbent undergoes almost complete carbonation at the reaction conditions tested.

146

Figure 52: Comparison of weight changes, when using (a) CaSO4 and (b) CaO solids in the TGA at 650°C, 10 atm, 30% H2 and CO.

147

Thus, it was discovered that CaSO4 is stable and does not undergo reductive decomposition in presence of H2 and CO at the carbonator reaction conditions employed in the CLP process. This works in the favor of the CLP process, as the reductive decomposition of CaSO4 would mean lowering of the yield of useful gaseous products by parasitic consumptions of H2 and/or CO. The prior literature gives evidence of CaSO4 decomposition at conditions of higher temperature (900°C and above) and atmospheric pressure, however we conclude that this decomposition does not occur at the CLP carbonator operating conditions. Therefore, for the purposes of the CLP system, CaSO4 seems to be the final stable solid product of the sulfur species, for the sulfur entering the system via various gaseous/solid phases.

Stability of CaSO4 at high purity H2 (carbonator exit)

It should also be noted that, according to the ASPEN process simulations for the CLP process conducted independently by CONSOL energy, the CO undergoes 99% conversion in the short residence time of the carbonator, leaving a product stream exiting the carbonator containing majority H2, along with H2O and small amount of other

91,39 gaseous species. This has been verified by fixed bed experiments. >90% of the CO2 produced as a result of WGS reaction is captured by the CaO sorbent, and therefore the exit gas consists of 60-70% H2, about 30-38% H2O and about 2% other impurities. On a dry basis, the H2 purity in the reactor exit is 95-97%, with ~1-2% unreacted CO, and ~1%

CO2. Thus, the actual CO partial pressure in the reactor is likely to be much lower than that tested here, and the actual H2 partial pressure much higher.

148

Therefore, a test was carried out to determine the formation of CaS at the conditions corresponding to the above discussion. On a dry basis, ~95% H2 (rest N2) was injected in the sample cell at 650°C and 10 atm. Steam was also present during the experiment, generated by the method outlined on page 2. The sample did not exhibit any weight loss during a 60 minute test. The weight of sample remained the same (~0.1 g) as the starting weight upon turning off the H2 gas at the end of the 60 minute period. The product sample was collected to be analyzed using XRD technique. The XRD analysis of this stored sample conducted on Rigaku SmartLab XRD identified CaSO4 as the only solid phase. The lack of weight loss at these test conditions also strongly indicates that CaSO4 is stable at the carbonator outlet conditions.

5.5.8 Treatment of purge stream

Once it has been established that CaSO4 does not adversely affect the desired reactions and/or reactive species in the CLP system, the focus now shifted to the safe disposal of the sulfurous species that leave the system via solids purge stream. It was concluded that the strongest oxidant for CaS is the oxygen that is present in the calciner, up to 5% in concentration. Therefore the tests were initiated for the CaS oxidation using oxygen/air at temperatures below the operating conditions of the calciner, in accordance with the conditions of the solid purge stream (identified in Figure 53 by the dotted line).

149

Figure 53: Process flow diagram of the CLP unit as applied to the coal-to-H2 process. Dotted line indicates the solid purge stream and possible treatment locations for the sulfurous species (original figure from reference 93).

150

It can be comprehended from this schematic that conditions of treatment of this purge stream for CaS oxidation are <900°C temperature and O2 concentration between 5-25%; as the upper limit on temperature is placed by the operating temperature of the calciner, and the upper limit of O2 concentration is placed by the O2 composition of the gas inlet of the calciner (gas stream coming from ASU diluted by recycled CO2).

Accordingly, the temperature range of 700-900°C and O2 concentration of 5-21% was selected for experiments.

Variation of CaS oxidation with O2 concentration and temperature

Isothermal experiments were carried out for CaS oxidation in the ranges of 700-900°C and 5-21% O2. For each experiment, the sample was ramped to desired temperature in inert N2 gas, and the gas flow was switched to desired concentrations upon reaching the isothermal conditions. The weight increase in the sample was measured as a function of time, and the reaction conversion was calculated considering the following oxidation reaction:

CaS + 2O2 → CaSO4 Rxn 15

The following Figure 54 shows the results obtained from the isothermal testing. The variation of reactivity with temperature is shown. Figure 54(a) shows the reaction conversion curves generated at various temperatures in the range considered at a fixed O2 concentration of 5%. Similarly, Figure 54(b) corresponds to a 10% O2 concentration and

Figure 54(c) indicates results obtained by using pure air (21% O2). The reaction rate is observed to be a strong function of temperature at all concentrations tested. At all

151 reaction conditions, the typical two step reaction conversion curve is observed, which becomes more prominent at higher O2 concentrations. This two-step reaction curve is consistent with the well-established mechanism of initial surface controlled reaction followed by a slower reaction rate, controlled by the diffusion of reacting gas through a product layer of CaSO4 formed on the surface. This result is in accordance with that published previously in literature (reference 100 and 97).

The two-step reaction can be highlighted further by plotting the dX/dt (rate) values against the reaction conversion, as seen in Figure 55. At lower conversion values, the reaction rates are high for all concentrations, as can be clearly seen from the Figure 55.

The inflection point occurs at progressively higher conversion values at higher temperatures.

152

Figure 54: CaS oxidation between 700 and 900°C at (a) 5% O2, (b) 10% O2 and (c) 21% O2.

153

Figure 55: reaction rate data as a function of reaction conversion X, for (a) 5% O2, (b) 10% O2 and (c) 21% O2. The graph clearly shows an initial fast reaction rate followed by a slow diffusion controlled regime of reaction.

154

According to results of Marbán et. al.97, reaction 11 is the prominent reaction below

<890°C, and this temperature is identified as the upper limit for suitability of CaS conversion to CaSO4. This temperature range has been termed as ‘regime of low SO2 release’ for the applications of pressurized fluidized bed combustors (PFBC) for the stabilization of sulfurous compounds. At these conditions, reaction 11 is the only observed reaction, which is in agreement with the results obtained in the present study.

Thermal decomposition of CaSO4 to CaO is reported elsewhere in literature at higher temperatures (~1175°C).101 The temperature range of 900-1175°C is termed as ‘regime of high SO2 release’ due to the observed formation of CaO and CaSO4 products simultaneously. However, no CaO formation was detected at the conditions tested in the present study (up to 950°C). Therefore, it is concluded that a very slim chance – if at all – of SO2 release exists in only the calciner reactor. However, the SO2 release is ruled out for the purge stream, as the temperature of this stream is slightly lower than the calciner operating temperatures.

Thus, it is concluded that the purge stream temperature must be maintained lower than

900°C to eliminate the chances of SO2 evolution by thermal decomposition of CaSO4.

Further, considering the kinetic advantage of temperature for operating the oxidation of

CaS, the reaction may be preferred to be conducted at just below the calciner operating temperature on the purge stream. Therefore, the oxidation is visualized to be carried out between the cyclone and the heat recovery steam generator unit on the purge stream. A slip stream from the gas inlet of the calciner may be introduced for the oxidation of the

155 purge stream to realize the advantage of faster kinetics obtained at higher O2 concentrations.

5.6 Commercial implications and conclusions

The objective of this work was to investigate the chemistry of the sulfur in the application of the CLP system for H2 generation from coal. The CLP system can be applied for pushing the WGS reaction forward in the H2 generation process from coal-derived syngas. In addition, the sulfurous species evolved from the coal are also captured or fixed in the solid state by reaction with the CaO sorbent. The fate of these sulfur species was investigated in this work by small lab-scale kinetic experiments, in order to eliminate one of the process uncertainties.

However, this CLP system can be applied to a wide variety of H2 and/or electricity generation methods, especially in the carbon constrained scenario. The multipollutant removal advantage of the CaO sorbent makes this a highly attractive technology. The high temperature calcination step affords the opportunity for high-quality heat recovery, and thus applying the CLP system for a coal-to H2 case necessarily results in a co- production of electricity. As stated in section 5.2, the preliminary process analysis indicates the advantage of operating the coal-to-H2 plant with the CLP system. While the base case 25.7 tph H2 power plant generates 7.2 MWe of net electric power, the CLP system applied to the identical coal-to-H2 process results in the generation of 320 MWe of net electric power.

156

The biggest byproduct of the process is the solid purge or waste stream exiting the process. The solid purge exiting the calciner for coal to H2 case is 68,675 kg/hr. This stream predominantly contains spent CaO, CaCO3 and CaSO4 along with some ash elements. This solid stream can be sold to the cement industry, potentially decarbonizing it. Since the cement industry produces clinker from limestone by calcining it, the purge stream from CLP can be used to offset this process. The energy of operating the cement

50 kiln to produce clinker can, in this manner, be potentially halved. The CaSO4 and ash present in the purge stream can also supplement the clay and silicate additives usually used to produce the cement.

This work concluded that the sulfur may exit the CLP system in the form of the solid purge stream, where it will be treated for complete oxidation to CaSO4. As stated above, the presence of CaSO4 in the purge stream makes it a viable candidate for the integration of the CLP technology with the cement industry. Therefore the results on experiments carried out on the carbonator block revealed that CaSO4 does not contribute to reducing the yield of the H2 product. Therefore, by eliminating this critical uncertainty of the process design; the present work bridges a technology gap, and presents the CLP system in a more favorable light than before.

Compared to a base coal-to-H2 plant, the CLP system uses ~53% more coal. This excess coal requirement comes from the coal-fired calciner, which requires auxiliary fuel to drive the endothermic regeneration of the spent sorbent. In addition to the extra coal being utilized by the CLP system, it also results in a ~310 MWe more coproduction of electricity over the base case. Therefore, in a carbon-constrained scenario, applying this

157 technology to produce H2 from Ohio coal would result in a significant increase in coal consumption. Pittsburgh #8 coal is the most heavily mined coal in Ohio. Although these simulations were performed on the Illinois #6 bituminous coal according to US-DOE’s guidelines, the results can be easily applied to Ohio’s Pittsburgh #8 coal, owing to the similar sulfur content of the two coals. However, to determine the increased use of Ohio coal by application of the CLP system along with reduced emissions, a rigorous analysis should be conducted on a plant-by-plant basis.

Since the application of CLP system to a coal-to-H2 plant results in a co-production of electricity from the heat recovery steam generation system, establishing a CLP-enabled coal-to-H2 plant is considered similar to a combined H2 plant with a pulverized coal fired boiler of similar commercial size. Since this process is developed for the purpose of higher efficiency generation of H2 from coal, the cost of individual pollutant removal such as sulfur/mercury/ash has not been calculated per ton of coal or per ton of pollutant removed. However, the use of CLP system results in removal of several process units of pollutant removal/gas cleanup by combining them in the calcium looping system. The following figure shows the base case coal-to-H2 plant with conventional sulfur removal blocks highlighted.

158

Figure 56: Base case coal-to-H2 plant with 2-stage Selexol and Claus plant for sulfur removal93

159

The first year comparison of cost of H2 (COH) and cost of electricity (COE) between the base and CLP plant is given below. It should be noted that these calculations are from the analysis performed by Connell et. al.93 in 2013, and are based on the then-current prices of natural gas and coal. However, even with the assumption of high natural gas pricing of

$6.21/GJ, it was concluded that the COH from the coal to H2 CLP plant is still significantly higher than for H2 produced from steam methane reforming (SMR).

Table 11: Cost comparison for H2 and electricity generation for coal to H2 plant, base case and CLP plant

Coal to H2 base Coal to H2 CLP plant plant

First year cost of H2 ($/ton of H2) $3150 $2770 First year cost of electricity ($/MWh) $105 $92.07

160

The SMR method of producing H2 may be modified by applying the CLP system to further improve the COH, by operating the calciner in a coal-fired configuration. The sulfur compounds in such a hybrid system may only enter the solid circulation loop via the oxycombustion of coal in the calciner. Thus the formation of calcium sulfide under reducing conditions is not anticipated in such a system.

From the technical standpoint, several key technology gaps need to be addressed next to propel the CLP system for H2 production towards commercialization. Work has already been initiated to probe the kinetics of the hydration reaction; as reactivation by hydration is a critical component in the success of CLP system which enables it to maintain low solid circulation rates. A high temperature steam hydrator reactor design is yet another technology gap that needs to be bridged. Solids handling and particulate removal efficiencies at high temperatures amenable to the CLP system also need to be studied for the overall success of this technology.

With the US-EPA’s proposed regulations for reducing CO2 emissions, the continued development of technologies such as CLP, which can achieve carbon abatement in addition to multipollutant cleanup, is essential.102 By eliminating the key technology gaps through rigorous scientific studies, the development of the CLP system can be strengthened, and a greater confidence may be instilled in future scale up efforts. With the backing of both government policies and scientific research initiative, the successful development of such technologies may be secured.

161

CHAPTER 6: Chemical Looping Applications: High Pressure Redox

Behavior of Iron-Oxide Based Oxygen Carriers

Reproduced with permission from Deshpande, N.; Majumder, A.; Qin, L.; Fan, L.-S.

High-Pressure Redox Behavior of Iron-Oxide-Based Oxygen Carriers for Syngas

Generation from Methane. Energy Fuels 2015, 29 (3), 1469–1478. Copyright [2015]

American Chemical Society.

6.1 Introduction

Transition metal oxides are one of the most technologically versatile materials that have found their applications in various fields. In electronics, they are used in making conductor and semiconductor materials.103 In electrochemistry, they have applications in solid oxide fuel cells, lithium-ion batteries etc.104,105 But perhaps they are the most widely used in the chemical industry as catalysts, catalyst precursors and oxygen donors in various processes.106 Selective oxidation, dehydrogenation107 and chemical looping108 constitute some of the most important processes that are based on the reduction-oxidation

(redox) properties of the transition metal oxides. In these processes, lattice oxygen from the metal oxides participates in the reaction while the vacancies left behind are replenished by molecular oxygen. The redox behavior of the metal oxides influences their crystal phases and their morphologies and consequently their optical, electrical and chemical properties.

162

In the chemical industry, partial and selective oxidation processes often need to be operated at conditions suitable for the downstream product applications and the process economics. For example, in the recent past, a portion of the chemical industry has shifted its focus towards natural gas or methane (CH4) for the synthesis of valuable chemicals like gasoline, MTBE, alcohols and oxygenates through such oxidation processes.109 A number of important processes like methanol, ammonia and Fischer-Tropsch synthesis, which are used to synthesize these valuable chemicals, use syngas as their feedstock.110

Syngas for these processes is preferably derived from CH4 because of the lower capital costs and the higher efficiency of the CH4 to syngas conversion systems as compared to coal derived syngas. CH4 is preserved in natural gas fields at high pressures. Also processes like Fischer-Tropsch synthesis and methanol synthesis, which use syngas as their feedstock, operate at elevated pressures between 2-4 MPa.111,112 Thus for the conventional syngas generation at ambient pressure, the syngas needs to be compressed prior to being introduced in the system. Hence, it is economically beneficial to carry out syngas generation processes at pressures compatible with downstream applications in order to minimize the energy losses associated with compressing the syngas feedstock for these processes. Studies on syngas generation at elevated pressures are, however, very limited. Gas-to-liquid (GTL) processes, like synthesis of gasoline, diesel and methanol, also require a hydrogen-rich syngas feed with a 2:1 ratio of hydrogen: carbon monoxide

113 (H2:CO). Existing syngas generation processes like steam methane reforming, autothermal reforming, and catalytic partial oxidation of methane are unable to achieve the required syngas quality in a single unit and need additional processing steps.110 Thus, the single step partial oxidation of CH4 over metal oxides offers an attractive alternative

163 to the existing CH4-to-syngas conversion methods, and the concept is widely utilized in the process of chemical looping reforming. The economic advantages, and the limited knowledge, of high-pressure partial oxidation of CH4 necessitate a comprehensive study of the impact of elevated pressures on the reactions involved.

Chemical looping has been regarded as one of the most promising technologies in the U.

1 S. Department of Energy’s CO2 capture roadmap. The technology is based on high temperature cyclic redox reactions of metal oxide based oxygen carriers between two or more reactors – reducer, combustor and oxidizer - for the conversion of carbonaceous

114 fuel to generate electricity and/or H2. It is designed to produce a sequestration-ready stream of CO2 from the reducer. So far, the chemical looping process has been used to

115,116,117 generate H2 and/or electricity using syngas, coal and biomass as feedstock.

Nevertheless, it is a highly versatile process and can be used for syngas generation using natural gas/CH4 as its feedstock. CH4 is converted to syngas in the reducer via oxygen transferred from the metal oxide based oxygen carriers. Such a process was conceptualized at the Ohio State University (OSU) for the utilization of shale gas, termed as the Shale gas to Syngas (STS) Process.118 The schematic of this process is seen in

Figure 57. Syngas with a 2:1 H2: CO ratio can be obtained by controlling the oxygen carrier circulation rate in the system, and thereby the extent of reduction of the oxygen carriers in the reducer. This process has been demonstrated experimentally at various scales at atmospheric pressure.118 Therefore, syngas generation via chemical looping can be developed into an efficient, economic and environment friendly process that overcomes the issues associated with the existing processes. However, experimental

164 kinetic investigations for the effect of pressure on this system are relatively sparse. Most existing studies on CH4-to-syngas conversion using metal oxides are focused on Nickel

(Ni) based complex oxides, due to its catalytic capabilities for the reforming

111,119,13 reaction. Some of the other metal oxides studied for CH4-to-syngas application include copper (Cu), iron (Fe), and manganese (Mn).120,121 This study has been conducted from the perspective of such redox systems using iron (Fe)-based oxides, which have the potential to be more economical if operated at elevated pressures. The bimetallic system has been investigated at OSU for the STS process in the form of iron-titanium complex metal oxides (ITCMO) particles designed for this process. The operating conditions of

STS process have been determined through detailed thermodynamic and process analysis.118 Therefore, it is essential that the ITCMO particles be tested for the effect of pressure on the reactions rates. The present study is intended to accomplish this objective.

Developing the solid oxide based redox system for high pressure partial oxidation of CH4 is a multi-optimization problem, which requires careful manipulation of each operating parameter to maximize performance, and to reduce the overall cost. These parameters include (but are not limited to) gas to solid loading ratio, choice of reactor operation such as moving bed vs fluidized bed, co-current and countercurrent gas solid flow, precise control of the gas and solids residence times in each reactor to obtain the desired oxidation state of the solids. Each of these parameters is equally crucial and deserves separate attention and in-depth analysis. Nevertheless, in this study, the reduction and oxidation kinetics of ITCMO oxygen carriers, developed at OSU, have been studied at

122 pressures ranging from 1-10 atm using H2/CH4 for reduction and air for oxidation.

165

Change in reaction kinetics may influence the reducer sizes, the processing capacity and consequently, the process economics. The purpose of this study is to demonstrate the effect of pressure on the reaction rates of the ITCMO particles for CH4-to-syngas conversion. Although the reducing environment of interest is CH4, H2 has been used as the reducing gas for a major part of the study as H2 reduction reaction is well understood and relatively easier to operate. It provides a clear understanding of the kinetics without the interference of coking, which is observed with CH4 as the reducing gas. Furthermore, the effect of pressure that is discussed in case of H2 can be extrapolated to other reducing environments. The results presented are that of kinetic experiments carried out in a thermogravimetric apparatus. It has been demonstrated in this work that higher pressures are kinetically favorable for H2 and CH4 reduction and to a lesser extent air re-oxidation of the metal oxides. This work also briefly discusses the advent of coking and its response to elevated pressures.

166

Figure 57: Schematic of Fe-oxide based system for syn-gas generation from partial oxidation of CH4

167

6.2 Thermodynamic analysis

The thermodynamic analysis of the reducer was conducted using HSC Chemistry

(OutoKumptu Research Oy, version 6.0). As stated earlier, the process of partially oxidizing CH4 for production of syngas using the oxygen carrier particles causes coking or C soot formation123. Operating the system at elevated pressures exacerbates this condition. To understand the thermodynamic equilibrium limits of operating this system at elevated pressures and its effect on the soot formation, the reducer species were simulated via Gibbs free energy minimization at elevated pressures (up to 10 atm). The

Fe-Ti bimetallic system is used for simulating the ITCMO oxygen carrier particles.

Various metal oxides are widely studied in literature for partial oxidation of CH4 for syngas production. Of these, the single metal oxides such as Ce, Ni, and Fe oxide systems were selected for their comparison with the ITCMO system. A temperature sensitivity analysis was performed for these systems. The parameters studied were syngas purity, and overall CH4 conversion. The syngas purity is defined as

푚표푙푒푠 표푓 퐶푂 + 푚표푙푒푠 표푓퐻 푆푦푛푔푎푠 푃푢푟𝑖푡푦 = 2 푚표푙푒푠 표푓 푎푙푙 푔푎푠푒표푢푠 푠푝푒푐𝑖푒푠

The temperature range selected was 900 to 1100°C as applicable to a partial oxidation system operation. An equimolar ratio of active metal oxide to CH4 was simulated.

Pressures of 1 and 10 bar were used. The results are shown in Figure 58.

Ce-CH4 system shows excellent syngas purity as well as high CH4 conversion (Figure

58a), making Ce-based systems thermodynamically most suitable for partial oxidation

168 applications. However, the slow kinetics of the Ce-oxides coupled with the high cost inhibits further development of Ce-based oxygen carriers for commercial applications.

The Ni-based system exhibits lower syngas purity as compared to the Ce system, although the overall CH4 conversion is high (Figure 58b). As compared to Ce and Ni systems, the Fe-based systems show superior CH4 conversions at the conditions employed here. The syngas purity as well as CH4 overall conversion is negatively affected by increase in pressure, in accordance to le-Chatelier’s principle for the volume expansion reactions of partial and complete oxidation of CH4. The CH4 conversion increases with temperature for all the systems, at both pressures considered here. For the

Fe and Fe-Ti system (Figure 58(c) and (d)), the syngas purity drops with increase in temperature, but the total CH4 conversion goes on increasing. In other words, at higher temperatures, complete oxidation of CH4 to CO2 and H2O is more favorable than at lower temperatures. The syngas purity obtained in the pure Fe system (Figure 58c) is much lower than that of Fe-Ti system (Figure 58d), indicating a higher propensity of complete oxidation of CH4 in presence of Fe-oxide alone.

169

Figure 58: Effect of temperature on the methane conversion and syngas purity for different metal oxide systems. (a) CeO2, (b) NiO, (c) Fe2O3 and (d) ITCMO system

170

Additionally, the ITCMO system was studied at isothermal and isobaric conditions.

Conversion of CH4 to syngas is carried out by partial oxidation using hematite (Fe2O3) as the reactive phase from the ITCMO oxygen carriers. The titanium oxide (TiO2) phase is assumed to be non-reactive. In its simplest form, the theoretical desirable reaction is

1/3 Fe2O3 + CH4 → CO + 2H2 + 2/3 Fe Rxn 24

Thus CH4 is partially oxidized to form syngas, a mixture of CO and H2, and the Fe2O3 is completely reduced to elemental Fe. However, thermodynamically, the reduction of

Fe2O3 goes through the progressively different reduced phases of iron oxide, namely magnetite (Fe3O4), wüstite (FeO), and finally the completely reduced form of elemental

Fe. All of these phases are likely to be present. Similarly, along with the formation of H2 and CO, CH4 oxidation also results in complete combustion (to CO2 and H2O) as well as formation of elemental C. Accordingly, the reactive system was simulated with the following species in the gaseous state: CH4, H2, CO, CO2, H2O, and following species in solid state: C, Fe, FeO, Fe3O4, Fe2O3, Fe2TiO5 and FeTiO3. Isothermal and isobaric systems were simulated at 950 ˚C, and 1, 5, and 10 atm. The solid loading was assumed to be in the forms of fully oxidized Fe and Ti metallic species. The gaseous input of CH4 was incrementally added, and the outlet species were analyzed for solid and gas equilibrium compositions at minimum Gibbs free energy, at fixed T and P values.

As expected, TiO2 remains largely unreacted while compounds of Fe undergo sequential reduction with increasing CH4 loading, from fully oxidized to fully reduced form, both in pure Fe-O as well as Fe-Ti-O complex phases. The equilibrium amount of FeTiO3 and

171

Fe2TiO5 are found to be negligible, and therefore for simplicity, only pure Fe-O phases are considered for the remainder of this discussion.

For all pressures, overall CH4 conversion was found to be >99% for the range of gas: solid ratios tested, which was varied between 0.05 to 1.5 of moles of CH4 per mole of

Fe2O3. This range was chosen due to the fact that all four oxidation states of Fe are found to exist in this range. The conversion was found to increase as the gas: solid ratio was increased. As expected, solid C formation is observed simultaneously with the formation of elemental Fe. This C amount is higher at higher pressures. This is in agreement with our experimental findings, discussed further in section 6.4.2. For example, at the

CH4:Fe2O3 ratio of 1.5, comparison of the C formed at 10 atm and 1 atm reveals that the equilibrium C amount at 10 atm is approximately 8 times that of 1 atm. The same comparison between 5 and 1 atm shows that equilibrium C formation at the same ratio is

4.5 times that of 1 atm. The C deposition is shown as a function of gas-solid ratio in

Figure 59 at 5 atm and 950 ˚C. The figure clearly shows the simultaneous onset of elemental Fe formation and C deposition.

172

Figure 59: Simulated equilibrium iron oxide phases and fractional carbon deposition as a function of inlet gas:solid ratios at elevated pressure. T = 950 ˚C, P = 5 atm.

173

In addition, this thermodynamic analysis reveals that increase in system pressure from 1 to 10 atm results in an increase in the formation of CO2 and H2O, along with a slight decrease in the formation of desirable H2 and CO, as well as overall CH4 conversion.

However, increase in system pressure is found to have a favorable impact on equilibrium

H2:CO ratio, which is desired ~2 for downstream processing such as Fischer-Tropsch synthesis. The carbon deposition can be managed by careful manipulation of the gas solid ratios in the moving bed reducer reactor system.118

The HSC Chemistry thermodynamic software (version 6.0, Outokumptu Research Oy) was used to calculate the equilibrium composition of the heterogeneous reaction mixture.

The Gibbs free energy minimization technique is used to calculate the composition of all reactants and product species.

Equilibrium composition of the specified system is a function of temperature, pressure, and the relative ratios of raw/starting materials. If complete reduction of the metal oxide particle is considered, 1 mole of Fe2O3 oxidizes 3 moles of CH4 according to reaction 24 above.

Due to the various oxide species of Fe, different gas compositions are in equilibrium with the solid at the reaction conditions at different Fe2O3 inlet loading. The goal is to maximize the amount of conversion of methane while using minimum loading of Fe2O3.

This optimum conversion of methane will achieve close to 2:1 ratio of H2 and CO in the syngas product.

The following results in Table 12 are obtained for optimum conversion of CH4.

174

Table 12: Methane optimum equilibrium conversion results for partial oxidation using ITCMO particles

Pressure

1 atm 10 atm 15 atm

900°C 99.986% 98.289% 96.833% Temperature 950°C 99.995% 99.388% 98.849%

The gas composition curves obtained display a sequential series of plateaus and slopes, which are characteristic of the different iron oxide compounds that are in equilibrium with the system at that point. For example, two such graphs are shown in below in Figure

60 and Figure 61.

175

2 100.002%

1.8 100.000%

O 2 1.6 99.998%

1.4

, , H and 99.996%

2 1.2 H2O(g) , , H

2 99.994% 1 H2(g)

99.992%

conversion

0.8 4 CO2(g)

99.990% CH 0.6 CO(g) 0.4 99.988% CH4 conversion moles of CO, CO 0.2 99.986% 0 99.984% 0 5 10 15

Fe2O3 to CH4 molar ratio

Figure 60: Gas composition profiles and methane conversion at 900C, 1 atm with respect to loading of Fe2O3

2 100.2%

1.8 100.0% O 2 1.6 99.8%

1.4 99.6% , , H and

2 99.4%

1.2 H2O(g) , , H

2 99.2% 1 H2(g)

99.0%

conversion

0.8 4

98.8% CO2(g) CH 0.6 98.6% CO(g) 0.4 98.4% CH4 conversion moles of CO, CO 0.2 98.2% 0 98.0% 0 2 4 6 8 10 12 14

Fe2O3 to CH4 molar ratio

Figure 61: Gas composition profiles and methane conversion at 900C, 10 atm with respect to loading of Fe2O3

176

Increasing the amount of metal oxide increases overall conversion of methane, however the complete combustion products such as CO2 and H2O also increase due to availability of excess oxygen, including but not limited to, the following reactions:

CO + Fe2O3 → CO2 + 2 FeO Rxn 25

CO + FeO → Fe + CO2 Rxn 26

H2 + Fe2O3 → H2O + 2 FeO Rxn 27

H2 + FeO → H2O + Fe Rxn 28

Also, according to Le Chatelier’s principle, the reactions such as (2) to (5) are favored at higher pressures over reaction (1) and its modifications. Therefore, the optimum conversion of CH4 decreases with increase in pressure.

Nevertheless, the success of this partial oxidation system rests equally on the kinetic factors affecting the reaction. The high pressure operation of Fe-based partial oxidation will require the basic understanding of the manner in which pressure affects the kinetics of each of the reactions involved. Therefore, in the following sections, the effect of pressure on the reduction and oxidation of Fe-Ti oxygen carriers is investigated.

177

6.3 Experimental setup, materials and procedure:

A magnetic suspension balance (MSB, Rubotherm GmbH, US-2004-00162) was used for the high pressure thermogravimetric analysis (TGA) experiments. The schematic of the

MSB setup is shown in Figure 39 in Chapter 5, section 5.4.2. Different pure gas bottles

(H2, CH4, N2, and Air) were connected through a battery of mass flow controllers and valves to the TGA assembly. A pressure transducer upstream of the sample cell measures and records the pressure of the sample during the experiment. Downstream of the sample cell, a back pressure regulator (BPR) is installed to regulate the pressure in the sample cell. The gas was preheated prior to entering the TGA assembly by means of heating tapes. The unique working principle of the MSB allows the sample weight and the balance to be connected via magnetic coupling, and therefore the balance is isolated from

(and not affected by) the reaction environment. This principle allows the use of high pressure and highly corrosive environments in the sample cell. The section of the TGA housing the magnetic coupling is maintained at 140oC by means of a heat jacket connected to an oil bath. The reaction temperature in the sample cell is maintained independently by means of an electric furnace.

178

The gas mass flow rates were fixed such that the gas space velocity experienced by the sample was constant at all pressure experiments (~1 min-1 for the volume of the sample cell). The sample was heated in the inert flow of N2. When the reaction pressure and temperature were achieved, the gases were switched to introduce the reaction mixture in the sample cell. The sample weight, temperature and pressure were recorded as a function of time.

Knowing the composition of the oxygen carrier particle, the theoretical maximum weight loss (complete reduction) and gain (complete oxidation) was calculated. These numbers were used to calculate the extent of each reaction. Specifically, for each sample the

푟푒푑 theoretical maximum weight loss during reduction (∆푤푚푎푥) is calculated on the basis of the active weight of the sample. During the experiment, the weight change due to reduction (∆푤푟) at any instant is then used to compute the extent of reduction by the following formula:

∆푤푟 푒푥푡푒푛푡 표푓 푟푒푑푢푐푡𝑖표푛 = 푟푒푑 푥100% ∆푤푚푎푥

Similarly, in case of oxidation,

∆푤표 푒푥푡푒푛푡 표푓 표푥𝑖푑푎푡𝑖표푛 = 표푥 푥100% ∆푤푚푎푥

Where

∆푤표= weight change due to oxidation

179

표푥 ∆푤푚푎푥= theoretical maximum weight gain during oxidation

Through the remainder of this article, a general term X is used to indicate the reaction conversion, or extent of reaction for reduction and oxidation alike, as applicable to the discussion. The X values were plotted for each experiment vs time (minutes), to obtain the thermogravimetric conversion curves. These curves were then used to compute the instantaneous rates of reaction, denoted by dX/dt, per minute.

Samples of ITCMO particles (0.1 g) were collected after selected experiments and mechanically crushed to a powdered form and sieved to the appropriate particle size. X-

Ray Diffraction (XRD) was performed on the powdered specimens with a Rigaku

SmartLab X-ray diffractometer. Additionally, specimens were also examined by

Scanning Electron Microscopy (SEM) using e-beam imaging of an FEI Helios

NanoLab600 DualBeam system. To obtain high quality cross-section imaging, focused ion beam (FIB) was generated using gallium ion source at an accelerating voltage of 30 kV for cross-sectional milling for in-situ SEM observation. The 2-D material mapping was obtained using Oxford Energy Dispersive X-ray Spectrometry (EDS) at an accelerating voltage of 20 kV. Additionally, a NOVA 4200e Quantachrome Brunauer-

Emmett-Teller (BET) analyzer was used to measure the pore volume and surface area of samples using N2 sorption.

6.4 Results and discussion

To study the effect of pressure on the rates of reactions involved in partial oxidation system, isothermal experiments of reduction and oxidation were conducted in reducing

180 environment of H2 and CH4, and the oxidizing environment of air. The rates of reactions were then compared by calculating the rates from the solid conversion curves obtained from the TGA.

6.4.1 Reduction in H2

The reduction of the oxygen carrier samples was conducted isothermally at 900°C using

H2 as the reducing agent, at various operating conditions of gas concentration and partial

pressure. The parameters studied for H2 reduction include partial pressure (푃푃퐻2), total

system pressure, and mole fraction of H2 (푌퐻2).

The weight loss of the sample corresponds to the total amount of oxygen lost by the sample during reduction. Therefore, the reduction conversion is calculated assuming

100% reduction at complete weight loss of sample, or the most reduced state of Fe. In this manner, reduction conversion was calculated and plotted. The rates of reaction are calculated graphically calculated at X = 0.5 and 0.75 reduction conversion values at the various conditions tested, and exhibit similar trends.

6.4.1.1 Constant partial pressure of H2 (푷푷푯ퟐ)

The isothermal isobaric experiments were conducted to observe the rates of reaction at different total gas pressures, with constant partial pressure of reducing agent. The TGA conversion curves were compared at different total pressures of the system, at three

different partial values of 푃푃퐻2 of 1, 1.5 and 3 atm. At a constant value of 푃푃퐻2, the

increase in the overall system pressure resulted in decrease in mole fraction of H2 (푌퐻2).

The rate of reaction is found to decrease with increase in total pressure of the system at

181 each value of constant 푃푃퐻2. Similar results have been reported in the past on metal oxide reduction reactions. For example, Garcı´a-Labiano et al. have reported decrease in

reduction reaction rate with an increase in system pressure at 푃푃퐻2 and 푃푃퐶푂 values of 1 atm for Fe, Cu, and Ni based metal oxides.124 The negative effect of pressure on various reactions has been previously observed by other researchers, and explained by factors such as increase in product gas volume upon reaction, or increased diffusion resistance through the product layer at higher pressures.125,126,127 Thus, the same set of data is plotted in two different ways. In Figure 62 the rates of reaction are plotted against the total system pressure, and in Figure 63 the same rate values are plotted against the

respective 푌퐻2 values. For example, comparing the experiments conducted at system pressure of 3 atm, the three different experiments at this pressure value would correspond

to the three different values of 푃푃퐻2 of 1, 1.5 and 3 atm. The rates of reduction for these three experiments fall on a vertical straight line of Figure 62, at x axis value of 3 atm. At

this system pressure of 3 atm, the mole fractions 푌퐻2 are, however, widely different.

푃푃퐻21 atm at a system pressure of 3 atm results in a 푌퐻2 value of 33%. 푃푃퐻2 = 1.5 atm at

the system pressure of 3 atm is at a 푌퐻2 = 50%. And finally, at 푃푃퐻2 = 3 atm, 푌퐻2 is obviously 100%. This is true for all the values of pressure along the x-axis of Figure 62,

i.e. the curve for 푃푃퐻2 = 3 atm is always at highest mole fractions of the reducing gas

(푌퐻2), which seems to be a crucial factor contributing to the superior rates of reaction observed. The entire set of experimental conditions is given in Table 13.

182

Figure 62: The effect of total system pressure on rates of reduction at X = 0.75, and constant partial pressure of H2. T = 900 ˚C

183

Table 13: YH2 as a function of total system pressure and PPH2 for section 6.4.1.1

푷푷푯ퟐ (atm)→ 1 1.5 3 Total Pressure (atm) ↓ 1 100% - - 3 33.33% 50% 100% 5 20% 30% 60% 7 - - 42.86% 8 12.50% 18.75% 37.50% 10 10% 15% 30%

184

The curves of rate vs system pressure seem to converge at higher pressure. This can be

explained by the fact that the 푃푃퐻2 values chosen here are relatively low, and therefore at

high system pressures the 푌퐻2 values are closer. By contrast, at lower system pressure, the

rate curves are further apart which correspond to the disparity in 푌퐻2 values at those conditions (see Table 13), which increases at lower pressure.

In Figure 63, the same data is plotted against mole fraction of H2 (푌퐻2). Following the

same logic, the rates can be compared at similar values of 푌퐻2. In the vicinity of 푌퐻2 = 30-

33%, even though the mole fraction of H2 is so similar, the system pressure values for the

three curves are 3, 5 and 10 atm. Therefore, it is observed that the higher 푃푃퐻2 values expectedly play a part in increasing the rate of reduction. When the rate values are plotted

in this manner as a function of 푌퐻2, the plots are linear; indicating a direct proportionality between the rate and mole fraction of reacting gas. However, the plots converge at lower

푌퐻2 values (which correspond to the high pressure experiments, towards the positive x- axis direction in Figure 62). Regardless of the manner of analysis, it is plainly seen that at higher values of constant partial pressure of the reducing gas a higher rate of the reduction of solid oxide ITCMO particles is achieved.

185

Figure 63: The effect of mole fraction of reducing gas (YH2) on rates of reduction at X = 0.75, and constant partial pressures. T = 900°C

186

6.4.1.2 Constant mole fraction of H2 ( 풀푯ퟐ)

Similarly, the reduction experiments were conducted at various pressures between 1 and

10 atm by keeping 푌퐻2 constant at 50% to study the rate of reduction of the ITCMO

particles at the varied pressures. In this case, the value of 푃푃퐻2 also inevitably increased with the increase in pressure. It is observed that as the pressure is increased, the slope of the conversion curves increases, indicating higher reaction rates. This is in contrast to the previously reported findings of Garcı´a-Labiano et al, who report a slight decrease in reaction rates with increase in pressure at constant mole fraction of CO at 10% value.124 It must be noted in the outset that the rate of reaction is determined by a combination of various factors, such as reactive gas partial pressure, change in diffusivity of reactant and product gas through the porous particle at elevated pressures, relative superficial velocity of gas with respect to solids, etc. As stated before, in the present study, a constant gas linear velocity (space velocity) was used for all experiments. The rates were determined graphically at a fixed conversion value for all the curves, and are plotted in Figure 64 at

X = 0.5 and at X = 0.75. From Figure 64 it can be concluded that there is more than a

100% increase in reaction rate as the pressure is increased from 1 to 10 atm, when operating at the same mole fraction of the reducing gas, namely, H2.

187

Figure 64: The effect of sysyem pressure on rates of reduction at X = 0.5 and X = 0.75, and constant mole fraction of reducing gas YH2 = 50%, T = 900 ˚C

188

6.4.1.3 Constant pressure of the system

Finally, the reduction reactions were studied for the reaction kinetics under constant pressure. For this set of tests, the pressure of the system was maintained at 5 atm and the

푌퐻2was increased from 30% to 100%, thus increasing partial pressure of the reducing gas

(푃푃퐻2). If the reaction is conducted under fixed pressure, the rate of reaction is found to

increase as expected, with an increase in 푃푃퐻2; this is shown in Figure 65.

Figure 65: The effect of partial pressure of reducing gas PPH2 on (a) conversion curves obtained and (b) rates of reduction at constant system pressure P = 5 atm. T = 900°C

189

6.4.2 Reduction in CH4

The same experiments as section 6.4.1 of this chapter were repeated with CH4 as the reducing gas instead of H2. As indicated in section 6.2, it was observed that the reduction of oxygen carrier particles with CH4 as the reducing agent results in the formation of elemental C. This is evidenced by the soot formation and weight increase of the sample beyond a certain point. The C deposition is observed when the oxygen carrier reaches a certain degree of reduction conversion, and is always after elemental Fe phase has been formed. Thus, in order to study the reaction kinetics of the reduction of oxygen carrier materials in presence of CH4, the reaction is arrested at or before the initiation of C deposition. Accordingly, experiments were conducted by adjusting the procedure and allowing the maximum possible reduction of particles, till the onset of C deposition.

190

Figure 66: Reduction conversion curves obtained using CH4 from the thermogravimetric analysis between 1 and 10 atm at constant mole fraction of reducing gas YCH4 = 50%. T = 950 ˚C

191

The reduction of Fe2O3 to Fe proceeds through sequential steps of various oxidation states. Here, complete loss of oxygen is considered as 100% conversion. Stage I corresponds to Fe2O3 to Fe3O4 conversion, which translates to 11% reduction conversion

(or X = 0.11). Stage II corresponds to Fe3O4 to FeO conversion, which translates to 33% reduction conversion (X = 0.33). At conversions higher than 33%, the stage III is initiated

120 which results in the formation of elemental Fe. Unlike reduction in H2, in case of CH4 reduction these three stages have three distinct reaction rates as seen in Figure 66.

Further, the different stages react differently to increase in pressure in terms of the rate of reaction. The rate of each reaction stage is studied at various pressures between 1 and 10 atm.

At higher pressures, the rate disparity between the three stages is less pronounced, giving a faster overall conversion obtained without three distinct rate stages. It was also observed that the reaction halted at lower conversions, owing to the higher amount of C deposition.

These conversion curves were used to compute the reaction rates at different pressures.

The evaluation of three separate rate values is warranted for the three distinct stages of reduction. Accordingly, the rate values were calculated from the conversion data obtained. These rate values are plotted in Figure 67. It can be appreciated that reaction rates for stages I and III go through a maxima in the range of the pressure tested here.

However, the rate of stage II increases exponentially with pressure. Since stage II is the slowest reaction stage, it is the overall rate determining step and therefore any change in the rate of stage II overwhelmingly affects the overall rate of the reduction reaction. For

192 example it can be seen from Figure 66 that at 10 atm, 33% reduction is achieved in almost 1/7th of the time taken at 1 atm; and similarly, 60% reduction is achieved in 1/3rd of the time.

193

Figure 67: The effect of system pressure on reaction rate for the three-step reduction with

CH4 as the reducing gas. YCH4 = 50%, T = 950°C, P = 1 to 10 atm

194

Increase in pressure, specifically at constant mole fraction of the reacting gas, results in increase in concentration of the active species in the gas phase. This increased concentration of gaseous species inevitably results in increased reaction rate, which is also reported earlier in section 6.4.1.2 (for H2). Specifically, the role of pressure in the reaction rate is observed to be particularly pronounced in case of reduction of ITCMO particles with CH4. The effect of pressure on reduction kinetics for chemical looping combustion (CLC) scenario has previously been investigated by Garcı´a-Labiano et al.124

They studied the reduction kinetics on Fe, Cu and Ni based oxygen carriers at pressures up to 30 atm. At a low constant mole fraction of 10% of H2 and CO, no significant increase in reaction rate was observed with an increase in pressure. However, it was noted that the actual reaction rate was influenced to a certain degree by several factors, of which an important factor was ‘gas dispersion’ that occurs particularly during the initial stage of the reacting gas introduction to the sample cell. The use of a constant molar flowrate across all pressures resulted in a progressively increased gas dispersion effect in the reaction cell at increased pressures. To minimize the progressive gas dispersion effect due to an increase in pressures in the present study, the ITCMO particles were reduced at constant space velocity across all pressures. The use of the constant space velocity leads to an increase in reaction rate of the reduction reaction of the ITCMO particle with increased pressure, as seen in Figure 67.

To explain the effect of gas-dispersion, Figure 68 can be seen. This figure shows the evolution of reduction conversion of the ITCMO particles between 1 and 10 atm, with constant space velocity and with constant molar flowrate. The dotted line indicates

195 reduction conversion curve at space velocity same as that at 1 atm. The dashed line denotes the reduction conversion curve at molar flowrate same as that at 1 atm. If space velocity is held constant, the advantage of operating the system at elevated pressure can be easily verified (comparison of the solid line and dotted line). On the other hand, if the molar flowrate is held constant, the kinetic advantage is not apparent due to the gas- dispersion or end-effect. At constant molar flowrate, a fixed reactor volume equilibrates over longer duration of time at higher pressure. Therefore, at the beginning the sample

‘sees’ much lower concentration of the active species than the setpoint. This results in a tradeoff between the positive effect of higher pressure, and the negative effect of lower concentration of the active species. In such experiments, the difference in reaction rates is found to diminish significantly, confirming that the negligible effect of increased pressure on reduction rate observed in reference 124 is attributed to the ‘gas dispersion’ effect with pressures.

196

Figure 68: Effect of gas dispersion in the reduction kinetics of ITCMO particles in presence of CH4.

197

6.4.3 Pressure correction

The sample weight measurement in the MSB is extremely sensitive to pressure changes in the cell. The system pressure is regulated by the BPR situated downstream of the sample cell. Usually, this BPR enables the system to be maintained at steady pressure value for all the experiments. Thus, all the changes in the sample weight measurement are

‘true’ weight changes, aka attributed to the reaction alone. However, the reduction reaction in presence of CH4 is a volume expansion reaction, where one mole of reactant gas is converted to three moles of product according to the reaction 24 above (other side reactions, such as complete combustion, also result in volume expansion).

The rate of this volume expansion is directly proportional to the rate of the reaction. In the fixed volume of the sample cell of MSB, rapid gas volume expansion thus results in a temporary increase in system pressure. The faster reaction results in more volume expansion, and thus has a more pronounced effect on the temporary system pressure change. The three stages in the CH4 reduction reaction occur at different reaction rates as shown in section 6.4.2. The difference in these rates is also reflected in the trend in the system pressure. During stage I, the pressure increases due to volume expansion (faster reaction = more pronounced effect on pressure). At the onset of stage II, there is sudden appreciable decrease in the rate of reduction reaction, and therefore the pressure buildup starts dissipating by returning to the setpoint value (and further shows a slight increase after stage III initiation). These pressure changes in the system inevitably affect the measurement of sample weight, causing it to change. Therefore, in order to discern the

198

‘true’ weight change of the sample due to reaction alone, it becomes necessary to correct the measured sample weight for fluctuation due to pressure effect.

Thus, a correlation was established between system pressure and sample weight in absence of reaction, by changing the pressure setpoint externally in inert gas and recording the changes in sample weight. This correlation was applied to the measured sample weight during reduction reaction to correct the weight to reflect reaction alone.

Figure 69 shows the example of this correction applied to compute the conversion.

Figure 69: Pressure data and Reduction conversion obtained based on the original data and the data obtained by applying pressure correction. Reducing gas = CH4 with YCH4 = 50%, T = 950°C, and P = 8 atm.

199

6.4.4 Air oxidation

The effect of pressure was also studied for oxidation reaction for the combustor block discussed in section 6.1. The oxidation reactions were carried out using air as the oxidizing agent. The sample ITCMO particles were subjected to reduction in H2 followed by air oxidation at constant pressure values (1, 5 and 10 atm). The oxidation conversion curves are compared in Figure 70. The increasing trend of reaction rate with pressure is apparent with air oxidation too, although it is not as pronounced as reduction reactions, mainly due to the fact that oxidation reaction with O2 is extremely fast to begin with, with the total oxidation completed in less than 8 minutes even at ambient pressure at the conditions tested here. This slight increase in oxidation rate with pressure is in agreement with a similar rate increase observed by Jin and Ishida in case of Ni-based oxygen carrier particles.26 In case of two different Ni-based particles, effect of pressure was reported to be more pronounced on reduction rates than oxidation, when H2 was used as reducing gas and O2 (air) was used as oxidizing gas, between 1 and 9 atmosphere.

200

Figure 70: Oxidation conversion curves obtained from the thermogravimetric analysis between 1 and 10 atm at YO2 = 0.1, T = 900 ˚C

201

6.4.5 XRD, SEM, EDS, and BET analysis

The samples were subjected to XRD analysis after complete reduction (H2) and oxidation, at the lowest and highest pressure tested, viz. 1 atm and 10 atm. Complete reduction of sample is unattainable in CH4, therefore, H2 was used as the reducing medium. The diffraction spectra are shown in Figure 71. The samples after reduction exhibit complete reduction in the iron oxide phase at both pressures tested. In addition, the peaks corresponding to TiO2 are diminished at elevated pressure as compared to that of ambient pressure, and the peak corresponding to Fe remains unchanged. The ambient pressure sample exhibits distinct TiO2 peaks along with two prominent peaks characteristic of Fe. In the case of oxidized samples, completely oxidized phases are identified as Fe2O3 and TiO2. In addition, trace amount of complex species are formed at ambient pressures namely Fe2TiO5 and FeTiO3. These complex species are also found to be significantly lesser at sample oxidized at a pressure of 10 atm.

These differences in the samples point toward possible mechanistic differences in the reactions conducted at ambient pressure vs high pressure, which might be contributing towards a faster reaction rate at high pressures. However, the exact nature of these differences may only be understood through an in-depth study of the morphological changes on the reaction surface at the microscopic level at elevated pressures.

202

Figure 71: XRD analysis of (a) reduced and (b) oxidized samples at 1 and 10 atm, 900°C.

Reducing environment is under H2 with YH2 = 0.5, and oxidizing environment is air

203

The reacted particles were studied using scanning SEM and elemental mapping according to the technique outlined in section 6.3. More details of this technique are available elsewhere.27 Figure 72(a) and (b) shows the reduced samples produced at 1 and 10 atm, respectively. Figure 72(a) shows a typical non-uniform microparticle produced at ambient pressure: a denser part of Ti-rich oxides with a particle size of 40 µm and a porous part containing comparable amount of Fe and Ti oxides with an average grain size of 1-2 µm. It is apparent that subjecting the oxygen carrier particles to high pressure results in more porous particles upon reduction, with an average grain size of 500 nm can be clearly seen in the cross section (Figure 72(b)). This strongly suggests that higher pressure results in higher surface porosity compared to ambient pressure, which may explain the difference in reaction rate. Figure 73 shows the difference between samples treated under 1 atm and 10 atm. After reduction, the particles processed at 1 atm have a grain size of 1 µm and higher pressures lead to smaller grain size of ~500 nm.

Consequently, a reaction pressure at 10 atm can largely promote the increase of overall surface area. These samples were further tested for surface area and pore volume measurements using a BET analyzer. The BJH pore size distribution method was used to find the values of total surface area and pore volume for the four samples tested here.

From the analysis, it was discovered that the increase in pressure from 1 to 10 atm resulted in increase in both surface area and pore volume values, for reduced as well as oxidized samples. For the reduced samples, the increase in pressure resulted in surface area change from 7.036 m2/g to 7.227 m2/g whereas the pore volume values increased from 0.014 cc/g to 0.022 cc/g. The same trend was observed for oxidized samples, where

204 the change was more pronounced, going from 4.726 m2/g to 15.507 m2/g of surface area values, as well as 0.025 to 0.117 cc/g of total pore volume.

Figure 72: SEM and EDS elemental mapping of cross sections of reduced particles.

Samples reduced under H2, YH2 = 0.5 and T= 900 ˚C. (a) 1 atm and (b) 10 atm

205

Figure 73: Surface grains in samples reduced under H2, YH2 = 50% and T= 900 ˚C. (a) 1 atm and (b) 10 atm

206

All of these observations serve to explain the increased reduction rates for oxygen carrier particles with increased pressure. Using this preliminary analysis, we conclude that reduction carried out at elevated pressures results in particles with superior surface & intra-particle morphology, as well as uniform small grain structure, which contribute to the high reaction rates observed. A detailed analysis of the relationship between the particle morphology and the reaction rates will result in greater understanding of the mechanism of high pressure redox reactions involved in Fe-based partial oxidation system. Such a study is currently underway and will be published separately.

It should be noted that upon re-oxidation no significant morphological difference was found in samples treated at 1 and 10 atm, which is consistent with the small difference in the oxidation reaction rates observed.

207

6.5 Conclusions

For chemical synthesis applications, Fe-based partial oxidation process can be applied to produce pressurized syngas from CH4 suitable for downstream GTL applications. One such process developed at OSU has been termed as the Shale gas to Syngas (STS) process. The use of the STS process for downstream GTL applications requires the system to be operated at elevated pressures. Thus, it is imperative to investigate the effect of pressure on the operation of such a system, specifically, on the metal oxide particles that supply the oxygen. This study was conducted in order to determine the effect of pressure on the rates of reduction and oxidation reactions for OSU’s ITCMO particles developed for applications such as the STS process. Extensive testing was carried out on the reducer block at the conditions amenable to such a process. It is found that increase in the mole fraction or partial pressure of the reducing gas has a favorable effect on the rate of the reduction reaction of the ITCMO particles in the range of conditions tested.

Overall, operating the system at higher pressure results in superior reduction reaction rates observed. Although the primary goal of the present work is to ascertain the effect of pressure on the reaction rates ITCMO particles for CH4 partial oxidation using the STS process, extensive parametric study for reduction was carried out using H2 as the reducing gas. It is found that the experiments carried out with CH4 exhibit the same trends with the variation of pressure and therefore the pressure effects can be extrapolated. By choosing the experimental methodology to eliminate any gas dispersion effects, the kinetic advantage of higher pressure operation was successfully realized in this study.

208

With CH4 as the reducing gas, the ITCMO particles exhibit three distinct reaction stages for reduction. Stage II, which is the slowest and therefore rate determining stage, is the most sensitive to change in pressure; undergoing a 10 fold increase in rate with the increase in pressure from 1 to 10 atm. Thus, any increase in pressure results in a favorable change in rate of reduction reaction by CH4. The improvement in reaction rate is also observed in case of oxidation of reduced ITCMO particles, albeit to a smaller degree.

The reacted particles were analyzed using SEM, XRD, and BET techniques to understand the morphological changes on surface and intra-particle level, and the role of these changes in the observed differences in reaction rates. This analysis indicates that conducting the reduction reactions at elevated pressures results in product particle which shows more uniformity with respect to grain sizes, porosity and reactive iron-oxide distribution, with increased surface and intra-particle porosity. The formation in this uniform particle significantly contributes to the superior reaction kinetics.

One of the major issues of syngas production using CH4 is the formation of C soot.

However, the soot deposition can be managed or eliminated entirely using the control of various other process parameters such as gas to solid loading, reactor residence times, gas injection location and mode of gas-solid contact. Therefore, the advantages of faster reaction kinetics at higher pressures can be realized by circumventing the soot deposition through precise process operation. The advantages of operating the Fe-based partial oxidation system at elevated pressures include increased processing capacity or reduced reactor sizes and capital cost. The optimum operating pressures of the process can be

209 determined through rigorous experimental testing on various scales, and thorough process and economic analyses.

210

CHAPTER 7: Chemical Looping Applications: Redox Reactivity of

Steam Oxidation for Chemical Looping Particles

7.1 Introduction

The concept of chemical looping for energy conversion systems has been around for some time now, the first mention of chemical looping for indirect fuel oxidation is found in the steam-iron process around 1939.128 In Chapter 6, it was discussed in the context of chemical looping partial oxidation (CLPO) for the oxidation of methane (CH4). The oxygen for partial or complete oxidation of carbonaceous fuel is supplied by means of metal oxide solid materials at high temperature. The metal oxide (MeOx) are reduced in the process, and can be re-oxidized in a separate reactor called the combustor in the presence of air. Thus, the carbon dioxide (CO2), and other combustion gases produced, are not diluted with N2 from air and can be obtained in high purity form. This process concept is known as chemical looping combustion and a simplified schematic of the same is represented in Figure 74.

Originally, this concept was proposed solely for the clean combustion of fossil fuels.

However, oxidation of reduced metal oxides may also be carried out in presence of steam

(H2O) as oxidizing agent, producing hydrogen (H2) as a byproduct of the process. In the case of some single metal oxides such as Fe, steam is inadequate to fully regenerate the metal oxides, and supplemental air oxidation of the metal oxides may be required for

211 restoring the metal oxides to their fully oxidized form, to be recycled back into the reducer. This concept forms the basis of the chemical looping gasification process. Any carbonaceous fuel may be used as the fuel for the reducer, resulting in CO2 and H2O as the complete combustion products. Many small and large scale demonstrations of the process concept have been carried out the world over. The process can be configured to use gaseous fuels such as syngas (CO + H2), or CH4; or solid fuels such as coal or biomass. The overall process schematic is shown in Figure 75. The fuel is fed to the reducer unit, where it contacts with metal oxide MeOx. The reaction is typically carried out on the solid oxide surface and results in complete combustion of the fuel. The chemical energy of the fuel fed to the reducer is thus ‘stored’ in the form of the reduced metal oxide, and is regenerated in the combustor/oxidizer block. The gas stream exiting this reducer is mainly composed of the combustion products of CO2 and H2O. The CO2 may then be purified by condensing out the H2O and used for other applications or compressed for sequestration. The reduced metal oxides, in the form of MeOz (where z< x) are transported to the second block, namely, oxidizer. Here, the metal oxide is partially oxidized to MeOy (where z

The partially oxidized metal oxides MeOy are further completely oxidized to MeOx form in a separate unit, namely the combustor. This oxidation is identical to that of the CLC

212 configuration, and is carried out using air as the oxidizing agent in a conventional gas- solid fluidized bed reactor. This reactor is exothermic in operation, and the heat of reaction is recovered in a heat recovery steam generation cycle for co-production of electricity along with H2. It is possible to operate this system in the hybrid configuration, by completely or partially bypassing the oxidizer block and feeding the reduced solids to the combustor directly. In this case, the heat evolved from the oxidation process increases and thus the net electricity generation increases at the expense of the H2 production from the oxidizer block. Thus, a tradeoff exists for the coproduction of H2 and electricity from the CLG configuration.

The three step syngas chemical looping (SCL) process was proposed at OSU, using iron oxide (Fe2O3) based oxygen carrier materials for clean combustion of coal-derived syngas to produce a highly pure and sequestration-ready CO2 stream. The process utilizes coal and may result in the co-production of H2 and electricity. Various bench scale studies have been conducted on the reducer block of the SCL process.115 The reducer is operated as countercurrent moving bed reactor. The solid products exiting the reactor are a mixture of Fe and FeO.in accordance with the thermodynamic equilibrium attained in contact with the specific partial pressures of the gaseous reactant/product mixture. A near complete conversion of the syngas fed to the reactor is achieved, along with a 50% conversion of the Fe2O3 metal oxides.

In order to be suitable candidates for H2 production by CLG application, metal oxides for

CLG application must possess the following desirable properties. The metal oxide must exhibit thermodynamic feasibility of producing H2 from steam at CLG operating

213 conditions. In addition, the reaction kinetics for the reduction and oxidation reactions should be sufficiently rapid to allow realistic reactor sizes and processing time. The metal oxides should possess high oxygen carrying capacity, to reduce the amount of inert weight circulating in the system. The metal oxides should maintain the reactivity over multiple cycles and should possess good mechanical strength and attrition resistance.

Nickel (Ni) and cobalt (Co) metals have been investigated in their mixed metal ferrites form for the chemical looping gasification applications.129 Without the support of reactive materials such as ferrites, single metal systems such as Ni/NiO and Co/CoO are

130 unsuitable for H2 production using chemical looping gasification scheme. High pressure H2 is reported to be generated from Fe2O3 supported on alumina (Al2O3) at a pressure 12 bar in a fixed bed reactor setup.131 FeO is known to form complexes with the alumina support which reduces the reactivity of the oxygen carrier samples, therefore, mixed metal ferrites are suggested as novel oxygen carrier materials.132,133

Based on the results obtained on the experimental bench scale setup as well as from the process simulations using Aspen Plus software, a 25 kWth unit was built for the continuous demonstration of the SCL process.6 This unit has demonstrated the excellent high purity H2 generation (>99.99% pure) over continuous operation of the integrated process for over 300 hours.7 Here, the oxidizer operating condition was that of countercurrent moving bed operation, and a temperature of 800 °C. This reactor is operated with dual injection of solids and with non-mechanical valves, in this manner, a robust operation of the integrated system was successfully demonstrated5.

214

The reactivity of the metal oxides towards steam needs to be maintained over cyclic operation of the system. Although steady state high purity H2 production has been achieved using the iron-titanium complex metal oxide (ITCMO) particles in the 25kWth unit on a sub-pilot scale, the redox reactivity and recyclability of the Fe-based oxygen carriers with respect to steam oxidation was not independently established. Thus, the following study was undertaken to verify the redox reactivity and recyclability of the Fe- based metal oxides in repeated cycles of steam environment. The thermogravimetric type of experiments were performed using the magnetic suspension balance (MSB) setup described in previous chapters, to verify the recyclability of the metal oxides in the controlled environment of steam oxidation. Fe and Co-based oxygen carriers were identified as primary candidates that can be used for the redox reactions in the three-step cycle, where H2 may be produced from H2O. Thus, Co-based samples were chosen for the comparison with the Fe oxide samples. These metal oxides were selected in the form of Fe2O3 and Co3O4, and supported on MgAl2O4. The samples thus prepared were subjected to reduction using H2 as reducing agent, and oxidation using H2O, to produce

H2. The reactivity and recyclability of these materials was tested in the MSB setup over multiple cycles. In addition, a few tests were conducted at elevated pressures to test the reactivity of the metal oxide solids under higher pressure.

215

Figure 74: Simplified block flow diagram of chemical looping combustion.

216

Figure 75: Simplified process schematic for the chemical looping gasification process using steam to produce H2

217

7.2 Thermodynamic analysis Fe and Co metal oxides have been analyzed for their reaction with steam, and the thermodynamic analysis is shown in Figure 76. The reaction of different oxidation states of the two metals is considered with H2O in the temperature range of 500 to 1000°C. The equilibrium constants are calculated using HSC Chemistry 6.0 software, by the Gibbs free energy minimization method. The constants are plotted as a function of temperature in Figure 76.The fully reduced metals of Fe and Co both react readily with steam at the temperatures of interest, to produce H2. Thus, the first oxidation state for Fe and Co, namely FeO and CoO are readily formed. For FeO to be oxidized further to Fe3O4 is also thermodynamically favorable. However, CoO is not oxidized further to Co3O4 in the presence of steam. Indeed, higher oxidation states of Co than CoO are difficult to be formed even with air oxidation. Therefore, the oxygen carrying capacity of Co-based oxygen carriers is considered 21% by weight, considering the redox transformations between Co and CoO. In comparison, the oxygen carrying capacity of Fe-based oxygen carriers is considerably higher, namely, 30% by weight, when considered on the basis of

Fe2O3, which is the highest oxidation state achievable readily with air oxidation.

Table 14 gives the heats of reaction for the various oxidation states of Co and Fe per mole of H2 produced.

218

-2

-4

-6

EquilbriumConstant, log(K) -8

-10 500 600 700 800 900 1000 Temperature, degC

Fe + H2O(g) = FeO + H2(g) 3FeO + H2O(g) = Fe3O4 + H2(g) 2Fe3O4 + H2O(g) = 3Fe2O3 + H2(g) Co + H2O(g) = CoO + H2(g) 3CoO + H2O(g) = Co3O4 + H2(g)

Figure 76: Thermodynamic equilibrium constants as a function of temperature for various oxidation states of Fe and Co based materials

219

Table 14: TheRmodynamic properties for reactions of the various oxidation states of Fe and Co with steam, per mole of H2 produced, at 900°C

ΔH ΔS ΔG Metal oxide kJ/mol H2 J/K kJ/mol H2

Fe -16.84 -10.53 -4.49

FeO -43.23 -43.59 7.91

Fe3O4 2.15 -92.24 110.36

Co 15.79 -13.48 31.60

CoO 45.42 -112.61 177.53

220

7.3 Materials and methods

Fe2O3 and Co3O4 were used as the active components of the solids, and MgAl2O4 was used as a support material. The support and active component were mixed in 50 wt% proportions and sintered at high temperatures.

The MSB setup shown in Figure 39 of Chapter 5, section 5.4.2. The detailed experimental procedure is same as that explained in the previous chapter for the partial oxidation of CH4. The isothermal oxidation-reduction experiments were carried out at

900 °C. The reducing environment was 50% H2 in N2, and oxidation environment was

50% steam, with N2 as the sweep gas. Water injection was conducted for 7 minutes, followed by flushing for 53 minutes. Total of 60 minutes of oxidation time was allowed.

The complete cycle was carried out for 2 hours. Up to 20 redox cycles were performed on each sample, and the experiment was carried out over multiple days. After each reduction step, the sample cell was flushed with inert N2 flow for 10 minutes. The samples collected after the redox cycles were analyzed using XRD and SEM techniques.

In addition, some single cycle reduction and redox tests were carried out on the Fe and

Co-based samples to determine the role of pressure in enhancing the reaction rates. Up to

5 atm of pressure was tested. The samples were ramped in inert N2 flow till reaction temperature and pressure was reached, thereafter, reactive gases were injected into the sample cell for a specified period of time. Half and full cycles of Co and Fe samples were conducted to obtain samples for XRD and SEM analysis. The samples were stored in airtight containers for analysis later. Some select Co samples were also subjected to cross

221 sectional analysis by Focused Ion Beam (FIB) milling technique described in earlier chapters.

7.4 Results and discussion

7.4.1 Fe-based oxygen carriers

The samples were subjected to reduction and steam oxidation reactions using ambient and 5 atm test conditions. The reaction conversion curves obtained are shown in Figure

77. Increasing the pressure resulted in improved kinetics in some cases, when the reaction rates at ambient pressure were sufficiently low. For Fe2O3 samples, increase in pressure resulted in no significant change in reduction reaction rate, as the reaction rates were found to be considerably fast even at ambient pressure conditions. After reduction, the reduced samples were analyzed using XRD technique and identified as completely reduced Fe form for all pressures. The re-oxidation of reduced Fe sample with steam is inherently slower than the reduction with H2. Thus, for re-oxidation using steam, a definite improvement in reaction rates was observed. This can be seen in Figure 77(b).

After oxidation reaction, the samples were similarly analyzed using XRD. The evidence of slight formation of Fe2O3 was found in the XRD spectra for oxidation at ambient pressures. However, this Fe2O3 formation is of no practical significance from the process standpoint, considering the large excess of steam that will be required given the thermodynamic unfavorability of the reaction. The other phases identified were Fe3O4 and MgAl2O4. No complexes were formed with the reaction between the reactive metal oxide and the support material, unlike Al2O3 which is known to form complex FeAl2O4, hindering further reaction132.

222

Figure 77: Effect of pressure on reaction rates for Fe samples, using at 900°C for (a) reduction with 50% H2 as reducing agent, and (b) oxidation with steam.

223

This sample was tested for the recyclability under redox environment involving steam.

The recyclability of the Fe-based oxygen carriers was tested for 20 consecutive cycles. In these cycles, the oxidation was halted after conversion corresponding Fe3O4 phase formation was achieved. After this, the sample cell was flushed with N2 for 10 minutes before beginning the next reduction step. As a result, the initial weight loss of the sample during the first reduction is larger than subsequent cycles, owing to the complete reduction from Fe2O3 phase to Fe. The following cycles correspond to a transformation between Fe3O4 and Fe. The overall reactivity of the samples is found to be maintained over the 20 redox cycles tested here. The 20 cycles of the samples are shown in Figure

78.

The XRD analysis after 20 cycles showed the inert support MgAl2O4 as well as Fe3O4 as the only two phases present, which agrees with the conversion data obtained from the thermogravimetric experiments.

224

Figure 78: The 20 cycle redox reactivity of Fe-based oxygen carriers at 900°C.

225

7.4.2 Co-based oxygen carriers

Similar experiments were performed with the Co3O4 based samples. As compared to Fe samples, the Co samples showed much slower kinetics of both reduction with H2 and oxidation with steam. For these samples, the increase in pressure showed more pronounced effect on the reaction rates as compared to the Fe samples. Specifically during reduction, the Co3O4 sample showed significantly faster reaction rates at 5 atm as compared to ambient atmospheric conditions. This is seen in Figure 79. No appreciable difference was observed in case of oxidation for Co samples with steam. The oxidation of

Co samples at elevated pressures proved to be problematic, and stable oxidation was never achieved. Furthermore, the loss of all the oxygen from a Co3O4 sample with 50% inerts should necessarily result in 13% weight loss during reduction. However, the samples never showed a 13% weight loss. The maximum amount of weight loss exhibited by the samples was ~8%, indicating that the starting sample was not entirely composed of fully oxidized phase of Co3O4, and rather contained some lower oxidation states of Co.

The XRD analysis of the initial sample showed evidence of CoO in addition to Co3O4, further confirming this fact. The XRD analysis of the reduced samples showed only Co as the active metal phase. The XRD analysis of the oxidized samples indicated that complete oxidation does not occur at the test conditions, with unreacted Co also being found in the samples along with CoO.

The Co-based samples were also subjected to extended number of redox cycles, with 20 cycles. The results of this cyclic testing are shown in Figure 80. As can be seen, the reactivity of Co samples was not constant over the cycles, and the weight loss and gain

226 varied for every cycle (the figure shows processed data which is adjusted for the same starting weight at the beginning of every reduction cycle.). The XRD analysis was performed on the sample after subjecting it to 20 cycles. The sample showed unconverted

Co at the end of the final oxidation, further substantiating the loss in reactivity observed during the thermogravimetric experiments.

The samples were tested for SEM and EDS analysis. The internal structure of the Co- based samples was examined using the FIB milling technique. The full technique is described in detail in the previous chapter. The Figure 81 shows the EDS mapping of the cross section of a reduced microparticle. The particles showed uniform distribution of the active metal and support material. Additionally, few separate microparticles of support material MgAl2O4 were also observed. The EDS mapping of one such microparticle is given in the section A.4.3 of the appendix.

227

Figure 79: Effect of pressure on reaction rates for Co samples, using at 900°C for (a) reduction with 50% H2 as reducing agent, and (b) oxidation with steam.

228

Figure 80: The 20 cycle redox reactivity of Co-based oxygen carriers at 900°C

229

Figure 81: SEM of the cross section of reduced Co sample at 5 atm, and 900°C. The EDS mapping clearly shows the presence of Co, Mg, Al, and O phases in a single microparticle tested here.

230

7.5 Conclusions

Chemical looping gasification is an attractive process scheme for the cogeneration of H2 and electricity from carbonaceous fuels such as coal. The iron oxide based syngas chemical looping (SCL) process has already been demonstrated at various scales of operation. The production of high purity H2 from steam (>99.99% H2) has been achieved in the reducer block of the three step SCL process through systematic testing. The Fe- based SCL process uses the iron-titanium complex metal oxide particles as the oxygen carriers for achieving this high purity H2. The favorable thermodynamics and kinetics of oxides of Fe in the redox system makes this the most commonly studied material for chemical looping gasification applications. Thus, it was essential to study the reactivity and recyclability of the Fe-based oxygen carriers under steam oxidation conditions. The present work was undertaken to achieve this purpose. Cobalt based oxygen carriers were also chosen for comparison purposes.

The MgAl2O4 material was used as support material for all the tests carried out under this work. This material was found to be an excellent inert support and did not interact in any way with the active metal oxide component of the oxygen carriers for Co as well as Fe samples. Excellent oxidation and reduction kinetics were observed for the Fe-based samples. The samples exhibited high oxygen carrying capacity, with full oxidation to

Fe3O4 phase being observed in the steam oxidation step. Further, the recyclability of the

Fe-based oxygen carriers was verified over 20 redox cycles with reduction under H2 and re-oxidation using steam. Also, increase in pressure resulted in faster reaction kinetics for the H2 production step for this Fe-based system. In comparison, the Co-based oxygen

231 carriers were found to exhibit substantially slower kinetics of reduction as well as oxidation. With steam, only oxidation up to CoO is observed, which results in substantially lower oxygen carrying capacity of the Co-based oxygen carriers.

Furthermore, the samples treated with 20 redox cycles exhibited a loss in reactivity, and the final oxidized sample showed the presence of unreacted Co. The co-based samples also showed favorable response to increase in pressure, with increased reaction rates observed at elevated pressures for reduction.

The SEM analysis of the samples showed that the Fe-based samples were more porous than their Co-based counterparts, showing larger amounts of surface area available for reactions. Generally, the reduced samples were observed to be more porous than the samples oxidized with steam for both systems. Overall, this study conclusively proves the excellent recyclability of the Fe-based oxygen carriers for the H2 production using syngas chemical looping process.

232

FUTURE DIRECTIONS

The chemical looping research field stands at an exciting juncture right now, with several important milestones in the offing in the near future.

For the calcium looping technology, some of the remaining uncertainties of the process need to be addressed before the road to scale up and possible commercialization is cleared. The biggest technological uncertainty for pre and post combustion calcium looping technology remains the design and successful operation of a fluidized bed steam hydrator that can be easily scaled up and integrated into the system. Presently, batch and semi-batch experiments reported in this study have yielded promising results, and point towards a fast fluidized bed reactor design as the most suitable mode of operation of this reactor. Research effort should be focused towards maximizing the reaction conversion of hydration (beyond current best of 70%) while minimizing the excess steam and the reaction time required. Additionally, the kinetics of the hydration reaction may be explored for the conditions amenable to the calcium looping process, for pre and post combustion application, to help maximize the hydration reaction conversion and optimize the reactor design.

The continuous demonstration of the integrated system over long term testing is next logical step in the development of the calcium looping technology. Once each individual

233 unit is optimized for the reactor performance, the high temperature solids circulation in the three reactor units will be the biggest engineering challenge which will have to be addressed. The particle capture efficiency of high temperature separation devices such as cyclones have been tested under cold conditions and given >99% efficiency for the

~10µm size calcium sorbent particles. This should be confirmed at actual operating conditions at high temperatures in the integrated operation. The long term removal of trace impurities and the buildup of inert species in the system will have to be studied in this integrated setup, and their implications on the purge and makeup rates verified. On the basis of the success of demonstration of the integrated operation of the three step calcium loop, several modifications to the current process simulations analysis will be made and a new techno-economic analysis should be performed to assess the impact of this carbon capture technology on the overall power generation process efficiency, cost of

CO2 avoided, and purity of other gaseous products in the case of pre-combustion applications.

Several of type I chemical looping technology applications are on the verge of being established at pilot scale demonstrations, especially the Syngas Chemical Looping 250 kWth pilot plant at the National Carbon Capture Center. This unit designed to demonstrate the chemical looping process concept through pressurized combustion of syngas. Excellent purity of H2 (>99.99%) has already been demonstrated at the 25kwth operation of the SCL sub-pilot unit with continuous operation. The successful demonstration of this technology at the 250 kWth plant will be the stepping stone of the launching of this technology platform towards large scale, commercial application.

234

At the same time, fundamental research effort is being expended simultaneously to develop superior metal oxide oxygen carrier particles suitable for this technology.

Through these fundamental studies, the actual reaction mechanisms on the gas-solid interface level can be better understood, which will lead to the development of engineered solid particles with excellent attrition resistance, superior mechanical strength and high reactivity at the same time. While the experience of large pilot scale demonstrations bolsters the technological and engineering knowhow and develops high confidence in operational expertise, the ongoing development through fundamental scientific studies of the excellent oxygen carrier particle possessing desired qualities will propel this technology platform towards successful commercialization.

235

APPENDIX: Supplemental Data

A.1 Calcium sorbent reactivation by hydration

A.1.1 Steam hydration TGA experiments

The hydration of CaO sorbent can be carried out in a specialized TGA setup using the

MSB apparatus to inject steam, as mentioned in Chapters 5 and 7. In this manner, investigation of hydration reaction rates is possible for various parameters such as steam partial pressure, temperature, etc. The following Figure 82 and Figure 83 shows examples of such tests carried out at 400°C.

236

80% 70% 60% 50% 40% Wt % change 30% conversion 20% 10% 0% reactionconversion and wt%change, % 0 10 20 30 40 time, minutes

Figure 82: Hydration of CaO sorbent, 50% steam, 3 atm, 400°C

100% 90% 80% 70% 60% 50% conversion 40% weight % change 30% 20% 10% 0% Reactionconversion wt%andchange, % 0 10 20 30 40 time, minutes

Figure 83: Hydration of CaO sorbent, 50% steam, 1.5 atm, 400°C

237

A.1.2 Decay in reactivity of CaO sorbent over continuous TGA testing

0.7 N2 calcination 0.6 CO2 calcination

0.5

0.4

0.3 conversion 0.2

0.1

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 cycle number

Figure 84: Carbonation-calcination cycles performed in the TGA, effect of calcination environment on sintering.

The calcination in CO2-rich environment shows more decay in reactivity of the sample.

This is exhibited in Figure 84 with samples derived from BR limestone. Calcination and carbonation continuous cycles were performed in inert N2 and CO2 rich environment.

Figure 85 shows the effect of increase in CO2 concentration on sintering of the calcite and dolomite sorbent.

238

BR calcite, 10% CO2 calc BR calcite, 40% CO2 calc dolomite, 40% CO2 calc dolomite 10% CO2 calc 0.9 0.8

0.7 0.6 0.5 0.4

0.3 conversionofCaO 0.2 0.1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 cycle number

Figure 85: Effect of CO2 concentration on sintering of the limestone and dolomite sorbent.

239

………. Equation A1

Equation A1 is given by Grasa and Abanades134 to explain the loss in reactivity of the

CaO sorbent over multiple cycles, Here, the decay in reactivity of CaO derived from MV limestone is fitted with equation A1 in Figure 86, with the parameters

푎∞ = 0.08 k = 1.2

Where 푎∞ is the residual conversion of the sample (approaching the asymptotic value), and k is an empirical constant to quantify the rate of decay. Different forms of equation

A1 have been proposed in the literature in an attempt to model the decay in reactivity of the CaO sorbent over multiple cycles.17

240

60% experimental conversion

50% equation

40%

30%

20% carbonationconversion 10%

0% 0 5 10 15 20 25 30 35 cycle number

Figure 86: Modeling the decay in reactivity by the equation given by equation A1 for MV sorbent

241

A.2 Fate of Sulfur

A.2.1 Soot formation

Soot

Figure 87: Carbon deposition observed at the coupling zone of the Rubotherm MSB

A.2.2 The Buoyancy change

The buoyancy change due to change in gases fed to the MSB setup causes an apparent weight change at elevated pressures, with CO and H2. 30% CO and ~20% H2 was injected in the sample cell containing CaSO4 sample, at 650 C and 5 atm. The results are shown in Figure 88.

242

Figure 88: Simultaneous addition of ~30% CO and ~20% H2 to the TGA in presence of CaSO4 solids. Temperature maintained at 650 C and pressure of 5 atm. The outlet gas concentrations are shown on primary y-axis and sample weight recorded by the MSB is shown on the secondary y-axis.

It was observed that the sample weight increased during the injection of CO and H2 simultaneously, however, this is only “apparent weight change” as it returns back to baseline after switching back to inert gases.

243

A.2.3 Investigation of CaSO4 reaction

Figure 89: Product XRD analysis after experiment with high purity (95%) H2, 650°C, 10 atm

244

A.3 Reduction of ITCMO particles under pressure

A.3.1 CH4 reduction at 900°C

80%

70%

60%

50% 50% CH4, 1 atm

40% 50% CH4, 1 atm 30% (repeat) 50% CH4, 3 atm 20% Redcutionconversion 50% CH4, 3 atm 10% (repeat) 50% CH4, 5 atm 0% 0 20 40 60 50% CH4, 10 atm time, minutes

Figure 90: ITCMO particles reduced under 50% CH4 at 900 °C, effect of pressure on reaction rates

245

A.3.2 H2 reduction at 900°C

100% 90%

80% 70% 60% 1 atm 50% 3 atm 40% 5 atm

30% 8 atm (repeat) Extent ofreduction, % 20% 10 atm 10% 0% 0 10 20 30 time, minutes

Figure 91: ITCMO particles reduced under 50% H2 at 900 °C, effect of pressure on reaction rates

100%

80%

3 atm, 50% H2 60% 5 atm, 30% H2

40% 8 atm, 18.75% H2 Conversion

8 atm repeat, 18.75% 20% H2 10 atm, 15% H2 0% 0 10 20 30 40 50 60 Time, minute

Figure 92: ITCMO particles reduced under constant partial pressure of H2 of 1.5 atm at 900 °C, effect of pressure on reaction rates

246

A.4 Steam oxidation of reduced oxygen carrier samples

A.4.1 Sample calculation of extent of oxidation for Fe-based oxygen carriers

We know that 50% of the sample is active Fe2O3. The initial weight taken after the buoyancy loss is

Initial wt = 92.584 mg

The following are the three approaches (assumptions) for calculating the extent of oxidation.

A: 50% of initial wt is Fe2O3

∴ Fe2O3 = 46.292 mg

∴ Fe = 32.404 mg = 0.5787 moles

If we take this weight as the “reactive wt”, we are making the assumption that complete reduction has occurred, and therefore all of the Fe in the sample is available for steam oxidation. (Here, we are NOT assuming that the support material undergoes no decomposition)

B: weight loss in the first reduction is purely from Fe2O3

Here too, there’s the underlying assumption that complete reduction has taken place.

The first ∆red1 = 92.584 - 79.534 = 13.05 mg of ‘O’ that came from Fe2O3

Therefore we calculate the wt of Fe = 30.45 mg = 0.5438 moles

247

It is important to note that we take the weight After H2 is turned off. By this time, the baseline weight has risen due to the buoyancy effect. Therefore, this weight is predictably higher than the “true” weight, and this results in lesser “reactive wt” of Fe. That is, B <

Here, in addition to the complete reduction assumption, we are also making the assumption that the support material is completely inert/did not contribute to the loss in weight in red1.

C: 50% of the initial wt is inert, and complete reduction takes place in red1

Therefore, wt of inert = 92.584/2 = 46.292 mg

We subtract this from the wt after red1 to obtain the wt of Fe.

Therefore Fe = 79.534 – 46.292 = 33.242 mg = 0.5936 moles

Note that method B and C work with identical assumptions and still yield different numbers. They take into consideration the weight after red1, which may not be the best assumption – because of the buoyancy effect. Also, compared to method A, these two methods have one additional assumption: that support material is completely inert.

Estimating the extent of oxidation every cycle:

Now that we know the “reactive wt” by methods A, B and C; we can calculate the extent of oxidation.

248

Here, we assume that the weight gain in every cycle is due to ‘O’ that has reacted with

Fe.

Table 15: Calculation of extent of steam oxidation of the Fe-based oxygen carriers by the three methods.

Cycle number Weight of O, mg mMoles of O FeOx, where x is as follows

Method A Method B Method C

1 13.318 .8324 1.438 1.531 1.402

2 11.609 .7256 1.254 1.334 1.322

3 11.750 .7344 1.269 1.350 1.237

4 11.233 .702 1.213 1.291 1.183

249

A.4.2 Raw data of 20 redox cycles of Fe-based oxygen carriers

Sample: Powdered form of Fe2O3-MgAl2O4 50% by weight

Reduction: 50% hydrogen in nitrogen 50 minutes, followed by 10 minutes of flushing

Oxidation: 50% steam in nitrogen. Water injection for 7 minutes, followed by flushing for 53 minutes. Total of 60 minutes

Temperature: 900°C

0.24

0.23

0.22 series1 0.21 Series2

weight, weight, g Series3 0.2 Series4 Series5 0.19

0.18 0 500 1000 1500 2000 2500 3000 Time, minutes

Figure 93: The 20 redox cycles of Fe-based oxygen carriers supported on MgAl2O4, each series corresponds to a new day of testing. Compiled data

250

Table 16: Raw weight data for the 20 redox cycles for Fe-based oxygen carriers initial wt 0.2317 mg moles of Fe 1.4272E-03 final wt 0.1974 mg

wt after (mg) cycle # red oxid ∆w (mg) mmoles of O FeOx 1 0.197429 0.223771 0.026342 1.6464E-03 1.154 2 0.202225 0.227135 0.02491 1.5569E-03 1.091 3 0.202696 0.227164 0.024468 1.5293E-03 1.072 4 0.202679 0.227600 0.024921 1.5576E-03 1.091 5 0.202291 0.225393 0.023102 1.4439E-03 1.012 6 0.202957 0.229069 0.026112 1.6320E-03 1.144 7 0.203606 0.227961 0.024355 1.5222E-03 1.067 8 0.203457 0.227962 0.024505 1.5316E-03 1.073 9 0.200044 0.224558 0.024514 1.5321E-03 1.074 10 0.201970 0.227010 0.02504 1.5650E-03 1.097 11 0.202781 0.227435 0.024654 1.5409E-03 1.080 12 0.202740 0.227181 0.024441 1.5276E-03 1.070 13 0.201448 0.225128 0.02368 1.4800E-03 1.037 14 0.202375 0.227703 0.025328 1.5830E-03 1.109 15 0.202685 0.227969 0.025284 1.5803E-03 1.107 16 0.199248 0.224803 0.025555 1.5972E-03 1.119 17 0.201567 0.227741 0.026174 1.6359E-03 1.146 18 0.203287 0.228956 0.025669 1.6043E-03 1.124 19 0.203472 0.228630 0.025158 1.5724E-03 1.102 20 0.203058 0.228619 0.025561 1.5976E-03 1.119

251

A.4.3 Co-based oxygen carriers upon re-oxidation

Figure 94: Reoxidized Co-based sample, SEM image of the surface. Sample reoxidized at 5 atm using steam, 900°C

252

Figure 95: EDS mapping of MgAl2O4 support used in steam oxidation experiments for chemical looping

253

REFERENCES

(1) Figueroa, J. D.; Fout, T.; Plasynski, S.; McIlvried, H.; Srivastava, R. D. Advances in CO2 Capture technology—The U.S. Department of Energy’s Carbon Sequestration Program. Int. J. Greenh. Gas Control 2008, 2 (1), 9–20. (2) Fan, L.-S. Chemical Looping Systems for Fossil Energy Conversions; John Wiley & Sons, 2011. (3) Fan, L.-S.; Zeng, L.; Wang, W.; Luo, S. Chemical Looping Processes for CO2 Capture and Carbonaceous Fuel Conversion – Prospect and Opportunity. Energy Environ. Sci. 2012, 5 (6), 7254–7280. (4) Fan, L.-S.; Li, F. Chemical Looping Technology and Its Fossil Energy Conversion Applications. Ind. Eng. Chem. Res. 2010, 49 (21), 10200–10211. (5) Tong, A.; Bayham, S.; Kathe, M. V.; 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. Energy 2014, 113, 1836–1845. (6) Sridhar, D.; Tong, A.; Kim, H.; Zeng, L.; Li, F.; Fan, L.-S. Syngas Chemical Looping Process: Design and Construction of a 25 kWth Subpilot Unit. Energy Fuels 2012, 26 (4), 2292–2302. (7) Tong, A.; Sridhar, D.; Sun, Z.; Kim, H. R.; Zeng, L.; Wang, F.; Wang, D.; Kathe, M. V.; Luo, S.; Sun, Y.; Fan, L.-S. Continuous High Purity Hydrogen Generation from a Syngas Chemical Looping 25 kWth Sub-Pilot Unit with 100% Carbon Capture. Fuel 2013, 103, 495–505. (8) Kim, H. R.; Wang, D.; Zeng, L.; Bayham, S.; Tong, A.; Chung, E.; Kathe, M. V.; Luo, S.; McGiveron, O.; Wang, A.; Sun, Z.; Chen, D.; Fan, L.-S. Coal Direct Chemical Looping Combustion Process: Design and Operation of a 25-kWth Sub-Pilot Unit. Fuel 2013, 108, 370–384. (9) Zeng, L.; He, F.; Li, F.; Fan, L.-S. Coal-Direct Chemical Looping Gasification for Hydrogen Production: Reactor Modeling and Process Simulation. Energy Fuels 2012, 26 (6), 3680–3690. (10) Bayham, S. C.; Kim, H. R.; Wang, D.; Tong, A.; Zeng, L.; McGiveron, O.; Kathe, M. V.; Chung, E.; Wang, W.; Wang, A.; Majumder, A.; Fan, L.-S. Iron- Based Coal Direct Chemical Looping Combustion Process: 200-H Continuous Operation of a 25-kWth Subpilot Unit. Energy Fuels 2013, 27 (3), 1347–1356. (11) Sit, S. P.; Reed, A.; Hohenwarter, U.; Horn, V.; Marx, K.; Proell, T. Cenovus 10 MW CLC Field Pilot. Energy Procedia 2013, 37, 671–676. (12) Penthor, S.; Mayer, K.; Kern, S.; Kitzler, H.; Wöss, D.; Pröll, T.; Hofbauer, H. Chemical-Looping Combustion of Raw Syngas from Biomass Steam Gasification

254

– Coupled Operation of Two Dual Fluidized Bed Pilot Plants. Fuel 2014, 127, 178–185. (13) Adanez, J.; Abad, A.; Garcia-Labiano, F.; Gayan, P.; de Diego, L. F. Progress in Chemical-Looping Combustion and Reforming Technologies. Prog. Energy Combust. Sci. 2012, 38 (2), 215–282. (14) Ströhle, J.; Orth, M.; Epple, B. Design and Operation of a 1 MWth Chemical Looping Plant. Appl. Energy 2014, 113, 1490–1495. (15) Abdulally, I.; Beal, C.; Andrus, H.; Epple, B.; Lyngfelt, A.; Lani, B. W. Alstom’s Chemical Looping Prototypes, Program Update. In Proceedings from 37th International Technical Conference on Clean Coal & Fuel Systems; Clearwater, FL, USA, 2012. (16) Shimizu, T.; Hirama, T.; Hosoda, H.; Kitano, K.; Inagaki, M.; Tejima, K. A Twin Fluid-Bed Reactor for Removal of CO2 from Combustion Processes. Chem. Eng. Res. Des. 1999, 77 (1), 62–68. (17) Blamey, J.; Anthony, E. J.; Wang, J.; Fennell, P. S. The Calcium Looping Cycle for Large-Scale CO2 Capture. Prog. Energy Combust. Sci. 2010, 36 (2), 260–279. (18) Sánchez-Biezma, A.; Ballesteros, J. C.; Diaz, L.; de Zárraga, E.; Álvarez, F. J.; López, J.; Arias, B.; Grasa, G.; Abanades, J. C. Postcombustion CO2 Capture with CaO. Status of the Technology and next Steps towards Large Scale Demonstration. Energy Procedia 2011, 4, 852–859. (19) Fan, L.; Gupta, H. Separation of Carbon Dioxide (CO2) from Gas Mixtures by Calcium Based Reaction Separation (CaRS-CO2) Process. 20060039853, February 23, 2006. (20) Fan, L.-S.; Ramkumar, S.; Wang, W.; Statnick, R. Carbonation Calcination Reaction Process for CO2 Capture Using a Highly Regenerable Sorbent. US8512661 B2, August 20, 2013. (21) Deshpande, N.; Phalak, N.; Fan, L.-S.; Sankaran, S. Carbon Dioxide (CO2) Capture from Coal-Fired Power Plants Using Calcium Looping. Chem. Eng. Educ. (22) Wang, W.; Ramkumar, S.; Li, S.; Wong, D.; Iyer, M.; Sakadjian, B. B.; Statnick, R. M.; Fan, L.-S. Subpilot Demonstration of the Carbonation−Calcination Reaction (CCR) Process: High-Temperature CO2 and Sulfur Capture from Coal- Fired Power Plants. Ind. Eng. Chem. Res. 2010, 49 (11), 5094–5101. (23) Chang, M.-H.; Huang, C.-M.; Liu, W.-H.; Chen, W.-C.; Cheng, J.-Y.; Chen, W.; Wen, T.-W.; Ouyang, S.; Shen, C.-H.; Hsu, H.-W. Design and Experimental Investigation of Calcium Looping Process for 3-kWth and 1.9-MWth Facilities. Chem. Eng. Technol. 2013, 36 (9), 1525–1532. (24) Phalak, N.; Ramkumar, S.; Deshpande, N.; Wang, A.; Wang, W.; Statnick, R. M.; Fan, L.-S. Calcium Looping Process for Clean Coal Conversion: Design and Operation of the Subpilot-Scale Carbonator. Ind. Eng. Chem. Res. 2012, 51 (30), 9938–9944. (25) Arias, B.; Diego, M. E.; Abanades, J. C.; Lorenzo, M.; Diaz, L.; Martínez, D.; Alvarez, J.; Sánchez-Biezma, A. Demonstration of Steady State CO2 Capture in a 1.7 MWth Calcium Looping Pilot. Int. J. Greenh. Gas Control 2013, 18, 237– 245.

255

(26) Plötz, S.; Bayrak, A.; Galloy, A.; Kremer, J.; Orth, M.; Wieczorek, M.; Ströhle, J.; Epple, B. First Carbonate Looping Experiments with a 1 MWth Test Facility Consisting of Two Interconnected CFBs. 21st Int. Conf. Fluid. Bed Combust. 2012, 421–428. (27) Florin, N. H.; Blamey, J.; Fennell, P. S. Synthetic CaO-Based Sorbent for CO2 Capture from Large-Point Sources. Energy Fuels 2010, 24 (8), 4598–4604. (28) Manovic, V.; Anthony, E. J. Lime-Based Sorbents for High-Temperature CO2 Capture—A Review of Sorbent Modification Methods. Int. J. Environ. Res. Public. Health 2010, 7 (8), 3129–3140. (29) Yu, F.-C.; Phalak, N.; Sun, Z.; Fan, L.-S. Activation Strategies for Calcium- Based Sorbents for CO2 Capture: A Perspective. Ind. Eng. Chem. Res. 2012, 51 (4), 2133–2142. (30) Liu, F.-Q.; Li, W.-H.; Liu, B.-C.; Li, R.-X. Synthesis, Characterization, and High Temperature CO2 Capture of New CaO Based Hollow Sphere Sorbents. J. Mater. Chem. A 2013, 1 (27), 8037–8044. (31) Liu, W.; Low, N. W.; Feng, B.; Wang, G.; Diniz da Costa, J. C. Calcium Precursors for the Production of CaO Sorbents for Multicycle CO2 Capture. Environ. Sci. Technol. 2010, 44 (2), 841–847. (32) Li, Y.; Sun, R.; Liu, H.; Lu, C. Cyclic CO2 Capture Behavior of Limestone Modified with Pyroligneous Acid (PA) during Calcium Looping Cycles. Ind. Eng. Chem. Res. 2011, 50 (17), 10222–10228. (33) Albrecht, K. O.; Wagenbach, K. S.; Satrio, J. A.; Shanks, B. H.; Wheelock, T. D. Development of a CaO-Based CO2 Sorbent with Improved Cyclic Stability. Ind. Eng. Chem. Res. 2008, 47 (20), 7841–7848. (34) Li, Z.; Cai, N.; Huang, Y. Effect of Preparation Temperature on Cyclic CO2 Capture and Multiple Carbonation−Calcination Cycles for a New Ca-Based CO2 Sorbent. Ind. Eng. Chem. Res. 2006, 45 (6), 1911–1917. (35) Arias, B.; Grasa, G. S.; Alonso, M.; Abanades, J. C. Post-Combustion Calcium Looping Process with a Highly Stable Sorbent Activity by Recarbonation. Energy Environ. Sci. 2012, 5 (6), 7353–7359. (36) Liu, W.; An, H.; Qin, C.; Yin, J.; Wang, G.; Feng, B.; Xu, M. Performance Enhancement of Calcium Oxide Sorbents for Cyclic CO2 Capture—A Review. Energy Fuels 2012, 26 (5), 2751–2767. (37) Sun, Z.; Chi, H.; Fan, L.-S. Physical and Chemical Mechanism for Increased Surface Area and Pore Volume of CaO in Water Hydration. Ind. Eng. Chem. Res. 2012, 51 (33), 10793–10799. (38) Wang, W.; Ramkumar, S.; Wong, D.; Fan, L.-S. Simulations and Process Analysis of the Carbonation–calcination Reaction Process with Intermediate Hydration. Fuel 2012, 92 (1), 94–106. (39) Ramkumar, S.; Fan, L.-S. Thermodynamic and Experimental Analyses of the Three-Stage Calcium Looping Process. Ind. Eng. Chem. Res. 2010, 49 (16), 7563–7573. (40) Wang, A.; Wang, D.; Deshpande, N.; Phalak, N.; Wang, W.; Fan, L.-S. Design and Operation of a Fluidized Bed Hydrator for Steam Reactivation of Calcium Sorbent. Ind. Eng. Chem. Res. 2013, 52 (8), 2793–2802.

256

(41) Schaube, F.; Koch, L.; Wörner, A.; Müller-Steinhagen, H. A Thermodynamic and Kinetic Study of the de- and Rehydration of Ca(OH)2 at High H2O Partial Pressures for Thermo-Chemical Heat Storage. Thermochim. Acta 2012, 538, 9– 20. (42) International Energy Outlook-2011; DOE/EIA-0484; 2011. (43) Ouyang, S.; Hsu, H. W.; Tong, L. T.; Liao, C. W.; Hu, R. Y. Z. Carbon Capture and Sequestration Technology Development in ITRI. Sustain. Environ. Res. 2011, 21 (1), 21–28. (44) Chiao, C. H.; Chen, J. L.; Lan, C. R.; Chen, S.; Hsu, H. W. Development of Carbon Dioxide Capture and Storage Technology - Taiwan Power Company Perspective. Sustain. Environ. Res. 2011, 21 (1), 1–8. (45) Shie, J. L.; Chang, C. Y.; Chiou, C. S.; Chen, Y. H.; Chen, Y. W.; Chang, C. C. Photocatalytic Reduction of Gaseous and Solution CO2 to Energy Products Using Ag/TiO2 and Cu/TiO2 in CuCl2 Solution. Sustain. Environ. Res. 2012, 22 (4), 237–246. (46) Lin, C. J.; Liou, Y. H.; Chen, S. Y.; Tsai, M. C. Visible-Light Photocatalytic Conversion of CO2 to Methanol Using Dye-Sensitized Mesoporous Photocatalysts. Sustain. Environ. Res. 2012, 22 (3), 167–172. (47) Rao, A. B.; Rubin, E. S. Identifying Cost-Effective CO2 Control Levels for Amine-Based CO2 Capture Systems. Ind. Eng. Chem. Res. 2006, 45 (8), 2421– 2429. (48) Li, F.; Fan, L.-S. Clean Coal Conversion Processes – Progress and Challenges. Energy Environ. Sci. 2008, 1 (2), 248–267. (49) Anthony, E. J. (Ben). Ca Looping Technology: Current Status, Developments and Future Directions. Greenh. Gases Sci. Technol. 2011, 1 (1), 36–47. (50) Dean, C. C.; Blamey, J.; Florin, N. H.; Al-Jeboori, M. J.; Fennell, P. S. The Calcium Looping Cycle for CO2 Capture from Power Generation, Cement Manufacture and Hydrogen Production. Chem. Eng. Res. Des. 2011, 89 (6), 836– 855. (51) George p. Curran; Carl e. Fink; Everett Gorin. CO2 Acceptor Gasification Process. In Fuel Gasification; Advances in Chemistry; AMERICAN CHEMICAL SOCIETY, 1967; Vol. 69, pp 141–165. (52) Balasubramanian, B.; Lopez Ortiz, A.; Kaytakoglu, S.; Harrison, D. P. Hydrogen from Methane in a Single-Step Process. Chem. Eng. Sci. 1999, 54 (15–16), 3543– 3552. (53) Barelli, L.; Bidini, G.; Gallorini, F.; Servili, S. Hydrogen Production through Sorption-Enhanced Steam Methane Reforming and Membrane Technology: A Review. Energy 2008, 33 (4), 554–570. (54) IEC, The World Egg Industry - A Few Facts and Figures. International Egg Comission, 2012. (55) Egg Industry Fact Sheet. United Egg Producers, 2012. (56) Ma, K.-W.; Teng, H. CaO Powders from Oyster Shells for Efficient CO2 Capture in Multiple Carbonation Cycles. J. Am. Ceram. Soc. 2010, 93 (1), 221–227.

257

(57) Sacia, E. R.; Ramkumar, S.; Phalak, N.; Fan, L.-S. Synthesis and Regeneration of Sustainable CaO Sorbents from Chicken Eggshells for Enhanced Carbon Dioxide Capture. ACS Sustain. Chem. Eng. 2013, 1 (8), 903–909. (58) Deshpande, N.; Yuh, B. Screening of Multiple Waste Animal Shells as a Source of Calcium Sorbent for High Temperature CO2 Capture. Sustain. Eng. Res. 2013, 23, 227–232. (59) Bargain Empty Ostrich Eggshell. 24. Floeck’s Country Ranch, Product Catalogue, 2012. (60) Phalak, N.; Deshpande, N.; Fan, L.-S. Investigation of High-Temperature Steam Hydration of Naturally Derived Calcium Oxide for Improved Carbon Dioxide Capture Capacity over Multiple Cycles. Energy Fuels 2012, 26 (6), 3903–3909. (61) Li, L.; Zhao, N.; Wei, W.; Sun, Y. A Review of Research Progress on CO2 Capture, Storage, and Utilization in Chinese Academy of Sciences. Fuel 2013, 108, 112–130. (62) Li, Y.; Zhao, C.; Duan, L.; Liang, C.; Li, Q.; Zhou, W.; Chen, H. Cyclic Calcination/carbonation Looping of Dolomite Modified with Acetic Acid for CO2 Capture. Fuel Process. Technol. 2008, 89 (12), 1461–1469. (63) Federal Register / Vol. 77, No. 64 / Tuesday, April 3, 2012 / Rules and Regulations. (64) Bowman, C. T. Control of Combustion-Generated Nitrogen Oxide Emissions: Technology Driven by Regulation. Symp. Int. Combust. 1992, 24 (1), 859–878. (65) Srivastava, R. K. Controlling SO2 Emissions–a Review of Technologies; United States Environmental Protection Agency, Office of Research and Development Washington, DC, 2000. (66) Spliethoff, H.; Greul, U.; Rüdiger, H.; Hein, K. R. G. Basic Effects on NOx Emissions in Air Staging and Reburning at a Bench-Scale Test Facility. Fuel 1996, 75 (5), 560–564. (67) Srivastava, R. K.; Hall, R. E.; Khan, S.; Culligan, K.; Lani, B. W. Nitrogen Oxides Emission Control Options for Coal-Fired Electric Utility Boilers. J. Air Waste Manag. Assoc. 2005, 55 (9), 1367–1388. (68) Radojevic, M. Reduction of Nitrogen Oxides in Flue Gases. Environ. Pollut. 1998, 102 (1, Supplement 1), 685–689. (69) Smoot, L. D.; Hill, S. C.; Xu, H. NOx Control through reburning1. Prog. Energy Combust. Sci. 1998, 24 (5), 385–408. (70) Kasper, J. M.; III, C. A. C.; Cooper, C. D. Control of Nitrogen Oxide Emissions by Hydrogen Peroxide-Enhanced Gas-Phase Oxidation of Nitric Oxide. J. Air Waste Manag. Assoc. 1996, 46 (2), 127–133. (71) Metz, B.; Davidson, O.; Coninck, H.; Loos, M.; Meyer, L. Carbon Dioxide Capture and Storage; IPCC, Cambridge University Press, UK, 2005. (72) Gupta, H.; Benson, S. A.; Fan, L.-S.; Laumb, J. D.; Olson, E. S.; Crocker, C. R.; Sharma, R. K.; Knutson, R. Z.; Rokanuzzaman, A. S. M. Pilot-Scale Studies of NOx Reduction by Activated High-Sodium Lignite Chars: A Demonstration of the CARBONOX Process. Ind. Eng. Chem. Res. 2004, 43 (18), 5820–5827.

258

(73) Gupta, H.; Fan, L.-S. Reduction of Nitric Oxide from Combustion Flue Gas by Bituminous Coal Char in the Presence of Oxygen. Ind. Eng. Chem. Res. 2003, 42 (12), 2536–2543. (74) Illán-Gómez, M. J.; Brandán, S.; Salinas-Martı́nez de Lecea, C.; Linares-Solano, A. Improvements in NOx Reduction by Carbon Using Bimetallic Catalysts. Fuel 2001, 80 (14), 2001–2005. (75) Zhao, Z.; Qiu, J.; Li, W.; Chen, H.; Li, B. Influence of Mineral Matter in Coal on Decomposition of NO over Coal Chars and Emission of NO during Char Combustion☆. Fuel 2003, 82 (8), 949–957. (76) Zhao, Z.; Li, W.; Li, B. Catalytic Reduction of NO by Coal Chars Loaded with Ca and Fe in Various Atmospheres. Fuel 2002, 81 (11–12), 1559–1564. (77) Yamashita, H.; Yoshida, S.; Tomita, A. Local Structures of Metals Dispersed on Coal. 2. Ultrafine FeOOH as Active Iron Species for Steam Gasification of Brown Coal. Energy Fuels 1991, 5 (1), 52–57. (78) Zhong, B. J.; Tang, H. Catalytic NO Reduction at High Temperature by de-Ashed Chars with Catalysts. Combust. Flame 2007, 149 (1–2), 234–243. (79) Illan-Gomez, M. J.; Linares-Solano, A.; Radovic, L. R.; Salinas-Martinez de Lecea, C. NO Reduction by Activated Carbons. 5. Catalytic Effect of Iron. Energy Fuels 1995, 9 (3), 540–548. (80) Chang, H.; Li, B.; Li, W.; Chen, H. The Influence of Mineral Matters in Coal on NO-Char Reaction in the Presence of SO2. Fuel 2004, 83 (6), 679–683. (81) Illan-Gomez, M. J.; Linares-Solano, A.; Radovic, L. R.; Salinas-Martinez de Lecea, C. NO Reduction by Activated Carbons. 4. Catalysis by Calcium. Energy Fuels 1995, 9 (1), 112–118. (82) Fan, L.-S.; Deshpande, N.; Phalek, N. United States Patent: 8877150 - Single- Step Process for the Simultaneous Removal of CO2, SOx and NOx from a Gas Mixture. 8877150, November 4, 2014. (83) Illán-Gómez, M. J.; Linares-Solano, A.; Radovic, L. R.; Salinas-Martínez de Lecea, C. NO Reduction by Activated Carbons. 7. Some Mechanistic Aspects of Uncatalyzed and Catalyzed Reaction. Energy Fuels 1996, 10 (1), 158–168. (84) Pevida, C.; Arenillas, A.; Rubiera, F.; Pis, J. J. Heterogeneous Reduction of Nitric Oxide on Synthetic Coal Chars. Fuel 2005, 84 (17), 2275–2279. (85) Mahajan, O. P.; Yarzab, R.; Walker Jr, P. L. Unification of Coal-Char Gasification Reaction Mechanisms. Fuel 1978, 57 (10), 643–646. (86) Illan-Gomez, M. J.; Linares-Solano, A.; Radovic, L. R.; Salinas-Martinez de Lecea, C. NO Reduction by Activated Carbons. 2. Catalytic Effect of Potassium. Energy Fuels 1995, 9 (1), 97–103. (87) Black, J.; Haslbeck, J. L.; Kuehn, N. J.; Lewis, E. G.; Pinkerton, L. L.; Simpson, J.; Turner, M. J.; Varghese, E.; Woods, M. C. Cost and Performance Baseline for Fossil Energy Plants; DOE/NETL-2010/1397; Pittsburgh, PA, 2010. (88) Wong, D. High Temperature Reactive CO2 Separation from Flue Gas Using Calcium Based Sorbents. Masters Thesis, The Ohio State University: Columbus, 2007.

259

(89) Wang, W. Experimental Results and Computer Simulations for Post-Combustion Carbon Dioxide Removal Using Limestone. Masters Thesis, The Ohio State University: Columbus, 2009. (90) Wang, W.; Ramkumar, S.; Fan, L.-S. Energy Penalty of CO2 Capture for the Carbonation–Calcination Reaction (CCR) Process: Parametric Effects and Comparisons with Alternative Processes. Fuel 2013, 104, 561–574. (91) Ramkumar, S.; Fan, L.-S. Calcium Looping Process (CLP) for Enhanced Noncatalytic Hydrogen Production with Integrated Carbon Dioxide Capture. Energy Fuels 2010, 24 (8), 4408–4418. (92) Ramkumar, S.; Phalak, N.; Fan, L.-S. Calcium Looping Process (CLP) for Enhanced Steam Methane Reforming. Ind. Eng. Chem. Res. 2011, 51 (3), 1186– 1192. (93) Connell, D. P.; Lewandowski, D. A.; Ramkumar, S.; Phalak, N.; Statnick, R. M.; Fan, L.-S. Process Simulation and Economic Analysis of the Calcium Looping Process (CLP) for Hydrogen and Electricity Production from Coal and Natural Gas. Fuel 2013, 105, 383–396. (94) Qiu, K.; Mattisson, T.; Steenari, B.-M.; Lindqvist, O. Thermogravimetric Combined with Mass Spectrometric Studies on the Oxidation of Calcium Sulfide. Thermochim. Acta 1997, 298 (1–2), 87–93. (95) Song, Z.; Zhang, M.; Ma, C. Study on the Oxidation of Calcium Sulfide Using TGA and FTIR. Fuel Process. Technol. 2007, 88 (6), 569–575. (96) Anthony, E. J.; Jia, L.; Qiu, K. CaS Oxidation by Reaction with CO2 and H2O. Energy Fuels 2003, 17 (2), 363–368. (97) Marbán, G.; Garcı́a-Calzada, M.; Fuertes, A. B. Kinetics of Oxidation of CaS Particles in the Regime of Low SO2 Release. Chem. Eng. Sci. 1999, 54 (1), 77– 90. (98) Wheelock, T. D.; Boylan, D. R. Reductive Decomposition of Gypsum by Carbon Monoxide. Ind. Eng. Chem. 1960, 52 (3), 215–218. (99) Diaz-Bossio, L. M.; Squier, S. E.; Pulsifer, A. H. Reductive Decomposition of Calcium Sulfate Utilizing Carbon Monoxide and Hydrogen. Chem. Eng. Sci. 1985, 40 (3), 319–324. (100) Davies, N. H.; Laughlin, K. M.; Hayhurst, A. N. The Oxidation of Calcium Sulphide at the Temperatures of Fluidised-Bed Combustors. Symp. Int. Combust. 1994, 25 (1), 211–218. (101) Uemiya, S.; Kobayashi, T.; Kojima, T. Desulfurization Behavior of Ca-Based Absorbents under Periodically Changing Condition between Reducing and Oxidizing Atmosphere. Energy Convers. Manag. 2001, 42 (15–17), 2029–2041. (102) US EPA. Clean Power Plan Proposed Rule http://www2.epa.gov/carbon- pollution-standards/clean-power-plan-proposed-rule (accessed Nov 24, 2014). (103) Hoffman, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Environmental Applications of Semiconductor Photocatalysis. Chem. Rev. U. S. 1995, 95 (1), 69–96. (104) Adler, S. B. Factors Governing Oxygen Reduction in Solid Oxide Fuel Cell Cathodes. Chem. Rev. 2004, 104 (10), 4791–4843.

260

(105) Poizot, P.; Laruelle, S.; Grugeon, S.; Dupont, L.; Tarascon, J.-M. Nano-Sized Transition-Metal Oxides as Negative-Electrode Materials for Lithium-Ion Batteries. Nature 2000, 407 (6803), 496–499. (106) Kung, H. H. Transition Metal Oxides: Surface Chemistry and Catalysis; Elsevier, 1989. (107) T Punniyamurthy, S. V. Recent Advances in Transition Metal Catalyzed Oxidation of Organic Substrates with Molecular Oxygen. Chem. Rev. 2005, 105 (6), 2329–2363. (108) 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. Energy Fuels 2009, 23 (8), 4182–4189. (109) Lunsford, J. H. Catalytic Conversion of Methane to More Useful Chemicals and Fuels: A Challenge for the 21st Century. Catal. Today 2000, 63 (2–4), 165–174. (110) Wilhelm, D. J.; Simbeck, D. R.; Karp, A. D.; Dickenson, R. L. Syngas Production for Gas-to-Liquids Applications: Technologies, Issues and Outlook. Fuel Process. Technol. 2001, 71 (1–3), 139–148. (111) Kobayashi, Y.; Horiguchi, J.; Kobayashi, S.; Yamazaki, Y.; Omata, K.; Nagao, D.; Konno, M.; Yamada, M. Effect of NiO Content in Mesoporous NiO–Al2O3 Catalysts for High Pressure Partial Oxidation of Methane to Syngas. Appl. Catal. Gen. 2011, 395 (1–2), 129–137. (112) Nagaoka, K.; Okamura, M.; Aika, K. Titania Supported Ruthenium as a Coking- Resistant Catalyst for High Pressure Dry Reforming of Methane. Catal. Commun. 2001, 2 (8), 255–260. (113) Rostrup-Nielsen, J. R. New Aspects of Syngas Production and Use. Catal. Today 2000, 63 (2–4), 159–164. (114) Welty, J. A. B. Apparatus for Conversion of Hydrocarbons. US2550741 A, May 1, 1951. (115) 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 (8), 2186–2199. (116) Fan, L.; Li, F.; Ramkumar, S. Utilization of Chemical Looping Strategy in Coal Gasification Processes. Particuology 2008, 6 (3), 131–142. (117) 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. Energy Fuels 2009, 23 (5), 2498–2505. (118) Luo, S.; Zeng, L.; Xu, D.; Kathe, M.; Chung, E.; Deshpande, N.; Qin, L.; Majumder, A.; Hsieh, T.-L.; Tong, A.; Sun, Z.; Fan, L.-S. Shale Gas-to-Syngas Chemical Looping Process for Stable Shale Gas Conversion to High Purity Syngas with a H2:CO Ratio of 2:1. Energy Environ. Sci. 2014, 7 (12), 4104–4117. (119) Rydén, M.; Lyngfelt, A.; Mattisson, T. Chemical-Looping Combustion and Chemical-Looping Reforming in a Circulating Fluidized-Bed Reactor Using Ni- Based Oxygen Carriers. Energy Fuels 2008, 22 (4), 2585–2597. (120) Go, K. S.; Son, S. R.; Kim, S. D. Reaction Kinetics of Reduction and Oxidation of Metal Oxides for Hydrogen Production. Int. J. Hydrog. Energy 2008, 33 (21), 5986–5995.

261

(121) Abad, A.; García-Labiano, F.; de Diego, L. F.; Gayán, P.; Adánez, J. Reduction Kinetics of Cu-, Ni-, and Fe-Based Oxygen Carriers Using Syngas (CO + H2) for Chemical-Looping Combustion. Energy Fuels 2007, 21 (4), 1843–1853. (122) Deshpande, N.; Majumder, A.; Qin, L.; Fan, L.-S. High-Pressure Redox Behavior of Iron-Oxide-Based Oxygen Carriers for Syngas Generation from Methane. Energy Fuels 2015. (123) Steinfeld, A.; Frei, A.; Kuhn, P. Thermoanalysis of the Combined Fe3O4- Reduction and CH4-Reforming Processes. Metall. Mater. Trans. B 1995, 26 (3), 509–515. (124) García-Labiano, F.; Adánez, J.; de Diego, L. F.; Gayán, P.; Abad, A. Effect of Pressure on the Behavior of Copper-, Iron-, and Nickel-Based Oxygen Carriers for Chemical-Looping Combustion. Energy Fuels 2006, 20 (1), 26–33. (125) Mess, D.; Sarofim, A. F.; Longwell, J. P. Product Layer Diffusion during the Reaction of Calcium Oxide with Carbon Dioxide. Energy Fuels 1999, 13 (5), 999–1005. (126) Chauk, S. S.; Agnihotri, R.; Jadhav, R. A.; Misro, S. K.; Fan, L.-S. Kinetics of High-Pressure Removal of Hydrogen Sulfide Using Calcium Oxide Powder. AIChE J. 2000, 46 (6), 1157–1167. (127) Agnihotri, R.; Chauk, S. S.; Misro, S. K.; Fan, L.-S. High-Pressure Reaction Kinetics of Hydrogen Sulfide and Uncalcined Limestone Powder. Ind. Eng. Chem. Res. 1999, 38 (10), 3802–3811. (128) Hurst, S. Production of Hydrogen by the Steam-Iron Method. Oil Soap 1939, 16 (2), 29–35. (129) Zafar, Q.; Mattisson, T.; Gevert, B. Integrated Hydrogen and Power Production with CO2 Capture Using Chemical-Looping Reforming Redox Reactivity of Particles of CuO, Mn2O3, NiO, and Fe2O3 Using SiO2 as a Support. Ind. Eng. Chem. Res. 2005, 44 (10), 3485–3496. (130) Svoboda, K.; Siewiorek, A.; Baxter, D.; Rogut, J.; Pohořelý, M. Thermodynamic Possibilities and Constraints for Pure Hydrogen Production by a Nickel and Cobalt-Based Chemical Looping Process at Lower Temperatures. Energy Convers. Manag. 2008, 49 (2), 221–231. (131) Nestl, S.; Voitic, G.; Lammer, M.; Marius, B.; Wagner, J.; Hacker, V. The Production of Pure Pressurised Hydrogen by the Reformer-Steam Iron Process in a Fixed Bed Reactor System. J. Power Sources 2015, 280, 57–65. (132) Liu, W.; Ismail, M.; Dunstan, M. T.; Hu, W.; Zhang, Z.; Fennell, P. S.; Scott, S. A.; Dennis, J. S. Inhibiting the Interaction between FeO and Al2O3 during Chemical Looping Production of Hydrogen. RSC Adv. 2014, 5 (3), 1759–1771. (133) Aston, V. J.; Evanko, B. W.; Weimer, A. W. Investigation of Novel Mixed Metal Ferrites for Pure H2 and CO2 Production Using Chemical Looping. Int. J. Hydrog. Energy 2013, 38 (22), 9085–9096. (134) Grasa, G. S.; Abanades, J. C. CO2 Capture Capacity of CaO in Long Series of Carbonation/Calcination Cycles. Ind. Eng. Chem. Res. 2006, 45 (26), 8846–8851.

262