<%BANNER%>

Design and Construction of a Solar Thermogravimeter

Permanent Link: http://ufdc.ufl.edu/UFE0044794/00001

Material Information

Title: Design and Construction of a Solar Thermogravimeter
Physical Description: 1 online resource (108 p.)
Language: english
Creator: Takagi, Midori
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: co2 -- cycle -- energy -- fuel -- solar -- syngas -- thermal -- thermochemical -- thermogravimeter -- water
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Liquid fuels such as gasoline can be produced from water and recycled CO2, using a two-step thermochemical cycle powered by concentrated solar energy. Unlike with fossil-derived fuels, the combustion of these fuels does not contribute to the atmospheric accumulation of CO2, as they are made from CO2 extracted from the air. The two-step thermochemical cycle consists of an oxidation step and a solar reduction step. In the oxidation step, steam and CO2 are reacted with a low-valence metal oxide to produce H2, CO and a higher-valence metal oxide. The H2 and CO gases are used to synthesize liquid hydrocarbon fuels. In the solar reduction step, the higher-valence metal oxide is thermally dissociated into the original lower-valence metal oxide and oxygen, the former of which is recycled into the oxidation step. Overall, the cycle bears a beautiful resemblance to photosynthesis. Research efforts are underway at the University of Florida to realize this innovative technology. One of the tasks is to investigate the fundamental mechanisms driving the reduction and oxidation reactions, and develop reaction kinetics models. Thermogravimetry, an analytical technique whereby the weight change of a reacting sample is measured as a function of time and temperature, is employed as a suitable method for this purpose. Due to the high temperatures required in the reduction step (~1500 ºC ), a unique solar thermogravimeter capable of attaining heating rates higher than those of conventional thermogravimeters was designed and is currently being constructed. The design of the solar thermogravimeter is presented in this thesis.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Midori Takagi.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Petrasch, Joerg.
Local: Co-adviser: Hahn, David W.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044794:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044794/00001

Material Information

Title: Design and Construction of a Solar Thermogravimeter
Physical Description: 1 online resource (108 p.)
Language: english
Creator: Takagi, Midori
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: co2 -- cycle -- energy -- fuel -- solar -- syngas -- thermal -- thermochemical -- thermogravimeter -- water
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Liquid fuels such as gasoline can be produced from water and recycled CO2, using a two-step thermochemical cycle powered by concentrated solar energy. Unlike with fossil-derived fuels, the combustion of these fuels does not contribute to the atmospheric accumulation of CO2, as they are made from CO2 extracted from the air. The two-step thermochemical cycle consists of an oxidation step and a solar reduction step. In the oxidation step, steam and CO2 are reacted with a low-valence metal oxide to produce H2, CO and a higher-valence metal oxide. The H2 and CO gases are used to synthesize liquid hydrocarbon fuels. In the solar reduction step, the higher-valence metal oxide is thermally dissociated into the original lower-valence metal oxide and oxygen, the former of which is recycled into the oxidation step. Overall, the cycle bears a beautiful resemblance to photosynthesis. Research efforts are underway at the University of Florida to realize this innovative technology. One of the tasks is to investigate the fundamental mechanisms driving the reduction and oxidation reactions, and develop reaction kinetics models. Thermogravimetry, an analytical technique whereby the weight change of a reacting sample is measured as a function of time and temperature, is employed as a suitable method for this purpose. Due to the high temperatures required in the reduction step (~1500 ºC ), a unique solar thermogravimeter capable of attaining heating rates higher than those of conventional thermogravimeters was designed and is currently being constructed. The design of the solar thermogravimeter is presented in this thesis.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Midori Takagi.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Petrasch, Joerg.
Local: Co-adviser: Hahn, David W.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044794:00001


This item has the following downloads:


Full Text

PAGE 1

1 DESIGN AND CONSTRUCTION OF A SOLAR THERMOGRAVIMETER By MIDORI TAKAGI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DE GREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Midori Takagi

PAGE 3

3 To my parents

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I would like to express my deepest gratitude to wards my supervisor, Dr. Jrg Petrasch, for his warm guidance and immense patience throughout my entire graduate degree. I would like to thank him for the countless times he spent answering my often repetitive questions and, furthermore, for generously taking time out of his incredibly busy schedule to even listen to my personal dilemmas. Aside fr om being one of the friendliest professors I have known, Dr. Petrasch also instilled in us a sense of discipline and responsibility in our actions and, for me, particularly, the importance of setting priorities and getting work done efficiently. His rigoro us and enthusiastic lectures in Advanced Solar Energy and Radiative Heat Transfer were just as interesting as they were challenging, and stimulated my curiosity towards solar energy even further. I would like to express my sincerest gratitude to wards my c o advisors, Dr. James Klausner and Dr. David Hahn, for their great kindness and patience in supervising me. As well as providing me with academic advice, Dr. Klausner helped me to identify the core of my problems and, most importantly, taught me not to for get the motivation for and positive mindset towards the common goal. Dr. Hahn always provided me with very generous encouragement and made sure that my academic life at UF was running smoothly. I would also like to thank Dr. Renwei Mei and Dr. Hitomi Green slet, who also provided me with kind words of encouragement and inspiration every time I spoke to them, and Dr. Herbert Ingley, who helped me to fulfill my curiosity towards solar energy through his humor filled and practice oriented lectures. I thank all my labmates: Anupam Akolkar, Ben Erickson, Abhishek Singh and Michael Tan, who gave me great company at an otherwise lonely Energy Park building.

PAGE 5

5 The ir friendliness and funny sense of humor always kept me in high spirits. I thank all the members of the sol ar fuels team, who constantly provided me with helpful advice and moral s upport. They are truly the nicest people in the world. Their clear thought processes, effective ways of communication and practical expertise gave me plenty to learn from. Particular thanks go to the following people who, on top of providing me with much advice in general, contributed to the TG in specific ways: Ben Erickson, who built the solar simulator which allows operation of the solar themogravimeter to be possible; to Ayyoub Meh dizadeh, who proposed the idea of using a wireless method for the transmission of temperature read ings; to Ben Greek, who pro posed the idea and design of the wireless temperature transmission circuit; to Kyle Allen, who provided helpful suggestions on desi gning the sample crucible; to Abhishek Singh, who provided much help on simulating radiative heat transfer to the TG crucible and to Nick Au Yeung, who provided much help and advice on coupling the TG with the mass spectrometer I would particularly like to thank Ms. Genevieve Blake, who always helped me spontaneously with ordering equipment. I thank the people involved in manufacturing custom made parts for the thermogravimeter; in particular, Ms. Cheryl Stanasek, Ms. Margaret Bishop and Mr. Driscoll at A bbess Instruments, who kindly put up with copious requests and e mail exchanges, and Mr. Paul Bergsma at Manufacturing Tooling & Engineering, who manufactured the most complex parts. I thank Mr. Jeff Studstill for kindly helping me with moving heavy items to the Energy Park. I cannot go without thanking the friends I gained at UF, who made my life in Gainesville into a most wonderful experience. I wish they would realize just how much

PAGE 6

6 of an impact they have had on my life, and how much I appreciated their company. I sincerely hope that our friendships will extend over a lifetime. The U.S. Government s Department of Energy is gratefully acknowledged for their funding towards this project. I would also like to gratefully acknowledge the College Women s Associ ation of Japan, who kindly provided me with a scholarship which made my studies in the U.S. possible. Finally, I thank my parents, who have always supported me unconditionally and allowed me to pursue whatever I wanted in life. I thank my grandparents, who gave me wise advice on life which always made good sense later on. Special thanks go to Jay who, despite being on a different continent for the last two years, has always been there for me every minute of the way

PAGE 7

7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 1.1 Solar Energy ................................ ................................ ................................ ..... 15 1.2 Synthesis of Clean, Liquid Hydrocarbon Fuels u sing Solar E nergy .................. 16 1.3 Thermal D issociation of H 2 O and CO 2 u sing Concentrated Solar E nergy ........ 19 1.3.1 Direct T hermolysis ................................ ................................ ................... 19 1.3.2 D issociation of H 2 O and CO 2 v ia Multistep Thermochemical C ycles ....... 20 1.3.3 D issociation of H 2 O and CO 2 u sing Two S tep Metal Oxide C ycles ......... 21 1.4 Solar Thermochemical Fuel Production P roject at the University of Florida ..... 24 1.5 Reaction Kinetics Studies and T hermogravimetry ................................ ............ 26 1.6 Solar T hermogravimeter ................................ ................................ ................... 28 1.6.1 Design Concept and A dvantages o ver Conventional T hermogravimeters ................................ ................................ ....................... 28 1.6.2 Comparison with Other Solar T hermogravimeters ................................ .. 30 2 EXPERIMENTAL DESIGN ................................ ................................ ..................... 32 3 THERMOGRAVIMETER DESIGN ................................ ................................ .......... 35 3.1 Overview of the TG A ssembly ................................ ................................ .......... 35 3.2 Component D etails ................................ ................................ ........................... 39 3.2.1 Vacuum C hamber ................................ ................................ .................... 39 3.2.2 Microbalance ................................ ................................ ........................... 41 3.2.3 Sample Crucible, Crucible Suspension and Thermocouple A ssembly .... 44 3.2.4 Wireless Circuit for Transmission of Temperature R eadings ................... 47 3.2.5 Quartz V essel ................................ ................................ .......................... 50 3.2.5.1 Thermal and mechanical properties ................................ ............... 50 3.2.5.2 Determination of wall thickness ................................ ...................... 52 3.2.5.3 Implosion hazard ................................ ................................ ............ 54 3.2.6 Radiation S hield ................................ ................................ ...................... 54 3.2.7 Cl amps and Vacuum S eals ................................ ................................ ..... 58 3.2.7.1 Functions ................................ ................................ ....................... 58 3.2.7.2 Seals ................................ ................................ .............................. 62 3 .2.7.3 Achievement of a sufficient vacuum seal ................................ ....... 62

PAGE 8

8 3.3 Using the TG for Investigation of Oxidation Reaction K inetics .......................... 65 4 THERMO GRAVIMETER CHARACTERISTICS ................................ ...................... 67 4.1 Time to Reach Target T emperature ................................ ................................ .. 67 4.1.1 One D imensional Transient Heat Conduction P roblem ........................... 67 4.1.2 Determination of the Effective Thermal C ......................... 69 4.1.2.1 Thermal conductivity of conduction heat transfer thro ugh the particle bed ................................ ................................ ............................. 70 4.1.2.2 Thermal conductivity of radiative heat transfer through the particle bed ................................ ................................ ............................. 72 4.1.3 Solution of th e O ne D imensional Transient Heat Conduction P roblem ... 73 4.2 Recommended Sample M ass ................................ ................................ ........... 77 4.2.1 Minimum Sample M ass ................................ ................................ ........... 77 4.2.2 Maximum Sample M ass ................................ ................................ .......... 78 4.3 Maximum Temperatures Attainable w ith t he TG ................................ ............... 79 5 CONCLUSION ................................ ................................ ................................ ........ 82 APPENDIX A CAD DRAWINGS ................................ ................................ ................................ ... 83 B BILL OF MATERIALS ................................ ................................ ........................... 101 C MATLAB CODE ................................ ................................ ................................ .... 103 LIST OF REFERENCES ................................ ................................ ............................. 105 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 108

PAGE 9

9 LIST OF TABLES Table page 3 1 Thermal and mechanical properties of quartz and borosilicate glass ................. 51 3 2 Wall thickness of quartz vessel calcu lated for a range of safety factors ............. 54 3 3 Thermal properties of aluminum and other metals ................................ ............. 58 3 4 Characteristics of some elasto meric materials ................................ ................... 62 3 5 % ................................ ................................ ................................ ........................ 65 B 1 Bill of materials for solar thermogravimeter ................................ ...................... 101

PAGE 10

10 LIST OF FIGURES Figure page 1 1 Synthesis of liquid hydrocarbon fuels using solar energy, H 2 O and CO 2 captured from the air. ................................ ................................ ......................... 18 1 2 Capture of CO 2 from the air using solar energy, demonstrated by Steinfeld et al. ................................ ................................ ................................ ........................ 19 1 3 Clean and sustainable solar thermochemical fuel production. ............................ 26 1 4 Thermogravimetric curves showing the thermal decomposition of CaCO 3 ........ 28 1 5 University o ................................ ................... 30 2 1 Experimental layout. ................................ ................................ .......................... 34 3 1 TG assembly. ................................ ................................ ................................ ..... 36 3 2 Partially cross sectioned view of the TG assembly ................................ ............ 37 3 3 Exploded view of the TG assembly. ................................ ................................ ... 38 3 4 V acuum chamber. ................................ ................................ .............................. 40 3 5 Microbalance and electronic module. ................................ ................................ 42 3 6 Position of the microbalance within the vacuum chamber. ................................ 43 3 7 Fixing the microbalance inside the vacuum chamber. ................................ ........ 43 3 8 Sample crucible and suspension mechanism. ................................ .................... 46 3 9 Sample crucible design. ................................ ................................ ..................... 47 3 10 Issue with transmitting thermoelectric signals. ................................ .................... 49 3 1 1 Wireless solution. ................................ ................................ ............................... 49 3 12 Quartz vessel. ................................ ................................ ................................ ..... 50 3 13 Hoop and axial stresses present in the vessel wall. ................................ ........... 53 3 14 Radiation shield. ................................ ................................ ................................ 56 3 15 Components of radiation shield. ................................ ................................ ......... 56 3 16 Route of cool ing water. ................................ ................................ ....................... 57

PAGE 11

11 3 17 Weld locations. ................................ ................................ ................................ ... 57 3 18 Slots created for assembly of the shield components. ................................ ........ 58 3 19 ................................ ................................ ................ 60 3 20 Exploded view of the components assembled above the vacuum chamber lid .. 60 3 21 Locations sealed against air leakage. ................................ ................................ 61 3 22 ............................. 61 3 23 Compression of gaskets using clamps. ................................ .............................. 64 3 24 Compression ratio of rubber gaskets as a function of their Shore hardness. ..... 65 3 25 Using the TG for the CO 2 splitting reaction. ................................ ........................ 66 3 26 The sample powder bed, heated radiatively from above. ................................ ... 68 3 27 Representation of the 1D transient heat conduction problem. ............................ 69 3 28 Modes of heat transfer through the particle bed. ................................ ................ 70 3 29 Com parison of experimental results with various analytical models developed for kcond. ................................ ................................ ................................ ........... 71 3 30 Comparison of experimental results with empirical models developed for kcond. ................................ ................................ ................................ ................ 72 3 31 Temperature variation with depth and time for the 5 mm particle bed described in 4.1.3. ................................ ................................ ............................. 75 3 32 Relationship between bed thickness and time s taken for bed surface and entire bed to reach 1500 C ................................ ................................ ................ 76 3 33 Irradiation by the solar simulator. ................................ ................................ ........ 80 3 34 Use of mirrors to re direct radiation onto the surface of the sample bed. ........... 81 A 1 Cross sectional drawing of TG assembly. ................................ .......................... 83 A 2 Top view of vacuum c hamber and vibration isolation mass. ............................... 84 A 3 Front view of vacuum chamber. ................................ ................................ .......... 85 A 4 Left side view of vacuum chamber. ................................ ................................ .... 86 A 5 Vacuum chamber lid. ................................ ................................ .......................... 87

PAGE 12

12 A 6 Right side view of vacuum chamber. ................................ ................................ .. 88 A 7 R ight side view of vacuum chamber and vibration isolation mass. ..................... 89 A 8 Crucible. ................................ ................................ ................................ ............. 90 A 9 Crucible platform. ................................ ................................ ............................... 91 A 10 Crucible suspension. ................................ ................................ .......................... 92 A 11 Suspension holder. ................................ ................................ ............................. 93 A 12 Quartz vessel. ................................ ................................ ................................ ..... 94 A 13 Base plate. ................................ ................................ ................................ .......... 95 A 14 Cooling cylinder. ................................ ................................ ................................ 96 A 15 Cylinder cap. ................................ ................................ ................................ ....... 97 A 16 North clamp. ................................ ................................ ................................ ....... 98 A 17 Clamp rod. ................................ ................................ ................................ .......... 99 A 18 South clamp. ................................ ................................ ................................ ..... 100

PAGE 13

13 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requ irements for the Degree of Master of Science DESIGN AND CONSTRUCTION OF A SOLAR THERMOGRAVI METER By Midori Takagi August 2012 Chair: Jrg Petrasch Major: Mechanical Engineering Liquid fuels such as gasoline can be produced from water and recycled CO 2 using a two step thermochemical cycle powered by concentrated solar energy. Unlike with foss il derived fuels, the combustion of these fuels does not contribute to the atmospheric accumulation of CO 2 as they are made from CO 2 extracted from the air. The two step thermochemical cycle consists of an oxidation step and a solar reduction step In the oxidation step, steam and CO 2 are reacted with a low valence metal oxide to produce H 2 CO and a higher valence metal oxide. The H 2 and CO gases are used to synthesize liquid hydrocarbon fuels. In the solar reduction step, the higher valence metal oxide i s thermally dissociated into the original lower valence metal oxide and oxygen, the former of which is recycled into the oxidation step. Overall, the cycle bears a beautiful resemblance to photosynthesis. Research efforts are underway at the University of Florida to realize this innovative technology. One of the tasks is to investigate the fundamental mechanism s driving the reduction and oxidation reactions, and de velop reaction kinetics models. Thermogravimetry, an analytical technique whereby the weight c hange of a reacting sample is measured as a function of time and temperature, is employed as a suitable

PAGE 14

14 method for this purpose. Due to the high temperatures required in the reduction step (~1500 C ), a unique solar thermogravimeter capable of attaining h eating rates higher than those of conventional thermogravimeters was designed and is currently being constructed. The design of the solar thermogravimeter is presented in this thesis.

PAGE 15

15 CHAPTER 1 INTRODUCTION 1.1 Solar Energy Climate change caused by the c onsumption of fossil fuels continues to pose a serious threat to human civilization. It has been widely accepted that in order to avoid an irreversible catastrophe, the increase in global temperature from pre industrial levels must be kept within 2 C Stu dies have shown that in order to fulfill this 2 C target, adopted by more than a hundred nations around the world, global CO 2 emissions must be reduced to at least 50 % of 1990 levels by 2050 [ 1 ], [2 ]. The threat of climate change is imminent, and this tog ether with the diminution of fossil fuel reserves continue s to urge the development of clean alternatives to fossil fuels. Among the various kinds of clean, renewable energy available, solar energy is certainly the most abundant. In the U.S., the potentia l capacities of non solar renewable s hardly equal the nation s energy consumption; however, it is estimated that filling an area of land approximately equal to the size of Arizona with concentrating solar power facilities can provide enough energy to meet the country s energy needs [3] Similarly, it has been estimated that installing concentrating solar power facilities in an area of the Saharan desert equal to the size of Germany can provide more than enough power to meet the energy demands of E urope and North Africa combined As plentiful as it is, however, solar energy suffers from intermittency and regional discrepancies in availability. To make solar energy more useful, ways of storing it for use during periods of cloudy weather and distributing it to areas of low insolation must be developed.

PAGE 16

16 1.2 Synthesis of C lean, L iquid H ydrocarbon F uels u sing S olar E nergy An incredibly fascinating solution to the storage and distribution issues pertaining to solar energy is the synthesis of clean, liquid hydrocarb on fuels using solar thermal energy [ 4 ]. The idea first involves using concentrated s olar energy to dissociate H 2 O into hydroge n and oxygen, and CO 2 into carbon monoxide and oxygen. The hydrogen and carbon monoxide gases (whose mixture is termed synthetic gas or syngas ) produced in these reactions are then used to synthesize liquid hydrocarbon fuels via the Fischer Tropsch process: 1 1 where alkanes with carbon numbers ranging between n = 7~22 are typical components of liquid fuels such as gasoline (n = 7~11), jet fuel (n = 9~15) and diesel (n = 12~22) [ 5 ]. Thus, solar energy is stored chemically in liquid fuels, which can be easily stored and distributed using already existing infrastructure. While solar derived hydrogen can itself be directly used as a fuel, its high diffusivity, high flammability and low energy density per volume still presents challenges for its storage and transportation. By creating liquid fuels from solar derived hydrogen an d carbon monoxide, however, su ch challenges are bypassed. Since burning these synthetic fuels still releases carbon dioxide into the air, the question may immediately arise as to whether or not they solve the problem of global warming. Clearly, if the CO 2 used to create these fuels is obtained from the combustion of fossil fuels, the net result of burning these synthetic fuels is that more CO 2 is added to the atmosphere. However, if CO 2 collected directly from the air is used the net effect is that there is no addition of CO 2 to the at mosphere. CO 2 is thus recycled as s hown in

PAGE 17

17 Figure 1 1 ve ry much akin to the way plants recycle CO 2 through photosynthesis. Solar derived hydrocarbon fuels made using CO 2 collected from the air are thus 100 % clean. Various technologies for capturing CO 2 f rom the air have been investigated [ 6 ] ,[7], [8] ,[9] In particular, Steinfeld et al. [6] have demonstrated a method of capturing CO 2 from the air using solar energy based on a thermochemical cycle consisting of a carbonation step and a calcination step. Th e cycl ic process is shown in Figure 1 2 wh ich is roughly reproduced from [6] In the carbonation step, air is passed through a fluidized bed of calcium oxide particles heated to 400 C The calcium oxide particles react with CO 2 in the air to form calcium carbonate, depleting the air stream of CO 2 The calcium carbonate particles are subsequently heated to 800 C in the calcination step where they release a stream of CO 2 gas and return to calcium oxide, which is used again in the carbonation step. In thei r study, h eating of the particles was conducted using arc lamps simulating concentrated solar energy.

PAGE 18

18 Figure 1 1 Synthesis of liquid hydrocarbon fuels using solar energy, H 2 O and CO 2 captured from the air. Combustion CO 2 H 2 O Solar dissociation CO + oxygen H 2 + oxygen Fischer Tropsch Process C n H (2n+2) (Liquid Fuels) Atmosphere

PAGE 19

19 Figure 1 2 Capture of CO 2 from the air using solar energy, demonstrated by Steinfeld et al. [6] 1.3 Thermal D issociation of H 2 O and CO 2 u sing C oncentrated S olar E nergy 1.3.1 Direct T hermolysis As described in 1.2, the synthesis of solar derived synth etic fuels first requires the dissociation of H 2 O and CO 2 using concentrated solar energy. One theoretically possible method of doing th is is direct thermolysis where H 2 O and C O 2 are thermally dissociated b y heating under concentrated solar energy. The re actions are expressed by: 1 2 1 3 CaO + CO 2 3 (at 400 C) CaO particles CaCO 3 particles CaCO 3 2 (at 800 C) CO 2 depleted air CO 2 + steam Air (containing CO 2 ) Steam Pure CO 2 Condensation of H 2 O Solar energy Carbonation step Calcination st ep

PAGE 20

20 Despite their apparent simplicity, direct thermolysis suffers from at least two major issues that hind er their practical feasibili ty [ 10 ]. The first issue is that the thermolysis of water and carbon dioxide require temperatures well above 2500 K for any significant yields of hydrogen and carbon monoxide to be obtained. While temperatures in excess of 3600 K are achievable with modern solar concentrators this presents difficulties in reactor design due to the limited range of materials that can remain physically and chemically stable at such high temperatures. The use of exotic or advanced materials increas es the cost of the solar reactor and thus of the fuel itself. The second issue is the need to separate the H 2 and CO gases from oxygen. The gases must be separated at the reaction temperature in order to prevent them from recombining with oxygen; however, no effective methods have so far been demonstrated. Ko gan et al. [ 11 ] a ttempted the separation of H 2 gas from oxygen using a porous membrane made from sintered zirconia powders; however, it was found that sintering of the powders proceeded further at the h igh temperatures, causing the pores to close While Ld [ 12 ] and Olalde [ 13 ] have demonstrated the separation of H 2 gas from oxygen by r apidly quenching the gases; quenching introduces a significant drop in the overall efficiency of the H 2 production proc ess and produces an explosive gas mixture. 1.3.2 D issociation of H 2 O and CO 2 via M ultistep T hermochemical C ycles The problems encountered wit h the direct thermolysis of H 2 O and CO 2 can be solved using multistep thermochemical cycles first studied by Funk and Rei nstrom [ 14 ]. In a multistep thermochemical cycle, more than one reaction is employed to dissociate water into hydrogen and oxygen, or carbon dioxide into carbon monoxide and oxygen while the net result remains the same as with direct thermolysis Temperatures required

PAGE 21

21 for these reactions are lower than those required for direct thermolysis, which alleviate s the issue r egarding selection of thermally stable reactor materials. Moreover, hydrogen and carbon monoxide are produced separately from oxygen eliminating the need for high temperature gas separation. One promising example of such multistep cycles developed for the dissociation of H 2 O is the sulfur iodine cycle or S I cycl 15 ], which consists of the following reactions: 1 4 1 5 1 6 Here, hydrogen and oxygen are produced in the third and second steps, respectively. In the third step, iodine is separated from the hydrogen gas by condensation and recycle d back into the first step. Likewise, water and sulfur dioxide produced in the second step are separated from the oxygen gas by condensation and fed back into the first step. The attractiveness of the S I cycle is derived from the low temperatures required for the endothermic second and third steps, which allow for efficient operation of the cycle using concentrated solar energy. The S I cycle, however, is still met with challenges. The high reactivit y of sulfur and iodine causes the formation of unwanted s ide products, which must be eliminated by supplying extra heat, leading to a reduction in process efficiency. Furthermore, reactor materials that are resistant to corrosion by these elements are required. 1. 3.3 D issociation of H 2 O and CO 2 u sing T wo S tep M e tal O xide C ycles Among the many multistep thermochemical cycles investigated are two step metal oxide cycles, initially proposed by Na kamura [ 16 ]. T hese cycles consist of an

PAGE 22

22 oxidation step and a reduction step In the exothermic oxidation step, a metal is reacted with steam to produce metal oxide and hydrogen, or reacted with carbon dioxide to produce metal oxide and carbon monoxide. This step is expressed as follows: 1 7 1 8 The metal oxides thus produced are then fed to the reduction step, where they are heated using concentrated solar energy and reduced to the original metal as follows: 1 9 The metal is subseque ntly recycled to the oxidat ion step, where it is used to dissociate yet more steam and CO 2 Again, H 2 and CO are produced separately from oxygen, thereby eliminating the issues associated with high temperature gas separation. The endothermic reduction step proceeds at temperatures l ower than those required for the direct thermolysis of H 2 O and CO 2 making these cycles relatively more feasible in terms of reactor material selection. Furthermore, in comparison with multistep cycles that i nvolve three or more steps, two step metal oxide cycles provide more simplicity, reducing the number of stages at which reaction products need to be separated and thus resulting in higher system efficiencies In some cases, metal oxides can be used instead of fresh metals in the oxidation step, as follo ws: 1 10 1 11 The metal oxide produced is of a higher valence to the reactant metal oxide. The reduction step is thus:

PAGE 23

23 1 12 An example o f a two step metal oxide cycle investigated for the production of hydrogen is the Zn/Z nO cycle [ 17 ]: 1 13 1 14 A n other is the FeO/Fe 3 O 4 cycle [ 16 ]: 1 15 1 16 In the Zn/ZnO cycle, the zinc recovered in the reduction step is in gaseous state and must be quenched in order to be separated from the oxygen gas, a process that is challengi ng and results in energy losses. Meanwhile, this issue is avoided in the FeO/Fe 3 O 4 cycle, where the wstite (FeO) recovered in the reduction step is in liquid state and can be separated easily from the oxygen gas. Temperatures required for reduction steps can be further lowered by conducting the step under low pressure. According to Le Chatelier s principle, if a chemical system at equilibrium experiences a decrease in pressure, the equilibrium shifts to counteract the change; i.e., any reaction whose resu lt is to increase the pressure of the system is favored. The reduction step of the FeO/Fe 3 O 4 cycle, for example, results in a pressure increase due to the evolution of oxygen gas. Therefore, if this step is conducted under vacuum, equilibrium will shift in the forward direction, favoring the dissociation of magnetite (Fe 3 O 4 ) to wstite and oxygen more than the recombination of wstite a nd oxygen to form magnetite. Thus, the reaction temperature necessary for the reduction

PAGE 24

24 step is lowered. Furthermore, mixed iron oxides such as Ni Fe Mn Fe Zn Fe and Mg Fe oxides or doped ferrites can have lower reduction temperature s while having increased melting points [18],[19],[20],[21],[22] As metal oxides that remain solid throughout the cycle are more advantageo us due to their ability to be used with high surface area (for instance, as powders and pellets or as coatings on high surface area structures such as honeycombs), this makes doped ferrites favorable candidate s for use in two step cycles. 1.4 Solar T hermoc hemical F uel P roduction P roject at the University of Florida Since December 2011, a solar thermochemical fuel production project funded by the U.S. Department of Energy has been underway at the University of F l orida. The project targets the development of a high temperature chemical reactor that will utilize concentrated solar energy to produce syngas, which can be used to synthesize liquid hydrocarbon fuels such as petroleum. The production of syngas will rely solely upon water and recycled CO 2 as feedstoc k, and will utilize two step thermochemical cycles based on metal oxides. Challenges to be addressed in the project include lowering the reduction temperature to below 1500 C in order to alleviate material selection issues and achieve high system efficien cies. In order to achieve this, doped ferrites will be investigated in particular, and the reduction step will be performed under vacuum. A flow diagram summarizing the envisioned solar fuel production and utilization scheme is shown in F i gure 1 3 Two ty pes of solar reactor configurations exist : one in which the metal oxide is heated by being directly exposed to solar radiation, and another in which it is heated indirectly : i.e., it is surrounded by or contained within a medium which is heated by solar ra diation. In the project at the University of Florida, a reactor of the latter configuration is

PAGE 25

25 used. While the indirectly heating solar reactor requires more solar energy to heat the metal oxide to the desired temperature, it avoids many issues associated with the directly heating reactor. The directly heating reactor requires, inevitably, a transparent glass window which allows direct access of concentrated solar radiation to the metal oxide. This creates difficulties for reactor operation, as glass window s are susceptible to thermal shock and need to be prevented from being contaminated by the fluidized metal oxide particles and condensable gases in the reacto r [ 23 ].

PAGE 26

26 Figure 1 3 Cle an and sustainable solar thermochem ical fuel production. 1.5 Reaction K inetics S tudies and T hermogravimetry One of the key tasks of the project is to identify the fundamental mechanisms driving the oxidation and reduction reactions, and develop kinetic models for the reactions. One laborato ry technique that is used to measure reaction rates is thermogravimetry In thermogravimetry, a sample is heated and its change in mass is L ow valence metal oxide Low valence metal oxide + H 2 O + CO 2 High valence metal oxide High valence metal oxide + CO + H 2 Liquid fuels H 2 O CO 2 High valence metal oxide Low valence metal oxide + O 2 Combustion Carbon capture Oxidation step Reduction step Atmos phere Solar energy Solar energy Solar energy

PAGE 27

27 recorded by a microbalance inside a thermogravimeter The mass change is recorded as a function of time and temperatu re. For instance, the decomposition of calcium carbonate (CaCO 3 ) to calcium oxide (CaO) and carbon dioxide (CO 2 ) produces thermogravimteric curves which look like Figure 1 4 A and Figure 1 4B (fi gures are roughly reproduced from [ 24 ] and [ 25 ]). Figure 1 4 A s hows the change in the mass of the CaCO 3 sample as a function of temperature. At around 800 C the mass shows a sudden decrease, indicating the occurrence of thermal decomposition. Meanwhile, Figure 1 4 B shows the change in the mass of the CaCO 3 sample a s a function of time, as the temperature is raised linearly to around 800 C and held constant. The mass decreases gradually as a result of thermal decomposition, and levels off after around 40 minutes, at which point the decomposition is complete. The gra dient of the mass curve provides data on the reaction rate, and the difference between the initial and final mass values provides information regarding the total mass that has reacted. Similarly, thermogravimetry can be used to analyze the rate and extent of the oxidation and reduction reactions, which are accompanied by an increase and decrease in sample weight, respectively. In particular, for the reduction step, thermogravimetric curves obtained under various levels of vacuum can provide insightful infor mation regarding the effect of low pressure on dissociation temperature and reaction kinetics. A thermogravimeter was thus desired for carrying out the kinetics studies. As conventional thermogravimeters typically have heating rates on the order of 10 C /m in (corresponding to time lengths on the order of an hour for samples to be heated to as high as 1500 C ), the design and construction of a thermogravimeter capable of attaining higher heating rates was sought after.

PAGE 28

28 Figure 1 4 Thermogravimetric curves showing the th ermal decomposition of CaCO 3 A ) sample mass as a function of temperature [24] B ) sample mass as a function of time [25] 1.6 Solar T hermogravimeter 1.6.1 Design C oncept and A dvantages o ver C onventional T hermogravimeters The central concept behind the design of the unique thermogravimeter is to allow the test sample to be heated radiatively using University of Florida s 5 6 kWe solar simulator ( Figure 1 5 [26] ). The solar simulator, consisti ng of seven xenon arc lamps 100 90 80 70 60 50 200 0 400 80 800 1000 1200 0 5 10 15 20 25 30 35 40 45 100 90 600 70 60 50 1400 1200 1000 800 600 400 200 0 Temperature (C) T ime ( min ) Mass ( %) Mass (%) Temperature (C) A ) B ) Mass Temperature

PAGE 29

29 and seven elliptical mirrors which focus their radiation onto a single spot, is capable of exposing targets to heat fluxes in excess of 5000 kW/m 2 Due to the sheer size of the radiative heat flux, powder samples of sizes typica lly used in conventional thermogravimeters (powder bed thicknesses less than 5 m m ) can, with the solar simulator, be heated to 1500 C at rates exceeding 100 K/s, while their top surfaces can be heated much faster at rates exceeding 1000 K/s. The solar th ermogravimeter also provides two advantages over conventional thermogravimeters. While conventional thermogravimeters are typically unable to reach temperatures above 2000 C the solar thermogravimeter can be heated to a maximum of 3063 K (2790 C corre sponding to the steady state temperature reached by a blackbody exposed to a heat flux of 5000 kW/m 2 ). The temperature range under which reaction kinetics studies can be performed is thus wider. Furthermore, the wavelength spectrum of the radiation emitted by the solar simulator closely resembles that of the solar radiation spectrum. This enables the study of the metal oxide s absorptivity towards real solar radiation. If the metal oxide is to be eventually used in a solar reactor where it is heated by dire ct exposure to solar radiation, the absorptivity of the metal oxide towards real solar radiation allows one to determine the exact quantity of solar radiation required to heat it to the desired temper ature. While t his may not be of immediate significance t o the University of Florida s solar thermochemical fuel production project, as the planned solar reactor is an indirectly heating reactor : i.e., the metal oxide is contained inside heated tubes, rather than exposed directly to radiation, this characteristi c of the solar simulator (and thus of the thermogravimeter) can prove convenient for any future experiments in which a material s surface behavior towards the solar radiative spectrum needs to be known.

PAGE 30

30 Figure 1 5. University of Florida s 56 kWe solar si mulator. [26] 1.6.2 Comparison with O ther S olar T hermogravimeters Schunk and Steinfe ld [ 27 ] h ave successfully investigated the reaction kinetics of the solar dissociation of ZnO to zinc and oxygen, using a solar thermogravimeter built at the Paul Scherrer Institute in Switzerland. The solar thermogravimeter at PSI has a resolution of 0.01 g and is capable of reaching temperatures above 2500 K at heating rates faster than 1000 K/s. The ZnO dissociation experiments were conducted using a solar furnace, utiliz ing real solar radiation focused onto the sample by means of a heliostat and a paraboloidal concentrator. Product gas compositions were analyzed using gas chromatography, IR based detectors and thermal conductivity based detectors. Temperatures were measur ed using a pyrometer and a thermocouple. The solar thermogravimeter described in this thesis aims to achieve a resolution of g The peak radiative flux concentration produced by the UF solar simulator (5000 suns) is equal to the solar flux concentration obtained at PSI, and allows comparable temperatures and heating rates to be achieved in the solar thermogravimeter.

PAGE 31

31 Further more, as the solar simulator is capable of providing a constant heat flux unlike real solar radiation which fluctuates with time, experiments may be carried out without temperature fluctuations occurring. The UF solar thermogravimeter also aims to enable t he quantitative analysis of oxygen arising from the dissociation reaction using a mass spectrometer, which will supplement the sample weight loss measurements in the analysis of reaction rates. The UF solar thermogravimeter utilizes a novel method for wire lessly transmitting sample temperature readings from a thermocouple This method solves temperature acquisition problems which have been encountered by other solar thermogravimeters currently under construction in the U .S The details of the wireless tempe rature transmission system are described in the following chapter.

PAGE 32

32 CHAPTER 2 EXPE RIMENTAL DESIGN The experimental design is shown in Figure 2 1 T he thermogravimeter (abbreviated as TG from here onwards) consists of a vacuum chamber (1) housing a 10 g resolution microbalance (2). The powder sample is placed in a zirconia crucible (3) suspended at a height of approximately 30 cm above the microbalance, and is heated radiatively from above. The heating is achieved by reflecting focused radiation produc ed by the solar simulator (4) onto the crucible by means of a water cooled, flat plate mirror (5). A quartz vessel (6) encases the sample crucible, allowing access of radiation to the sample while sealing the sample inside the vacuum. A water cooled radiat ion shield (7) protects the vacuum chamber and balance from exposure to any radiation which may bypass the crucible. The sample mass is transmitted from the microbalance to an on line computer ( 8 ) outside of the vacuum chamber via an electronic module (9) To prevent exposure of the microbalance to hot product gases, vacuum is pulled through a side tube at the top of the quartz vessel ( 10 ) instead of through the vacuum chamber wall. Any gases in the TG are thus subject to an upward flow. A mass spectromet er (1 1 ) is connected in betwee n the TG and the vacuum pump (1 2 ) and collects a sample of the product gases. The constituents of the product gases and their quantities are analyzed to provide supplementary data for determining the sample reaction rates. To aid the quantitative analysis of these gases, an inert reference gas such as argon is supplied through a por t in the vacuum chamber wall (1 3 ) via a mass flow controller (14) The product gases ar e passed through a cold trap (1 5 ) installed immediately next to the quartz vessel, so that any metal vapors potentially arising from the experiment can be condensed and trapped before they are able to

PAGE 33

33 contaminate components downstream. A vacuum valve (1 6 ) is used to control the intensity of the vacuum level inside t he TG, and a 10 4 torr (approx. 10 7 bar) res olution digital vacuum gauge (1 7 ) installed in the vacuum chamber wall is used to measure the actual vacuum level. Temperature measurement of the sampl e is achieved by a pyrometer (1 8 ) and a type B thermocoup le (19) placed underneath the sample crucible, capable of measuring temperatures up to 1700 C The thermoelectric signal from the thermocouple is transmitted t o the data acquisition system ( 20 ) u sing a novel wireless circuit ( 21 ) conceived and designed by Be n Greek. The TG sits on a concrete slab (2 2 ) which absorbs vibrations from the solar simulator vacuum pump and other equipment. The TG is adjusted in place by means of a height adjustable table (2 3 ) and an x y table (2 4 ), such that the focus of the radiat ion coincides precisely with the position of the sample crucible.

PAGE 34

34 Figure 2 1 E xperimental layout. (1) Vacuum chamber; (2) Microbalance: (3) Crucible; (4) Solar simulator; (5) Mirror; (6) Quartz vessel; (7) Rad iation shield; (8) Computer; (9) Electronic module ; (10) Vessel side tube; (11) Mass spectrometer; (12) Vacuum pump; (13) Gas feedthrough; (14) Mass flow controller; (15) Cold trap; (16) Vacuum valve; (17) Vacuum gauge; (18) Pyrometer; (19) Thermocouple; ( 20) Data acquisition board; (21) Wireless temperature transmission circuit; (22) Concrete slab; (23) Height adjustable table; (24) X Y table. (22) (23) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (15) (16) (17) (18) (20) (24) (21) Comp. Elec. box Mass spec. Vac. pump DAQ (14) Ar (19)

PAGE 35

35 CHAPTER 3 THERMOGRAVIMETER DES IGN 3.1 Overview of the TG A ssembly Figures 3 1 3 2 and 3 3 show, respectively, 3 D CAD drawings of the outside view, partially cross sectioned view and exploded view of the TG assembly. Each component is marked with the same number in Figures 3 2 and 3 3 A s described in Chapter 2, the microbalance (1) is housed inside the vacuum chamb er (2). The zirconia crucible (3) is su spended above a vertical zirconia su spension (4) which, in turn, is held by an aluminum holder (5) screwed to the load receptor (6) of the balance. The vertical support, while serving to distance the crucible from the balance and thus protecting the balance from exposure to radiation, also encases the type B thermocouple (7). The thermocouple junction is in contact with the crucible, and its two wire ends are connected to a circuit board (8) which transmits the thermoe lectric signals wirelessly. The water cooled, aluminum radiation shield (9) sits on the surface of the vacuum chamber lid. A fluorocarbon (Viton) o ring (10) is placed between the radiation shield and the lid to provide a vacuum seal. The quartz vessel (11 ) fits over the radiation shield and its flange is sealed on both sides by a pair of fluorocarbon (Viton) gaskets (12). Lastly, a pair of clamps (13) are fitted over the components and screwed into the vacuum chamber lid. This compresses the aforementioned o ring and gaskets, thus sealing the vacuum enclosure. A detailed description of each component is given in the following section. CAD drawings of the components and the bill o f materials are included in Appendi ces A and B, respectively. The CAD drawings in Appendix A have units of millimeters except for Figures A 2 to A 7, which are in inches.

PAGE 36

36 Figure 3 1 TG assembly.

PAGE 37

37 Figure 3 2 Partially cross sectioned view of the TG assembly. (1) Microbalance; (2) Vacuum chamber; (3) C rucible; (4) Crucible suspension; (5) Suspension holder; (6) Balance load receptor; (7) Thermocouple (hidden underneath crucible) ; (8) Circuit board; (9) Radiation shield; (10) O ring; (11) Quartz vessel; (12) Gaskets; (13) Clamps. (1) (3) (2 ) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)

PAGE 38

38 Figure 3 3 Exploded view of the TG assembly. (1) Microbalance; (2) Vacuum chamber; (3) Crucible; (4) Crucible suspension; (5) Suspension holder; (6) Balance load receptor; (7) Thermocouple; (8) Circuit board; (9) Radiation shield; (10) O ring; (11) Quartz vessel; (12) Gaskets; (13) Clamps. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)

PAGE 39

39 3.2 Component D etails 3.2.1 Vacuum C hamber The vacuum chamber is a custom made, stainless steel vacuum chamber manufactured by Abbess Instruments. It has internal dimensions of 13 (length) x 11 (width) x 7 (hei ght) and thus an internal volume of 16.4 liters. It weighs approximately 60 kg. Figure 3 4 shows the vacuum chamber with ( A ) its lid closed and ( B ) its lid open. The lid (1) has hinges on one end and is se cured with a pair of latches As the lid is closed, an o ring ( 2 ) is compressed and provides a vacuum seal. A 20 mm diameter through hole ( 3 ) is processed in the lid to allow the vertical crucible support to pass through Eight tapped M8 blind holes ( 4 ) surround the through hole and allow for the radiation shield and clamps to be screwed onto the lid. Additionally, an o ring gland ( 5 ) is processed in the top surface of the lid for installation of the fluorocarbon (Viton) o ring which seals the interface between the lid and the radiation shiel d. Four tapped M6 blind holes ( 6 ) are processed in the bottom interior surface of the chamber to allow the balance to be screwed into its correct location. A Super Bee digital vacuum gauge from InstruTech ( 7 ) is installed in the side wall of the chamber and measures pr essures ranging from 10 4 torr to 1000 torr (approx. 10 7 bar to 1 bar). The pressure readings are transmitted to a computer usin g a serial cable. A gas feedthrough ( 8 ) is installed in the same wall to allow the introduction of gases into the chamber. A 25 pin, D sub electrical feedthrough ( 9 ) is installed in another wall to allow for any electrical wires to be passed through the wall without compromising the vacuum, thereby connecting devices inside and outside of the chamber t ogether. A vacuum ball valv e ( 10 ) is installed in the same wall and can be connected to the vacuum pump to evacuate the chamber if necessary; however, this is not desired as the vacuum should be drawn via

PAGE 40

40 the quartz vessel instead, as explained in Chapter 2. A vent ball valve (1 1 ) i s also installed in this wall and is opened gently to allow air back into the chamber after an experiment under vacuum. The vacuum chamber cannot be evacuated alone due to the presence of the through hole in the lid. It can only be evacuated once all compo nents (radiation shield, quartz vessel, clamps, gaskets and o ring) have been properly installed. If, under any circumstance, one wishes to use the vacuum chamber on its own, a plate can be manufactured and screwed on top of the lid to cover the through ho le, using the o ring to seal the space in between. Finally, it should be noted that during assembly of the TG, the hinges of the chamber lid need to be removed so that the lid can be installed in such a way as to let the crucible support fit thr ough its 20 mm diameter through hole. The hinges can be reattached once the lid has been installed. Figure 3 4 Vacuum chamber. A ) With lid closed B ) With lid open. (1) Lid; (2) O ring; (3) Through hole; (4) Tapped M8 blind holes; (5) O rin g gland; (6) Tapped M6 blind holes; (7) Vacuum gauge; (8) Gas feedthrough; (9) Electrical feedthrough; (10) Vacuum ball valve; (11) Vent ball valve. A ) B ) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

PAGE 41

41 3.2.2 Microbalance The microbalance is a WZA225 CW model from Sartorius with a resolution of 10 It has a maximum weight capacity of 220 mg; however, as the sample crucible and its supporting components are mounted onto it, the maximum allowable weight of the powder sample is 220 mg minus the total weight of these components. Furthermore, in order to allow its automatic internal calibration system to function (as is necessary for the TG), the total weight of the mounted components must be kept under 110 g. Mounting, for example, a crucible and supporting components weighing 80 g in total, allows th e internal calibration system to work and leaves a weighing range of 140 g for the sample. The balance comes with an electronic module, as sho wn in Figure 3 5 The balance (1) is connected to the electronic module (2) via a unique cable (3) designed by Sar torius. To transmit the mass readings to a computer, the computer (4) is connected to the electronics module with a serial cable (5). The load receptor (6) of the balance has three tapped M3 holes which can be used for mounting the crucible and its support ing components on top of it. As the vacuum chamber is only large enough to accommodate the microbalance, the electronic module must be placed outside the chamber and must therefore be connected to the microbalance through the chamber wall. This is achieved by cutting the connecting cable (3) in half and connecting its wires to the electrical feedthrough installed in the chamber wall (7), as shown i n Figure 3 6 This cannot be achieved by using two individual connectors; only the unique cable designed by Sar torius will work and must therefore be cut. A total of 15 individually discernible wires are present in the unique cable. Finally, the location of the microbalance inside the vacuum chamber must be such that the center of its load receptor (8) is coinciden t with the cente r of the 20 mm diameter through hole in the

PAGE 42

42 chamber lid. Positioning the balance so that space is left on all four sides according to the dimensions in Figure 3 6 allows this to be achieved. As shown in Figure 3 7 the balance is fixed in t he correct position using two aluminum pieces (9). The pieces screw onto the balance via tw o tapped M6 holes (10) present in the balance, and onto the bottom interior surface of the chamber via its four tapped M6 holes (11). Figure 3 5 Micro balance and electronic module. (1) Microbalance; (2) Electronic module; (3) Sartorius cable; (4) Computer; (5) Serial cable; (6) Load receptor. (1) (2) (4) (3) (5) (6)

PAGE 43

43 Figure 3 6. Position of the microbalance within the vacuum chamber. (7) Electrical feedthrough; (8) Center of load receptor. Figure 3 7. Fixing the microbalance inside the vacuum chamber. (9) Aluminum pieces; (10) Tapped M6 holes in balance; (11) Tapped M6 holes in vacuum chamber base. (7) (8) Electronic Module Comput er Balance Vacuum chamber + (9) (10) (11)

PAGE 44

44 3.2.3 Sample C rucible, C rucible S uspension and T her mocouple A ssembly The sample crucible and its suspension mechanism are sho wn in Figure 3 8 T he assembly is shown in ( A ) and an exploded view is shown in ( B ). The 20 mm diameter crucible (1) sits on top of a circular platform (2) which in turn sits on top of two hollow suspension rod s (3). The crucible, platform and rod s a re made from zirconia w hich has a melting point of 2 715 C All three items are custom made by Ortech. The rod s are supported by an aluminum holder (4), manufactured by Manufacturing Too ling & Engineering. The aluminum holder is screwed onto the load receptor (5) of the balance with M3 bolts. The height of the crucible can be changed by adjusting the suspension rod s up and down inside the aluminum holder. The rod s can then be secured in p osition by screwing the side screw on the holder. The thermocouple (6) is an unsheathed, bare wire type B thermocouple from Omega, capable of measuring temperatures up to 1 8 00 C The positive and negative leads of the thermocouple are inserted through the double bored platform and rod s and are thus electrically insulated from each other. The lead ends pass through an opening in the holder (7) and are connected to the wireless circuit (8) described in the next section. The thermocouple tip (9) remains exp osed above the surface of the platform, and comes into contact with the bottom surface of the crucible when the crucible is fitted over the platform. The close up drawing and cross sectional dr awing in Figure 3 9 A and Figure 3 9B sho w the crucible assemb ly more clearly. The platform ( 10 ) was designed so that the crucible and its contents will not fall apart in the event that thermal shock caused the crucible to break in half. The crucib le is designed with a flange ( 11 ) at the bottom so that any metal liquid which may unexpectedly form and bubble over the top rim of the crucible can be captured and prevented from falling through the radiation shield and onto the microbalance. The

PAGE 45

45 crucible was designed to fit over the platform rather than sit on it, so that it has more stability when it sits on top of the thermocouple tip (12) and any liquid metal spilling over the top of the crucible cannot come into contact with the platform itself. Keeping potentially arising metal liquids away from the platform and therm ocouple was important, as the liquid metal could otherwise solidify on these components, creating the need for costly and time consuming replacements. This was considered thanks to Kyle Allen, who pointed out the issue regarding liquefied samples spilling out of thermogravimeter crucibles and suggested ideas to protect neighboring components from such spills. A suitable thickness t (13 ) for the crucible base will be determined. In order to minimize the temperature difference across its top and bottom surf ace, and thus minimize the deviation of the thermocouple measurement from the actual sample temperature, the crucible base should be made as thin as possi ble. However, a thinner base is more su sceptible to failure from thermal expansion. Thus, a thermal st rain analysis will be conducted to determine the minimum allowable thickness t

PAGE 46

46 A ) B ) Figure 3 8 Sample crucible and suspension mechanism. A ) Assembly B ) Exploded view. (1) Crucible; (2) Platform; (3) Suspension rods; (4) Suspen sion holder; (5) Load receptor; (6) Thermocouple; (7) Holder opening; (8) W i reless circuit; (9) Thermocouple tip (1) (2) (3) (4) (5) (6) (7) (8) (9)

PAGE 47

47 A) B ) F igure 3 9 Sample crucible design. A) Close up view B ) Cross sectional view. (10) Platform; (11) Fla n ge; (12) Thermocou ple tip; (13) Crucible b ase thickness t 3.2.4 Wireless C ircuit for T ransmission of T emperature R eadings The wireless circuit used for transmission of temperature readings is an outstanding feature of the TG proposed and designed by Ben Greek. The signif icance of this circuit and its principle of operation are explained in this section. As shown in Figure 3 10 d u r ing the early design phase of the TG, it was planned that the two thermocouple leads (1) described in 3.2.3 would be connected directly to the electrical feedthrough (2) in the vacuum chamber wall for transmission of the thermoelectric signals to the outside of the chamber. However, this set up was problematic. Ayyoub Mehdizadeh alerted the issue that tensional forces arising in the thermocouple leads would pull on the crucible support and affect the fine weight measurements taken by (10) (11) (13) t (12)

PAGE 48

48 the microbalance. To solve this issue, Ayyoub proposed that the thermoelectric signals be transmitted wirelessly. A wireless thermocouple was first sought after; ho wever, the signal transmitters of such products were too large in size and weight to be mounted on top of the microbalance, which, as mentioned in section 3.2.2, could only handle up to 110 g worth of components if the automatic internal calibration system of the balance was to be used. Furthermore, it could not be guaranteed that the signals would successfully pass through the thick stainless steel walls of the vacuum chamber. Ben subsequently proposed an alternative method for transmitting the thermoelect ric signals wirelessly. The unique meth od is shown in Figure 3 11 The two t hermocouple leads, instead of connecting directly to the electrical feedthrough, are connected to a circuit board (1) attached to the rod holder. On the circuit board, the thermoel ectric signals produced in the thermocouple are amplified and converted into a frequency signal, where the frequency is a function of the magnitude of the thermoelectric signal (and thus of the measured temperature). The frequency signal is then supplied t o an infra red LED (2) featured on the circuit board, which pulsates at a frequency equal to that of the incoming signal. At a short distance from the LED, a phototransistor (3) is placed. The phototransistor receives the pulsating infra red signal, conver ts it into a pulsating voltage signal and sends this to the data acquisition system via the electrical feedthrough (4). The frequency of the received voltage signal is then translated into a temperature reading. The small and lightweight circuit board easi ly mounts onto the aluminum holder and moves up and down with the crucible and suspension. Therefore, it does not interfere with the weight measurements. This novel and convenient wireless

PAGE 49

49 transmission system is a major advantage, as other solar TGs have s uffered from issues regarding transmission of temperature readings. Figure 3 10 Issue with transmitting thermoelectric signals. (1) Thermocouple leads; (2) Electrical feedthrough. Figure 3 11 Wireless solution. (1) Circuit board; (2) LED; (3) Phototransistor; (4) Electrical feedthrough. (1) (2) (1) (2) (3) (4)

PAGE 50

50 3.2.5 Quartz V essel The quartz vessel is custom made by Quartz Sc ientific. As shown in Figure 3 1 2 it has a cylindrical body and a spherical top to minimize locations where stress concentration may occ ur. The inner and outer diameters of the cylindrical and spherical portions are 75 mm and 80 mm, respectively. The side tube (1), where vacuum is drawn from, has an outer diameter of 19 mm such that it can be connected to the inlet tube of the glass vacuum trap using a 3/4 (19.05 mm) Swagelok vacuum union. The flange (2) has a diameter of 140 mm and a thickness of 5 mm. It is sealed on both sides by a pair of fluorocarbon gaskets. Figure 3 12 Quartz vessel. (1) Side tube; (2) Flange. 3.2.5.1 Therma l and m echanical p roperties The vessel is exposed to intense thermal radiation. Furthermore, as it is bounded by the atmosphere on one side and vacuum on the other, it is exposed to a compressive force. It was therefore required that the glass material use d for the vessel possessed a high service temperature, a strong resistance against thermal shock and a (1) (2)

PAGE 51

51 high compressive strength. While borosilicate glass is used to fabricate most laboratory glassware, quartz was chosen as the material for the vessel, as it outperforms borosilicate glass in all three cr iteria. Table 3 1 shows a comparison of the thermal and mechanical properties of quartz and borosilicate glass. Quartz has a service temperature of 1100 C, twice as high as that of borosilicate glass. Beyon d this point, thermal stress begins to appear within the quartz glassware. The thermal shock parameter, R T is calculated for both materials using the follow ing formula [ 28 ]: 3 1 where k = thermal conductivity, T = tensil e strength, E The thermal shock parameter indicates that quartz is 13 times more resistant to thermal shock than borosilicate glass, largely owing to its low coefficient of thermal expansion, which equals 1/8 that of borosilicate glass. Furthermore, the mean compressive strength of quartz exceeds twice that of borosilicate glass. Table 3 1 Thermal and mechanical properties of quartz and borosilicate glass. Property Quart z [ 29 ] Borosilicate glass [ 30 ] Service temperature (C) 1100 500 Compressive strength (MPa) 1600 2000 600 1000 Tensile strength (MPa) 70 120 80 150 Thermal conductivity (Wm 1 K 1 ) 4.8 2.0 3.8 0.17 0.20 Coefficient of thermal expansio n (K 1 ) 0.53 3 6 62 75 65 85 Thermal shock parameter (Wm 1 ) 10400 790 *Calculate d using f ormula 3 1

PAGE 52

52 3.2.5.2 Determination of w all t hickness The thickness of the quartz vessel wall required to withstand the compressive force of the at mosphere was determined as follows. As shown in Figure 3 13 the vessel wall is subjected to compressive hoop stress in the spherical and cylindrical walls, and compressive axial stress in the cylindrical wall. For a vessel of radius R, thickness t and a p ressure difference of across its wall, it can be easily shown that these stresses are expressed by: 3 2 3 3 3 4 The largest stress oc curring in the vessel is thus the compressive hoop stress in the cylindrical wall. In order for the vessel to resists this stress, the compressive strength of the vessel must be greater than th is stress by an appropriately chosen safety factor. This leads to the relation: 3 5 yielding the following condition for the thickness of the vessel wall: 3 6 The inner diameter of the cylindrical wall was designed to be 75 mm, in order for the vessel to fit over the cylindrical radiation shield. Substituting, therefore, R = 0.0375 m, = 0.1013 MPa and compressive strength of quartz = 1600 MPa, the required wall

PAGE 53

53 thickness was calculated for safety factors ranging from 1 to 10, as shown in Table 3 2 It was seen that even for a safety factor of 10, much greater than the safety factor of 3.5 4 commonly used in industry for vessel design, the required wall thickness did not even reach 1 mm. To reduce the cost of the custom made quartz vessel, it was decided tha t standard cylindrical tube sizes used by the manufacturer should be used. The cylindrical and spherical walls were thus designed to have an outer diameter of 80 mm, resulting in a wall thickness of 2.5 mm. While the pressure calculation shows that this th ickness is sufficient, however, it should be noted that exposure of the vessel to intense thermal radiation may still make the vessel susceptible to implosion. Figure 3 13 Hoop and axial stresses present in the vessel wall. 1 1 1 1 2 2 3 3

PAGE 54

54 Table 3 2 Wall thickness of quartz vessel calculated for a range of safety factors. Safety factor Wall thickness (mm) 1 0.002 2 0.005 3 0.007 4 0.009 5 0.012 6 0.014 7 0.017 8 0.019 9 0.021 10 0.024 3.2.5.3 Implosion hazard W hile the vessel wall thickness is designed to resist the atmospheric pressure, it should be noted that exposure of the vessel to intense thermal radiation may still make the vessel susceptible to implosion. Even though the vessel should collapse inward in this event, glass shards may still pose a safety hazard. Care should be taken to evacuate the TG very slowly, and to heat the TG gradually to avoid thermal shock. Furthermore, a fan may be used to cool the surface of the vessel and slow its temperature inc rease. 3.2.6 Radiation S hield The water cooled, aluminum radiation shield is custom made by Manufacturing Tooling & Engineering. As shown i n Figure 3 14 th e water is supplied through an inlet (1) in the side of the shield and exits from a neighboring out let (2) The tubes carrying the water have a 1/2 outer diameter and are connected to the inlet and outlet using 3/8 Swagelok NPT tube fittings (3) model B 810 1 6. The radiation shield consists of three pieces welded together, shown in Figure 3 15 : a co oli ng cylinder (1), base plate (2) and cap (3). Cooling channels are machined inside the cooling cylinder, the top surface of the base plate and the bottom surface of the cap, such that when the three components

PAGE 55

55 are joined together, they connect to form a single, series chan nel. Figure 3 16 shows the route of the water as it travels through this channel. Upon entering through the side of the cooling cylinder, it first descends into a cooling channel in the base plate where it travels circumferentially along the plate, travels radially towards the center and rises through one of six vertical channels in the cylinder. After reaching the top of the cylinder, it travels circumferentially along a channel in the cap, descends through another vertical channel in th e cylinder and reaches another channel in the base plate. In this manner, the cooling water travels up and down repeatedly through the radiation shield, gradually making its way around it before exiting at the outlet. While a parallel flow channel design c ould have also been used, a series flow channel was chosen to guarantee that the water will pass through every part of the channel, thus decreasing the risk that any hot spots will occur in the shield as it is exposed to radiation. The pieces were welded t ogether at four locations indicate d in Figure 3 17 : the inner ( 1 ) and outer ( 2 ) edges of the cap and cyli nder interface, and the inner ( 3 ) and outer ( 4 ) edges of the cylinder and base plate interface. The water is thus strictly contained within the radiati on shield, and will by no means leak into the vacuum chamber. A s shown in Figure 3 18 the cap and base plate were designed to slot into the cooling cylinder via projections on their surface, so that they may be properly aligned with respect to the cylinde r before welding. Aluminum wa s chosen as the material of the shield, over alternatives such as copper and stainless steel. Table 3 3 compar es the thermal properties of these metals. Aluminum has a thermal conductivity 15 times higher than that of stainless steel, and is thus better able to transfer heat to the cooling water. While copper is even better at conducting heat, this very property makes it difficult to weld and

PAGE 56

56 was thus avoided. However, copper and stainless steel both have a melting point higher than that of aluminum. While it is unlikely that the radiation shield surface would experience temperatures above its melting point, copper and stainless steel should be considered as alternatives in the event that this should happen. Figure 3 14 Radiation shield. (1) Water inlet; (2) Water outlet; (3) Swagelok tube fittings. Figure 3 15 Components of radiation shield. (1) Cooling cylinder; (2) Base plate; (3) Cap. (1) (2) (3) (1) (2) (3)

PAGE 57

57 Figure 3 16 Route of cooling water. Figure 3 17 Weld loc ations. (1) (2) Inner and outer edges of cap and cylinder interface; (3) (4) Inner and outer edges of cylinder and base plate interface. (1) (2) (3) (4)

PAGE 58

58 Figure 3 18 Slots created for assembly of the shield components. Table 3 3 Thermal properties of aluminum an d other metals. Aluminum Copper Stainless steel Thermal conductivity ( Wm 1 K 1 ) [ 31 ] 250 401 16 Melting point (C) [ 32 ] 660 1084 1510 3.2.7 Clamps and V acuum S eals 3.2.7.1 Functions The stainless steel clamps are custom made by Manufacturing Tooling & Engineering Figure 3 19 s hows the two clamps: a north clamp and a south clamp. The north clamp has four projections on its side which slot into four holes in the south clamp. The clamps serve to hold the chamber top components (radiation shield, quartz vessel) in place, as well as to compress the gaskets and o ring used to seal the vacuum enclosure. Figure 3 20 shows an exploded view of the components which are mounted on top of the vacuum chamber lid, and how the clamps are designed to hold

PAGE 59

59 thes e components in place. During assembly, the radiation shield (1) is first placed on top of the vacuum chamber lid (2). A fluorocarbon (Viton) o ring (3) with an AS568 American Standard o ring size of 2 226 is installed in between the bottom surface of the radiation shield and the top surface of the lid. The quartz vessel is then fitted over the radiation shield. Its flange sits inside the recess of the radiation shield and is sealed on top and bottom by a pair of 1/8 thick fluorocarbon (Viton) gaskets (4). The north (5) and south (6) clamps are then slotted into each other to form one circular clamp, and fitted over these components. Finally, eight M8 fla t head socket cap screws are used to screw the clamps down. The screws pass th rough the holes in the clamps ( 7 ) and the holes in the radiation shield ( 8 ) before screwing into the tapped blind holes in the to p surface of the chamber lid ( 9 ). Once the clamps are screwed in place, this compresses the aforementioned o ring and gaskets. As indicated in Figure 3 21, locations from which air can otherwise leak into the TG are thereby sealed from atmosphere. The clamps were designed to consist of two pieces ( north & south ), as one continuous ring shaped clamp would not fit over the vessel due to obstruction b y the side t ube present in the vessel wall The clamps are to be slotted into each other below the side tube. As shown in Figure 3 2 2 the south clamp has an opening in its side to allow access of the cooling water tubes to the radiation shield. The clam ps must therefore be oriented such that this opening is coincident with the location of the tube fittings. It should also be noted that, during assembly, the tubes must be fitted to the radiation shield before the clamps are installed. Due to the small siz e of the opening, the tubes cannot be fitted to the shield after the clamps have been installed.

PAGE 60

60 Figure 3 19 North and south clamps. Figure 3 20 Exploded view of the components assembled above the vacuum chamber lid. (1) Rad iation shield; (2) Vacuum chamber lid; (3) O ring; (4) Gaskets; (5) North clamp; (6) South clamp; (7) Through holes in clamps; (8) Through holes in radiation shield; (9) Blind holes in chamber lid. North South (1) (2) (3) (4) (5) (6) (7) (8) (9)

PAGE 61

61 Figure 3 21 Locations sealed against air leakag e. Figure 3 22 Opening in south clamp for access of cooling water tubes. Air Air Air Air Opening

PAGE 62

62 3.2.7.2 Seals The o ring and gaskets mentioned in the previous section are made out of fluorocarbon. This elastomer comes under several different names: FKM, FPM, fluoro rubber fluoroelastomer and Viton. Viton is a trademark of DuPont Performance Elastomers L.L.C. Among the various elastomeric materials available, fluorocarbon was chosen due to its high service temperature (200 C) and particularly low permeability to gas. Tabl e 3 4 sho ws a comparison of the service temperatures and gas permeability characteristics of some common elastomeric materials. While silicone rubber also has a high service temperature, it has a poor resistance against permeation of gas and was thus avoid ed. Table 3 4 Char acteristics of some elastomeric mate rials. [ 33 ] Rubber type Abbreviation Trade names Max imum service temperature (C) Resistance to gas permeation Natural NR 70 Poor Styrene butadiene SBR B u na S 80 Fair Butyl IIR Butyl 90 Excellent Ethylene propylene EPDM 120 Fair Nitrile NBR Buna N 80 Very good Hydrogenated nitrile HNBR 140 Very good Chloroprene CR Neoprene 90 Good Chlorosulfonated polyethylene CSM Hypalon 90 Good Polyacrylate ACM Hycar/ Vamac 150 Good Polyurethane PU 80 Excellent Silicone Q 200 Poor Fluorosilicone MFQ 175 Poor Fluorocarbon FKM/FPM Viton 200 Very good 3.2.7.3 Achievement of a sufficient vacuum seal Figure 3 2 3 shows a cross sectional view of the clamps compressing the two gaskets sealing the top and bottom of the quartz vessel flange. In order to provide a

PAGE 63

63 sufficient vacuum seal, rubber gaskets must be compressed by a certain ratio of its initial thickness, or compression ratio The necessary compression ratio for the gaskets in Figure 3 23 was th us determined, and subsequently, the dimension X (length of space beneath the clamp surface in contact wit h the upper gasket) required to achieve this compression ratio was found. The minimum compression ratio required for a vacuum tight seal depends on the cross sectional shape and hardness of the gasket. Fi gure 3 24 region 1 shows the minimum required compression ratio as a function of gasket hardness, expressed on the Shore hardnes s scale [ 34 ]. Gaske ts of square cross section need the lowest compression ratio (represented by the bottom boundary curve of region 1), followed by those of semi circular cross section. Gaskets of circular cross section need the highest compression ratio and are represented by the top boundary curve of region 1. For gaskets with a Shore hardness value greater than 50, the leak rate becomes tolerable at a compression ratio of about 15 % for any gasket shape. On the other ha nd, Figure 3 24 curve 2 shows the maximum permissible compression ratio. The gaskets used for the solar TG, obtained from McMaster Carr ( thickness: inner diameter: outer diameter: ) have a Shore hardness of 75. Thus, from Figure 3 24 it was seen that a compression ratio between 10 % and 20 % would be appropriate for these gaskets. The dimension X in Figure 3 23 was therefore designed to achieve a compression ratio of 15 %, an d its tolerances were determined so that the actual compression ratio will not be less than 10 % or greater than 20 %. From Figure 3 23 the length of space X is expressed by: 3 7

PAGE 64

64 Denoting the compression ratio by CR, this can be rewritten as: 3 8 The initial thickness of each gasket, 1/8 is equal to 3.175 mm. Substituting CR = 0.1 (10 %), 0.15 (15 %) and 0.2 (20 %) into this expression yields values of X as show n in Table 3 5 As a result, the length X was designed to be 20.4 mm, with tolerances (+) 0.315 mm / ( ) 0.320. Figure 3 23 Compression of gaskets using clamps. Clamp Quartz vessel Gaskets X

PAGE 65

65 Figure 3 24 Compression ratio of rubber gaskets as a function of their Shore hardness. (1) Minimum compression ratio needed for a vacuum tight seal; (2) Permissible compression ratio; (3) Maximum compression ratio. [34] Table 3 5 Values of X for achieving gasket compression ratios of 10 %, 15 % and 20 %. Compression ratio Length X required to achieve compression ratio 0.10 20.7159 0.15 20.3975 0.20 20.0800 3.3 Using the TG for I nvestigation of O xidation R eaction K inetics The solar TG, while primarily designed for reduction reactions, can also be us ed to test the CO 2 splitting oxidation reaction. To use the TG for investigating the CO 2 splitting reaction, carbon dioxide should be supplied through the gas port in the vacuum chamber wall, along with the chosen reference gas used for analysis of the eff luent gases, using a manifold. As the carbon dioxide gas will inevitably come into contact with the microbalance, it must not be pre heated, since the microbalance cannot be operated at temperatures above 40 C. Instead, the carbon dioxide must be heated u pon contact with the irradiated sample. The hot effluent gases should be pumped out from the side tube in the quartz vessel, just as with a reduction experiment. As a result, the flow of

PAGE 66

66 gas inside the TG will be upward preventing hot gas from travellin g down towards the microbalance ( Figure 3 25 ). The pressure downstream of the TG can be controlled by adjusting the vacuum valve Although the quantity of carbon monoxide produced should be small, care should be taken to safely vent the gas outdoors as nec essary. The solar TG is not suited for handling steam and cannot be used to run the water splitting oxidation reaction. Figure 3 25 Using the TG for the CO 2 splitting reaction. Gas feedthrough Pathway of gas

PAGE 67

67 CHAPTER 4 THERMOGRAVIMETER CHA RACTERISTICS 4.1 Time to R ea ch T arget T emperature 4.1.1 One D imensional T ransient H eat C onduction P roblem Figure 3 26 shows a schematic diagram of the sample powder bed contained inside the crucible and heated radiatively from above. The top surface of the bed, exposed directly to t he incoming radiation, heats up most rapidly and reaches the target temperature fastest. The bottom surface of the bed heats up most slowly. The time it takes for the whole sample bed to reach the target temperature can be found by obtaining the temperatur e variation with depth and time. This is done by solving the 1D transient heat conduction equation: 3 9 where T denotes temperature, z denotes the depth from the top surface of the bed, t denotes time a is the thermal dif fusivity of the bed. This 1D transient conduction problem is represented in Figure 3 27 The bed is subjected to two boundary conditions. At the top surface of the bed (z = 0), the net heat flux entering the surface is equal to what is absorbed of the inco ming radiative flux minus what is re radiated to the surroundings, and is expressed by: 3 10 where k = effective thermal conductivity of the bed, face absorptivity of the sample, q = radiative flux incident on the sample (controlled to equal target 4 / ),

PAGE 68

68 Stefan Boltzmann constant. The bottom surface of the bed (z = L) is insulated by the ceramic cr ucible and is thus subjected to the following boundar y condition: 3 11 Additionally, the initial condition of the bed is such that: 3 12 Figure 3 26 The sample powder bed, heated ra diatively from above. Powder bed Crucible Radiation z = 0 (Irradiated) z = L (Insulated) z Boundary condition at top surface of bed: Boundary condition at bottom surface of bed: Initial condition: Powder bed

PAGE 69

69 Figure 3 27 Representation of the 1D transient heat conduction problem. 4.1.2 Determination of the E ffective T hermal C onductivity, k In the two boundary conditions described in 4.1.1, the effective thermal c onductivity of the bed, k differs from the solid thermal conductivity of the sample material itself. As the sample bed consists of powder, heat transfer through the bed consists of two modes: conduction through the solid phase of the powder particles, a nd radiative heat transfer from the surface of one particle to another. Heat transfer due to convection is neglected, as this is negligible in a vacuum environment. The modes of heat transfer are shown in Figure 3 28 where the powder bed is approximated b y a packed bed of spherical particles. The effective conductivity of the bed is thus the sum of the conductivities associated with these two heat transfer modes, and is expressed by: 3 13 w here k cond is the thermal conducti vity of conduction through the particle bed, and k rad is the thermal conductivity of radiative heat transfer through the particle bed, called the radiant conductivity The methods for determining the two conductivities are explained in the following sectio ns.

PAGE 70

70 Figure 3 28 Modes of heat transfer through the particle bed. 4.1.2.1 Thermal conductivity of conduction heat transfer through the particle bed Various analytical and empirical models exist for determining the thermal conductivity of conduction heat transfer through a packed bed of s pherical particles [ 35 ]. The models give k cond as a function of the solid conductivity of the particles, k s the conductivity of the fluid between the solid part icles, k f and the porosity of the particle bed, Among them, three in particular Hadley s weighted av erage [ 36 ], an analytical model, an empirical model by Krupiczka [ 37 ] and an empirical model by Zehnder and Schlnder [ 38 ] have been demonstrated to closely approxim ate the results of experiments conducted with packed beds of = 0.38 for k s /k f ratios between 1 and 10 4 as shown in Figure s 3 29 and 3 30 [ 3 5 ]. Furth ermore, the three models in particular yielded highly consistent values of k cond in a trial calculation conducted for a packed bed of ferrite particles, using bed porosity = 0.5, k s = 4 Wm 1 K 1 (for ferrites [ 39 ]), k f = 0.01 Wm 1 K 1 (for air at 10 4 bar [ 40 ]) and th us a k s /k f ratio of 400. As these three models were shown to be the most reliable, one of them Hadley s weighted average model was chosen as the representative model to be used for the Powder bed Radiation Powder particle Conduction heat tra nsfer Radiative heat transfer

PAGE 71

71 determination of k cond in the 1D heat conduction problem. Hadley s model is given below: 3 14 3 15 3 16 where k s = solid conductivity of the particles, k f = conductivity of the fluid between the particles, = porosity of the particle bed. Figu re 3 29 Comparison of experimental results with various analytical models developed for k cond [ 35 ]

PAGE 72

72 Figure 3 30 C o mparison of experimental results with empirical models developed for k cond [ 35 ] 4.1.2.2 Thermal conductivity of radiative heat transfer t hrough the particle bed Assuming the particle bed to be an optically thick medium, the radiative heat transfer through the bed can be considered similar to heat diffusion as shown by the Rosseland approximation [ 41] In this case, a radiant conductivity k rad is defined. The following correlati on [ 41 ] w as used for the determination of k rad in the 1D heat conduction problem: 3 17 3 18 w here D = particle diameter, Boltzmann constant, T = mean temperature of the bed,

PAGE 73

73 k s = solid conductivity of the particles, = bed porosity, a 1 = 0.5756, a 2 = 1.5353, a 3 = 0.8011, a 4 = 0.1843. 4.1.3 Solution of the O ne D imensional T ransient H eat C ond uction P roblem The temperature variation with depth and time was solved numerically using the heat equation, boundary conditions and initial condition in 4.1.1. The effective thermal conductivity of the powder bed, k included in the boundary conditions, was obtained using the models in 4.1.2. The temperature variation with depth and time, and thus the time taken for a uniform temperature distribution to be reached within the bed, depends on the following factors: the powder material used, the particle di ameter of the powder, the porosity of the bed, the total depth (thickness) of the bed, the target temperature, the radiative flux provided, the absolute pressure of the vacuum used. The Matlab code included in Appendix C computes, for a bed of 150 diam eter ferrite particles with a porosity of 50 % in a vacuum of 10 4 absolute pressure, the time taken to reach uniform temperature at 1500 C for a giv en bed thickness (total depth) when exposed to a heat flux of 5000 kW/m 2 It also gives the temperature v ariation with depth and time, up to 30 minutes. The parameters in the code including those listed

PAGE 74

74 above and the supplied heat flux, can be modi fied to suit different particle beds and experimental condition s Figure 3 31 shows the temperature variation wi th depth and time for a 5 mm thick ferrite bed as described above, w hen exposed to a heat flux of ( A ) 5000 kW/m 2 and ( B ) 560 kW/m 2 ( 1 % greater than the flux required for the bed to reach a steady state temperature of 1500 C ). Figure 3 32 sh ows, for varyi ng thicknesses of the same ferrite bed, the time taken for the top surface and entire bed to reach 1500 C when exposed to a heat flux of ( A ) 5000 kW/m 2 and ( B ) 560 kW/m 2 It can be seen that for a 5 mm thick bed, the bed surface and entire bed reach 150 0 C after 0.07 s econds and 78 seconds respectively, when exposed to a heat flux of 5000 kW/m 2 When it is exposed to a heat flux of 560 kW/m 2 the bed surface and entire bed reach 1500 C within 5 and 9 minutes, respectively. The times are even shorter f or a bed height of 1 mm, with the surface and entire bed reaching 1500 C in 0.07 seconds and 4 seconds, respectively, when exposed to a heat flux of 5000 kW/m 2 This translates into a heating rate of over 20,000 K/s for the surface and over 300 K/s for th e entire bed. While a heat flux of target 4 is enough to heat the bed to its target temperature T target greater heat fluxes can be supplied if higher heating rates are desired. Once the bed reaches its target temperature th e temperature can be mai ntained constant by reducing the flux back to 4

PAGE 75

75 A) B) Figure 3 31 Temperature variation with depth and time for the 5 mm particle bed described in 4.1.3. A ) H eat flux = 5000 kW/m 2 B ) H eat flux = 1 % above that required for a steady state temperature of 1500 C to be reached (q = 1773 4 1.01). Time (min) Time (min) Temperature (K) Temperature (K) Depth (mm) Depth (mm)

PAGE 76

76 A ) B ) Figure 3 32 Relatio nship between bed thickness and time s taken for bed surface and entire bed to reach 1500 C A ) H eat flux = 5000 kW/m 2 B ) H eat flux = 1 % above that required for a steady state temperature of 1500 C to be reached (q = 1773 4 1.01). 1500 C 1500 C

PAGE 77

77 4.2 Recommended S ample M ass 4.2.1 Minimum S ample M ass In order for the TG to be able to detect the mass change of the sample, it must be ensured that the initial sample mass must be e qual to or greater than a minimum mass. For the reduction reaction, this minimum mass depends on the molar mass of the sample, the number of moles of oxygen released for every mole of sample reacted, and the resolution with which the extent of the reductio n reaction is to be measured. This is explained with the example of the reduction of magnetite (Fe 3 O 4 ) to ferrous oxide (FeO), as follows. This reduction reaction is expressed by: 3 19 Here, it is supposed that the conversio n of the magnetite sample is to be measured with a resolution of 1 % of its mass; in other words, for every 1 % of the sample mass that undergoes conversion, the resulting decrease in mass must be greater than the resolution of the microbalance. The decrea se in mass is equal to the mass of oxygen gas evolved. For a magnetite sample weighing X g, the decrease in mass accompanying a 1 % conversion of the sample is thus expressed as follows: 3 20

PAGE 78

78 where 231.52 g/mol is the molar mass of Fe 3 O 4 This mass decrement must be sufficiently larger than the resolution of the microbalance, 10 Taking a safety factor of ten, the minimum required sample mass, X min is calculated as follows: 3 21 Thus, using at least 0.14 g of Fe 3 O 4 will ensure that e very 1 mass % c onversion of the sample is detected by the microbalance. 4.2.2 Maximum S ample M ass In order to analyze the rate of O 2 gas evolution as a secondary method of determining the reduction reaction rate, the rate of O 2 gas evolution must be such that the flow rate of O 2 gas diverted to the mass spectrometer is greater than the minimum detectable flow rate of the mass spectrometer (given in standard cubic centimeters per minute). While the rate of O 2 gas evolution is dependent on the r eaction kinetics and is thus difficult to predict, increasing the sample mass can increase this rate, thus increasing the probability that the mass spectrometer would detect the O 2 gas. However, using a larger sample mass has a detrimental effect on the ex periment: the greater the mass, the thicker the powder bed becomes, and the time taken for the whole bed to reach uniform temperature increases. The time delay between the point at which the top surface of the bed reaches the reaction temperature and the p oint at which the bottom surface of the bed reaches the reaction temperature is increased. This time delay must be significantly less than the time taken for the reaction to complete. Ideally, in order to obtain an accurate reaction rate, all the particles composing the powder bed should react in unison. Therefore, there is a limit to how much one can increase the

PAGE 79

79 sample mass, and this cannot be known until the reaction rate of the sample has been found. It is recommended that, during a preliminary experime nt, the time taken for a single powder particle to complete its reaction is found (either by timing one powder particle, or by timing a particle thick layer of powder and averaging for each particle). If the reaction takes, say, 30 minutes to complete, and a 5 mm thick bed can reach uniform temperature in 3 minutes, a 5 mm thick bed can be considered permissible. However, if the reaction takes only 5 minutes to complete, it would be better to use a smaller bed thickness so that uniform tempera ture is reache d much faster. P lot s such as the one s in Figure 3 32 ma y be used to determine the maximum allowable sample mass. 4.3 Maximum T emperatures A ttainable with the TG W hile temperatures above 1500 C are not likely to be used in the ARPA E project, the TG can be used to heat samples to high er temperatures as necessary The steady state temperature, T, to which a sample can be heated, is a function of the absorptivity of the that can be supplied t and is expressed by : 3 22 At full power, the solar simulator can provide a radiative heat flux in excess of 5000 kW/m 2 onto a plane target situated at its focal point (F igure 3 33 A ). A blackbody, with an absorptivi ty and emissivity equaling 1, can be heated to as high as 3064 K (2791 C ) according to the above relation. The TG sample bed, however, cannot intercept the full 5000 kW/m 2 flux without the use of a mirror as the surface of the sample bed is perpendicular to the direction of the radiation beam ( Figure 3 33 B ). To

PAGE 80

80 maximize the flux that is intercepted by the sample, mirrors can be used to re direct the beam onto the sample bed from above, as shown in Figure 3 3 4 Plan e mirrors, elliptical mirrors and hyperbo lic mirrors can fulfill this function. The maximum flux intercepted by the sample and thus its maximum attainable temperature is then dependent on the geometry of the mirror and its reflectivity. A more significant limiting factor is the melting point of t he crucible and suspension and the service temperature of the thermocouple. The present TG design features a zirconia crucible and suspen sion with a melting point of 2 715 C and a type B thermocouple capable of measuring temperatures up to 1800 C ; therefo re, TG analyses are limited to 1800 C If measurements above 2000 C are required, a type C thermocouple capable of measuring temperatures up to 2320 C may be used. Type C thermocouples, however, degrade rapidly in oxidizing environments and must never b e used in the presence of oxygen. Its use is therefore limited to experiments in which an inert atmosphere or a very deep vacuum may be used. Figure 3 33 Irradiation by th e solar si mulator A ) P lane target B ) C rucible. A ) B ) Plane target Crucible

PAGE 81

81 Figure 3 34 Use of mirrors to re direct radiation ont o the surface of the sample bed A ) P lane mirror B ) H yperbolic mirror C ) E lliptical mirror. TG TG TG Mirror Mirror Mirror A ) B ) C )

PAGE 82

82 CHAPTER 5 CONCLUSION A solar thermogravimeter was designed and is currently under construction. The solar TG will be us ed to study the mechanisms of reduction and oxidation reactions employed in the solar thermochemical production of synthetic gas. The solar TG is heated radiatively using University of Florida s 5 6 kWe solar simulator. While maximum reachable heating rates and temperatures depend on multiple variables including sample material surface properties powder bed thickness and other experimental conditions, heating rates exceeding 100 K/s (over 1000 K/s for top surface of sample powder bed) can be obtained for pa rticle beds and a blackbody can be heated to a maximum of approximately 3000 K. Temperature measurements can be conducted up to the service temperature of the thermocouple used; i.e 1800 C with a type B thermocouple and 2320 C with a type C thermocouple The TG can be evacuated to pressures as low as 10 4 bar. Changes in the mass of the sample are recorded by a 10 resolution microbalance. Product gases are identified and quantified using a mass spectrometer connected to the TG. Temperature measurements of the sample are taken using a pyrometer and a thermocouple placed underneath the sample crucible. The TG makes use of a novel system to transmit thermocouple temperature readings wirelessly, solving practical issues that are otherwise encountered.

PAGE 83

83 APPENDIX A CAD DRAWINGS Figure A 1. Cross sectional drawing of TG assembly. (Dimensions in mm)

PAGE 84

84 Figure A 2. Top v iew of vacuum chamber and vibration isolation mass. (D rafted by Mr. Tom Driscoll of Abbes s Instruments; dimensions in inches )

PAGE 85

85 Figure A 3. Front view of vacuum chamber. (Drafted by Mr. Tom Driscoll of Abbess Instruments ; dimensions in inches )

PAGE 86

86 Figure A 4 Left side view of vacuum chamber. (Drafted by Mr. Tom Driscoll of Abbess Instruments; dimensions in inches )

PAGE 87

87 Figure A 5. Vacuum chamber lid. (Drafted by Mr. Tom Driscoll of Abbess Instruments; dimensions in inches )

PAGE 88

88 Figure A 6. Right side view of vacuu m chamber. (Drafted by Mr. Tom Driscoll of Abbess Instruments ; dimensions in inches )

PAGE 89

89 Figure A 7. Right side view of vacuum chamber and vibration isolation mass. (Drafted by Mr. Tom Driscoll of Abbess Instruments ; dimensions in inches )

PAGE 90

90 Figure A 8. Crucible. (Dimensions in mm) determined by thermal stress analysis t

PAGE 91

91 Figure A 9. Crucible platform. (Dimensions in mm)

PAGE 92

92 Figure A 10. Crucible suspension. (Dimensions in mm)

PAGE 93

93 Figure A 11. Suspension holder. (Dimensions in mm)

PAGE 94

94 Figure A 12. Quartz vessel. (Dimensions in mm)

PAGE 95

95 Figure A 13 Base plate. (Dimensions in mm)

PAGE 96

96 Figure A 14. Cooling cylinder. (Dimensions in mm)

PAGE 97

97 Figure A 15. Cylinder cap. (Dimensions in mm)

PAGE 98

98 Figure A 16. North clamp. (Dimensions in mm)

PAGE 99

99 Figure A 17. Clamp rod. (Dimensions in mm)

PAGE 100

100 Figure A 18. South clamp (Dimensions in mm)

PAGE 101

101 APPENDIX B BILL OF MATERIALS Table B. Bill of materials for solar thermogravimeter. Category Item Subcomponents Vendor Individual price Qty Total price TG body Vacuum chamber & accessories Vacuum chamber Vacuum gauge Vacuum pump Vit on o ring Vibration isolation mass Vacuum hoses Abbess Instruments $14,967.00 1 $14,967.00 Microbalance Microbalance Electronic box Sartorius $8,930.00 1 $8,930.00 Crucible & supports Crucible Crucible platform Suspensions Ortech 1 Quartz vessel Technical Glass Products $392.90 5 $1,964.50 Radiation shield components & clamps Base plate Cooling cylinder Cylinder cap Clamps Manufacturing, Tooling & Engineering (MTE) $4,175.00 1 $4 175.00 Viton gasket McMaster Carr $9.13 2 $18.26 Cooling w ater tubing McMaster Carr $40.00 1 $40.00 Thermocouple Omega $100.00 1 $100.00 Wireless circuit components Mouser $5.00 1 $5.00 TG support Table McMaster Carr $500.00 1 $500.00 Connections to mass spectrometer Graham condenser Sigma Aldrich $100 .00 1 $100.00 Flask Sigma Aldrich $100.00 1 $100.00

PAGE 102

102 Table B 1 Bill of materials for solar thermogravimeter. (Continued) Category Item Subcomponents Vendor Individual price Qty Total price Connections to mass spectrometer Vacuum connectors Swagelok $ 50.00 6 $300.00 Manifold tee Swagelok $50.00 1 $50.00 Ball valves Swagelok $50.00 2 $100.00 Mass flow controller Alicat $1,000.00 1 $1,000.00 Mirror Metal substrate McMaster Carr $50.00 1 $50.00 Cooling water tubing McMaster Carr $40.00 1 $40 .00 Mirror support Framing extrusions McMaster Carr $50.00 1 $50.00 Brackets McMaster Carr $5.00 10 $50.00 Other Pyrometer Omega $1,000.00 1 $1,000.00 TOTAL: $33,539.76 *Items which have already been purchased/ordered; prices are accurate. Rest ar e estimates.

PAGE 103

103 APPENDIX C MATLAB CODE L=input('Depth of bed in mm?:'); %Prompts input of powder bed depth in millimeters M=L/1000; %Gives bed depth in meters dz=0.0001; %Each depth step is 0.0001 meters (0.1 mm) Nz=M/dz; %Gives number of depth steps maxtim e=1800; %Maximum time in seconds, until which the temperature distribution is to be obtained (30 min) dt=0.01; %Each time step is 0.01 seconds Nt=maxtime/dt; %Gives number of time steps (1800) abs=0.8; %Surface absorptivity of sample (ferrite) to solar rad iation c=750; %Specific heat capacity of sample, J/kg/K d=0.00015; %Particle diameter of sample powder, m emi=0.8; %Surface emissivity of sample ks=4; %Solid thermal conductivity of sample, W/m/K kf=0.01; %Thermal conductivity of surrounding fluid (air at 10^ 4 bar), W/m/K por=0.5; %Porosity of bed rhos=5000; %Density of solid sample, kg/m^3 rhop=rhos*(1 por); %Density of porous sample, kg/m^3 sigma=5.67*10^( 8); %Stefan Boltzmann constant, W/m^2/K^4 Ts=1773; %Temperature of bed at steady state (desired tem perature), K a0=10^( 1.084 6.778*(por 0.298)); %Determine constant a0 to be used in calculating kcond f0=0.8+0.1*por; %Determines constant f0 to be used in calculating kcond krat=ks/kf; %Determines ks/kf ratio to be used in calculating kcond kcond=kf*(1 a0 )*(por*f0+krat*(1 por*f0))/(1 por*(1 f0)+krat*por*(1 f0))+kf*a0*(2*krat^2*(1 por)+(1+2*por)*krat)/((2+por)*krat+1 por); %Solid conductivity through porous bed, W/m/K krad=4*d*sigma*Ts^3*(0.5756*emi*atan(1.5353/emi*(ks/4/d/sigma/Ts^3)^0.8011)+0.18 43); %Radi ant conductivity through porous bed, W/m/K keff=kcond+krad; %Effective thermal conductivity through porous bed, W/m/K alpha=keff/rhop/c; %Thermal diffusivity of porous bed, m^2/s q= 5000000 ; % H eat flux supplied to bed surface from simulator, W/m^2 T=298*one s(Nz+3,Nt+1); %Temperature matrix with rows and columns representing depth and time respectively; # of rows includes ghost rows at top and bottom surfaces, initial temperature is 298K for i=1:Nt T(end,i)=T(end 2,i); %Forces boundary condition at insulated bottom surface; i.e. keff*[T(end,i) T(end 2,i)]/[2*delta z]=0 (no heat flux) T(1,i)=T(3,i) 2*dz*( abs*q/keff+emi*sigma*T(2,i)^4/keff); %Forces boundary condition at irradiated top surface, i.e. [T(3,i) T(1,i)]/[2*delta z]= abs*q/keff+emi*sigma*T(2,i)^4 depth_2D=(T(3:end,i) 2*T(2:end 1,i)+T(1:end 2,i))/dz^2; %Calculates second derivative of temperature with respect to depth time_1D=alpha*depth_2D; %Calculates first derivative of temperature with respect to time using the above matrix and heat equation r elation T(2:end 1,i+1)=time_1D*dt+T(2:end 1,i); %Calculates temperature after delta t

PAGE 104

104 end figure(1) %Plot temperature variance with depth and time imagesc([0 maxtime/60],[0 L],T); title('Temperature plot (imagesc)') %Label time axis from 0 to 30 min, dept h axis from 0 to L colorbar for i=1:Nt+1 if T(end 1,i) >= Ts disp((i 1)*dt); %Gives time in seconds, after which uniform temperature has been reached disp((i 1)*dt/60); %Gives the above time in minutes break; end end

PAGE 105

105 L IST OF REFERENCES [1] Meinshausen M Meinshausen N, Hare W, Raper SCB, Frieler K, Knutti R, Frame DJ, Allen MR. Greenhouse gas emission targets for limiting global warming to 2 C Nature. 2009;458:1158 1162. [2] Allen MR, Frame DJ, Huntingford C, Jones CD Lowe JA, Meinshausen M, Meinshausen N. Warming caused by cumulative emissions towards the trillionth tone Nature. 2009;458:1163 1166 [3] MacKay DJC. Sustainable Energy without the hot air. UIT Cambridge Ltd.; 2009. [4] Loutzenhiser PG, Stamatiou A, Vi llasmil W, Meier A, Steinfeld A. Concentrated solar energy for thermochemically producing liquid fuels from CO 2 and H 2 O. Journal of The Minerals, Metals and Materials Society. 2011;63(1):32 34. [5] Bawase MA, Reve SD, Shete SV, Saraf MR. Carbon number dist ribution by gas chromatography for identification of outlying diesel sample. Paper presented at the 2 nd National Conference of Metrology Society of India on Advances in Metrology, Pune, India, 2012. [6] Nikulshina V, Gebald C, Steinfeld A. CO 2 capture from atmospheric air via consecutive CaO carbonation and CaCO 3 calcination cycles in a fluidized bed solar reactor. Chemical Engineering Journal. 2009;146:244 248. [7] Gebald C, Wurzbacher JA, Tingaut P, Zimmermann T, Steinfeld A. Amine based nanofibrillated c ellulose as adsorbent for CO 2 capture from air. Environmen tal Science & Technology. 2011;45:9101 9108. [8] Lackner KS. Capture of carbon dioxide from ambient air. The European Physical Journal Special Topics. 2009;176(1):93 106. [9] Stolaroff JK, Keith D W, Lowry GV. Carbon dioxide capture from atmospheric air using sodium hydroxide spray. Environmental Science & Technology. 2008;42:2728 2735. [10] Steinfeld A. Solar thermochemical production of hydrogen a review. Solar Energy. 2005;78:603 615. [11] Koga n A. Direct solar thermal splitting of water and on site separation of the products IV. Development of porous ceramic membranes for a solar thermal water splitting reactor. International Journal of Hydrogen Energy. 2000;25:1043 1050. [12] L d J, Villerm aux J, Ouzane R, Hossain MA, Ouahes R. Production of hydrogen by simple impingement of a turbulent jet of steam upon a high temperature zirconia surface. International Journal of Hydrogen Energy. 1987;12:3 11.

PAGE 106

106 [13] Olalde G, Gauthier D, Vialaron A. Film bo iling around a zirconia target. Application to water thermolysis. Advanced Ceramics. 1988;24:879 883. [14] Funk JE, Reinstrom R. Energy requirements in production of hydrogen from water. Industrial and Engineering Chemistry, Process Design and Development. 1966;5(3):336 342. [15] Mathias PM, Brown LC. Thermodynamics of the sulfur iodine cycle for thermochemical hydrogen production. Presentation presented at the 68 th Annual Meeting of the Society of Chemical Engineers, Japan, 2003. [16] Nakamura, T. Hydrogen production from water utilizing solar heat at high temperatures. Solar Energy. 1977;19:467 475. [17] Steinfeld A. Solar hydrogen production via a two step water splitting thermochemical cycle based on Zn/ZnO redox reactions. International Journal of Hydro gen Energy 2002;27:611 619. [18] Kaneko H, Aoki H, Ishihara H, Takahashi Y, Hasegawa N, Suzuki A, Tamaura Y. Conversion of concentrated solar energy to chemical energy (54) Two step water splitting cycle with Ni ferrite/zirconia system. Nippon Kagakkai Ko en Yokoshu 2004;84: 361. [19] Ehrensberger K, Frei A, Kuhn P, Oswald HR, Hug P. Comparative experimental investigations of the water splitting reaction with iron oxide Fe 1 y O and iron manganese oxides (Fe 1 x Mn x ) 1 y O. Solid State Ionics. 1995;78(1 2):151 16 0. [20] Kaneko H, Kodama T, Gokon N, Tamaura Y, Lovegrove K, Luzzi A. Decomposition of Zn ferrite for O 2 generation by concentrated solar radiation. Solar Energy. 2004;76:317 322. [21] Tamaura Y, Steinfeld A, Kuhn P, Ehrensberger K. Production of solar hydrogen by a novel, 2 step, water splitting thermochemical cycle. Energy. 1994;20(4):325 330. [22] Kodama T, Gokon N. Thermochemical cycles for high temperature hydrogen production. Chemical Reviews. 2007;107:4048 4077. [23] Z Graggen A, Steinfeld A. Radiative exchange within a two cavity configuration with a spectrally selective window. Journal of Solar Energy Engineering. 2004;126:819 822. [24] Digitalfire Reference Database. http://digitalfire.com/4sight/material/calcium_carbonate_173.html 2008. [25] Gergely V, Curran DC, Clyne TW. The FOAMCARP process: foaming of aluminium MMCs by the chalk aluminium reaction in precursors. Composites Science and Technology. 2003;63:2301 2310.

PAGE 107

107 [26] Petrasch J. Solar fuels from thermochemical cycles at low pressures. Florida Energy Systems Consortium. http://www.floridaenergy.ufl.edu/wp content/uploads/UF Petrasch.pdf 2011. [27] Schunk LO, Steinfeld A. Kinetics of the thermal dissociation of ZnO exposed to concentrated solar irradiation using a solar driven thermogravimeter in the 1800 2100 K range. AIChE Journal. 2009;55(6):1497 1504. [28] Krupke WF, Shinn MD, M a rion JE, Caird JA, Stokowski SE. Spectroscopic, optical, and thermomechanical properties of neodymium and chromium doped gadolinium scandium gallium garnet. Journal of the Opti cal Society of America B. 1986;3(1):102 114. [29] Momentive. Properties of Fused Quartz. http://www.momentive.com/Products/Main.aspx?id=20339 2012. [30] MakeItFrom.com. Material Properti es Data: Borosilicate Glass. http://www.makeitfrom.com/material data/?for=Borosilicate Glass 2012. [31] The Engineering Toolbox. Thermal conductivity of some common materials and gases. http://www.engineeringtoolbox.com/thermal conductivity d_429.html 2012. [3 2] The Engineering Toolbox. Metals Melting Temperatures. http://www.engineeringtoolbox.com/melting temperature metals d_860.html 2012. [33] Stffl Rudolf GmbH. Elastomer properties overview. http://www.stoeffl.at/download/elastomers properties.pdf 2009. [34] Roth A. Vacuum Sealing Techniques. AIP Press; 1994. [35] Kaviany M. Principles of Heat Transfer in Porous Media. Springer; 1995. [36] Hadley GR. Thermal conductivity of packe d metal powders. International Journal of Heat and Mass Transfer. 1986;29(6):909 920. [37] Krupiczka R. Analysis of thermal conductivity in granular materials International Chemical Engineering. 1967;7(1):122 143. [38] Zehnder P, Schlnder EU. Thermal con ductivity of granular materials at moderate temperatures. Chemie Ingenieur Technik. 1970;42:933 941. [39] Goldman A. Handbook of modern ferromagnetic materials. Kluwer Academic Publishers; 1999. [40] Cremers CJ, Fine HA. Thermal Conductivity, Volume 21. Sp ringer; 1991. [41] Kaviany M. Principles of Heat Transfer. John Wiley & Sons, Inc.; 2002.

PAGE 108

108 BIOGRAPHICAL SKETCH Midori Takagi was born in Kanagawa, Japan. After receiving her B.S. degree in mechanical engineering in 2009 from Toh oku University in Sendai, Jap an and studying environmental materials science at the graduate school of the same university, she moved to University of Florida to fulfill her interest in solar energy. She received her M.S. in mechanical engineering in Summer 2012.