USF Libraries
USF Digital Collections

Low voltage electrochemical hydrogen production

MISSING IMAGE

Material Information

Title:
Low voltage electrochemical hydrogen production
Physical Description:
Book
Language:
English
Creator:
Weaver, Eric P
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Energy
Ruthenium oxide
Sulfuric acid
Solar
Sulfur
Electrolyte
Dissertations, Academic -- Chemical Engineering -- Masters -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Hydrogen production is dependent on natural gas, 90% in the U.S. and 48% of the world's production. Natural gas supply is dwindling and it's price is increasing. Greenhouse gases and air pollutants are emitted when natural gas is used. In a single product production facility, coal is not competitive with natural gas for hydrogen production at current prices. Hydrogen production by direct electrochemical dissociation of water requires a relatively high voltage.Techniques have been developed for manufacturing hydrogen as a lucrative byproduct of IGCC electric power generation, refinery sulfur production and sulfuric acid production for fertilizer production. ^Laboratory experiments have been conducted on small systems to advance the technology and full size commercial plants have been conceptualized and analyzed to establish economic viability.In this thesis, a low voltage electrochemical hydrogen production technique has been developed that entails scavenging of the anode with sulfur dioxide. In an electrochemical cell hydrogen is produced at the negative electrode while the positive electrode is bathed in sulfur dioxide which is oxidized with water to sulfuric acid. The presence of SO2 substantially reduces the equilibrium voltage relative to that required for the direct dissociation of water into hydrogen and oxygen. Also sulfuric acid is a more valuable byproduct than oxygen. More sulfuric acid is produced than any other chemical commodity in the U.S. and is a major economic indicator. ^Hydrogen produced by the electrochemical route being discussed in this thesis illustrates industrial possibilities for large scale-up, economical hydrogen production.In an electrochemical cell, an equilibrium voltage of 1.23 volts is required to decompose water into hydrogen and oxygen. The presence of sulfur dioxide to scavenge the anode can reduce the equilibrium voltage from 1.23 volts to 0.17 volts. The equations shown below are reactions showing the energy requirements.2H2O→ 2H2 + O2 - 4 Faradays @ 1.2V→ 2SO2 + 4H2O 2H2SO4 + 2H2 - 4 Faradays @ 0.17V The thermochemical free energy is reduced from 113kcal/mole to 15kcal/mole if sulfur dioxide is used as a scavenger.In this work, extensive studies to determine the most effective electrodes and catalysts have been carried out. ^The possibilities for photo electrochemical implementation have been investigated and cell design optimization has been performed Experimental methods and results will be presented and discussed.
Thesis:
Thesis (M.S.Ch.E.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Eric P. Weaver.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 95 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001920824
oclc - 190561107
usfldc doi - E14-SFE0001849
usfldc handle - e14.1849
System ID:
SFS0026167:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200385Ka 4500
controlfield tag 001 001920824
005 20081006115121.0
007 cr mnu uuuuu
008 080110s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001849
035
(OCoLC)190561107
040
FHM
c FHM
049
FHMM
090
TP145 (ONLINE)
1 100
Weaver, Eric P.
0 245
Low voltage electrochemical hydrogen production
h [electronic resource] /
by Eric P. Weaver.
260
[Tampa, Fla] :
b University of South Florida,
2006.
520
ABSTRACT: Hydrogen production is dependent on natural gas, 90% in the U.S. and 48% of the world's production. Natural gas supply is dwindling and it's price is increasing. Greenhouse gases and air pollutants are emitted when natural gas is used. In a single product production facility, coal is not competitive with natural gas for hydrogen production at current prices. Hydrogen production by direct electrochemical dissociation of water requires a relatively high voltage.Techniques have been developed for manufacturing hydrogen as a lucrative byproduct of IGCC electric power generation, refinery sulfur production and sulfuric acid production for fertilizer production. ^Laboratory experiments have been conducted on small systems to advance the technology and full size commercial plants have been conceptualized and analyzed to establish economic viability.In this thesis, a low voltage electrochemical hydrogen production technique has been developed that entails scavenging of the anode with sulfur dioxide. In an electrochemical cell hydrogen is produced at the negative electrode while the positive electrode is bathed in sulfur dioxide which is oxidized with water to sulfuric acid. The presence of SO2 substantially reduces the equilibrium voltage relative to that required for the direct dissociation of water into hydrogen and oxygen. Also sulfuric acid is a more valuable byproduct than oxygen. More sulfuric acid is produced than any other chemical commodity in the U.S. and is a major economic indicator. ^Hydrogen produced by the electrochemical route being discussed in this thesis illustrates industrial possibilities for large scale-up, economical hydrogen production.In an electrochemical cell, an equilibrium voltage of 1.23 volts is required to decompose water into hydrogen and oxygen. The presence of sulfur dioxide to scavenge the anode can reduce the equilibrium voltage from 1.23 volts to 0.17 volts. The equations shown below are reactions showing the energy requirements.2H2O 2H2 + O2 4 Faradays @ 1.2V 2SO2 + 4H2O 2H2SO4 + 2H2 4 Faradays @ 0.17V The thermochemical free energy is reduced from 113kcal/mole to 15kcal/mole if sulfur dioxide is used as a scavenger.In this work, extensive studies to determine the most effective electrodes and catalysts have been carried out. ^The possibilities for photo electrochemical implementation have been investigated and cell design optimization has been performed Experimental methods and results will be presented and discussed.
502
Thesis (M.S.Ch.E.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 95 pages.
590
Adviser: Burton Krakow, Ph.D.
653
Energy.
Ruthenium oxide.
Sulfuric acid.
Solar.
Sulfur.
Electrolyte.
690
Dissertations, Academic
z USF
x Chemical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1849



PAGE 1

Low Voltage Electrochemical Hydrogen Production by Eric P. Weaver A thesis submitted in partial fulfillment of the requirements for the degree of Department of Chemical Engineering Master of Science in Chemical Engineering College of Engineering University of South Florida Co-Major Professor: Burton Krakow, Ph.D. Co-Major Professor: John Wolan, Ph.D. Elias Stefanakos, Ph.D. Date of Approval: November 1, 2006 Keywords: energy, ruthenium oxide, sulfur ic acid, solar, sulfur, electrolyte Copyright 2006, Eric P. Weaver

PAGE 2

DEDICATION This work is dedicated to my wife, Sherida, and my daughters; Stefanie-Jo, Sharisse and Salisha, for their patience and support duri ng the development of this research. Willing is not enough; we must do. Johann Wolfgang von Goethe

PAGE 3

ACKNOWLEDGEMENTS I would like to express my appreciation to my committee members Dr. Burton Krakow, Dr. Elias Stefanakos and Dr. John Wo lan for their guidance and support for this work. A sincere added thanks goes to Dr Burton Krakow for his tireless hours of assistance in the lab. A special thanks to my friend and coll eague Maheshkumar Chettiar for laying a good foundation for this research and his ment orship in patience. Also from the beginning of this journey I would like to thank Dr. Venkat Bethanabotla for sharing limited lab space in the beginning and technical assistance when requested. I would like to thank Dr. Sesha Sriniv asan and Dr. Nikolai Kislov for their technical support, advice in the lab and characterization of thin films. I would also like to thank Jay Bieber with University of South Floridas Nanomaterials and Nanomanufacturing Research Center and Subr amanian Krishnan for characterization of thin films. Also from NNRC I would lik e to thank Richard Ev erly for assistance depositing thin films. I owe a debt of gratitude to George Moore for his help with economic analysis. Finally I would like to thank Matt Smith and my other collea gues with the Clean Energy Research Center for assi stance and advice in the lab.

PAGE 4

i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v ABSTRACT viii CHAPTER 1 INTRODUCTION 1 1.1 Hydrogen Production Overview 1 1.1.1 Natural Gas Reforming 1 1.1.2 Oil Reforming 2 1.1.3 Coal Gasification 3 1.1.4 Water Electrolysis 4 1.1.5 Alternative Hydrogen Sources 5 1.2 Scope of this Thesis 6 CHAPTER 2 ECONOMIC ANALYSIS 7 2.1 Overview 7 2.2 Electrochemical Hydrogen Sulfide Dissociation 8 2.2.1 IGCC Power Plants 8 2.2.1.1 Sulfur Removal 9 2.2.1.2 Proposed Process 9 2.2.1.3 Economics 10 2.2.1.4 Carbon Dioxide Sequestration Economics 12 2.2.1.5 Technical Approach 14 2.2.2 Refineries 15 2.2.2.1 Economics 15 2.3 Electrochemical Hydrogen a nd Sulfuric Acid Production 17 2.3.1 Sulfuric Acid Production 17 2.3.2 Proposed Process 18 2.3.3 Plant Design and Economic Analysis 19 2.3.4 History 20 2.3.5 Technical Approach 22 CHAPTER 3 FOUR-INCH DIAM ETER ELECTRODE CELL 23 3.1 Overview 23 3.2 Design Issues 23 3.3 Design Changes 24 3.4 Flow System 26

PAGE 5

ii 3.5 Control System 27 3.6 Experimental Results 27 CHAPTER 4 ELECTRODE AND CATALYST TESTING 30 4.1 Overview 30 4.2 Lead Electrodes 31 4.2.1 Catalyst Deposition 31 4.2.2 Experimental Setup 32 4.2.3 Control and Data Acquisition 33 4.2.4 Experimental Results and Discussion 34 4.3 Carbon Electrodes 35 4.3.1 Catalyst Deposition 35 4.3.2 Electrode Characterization 36 4.3.3 Experimental Results and Discussion 39 4.4 Silicon Electrodes 40 4.4.1 Cleaning 41 4.4.2 Metal and Catalyst Deposition 42 4.4.3 Electrode Characterization 45 4.4.4 Resistance Testing 48 4.4.5 Experimental Results and Discussion 48 4.5 316 SS Electrodes 49 4.5.1 Gold Plating 49 4.5.2 Catalyst Deposition 50 4.5.3 Electrode Characterization 51 CHAPTER 5 TWO-INCH ELECTRODE CELLS 53 5.1 Overview 53 5.2 Liquid Electrolyte Cell 53 5.2.1 Design Considerations 53 5.2.2 Flow System/Operation 56 5.3 Solid Electrolyte Cell 58 5.3.1 Design Considerations 58 5.3.2 MEA Development 60 5.3.3 Flow System/Operation 61 5.3.4 Monitoring Sulfur Dioxide Diffusion 62 5.4 Goals and Limitations 62 5.5 Experimental Results and Discus sion of Preliminary Experiments 63 5.5.1 Experiment 1 64 5.5.2 Experiment 2 64 5.5.3 Experiment 3 64 5.5.4 Experiment 4 65 5.5.5 Experiment 5 65 5.5.6 Experiment 6 65 5.5.7 Experiments 7-9 66 5.5.8 Experiments 10 and 11 67

PAGE 6

iii 5.5.9 Conclusions of Preliminary Experiments 68 5.6 Experimental Results and Di scussion of Final Experiments 68 5.6.1 Experiment 1 69 5.6.2 Experiment 2 69 5.6.3 Experiment 3 70 5.6.4 Experiment 4 70 5.6.5 Experiment 5 72 5.6.6 Experiment 6 73 5.6.7 Experiment 7 73 5.6.8 Experiment 8 73 5.6.9 Experiment 9 75 5.6.10 Experiment 10 75 5.6.11 Experiment 11 76 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS 78 REFERENCES 80 APPENDICES 83 Appendix A: Data Acquisition Hardware and Programs 84 Appendix B: Characterization Equipment 87 Appendix C: Cell Drawings 91

PAGE 7

iv LIST OF TABLES Table 2.1 Hydrogen Production 7 Table 2.2 IGCC Economic Analysis 11 Table 2.3 CO 2 Sequestration Economics 13 Table 2.4 Refinery Electrolysis Plant Costs 16 Table 2.5 Refinery Sulfur Processing Comparison 16 Table 2.6 Sulfuric Acid Pl ant Economic Comparison 19 Table 2.7 Sulfur Cycle Energy Requirements 21 Table 5.1 Experimental Parameter Comparison 77

PAGE 8

v LIST OF FIGURES Figure 2.1 FutureGen Power Plant Schematic 8 Figure 2.2 IGCC with Claus Proces s and Electrolytic Process 10 Figure 2.3 Viscosity of Sulfur, Conductivity of CsHSO 4 vs Temperature 14 Figure 2.4 Refineries with Claus Process and Electrolytic Process 15 Figure 2.5 Comparison of Curre nt Acid Process and Electrochemical Alternative 18 Figure 2.6 Four-Inch Diameter Electrode Cell 22 Figure 3.1 New Four-Inch Electrode Cell Base 24 Figure 3.2 Four-Inch Electrode Cell and New Base 24 Figure 3.3 Four-Inch Electrode Cell with New H 2 SO 4 Inlet Tube 25 Figure 3.4 Four -Inch Diameter Electrode Cell Flow System 26 Figure 3.5 Current Voltage Curve Comparison 28 Figure 4.1 Electrode Testing Apparatus 33 Figure 4.2 Lead Electrode Results 34 Figure 4.3 SEM Image of Carbon Electrode with RuO 2 Deposition 37 Figure 4.4 EDS Spectrum RuO 2 on C 38 Figure 4.5 SEM Image of Carbon Elect rode with Ru Map Overlay 38 Figure 4.6 EDS Map of Ru on Carbon Electrode with EDS Map of O 2 Carbon Overlay 39 Figure 4.7 Carbon Electrode Re sults, Carbon Electrode 39 Figure 4.8 Carbon Electrode Results, RuO 2 Coated Cathode 40 Figure 4.9 Tube Furnace 43 Figure 4.10 Wafer Holder 43 Figure 4.11 SEM Image of Si Electrode with RuO 2 Deposition 45 Figure 4.12 EDS Spectrum RuO 2 on Si 46 Figure 4.13 EDS Map of Ru on Si Electrode with EDS Map of O 2 on Si Overlay 46

PAGE 9

vi Figure 4.14 XRD Spectra for Sol-Gel RuO2 Deposition on Si 47 Figure 4.15 Resistance Test Jig 48 Figure 4.16 Resistance Test Results 49 Figure 4.17 XRD Spectra of RuO 2 Deposited on Au on 316 SS 52 Figure 5.1 Two-Inch Low Pressure Liquid Electrolyte Cell 54 Figure 5.2 Two-Inch 80 psig Liquid Electrolyte Cell 55 Figure 5.3 PFD for Two-Inch Li quid Electrolyte System 57 Figure 5.4 Three Compartment Solid Electrolyte Cell 59 Figure 5.5 Membrane Hot Press Diagram 60 Figure 5.6 PFD for 3 Compartment Cell 61 Figure 5.7 Preliminary Two-Inch Cell Comparison 66 Figure 5.8 Hybrid 3 Compartment Cell 67 Figure 5.9 Total Current vs Voltage Curves: Experiments 1 Through 4 71 Figure 5.10 Current Density vs Voltag e Curves: Experiments 1 Through 4 71 Figure 5.11 Total Current vs Voltage Curves: Experiments 4 Through 8 74 Figure 5.12 Current Density vs Voltag e Curves: Experiments 4 Through 8 74 Figure 5.13 Total Current vs Voltage Curves: Experiments 6 and 9 Through 11 75 Figure 5.14 Current Density vs Voltage Cu rves: Experiments 6 and 9 Through 11 76 Figure A.1 Power Supply Control Labview User Interface 84 Figure A.2 Power Supply Control Labview Block Diagram 85 Figure A.3 Data Aquisition Cont rol Labview User Interface 85 Figure A.4 Data Aquisition Cont rol Labview Block Diagram 86 Figure B.1 Hitachi S-800 Sca nning Electron Microscope 87 Figure B.2 Hitachi S-800 Specifications 87 Figure B.3 X'Pert Diffractometer 88 Figure B.4 X'Pert Diffract ometer Specifications 88 Figure B.5 Anatech Limited Hu mmer X Sputter Coater 89 Figure B.6 Anatech Limited Hummer X Sputter Coater Specifications 89 Figure B.7 Plasma Sciences CRC-100 Sputter Tool 90 Figure B.8 Plasma Sciences CRC-100 Sputter Tool Specifications 90

PAGE 10

vii Figure C.1 Low Pressure Cell Anode Compartment 91 Figure C.2 Low Pressure Cell Electrode Assembly 92 Figure C.3 Low Pressure Cell Cathode Compartment 92 Figure C.4 80 psig Cell Anode Compartment 93 Figure C.5 80 psig Cell Electrode Assembly 94 Figure C.6 80 psig Cell Cathode Compartment 95

PAGE 11

viii LOW VOLTAGE ELECTROCHEMI CAL HYDROGEN PRODUCTION Eric P. Weaver ABSTRACT Hydrogen production is dependent on natural gas, 90% in the U.S. and 48% of the worlds production. Natural gas supply is dwindli ng and its price is increasing. Greenhouse gases and air pollutants are emitted when natural gas is used. In a single product production facility, coal is not competitive with natural gas for hydrogen production at current prices. Hydrogen produc tion by direct electrochemical dissociation of water requires a relatively high voltage. Techniques have been developed for manufacturing hydrogen as a lucrative byproduct of IGCC electric power generation, refi nery sulfur production and sulfuric acid production for fertilizer produc tion. Laboratory experiment s have been conducted on small systems to advance the technology a nd full size commercial plants have been conceptualized and analyzed to establish economic viability. In this thesis, a low vo ltage electrochemical hydrogen production technique has been developed that entails scavenging of the anode with sulfur dioxide. In an electrochemical cell hydrogen is produced at the negative electrode while the positive electrode is bathed in sulfur dioxide which is oxidized with water to sulfuric acid. The presence of SO 2 substantially reduces the equilibrium voltage relative to that required for the direct dissociation of wate r into hydrogen and oxygen. Also sulfuric acid is a more valuable byproduct than oxygen. More sulfuric acid is produced than any other chemical commodity in the U.S. and is a major economic indicator. Hydrogen produced by the electrochemical route being disc ussed in this thesis illustra tes industrial po ssibilities for large scale-up, economi cal hydrogen production.

PAGE 12

ix In an electrochemical cell, an equilibrium voltage of 1.23 volts is required to decompose water into hydrogen and oxygen. The presence of sulfur dioxide to scavenge the anode can reduce the equilibrium voltage from 1.23 volts to 0.17 volts. The equations shown below are reactions showing the energy requirements. 2H 2 O 2H 2 + O 2 4 Faradays @ 1.2V 2SO 2 + 4H 2 O 2H 2 SO 4 + 2H 2 4 Faradays @ 0.17V The thermochemical free energy is reduced from 113kcal/mole to 15kcal/mole if sulfur dioxide is used as a scavenger. In this work, extensive studies to determine the most effective electrodes and catalysts have been carried out. The possibilities for photo electrochemical implementation have been investigated and cell design optimization has been performed Experimental methods and results will be presented and discussed.

PAGE 13

1 CHAPTER 1 INTRODUCTION 1.1 Hydrogen Production Overview The utmost importance to the evolution of the Hydrogen Economy, is a sustainable supply of hydrogen. Current met hods of industrial scale hydrogen production are not sustainable. The primary source of industrial hydrogen is from steam reforming of natural gas also known as steam methane reforming (SMR). SMR accounts for approximately 48% of the worlds production of hydrogen and 90% of the United States production [1]. Oil reforming and coal gasification come in next at 30% and 18% respectively. Electrolysis ranks last in curre nt industrial hydrogen production techniques at 4% [1]. Currently all of the above techniques involve CO 2 emissions thereby defeating the main purpose of the proposed switch to hydrogen. Ev en the electrolysis of water under current techniques involves CO 2 emissions because the electricity for electrolysis is provided by power plants where the electricity is predomin antly produced from coal and natural gas. The technical aspects of the preceding t echniques, along with other techniques which are not yet performed at industrial scale will be briefly explained in the following sections. The advantages and disadvantages of the techniques will also be discussed. 1.1.1 Natural Gas Reforming Natural gas reforming is also known as steam methane reforming due to its main constituent, methane. The first step in SM R is to pass methane and steam over a nickel catalyst at high temperature and pres sure, 750C -1000C and 15 atm 25 atm respectively [2]. The resultant reaction is displayed in equation 1. CH 4 + H 2 O CO + 3H 2 (1)

PAGE 14

2 The second step then oxidizes the CO using H 2 O as steam at 200C-475C, resulting in more hydrogen and CO 2 as equation 2 displays. This is known as the watergas shift reaction. CO + H 2 O CO 2 + H 2 (2) SMR is currently the most economical way to produce hydrogen and produces less CO 2 than other hydrocarbon based techniques. Yet, there are several issues against the use of SMR to fuel the Hydrogen Econom y. The first issue is the production of CO 2 One of the key objectives in developing hydrogen as the main energy carrier is the reduction of greenhouse gases. With the development of CO 2 sequestration techniques this first issue could be corrected. However, this would probably affect SMRs status as the most economical technique for producing hydrogen because CO 2 sequestration adds considerable costs [3]. The next draw back is the instability of the cost of natural gas. The Cost of natural gas has a history of fluctuation and is projected to increase [4]. This situation would certainly be exacerbated by the in creased demand introduced by large scale hydrogen production. The final negative aspect of SMR as a long term hydrogen production technique is the supply of natural gas. There is much debate as to how long the worlds reserves of natural gas will last. The Energy Information Agency from the Department of Energy concludes that even counting on reserves that are not at present ec onomically viable; the world possesses approximately 60 years worth of reserves at predicted consumption trends [4]. 1.1.2 Oil Reforming In oil reforming heavy hydrocarbons are first cracked or split into lighter hydrocarbons. They are then reformed sim ilarly to natural gas. Hydrogen from oil produces even more CO 2 than natural gas. There is gr owing demand for oil and therefore prices are rising. There is polit ical unrest in the Middle East where the greatest reserves of oil are located. Oil reserves are beli eved to be less than that of natural ga s The EIA estimates approximately 20 years worth of rese rves at predicted consumption trends [4].

PAGE 15

3 1.1.3 Coal Gasification Experiments with coal gasification are documented as early as the 1600s [5]. Several reactions occur to form syngas from coal. Reactions 3-6 below are the primary reactions for CO production. C + O 2 CO 2 (3) C + CO 2 2CO (4) 2C + O 2 2CO (5) Reaction 3 and 4 are exothermic. When te mperatures are too great the solid waste ash unreacted minerals from the coal, is possibly fu sed. To alleviate this problem, steam is added as shown in reaction 6. C + H 2 O CO + H 2 (6) The final step is the water-gas shift reaction as with steam methane reforming equation 2 [5]. CO + H 2 O CO 2 + H 2 (2) The temperatures and pressures for these re actions depend on the type of gasifier. Fixed bed gasifiers have a high temperature around 2100 F and an exit gas temperature around 800 F -1200 F. The operating pressure is 435 psig or greater. Fluidized bed gasifier has uniform temperat ure because it operates like a continuously stirred tank reactor (CSTR). Gas exits at the temperat ure of the reactions between 1700 F and 1900 F. Fluidized bed reactors operate between 5 and 435 psig. En trained flow gasifiers also operate like a CSTR. They have an oper ating temperature between 2300 F and 3200 F. They have the highest operating pressure of the gasifiers at great er than 725 psig. Coal gasification is considered by some, esp ecially in the U.S, as the bridge to the Hydrogen Economy [6]. Reserves of coal are great er than other fossil fuels in the United States and throughout the world [4]. A dditionally, coals cost is lower than other hydrocarbons [4]. Unfortunately coal has the largest car bon to hydrogen atomic ratio ranging from 2:1 to 1:1 compared to 1:2 for oil and 1:4 for methane. Coal reserves are larger than other fossil fuels but they are still not very plentiful.

PAGE 16

4 At current usage rates the EIA estimates approximately 190 years of reserves, but with projected consumption rates, that drops to around 90 years of reserves [4]. 1.1.4 Water Electrolysis Water electrolysis closes out the list of current indus trial hydrogen production techniques Water is dissociated in an electroc hemical reactor via an electrical bias across two electrodes in an aqueous alkaline solution. Electrons pass through the circuit created by a power source to the cathode. Hydrogen is created at the cathode. The reaction is represented by equation 7. 2H 2 O + 2e H 2 + 2OH (7) The hydroxyl ions pass through the electrolyte to the anode forming oxygen, water and returning the electrons to the circuit. The resultant reaction is de picted in equation 8. 2OH O 2 + H 2 O + 2e (8) The overall reaction is equation 9. H 2 O H 2 + O 2 (9) The theoretical voltage needed for this reaction is 1.23 volts. The applied voltage will be higher than the theoretical volta ge, typically 1.65-1.8 volts [7], due to overvoltage, an effect caused by system resistance, reaction kineti cs at the electrode surfaces and concentration gradients. Currently, water electrolysis is the mo st expensive industr ial hydrogen production technique [7]. This will change in the near future because the cost of fuels is projected to go up and the capitol cost of electrolyzers is projected to go down [8] [9] [10]. Water electrolysis itself produces no harmful bypr oducts or waste. The only products are hydrogen and oxygen. The problem with electrolys is lies in the amount of power needed and the source of the electric power The voltage is too high for single band gap and thin film semiconductors whic h produce between 0.5 and 0.7 volts These are the most common solar cells used curren tly for photovoltaic cells. In order to use these cells they must be connected in series to obtain the required voltage. This means it takes 3 to 4 cells to meet the voltage requirements. Therefore, these cells can not be placed directly in the electrolyte and used as the electrodes.

PAGE 17

5 Both of these facts cause an increase in cost and a loss of efficiency because the single crystal silicon solar cells producing 0.5 volts are the mo st efficient solar cells made commercially. If electri city from power plants is used, carbon dioxide is still released because hydrocarbons are currently used as fuel in the majority of pow er plants. If CO 2 sequestration is implemented, benefits are incr eased because sequestration is much easier and cost effective at point-sources like power plants. Most people envision mass hydrogen production with electrolysis and al ternative energy sources like wind and solar With the current pricing structure, alternative sources can not compete economically Some people propose nuclear as the power source of choice. Even if the environmental arguments are addressed, the political aspe cts of nuclear power leave its future questionable. 1.1.5 Alternative Hydrogen Sources Research on using biomass and waste to produce hydrogen is at this time a highly active area. Biomass include s crops grown specifically as fuel stock, waste from agricultural crops, wood chips and many other organic wastes. Researchers are looking into sugar cane waste, orange peel waste, fast growing trees and grasses along with other sources of biomass in Florida. Biomass and so lid waste can be gasified similar to coal or added to coal. Unfortunately it is far le ss efficient. Research to improve the conversion of biomass to syngas and production of hydr ogen from syngas is being done and can make this a more viable option [11]. Work is being done with multijunction semiconductors to increase the voltage output in a photoelectrochemical cell. These cells can be more expensive to make and are less efficient than single junction solar cells. The U.S. Nati onal Renewable Energy Laboratory (NREL) has improved on these efficiencies using a combination of photovoltaic and photoelectrochemical cells [11]. Biological techniques are being studied that adapt photosynthesis for hydrogen production which include biophotolysis of water by microalgae and cyanobacteria.

PAGE 18

6 Other biological techniques study bacteria that decompos e organic compounds, such as sugar or starch bearing waste waters, and produce hydrogen. So far costs are high and efficiencies are low 1.2 Scope of this Thesis The first intent of this work is to determine if two of the hydrogen production techniques of interest to this group are econom ically viable. The technical and historical importance of these two techniques will be discussed with the economic analysis. The two techniques that economical analyses are done on are electrochemical dissociation of H 2 S to hydrogen and sulfur and the electrochemical oxidation of SO 2 with H 2 O to produce hydrogen and su lfuric acid. Next, improvements on an existing electrolysis cell for SO 2 oxidation to inhibit the flow of sulfur dioxide to the cathode co mpartment are evaluated. Previous work done with this cell suggested the use of tungsten carbide as a catalyst [12]. Tests with tungsten carbide electrodes will be discussed. Redesign of the control and data acquisition systems for this cell and late r cells will be presented. An electrolysis cell redesign to implement sm aller electrodes to facilitate the use of catalyst deposition techniques available to the group will be described. These catalyst deposition techniques and results from experime nts with these catalysts will be discussed. The conversion of this cell desi gn to a cell utilizing a polymer electrolyte for experiments similar to historical work [13] will be presented. Conclusions developed from experimental results will be drawn out and suggestions for further work will be presente d. Some of the technical aspects of the future work will be discussed.

PAGE 19

7 CHAPTER 2 ECONOMIC ANALYSIS 2.1 Overview Vast amounts of hydrogen will be needed to fuel the Hydrogen Economy. Two of the techniques researched by our group can be utilized in conjunction with the production of sulfuric acid which has higher production than any chemical commodity in the United States. This could constitute mo re hydrogen than is currently produced by all techniques. Sulfuric acid is manufactured by oxidizi ng sulfur dioxide with oxygen from air. If the oxygen were derived from water by the electrochemical route being developed in this project, large quantities of hydrogen would be produced as a valuable byproduct. Table 2.1 demonstrates how much hydr ogen can be produced as a result of commercialization of this and other byproduct reactions of sulfur compounds that our group is studying. It shows that these pro cesses can produce almost 3 times the current annual production of H 2 in the United States within current sulfuric acid production levels. Our attention currently stresses the oxidations of SO 2 and H 2 S, which by themselves can produce 1/3 more hydrogen than is produced now. Table 2.1 Hydrogen Production Reaction Hydrogen Yield (10 9 kg) Free Energy (Kcal/mol) Equilibrium Voltage (V) H 2 S S+H 2 @ 400K 0.8 9 0.19 2H 2 O+S SO 2 +2H 2 1.6 2H 2 O+SO 2 H 2 SO 4 +H 2 0.8 15 0.17 Total 3.2 Merchant Hydrogen Production in 2000 =1.2 X 10 9 Kg

PAGE 20

2.2 Electrochemical Hydroge n Sulfide Dissociation The economic analysis of hydrogen sulf ide dissociation was performed to compare our electrochemical process to the typical Claus Process in IGCC power plants and refineries. Both analyses provide legitim ate reason for further i nvestigation into this process. 2.2.1 IGCC Power Plants The FutureGen program that President Bush supports is based on integrated gasification combined-cycle technology (IG CC). Figure 2.1 illustrates a FutureGen power plant. In an IGCC'S gasifier, carbon-based raw ma terial reacts with steam and oxygen at high temperature and pressure to produce comb ustible synthesis gas. The gasifier's high temperature vitrifies inorganic materials into a course, sand like materi al, or slag that is sold for road building. The synthetic fuel leaves the gasifier and is further cleaned of impurities. It is used in the system to run primary and secondary gas and steam turbines, similar to a natural gas combined-cycle generating system. Figure 2.1 FutureGen Po wer Plant Schematic 8

PAGE 21

9 The primary environmental benefits include increased efficiency and nearly zero air pollution. Most pollutants are removed before combusti on and are not created when the fuel is burned. In the case of sulfur, it is collected in a form that can be sold. This is a big change for conventional co al plants, where even clean ones produce a lake-sized impoundment of sulf uric slurry by pulling sulfur compounds from the stack flue gas. IGCC power plants are the cleanes t coal-based power generation facilities in the world. The capital cost for an IGCC power plant is greater than for a conventional plant. This is partly justified by higher effici ency, lower emissions and the potential for producing byproducts. However, the bottom lin e is that conventional coal and natural gas fired power plants can produce electric pow er at a lower net cost. While increased fuel costs and environmental regulations may eventually close this gap, technical improvements are needed to help achieve this and remove the economic barrier to deployment of these cleaner and more efficient systems. 2.2.1.1 Sulfur Removal IGCC power plants now in operation extract the sulfur from the synthesis gas as hydrogen sulfide. The hydrogen sulfide is extracted along with carbon dioxide in a stream called acid gas. Partial oxidation of th e acid gas with air yields elemental sulfur and water (Claus Process) with a waste stream of dilute carbon dioxide in nitrogen. The process is illustrated in figure 2.2. There is one IGCC plant that uses complete oxidation of hydrogen sulfide to sulfuric acid. The sulfur recovery sy stems of IGCC power plants can be improved and thereby produce an addi tional revenue stream that will lower the cost of IGCC electricity. 2.2.1.2 Proposed Process Electrolytic oxidation of the extracted hydrogen sulfide can yield sulfur and hydrogen. The carbon dioxide is separated before electrolysis which leaves a concentrated carbon dioxide waste stream which is easier to sequester. This process is illustrated in Fig. 2.2. The value of the hydroge n makes the system more profitable.

PAGE 22

Figure 2.2 IGCC with Claus Proce ss and Electrolytic Process The energy benefit of electro lytic decomposition of the H 2 S as compared to decomposing water is illustrated in Equations (10) and (11). H 2 S(g) S(l) +H 2 (g) 2 Faradays @0.19V G o = 8.9 kcal/mole @ 400K (10) H 2 O(l) H 2 (g) + O 2 (g) 2 Faradays @ 1.2V G o = 57 kcal/mole (11) 2.2.1.3 Economics The scale of the plant developed for this techno-economic analysis is for replacement of the sulfuric acid producti on plant at TECOs Polk Power Station IGCC power plant in Florida. It is a nominal 250 MW (net) IGCC power plant. This is a comparable size to the Wabash River IG CC power plant which has a nominal 262 MW (net) rating. The plant design includes separation of the acid gas into its two main constituents, CO 2 and H 2 S. The separation of CO 2 and H 2 S minimizes the flow thr ough the electrolyzer and facilitates CO 2 sequestration. A waste stream from th e electrolyzer is separated into its components for recycling or disposal. 10

PAGE 23

11 The economics of this process rely h eavily on power costs and electrolyzer pricing. When dealing with electrolysis, cost optimization of these two parameters is needed. There is currently not enough technical inform ation on current density and voltage requirements to perform this optimi zation. A reasonable cu rrent density of 2000 A/m 2 at 0.7 volts is used. It is assumed that similar costs as those published in literature for water electrolyzers will apply. Literature values for electrolyzers currently range from $2280/m 2 -$2850/ m 2 Costs are projected to drop to $356/ m 2 -$855/ m 2 as technological advances are ma de [8] [9] [10]. Table 2.2 s hows the economics calculated at the short term cost of $2280/m 2 Table 2.2 IGCC Economic Analysis POWER PLANT PARAMETERS IGCC Plant Capacity (Gross MW) 315 Sulfur Production (Tons/day) 63 Coal sulfur content (%) 2.5 ELECTROLYSIS PLANT INVESTMENTS Apparatus to Remove Carbon Dioxide $586,849 Electrolyzer $4,560,000 Balance of Plant $6,213,151 Total Electrolysis Plant Investments $11,360,000 ANNUAL CAPITAL AND O&M COSTS Annualized Capital Costs $1,730,000 Labor $516,000 Catalysts, water and other operating costs $632,417 Electricity $2,500,000 Total Annual Capital and O&M Costs $5,378,417

PAGE 24

12 Table 2.2 Continued HYDROGEN PRODUCTION COSTS Costs to Produce 1 Ton of S + 125lb. of H 2 $245 Avoided Cost of Claus Process/Ton of S $137 Net Production Cost of 125 lb. of H 2 $108 Net Production Cost of 1 lb. of H 2 Gas $0.86 Hydrogen Liquefaction Cost ($/lb.) $0.55 Net Production Cost of 1 lb. of Liquid H 2 $1.41 MARKET PRICE OF LIQUID H 2 ($/LB.) $2.00 GROSS PROFIT ON LIQUID H 2 ($/LB.) $0.59 Savings to Electrical Production Cost ($/MWh) $0.63 2.2.1.4 Carbon Dioxide Sequestration Economics The electrolytic production of hydrogen and sulfur never generates any carbon dioxide. However, there is carbon dioxide in the hydrogen sulfide feedstock extracted from the synthesis gas. Processing th e feedstock for byproduct hydrogen production can facilitate sequestering of this carbon dioxide at a cost far lo wer than conventional flue gas separation and sequestering. The proposed separation process costs approximately $21/ton C. The DOE website on CO 2 sequestration estimates carbon storage, transportation, and sequestrati on to cost approximately $50/ ton C [3]. The economics of CO 2 sequestration are laid out in Table 2.3. The CO 2 capture cost of $21/ton C is not listed in Table 2.3 as a cost because it is already accounted for in the cost of H 2 production. When the acid gas of an IGCC power plan t is partially oxidized with air in the Claus process, the rejected CO 2 is mixed with a large quantity of nitrogen. For sequestering, the carbon dioxide has to be se parated from the nitrogen or sequestered together with the nitrogen. These are both undesirable. The DOE website on CO 2 sequestration estimates carbon capture in dilu te streams to cost approximately $150/ton C [3]. If carbon emission reductions are mandated as they are in Europe, CO 2 trading prices will likely follow prices in Europe.

PAGE 25

13 The price for CO 2 credits on the European Climate Exchange as of 27 Jun, 2005 is $38.01/ ton CO 2 [14]. This is the value for captured CO 2 used in Table 2.3. The table clearly shows that sequestration of the concentrated CO 2 captured in the proposed process is economically beneficial at th e European trading price for CO 2 It also shows that capture and sequestration of the diluted CO 2 from the current process would lose money at the European trading price for CO 2 Table 2.3 CO 2 Sequestration Economics Projected Carbon Sequestration Economics Proposed Process Current Process IGCC Plant Capacity (Gross MW) 315 315 Acid Gas Stream Carbon (tons/hr) 3.04 3.04 Carbon Costs Carbon Capture Cost ($/ton C) 0 150 Additional Cost for Storage, Transportation and Sequestration ($/ton C) 50 50 Carbon Value CO 2 Trading Price ($/ton CO 2 ) 38.01 38.01 CO 2 Trading Price ($/ton C) 139.38 139.38 Profits or Losses Profits From Collection, Storage, Transportation and Sequestration ($/ton C) 89.38 -60.62 Profits From Collection, Storage, Transportation and Sequestration ($/MWh) 0.86 -0.58 The cost for IGCC power is greater than for conventional electric power. Rule of thumb estimates say the difference is ar ound $10-$20/MWh. Tables 2.2 and 2.3 show the proposed process can help reduce that cost difference. The initial profits from H 2 sales reduce that difference by $0.63/MWh. Long term projections for H 2 profits will reduce that difference by $1.38/MWh.

PAGE 26

If carbon emission reductions are mandated, as they are in Europe, sequestration of the concentrated CO 2 captured in the proposed process would provide an additional economic benefit of $0.86/MWh. This gives a near term savings of $1.49. 2.2.1.5 Technical Approach Hydrogen sulfide will be decomposed at a temperature at which sulfur is a low viscosity liquid so that it can run out of the electrolytic cell quickly and easily. The solid line curve in figure 2.3 gives the viscosity of liquid sulfur and shows that it is minimized near 150C. We will operate near this temperature. To do this we are seeking to exploit some recent developments in solid state electrolytes. These involve inorganic crystals whose prot on conductivities rise rapidly with temperature. One of these is cesi um hydrogen sulfate whos e conductivity is shown by the dashed curve in Figure 2.3. The conductivity scale is logarithmic. The conductivity goes up by about 5 orders of magnitude between 120 and 150C and the material is a good conductor at the te mperature where we want to work. 6 7 8 9 10 11 12 120 130 140 150 160 Temperature (Degree Celsius) -8 -7 -6 -5 -4 -3 -2 -1 0 log[ (1/ cm)] of CsHSO4 Viscosity of Sulfur log of conductivity ( ) of CsHSO4 Viscosity of Sulfur (Centipoises) Figure 2.3 Viscosity of Sulfu r, Conductivity of CsHSO 4 vs Temperature 14

PAGE 27

2.2.2 Refineries Refineries create and use hydrogen using SMR. They also have a hydrogen sulfide waste stream that needs treatment. They currently use the Claus process similar to IGCC power plants. The electrochemi cal split of hydrogen sulfide can produce hydrogen for some of the needs thereby reducing CO 2 emissions from SMR. A graphical comparison of the two processes is depicted in figure 2.4. An economic analysis was performed to compare the existing technology to the electrochemical approach. Refinery Air H2S Partial Oxidation S H e a t H2O N2 Refinery H2S Electrolyzer S D C H2 Claus Process Electrolysis Figure 2.4 Refineries with Claus Pr ocess and Electrolytic Process 2.2.2.1 Economics The same sulfur capacity plant of 63 tons per day was evaluated for the refinery analysis. The plant is considerably smaller due to a cleaner hydrogen sulfide stream. Additional separation equipment is not needed to purify the hydroge n sulfide before the electrolyzer. This lowers the capitol cost of the plant for refineries as compared to IGCC power plants. The same assumptions about electrolyzer costs were made. The costs used for the analysis are di splayed in table 2.4. The comparison of costs for the two processes and the determined savings are in table 2.5. 15

PAGE 28

16 Table 2.4 Refinery Electrolysis Plant Costs ELECTROLYSIS PLANT INVESTMENTS Electrolyzer $4,560,000 Balance of Plant $3,428,365 Total Electrolysis Plant Investments $7,988,365 ANNUAL CAPITAL AND O&M COSTS Annualized Capital Costs $1,216,680 Labor $336,195 Catalysts, water and other operating costs $625,149 Electricity $1,821,315 Total Annual Capital and O&M Costs $3,999,339 Table 2.5 Refinery Sulfur Processing Comparison Typical Refinery Costs/DayCurrent Costs to Produce 63 Ton of S @$137/T $8,631 Hydrogen Cost (7875# @$0.60) $4,725 Total Cost/Day $13,356 Electrochemical Hydrogen Production$/Day Hydrogen & Sulfur Production Cost $11,938 Daily Electrochemical Savings $1,418 Potential Yearly Saving (335days/year) $475,030 The preceding table shows the process looks attractive. Additional savings could be made in the event of car bon taxes due to reduced CO 2 emissions.

PAGE 29

17 2.3 Electrochemical Hydrogen and Sulfuric Acid Production The economic analysis of hydrogen and sulfuric acid production was performed to compare our electrochemical process to a typical sulfuric acid production plant in the manufacture of fertilizers. The analysis provides legitimate reason for further investigation of this process. 2.3.1 Sulfuric Acid Production As stated in the introduction to this chapter sulfuric ac id is typically produced by first oxidizing sulfur to sulfur dioxide. Air is initially passed thr ough a dryer to eliminate moister that could condense and corrode equipm ent further down the line. The dryer is an absorption tower with sulfuric acid which is very hygroscopic and therefore easily absorbs the water in the air. The air is then compressed to a level high enough to provide a pressure difference through the entire plant. Excess air is then fed to a burner with sulfur that produces sulfur dioxide and h eat. The reaction in the burner is given by equation 12. S + O 2 SO 2 + Heat (12) The sulfur dioxide is then passed into a catalytic converter ov er a catalyst. The sulfur dioxide is oxidized by the some of th e excess air to produce sulfur trioxide and more heat. This oxidation r eaction is shown in equation 13. SO 2 + O 2 SO 3 + Heat (13) These gases now consist of air, sulfur dioxide and sulfur trioxide. They are passed through what is typically called an interpass absorber where sulfur trioxide is removed and the sulfur dioxide a nd air are returned to the catalytic converter. This shifts the equilibrium of th e sulfur dioxide oxidation reaction and improves the conversion. The final product from the catalytic convert er and the sulfur trioxide from the interpass absorber are passed through the fi nal absorber with water. This produces between 93% to 98% sulfuric acid by wei ght. The reaction from this mixture is expressed in equation 14. SO 3 + H 2 O H 2 SO 4 + Heat (14)

PAGE 30

As can be seen by equations 12, 13 and 14, the current process for sulfuric acid manufacture produces a lot of heat. As stated earl ier, this heat is used to make electricity. This amount of heat produces more energy th an the sulfuric acid plant consumes. The electricity can be used in the balance of the fertilizer plant or sold to the local utility. This is denoted as a negative energy cost in the economic analysis. 2.3.2 Proposed Process A low voltage electrochemical hydrogen production technique has been developed involving scavenging of the anode with sulfur dioxid e. In an electrochemical cell with a sulfuric acid elec trolyte hydrogen is produced at the negative electrode while the positive electrode is bathed in sulfur dioxide which is oxi dized to sulfuric acid. The presence of SO 2 to scavenge the anode substantia lly reduces the equilibrium voltage relative to that required for the direct dissociati on of water into hydrogen and oxygen as can be seen in equations 15 and 16. SO 2 +2H 2 O H 2 SO 4 + H 2 @ 0.17 V G 0 = 15 kcal/mol (15) 2H 2 O 2H 2 + O 2 @ 1.2 V G 0 = 113 kcal/mol (16) Sulfuric acid is a more valuable byproduct than oxygen. The differences between a typical sulfuric acid plant and the proposed electrochemical plant are shown in figure 2.5. Figure 2.5 Comparison of Current Acid Pr ocess and Electrochemical Alternative 18

PAGE 31

2.3.3 Plant Design and Economic Analysis The concept plant for the economic analysis is based on a replacement plant for current 3500 tpd H 2 SO 4 plant. It is assumed that th e sulfur burner, heat exchangers, piping, etc. are unchanged on a cost basis. The plant concepts were designed and priced by Chemcad modeling, discussion with industr y, and plant design estimations [15]. The SO 2 /N 2 separation unit was modeled in Chemcad and priced using plant design estimates [15]. A combination of co mpression, cooling and absorption were used in the model for SO 2 /N 2 separation. The sulfuric acid concentration unit was modeled in Chemcad and priced as the SO 2 /N 2 separation unit. The concept for the sulfuric acid concentration unit is for vacuum distillation. The pricing compared very closely to the vacuum flash system by Aker Kvaerner Chemetics. Discussions with Aker Kvaerner Chemetics were the starting point for design parameters of the distillation system. The economics of this process rely h eavily on power costs and electrolyzer pricing. When dealing with electrolysis, cost optimization of these two parameters is needed. There is currently not enough technical inform ation on current density and voltage requirements to perform this op timization. A current density of 2000 A/m 2 at 0.7 volts is used. This current density is reporte d in literature. It is assumed that similar costs as those published in literature for water electrolyzers will apply. Literature values for electrolyzers currently range from $2280/m2-$2850/ m 2 Costs are projected to drop to $356/ m2-$855/ m 2 as technological advances are made [8] [9] and [10]. Table 2.6 shows the economics calculated at the short term cost of $2280/m 2 Table 2.6 Sulfuric Acid Plant Economic Comparison Economic Evaluation Assumptions (SO2 Electrolysis) Production (tons/day 98% H2SO4) 3500 Production (tons/day H2) 71.4 Plant Operation Life (yrs) 30 Electrical Energy Cost (2002 $/MWh) 40 Cost of Capital (%) 15 19

PAGE 32

Table 2.6 Continued Economic Evaluation Current Process Electrochemical Process Total Capital (MM$) 51.2 133.8 Annualized Cost(MM$) Capital 7.8 20.4 O & M 32.6 42.1 Utilities -7.5 21.1 Total cost of Products ($/ton H2SO4) 27 ($/ton H2SO4 + 40 lb H2) 66.44 Cost Difference 42.44 gH2 Cost($/lb H2) 1.04 LH2 Cost ($/lb H2) 1.68 2.3.4 History In the 1970s and early 1980s Westinghouse in the U.S. and the European Commissions Joint Research Centre (JRC) in Ispra, Italy investigated what became known as the Westinghouse cycle or Mark 11. The sulfur cycle is so called because like the main focus of this thesis sulfur dioxi de is used to depolarize the anode in an electrolyzer to form hydrogen and sulfuric aci d, but unlike this thesis the sulfuric is decomposed to water, oxygen and sulfur dioxide so the sulfur dioxide can be recycled. The sulfur cycle is represented by equations 17, 18 and 19. SO 2 +2H 2 O H 2 SO 4 + H 2 (17) H 2 SO 4 SO 3 + H 2 O (18) SO 3 SO 2 + O 2 (19) Equations eighteen and nineteen are very energy intensive. Equation eighteen requires temperatures of ar ound 700 K. Nineteen requires a temperature of 1200 K [16]. 20

PAGE 33

21 The economic analysis for this thesis did not consider the requirements for equations 18 and 19 because at current and near term production needs for hydrogen it is more economical to sell the su lfuric acid. In the future it may be necessary to decompose the acid if this technique is used to pr oduce more hydrogen and therefore more sulfuric acid than the market can handle. At that time it would be essential for solar heat collection or possibly nuclear wast e heat to be used for the he at requirements of sulfuric acid decomposition in order to maintain favorable economics with such high energy requirements. Westinghouse discussed both of the possibilities for thermal inputs [17] [18]. The energy requirements for the Westinghouse process we re estimated in an article by D. Van Velzen and placed in the table 2.7 [16]. Table 2.7 Sulfur Cycle Energy Requirements 65% H 2 SO 4 55% H 2 SO 4 Electricity (0.62V) 315 KJ/mol 315 KJ/mol Concentration Step 169 KJ/mol 240 KJ/mol Thermal Dissociation 272 KJ/mol 272 KJ/mol Total 756 KJ/mol 827 KJ/mol Investigations in Germany determined wh at might be acceptable parameters for an electrolytic cell. They developed a 3 co mpartment cell that had a center compartment with flowing 30% H 2 SO 4 between two cation exchange membranes. They used graphite felt for the electrodes. The cathode felt was pla tinized. They used a homogeneous catalyst HI in the anode compartment. The German group also did work with optimizing the carburization of WO to form WC electrodes which they determined looked promising as catalytically active cathodes [19]. Westinghouse reported results better than those used for the economics of this thesis. They reporte d achieving 200 mA/cm 2 at only 0.6 V. They used a bipolar membrane electrolyzer with carbon porous flow through electrodes. They reported using ruthenium oxide as the cataly st on both anode and cathode.

PAGE 34

2.3.5 Technical Approach There are two different technical approach es for the electrolyzer considered in this thesis. The first studied involved only a liquid electrolyte, sulf uric acid. The first cell was designed strictly for this purpose. It can be seen in figure 2.6. This design easily facilitates the switch to a low band gap single junction solar cell when the voltage requirements are met by the optimized design pa rameters. This design requires a flow gradient to restrict flow of sulfur dioxide from th e anode compartment to the cathode compartment. This cell was operated at am bient temperature and 100 psi. The design parameters are discussed in an earlier thesis [12]. 1,8,15 1/8" NPT Fittings 2 Sensor Ring 3 Glass Window 4 Cathode 5 Barrier/ Anode 6 Steel Ring 7,9 Teflon Housing 10 Steel Base Ring 11 Bolts 12 Viton O-Ring 13 Anode 14,17 Capillary Tube 16 Teflon Insulator/Barrier Figure 2.6 Four-Inch Diameter Electrode Cell It was determined that smaller electrodes might facilitate cata lyst deposition better than the four inch electrodes for the group. A cell to utilize two inch diameter or two inch square electrodes was developed. This second design was easily manipulated for use with liquid electrolyte as the first design, solid electrolyte, or a combination of both. It was designed for use of materials readily available locally and ma terials that could be adapted with out use of the machine shop. Diag rams of the cells for two inch electrodes will be shown in the experimental results section. These cells were operated at ambient temperature and 0-90 psig. 22

PAGE 35

23 CHAPTER 3 FOUR-INCH DIAMETER ELECTRODE CELL 3.1 Overview The first work done towards this thesis was a continuation of the work done by Chettiar [12]. This work was done with the fo ur inch diameter cell briefly described at the end of the last chapter. It had been pr oven with the four inch cell that process was producing relatively pure hydrogen at potenti als lower than water electrolysis. It appeared the use of tungsten carbide lowere d the overpotential but the adhesion of the evaporated tungsten carbide was not sufficient. 3.2 Design Issues Where the previous work had let off there were still some design issues with the four inch cell. First the diffusion of sulf ur dioxide to the cathode was significantly slowed down but was never stopped. Another issue is the fact there was still considerable overpotential adherent with the four inch cell. The previous thesis by Chettiar concluded that tungsten carbide and ru thenium oxide as catalysts could reduce some of this overpotential [12]. Finally the window for the four inch cell was sus ceptible to cracking and therefore a Teflon and aluminum replacement ha d to be installed. The lack of a window requires a system for sensing the hydrogen pocket in the cell as described by Chettiar [12]. This adds to the comp lexity of the overall system and the lack of a window means design changes will be needed to facilitate the use of a solar cell to replace the electrodes.

PAGE 36

3.3 Design Changes The first design change was a remodel of th e stand for the four inch cell. It had been determined that the most probable reason for the glass window breakage was the added stress to two of the bolts for sealing the system caused by use of those same two bolts for attachment of the cel l to the stand, see figure 2.6. The stand was redesigned with a ring that the entire base of the cell could sit on therefore creating no stress points in the cell. See figures 3.1 and 3.2. A window rated at 500 psi instead of 200 psi was also purchased for added assurance. Figure 3.1 New Four-Inch Electrode Cell Base 24 Figure 3.2 Four-Inch Electrode Cell and New Base

PAGE 37

The next two design changes were implemented to deter sulfur diffusion to the cathode compartment. It was noted that su lfur build up on the cathode due to sulfur dioxide diffusion starte d at the bottom of the cathode. It was postulated that possibly a flow gradient was created by hydrogen bubble ev olution and introduction of sulfuric acid to the cathode compartment through a tube facing up the cathode, see figure 2.6 item 17. This flow gradient could po ssibly have been pulling fluid from the bottom of the cell and/or forcing sulfuric acid over the top of the cathode to the a node compartment then out the exit at the botto m of the anode compartment. This flow of acid from the cathode to the anode compartment could force the sulfuric acid and sulfur dioxide from the center of the anode to the bottom then through the electrolyte bridge to the cathode compartment. Another possibility determined was that due to the fact that liquid sulfur dioxide is denser than 20% sulf uric acid, the sulfur dioxide introduced at the center of the anode could be sinking to the bottom of the anode compartment. This would leave a short path across the electrolyte bridge to the cathode compartment. The first design change to stop sulfur dioxide diffusion due to the previously stated reasons was to extend the insulator be tween the two electrodes to the bottom of the cell and seal off that section of the electrolyte bridge, see figure 2.6 item 16. The second change to slow sulfur dioxide diffusion was to place a 180 degree bend in the end of the cathode sulfuric acid inlet to deter an upward flow gradient of sulfuric acid, see figure 3.3. Figure 3.3 Four-Inch Electrode Cell with New H 2 SO 4 Inlet Tube 25

PAGE 38

The last change implemented to the four inch cell was due to previous misfortune with catalyst deposition. Instead of deposit ing tungsten carbide as a catalyst, tungsten carbide electrodes were purchased. 3.4 Flow System The same flow system as described by Chettiar [10] was used in the operation of the four inch cell. The flow diagram as depicted by Chettiar [10] is seen in figure 3.4. Liquid sulfur dioxide and dilute sulfuric ac id are pumped into the bottom center of the cell using two separate piston metering pumps. Figure 3.4 Four-Inch Diameter Electrode Cell Flow Diagram A third metering pumps creates a flow gradient from the cathode compartment to the anode compartm ent by pumping dilute sulf uric acid into the cat hode compartment at a greater rate than it is pumped into the anode compartment. Pressure in the cell is controlled by a 100 psi back pressure regulator in the product acid line at the bottom of the cell. Pressure is monitored by a teflon pressure transducer monitored by a Labview program. 26

PAGE 39

27 Hydrogen is released from the cell th rough a micrometering valve and bubbled through water into a gas collec tion cylinder. Gas samples can be taken before or after bubbling through water through a septum or collected in an infrared cell. 3.5 Control System The control system discussed by Chettiar fo r earlier work with the four inch cell was not implemented for the experiments with the four inch cell for this thesis. A computer crash required manual control and data acquisition for the experiments done with the four inch cell. An Agilent 3640A dc power supply was used for the electrodes. An analog dc power supply and multimeter were used for the sensor ring and the pressure transducer. A voltage slightly higher th an the voltage applied to th e electrolysis electrodes is supplied to the sensor ring an it is monitore d for amperage by the multimeter. The sensor ring serves three purposes; sense sulfur dioxi de escaping the anode compartment, reject hydrogen ions to the cathode compartment and oxidize escaping sulfur dioxide before it leaves the cell. Twelve volts dc is supplied to the pressure transducer with the analog power supply. An output voltage between 0 and 5 volts is read with the multimeter and multiplied by 30 to determine the pressure in the cell. 3.6 Experimental Results Only one successful run was made with the four inch cell. Several seemingly unsuccessful runs did manage to prove that two of the design changes were successful. Unless there were issues with the flow system or pressure regulation, migration of sulfur dioxide to the cathode compartment had been deterred. There were several issues with pressure regulation in the unsuccessful runs, of ten resulting in excessive pressure, due to blocked pressure regulators. The new stand design and higher pressure rated glass resulted in no damage to the glass window.

PAGE 40

The one successful run with the four inch cell implemented all of the changes discussed in section 3.1.2. It was deemed a success because it was the longest run with no sulfur dioxide diffusion to the cathode compartment. It was the first run implementing the t ungsten carbide electro des. This design change was not deemed successful. A curre nt voltage comparison chart is shown in figure 3.5. It compares the last run of Chettiar [12] to the run with tungsten carbide electrodes. Current Voltage Comparison0 0.02 0.04 0.06 0.08 0.1 0.12 0.60.70.80.911.11.21.31.4Voltage (V)Current (A) WC Results Chettiar Results Figure 3.5 Current Voltage Curve Comparison The results at lower voltages look deceptively encouraging. Due to time constraints the current at the lower voltages wa s not allowed to settle to its lowest point as was later done at the highe r voltages. Possibilities for the lower curr ent with the tungsten carbide electrodes include the fact that platinum from the previous run is a better cathodic catalyst than tungsten carbide and mu ch less anodic surface area due to much lower porosity of tungsten carbide than carbon. Another drawback to the use of th e tungsten carbide electrodes was the production of a purple substance. Oxides of cobalt and manganese are known to be blue to purple in color. Cobalt and manganese are known binders used to hold tungsten carbide compounds together. The electrodes we re treated electrochemically, soaked in ambient temperature sulfuric acid for days and treated in boiling sulfuric acid for hours to attempt to leach out the binders to no avail. 28

PAGE 41

29 It was hoped the leaching of the binder would eliminate contamination of the sulfuric acid with the oxide and increase por osity of the tungsten car bide. Sulfuric acid contamination did not noticeably diminish after extensive treatment nor did porosity noticeably increase. This led to the work investigating electrode materials and catalysts described in the following chapter.

PAGE 42

30 CHAPTER 4 ELECTRODE AND CATALYST TESTING 4.1 Overview Results from the work by Chettiar with the four inch cell and results from preliminary work for this thesis led to th e decision to do some studies of electrode materials and catalyst deposition techniques be fore proceeding. All the previous work showed that the biggest apparent problem was with the anode material. All metals tested showed effects of corrosion on the anode when a potential was applied to the electrodes. It was believed a charge build up due to inhibited ion diffusion in porous carbon was a draw back to carbon electrodes. Carbon electr odes were still considered due to there extensive use in other research and commercial systems. Literature reviews led to the decision to continue to try tungsten carbide and ruthenium oxide as catalysts [ 19] [20]. It was decided to attempt to try lead electrodes due to leads relatively low melting point ( 327.4 C) and malleability. It was believed these attributes could lead to a successful way to attach tungsten carbide and ruthenium oxide mechanically to an electrode. Gold is known to be resistant to sulfuric acid and this fact was proven by earlier work when gold was used on the carbon anode and showed no deterioration. To decrease the porosity of the anode it was decided to try gold deposition on 316 SS for a protective coating. The ultimate goal of this research is to lower the voltage required for electrolysis low enough that the voltage can be supplied by si ngle junction silicon solar cell. Due to the fact silicon will eventually be used it was decided to start testing its durability as an electrode in this corrosive envi ronment. It was also decided to try different catalyst and metal deposition techniques as might be need ed on the silicon substrate in the future.

PAGE 43

31 4.2 Lead Electrodes Lead electrodes were fabricated by me lting lead shot in a combustion boat 100 mm long and 20 mm wide. The lead shot was melted under nitrogen to prevent oxidation. The first electrodes showed considerable dark gray to black residue. Wire brushing and melting repeatedly reduced the amount of dark gray to black residue. It was noted that the lead shot had the same residue. The lead shot was soaked five minutes in aqua regia which is a mixture of concentrated hydrochloric acid and nitric acid in a 3 to 1 ratio. This treatment formed a white residue on the pellets th at was removed by stirring in deionized water. If the pellets were exposed to air they turned gray. If the shot was not going to be immediately place in tube furnace under nitrogen it was stor ed in 20% acetic acid to prevent oxidation. The treatment also reduced the residue on electrodes but did not eliminate it completely. 4.2.1 Catalyst Deposition Prior to deposition lead electrodes were formed as described in the previous section. To deposit tungsten carbide approximately 0.3 grams of 99% pure 10 micron tungsten carbide from Sigma Aldrich was placed in the bottom of the same combustion boat as was used to make electrodes. It was then placed in tube furnace under nitrogen flow and heated to 300 C to ensure the powder was dry. After removal from the tube furnace the tungsten carbide powder was spr ead across bottom of the boat and covered with the lead electrode made earlier. The tungsten carbide was placed at the bottom of the boat due to the fact its density 15,700 kg m -3 is greater than that of lead at 11,340 kg m -3 The boat was then placed in the tube furnace which was then purged with nitrogen before the atmosphere was changed to hydr ogen to ensure no oxidation and reduce any oxides already formed. The temperature was raised to 400 C for approximately 1 hour to make certain the lead was completely melted. The temperature was then lowered slowly to 260 C before the furnace was shut off and the atmosphere returned to nitrogen. This appeared to leave a fairly even coverage of tungsten carbide on one side of the lead electrode.

PAGE 44

32 Ruthenium oxide 99.9% pure also from Sigm a Aldrich was deposited in a similar manner to tungsten carbide. Ruthenium oxide is less dense than lead 7050 kg m -3 compared to 11340 kg m -3 respectively. There were conc erns that spreading ruthenium oxide only on top of the electrode would not break the surface tens ion and stick to the lead. Ruthenium oxide, approximately 0.2 gram s, was dried then half was placed in the bottom of the boat and the second half on top of the electrode. The boat was then placed in the tube furnace under hydrogen. The temperature was raised to 500 C for 45 minutes and lowered slowly to 260 C before the oven was turned off. The hydrogen reduced the ruthenium oxide to ruthenium. This re sulted in good adhesion. One side was totally covered and the other side showed approximately 1/3 rd coverage. The following ruthenium oxide depositions were done under a nitrogen atmosphere to prevent the re duction of ruthenium oxide. Th e first deposition resulted in only approximately 1/8 th of one side of the electrode had ruthenium adhered to it. The next deposition resulted in total coverage on one side and approximately 1/3 rd coverage on the other just like the ruthenium deposition. 4.2.2 Experimental Setup The experiments were carried out in a sy stem of beakers set up to slow diffusion of sulfur dioxide diffusion from the anode to the cathode compartment. The anode beaker was 30 ml beaker filled with glass b eads. The anode beaker was submerged in a larger cathode beaker. The 30 ml anode beaker was filled to the 25 ml mark with glass beads and 20% sulfuric acid. Sulfur dioxide gas was then slowly bubbled into the 30 ml beaker for one hour using a micrometering valve. Saturation of sulfuric acid with sulfur dioxide was assumed. One of the lead electrodes describe d in the previous sect ion was then drilled for attachment of a 316 SS electrical lead. The lead electrode with no deposition was wire brushed to remove any oxide. The 316 SS lead was attached and the electrode was place in the saturated sulfuric acid solution. The 30 ml beaker was then filled to capacity with glass beads.

PAGE 45

The 30 ml beaker setup and an oversized carbon cathode was then placed in a 250 ml beaker see figure 4.1 The 250 ml beaker was the filled with 20% sulfuric acid to approximately 6.35 mm over the top of the 30 ml beaker to create an electrolyte bridge between the anode compartment and cathode comp artment. There was no flow or stirring in this setup. Figure 4.1 Electrode Testing Apparatus 4.2.3 Control and Data Acquisition Voltage was supplied to the electrodes with an Agilent 3640A dc power supply. The power supply was controlled by a progr am written in a graphical programming language for Labview, a software package from National Instruments. The program was written to control two Agile nt 3640A power supplies. The building blocks for the program are drivers for the Agilent 3640A downloaded from National Instruments website. The drivers were written into th e program to allow voltages and amperages to be constantly monitored from the deskt op computer and constant manipulation of voltages from the computer if needed. The pr ogram also allowed data to be saved to a file that could be opened later for data analysis. A screen shot of the user interface and the block diagram can be seen in appendix A. 33

PAGE 46

Voltages were also monitored by a Na tional Instruments graphical program receiving data from a NI PCI-4351 data acqui sition card linked to a TBX-68T terminal block, see appendix A. The program reads vol tages from the data acquisition devices. One part of the programs designed for use in th e four inch cell mon itors voltages supplied by a pressure transducer that monitors system pressure. The voltage is correlated to pressure within the program. Several channels can be read for meas uring several voltage s or pressures if needed. Voltages and pressures are shown nume rically and graphically. Data is sent to a file for later data analysis. Voltage is m onitored by this second program to verify the output of the power supply program and due to a desire to monitor residual voltage after the power supplies are tu rned off and disconnected from the system. A screen shot of this user interface and the block diagra m can also be seen in appendix A. 4.2.4 Experimental Results and Discussion Experiments were run with the setup described in section 4.2.2. Voltage was applied and monitored using th e systems in the previous s ection. Voltage was raised from 0.5 V dc to 1.5 V dc in 0.25 increments at similar time increments for each experiment. A plot of the results is shown in figure 4.2. Electrode Comparison0 0.5 1 1.5 2 2.5 00.511.52 Voltage (V)Current Density (mA/Cm2) Pb/WC Anode Carbon Cathode 4/18/05 Pb Anode Carbon Cathode 4/26/05 Pb/Ru Anode Carbon Cathode 5/01/05 Pb/RuO2 Anode Carbon Cathode 5/10/05 Figure 4.2 Lead Electrode Results 34

PAGE 47

35 The results are shown as voltage versus current density. Current density is the current as read by the power supply divided by the cross sectional area of the electrode. The active area of the anode, the smaller electr ode, was used. The active area for the lead only electrode was considered to be both sides of the electrode. The active area of the tungsten carbide coated lead electrode was consid ered to be only one side of the electrode because an oxide covered the other side of the electrode. The active area of the ruthenium and ruthenium oxide coated lead el ectrodes was considered to be one side and 1/3 rd of the other side with the other 2/3 rds of one side covered with an oxide. The ruthenium coated electrode showed the lowest amperage. The electrodes with ruthenium oxide and tungsten carbide showed better results at lower voltages. It is believed that this is due to the catalytic activ ity. The seemingly better results of the lead electrode at higher voltages ar e possibly due to the oxidation of lead not from sulfur dioxide oxidation to sulfuric acid. Only the best results from the tests with lead have been presented. There were several issues with sporadic re sults, contamination of the surface of the lead and adhesion of the catalysts to the lead. These issues were in conflict with the reasoning for attempting to use lead electrodes therefore th e tests with lead elec trodes were abandoned. 4.3 Carbon Electrodes To attempt to get away from corrosion issues it was decided to try carbon as a substrate for catalyst deposition. Electrode s were fabricated from a carbon sample donated by Asbury Carbons. Four inch by inch electrodes inch th ick were cut out of the sample. A hole was the drilled in the el ectrode to facilitate the connection of an electrical lead. The same e xperimental setup, control system and data acquisition system were used as with the experiments with lead. 4.3.1 Catalyst Deposition It was decided to try deposition of ruth enium oxide on carbon electrodes. A solgel technique describe in an article by Cons tantinou et al in Catalysis Letters was used [21].

PAGE 48

36 The article describes a typical sol-gel techniqu e in which a precursor, usually a salt and in their case RuCl 3 is dissolved in a solvent. The solution is then placed on a substrate and the temperature is raised to evaporate the so lvent. The sample is then calcined at high temperature. Specifically for these experiments the ruthenium oxide was deposited as follows. Approximately 0.251 grams of RuCl 3 was placed in 10 ml Isopropanol. This was supposed to create a 0.121 M solution. This actually created a heterogeneous mixture with some of the RuCl 3 dissolving into solution but a lot of it remained as solid particles in the mixture. The mixture was painted on to the carbon electrode s. The electrodes were dried at 80 C for approximately 45 mi nutes in a tube furnace exposed to outside air. The temperature was then slowly raised to 500 C. The temperature was maintained for 1 hour. The set temperature is changed to 22 C and the temperature was allowed to drop naturally. 4.3.2 Electrode Characterization To determine if RuO 2 was successfully deposited th e carbon electrodes were characterized using a Hitachi S-800 Scanning Electron Microscope to perform scanning electron microscopy (SEM) and Energy Disp ersive X-ray Spectroscopy (EDS) on the sample. The Hitachi S-800 SEM is owned and maintained by the University of South Florida Nanomaterials and Nanomanufacturing Research Center. A photograph and specifications are in appendix B. A brief description of SEM as depicted by Iowa State University College of Engineering Department of Materials Scien ce and Engineering was addapted as follows [22]. A beam of elect rons is generated by an electron gun. The system is run at high vacuum to ensure no interference of gas molecules with the instable electron beam and to prevent damage to the electron gun by reaction of gas molecules. The beam is attracted through an anode, condensed by a condenser lens and focused as a very fine point on the sample by an objective lens. Scan coils are energized and create a magnetic field which deflects the beam back and fort h in a controlled pattern.

PAGE 49

The electrom beam hits the sample, producing secondary electrons, auger electrons, X-rays and back scattered electr ons. The secondary electrons and the back scattered electrons are collected by detect ors which convert them to a voltage and amplify them. The amplified voltage is app lied to a grid of a cathode ray tube, CRT. This causes the intensity of a spot of light to change. The image consists of thousands of these spots of light of varying intensity. This corresponds to the topography of the sample. A succinct definition of EDS was adapted from the University at Buffalo, School of Dental Medicine website [23]. As previously stated X-rays are emitted when the electron beam hits the sample. They are creat ed by shell transitions created within the individual elements caused by the energy of the electron beam. Each X-ray has the has an energy level consistant with its shell level and parent element. Detection and measurement of this energy allows for elementa l analysis. This syst em can also be used to map the surface of the sample and show the distribution of elements. An SEM image of the sample is seen in figure 4.3. The EDS spectrum and elemental analysis are shown in figure 4.4. Figure 4.3 SEM Image of Carbon Electrode with RuO 2 Deposition A map of the surface of th e carbon electrode was also taken to determine the distribution of elements. The map of Ru atom s is overlayed on the SEM image in figure 4.5 with Ru depicted in yellow. Figure 4.5 show s a fairly even distribution of Ru atoms. 37

PAGE 50

Figure 4.6 then shows the map of Ru overlayed on the map of O 2 again with Ru depicted in yellow and O 2 depicted in green. This figure shows Ru and O 2 are evenly distributed therefore the Ru compound present is RuO 2 The amount of Ru depicted in figures 4.4-4.6 is low because the electron b eam of the Hitachi S800 Scanning Electron Microscope does not have e nough energy to disperse k she ll Ru electrons. Therefore only l shell electron transitions are detected. EDS Spectrum0 100000 200000 300000 400000 500000 600000 01234567Energy (kEv)Intensity (Counts)C Ka O Ka S Ka Ru LI Ru La Ru Lb Ru Lg 8 Figure 4.4 EDS Spectrum RuO 2 on C Figure 4.5 SEM Image of Carbon Electrode with Ru Map Overlay 38

PAGE 51

Figure 4.6 EDS Map of Ru on Carbon Electrode with EDS Map of O 2 Carbon Overlay 4.3.3 Experimental Results and Discussion The set up of the experiments with carbon was the same as the experiments with lead. The operation was slightly different Only two voltages were used for these experiments. This was done to assure that water was not being di rectly decomposed. Each experiment was run for hour at 0.5 volts then hour at 1 volt. Ruthenium oxide was used as both an anodic catalyst and cat hodic catalyst. Different combinations of carbon and ruthenium oxide coated electrodes with and with out sulfur dioxide were used. The results for carbon cathodes are plotte d in figure 4.7. The results for ruthenium oxide coated cathodes are plotted in figure 4.8. Carbon Cathode Comparison0 0.05 0.1 0.15 0.2 00.20.40.60.811.2Voltage (V)Current Density (mA/cm2) CRuO2_SO2 C-RuO2_No SO2 C_SO2 C_No SO2 Figure 4.7 Carbon Electrode Results, Carbon Cathode 39

PAGE 52

C_RuO2 Cathode Comparison0 0.1 0.2 0.3 0.4 0.5 0.6 00.20.40.60.811.2Voltage (V)Current Density (mA/cm2) C-RuO2_SO2 C-RuO2_No SO2 C_SO2 C_No SO2 Figure 4.8 Carbon Electrode Results, RuO 2 Coated Cathode Results appear to show catalytic activ ity for hydrogen evolution on the cathode as well as sulfuric acid production at the anode. A larger benefit app ears to occur at the cathode. These results considered positive a nd led to further work with the sol-gel deposition technique for ruthenium oxide and carbon electrodes which will be discussed later. 4.4 Silicon Electrodes In another attempt to stay away from corrosive metals silicon electrodes were tested. The electrodes were made from two inch n-Si 0.4-0.6 ohm-cm 0.55 mm thick and two inch p-Si 7.2-10.8 ohm-cm 0.55 mm th ick. These were the lowest resistivity wafers the group had easy access to. Low resi stivity wafers were desired to minimize voltage loss through the wafer. Voltage loss is calculated simply with Ohms Law, which is depicted in equation 20. V= IR (20) In equation 20 V is voltage loss in vo lts, I is current in amperes and R is resistance in ohms. Resistan ce is calculated using the quoted resistivity of the material the current is being passed through using equation 21. R= LA -1 (21) In this equation R is ag ain resistance in ohms, is resistivity in ohm-cm, L is length in cm of the resistive path and A is the cross sectional area of the resistive path in cm 2 40

PAGE 53

41 The experiments and tests were set up to lower the resistance of the electrodes. Therefore they were set up to use the 2 inch face as the cross sectional area of the wafer and the 0.55 mm thickness as the length. 4.4.1 Cleaning Critical to any deposition is the cleanline ss of the substrate. This is important because it might not be known what effect the contaminants could have on adhesion of the material deposited. Even more critical for deposits on silicon where low resistance is desired is the removal of silicon dioxide from the surface. Silicon dioxide has resistivity in the range of 10 12 -10 18 ohm-cm. With resistivities as high as just st ated, Angstroms of native oxide can cause considerable voltage loss thr ough that layer. One angstrom of native oxide on a two inch wafer can result in an added re sistance of 4930 ohms which corre lates to a voltage loss of 493 volts. Fortunately due to impurities in th e oxide layer and diffusion of metals used for contacts or catalysts the effects are not ne ar as bad as possible with pure native oxide. Initially wafers were cleaned with a ve ry basic cleaning technique. They were rinsed with Acetone to clean organics and th en soaked in 100:1 HF buffered oxide etch. The initial resistances measured were quite hi gh. The cleaning process was considered as one of a few possible reasons for these high resistance measurements. Another more detailed cleaning system was found on the website from Holon Institute of Technology in central Israel [24]. The procedure begins with a solvent clean with DI rinse. That step is followed by a RCA-1 clean with rinse. The procedure is concluded with a HF dip, DI rinse and blow dry. The solvent clean uses two solvents because sometimes the acetone only treatment can leave a residue. The solvent clean is used to remove oils and organics from the surface of the wafer. The wafer is firs t placed in 55 C acetone for 10 minutes. The wafer is then placed in ambient methanol fo r 2-3 minutes. After the methanol soak the wafer is soaked in DI and blown dry with nitrogen.

PAGE 54

42 The RCA-1 process is so named because it was developed at RCA in the 1960s by Werner Kern [25]. The solution for an RCA-1 clean consists of 5 parts water, 1 part 27% ammonium hydroxide and 1 pa rt 30% hydrogen peroxide. The RCA-1 process is also used to remove organics. The solution is prepared by first adding ammonium hydroxide to water and heating to approximately 70 C. Hydrogen peroxide is then added to the so lution. Once it bubbles for 1-2 minutes the solution is ready. Once the solution is ready the wafer can be placed in the solution for 15 minutes. After the allotted time the wafer is removed to a beaker with overflowing DI water. After several changes of water the wafer should be removed under running DI water to deter organics on the surface of water from depositing on the wafer. The last step is the HF dip. The hydrofl uoric acid is mixed to a 2% solution with DI water. The wafer is placed in the HF solution for only 2 minutes. The wafer is removed from solution and rinsed under running DI The wafer is then tested to see if the surface is hydrophobic. An oxide free surf ace is hydrophobic and water will bead on the surface. 4.4.2 Metal and Catalyst Deposition The first deposition tried with silicon was the deposition of ruthenium oxide on the n-Si wafers. The sol-gel technique used with the carbon elec trodes was also used with the n-Si wafers. The wafers were cleaned using the firs t cleaning technique described in the previous section. The deposition technique was slightly changed for the ruthenium deposition on silicon. The deposition was performed in a ni trogen rich atmosphere. A nitrogen rich atmosphere was chosen in hopes to create a reaction with the sili con dioxide if there was not enough oxygen to react with all the ru thenium. It was hoped that the reaction depicted by equation 22 would occur. 4RuCl 3 + 3SiO 2 + O 2 4RuO 2 + 3SiCl 4 (22)

PAGE 55

It was determined by XRD that RuO 2 was formed. The results will be shown in the results section. Much more study has to be done to determine if reaction 22 actually occurred. After a 45 minute etch in the buffered oxi de etch the wafers were immediately painted on both sides with the sol-gel soluti on. The wafer was placed in a tube furnace positioned vertically, see figure 4.9. Figure 4.9 Tube Furnace Due to the fact the diameter of the silicon wafer was the same as the largest diameter tube that could fit in our furnace no tube was used. An apparatus was made to suspend the wafer in the middle of the furnace, see figure 4.10 43 Figure 4.10 Wafer Holder

PAGE 56

44 The wafer was placed in the holder and suspended in the tube furnace. Nitrogen was turned on to flush the furnace. The set temperature of the furnace was slowly raised to 80 C. The 80 C set temper ature was maintained for one hour. The temperature was then slowly raised to 335 C. The temperature was chosen because it is just below the decomposition temperature of RuCl 3 360C. It was hoped that this would allow the RuCl 3 to react with the SiO 2 before decomposing. The 335 C temperature was maintained for one hour. The temperature was then raised from 335 C to 385C. This temperature was maintained for 30 minutes. The temperature was finally raised to 500 C and maintained for 30 more minutes. The furnace was then allowed to cool naturally by lowering the set temperature to 22 C. Upon cooling the wafer was rinsed a nd scrubbed vigorously under DI, rinsed and scrubbed vigorously under Isopropanol a nd dried with nitrogen. This procedure was repeated up to 3 times per wafer. The first metal depositions on silico n were done with a Hummer X Sputter Coater from the University of South Fl oridas Nanomaterial s & Nanomanufacturing Research Center (NNRC), see appendix B for pictures and specificati ons At the time our main concern was resistance. The Hummer X was used due to its capability to ion etch prior to deposition. It was believed this could help re duce the resistance due to the native oxide formed immediately after chemi cal etching when the wafer is exposed to air. The Hummer X has the capability to pump down to approximately 35 mtorr. This is not a very low vacuum pressure and a llows for oxygen to be left in the chamber. Wafers were etched using the first cleaning procedure. A 60-70 nm layer of AuPd was deposited on both sides of multiple wafers. Some of the wafers were ion etched before deposition and some were not. The results displayed in figure 4.16 show the Hummer X AuPd depositions did not lower the resistan ce as low as desired. The AuPd also did not hold up to sulfuric acid. The Pd was attacked by the sulfuric acid and the AuPd flaked off the wafer.

PAGE 57

The poor results with the Hummer X le d to the use of another NNRC sputter system, the CRC 100 Sputter System by Plasma Sciences Inc, see appendix C for picture and specifications. This system was chosen due to its ability to reach higher vacuum, approximately 0.001 mtorr, and its abil ity to also ion etch. It was determined early on the system could not be used to ion etch due to the configuration used by NNRC. Multiple wafers could be coated at the same time. Deposits were made on the n-Si and p-Si wafers described earlier. Wa fers were prepared using the second more detailed cleaning procedure. A layer of approximately 2000 angstroms of silver was deposited on the side of the wafer that woul d not come in contact with sulfuric acid. A layer of approximately 100-200 angstroms of pt was deposited on the reactive side of the wafer. Results from this procedure show the best resistance results to date. 4.4.3 Electrode Characterization To determine if RuO 2 was successfully deposited the silion electrodes were characterized using the Hitachi S-800 Scanni ng Electron Microscope as described with the carbon electrodes. The SEM image is s hown in figure 4.11. The EDS spectrum is shown in figure 4.12. Figure 4.11 SEM Image of Si Electrode with RuO 2 Deposition 45

PAGE 58

EDS Spectrum0 500000 1000000 1500000 2000000 2500000 012345678Energy (kEv)Intensity (Counts)C Ka Ru Lb Si Ka Ru La Ru Ll Ru Lg O Ka Figure 4.12 EDS Spectrum RuO 2 on Si The EDS map of Ru over laid on the EDS map of O 2 is shown in Figure 4.13 again with Ru depicted in yellow and O 2 depicted in green. Figure 4.13 EDS Map of Ru on Si Electrode with EDS Map of O 2 on Si Overlay 46

PAGE 59

It can be seen that the coverage of Ru is not consistent. O 2 appears more consistent due to the native oxide on the Si. Further characterization of the RuO 2 deposition on Si was carried out using the NNRCs Panalytical X'Pert Diffractometer. A picture and specifications for the Diffractometer can be found in appendix B. A very brief description adapted from the Materials Analytical Services Inc website follows [26]. X-ray Diffraction (XRD) is a powerful tool for characterizing crystalline materials. It is nondestructive to the sample. XRD provi des information on structure, crystal orientation, crysta llinity, etc. X-ray diffr action peaks are produced by constructive interference of a monochromatic beam scattered from each set of lattice planes at specific angles. The X-ray diffrac tion pattern is a fingerprint of a specific material. When an X-ray pattern is taken of a sample the pattern can be compared to a data base to determine the makeup of crystals in the sample. The results of the XRD of the RuO 2 deposition on Si after multiple coats are shown in figure 4.14. Multiple layers show improved peak matching due the thicker coat. The vertical lines in figure 4.14 show the location of the literature values for the RuO 2 [110] and [101] peaks. 0 50 100 150 200 250 300 20 25 30 352 Theta (degrees)Intensity (a.u.) RuO2on Si 2 coats RuO2on Si 3 coats RuO2 [110] peak RuO2 [101] peak Figure 4.14 XRD Spectra for Sol-Gel RuO2 Deposition on Si 47

PAGE 60

4.4.4 Resistance Testing The resistance was tested with an Agilent LCR meter. A jig was set up to maintain consistent pressure, see figure 4.15. The jig holds the wafer between two copper blocks. This setup was used instead of a four point probe system because multiple layers were going to be used on most materials. A four point probe system measures sheet resistance and is correlated to bulk resistance. It was determined with multiple layers this would not be a good approach. Figure 4.15 Resistance Test Jig 4.4.5 Experimental Results and Discussion Tests of RuO 2 deposition adhesion showed excelle nt adhesion. The wafers were scoured under DI water and Isopropanol w ith a neoprene glove. Adhesion was also tested with a tape test, where a piece of tape is placed on the wafer and removed to see if the deposited material will stick to the tape or the substrate. Long term tests of wafers coated in RuO 2 and placed in sulfuric acid show no visual change and no change in resistance. Results of resistance testing of wafers prepared under air and nitrogen rich atmospheres are shown in figure 4.16. 48

PAGE 61

Deposits made by the Hummer X sputter system passed the tape test. It was noticed during setup for electrolysis the Au Pd deposition was scraped off in places. As stated earlier the AuPd was attacked by sulfuric acid and peeled off the wafer. Results of resistance test s are shown in figure 4.16. Deposits made by the CRC 100 sputter system showed excellent adhesion. The depositions passed the tape test and could not be scraped off. The wafers with Pt and Ag deposited with the CRC 100 sputter system showed the best results in resistances tests. Results are displayed in figure 4.16. Wafer Deposition Resistance 0 0.5 1 1.5 2 2.5 N2AirEtchNo Etchn-Sip-Si n-Si RuO2 n-Si RuO2 Hummer X Hummer X CRC 100 CRC 100 Resistance (ohms) Wafer Deposition Resistance Figure 4.16 Resistance Test Results 4.5 316 SS Electrodes 316 Stainless Steel electrode s were chosen early on because they were shown to hold up to sulfuric acid. They did not hold up when 316 SS was used as an anode. Considerable corrosion occurred with 316 SS a nodes. Gold electropl ating turned out to be a good seed layer for platinum depositions on cathodes, so it was decided to try gold plating of the anodes to inhibit corrosion. 4.5.1 Gold Plating Acid Gold Strike by Technic Inc. was used as a seed layer for gold depositions also. Acid Gold Strike can only be used as a seed layer. 49

PAGE 62

50 It is far too porous to be used as a protect ive coating and can only be applied at very limited thickness. Techni-Gold 25 E by Technic Inc. was plated on top of the seed layer to protect the 316 SS substrate. 316 SS was cleaned as previously reporte d by Chettiar [12]. The 316 SS was first cleaned with a detergent, Mean Gr een Industrial Strength cleaner. The electrodes were then soaked in 20 % NaOH-H 2 O for approximately 10 minutes, rinsed with DI water and blow dried. They were then rinsed with 20 % HClH 2 O, rinsed with DI water and blow dried. Acid Gold Strike was first deposited. A carbon anode was used. The Acid Gold Strike solution temperature was maintained at approximately 90F. The solution was stirred at medium speed. A current of 0.25 amps was maintained by varying voltage from 1.5 volts to 2 volts for approximately 2 minutes. Techni-Gold 25-E was deposited immediat ely after Acid Gold Strike. A new carbon anode was used. The solution temper ature was maintained at approximately 110F. The solution was stirred at medi um speed. A current of 0.04 amps was maintained by varying voltage from 0.46 volts 0.6 volts for 30 minutes for cathodes and 90 minutes for anodes. 4.5.2 Catalyst Deposition The thin layers of RuO 2 deposits on silicon led to further investigation of literature concerning sol-gel RuO 2 techniques. Upon further r eading it was realized that most research was done with ru thenium chloride hydrate, RuCl 3 -xH 2 O dissolved in isopropanol as the precursor not RuCl 3 [27] [28]. RuCl 3 -xH 2 O was ordered from Sigma Aldrich. Another source discussed dissolving RuCl 3 in 20 % HCl and drying prior to dissolving in isopropanol [29]. This techni que was attempted and the solution showed better solubility of the solid. The first step in the new technique fo r preparing sol-gel solution was to grind 0.4355 g RuCl 3 using a mortar and pestle. The RuCl 3 was then dissolved in 20 % HCl. The solvent was boiled off and the RuCl 3 was dried in an oven.

PAGE 63

51 The solid is then dissolved in isopropanol and shows better solubility than non pretreated RuCl 3 The 0.121 M solution is still heter ogeneous but much less solid settles out of solution. The same steps for calcining the RuCl 3 as used with silicon were followed. The formed ruthenium oxide showed good adhesion. The deposits held up to cleaning in DI water and isopropanol. They also passed the tape tests. The new technique did not appear to improve the thickne ss of the layers. Th e catalytic affect of the deposits will be discussed in the next chapter on the two inch cell. The final deposition technique used was sputter coating Pt on the gold plated electrodes in the CRC 100 sputter system. Pt had been successfully deposited on new electrodes and used electrodes that had been cleaned. The used electrode were first cleaned in toluene, dried and rinsed with DI water. The electrodes were then soaked in 20% NaOH for approximately 20 minutes followed by a DI water rinse and dry. The electrodes were finally rinsed with 20 % HCl, DI water and then dried. Once again a layer of approximately 100200 angstroms of Pt was deposited. The deposition showed very good adhesion. It pa ssed the tape test and could hold up to cleaning in DI and solvents. 4.5.3 Electrode Characterization Characterization of the RuO 2 deposition on Au on 316 SS was carried out using the NNRCs Panalytical X'Pert Diffractometer The results of the XRD of the RuO 2 deposition on Au on 316 SS are shown in figur e 4.17. The vertical lines in figure 4.17 show the location of the literature values for the RuO 2 [110] and [101] peaks.

PAGE 64

-50 0 50 100 150 200 250 300 202224262830323436 2 Theta (degrees)Intensity (a.u.) RuO2 [110] peak RuO2 [101] peak RuO2 on Au on 316 SS Figure 4.17 XRD Spectra of RuO 2 Deposited on Au on 316 SS 52

PAGE 65

53 CHAPTER 5 TWO-INCH ELECTRODE CELLS 5.1 Overview The deposition techniques from the prev ious chapter had a large part in the decision to develop the two inch electrode cell. Some of the equipment available to our group for deposition had size limitations. Anothe r determining factor was the availability of supplies. The housing of the two inch cel l was polycarbonate. Polycarbonate has only a fair rating with sulfuric aci d but it is much easier to pick up locally and can take the place of both Teflon and glass because it is clear The gasket material was changed from Viton to EPDM also because of local availabi lity. Other design issues will be discussed in the individual sections where they were implemented either the liquid electrolyte cell or the solid electrolyte cell. 5.2 Liquid Electrolyte Cell s The two inch electrode liquid electroly te cells were designed with knowledge gained from operating the 4 inch electrode ce ll. They were developed to try a benefit from the lower conductivity of aqueous sulfur ic acid than Nafion. The liquid electrolyte cells were designed with one main similarity to the four inch diameter electrode cell. They were built to facilitate an easy conversi on to a solar chemical cell. They were built so the electrodes could easily be replaced by a 2 inch silicon wafer designed as a solar cell. The first liquid electrolyte cell was r un at ambient temperature and pressure. The second liquid electrolyte cell wa s run at ambient temperature and approximately 90 psig. 5.2.1 Design Considerations The cell was built with limited volume to for ce the reactants against the electrode, especially in the anode compartment.

PAGE 66

The anode compartment was packed with carbon fiber to create more surface area and further reduce cell volume. See appe ndix C for 2 dimensional drawings. The electrolyte bridge was placed to give sulfur dioxide entering the anode com partment the longest path length possible to the cathode compartment, see figure 5.1. Due to the possibility that all the sulfur di oxide might not be elec trolyzed, the outlet was placed lower than the electrolyte bridge and a flow gradient was maintained from the cathode compartment to the anode compartment to keep sulfur dioxide from entering the cathode compartment and sweep unreacted sulfur dioxide out with the product acid. Figure 5.1 Two-Inch Low Pre ssure Liquid Electrolyte C ell In the low pressure cell the electrodes were m ade to extend beyond the bottom of the cell so electrical connecti on can be made outside of th e cell so corrosion is not an issue. The gold plated 316 SS electrodes are placed back to back and insulated from one another and sealed with apiez on grease. When silicon electr odes are used copper current collectors are pressed to the back side of each electrode and extended outside of the cell. 54

PAGE 67

The current collectors are insulated from each other with a thin sheet of Teflon. The backside of the electrodes and current collect ors are sealed from sulfuric acid by EPDM rubber. The 80 psig cell needed some design changes to prevent leaks. The entire cell body and gaskets were extended to from 4 X 4 to 4 X 6 to facilitate extra bolts at the bottom and top of the cell. See appendix C fo r 2 dimensional drawings of each section of both cells. The electrodes were expanded from 2 X 4 to 4 X 4 to allow for more gasket and electrode contact ar ea to stop leaks. The 80 psig cell can be seen in figure 5.2. Figure 5.2 Two-Inch 80 psig Liquid Electrolyte Cell 55

PAGE 68

56 5.2.2 Flow System/Operation The flow is considerably less complex for the two inch liquid electrolyte cell than the 4 inch diameter electrode cell, see figur e 5.3. A big part of lowering the complexity is the use of gaseous SO 2 to saturate the acid entering the anode compartment instead of pumping liquid SO 2 directly into the cell. This makes it possible to operate at a lower pressure. Previously the cell was oper ated at 100 psi to assure the liquid SO 2 did not flash from rapid pressure drop and expand rapidl y to vapor causing valves or fittings to freeze up. This low pressure operation was later changed due to some minor separation of dissolved SO 2 to form gas bubbles that in creased the likelihood of SO 2 migration to the cathode compartment. Difficulties also arose from the low pressure SO 2 coming out of solution before the pump causing the pump to lose prime and stop pumping. The 30% sulfuric acid was saturated with SO 2 prior to operation by bubbling SO 2 into a container of sulfuric ac id. Several inches of water column pressure are maintained by allowing enough gas pressure to push several inches of sulfuric acid up the vent tube. Saturation is assured when several inches of water column pressure is maintained for several hours, because if pressure dropped SO 2 was being absorbed. The vent tube was closed in the 80 psig experiments to maintain a saturation cell pressu re equal to the vapor pressure of SO 2 approximately 35 psig. Saturation is assured when the system equalizes at the vapor pressure of SO 2 and the SO 2 gas feed no longer bubbles into the pressurized cell. Positive pressure is maintain ed during electrolysis by slowly bubbling SO 2 into solution through the micro metering valve V-1 while the SO 2 pump is pumping. After saturation is assured the electrolyzer is placed in the system. To fill the electrolyzer with electrolyte valves V-2 and V-3 are opened and the catholyte pump is turned on to approximately 1 ml/min. Once the electrolyzer is full, sulfuric acid will come out into the product acid tank. During 80 psig operation a 80 ps ig back pressure regulator is placed in line at the product acid tank to maintain pressure in the cell. The catholyte pump can be turned back to the operating pump rate of 0.03 ml/min. The gas pocket at the top of the cathode compartment is maintained by manipulating valve V-8 and the catholyte pump rate if needed.

PAGE 69

Figure 5.3 PFD for Two-Inch Liquid Electrolyte System Once the electolyzer is full, voltage can be applied to the electrodes. The same control and data acquisition system are used as was described in section 4.2.3. Once a potential difference is across the electrodes SO 2 can be introduced to the anode compartment. Valves V-4 and V-5 are turned on and the SO 2 pump is turned on to 0.01 ml/min. Upon completion of the elec trolysis run the system can be switched to water for rinsing the system. It is best to maintain a potential difference across the electrode when the system is being cleaned. It provides extra time to oxidize SO 2 in the system. The first step towards shutdown is to close valv e V-4 and open valve V-6. This allows water to enter the anode chamber. After considerable drop in amperage it is ok to rinse the entire system. To do this valve V-2 is closed and valve V-7 is opened. After amperage is nearly nonexistent the control system can be shut down. The pump rate of both pumps can be increased to 1 ml/min each, to decrease rinsing time. Typically 250 ml is pumped through each pump to assure the pumps and the system are clean. 57

PAGE 70

58 5.3 Solid Electrolyte Cell Work was started with solid polymer elec trolyte because it has been reported that solid polymer cells were successfully used in the past [13] [19]. Processes have been borrowed from the manufacture of typical PEM electrolyzers used for electrolyzing water. Certain design changes had to be made to make PEM style electrolyzers work. 5.3.1 Design Considerations The biggest issue with using solid poly mer electrolyte is the diffusion of SO 2 through the electrolyte. A key in determin ing how conductive a solid polymer will be is the amount of water up take in the matrix of the polymer. This causes problems with sulfur dioxide because sulfur dioxide is very soluble in water and diffuses very quickly through water. This ability for rapid diffusion is defeated in the liq uid electrolyte cell with a flow gradient against the diffusion. Convective transport of sulfuric acid in the liquid electrolyte cell is able to overcome the diffusive transport of sulfur dioxide. Nafion is not porous. It hol ds its water trapped in the matrix. There is no flow. Therefore this was overcome by the addition of a third cell. Sulfur di oxide is allowed to diffuse through the first layer of Nafion wher e it is met by a perpendicular flow through the center compartment rapid enough to convectively sweep away the SO 2 before it can diffuse through that central compartment, see figure 5.4. This can be accomplished because the rate of diffusion of hydrogen ions in sulfuric acid is much greater than the rate of diffusion of sulfur dioxide. Hydrogen Ions diffuse through the center compartment rather than being swept away like SO 2 Electrodes for PEM electrolyzers are considerably different. The porous electrodes along with catalysts are mechanically pressed and stuck to the membrane creating one cohesive unit ca lled a membrane electrode a ssembly (MEA). MEA design and manufacture will be discu ssed in the next section. With the actual electrode attached to the membrane this leaves what in the liquid electrolyte cell is considered an electrode to be a current collector, a means to pass current (electrons) to the actual electrode.

PAGE 71

The anode current collector has two hol es, see figure 5.4. The bottom one allows the reactants to enter the small anode chamber packed with carbon the makes the electrical contact to the electrode that is pressed to the solid electrolyte. The reactants are forced by flow to the reactive interface where the electrode and catalyst contact the solid electrolyte. The top hole in the current collector allows the product acid and any unreacted material to exit the cell. The cathode current collector has severa l holes to allow hydr ogen, created at the interface of the solid el ectrolyte and porous electrode, to escape. Some experiments the cathode compartment from a water electrolysi s cell designed by Heliocentris was used. The borrowed parts included the hous ing, current collector, and MEA. Figure 5.4 Three Compartment Solid Electrolyte Cell The center compartment is simply two pieces of EPDM with capillary tubes sealed in packed with glass fiber. 59

PAGE 72

The EPDM center compartment serves as the seal to seal each MEA to its respective current collector. Sulfuric acid is pa ssed through the capillary tubes creating the perpendicular convective flow described. 5.3.2 MEA Development MEAs were produced similar to what is ty pically discussed in lit erature [30] [31]. Catalyst inks which in this case are a combination of Liquion LQ-1005 Nafion solution from Ion Power, Inc and Rutheniu m oxide 99.9% pure from Sigma Aldrich, 99% pure 10 micron tungsten carbide also from Sigm a Aldrich or 5% Pt on activated carbon are air brushed or painted on 1.5 Teflon bla nks. The inks are allowed to dry on the Teflon blanks. Typically the i nks are hot pressed on both sides of a piece of Nafion. In this case only one side is hot pressed because one side is a reactive surface and the other is the center compartment wher e no reaction takes place. Nafion membrane N112 is cleaned by hour in boiling and stirred 3% hydrogen peroxide, hour in boiling and st irred DI water and hour in boiling and stirred 0.5M sulfuric acid. The membrane is dried in a 200 C oven for two minutes. The Nafion membrane and coated Teflon blank are pressed between two preheated 2 diameter copper blocks prot ected by 2.5 diameter Teflon. The copper blocks are preheated to 200 C. The system is stacked as per figure 5.5 and hot pressed at 1000 psi for 2 minutes. Figure 5.5 Membrane Hot Press Diagram 60

PAGE 73

Upon cooling the system is disassembled. The result of this process is the catalyst and electrode material on the Teflon bla nk are transferred to the Nafion. 5.3.3 Flow System/Operation The process flow diagram for the 3 compartment system is depicted in figure 5.6. The saturation system works the same for the three compartment system as it does for the 2 inch liquid electrolyte system. The cathode compartment in the three compartment system no longer has flow. The first step after saturation is to fill the cathode compartment. The cathode compartment is f illed with a syringe. Th e second step is to open valves V-2 and V-3 and run the center compartment pump at 1ml/min. Figure 5.6 PFD for 3 Compartment Cell Once flow is running through the center co mpartment the voltage can be applied to the electrodes. After there is voltage to the electrodes valves V-4 and V-5 can be opened and the SO 2 pump started at 0.01 ml/min. Upon completion of the elec trolysis run the system can be switched to water for rinsing the system. It is best to maintain a potential difference across the electrode when the system is being cleaned. It provides extra time to oxidize SO 2 in the system. 61

PAGE 74

The first step towards shutdown is to close valve V-4 and open valve V-6. This allows water to enter the anode chamber. After considerable drop in amperage it is ok to rinse the entire system. To do this valve V-2 is closed and valve V-7 is opened. After amperage is nearly nonexistent the control system can be shut down. The pump rate of both pumps can be increased to 1 ml/min each, to decrease rinsing time. Typi cally 250 ml is pumped through each pump to assure the pumps and the system are clean. 5.3.4 Monitoring Sulfur Dioxide Diffusion It is important that sulfur dioxide does not diffuse to the cathode compartment. Sulfur is known to poison cathode catalysts especially when Pt is used. Sulfur is one of the possible products when SO 2 is reduced. Another possible product is H 2 S. H 2 S is a deadly gas. It is not desired to have such a dangerous in the product. The center compartment acid and the cathode compartment acid are closely monitored in this three compartment se tup. Samples are taken from the center compartment and cathode compartment during op eration. The samples are placed in a 1 mm path length cuvette and absorbance is measured in an Ocean Optics UV-VIS system. This system is used because SO 2 has high absorbance in this li ght range and sulfuric acid and water do not. 5.4 Goals and Limitations The ultimate goal of this research is to solar chemical hydrogen production. This sets a limitation for the research. The max solar insulation is 1000 watts/m 2 If we rather grandly assume 10 % efficien cy this becomes 100 watts/m 2 of available power. Conversion to cm 2 gives .01 watts/cm 2 We now consider the equation for power as seen in equation 23. IVP (23) P in this case is the power density in watts/cm 2 V is the voltage in volts and I in this case is the current density in amps/cm 2 62

PAGE 75

The target voltage for a single junction s ilicon solar cell is 0.5 volts and for and CdTe cell is 0.7 volts. Plugging these number s and the maximum possible power density into equation 23 leaves a max current density of 0.014-.02 A/cm 2 Another limitation to the max curre nt density is the choice to use SO 2 saturated H 2 SO 4 at ambient pressure to de liver the reactants to the ce ll. The amount of reaction governs the current density in an electrochemi cal cell. The amount of reactants can limit the amount of reaction. Therefore we need to know how much limiting reactant is getting to the cell at the experimental flow rate. The acid concentration used for maximum conductivity is 30%. Solubility data for SO 2 in 30% H 2 SO 4 was reported by Hayduk [29]. The solubility of SO 2 in 30% H 2 SO 4 is approximately 0.105 g/g. A current dens ity at the operating flow rate can now be calculated assuming negligible volume ch ange of the solvent when saturated. The operating flow rate for the anode compartment in the experiments is 0.01 ml/min. The density of 30 % H 2 SO 4 is 1.219 g/ml. Therefore the mass flow rate is 0.0129 g/min. This makes the mass flow rate of SO 2 in the approximately 0.00136 g/min. These flow rates now have to be converted to moles/sec so Faradays law can be used to determine the current. The molar flow rate is 3.53 x 10 -7 mol/sec. Faradays law is given in equation 24. F z I n ( In thi 24) s case because SO2 and H2 are equamolar n is the moles of SO2. The letter I is the f .5 Experimental Results and Discussion of Preliminary Experiments r the diffusion of SO2 to the cathode compartment. That design goal was achieved. current given in amps which is also known as coulombs/sec. The letter z is the number o moles of electrons per mole product, in this case 2. Finally F is the number of coulombs per mole of electrons or approximately 96500. Equation 24 can be rearranged and solved for the current. The total current is 67.5 mA. The current divided by our electrode area gives the current density. The current density max limit due to SO 2 is approximately 5.0 mA/cm 2 at the experimental flow rate. 5 A main concern of both designs for two inch electrodes was to dete 63

PAGE 76

64 any evidence that The first experiment was run using the liquid electrolyte cell as depicted in figure gold plated 316 SS. The anode was gold plated 316 SS with RuO2 efore and the RuO2 was a little depleted. After ent 2 The second experiment followed immediately after the first. The same d except the anode compartment was pack ed with carbon fiber. As than y is at both the cathode The third experiment was also done in the liquid electrolyte cell and was set up as .2.1. The experiments discussed in this section had run times varying from 30 minutes to 5 hours with the average being approximately 2 hours. There was never SO 2 was in the cathode compartment. 5.5.1 Experiment 1 5.1. The cathode was deposited as describe in section 4.5.2. RuO 2 was not used on the cathode because tests performed in beakers showed poor adhesion of RuO 2 after electrolysis despite seeming mechanically sound before electrolysis. As figure 5.7 shows this is the poorest results of any of th e liquid electrolyte experiments. The anode had been used b experiment 2 it was determined the RuO 2 adhesion was suspect on the anode after electrolysis. 5.5.2 Experim configuration was use can be seen in figure 5.7 this greatly improve d the current density. This is more likely due to the extra surface rather than increased flow rate through lower volume compartment. This has been determined b ecause increased flow velocity through the same volume does not show as much benefit as is produced here. The current densit increase by greater than a factor of 5 through use of the anode carbon. Upon disassembly of the cell it wa s noted that more of the RuO 2 has come off of the electrode. Despite good adhesion prior to electrolysis, it appears th and anode RuO 2 deposits are adversely affected by electrolysis. 5.5.3 Experiment 3 described in section 5

PAGE 77

65 en in figure 5.7. It showed one of the best results at higher The fourth experiment was set up the sa me as the third except the electrode with liquid EPDM and allo wed to cure for several days. It showed the mos ising For the fifth experiment it was decide d to get a baseline experiment with no other li quid electrolyte experiment. It was again set up as figure 5.1 ith t The sixth experiment is the first solid electrolyte three compartment cell the onl y one presented that is assemble d exactly as is depicted in d as This experiment used Pt on an n-Si wafer for the cathode and Pt on a p-Si wafer for the anode. The results can be se voltage. There were some minor adhesion issues with the Pt on Si after electrolysis. It is believed they were caused by the sealing grease used. It was also apparent upon disassembly that the sealant between the electrodes was not sufficient and the Ag backside contacts were affect ed by the sulfuric acid. 5.5.4 Experiment 4 assembly was sealed t promising result at lower voltage as can be seen in figure 5.7. This is prom for the long term research because ultimately Pt on silicon will serve as the Schottky barrier and cathodic catalyst in the solar chemical cell. A comparison can not be made to experiment 3 because that experime nt was only done at high voltage. 5.5.5 Experiment 5 catalyst. This was an depicts. This was the longest run. It was the best proof that the diffusion problem w SO 2 has been solved. As was expected this setup shows the lowest current density of the liquid electrolyte cells so far excluding the firs t experiment which is the only run with ou carbon packed in the anode compartment. 5.5.6 Experiment 6 experiment. It is also figure 5.4. The MEAs both have 5% Pt on ac tivated carbon electrodes hot presse described in section 5.3.2. This was the longest 3 compartment run at 2.5 hours. It showed no diffusion of SO 2 through the second layer of membrane. This is the only experiment that showed a benefit from increasing the flow rate.

PAGE 78

66 a hen the flow rate was The flow rate to the anode was increased by a factor of 8 which increased the current by factor of 7. The other experim ents all showed little to no effect w increased. Preliminary I vs V Curves0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3Voltage (V)Current Density (mA/cm2) 1 RuO2 2 RuO2 Fiber 3 Pt-Si 4 Pt-Si 5 Au 6 3 Comp 7 Pt-Au 8 Pt-Au Figure 5.7 Preliminary Two-Inch Cell Comparison Experiments 7-9 have been lumped togeth er because there is relatively small All three experiments utilize gold plate 316 SS electrodes with n is is 5.5.7 Experiments 7-9 changes between them. 100-200 angstroms Pt deposited on them as described in section 4.5.2. The cell desig slightly changed between experim ents 7 and 8 from that depicted in figure 5.1. A spacer was added to enlarge the cathode compartment because excessive bubble growth in experiment covered the cathode and blocke d part of the electrolyte bridge. It is reasoned that the cu rrent density drop from experi ment 7 to experiment 8 cause by a contaminated and depleted Pt surface.

PAGE 79

67 At the e d SO2 diffused into the cathode nt 7 roughly cleaned and a new de Once again experiments are lumped t ogether for discussion due to their were run from a hybrid cell that was a combination of a m section nd of experiment 7 valves were manipulated to try to improve hydrogen bubble evolution. This changed the flow in the system an compartment. The SO 2 was reduced and sulfur was de posited on the cathode. In an attempt to clean the sulfur some of the catalyst was removed. Experiment 9 seems to confirm the assumptions about changes from experime to experiment 8. The cathode from expe riments 8 and 9 was t ho position of Pt was done. This in combination with the larger cathode compartment gave the best curren t density results of this work. 5.5.8 Experiments 10 and 11 similarities. These experiments commercial water electrolyzer designed by Heli ocentris and design concepts fro 5.3.1, see figure 5.8. The commercial cathode compartment and catalyst were used in conjunction with a center compartment and anode similar to the ones depicted in figure 5.1. Figure 5.8 Hybrid 3 Compartment Cell As can be seen in figure 5.7 slightly better curr ent densities were reco rded but at a higher l experiment. voltage than for the other thre e compartment cel

PAGE 80

68 liminary experiments that stringent enough scientific method was not followed. Even with this decision the eriments to be set up that would allow m r ch ese results that no more three compar time. ssf ul deposition technique at this time. It ethod. The experiments were run for much longer times and were not stopped unless hode compartment. de compartment was when the flow system was disrupt e 5.5.9 Conclusions from Preliminary Experiments It was decided from the results and discus sion of the pre a preliminary results did allow for a narrower se t of exp ore ease to follow a rigid methodology. All three experiments with the three co mpartment cell resulted in much lowe current densities than the liqui d electrolyte cell excluding the experiment 1 that had mu less anode surface area. It was determined from th tment cell experiments would be run. As discussed in sections 5.5.1 a nd 5.5.2 there is an issue with RuO 2 adhesion during electrolysis. It was d ecided to not proceed with RuO 2 experiments at this Sputtered Pt seems to be the only succe was determined the rest of the experiments would be run with Pt as the cathodic and anodic catalyst. The Pt on gold plated 316 SS are the sturdiest electrodes and easiest to seal. It was decided to run the majority of the experiments with the 316 SS electrodes until an optimum was determined then change to Pt on Si to make a final comparison. 5.6 Experimental Results and Disc ussion of Final Experiments The final experiments were run follow ing a much more rigorous scientific m some mishap with the flow system allowed SO 2 to diffuse to the cat The experiments were run until the amperage from the cell leveled off for several minutes. This was done because when voltage is applied or raised the current typically spikes because the system is thrown out of e quilibrium. Therefore, time must be allowed for the system to return to equilibrium. These experiments again proved, barring mishaps, SO 2 diffusion has been eliminated. These experiments ran for 12-50 hours. The only time sulfur diffused to the catho ed or if anode packing material shif ted allowing a straight path to the cathod compartment for SO 2

PAGE 81

69 run at atmospheric pressure and the H2SO4 was saturated at only a few in SO2, to the pr eliminary experiments. It was run at ssure. The flow rate to the anode was 0.1 ml/min. The flow rate to the in. It was run with the ce ll depicted in figur e 5.1. The electrodes de. ue to the anode compartment. It was assumed the bubble was SO2 because bubbles could n. The raise pressure caused leaks so larger electr odes were designed to a ssure more surface area ts in hope to stop leaks. Also clamps were added at the top and bottom of the cell where there were no bolts. The only drastic change was a decision to r un at higher pressure. This allows for a higher concentration of SO 2 but the results still show explainable trends. The first experiment was ches of water column pressure. Th e remainder of the experiments were run at approximately 70 psig and the H 2 SO 4 was saturated at the vapor pressure of approximately 35 psig. This should cause approximately a 2.4 times increase in SO 2 solubility according to literature [32]. 5.6.1 Experiment 1 Experiment one was run similar atmospheric pre cathode was 0.3 ml/m were 2 x 4 Pt on gold plated 316 SS electrodes. There was one anode and one catho The anode compartment was packed with 5 pi eces of carbon fiber th at was gold plated with orostrike plating solution. The fibers also had a 100-200 angstrom coat of Pt for catalytic effect. This run was considered the baseline run for comparison to the remainder of the final experiments. The result s of this experiment can be seen in figure 5.9. Difficulty was had maintaining prime in the SO 2 saturated H 2 SO 4 pump d bubble formation in the inlet line to the pum p. An occasional bubble was noticed coming from also be seen coming out of the back pressure regulator in the inlet line to the cell. For this reason the remainder of the experiments were run at approximately 80 psig and the H 2 SO 4 was saturated at the vapor pressure of SO 2 approximately 35 psig. 5.6.2 Experiment 2 Experiment two required the changes explai ned in the previous sectio in for contact with gaske

PAGE 82

70 depicted by figure 5.1. le e compartment was packed ith the Pt and Au coated carbon fiber and the anode was packed with Au coated fiber. for this run was similar to the previous runs as seen in figure 5.9. s ach eactant diffusion limited. This led to a e re estriction te bridge caused by the narrowness of the anode compartment. The active area of the electrode was not changed because the extra surface area was covered by gasket material. With the he lp of some clamps the leaks were stopped. The cell for this run was actually a comb ination of figure 5.1 and 5.2. The large electrodes from figure 5.2 were used in the cell The planned change for this run was to double the surface of the anode by adding a second anode. The same 5 pieces of Pt and Au coated carbon fiber were used to pack the compartment between the electrodes. Figure 5.9 and figure 5.10 show that litt benefit was gained from doub ling the anode surface area. 5.6.3 Experiment 3 Experiment 3 was run with the same cell design described for experiment 2. Instead of two anodes two cathodes were us ed. The cathod w Total current Figure 5.10 shows current density calculat ed for hydrogen production. There is a significant drop in current density because th e surface area of the cathode doubled. Thi shows the anode reaction is limiting the current. Manipulating the anode surface area in e xperiment 2 showed little change in current density. This shows the system is not reaction rate limited. At the end of e experiment the flow rates were manipulated and showed only minor changes in current density. This seems to suggest the system is not r decision to manipulate the elec trolyte bridge to increase i on diffusion. The electrolyt bridge is restricted by the na rrowness of the anode compartment. This was done to assu shorter diffusion path for reactant s and restrict diffusion of SO 2 to the cathode compartment. 5.6.4 Experiment 4 As stated at the end of the last sectio n it was decided to manipulate the r in the electroly

PAGE 83

71 s experiment by changing the thickness of the anode gasket. The This was done for thi anode gasket was changed from approxi m ately 0.04 inches to 0.125 inches. Current vs Voltage 0 10 20 30 40 50 60 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2Voltage (V)Current Density (mA/cm2) 14.69 psi 0.04" 1 Anode 1 Cathode 80 psi 0.04" 2 Anodes 1 Cathode 80 psi 0.04" 1 Anode 2 Cathodes 80 psi 0.125" 1 Anode 1 Cathode Figure 5.9 Total Current vs. Voltage Curves: Experiments 1 Through 4 Current Density vs Voltage 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 05 0.6 0.7 08 0.9 1 1.1 1.2Voltage (V)Current (mA) 14.69 psi 0.04" 1 Anode 1 Cathode 80 psi 0.04" 2 Anodes 1 Cathode 80 psi 0.04" 1 Anode 2 Cathodes 80 psi 0.125" 1 Anode 1 Cathode Figure 5.10 Current Density vs Voltage Curves: Experiments 1 Through 4

PAGE 84

72 Along with the change in anode gasket ch anges in cell design were implemented because one of the end plates was cracked dur ing the last run. Thicker polycarbonate was used and the height of the end plates and gaskets were increased so the clamps in experiments 2 and 3 could be replaced by bolts The same electrodes as experiments 2 and 3 were used. This cell is depicted in figure 5.2. The added thickness of the anode compartm ent required additional carbon fiber to be placed in the anode compartment to restrict SO2 diffusion. Because the thickness of the anode compartment was tripled the flow from the cathode compartment was tripled from 0.3 ml per minute to 0.9 ml/min to maintain the same cross sectional flow to inhibit SO2 diffusion. For the purpose of limited change no more Pt was added to the carbon fibers. The fibers were gold plated to improve conductivity. The thicker anode compartment showed improved current density, see figure 5.9. This seemed to confirm the assumption that the narrowness of the electr olyte bridge restricti ng ion diffusion is the ent 4 showed that the increased flow fr e seen in rate determining step. It was decided to enlarge the anode compartment again for the next experiment. The results from experiment three can also be seen as the base line for figure 5.10. 5.6.5 Experiment 5 Manipulations of flow rates at the end experim om the cathode was probably not nee d, but for consistency and safety from SO 2 diffusion 0.9 ml/min will be maintained for the remainder of the experiments but further increases were not deemed necessary. Th e anode compartment was enlarged from experiment 4 by changing from the 0.125 inch ga sket to a 0.25 inch gasket. Additional carbon fiber was gold plated and added to the anode compartment. The larger anode compartment again proved beneficial. The results can b figure 5.10. The restriction from ion diffusion seemed to be lessoned so it was decided to again try to manipulate the surface area.

PAGE 85

73 .6.6 Experiment 6 de was added to the outside of the anode compartment for ensity calculated for hydrogen production ctually went down as shown in figure 5.12. Th is confirms the findings of experiment 6 at the anode reaction seems to be the limiti ng reaction. Another limiting factor for this un is the fact that for the first time all the SO2 being pumped in to the cell is being used ge. ue to e to elying more de. Once all xcess SO2 is used up the current will eventually drop to 67.5 if the solubility data is current distribution because no SO2 will make it to coated car bon was placed in the cathode compartment because it is a better hydrogen catalyst than sulfuric acid catalyst. 5 A second ano experiment 6. The same carbon fiber packi ng material from experiment 5 was used except one Pt coated piece was lost and replaced with a gold plated piece. This time increasing the anode surface area showed improvement as can be seen in figure 5.10. This shows that sulfuric acid pr oduction is the rate limiting reaction. 5.6.7 Experiment 7 A second cathode was added to the out side of the cathode compartment for experiment 7. The rest of the setup is the sa me as experiment 6. The total current went up as seen in figure 5.11, but the current d a th r up at the highest volta The current of 71.18 is higher than the curr ent possible at the flow rate d solubility limits because the c oncentration is higher than what is being pumped in du unreacted SO 2 from lower voltages and currents. Th is means the system is r on SO 2 diffusion than the convective flow from the input pump rate. This will most likely cause uneven current distribution on the electro e correct. This will cause very uneven the top of the electrode because it will all be reacted before it gets there. 5.6.8 Experiment 8 It was decided to try packing the cathode compartment with gold plated carbon to increase the linear flow through the compartm ent to hopefully sweep sulfate and bisulfate ions and hydrogen away from the cathode and hopefully help hydrogen ion diffusion to the cathode surface. The Pt

PAGE 86

74 ty Total current was slightly lower, see figure 5.11, probably due to more difficul with hydrogen evolution due to the carbon packed compartment. The same restriction at 1.2 volts from higher amperage and co mplete electrolysis of input SO 2 as experiment 7 will hold. Current vs Voltage 0 10 20 0.5 0.6 0.7 0.8 0.9 1 1.1Voltage (V)C 30 50 60 1.2urreA) 40nt (m 70 80 0.125" 1 Anode 0.25" 1 Anode 0.25" 2 A nodes 0.25" 2 Anodes 2 Cathodes 0.25" 2 Anodes 2 Cathodes-Fiber Figure 5.11 Total Current vs Voltage Curves: Experiments 4 Through 8 Current Density vs Voltage0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 50 60 70 80 911 11Voltage (V)Current Density (mA/cm2) 2 0.125" 1 Anode 0.25" 1 Anode 0.25" 2 Anodes 0.25" 2 Anodes 2 Cathodes 0.25" 2 Anodes 2 Cathodes-Fiber Figure 5.12 Current Density vs Voltage Curves: Experiments 4 Through 8

PAGE 87

75 5.6.9 Experiment 9 Experiment 9 was run to give a base line for the cell with out the use of SO2 to lower the voltage requirement for hydrogen pr oduction. This experiment was run with the same setup as experiment 6. The result s were as expected w ith very low current density because the cell was operated at the same voltages which are too low for direct dissociation of water. The re sults can be seen in figure 5.13. 5.6.10 Experiment 10 The anode compartment was expanded from 0.25 inches to 0.58 inches for this experiment by changing the 0.25 inch gasket to two 0.04 inch gaskets, one on either side of a 0.5 inch piece of polycarbona te. Another piece of Pt co ated carbon fiber was lost, so eets and 70 u coated carbon fiber sheets. The highest curr ents to date resulted and are compared to experiments 6, 9 and 11 in figure 5.13 and current densities in figure 5.14. the system was run with only three pieces of Pt and Au coated carbon fiber sh A Current vs Voltage-0.25 19.75 39.75 59.75 79.75 99.75 119.75 139.75 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 0.25" 2 Anodes Current Density (mA/cm2) 0.5" 2 Anodes 0.5" 2 Anodes Pt Si 0.25" 2 Anodes No SO2" Voltage (V) F igure 5.13 Total Current vs Voltage Curv es: Experiments 6 and 9 Through 11 g Once again the max current due to input flow was topped at higher voltages. With this experiment it started at 1.1 volts. Once again this means the system was relyin on unreacted SO 2 from lower voltages and amperages.

PAGE 88

76 f the electrolyte bridge is ow consistent from cathode to anode so both compartments and the height of acid above to be changed. It was decided to run the final run with this e To further test the expansion of the cro ss sectional area of the electrolyte bridge the cell will need to be rede signed. The cross sectional area o n the electrodes will need design with silicon electrodes with the same 100 to 100 angstrom deposition of Pt on th anode and cathode. Current Density vs Voltage8.75 9.752) -0.25 0.75 1.75 2.75 3.75 4.75 5.75 6.75 7.75 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2Voltage (V)Current Density (mA/cm 0.25" 2 Anodes 0.5" 2 Anodes 0.5" 2 Anodes Pt Si 0.25" 2 Anodes No SO2" Figure 5.14 Current Density vs Voltage Curv es: Experiments 6 and 9 Through 11 5.6.11 Experiment 11 k de was coated with silver to improve conducti vity. The Si electrodes were sealed to the electrod SS electrodes. This could have inhibite As stated above this experiment was run with Pt coated Si electrodes. The bac si es from the previous runs with api ezon wax. Current density went down from experiment 9 as can be seen in figure 5.11. It was noticed that hydrogen bubbles grew much larger on the Si electrodes than the gol d plated 316 d current because the hydrogen bubbles become an insulator against ion flow. As with experiments 7, 8 and 10 the system wa s restricted at higher voltages. This time like experiment 10 the high amperages were at 1.1 volts and 1.2 volts. Below is a table which summarizes all the final experiments.

PAGE 89

771 Exp. 2 Exp. 3 Exp. 4 Exp. 5 Exp. 6 Exp. 7 Exp. 8 Exp. 10 Exp. 11 Table 5.1 Experimental Parameter Comparison Parameters Exp. Pressure (psig) 0 ~80 ~80 ~80 ~80 ~80 ~80 ~80 ~80 ~80 Flow Rate (ml/min) Anode Cathode 0.01 0.03 0.010 0.03 0.01 0.09 0.01 0.09 0.01 0.09 0.01 0.09 0.01 0.09 0.01 0.09 0.01 0.09 0.01 0.09 Anode Thickness (in) 0.04 0.04 0.04 0.125 0.25 0.25 0.25 0.25 0.5 0.5 Anode Packing 5 Pt 5 Pt 5 Au 5 Pt 10Au 5 Pt 25Au 4 Pt 26Au 4 Pt 26Au 30Au 4 Pt 70Au 4 Pt 70Au Cathode Packing NA NA 5 Pt NA NA NA NA 4 Pt 70Au NA NA electrode total A/C 1/1 2/1 1/1 1/1 2/1 2/2 2/2 2/1 2/1 Electrode Pt Pt Pt Au Pt Au Pt Au Pt Au Pt Au Pt Au Pt Au Pt Material Au Au 316 316 316 316 316 316 316 316 316 Si

PAGE 90

78 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS beneficial to continue res earch on both low voltage hydroge n production techniques. It appears that H2ecsise are economasapro fo power plants and refineries than the Claus proc sulfuric acid and hydrogen production is a vi able alternative to typical sulfuric acid production. It can be concluded that roblems with SO lleviated. The flow gradient with liquid electrolyte s eems to limit the diffusion of SO2 with less harm to the overall resistance of th e system than the addition of an intermediate chamber in the solid electrol yte cell. This seems logical because Nafion has a much lower conductivity than sulfuric acid and two pieces ar e needed for the 3 compartment cell. The lower conductivity is typically over come by the thin sheets of membrane that are used and the intimate contact made be tween the electrode, catalyst and membrane when the MEAs are hot pressed. There in lies part of the solution fo r bettecurrent density in the three compartment cell. First of all thinner memb raneshould be used. Other membranes with better conductivity should be looked into. Better MEA production processes should be developed or commercial MEAs purchased. Just as catalyst deposition is an issu e in MEA production it is also a major draw back to electrode production for the liquid el ectrolyte cells. It appears Pt has a good effect on the cathode which is beneficial because it will make a good Schottky barrier when the solar cells are developed. The short comings happen at the anode. It appears that the corrosion problems at the anode have been overcome by electro The economics presented within this thes is show reasonable evidence that it is S el troly can b mo ical w te tre tment cess r ess. It also appears th at electrochem ical the in itial p 2 diffus ion have been a r s plating gold.

PAGE 91

79 Unfortunately gold does not appear to be catalytically active fo r sulfuric acid production. Extensive work needs s well as l izing the cross sectional area of the electrolyte bridge to inhibit SO2 gn .2 volts, than can be maintained at the flow rate and mes due to t is. This suggests th e need for increased uce to be done to develop an anode catalyst. This work a historical shows that RuO 2 is a candidate for the anode cat alyst. The issues with RuO 2 loss after long term electrolysis needs to be further investigat ed and corrected. Historica data also suggests WC as a good anodic and cathodic catalyst candidate. Deposition techniques for WC need to be investigated. The experiments manipulating the electr olyte bridge show that the original design minim diffusion adversely affected the performance of the cell. They show that more desi changes increasing the cro ss sectional area in the anode compartment and cathode compartment are needed to maximize the current density while still inhibiting SO 2 diffusion. It also appears that is may be necessa ry to return to the usage of liquid SO 2 to increase the concentration of SO 2 in the anode compartment. The last experiments with two cathodes and anode compartment a nd one cathode and anode compartment show greater current, at 1.1 and 1 concentration of SO 2 used in these experiments. This is possible at the short ti residual SO 2 not reacted at lower voltages and cu rrents. Increasing the flow rate can increase the amount of SO 2 but is does not appear to show a large improvement in curren according experiments not reported in this thes concentration is preferable to increased flow rate. The concentration can be increase in the SO 2 saturated H 2 SO 4 by increasing pressure, but it w ould be much easier to introd liquid SO 2

PAGE 92

80 ns e rogen Energy Transition: Moving To ward the Post Petroleum Age in nsportation. pp. 73-92. Elsevier Academic Press, Burlington, MA, 2004. y. REFERENCES [1] Joan Ogden. Where Will the Hydrogen Come From? System Consideratio and Hydrogen Supply. In Daniel Sperli ng and James S. Cannon. (Eds.). Th Hyd Tra [2] Joseph J. Romm. The Hype About Hydrogen. pp. 71-72, Island Press, Washington D.C., 2004. [3] Carbon Capture Research. US Department of Energy Office of Fossil Energ 01 Nov. 2004. Department of Energy. 29 June 2005. http://www.fossil.energy.gov/programs/ sequestration/capture/index.html [4] Official Energy Statistics From The US Government. Energy Information Administration. U.S. Department of Energy. 19 May 2006. http://www.eia.doe.gov/ [5] Bruce G. Miller, Coal Energy Systems. pp. 247-249, Elsevier Academic Press, Burlington, MA, 2005. [6] Office of Fo ssil Fuels. Hydrogen and Clean Fuels Research. US Department of Energy. December 2005. U.S. Department of Energy. 10 October 2006. http://www.fossil.energy.gov/programs/fuels/index.html. [7] H. Wendt and G Kreysa. Electrochem ical Engineering, Science and Technology in Chemical and Other Industries. pp. 317-318, Springer, Verlag, Germany, 1999. [8] Board on Energy and Environmental Systems. The Hydrogen Economy: Opportunities, Costs, Barrie rs, and R&D Needs. The National Academies Press. 2004. National Academy of Engineers. 29 June 2005. http://www.nap.edu/openbook/0309091632/html/221.html [9] Braun, Harry. The California P hoenix Project Plan. Phoenix Project Foundation. 08 Nov 2003. Phoenix Pr oject Foundation. 29 June 2005. http://www.phoenixprojectfoundatio n.us/user/California%20PPP.pdf

PAGE 93

81 [10] Office of Nuclear En ergy, Science and Technology. Nuclear Hydrogen R&D Plan. US Department of Energy Hydrogen Program. March 2004. U.S. Department of Energy. 29 June 2005 http://www.hydrogen.energy.gov/pd fs/nuclearenergyh2plan.pdf Rebecca L. Busby, Hydrogen and Fuel Cells a Comprehensive Guide. pp. 193, 203-204, Penn Well Corporation, Tulsa, OK, 2005. [11] sis, Univers ity of South Florida, Tampa, FL, US. f [14] ancial Instrum ents. ge. 27 June 2005. [12] Maheshkumar Chettiar, Co-Production of Hydrogen and Sulfuric Acid by Electrolysis. Masters The ent of an Electroly tic Cell [13] B.D Struck, H, Neumeister, A. Naoumid is, Developm for the Anodic Oxidation of Sulfur Dioxide and the Cathodic Production o Hydrogen W ithin the Sulfuric Acid Hybrid Thermochemical Cycle. International Journal of Hydrogen Energy, Vol 7, No. 1, pp 43-49, 1982. European Clim ate Exchange Market Da ta. ECX Carbon Fin European Climate Exchan http://www.europeanclimateexchange.com/indexflash.php W.D. Seider. Product an [15] d Process D e sign Principles Synthesis, Analysis, and [16] lsevier, Amsterdam, [17] ber 1982. Westinghouse Electric Corp. Pittsburgh, PA. 1982. mical : ia, 20arge, Evaluation. pp. 185-187, 443-557, John Wiley and Sons Inc.,U.S., 2004. Hartm ut Wendt. Electrochemical Hydrogen Technologies Electrochemical Production and Combustion of Hydrogen. pp. 306,330, E 1990. Solar Therm al Hydrogen Production Process: Final Report, January 1978Decem [18] Studies of the Use of Heat From High Temperature Nuclear Sources for Hydrogen Production Processes. Wes tinghouse Astronuclear Lab. Pittsburgh, PA. 1976. [19] B.D. Struck et al. Electrolytic Cell for the Thermochemical-Electroche Sulphur Hybrid Cycle f or Water Splitt ing. Hydrogen Energy Progress VIthustr Proceedings of the 6 World Hydrogen Energy Conference, Vienna, A 4 July 1986. pp. 739-743. Pergamon Press, New York, 1986. 2 [20] Economic Comparison of Hydrogen Pr oduction Using Sulfuric Acid E lectrolysis and Sulfur Cycle Water Decompositi on Final: F inal Report, June 1978. Advanced Systems Division. Westinghouse Electric Corporation. South L PA. 1978.

PAGE 94

82 ers, Vol. 100, Nos. 3-4, pp. 125-133. April 2005. 2] Understanding How the SEM Works a nd How to Use It On a College Level. [21] I. Constantinou et al. Electrochem ical Promotion of RuO2 Catalysts for the Cumbustion of Toluene and Ethylene. Catalysis Lett [2 Materials Science & Engineering Department Iowa State University. 21 October 2006. http://mse.iastate.edu/microscopy/path.html SEM/EDS : Scanning Electron Microsc opy With X-ray Microanalysis. S [23] chool of Dental Medicine. Univer sity at Buffalo. 21 October 2006. http://www.sdm.buffalo.edu/scic/sem-eds.html Cleaning Procedures for Silicon Wa fers. Holon Institute of [24] Technology http://www.hait.ac.il/staff/lapsker/50082/Cleaning%20Proced ures%20for%20Sili con%20Wafers.doc [2 5] W. Kern and J. Vossen. Thin Film Processes. Academic Press, New York, 1978. [26] s, Inc. 21 October 2006. http://www.mastest.com/xrdxrr.htm# XRD/XRR. Semicon ductors & Microelectronics Materials Analytical Service lls. Journal of Applied Electrochemistry, 35, pp. 931-938, 2005. [28] l ectrochemistry. Advances in Colloid and Interface Science, 64, pp.173-251, 1996. [29] 05. 05. [31] ero et al. Developm ent and Performance Ch aracterization of New Electrocatalysts for PEMFC. Journal of Power Sources, 106, pp. 206-214. 2] ility of Sulfur Dioxide in Aqueous Sulfuric Acid Solutions. Journal of Chemi cal Engineering Data, 93, pp.506-509, 1988. [27] L. Declan Burke and Nageb S. Naser. Metastability and Electrocatalytic Activity of Ruthenium Dioxide Cathodes Used in Water Electrolysis Ce S. Ardizzone and S. Tras atti, Interfacial Property of Oxides with Technologica Impact in El Neelkanth G. Dhere, and Anant H. Jahagirdar. Photoelectrochemical Water Splitting for Hydrogen Production Using Combination of CIGS2 Solar Cell and RuO 2 Photocatalyst. Thin Solid Films, 480-481, pp. 462-465, 20 [30] X. G. Yang and C. Y. Wang. Nanos tructured Tungsten Carbide Catalysts for Polymer Electrolyte Fuel Cells. Applied Physic Letters, 86, 224104, 20 M. J. Escud 2002. [3Walter Hayduk et al. Solub

PAGE 95

83 APPENDICES

PAGE 96

84 Appendix A: Data Acquisiti on Hardware and Programs Figure A.1 Power Supply Cont rol Labview User Interface

PAGE 97

Appendix A: (Continued) error in ( no error ) OVP Enabled ( True ) V olta g e Level ( 0.0 Volts ) Out p ut Enabled ( False ) Current Limit ( 0 Am p s ) error out Over Volta g e Protection OVP Limit ( 22.0 Volts ) V ISA session du p VISA session 5000 Miliam p s 2 Measurement T yp e ( Current ) 1 error in ( no error ) 2 OVP Enabled ( True ) 2 V olta g e Level ( 0.0 Volts ) 2 Out p ut Enabled ( False ) 2 Current Limit ( 0 Am p s ) 2 error out 2 Over Volta g e Protection 2 OVP Limit ( 22.0 Volts ) 2 V ISA session 2 du p VISA session 2 5000 Milliam p s Measurement T yp e ( Current ) 2 Measurement 2 Measurement T yp e ( Volta g e ) 2 Measurement 1 Measurement T yp e ( Volta g e ) 1 a g e364xa Measure.vi Build Arra y Build Arra y Last Time Lo gg ed Last Date Lo gg ed 1000 1000 Figure A.2 Power Supply Cont rol Labview Block Diagram Figure A.3 Data Aquisition Control Labview User Interface 85

PAGE 98

Appendix A: (Continued) Device V ota g e In p ut Limits low limit hi g h limit 7 6 5 4 3 2 1 0 Plot.Name A ctPlot Plot.Name A ctPlot Plot.Name A ctPlot Plot.Name A ctPlot Plot.Name A ctPlot Plot.Name A ctPlot Plot.Name A ctPlot Plot.Name A ctPlot V OLTAGE Channels Build Arra y 5000 V OLTAGE Latest Readings 1 1 Last Time Lo gg ed Last Date Lo gg ed Build Arra y V OLTAGE 2 Latest Readings 2 30 30 Plot.Name A ctPlot Plot.Name A ctPlot V OLTAGE 2 Device 2 V olta g e In p ut Limits 2 low limit hi g h limit Channels 2 1 2 Figure A.4 Data Aquisition Co ntrol Labview Block Diagram 86

PAGE 99

87 tion Equipment Following are pictures and specifications of the characterization equipment used in the thesis. They can be found on the website for the University of South Florida Nanomaterials and Nanomanuf acturing Research Center, http://nnrc.eng.usf.edu/Labs&R esources/toollist.asp?id=2. Appendix B: Characteriza Figure B.1 Hitachi S-800 Sc anning Electron Microscope Specifications : 2 nm Resolution 300,000X mag. Cold cathode field emission source Accepts specimens up to 25 mm dia., by 20 mm height EDAX-Phoenix EDS System Figure B.2 Hitachi S-800 Specifications

PAGE 100

Appendix B: (Continued) Figure B.3 X'Pert Diffractometer Specifications : Selectable line or point focus 1.8kW seal ed ceramic copper x-ray tube source Choice of PreFIX incident beam optics include: For Line Focus ApplicationsHybrid 4-bounce Monchromator with automatic attenuator For Point Focus ApplicationsGe [440] Monochromator rossed slits Collimator iffracted beam optics include: ree boun ce (022) channel cut Ge crystal to onochromator iving Slit with Fixe d Anti scatter Slit and a curved crystal monochromator lly encoded sample positioning enables a 1/2 circle Eulerian cradle with motorized sample stage enables sample tilts of +/90, in-plane rotation of 360, in-plane X and Y translations of 100 mm, and vertical Z displacement of 11 mm 2 sealed proportional detectors with a large dynamic range Fixed Divergence Slit Module X-Ray Mirror with automatic attenuator C d Choice of PreFIX Triple axis setup utilizes a th provide an acceptance an gle of 12 arc seconds llimator with a Flat Crystal M The parallel plate co Programmable Rece High resolution goniometer with optica minimum step size of 0.0001 F igure B.4 X'Pert Diffractometer Specifications 88

PAGE 101

Appendix B: (Continued) Figure B.5 Anatech Limited Humm er X Sputter Coater Figur B Specific ations : Max V Modes: Plate, E Base Vacuum: 0 Target: 3 inc Max Curent: 20 mA r oltage: 30 00 V.D.C. h, Plasma tc 2 mT Gold Palladium h e.6 A natech Limited Humm er X Sputter Coater Specifications 89

PAGE 102

Appendix B: (Control) Figure B.7 Plasma Sciences CRC-100 Sputter Tool Specifications : Power supplies: DC1 50watts max, RF 200watts .125 thick maximum r ce: 24 adjustable ghing pump Alcatel 5081 turbo pump, 15min pumpdown time. Cooling water: .5 lpm Gases: Argon Target: 2 diameter fixed x Sample size: up to a 4 wafe Target to sample distan Vacuum system: Welsh duo rou Figure B.8 Plasma Sciences CRC100 Sputter Tool Specifications Power: 120Vac 15amps 90

PAGE 103

91 ings Following are two dimensional drawings of the individual sections of the low pressure two inch cell. Appendix C: Cell Draw Figu ell Anode Compartment re C.1 Low Pressure C

PAGE 104

Appendix C: (Continued) Figure C.2 Low Pressure Cell Electrode Assembly F 92 igure C.3 Low Pressure Cell Cathode Compartment

PAGE 105

Appendix C: (Continued) 93 Following are two dimensional drawings of th e individual sections of the 80 psig two inch cell. Figure C.4 80 psig Cell Anode Compartment

PAGE 106

Appendix C: (Continued) Cathode Anode Spacer Spacer Spacer Spacer One Piece Insulator F igure C.5 80 psig Cell Electrode Assembly 94

PAGE 107

Appendix C: (Continued) Figure C.6 80 psig Cell Cathode Compartment 95