UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Enhancing aluminum corrosion for hydrogen generation Skrovan, John 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_fall_skrovan_john.pdf [ 2.59MB ]
Metadata
JSON: 24-1.0067321.json
JSON-LD: 24-1.0067321-ld.json
RDF/XML (Pretty): 24-1.0067321-rdf.xml
RDF/JSON: 24-1.0067321-rdf.json
Turtle: 24-1.0067321-turtle.txt
N-Triples: 24-1.0067321-rdf-ntriples.txt
Original Record: 24-1.0067321-source.json
Full Text
24-1.0067321-fulltext.txt
Citation
24-1.0067321.ris

Full Text

Enhancing Aluminum Corrosion For Hydrogen Generation by John Skrovan B. Sc. University of Texas at Austin, 1990 M. Sc. University of Texas at Austin, 1994 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Materials Engineering) UNIVERSITY OF BRITISH COLUMBIA (Vancouver) July 2009 ©John Skrovan 2009  Abstract Aluminum powder when ball milled with secondary particles will corrode in water releasing hydrogen. With corrosion approaching 90% of the available aluminum within 20 minutes, this process is of interest as a portable hydrogen source. Electrochemical polarization and hydrogen capture tests are used to study changes in solution pH and temperature during the reaction, as well as calculate the activation energy for corrosion of the ball milled powder. Evidence that the ball milled aluminum is increasing the solution pH is presented along with tests indicating the pH shift is not sufficient to account for the increase in corrosion rate. The effect of solution temperature on reaction products and corrosion rate for aluminum powders is measured, and the hypothesis that the exothermic nature of the reaction, combined with a deformed surface, is creating high temperature micro-environments is discounted. Activation energy for the rate limiting step in the corrosion of ball milled aluminum is calculated to be between 72 – 74 kJ/mol, similar to that seen for aluminum disks at a similar pH. Finally BET measurements show an increase in surface area of the aluminum particles after ball milling between 10x-20x. The amount of hydrogen evolved in the first hour is seen to correlate almost exactly with the aluminum surface area. The addition of alumina powder without ball milling is shown to increase the corrosion rate of aluminum powders by an order of magnitude or greater and to delay or prevent passivation of the aluminum. Two models are proposed to explain this observation and tests run to support them. The high surface area (10m2/g) of alumina is thought to provide an alternative deposition site for hydroxide formed during aluminum ii  corrosion and the positive surface charge alumina acquires at pH 7 to provide a source of protons for hydrogen evolution. The hydrogen exchange current densities on aluminum and platinum surfaces are shown to increase by over an order of magnitude in the presence of alumina particles. The acceleration of aluminum corrosion is only seen with electrical contact between the aluminum and alumina, but contact is not required to delay passivation.  iii  Table of Contents Abstract ............................................................................................................................... ii Table of Contents .............................................................................................................. iv List of Tables ..................................................................................................................... vi List of Figures .................................................................................................................. vii List of Equations ............................................................................................................. xvi Acknowledgements ........................................................................................................ xviii Nomenclature .................................................................................................................. xix Chapter 1 Introduction ...................................................................................................... 1 Chapter 2 Literature Review .............................................................................................. 4 2.1 Aluminum and Its Corrosion Properties in Water .......................................................... 5 2.1.1 Production and Recycling of Aluminum ...................................................................................... 5 2.1.2 Hydroxide Formation During Aluminum Corrosion.................................................................... 7 2.1.3 Growth and Dissolution of Passive Oxide Films ....................................................................... 12  2.2 Hydrogen Production from Aluminum........................................................................... 23 2.2.1 Reaction of Al with Alkaline Aqueous Solution ........................................................................ 23 2.2.2 Aluminum – Water Reaction in Neutral Solutions .................................................................... 26  2.3 Hydrogen ........................................................................................................................... 31 2.3.1 Hydrogen Production ................................................................................................................. 32 2.3.2 Hydrogen Storage ...................................................................................................................... 37  2.4 Review Summary .............................................................................................................. 44  Chapter 3 Research Objectives ........................................................................................ 47 Chapter 4 Experimental Aspects ..................................................................................... 49 4.1 Materials ............................................................................................................................ 49 4.2 Sample Preparation .......................................................................................................... 51 4.3 Hydrogen Evolution Test Apparatus .............................................................................. 53 4.4 Electrochemical Investigations ........................................................................................ 54 4.5 Materials Characterization .............................................................................................. 57 4.5.1 Scanning Electron Microscopy (SEM / EDS) ............................................................................ 57 4.5.2 X-Ray Diffractometry (XRD) .................................................................................................... 57 4.5.3 Surface Area Measurement by BET Analysis............................................................................ 57 4.5.4 Transmission Electron Microscopy (TEM)................................................................................ 58  4.6 Errors and Reproducibility .............................................................................................. 58  Chapter 5 Corrosion of Ball Milled Aluminum .............................................................. 61 5.1 Effect of Secondary Particles ........................................................................................... 63 5.1.1 Electrochemical Tests ................................................................................................................ 64 5.1.2 Hydrogen Evolution ................................................................................................................... 66  iv  5.2 Activation Energy ............................................................................................................. 69 5.2.1 Aluminum Disk .......................................................................................................................... 70 5.2.2 Ball Milled Powders................................................................................................................... 75  5.3 pH Effects .......................................................................................................................... 82 5.3.1 Electrochemical Tests ................................................................................................................ 82 5.3.2 Hydrogen Evolution from Pure Aluminum ................................................................................ 85  5.4 Temperature Effects ......................................................................................................... 87 5.4.1 Hydrogen Evolution for BM Powders ....................................................................................... 87 5.4.2 Electrochemical Tests ................................................................................................................ 88 5.4.3 Hydrogen Evolution Rate for Al Powders ................................................................................. 90 5.4.4 XRD Characterization of Corrosion Products ............................................................................ 91  5.5 Surface Area Effects ......................................................................................................... 93 5.5.1 Changing Amount of Secondary Particles ................................................................................. 95 5.5.2 Surface Area of Secondary Particles .......................................................................................... 98 5.5.3 BET for Varying Oxide Percentage ..........................................................................................100 5.5.4 Hydrogen Evolution for Constant Surface Area .......................................................................102 5.5.5 TEM Images of BM Aluminum ................................................................................................106  5.6 Summary .......................................................................................................................... 108  Chapter 6 Enhancing Corrosion of Al without Ball-Milling ....................................... 111 6.1 Corrosion of Aluminum Powder with added Alumina ................................................ 112 6.1.1 Electrochemical Tests ...............................................................................................................112 6.1.2 Hydrogen Evolution ..................................................................................................................114  6.2 Subsequent Runs ............................................................................................................. 117 6.3 Proton Donor Model ....................................................................................................... 122 6.3.1 Effect of Iso-electric Point ........................................................................................................124 6.3.2 Surface Area Effect ...................................................................................................................125 6.3.3 Cathodic Slopes ........................................................................................................................126 6.3.4 Effect on Hydrogen Evolution off Platinum .............................................................................128 6.3.5 Electrical Contact with Aluminum............................................................................................130  6.4 Surface for Hydroxide Deposition Model ..................................................................... 132 6.4.1 Effect on Hydrogen Evolution Curves ......................................................................................134 6.4.2 Leaching Model ........................................................................................................................139  6.5 Summary .......................................................................................................................... 145  Chapter 7 Conclusions ................................................................................................... 147 7.1 General Conclusions ....................................................................................................... 147 7.2 Importance to Industrial Applications .......................................................................... 150  Chapter 8 Recommendations for Future Work ............................................................ 151 References ...................................................................................................................... 153 Appendices ...................................................................................................................... 164 Appendix A - BET Plots ....................................................................................................... 164 Appendix B - Activation Energy .......................................................................................... 177 Appendix C – Repeat Graphs .............................................................................................. 179  v  List of Tables  Table 1 Activation energies for the series of tests of Al corrosion in water. _________ 81 Table 2 Surface areas for BM and as received powders from BET.________________ 99  vi  List of Figures Figure 1 Schematic of processes that lead to the formation of bi-layer passive films on metal surfaces. (Redrawn from MACDONALD, 1992) __________________________ 9 Figure 2. E-pH diagram for Al-water at 25oC.________________________________ 11 Figure 3 Schematic of physico-chemical processes that occur within a passive film according to the point defect model. m=metal atom, MM=metal cation in cation site, OO=oxygen ion in anion site, VMΧ’=cation vacancy, V-O=anion vacancy, VM=vacancy in metal phase. (Redrawn from MACDONALD, 1999) ___________________________ 14 Figure 4 Point defect model for the pitting of Al. Under anodic polarization, aluminum ions are driven to the Al2O3:solution interface and into solution, while aluminum vacancies are driven to the Al:Al2O3 interface where they aggregate to form voids as precursors to pits. Oxygen ions are also driven to the Al:Al2O3 interface, while oxygen vacancies migrate to the Al2O3:water interface to be filled by water molecules. (Redrawn from BUNKER, 2002) ___________________________________________________ 16 Figure 5 Uniform Corrosion Model for Al in hot water. (Redrawn from BUNKER, 2002) _____________________________________________________________________ 19 Figure 6 Aluminum particles – as received. __________________________________ 50 Figure 7 Alumina particles – as received. ___________________________________ 50 Figure 8 Al/Al2O3 mixture ball milled for 15 minutes. __________________________ 51 Figure 9 SPEX 8000 Shaker Mill.__________________________________________ 52 Figure 10 Ball mill vial and alumina balls. __________________________________ 53 Figure 11 Hydrogen evolution experimental setup. ____________________________ 54 Figure 12 Electrochemical cell. ___________________________________________ 55 vii  Figure 13 Working electrode sample holder. _________________________________ 55 Figure 14 Luggin probe tip with Vycor frit and sample holder. ___________________ 56 Figure 15 Potentiostat/Gavanostat model 273A. ______________________________ 56 Figure 16 Hydrogen evolution from 0.5g of aluminum powder in 55oC DI water with as received and ball milled (50 wt% Al2O3) powders. ____________________________ 61 Figure 17 1g of aluminum powder ball milled with 1g of KCl and placed in 30 mL of tap water at 55oC. _________________________________________________________ 62 Figure 18 Potentiodynamic scans of Aluminum powder ball milled with Cu, MgO, SiO2, and Al2O3 in a 0.1M Na2SO4 solution at 25oC. ________________________________ 65 Figure 19 Potentiodynamic scans of Aluminum powder BM with Al2O3, SiO2, and 2% additions of MgO in a 0.1M Na2SO4 solution at 25oC.__________________________ 66 Figure 20 Hydrogen evolution from 0.5 g of aluminum powder BM with Al2O3, SiO2, and MgO in 50ml of DI water at 70oC. _________________________________________ 67 Figure 21 Hydrogen evolution from 0.5 g of aluminum powder BM with 50% Cu in a 70oC 0.1 M Na2SO4 solution. _____________________________________________ 68 Figure 22 Hydrogen evolution from 0.5 g of aluminum powder BM with 0.5 g of Si3N4, Al2O3, and SiC in 100ml of DI water at 60oC. ________________________________ 69 Figure 23 Potentiostatic test of current verses time for an aluminum disk in 0.01M borax solution held at +200mV for 12 hours at each temperature step from 10oC to 85oC. __ 71 Figure 24 Arrhenius plots for the aluminum disk in 0.01M Borax solution for temperature steps from 10oC to 70oC and then 85oC and back down to 10oC. Ea=64 kJ/mole. _____ 72 Figure 25 Arrhenius plot for aluminum disk held at -200mV in 0.01M Borax solution for temperature steps from 10oC to 55oC. Ea=76.8 kJ/mole. ________________________ 73  viii  Figure 26 Potentiodynamic scans of a non-milled pellet of 50% Al and 50% alumina in a .1M Na2SO4 solution at temperatures from 5oC to 55oC after cycling two times. _____ 74 Figure 27 Arrhenius plot for non-milled Al/alumina pellet in 0.1M NaSO4 solution. __ 74 Figure 28 Potentiodynamic scans of a non-milled pellet 100% Al in a 0.1M Na2SO4 solution at temperatures from 5oC to 55oC after cycling two times. _______________ 75 Figure 29 Potentiostatic test for BM Al in 0.01M Borax at +200mV at various temperatures. _________________________________________________________ 76 Figure 30 Arrhenius plot for BM Al in 0.01M Borax at +200mV from 15oC to 55oC. Ea=23.9 kJ/mole._______________________________________________________ 77 Figure 31 Potentiodynamic scans for BM Al/Alumina pellet in 0.1M Na2SO4 solution at temperatures from 5oC to 35oC after initial series run up to 55oC. ________________ 78 Figure 32 Arrhenius plot for BM Al/Alumina pellet in .1M Na2SO4 solution after initial series of runs to 55oC. ___________________________________________________ 79 Figure 33 Arrhenius plot for BM Al/Alumina pellet in 0.1M Na2SO4 solution during initial run from 5oC to 45oC. ______________________________________________ 80 Figure 34 Potentiodynamic scan of an aluminum disk, an aluminum powder pellet, and a BM al/alumina pellet in water at 25oC. _____________________________________ 83 Figure 35 Potentiodynamic scans of an aluminum disk and a BM al/alumina pellet in a 0.01M Borax solution at 25oC. ____________________________________________ 84 Figure 36 H2 evolution from aluminum powder in DI water (pH 7) and NaOH solutions (pH 10 and 11). ________________________________________________________ 86 Figure 37 Hydrogen evolution rate for 0.5g Al BM with 0.5g Alumina in 100mL DI water at 55oC, 65oC, and 85oC starting temperatures. _______________________________ 88  ix  Figure 38 Potentiodynamic curve of aluminum powder in water at 25oC and 50oC. __ 89 Figure 39 Potentiostatic test of aluminum disk held at +200mV in 0.01M Borax solution at 70oC and 85oC. ______________________________________________________ 89 Figure 40 Hydrogen evolution with time for aluminum powder 0.5g in DI water at 55oC, 70oC, and 95oC.________________________________________________________ 90 Figure 41 Percentage of aluminum corroded with time for 0.5g of aluminum powder in DI water at 55oC, 70oC, and 95oC. _________________________________________ 91 Figure 42 XRD pattern for aluminum disks held in 0.01M Borax solution for 24 hours at 55oC and 85oC. ________________________________________________________ 92 Figure 43 Surface area in m2 g-1 for a spherical particle of aluminum of given radius. 94 Figure 44 Percentage of 0.5g of aluminum that corrode with time in 65oC water after BM with 25%, 50%, or 75% alumina by weight, or 50% KCl that is subsequently washed from the powder. _______________________________________________________ 95 Figure 45 Volume of hydrogen gas evolved from 0.5g of aluminum in 65oC water after BM with 25%, 50%, or 75% alumina, or 50% KCl that was subsequently washed from the powder. ___________________________________________________________ 96 Figure 46 Volume of hydrogen gas evolved from 0.5 g of aluminum in 65oC water that was BM with either 50% or 25% by weight of Al2O3, Si3N4, or SiC. _______________ 97 Figure 47 Isotherm plot for 8g of aluminum powder after 72 hours under vacuum at 150oC. _______________________________________________________________ 98 Figure 48 BET plot for 8g of Aluminum powder giving a surface area of 0.59 m2 g-1. _ 99 Figure 49 BET plot for 1.97g of BM aluminum/alumina powder 50%/50% with a surface area of 5.3 m2 g-1. _____________________________________________________ 100  x  Figure 50 Surface area calculated from BET for aluminum BM with 25%, 50%, or 75% Al2O3 and 100% Al2O3 that was not BM. ___________________________________ 101 Figure 51 Volume of hydrogen gas evolved from aluminum powder in 65oC water after BM with 25%, 50%, or 75% alumina, or 50% KCl and subsequently washed, with all volumes scaled to an equal surface area of 6.88 m2. __________________________ 103 Figure 52 Volume of hydrogen gas evolved from aluminum powder and washed Al in 65oC water with surface area of aluminum equal to 6.88 m2. ___________________ 104 Figure 53 Rate of hydrogen evolution per minute for equal surface area of aluminum powder and washed Al (6.88 m2) in 65oC water. _____________________________ 105 Figure 54 Potentiodynamic scans of a washed aluminum pellet in a 0.1M Na2SO4 solution at temperatures from 5oC to 55oC after cycling two times. ______________ 106 Figure 55 Bright field of 10-20 nm particles from BM Al/Al2O3. _________________ 107 Figure 56 Dark field of 10-20 nm particles from BM Al/Al2O3. __________________ 107 Figure 57 Dark field of several particle clusters. _____________________________ 108 Figure 58 Potentiodynamic curve of an aluminum/alumina mixture in water compared to aluminum powder and a BM aluminum/alumina mixture. ______________________ 113 Figure 59 Potentiodynamic scans for aluminum/alumina pellets in DI water with the amount of aluminum varying from 33% to 75%. _____________________________ 114 Figure 60 Hydrogen evolution from 0.5g of aluminum powder in 75ml of 65oC DI water with 5g of alumina powder added. ________________________________________ 115 Figure 61 Volume of hydrogen released in one, two, and three hours after immersion in 65oC water as a function of the number of grams of Al2O3 powder added to 1 gram of aluminum powder._____________________________________________________ 116  xi  Figure 62 Volume of hydrogen released with time for 0.5g Al in 70oC water with additions of Al2O3, SiC, CuO, La2O3, and SiO2. ______________________________ 116 Figure 63 Volume of hydrogen released from 5g Al2O3 in 75ml of 65oC DI water with the addition of 0.5g Al powder at the start of each run, run 6 started one day later. ____ 117 Figure 64 Volume of hydrogen evolved when 0.5g of Al powder is added every 30 minutes to 5g alumina in 75ml of 65oC DI water. ____________________________ 118 Figure 65 Volume of hydrogen released with the addition of 0.5g Al every 5 minutes to 5g Alumina in 75ml of DI water in insulated beaker. ____________________________ 119 Figure 66 Cathodic polarization scans of aluminum powder in 0.045 M KCl and 0.5 M boehmite solutions at 25oC. _____________________________________________ 120 Figure 67 Hydrogen evolution with time from 0.5 g of aluminum powder in DI water with 5g of Al2O3 compared to 0.5 g of aluminum powder in a 0.5 M Boehmite solution with and without the 5 g of alumina. __________________________________________ 121 Figure 68 Volume of hydrogen evolved with time for 0.5 g of Al powder in 100ml DI water at 70oC with and without the addition of 5 g Si3N4. ______________________ 124 Figure 69 Difference in hydrogen evolution from 0.5g Al powder in 70oC DI water, between high surface area particles Al2O3 and Si3N4 ( >8m2/g) and low surface area particles SiC (<0.8m2/g). _______________________________________________ 125 Figure 70 Cathodic polarization scan for an aluminum powder pellet and an aluminum/alumina pellet in a 0.045M KCL solution at 25oC. ___________________ 127 Figure 71 Linear fit to Tafel region of cathodic slopes for calculation of hydrogen exchange current density (hio) for aluminum powder and aluminum/alumina mixture in 0.045M KCL solution at 25oC. ___________________________________________ 128  xii  Figure 72 Linear fit to Tafel region for calculation of hydrogen exchange current density (hio) for platinum foil and platinum surrounded by alumina powder in a 0.1M KCl solution at 25oC. ______________________________________________________ 129 Figure 73 Linear fit to Tafel region for calculation of hydrogen exchange current density (hio) for platinum foil and platinum surrounded by alumina powder in a 0.1M KCl solution at 55oC. ______________________________________________________ 130 Figure 74 Volume of hydrogen released from 0.5 g of Aluminum powder with and without contact with 5 g Alumina powder in 70oC water. _____________________________ 131 Figure 75 Volume of hydrogen evolved with time from 0.5g of Aluminum powder by itself and mixed with 0.5g of Al2O3 at temperatures from 60oC to 100oC for initial reaction. 134 Figure 76 Volume of hydrogen evolved with time from 0.5g of aluminum powder by itself and mixed with 0.5g of Al2O3 at temperatures from 60oC to 100oC for long term immersion. ___________________________________________________________ 135 Figure 77 Volume of hydrogen from 0.5g Al powder at 70oC in water with 5 g of Al2O3 mixed and held separate (no contact). _____________________________________ 136 Figure 78 Long term effect on hydrogen evolution when 5g Al2O3 is not in contact with 0.5g Al powder in 70oC water. ___________________________________________ 137 Figure 79 Si3N4 particle as received showing no Al. __________________________ 138 Figure 80 Si3N4 particle after reaction for 3 days in 70oC water with Al. __________ 138 Figure 81 The shrinking core model predicts a linear fit of 1 – (Uƒ)1/3 with time for aluminum powder in 70oC water. _________________________________________ 139 Figure 82 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder in 70oC water. _________________________________________________ 140  xiii  Figure 83 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder in 95oC water. _________________________________________________ 141 Figure 84 Showing the fit to a shrinking core model with time for 0.5g of washed aluminum powder in 70oC water. _________________________________________ 142 Figure 85 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder hand mixed with 0.5g of Al2O3 in 100oC water.________________________ 142 Figure 86 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder hand mixed with 5g of Al2O3 in 60oC water. __________________________ 143 Figure 87 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder not in contact with 5g of Al2O3 in 70oC water. ________________________ 143 Figure 88 The fraction of aluminum that has not yet reacted when a linear fit to the shrinking core model (1 – (Uƒ)1/3) falls below a R2 of 0.999 and R2 of 0.99. ________ 144 Figure 89 Isotherm plot for BM aluminum 50%/ Alumina 50%. _________________ 164 Figure 90 BET Plot of Al2O3 with a surface area of 9.64 m2 g-1. _________________ 165 Figure 91 BET plot for Washed Al powder with a surface area of 13.76 m2 g-1. _____ 166 Figure 92 BET plot for as received aluminum powder with a surface area of 0.61 m2 g-1. ____________________________________________________________________ 167 Figure 93 BET plot of BM Al/Al2O3 50%/50% mixture with a surface area of 5.33 m2 g-1. ____________________________________________________________________ 168 Figure 94 BET plot of BM Al/Al2O3 25%/75% mixture with a surface area of 7.85 m2 g-1. ____________________________________________________________________ 169 Figure 95 BET plot of BM Al/Al2O3 75%/25% mixture with a surface area of 3.67 m2 g-1. ____________________________________________________________________ 170  xiv  Figure 96 BET plot of BM washed Al/KCL mixture with a surface area of 13.01 m2 g-1. ____________________________________________________________________ 171 Figure 97 BET plot of BM washed Al/KCl mixture with a surface area of 13.76 m2 g-1.172 Figure 98 BET plot of Al2O3 powder with a surface area of 9.64 m2 g-1.___________ 173 Figure 99 BET plot of SiC powder with a surface area of 0.73 m2 g-1. ____________ 174 Figure 100 BET plot of Si3N4 powder with a surface area of 8.60 m2 g-1. __________ 175 Figure 101 BET plot of NaCl with a surface area of 1.39 m2 g-1. ________________ 176 Figure 102 Potentiostatic test of an aluminum disk at -200mV in 0.01M Borax held for 12 hours at each temperature from 10oC to 85oC._______________________________ 177 Figure 103 Arrhenius plot for al disk held at +200mV in 0.01M borax solution re-run a second time.__________________________________________________________ 178 Figure 104 Arrhenius plot for BM Al in 0.1M NaSO4 at 5oC and 15oC using separate pellets with current taken from Tafel fit at Eoc. ______________________________ 178 Figure 105 Hydrogen released from 0.5 g Al when BM with KCl or alumina and put in 70oC DI water. _______________________________________________________ 179 Figure 106 Hydrogen released from 0.5 g Al when BM with KCl or alumina and put in 70oC tap water. _______________________________________________________ 180 Figure 107 Potentiodynamic curve of aluminum BM with CuO in 0.1M Na2SO4 solution. ____________________________________________________________________ 181 Figure 108 2 Potentiodynamic scans for aluminum BM with Al2O3 in a 0.1M Na2SO4 solution at 25oC. ______________________________________________________ 182  xv  List of Equations  Eq. 1 Al → Al3+ + 3e- ....................................................................................................... 7 Eq. 2 O2 + 2H2O +4e- → 4OH- ....................................................................................... 7 Eq. 3 2H+ + 2e- → H2 ...................................................................................................... 7 Eq. 4 Al3+ + 3OH- → Al(OH)3 ......................................................................................... 8 Eq. 5 Al(OH)3 + 3H3O+ → Al3+ + 6H2O ......................................................................... 9 Eq. 6 Al2O3 + 6H3O+ → 2Al3+ + 9H2O ........................................................................... 9 Eq. 7 Al(OH)3 + OH¯ → Al(OH)4¯ ................................................................................. 9 Eq. 8 Al2O3 + 3H2O + 2 OH¯ → 2 Al (OH)4¯ ................................................................. 9 Eq. 9 2Al + 6H2O → 3H2 + 2Al(OH)3 ....................................................................... 10 Eq. 10  2Al + 4H2O → 3H2 + 2AlOOH .................................................................... 10  Eq. 11  Al2O3 + H2O → 2AlOOH ................................................................................. 19  Eq. 12  6AlOOH +2Al → 4Al2O3 + 3H2 ....................................................................... 20  Eq. 13  Al +2H2O → AlOOH +3/2H2 ........................................................................... 20  Eq. 14  Al + 4OH¯ → Al(OH)4¯ + 3e¯ ......................................................................... 21  Eq. 15  Al + 3OH¯→ Al(OH)3 + 3e¯ ............................................................................ 21  Eq. 16  Al(OH)3+ OH¯ → Al(OH)4¯ ........................................................................... 21  Eq. 17  2H2O + 2e¯ = H2 + 2OH¯................................................................................ 21  Eq. 18  2Al + 6H2O + 2NaOH → 2NaAl(OH)4 + 3H2 ................................................. 24  Eq. 19  NaAl(OH)4 → NaOH + Al(OH)3 ...................................................................... 24  Eq. 20  2Al + 6H2O → 2Al(OH)3 + 3H2 ....................................................................... 24  Eq. 21  2Al + 3H2O → Al2O3 + 3H2 ............................................................................. 29 xvi  Eq. 22  Al + 2H2O → AlOOH +1.5 H2 ......................................................................... 32  Eq. 23  27 kg + 36 kg → 60 kg + 3 kg .......................................................................... 32  Eq. 24  CH4 + H2O → CO + 3H2 ................................................................................. 34  Eq. 25  CO + H2O → CO2 + H2 ................................................................................... 34  Eq. 26  2H2O → 2H2 + O2 ............................................................................................ 36  Eq. 27  Al + 3H2O → 1.5H2 + Al(OH)3 ........................................................................ 53  Eq. 28  Al + 2H2O → 1.5H2 + AlOOH ......................................................................... 53  Eq. 29  k = Ae-Ea/RT ........................................................................................................ 69  Eq. 30  1 – (Un-reacted Fraction)1/3 ........................................................................... 139  xvii  Acknowledgements  I wish to give my sincere thanks to Dr. Akram Alfantazi and Dr. Tom Troczynski from the department of Materials Engineering at UBC for their guidance and encouragement throughout this research project. I would also like to thank all of my supervisory committee members for their help and support.  My thanks are also extended to all the other people that keep this department running and provide so much assistance, the staff in stores, the machine shop, and in the office, the other faculty that keep labs running, provide training, and oversee equipment.  Special thanks go to Edith Czech and Dr. Edouard Asselin for all the training, knowledge, and many discussions.  I would further like to acknowledge the financial support received from the Natural Sciences and Engineering Research Council of Canada (NSERC), Global Hydrogen Technologies, and the University of British Columbia.  xviii  Nomenclature BM CE E Eº ē hio i ρ IEP M MW OCP RE SEM SMR WE XRD PDM bl ol  Ball Milled Counter electrode Potential (Volts) Standard potential of a cell reaction when that reaction involves the oxidation of molecular hydrogen to solvated protons (Volts) Electron hydrogen exchange current density Current density (A/m2) Density (kg/L) Iso-electric Point Molarity, Moles of solute / Liter solution Molecular weight (g/mol) Open circuit potential Reference electrode Scanning Electron Microscope Steam methane reforming Working electrode X-ray Diffraction Point Defect Model Barrier layer Outer layer  xix  Chapter 1 Introduction  With the current concern about global warming, air pollution, and diminishing reserves of fossil fuels, there has been increased emphasis on transitioning to a hydrogen economy. However, if hydrogen is to be widely used as a future energy carrier, storage will be needed to meet time-varying demands for fuel and allow for mobile energy sources. By far the most common storage solution for small systems is currently steel pressure cylinders. Industry has set a target of a cylinder with a gravimetric storage density of 6 mass % and a volumetric storage density of 30 kg/m3 (ZUTTEL, 2003). Aluminum being a reactive metal, it will readily form aluminum hydroxide when placed into water with a resulting release of H2 gas (VARGEL, 2004). Looking at a mass balance equation, we can calculate that aluminum could have an equivalent storage density of 11 mass%, well above current storage targets. However, in practice Al is almost always found with a protective Al2O3.nH2O oxide/hydroxide layer covering the bulk material due to reaction with oxygen and water vapor in the atmosphere. This layer is non-soluble and prevents the Al-water reaction. In near neutral pH pure water systems, aluminum is reported to form various aluminum hydroxides depending on water temperature and pressure, all hydroxides form a passivating layer that grows to at most a couple of microns thickness before the reaction stops (ALWITT, 1976). For this reason aluminum has not traditionally been viewed as a practical source for hydrogen generation.  1  In the 1990’s, professor Asok Chaklader of U.B.C. discovered that, after milling aluminum with aluminum oxide, the resulting mixture when submerged in water at near neutral pH would continuously produce hydrogen (CHAKLADER, 2002). This reaction was dependent on temperature, pH, proportion and particle size of the ingredients and the mixing conditions. The mixture would react over the entire range of wt% of additives and proceed for the pH range of 9>pH>4. This effect has a number of advantages for point of use hydrogen production. Only water and the aluminum/alumina powders are needed so there are no concerns with toxic materials in the event of a leak or spill, and the storage and transport of powder can be done at atmospheric pressure and room temperature. To increase both the rate of hydrogen release and the extent of reaction, a series of alternative additives to aluminum oxide has been investigated. To some extent almost any small hard particle can be used during milling, but the best results to date came from a 50 weight percent mixture of KCl with aluminum (CZECH, 2006). To improve the overall hydrogen per gram yield, the salt can be subsequently washed from the powder to leave just reactive aluminum. To further improve reaction kinetics, the water is generally heated to above 50 degrees C. The lack of other trace gases like S, CO, or CO2, which can bind with the catalysts used in fuel cells, is also an advantage over hydrogen from steam reforming natural gas. This gives a significant advantage over other Al-water split systems (ANDERSEN, 2004) which rely on sodium hydroxide additives. In general, the production of hydrogen from an aluminum-water reaction requires that the protective oxide layer be efficiently and continuously removed, either by use of high temperature in 2  molten systems or elevated pH regions, for example using sodium hydroxide. Both of these methods increase the danger to users from inadvertent leaks or spills.  The  advantage of the milled system is its lower operating range (<90oC) and its ability to proceed without corrosive additives. All materials and byproducts can be safely handled without special procedures. Other systems using different powders have been proposed such as freshly ground iron powder (WERTH, 2000). However, Werth is still using relatively high temperature steam (250oC), and the iron must be freshly ground to enhance its reactivity. In contrast the aluminum-oxide mixture can be stored for months with only minor reduction in hydrogen formed and reacted in water at 50oC (CZECH, 2006). Thus the initial focus of this thesis is investigating the mechanism responsible for the increased corrosion observed after milling of aluminum powder. Electrochemical polarization and hydrogen capture tests are used to study changes in solution pH and temperature during the reaction, as well as calculate the activation energy for corrosion of the ball milled powder. Changes in the surface area of the milled powders are measured by BET. While these tests lead to the conclusion that it is the change in surface area with milling that plays the major role in determining the extent of corrosion for ball milled powders, it was subsequently discovered that alumina particles act to accelerate aluminum corrosion when in physical contact without the need for milling. In the second part of the thesis, data are produced for the rate and overall extent of aluminum corrosion when mixed with alumina powder. Two models are proposed to explain these data and tests run to test the predictions of the models.  3  Chapter 2 Literature Review This review is comprised of three main sections. The first section discusses the oxide and hydroxide layers which cover an aluminum surface after exposure to air and when aluminum reacts in water. Aluminum is a widely used material because these layers generally provide excellent corrosion resistance, and has been investigated for many years with regards to increasing its corrosion resistance in commercial alloys. Also briefly discussed is the production and recycling of aluminum which affects the economic viability of hydrogen generation from aluminum corrosion compared to other methods of producing and storing hydrogen. Only with the recent push to a hydrogen economy has the accelerated corrosion to generate hydrogen been strongly pursued. The second section of this review summarizes the recent work on aluminum to produce hydrogen. This includes melting and mechanical milling of alloying elements that prevent passivation, ball milling to increase surface area, the use of high temperature reactions or strongly alkaline solutions, and the addition of oxide in a sintering process to weaken the protective layer. Very little work has been done that directly addresses the focus of this thesis, but this section attempts to outline and compare all of the closest competing technologies. The final section of this review covers the various methods of hydrogen production and storage not involving aluminum to establish the commercial relevance of using aluminum as a hydrogen source.  4  2.1 Aluminum and Its Corrosion Properties in Water Aluminum is a silvery white metal that is not soluble in water under normal conditions. It is the most abundant metal in the Earth’s crust and the third most abundant element after oxygen and silicon, comprising about 8.3% by weight of the Earth’s solid surface (GREENWOOD, 1997). Aluminum is the most widely used metal after iron (HETHERINGTON, 2007) due to its low density (1/3 that of iron) and its ability to resist corrosion because a thin surface layer of aluminum oxide forms when the metal is exposed to air which prevents further oxidation. Almost all metallic aluminum is produced from the ore bauxite (AlOx(OH)3-2x) which occurs as a weathering product of low iron and silica bedrock in tropical climatic conditions (GUIBERT, 1986).  2.1.1 Production and Recycling of Aluminum Hans Christian Orsted first produced an impure form of aluminum in 1825 by reacting aluminum chloride with potassium amalgam (GRJOTHEIM, 1977). Friedrich Wohler is credited with isolating aluminum by reacting aluminum chloride with potassium vapor in 1827 (CHADWICK, 1958). The first commercial production was developed in 1854 by Henri Deville and then in 1886 the standard industrial method of reducing alumina was discovered by Charles Martin Hall and Paul Louis Toussaint Heroult working independently. The Hall-Heroult process became feasible with the availability of electric power and is the basic electrolytic process used today (CHADWICK, 1958). While aluminum in its native form has been reported in Russia, China, and in lunar soil (FRANK, 1997), it is so rare that all commercial production is done by refining 5  aluminum ores, primarily bauxite. “Bauxite describes sedimentary rocks that contain economically recoverable quantities of the aluminum minerals gibbsite, boehmite and diaspore” (FRANK, 1997). In 2003, six countries accounted for 83% of the world’s Bauxite production; Australia (38%), Guinea (11%), Jamaica (9%), Brazil (9%), China (8%), and India (6%), (PLUNKERT, 2004). Four or five tons of bauxite are used to produce 2 tons of alumina since a typical bauxite ore is usually 30-60% gibbsite and boehmite, with the remainder being oxides of Fe, Si, Ti, and Ca along with other minor constituents (AUTHIER-MARTIN, 2001). Roughly two tons of alumina are needed to produce one ton of aluminum and the cost of alumina at U.S. smelting facilities ranged from 265-700$/metric ton in 2006 (PLUNKERT,2007). Aluminum electrolysis with the Hall-Heroult process consumes a lot of energy, with a worldwide average specific energy consumption of approximately 15 kilowatthours per kilogram of aluminum (VAUGHAN, 2007). Theoretical power requirements for electrolysis of aluminum in a Hall-Heroult cell were calculated to be 6 kWh/kg Al (CHOATES, 2003), but state of the art plants are still running around 12.8 kWh/kg Al. In addition to the electricity used, the process also emits CO2, CO, CF4, and C2F6. Estimates by (RICHARDS, 1994) calculate that the aluminum industry could be contributing up to 1.5% of the total CO2 sourced from fossil fuels in the world. This is important to consider because much of the push behind using hydrogen in fuel cells to produce electricity is predicated on the lack of greenhouse gasses and certain cost penalties are acceptable for “green” energy. As will be discussed further in the third section, hydrogen from aluminum is not as “green” as other sources. One way to avoid this CO2 penalty is to use recycled aluminum, particularly that which is headed for landfills. 6  Recycling Since the late 1960s, the growing use of aluminum beverage cans has made the recovery of aluminum via recycling an important part of the aluminum industry. In 1995, 33% of the Reynolds Metal Company’s Aluminum was from recycled feed (OYE, 1999). Recycling aluminum requires 96% less energy than direct production from ore (HABASHI, 1997). Recycling also requires much less water than primary production which uses water to produce alumina. In 2001, approximately 11 billion liters of water were used to make aluminum cans that were not recycled (GITLITZ, 2002). In 2003, the scrap Al market was about 45% post-consumer scrap (CHOATE, 2005). In the US, about 22% of all primary aluminum goes toward making aluminum cans, and in 2001 an estimated 50.7 billion cans were not recycled which is the energy equivalent of 16 million barrels of crude oil (GITLITZ, 2002). This highlights the benefit of establishing additional uses for aluminum scrap.  2.1.2 Hydroxide Formation During Aluminum Corrosion When aluminum is immersed in water the corrosion at the metal/electrolyte interface is an electrochemical reaction involving the transfer of charge from different parts of the surface acting as the anode and cathode. At the anodic areas the metal is oxidized in a dissolution reaction: Eq. 1  Al → Al3+ + 3e-  At the cathodic areas dissolved oxygen is reduced or hydrogen is evolved: Eq. 2  O2 + 2H2O +4e- → 4OH-  Eq. 3  2H+ + 2e- → H2 7  There are also chemical reactions present that do not involve a charge transfer such as the precipitation of metal hydroxides: Eq. 4  Al3+ + 3OH- → Al(OH)3  The oxide film on the aluminum typically forms a barrier for electronic and ionic charges which protects the metal from oxidation and dissolution. For the past 40 years one of the principal tools used by corrosion researchers to predict the behavior of metals in solution has been the Pourbaix diagram, which defines regions of active corrosion, immunity, and passivation with respect to pH and potential. There are two types of passivity, thin film and thick film, which are seen for metals (KELLY, 2003). Thin film passivity can be defined as: “A metal is passive if, on increasing its potential to more positive values, the rate of dissolution decreases, exhibiting low rates at high potentials.” (KELLY, 2003) Aluminum is an example of the second type of passivity, thick film defined as: “A metal is passive if it substantially resists corrosion in an environment where there is a large thermodynamic driving force for its oxidation.” (KELLY, 2003) The descriptions ‘thin’ or ‘thick’ refer to which layer of a bi-layer film coating the metal controls the corrosion properties. The passive film will consist of a thin (1-3 nm) homogeneous primary passive film known as the barrier layer (bl) situated between the metal surface and the thicker porous precipitated upper layer film or outer layer (ol) (MACDONALD, 1990, 1992). See Figure 1.  8  Primary  Porous, Precipitated  Passive  Metal  Upper Layer Film  Film Mχ+  M (OH)δ  (Transmission)  Film Growth  Solution δH+  Mδ+ (Dissolution of Film) Oxygen Injection  Mδ+ H2O  Figure 1 Schematic of processes that lead to the formation of bi-layer passive films on metal surfaces. (Redrawn from MACDONALD, 1992)  From the potential-pH equilibrium diagram for the aluminum-water system at 25oC, we can see that the immunity region for aluminum is very electronegative (-1.8V SHE) (POURBAIX, 1966). For neutral pH water 4<pH<9 there is a stable hydroxide formed for all potentials above -1.8V SHE. Below a pH of 4, aluminum oxide or hydroxide will dissolve as trivalent Al+++ ions: Eq. 5  Al(OH)3 + 3H3O+ → Al3+ + 6H2O  Eq. 6  Al2O3 + 6H3O+ → 2Al3+ + 9H2O  Water molecules will bind to the Al3+ to form stable hexahydrated aluminum ions [Al(H2O)6]3+, because of the small radius and high charge of the aluminum cation. (LAWSON, 1974), (DOWNS, 1993) In alkaline solutions the aluminum oxide or hydroxide dissolves as aluminate ions AlO2or Al(OH)4-. Eq. 7  Al(OH)3 + OH¯ → Al(OH)4¯  Eq. 8  Al2O3 + 3H2O + 2 OH¯ → 2 Al (OH)4¯ 9  The corrosion resistance of aluminum in general is a function of the passivating nature of the hydroxide films that are formed. It is important to remember that Pourbaix diagrams only show the most thermodynamically stable hydroxide (for example Al2O3.3H2O), (see Figure 2 which was generated using the HSC Chemistry 5 software from Outokumpu Research); kinetics of film growth can often result in the presence of meta-stable films. The dissolved product can vary from [Al(OH)]2+ to more complicated de-protonated hydroxo-complexes such as: [Al(OH)(H2O)5]2+ [Al(OH)2(H2O)4]+ , [Al(OH)3(H2O)3] or [Al(OH)4(H2O)2]¯ (TOMCSANYI, 1989). In near neutral pH pure water systems, aluminum is reported to form various aluminum hydroxides depending on water temperature and pressure all of which form a passivating layer that grows to at most a couple of microns thickness before the reaction stops (ALWITT, 1976). For temperatures below 100oC in pure water there are two main hydroxides formed. Eq. 9 shows the predominate reaction observed for water temperatures above 40oC (DENG, 2008), although Hart (1957) reports that AlOOH (Eq. 10) is still formed initially. Eq. 9  2Al + 6H2O → 3H2 + 2Al(OH)3  Eq. 10  2Al + 4H2O → 3H2 + 2AlOOH  At temperatures below 60oC and at 1 atm. the first film grown is an amorphous AlOOH next to the metal surface, then orthorhombic AlOOH and finally a crystallized (bayerite) Al(OH)3 forms next to the solution (HART, 1957). Al(OH)3 forms more rapidly at lower temperature and in hot water (above 70oC) may not form for many hours. In these systems Bunker (2002) reports only a hydration front of AlOOH. 10  Figure 2. E-pH diagram for Al-water at 25oC.  Aluminum alloys left in sea water showed average weight loss of less than 100 mg/m2 after periods of up to 10 years (DAVIS, 1999). Davis also reports that the slight reaction between aluminum alloys and high-purity water ceases almost completely within a few days and after the protective oxide film forms the amount of metal dissolved becomes negligible, but at temperatures above 200oC a high-purity aluminum sheet (1.62 mm thick) disintegrates within a few days to form aluminum oxide. For elevated temperature high purity water, the best resistance is found from aluminum-nickel-iron alloys which can withstand temperatures up to 315oC (DAVIS, 1999). Both the rate and type of attack are sensitive to solution temperature. The rate of corrosion will increase with temperature until 60-70oC with the dominating corrosion tendency to be pitting. At higher temperatures the pitting depth sharply decreases and up  11  to 150oC the propensity of pitting corrosion progressively disappears (VARGEL,2004). Globally, aluminum’s corrosion resistance in freshwater is better at a temperature between 70 and 100oC than at room temperature. Above 150oC intercrystalline corrosion develops rapidly at the surface and above 250oC a sharp increase in corrosion rate is seen with the kinetics changing to a linear law (VARGEL, 2004).  2.1.3 Growth and Dissolution of Passive Oxide Films One possible model describing the formation and growth kinetics of the passive film postulates film growth due to the transport of metal cations (Al3+) across the oxide film to the film/solution interface where they react with the electrolyte. This migration of the cations is assisted by the high electric field created by oxygen atoms on the outer amorphous oxide layer (SHIMIZU, 1991). Oxide growth occurs at both the oxide/solution interface (by Al3+ metal ion migration outwards) and the metal/oxide interface (by O2- ions migration inward) (SHIMIZU, 1999) (DIGNAM, 1962). The ratelimiting step for the film growth is the emission of metal cations from the metal (CHAO, 1981a,b). This model describes the growth of a precipitated outer layer on the surface of aluminum coated with a native oxide (Figure 1).  Point Defect Model The work done in this thesis looks at corrosion of aluminum in water below 100oC which is considered to be dominated by pitting (VARGEL, 2004). Before a discussion on pitting and the growth and dissolution of passive films, it is necessary to  12  consider kinetics, and this brings up the point defect model (PDM). Asselin (2007) writes: “The PDM is the leading model to describe the kinetics of passivity. It was developed by Macdonald and co-workers in the early 1980s. There are three fundamental, well established, experimental observations which have framed the PDM and they are, very generally [Macdonald 1992]: A. As already mentioned, passive films are bi-layered and passivity is attributed to the inner bl which is a highly defective, defect semiconductor in which vacancies act as the electronic dopants [Macdonald 1999]. These vacancies are annihilated and generated at the interfaces – the metal (m)/barrier layer (bl) interface and the bl/solution (s) interface which are separated by a few nanometers. B. Under steady state conditions, the bl thickness and the logarithm of the currents both vary linearly with applied voltage C. The bl grows into the metal while the ol grows out from the metal surface”  13  Metal  Barrier Layer (MOχ/2)  (1) m + VMΧ’ → MM + VM + Χe’  Precipitated Outer Layer / Solution (3) MM → Mδ+ (aq) + VMΧ’ + (δ-Χ)e’  (4) V-O + H2O → OO + 2H+  (2) m → MM + (Χ/2) V-O + Χe’  (5) MOΧ/2 + ΧH+ → Mδ+ + Χ/2 H2O + (δ-Χ)e’ VMΧ’ V-O  Figure 3 Schematic of physico-chemical processes that occur within a passive film according to the point defect model. m=metal atom, MM=metal cation in cation site, OO=oxygen ion in anion site, VMΧ’=cation vacancy, V-O=anion vacancy, VM=vacancy in metal phase.  (Redrawn from  MACDONALD, 1999)  From Figure 3 we can make the following statements. During film growth cation vacancies produced at the film/solution interface travel to the metal/film interface and are consumed. Anion vacancies are formed at the metal/film interface and are consumed at the film/solution interface. Reaction (1) where a cation vacancy is eliminated by injection of a lattice cation, reaction (3) where a cation vacancy is created by oxidative dissolution, and reaction (4) where an anion injection eliminates an oxygen vacancy, are lattice conservative processes. The growth and dissolution of the barrier layer which result in movement of the barrier layer boundaries are thus governed by reaction (2) the injection of a lattice cation and creation of an oxygen vacancy, and reaction (5) the chemical  14  dissolution of oxide. A steady state film thickness implies that the reaction rates of (2) and (5) are equal. (MACDONALD, 1999)  Pitting Pitting corrosion of aluminum in water develops at preferential sites where the natural oxide film is less resistant because of heterogeneous features such as impurities or defects related to localized thinning or rupture of the natural oxide film. These sites are anodic with respect to their vicinity, and corrosion pits can develop. While pit initiation is localized, the kinetics may be controlled by the transport of species through the passive oxide. In the point defect model, anodic polarization drives aluminum vacancies through the film to the Al:Al2O3 interface and oxygen vacancies are driven to the Al2O3:solution interface. This is equivalent to moving Al3+ ions to the solution interface, and O2- is driven to the metal surface (BUNKER, 2002). When the flux of aluminum vacancies reaching the Al:Al2O3 interface exceeds the rate at which Al is oxidized to Al3+ to fill the vacancies, the vacancies can nucleate and grow localized voids. These voids weaken the overlaying oxide and serve to initiate pits. See Figure 4 from (BUNKER, 2002). Once a pit is formed it will be anodic with respect to the surroundings and see enhanced corrosion of the aluminum until a passive film is re-grown.  15  H2O  Al3+ Solution Al2O3  O2VAl  Al  VO  Vacancy condensates  Figure 4 Point defect model for the pitting of Al. Under anodic polarization, aluminum ions are driven to the Al2O3:solution interface and into solution, while aluminum vacancies are driven to the Al:Al2O3 interface where they aggregate to form voids as precursors to pits. Oxygen ions are also driven to the Al:Al2O3 interface, while oxygen vacancies migrate to the Al2O3:water interface to be filled by water molecules. (Redrawn from BUNKER, 2002)  Pitting can also be initiated by anions that penetrate into the defects. Chlorides have the highest propensity to penetrate into the natural oxide layer since they are small and very mobile (BRITTON, 1930). They can then decrease the resistance of the film by replacing oxygen atoms in the lattice; this creates defects in the Al2O3 that increase the concentration of aluminum vacancies in the film, which facilitates the release of aluminum atoms to diffuse into the water (HEINE, 1965) (BUNKER, 2002). Ovari (1988) states the general form of aluminum corrosion in neutral solutions is pitting 16  corrosion and is most commonly seen in the presence of Cl- ions. Cl- ions can also facilitate pitting by helping to maintain an acidic environment within the pit electrolyte. Locally dissolved and hydrolyzed aluminum ions produce an acidic environment (pH 1 to pH 3) and the Cl- ions prevent repassivation by maintaining charge neutrality (MCCAFFERTY, 2003), (SZKLARSKA, 1992). AlCl3, Al(OH)2Cl and Al(OH)Cl2 salts were found within Al pits by (SZKLARSKA, 1999) and their precipitation on the bare Al surface may prevent locally the metal from passivating (BECK, 1984). There is a pitting potential above which pitting will occur all over forming a roughened surface. Gross pitting occurs when substances inhibit tunnel nucleation on the free surface, like OH- ions in alkaline or sulfate ions in neutral solutions (BAUMGARTNER, 1990). Unlike the case for Cl- ions, aluminum and its alloys are relatively insensitive to the concentrations of oxygen present in most aqueous solutions (UHLIG, 1948). In general, high concentrations of dissolved oxygen tend to stimulate attack somewhat, especially in acid solutions, but they also contribute to the formation of the natural oxide layer. Vargel (2004) reports that the corrosion resistance of aluminum does not differ significantly between aerated and deaerated water. The presence of oxygen leads at most to a more localized corrosion, but has no influence on pitting depth. DiBari (1971) reports that dissolved oxygen has a negligible effect on the rate of corrosion of high purity aluminum unless cathodic potentials are applied. For this reason, nitrogen purging of the solution to remove oxygen was performed when measuring cathodic slopes to determine hydrogen exchange current densities, but not for general hydrogen evolution tests.  17  Uniform Corrosion Model Bunker (2002) writes that “while defect migration may be required for pitting, quantitative predictions of initiation kinetics are not consistent with published literature regarding how easy it is to create and move aluminum or oxygen vacancies in aluminum oxide.” Studies by Le Gall (1995) indicated negligible (D=10-80cm2/s) aluminum and oxygen vacancy diffusion in sapphire, many orders of magnitude less than those consistent with known pitting kinetics (D=10-20cm2/s) (BUNKER, 2002). This indicates that the native oxide must be substantially different than bulk α-Al2O3. One explanation is that the amorphous oxide has a structure similar to highly defective γ-Al2O3 with a spinel structure vAl8O12 (where v represents an aluminum vacancy) (SOHLBERG, 2001). Six Al cations occupy octahedral sites and two occupy tetrahedral sites with the Al vacancy also involving a tetrahedral site. Vacancy transport among the available tetrahedral sites is facilitated since one-third are vacant. Another possibility is the defects responsible for pitting are created in the film from environmental exposure. Electrical conductivity of passive oxide films on Al is increased by a factor of over 106 upon exposure to water (SULLIVAN, 1999), and Sohlberg (2001) suggests that water can be incorporated into the oxide lattice while preserving the spinel structure for Al2O3-nH2O compositions up to n=0.6. Studies by Bunker (2002) of aluminum corrosion in hot water show the water reacting with the native oxide to create oxyhydroxide and hydroxide phases on the aluminum surface with substantially different properties than the native oxide. This hydration of thin oxide films during corrosion leads to the uniform corrosion model in Figure 5.  18  Al2O3 + H2O → 2AlOOH Hydration front  Al2O3 Stage 1: Induction (no growth)  Al  2Al + 4H2O → 2AlOOH + 3H2 AlOOH  OH-  Al2O3 Al  H2 Stage 2: Rapid AlOOH Growth  Figure 5 Uniform Corrosion Model for Al in hot water. (Redrawn from BUNKER, 2002)  In the uniform corrosion model (Figure 5) there is an initial induction time with no apparent corrosion during the hydration (Eq. 11) of the native oxide. Eq. 11  Al2O3 + H2O → 2AlOOH  Hydration modifies the oxide lattice (Al-O-Al bonds are disrupted to form Al-OH species), and lowers its effective cross-link density by replacing O2- ions with more mobile species which leads to enhanced diffusion and rapid film growth. These species include protons which can hop from oxygen to oxygen, hydroxide ions which migrate between oxygen vacancies, and water molecules which fill interstices in the structure that are opened up during hydration. Hydroxide ions are anionic defects that migrate to the Al:Al2O3 interface under an anodic potential. Extensive hydration eventually produces a pseudoboehmite layer (AlOOH) that, after an induction time lasting minutes to hours, 19  depending on the temperature, begins a period of rapid growth (BUNKER, 20002). This is followed by a period of slower growth of Al(OH)3 on top of the pseudoboehmite.  Hydrogen Bubble Model of Pitting The hydrated oxide is predominantly AlOOH except at the Al:Al2O3 interface, where the hydroxide is reduced by Al to regenerate the oxide and generate H2 (Eq. 12) (DENG, 2008). Eq. 12  6AlOOH +2Al → 4Al2O3 + 3H2  If the generation of H2 bubbles at the interface is rapid compared to the slower diffusion of hydrogen through the Al or oxide, then the bubbles may rupture the overlaying film creating pits or blisters (DENG, 2007, 2008). After the film is broken Al and water make contact and react (Eq. 13). Eq. 13  Al +2H2O → AlOOH +3/2H2  Once all of the water next to the aluminum surface has reacted the surface is now covered again by a layer of AlOOH which again reacts by Eq. 12 to form a new protective oxide layer and then utilizes Eq. 11 to start the cycle over again. In this “hydrogen bubble” model for pitting, the pitting potential corresponds to the voltage required to drive hydroxide ions to the Al:Al2O3 interface fast enough to promote bubble nucleation. Under ambient conditions OH- (or H2O molecules) are the mobile species in the film, rather than H+, O2-, or Al3+ ions (BUNKER, 2002).  High pH Environments In an alkaline solution there is movement of OH- ions through the native surface oxide toward the aluminum and an electrochemical film formation at the aluminum/film 20  interface along with chemical film dissolution due to OH- attack at the film/solution interface (PYUN, 2000). The anodic dissolution of Al in alkaline solutions comes from direct metal dissolution of Al (Eq. 14); Eq. 14  Al + 4OH¯ → Al(OH)4¯ + 3e¯  and indirect metal dissolution by consecutive oxide film formation (Eq. 15) and dissolution (Eq. 16). Eq. 15  Al + 3OH¯→ Al(OH)3 + 3e¯  Eq. 16  Al(OH)3+ OH¯ → Al(OH)4¯  The corrosion rate in alkaline solutions is thought (PYUN, 2000; EMERGUL, 2000; ARMSTRONG, 1996) to be governed by the formation of Al(OH)4- at the metal/oxide interface. In weak alkaline solutions, the cathodic reaction for Al corrosion is primarily the reduction of water to hydrogen (Eq. 17).  Eq. 17  2H2O + 2e¯ = H2 + 2OH¯  The reaction rate is strongly potential dependent (EMERGUL, 2000), but independent of pH and film thickness (ARMSTRONG, 1996). When the electrons e- react with water at the metal/film interface, the formation of OH- ions increases the local alkalinity causing hydration, dissolution and film structure changes (EMERGUL, 2000). Impedance measurements by Emregul (2000) indicate there is an anhydrous surface film 2-3 monolayers thick which is porous and has an accumulation of hydrogen gas molecules in the pores. On top of this lies a hydrated layer with the steady-state thickness dependent on hydroxide ion concentration and potential.  21  Mechanical Deformation and Microstructure The corrosion behavior of aluminum and its alloys can be affected by the number and type of constituents, grains size and grain orientation, defect concentration and type, as well as the metallurgical treatment. Constant stress below the yield strength does not affect the rate of attack of most commercial aluminum alloys under normal condition of use (MEARS, 1945). Leth-Olsen (1998a,b) and Afseth (2001,2002) report that plastically deformed aluminum corrodes faster in water solutions than strain-free aluminum. Afseth (2001, 2002) describes a drastic loss of corrosion resistance after annealing an aluminum alloy AA3005 with a highly deformed near-surface region approximately 1 um thick caused by shear deformation during rolling. This increased corrosion was attributed to a resulting microstructure where the near-surface region contains a higher density of fine intermetallic particles which increased the number of potential corrosion initiation sites on the metal surface, and lower solid solution levels of the more noble alloying elements than the underlying bulk aluminum alloy. This would not be the case in a system where pure aluminum is used.  Summary In conclusion, corrosion of aluminum in many environments and forms has been studied extensively for many years. There are several models to describe the growth of hydroxide layers on aluminum in water, and while there is agreement that the formation of pits plays a central role in overcoming passivation and promoting corrosion, the mechanism behind film growth and pit initiation is dependent on diffusion rates of ionic species and vacancy migration. In general there has been little focus on aluminum 22  powders left in water because of a lack of practical application. Many of the known corrosion issues deal with impurities or added elements to the metal or the solution. For pure aluminum in de-ionized water at moderate temperatures, corrosion is perceived to proceed slowly and in general to be prevented by the formation of a passive hydroxide layer.  2.2 Hydrogen Production from Aluminum It is well known that alkali metals Li, Na, and K react in the presence of water producing hydrogen, but the metals are relatively expensive and the byproducts of reaction are strong alkalis which can be difficult to deal with in portable applications (DENG, 2008). The use of aluminum to produce hydrogen has gained interest in recent years because aluminum’s low density and tri-valence state result in high effective hydrogen content per mass of reactants, and the by-products have minimal environmental impact (WANG, 2008). There is also an established process and infrastructure to recycle the reacted aluminum hydroxide back to aluminum through the Hall-Heroult process. The lack of such a recycle process is a major drawback of some of the competing hydride storage systems such as NaBH4. However the strong affinity for oxygen creates a protective oxide layer that becomes the major hurdle to realizing continuous hydrogen generation through aluminum corrosion.  2.2.1 Reaction of Al with Alkaline Aqueous Solution Aluminum is attacked in acid and alkaline solutions as soon as the oxide film is eliminated. The dissolution of the oxide films is faster in alkaline solutions (POURBAIX,  23  1966) so this is the primary method adopted by researchers looking to generate hydrogen for portable fuel cells (BELITSKUS, 1970), (SOLER, 2007a,b), (HIRAKI, 2007), (JUNG, 2008). Hydroxide ions (OH-) in strongly alkaline solutions break down the protective oxide layer on the aluminum surface forming AlO2- and causing the aluminum to be readily dissolved even at room temperature with hydrogen generation resulting (Wang, 2008). Sodium hydroxide (NaOH) is the most commonly used alkali, and patents for its use to generate hydrogen from aluminum have been filed multiple times over the last 100 years (BRINDLEY, 1909a,b), (GILL, 1955), (CHECKETTES, 1998), (ANDERSEN, 2003a,b), (ANDERSEN, 2004), (HIRAKI, 2007). The initial reaction (Eq. 18) consumes NaOH and produces 2.8 wt% hydrogen (DENG, 2008), Eq. 18  2Al + 6H2O + 2NaOH → 2NaAl(OH)4 + 3H2  but when the NaAl(OH)4 concentration exceeds the saturation limit it undergoes a decomposition reaction (Eq. 19) that produces a precipitate of Al(OH)3 and regenerates the alkali (BELITSKUS, 1970), (SOLER, 2007b), (LI, 2002). Eq. 19  NaAl(OH)4 → NaOH + Al(OH)3  The overall reaction of the Al is given by Eq. 20. Eq. 20  2Al + 6H2O → 2Al(OH)3 + 3H2  However, the use of NaOH to produce hydrogen has the disadvantage that the NaOH is extremely corrosive and not suitable for hydrogen production in vehicles or in household power systems (ANDERSEN, 2003a). The kinetics of the aluminum-water reaction with the addition of alkalis such as NaOH have been studied by many researchers and the activation energy for this reaction 24  has been reported between 42.5 and 68.4 kJ/mol (HIRAKI, 2005), (ALEKSANDROV, 2003), (HU, 2003), (ZHUK, 2006). Studies to optimize the reaction found an optimum temperature of 70-90oC and a NaOH concentration of 5.75 M, to lead to the highest controllable rate of hydrogen production with the minimal mass of NaOH and H2O (STOCKBURGER, 1992). If aluminum alloys are used the composition will play a role in the hydrogen yield and in a study of 16 different alloys, Al/Si was found to show the highest initial hydrogen generation rate (SOLER, 2007a). Martinez (2005) studied the effect of the molar ratio of NaOH/Al on hydrogen generation at 23oC from soft drink can waste. Higher ratios increased the initial hydrogen generation rate, but not the total volume of hydrogen produced. Martinez (2007) also tied the aluminum can-based hydrogen production setup with a proton exchange membrane fuel cell (PEMFC) and compared it to an electrolyzer driven by solar energy and concluded that aluminum cans have a better performance (WANG, 2008). Other hydroxides have also been studied for reacting Al. Potassium hydroxide (KOH) shows a synergistic effect on hydrogen liberation performance when increasing temperature and base concentration at the same time, but a drawback to this method is the consumption of KOH due to its reaction with CO2 in air (SOLER, 2005). A recent study compared the hydrogen generation performance of three different hydroxides: NaOH, KOH, and Ca(OH)2 with NaOH giving the fastest aluminum consumption (SOLER, 2007a). Combining sodium borohydride (NaBH4) with aluminum in alkaline solutions enhances both hydrogen production rate and conversion yield (SOLER, 2007b). This was attributed by Soler (2007b) to both the pH increase caused by the hydrolysis of NaBH4 and the catalytic effects of some aluminum alloys on the hydrolysis of NaBH4. However, 25  NaBH4, a complex hydride made from borax, is quite expensive which makes the hydrogen not economically viable compared to alternative sources.  2.2.2 Aluminum – Water Reaction in Neutral Solutions Aluminum can also react in plain water, but because of passivation the rate is generally too limited for practical use. Cutting, drilling, and grinding of the aluminum can be used to remove the oxide layer and expose fresh aluminum to the solution, however the hydrogen generation stops immediately after the machining stops due to rapid repassivation (UEHARA, 2002). For portable systems there would be energy lost to grinding, but this was not addressed in the paper.  Ball Milling Another approach is to use metal particles with small sizes to increase the specific exposed surface area (IIYIN, 2000), (IVANOV, 2001). One approach to producing fine metal powder is the use of high energy ball milling, but the size reduction of the particles is strongly dependant on the mechanical properties of the metals (GROSJEAN, 2005, 2006). In addition to changing the particle size, ball milling is reported to induce pitting corrosion by creating numerous defects and fresh surfaces on metals (GROSJEAN, 2004, 2005, 2006), (FAN, 2007a,b, 2008). There are major differences between the work on ball milled systems reported by Grosjean and Fan, and that done here at UBC. Grosjean (2004, 2005, and 2006) was ball milling Mg, which has different mechanical properties during milling and, because it has a valence state of 2, it will release less hydrogen per mole of Mg reacted and suffer from both a wt% hydrogen yield and a cost basis for the  26  hydrogen compared to an aluminum system. Fan (2007a,b, 2008) was ball milling Al-Bihydride or Al-Bi-salt mixtures and attributed the improved aluminum corrosion to three factors: 1. The additives used (MgH2, CaH2, LiCl, MgCl2, KCl, etc.) help to decrease the mean size of the mixture particles. 2. The exothermic dissolution of the salt additive will increase solution temperature. 3. The hydrolysis of the additives offer conductive ions that assist the microgalvanic cell of the Al-Bi composite. This work by Fan is closely related to the work done by Czech (2006) at UBC with several differences worth noting. Factor (1) is believed to be critical in this thesis for increasing the Al surface area, Fan does not report the change in size for the Al particles or determine to what extent the salt is responsible for the increased corrosion. Factor (2) was also reported by Czech and helps decrease the induction period before reaction begins, but has not been shown to increase overall yield. Factor (3) relates to the Al-Bi micro-galvanic cell that Fan is reporting to be critical to the increased corrosion and the salts assist that corrosion. Fan also reports that Al milled with 20% NaCl does not produce hydrogen at all. This is directly contradicted by Czech and data from this thesis. Notwithstanding these discrepancies, the system presented by Fan would suffer a weight penalty because of additives to the aluminum and increased cost.  Ball Milling with Oxides Czech (2006) expanded upon the initial discovery by Chaklader (2002) with optimized milling conditions. Many additives (Al2O3, SiO2, NaCl, KCl, etc.) were found 27  to work to prevent Al galling and decrease particle size. Reactivity of the aluminum was found to increase with milling time up to one hour, with decreasing returns after 15 minutes. After one hour increased oxidation of the aluminum began to decrease hydrogen yield. Other metals were not used as additives so no galvanic coupling contributed to the aluminum corrosion. The use of Al2O3 in aluminum matrix composites has been reported to show increased corrosion around the oxide inclusions, because of de-alloying in the nearby region and formation of a galvanic couple with the more noble regions (SALAZAR, 1999). This would not be expected for pure aluminum. Milling with salts showed decreased induction times and faster hydrogen generation compared to oxides, and the best wt% yield was from ‘washed’ aluminum systems where the salt used in milling is removed by repeated rinsing with cold water. There were 5 primary conclusions from this work (Czech, 2006): 1. Milling in the presence of a second phase is critical for increasing the hydrogen evolution and milling parameters must be optimized. 2. The second phase can be removed after milling without loss of activity. 3. Structural defects and microstrains induced in Al are not the reason for increased reactivity in water. 4. Higher temperatures favored the reaction. pH shifts from neutral in either direction within the range of 3.5 to 9 increase the induction time and delay the H2 generation. A pH shift upward was observed during every reaction and is thought to promote the aluminum corrosion. 5. Local temperature increases at the reaction site are expected and lack of repassivation of the active metal surface is attributed to local alkalinization. 28  High Temperature High temperatures can be used to react or burn aluminum with water vapor under pressure and this reaction (Eq. 21) has been studied for use in propulsion systems (IVANOV, 1994), (INGENITO, 2004). Eq. 21  2Al + 3H2O → Al2O3 + 3H2  To facilitate this reaction, nano-scale aluminum powders are used along with watersoluble polymers to inhibit the water from evaporating during combustion (IVANOV, 2000), (SHAFIROVICH, 2006). Work by Diakov (2007) and Shafirovich (2007) also show that this highly exothermic combustion enhances hydrogen production from NaBH4 without the need for expensive precious metal catalysts and increase the hydrogen yield over hydrolyzing NaBH4 by itself (RICHARDSON, 2005). The high temperatures and extremely rapid hydrogen release in these systems would, of course, present engineering challenges to safely incorporate into portable consumer products.  Alloying with Other Metals Wang (2008) states: “Another widely used strategy to enhance the hydrogen generation is to form a corrosion cell by coupling two or more dissimilar metals or metal alloys together in the presence of electrolytes (GROSJEAN, 2005, 2006), (BLACK, 1976), (SERGEV, 1980, 1981), (IZURU, 2003).” Kravchenko (2005) and Fan (2007a,b) report that Al alloys doped by approximately 20 wt% with Ga, In, Bi, Sn, etc. would continuously react with water with the secondary metal acting as the cathode in a microgalvanic cell with Al as the anode. Woodall (2007) reports that the gallium in an 29  Al-Ga alloy hinders the formation of a skin on the aluminum surface after oxidation which allows the corrosion to continue to completion. The alloys can be made by melting or mechanical milling, with milling being preferred because it avoids unnecessary vaporization loss of low-melting point metals during the alloying process. It also avoids air pollution and creates more defects on the alloy surface (FAN, 2007a). However, these alloys are not readily available and they are unstable. Storage in liquid nitrogen is necessary to retain their chemical activity (KRAVCHENKO, 2005). They are also expensive with the price of metal Ga about 200 times that of Al (DENG, 2008).  Oxide Modified Al Powder Deng (2005a,c) proposed the concept of “ceramic oxide surface modification of metal Al”, in which Al and Al(OH)3 are ball milled in an ethanol solution to uniformly mix them. The mixture is then dried and cold pressed to form green billets which are sintered at 600oC. During heating the Al(OH)3 decomposes into a γ-Al2O3  phase  (DENG, 2001a,b), and results in a γ-Al2O3-modified Al powder (GMAP) (DENG, 2008). The reaction of GMAP increases with water temperature, and no change in pH was observed. At 50oC, over 60% of the available hydrogen was evolved in 6 hours for GMAP mixture of 37 wt% and 70 wt% γ-Al2O3, and roughly 50% hydrogen was evolved for the 20 wt% γ-Al2O3  mixture (DENG, 2008). Deng (2007) has presented a  physicochemical model to explain the continuous reaction of GMAPs in pure water that is based on the uniform corrosion model presented by (BUNKER, 2002) and where the sintering with γ-Al2O3 is thought to weaken the native oxide coating on the particles and allow for easier blistering or rupture by hydrogen bubbles. While Deng does report using 30  alumina to encourage aluminum corrosion, there are several inconsistencies between his work and what was found in this thesis. First despite showing TEM pictures of the original Al powder that is much smaller (<1um) (DENG,2008) than that used in this work, Deng (2007) reports that pure Al powder does not react with water at temperatures up to 75oC. Not enough detail on initial powder was provided to determine differences with this work. Deng also does not report any interaction for alumina (that has not been milled and sintered) with aluminum and yet the reported hydrogen evolution rates are below that seen in this work. So while the proposed mechanism explaining the corrosion of Al powder seems sound, the need for milling and sintering seems lacking.  2.3 Hydrogen While hydrogen is the most abundant of the chemical elements and constitutes roughly 75% of the universe’s elemental mass (PALMER, 1997), it is relatively rare in elemental form on Earth existing as a diatomic gas H2 in the Earth’s atmosphere in a concentration of about 1ppm by volume. Hydrogen is mostly found in water and organic compounds where it is the third most abundant element on Earth, and will form compounds with most elements. The name was given in 1783 by Antoine Lavoisier after he and Laplace reproduced the earlier work (1766) of Henry Cavendish who produced hydrogen while experimenting with acids and mercury (EMSLEY, 2001). Although Cavendish is usually given credit for its discovery it had been produced and described in the early 1500’s by T Von Hohenheim from the mixing of metals and strong acids and again in 1671 when Robert Boyle reacted iron filings with dilute acids. (WINTER, 2007).  31  Hydrogen gas is highly flammable and will burn at concentrations as low as 4% H2 in air releasing 286 kJ/mol. However most of the interest in using hydrogen as an energy carrier comes from the ability of a fuel cell to electrochemically convert hydrogen and oxygen into water and produce electricity. The electrochemical nature of a fuel cell means that it is not limited by Carnot efficiencies that apply to combustion engines.  2.3.1 Hydrogen Production Hydrogen is made at large scale today (mostly from natural gas) for use in chemical processes such as oil refining and ammonia production (LARMINIE, 2003). About 1% of U.S. primary energy use (~ 5% of U.S. natural gas use) goes to hydrogen production for chemical applications (OGDEN, 1999). A variety of hydrogen production processes are commercially available today, including thermochemical methods, which are used to derive hydrogen from hydrocarbons, and electrolysis of water, during which electricity is used to split water into its constituent elements, hydrogen and oxygen. Future potential methods of hydrogen production involving direct conversion of sunlight to hydrogen in electrochemical cells or biological hydrogen production are being researched at a fundamental-science level.  Aluminum Water Split System The basic equation for the water split reaction is: Eq. 22  Al + 2H2O → AlOOH +1.5 H2  Expressed on a mass basis we have: Eq. 23  27 kg + 36 kg → 60 kg + 3 kg  32  From Eq. 23 we see that we need 9 kg of aluminum for every kg of hydrogen assuming complete reaction of available aluminum. As a matter of practice hydrogen production has typically been approximately 90% of theoretical so the minimum cost for each kg of hydrogen is the cost of 10 kg of aluminum plus the energy cost of milling and processing which have not been estimated at this point. Vaughan (2007) reports the cost of aluminum at 1.81 USD/kg which would give a cost of 18$/kg for hydrogen produced by the aluminum water split reaction. As we shall see, this is much higher than costs from other methods of hydrogen production. While the water split reaction results in no greenhouse gas emissions the production of aluminum is not clean. From Vaughan, “According to David Creber of Alcan, the industrial average CO2 production is 1.74 kg CO2 / kg Al produced.” Also, cryolite (Na3AlF6) is routinely oxidized during aluminium production causing CF4 and C2F6 gas emissions. The global warming potential of these C-F gases are 5,700 and 11,900 respectively. That is, every CF4 molecule will warm the earth to the same extent as would 5,700 CO2 molecules (U.S. Department of Energy, 2003). As CO2 equivalents, these emissions are on the order of 1.5 kg CO2-eq/ kg Al produced. While there are significant improvements in recent years which bring the combined total down to around 2 kg CO2-eq / kg of Al this still means 20 kg CO2 / kg H2. In addition to the gas emissions from aluminum ore refining, the process takes around 15 KWh/kg Al (VAUGHAN,2007) of electricity which may also have associated greenhouse gas emissions if not from clean sources. This means another 150 KWh/kg of H2. One kg of H2 is equivalent to 141. MJ (HHV) (OGDEN, 1999) which would equal 39.4 KWh, so the efficiency of energy produced/energy used is only about 33  26%. The use of recycled aluminum would avoid all of these emissions and would lower the cost to around 7$/kg H2, but would only be available for small scale use since aluminum is a heavily recycled metal due to its high processing cost.  Natural Gas Reforming Hydrogen is made thermochemically by processing hydrocarbons (such as natural gas, coal, biomass, or wastes) in high-temperature chemical reactors to make a synthetic gas or “syngas”, composed of H2, CO, CO2, H2O, and CH4. The syngas is further processed to increase the hydrogen content, and hydrogen is separated out of the mixture at the desired purity (OGDEN, 1999). Catalytic steam reforming of methane (the main component of natural gas) is a well-known, commercially available process and accounts for >90% of the hydrogen manufactured today in the United States (HEYDORN, 1994). The steam reforming reaction Eq. 24  CH4 + H2O → CO + 3H2  is endothermic and requires external heat input. Thus operating temperatures are usually 700oC - 800oC. After reforming, the resulting syngas is sent to one or more shift reactors, where the hydrogen concentration is increased via the water-gas shift reaction Eq. 25  CO + H2O → CO2 + H2  At this point the gas is mostly H2 (70%-80%) plus CO2, CH4, and small quantities of H2O and CO. Water can be beneficial in hydrogen for use in PEM fuel cells, but even small amounts of CO can poison the fuel cell by binding with the platinum catalyst, and its concentration must be kept lower than about 10 parts per million (LARMINIE, 2003). The hydrogen from steam reformers is typically purified with pressure swing adsorption 34  (PSA) systems or palladium membranes to produce hydrogen at <99.999% purity (OGDEN, 1999). The energy conversion efficiency [=hydrogen out (HHV)/ energy input (HHV)] of large-scale systems is between 75% and 80% (KATOFSKY, 1993). From Eq. 24 and Eq. 25 above, we can see that one CO2 molecule is released for every four H2 molecules produced, or 5.5kg CO2 per kg of H2. To make this process more environmentally friendly it has been suggested the CO2 could be sequestered underground in secure geological formations such as deep saline aquifers or depleted gas fields (WILLIAMS, R.H., 1998). The cost of hydrogen will depend on the cost of natural gas and to what extent economies of scale reduce the cost of steam methane reformer (SMR) plants. Hydrogen from local steam methane reformers could be produced, compressed to 41.4 MPa, and stored for delivery to 34.5 MPa vehicle tanks for a cost less than the wholesale cost of gasoline on an equivalent cost per mile basis (THOMAS, 2000). Solid hydrocarbon feedstocks such as biomass, coal, or wastes can be gasified at high temperature to produce a syngas, which can then be processed to produce hydrogen. In the early part of the last century before the availability of low-cost natural gas, coal gasification was the preferred method of hydrogen production in the United States, and it is still practiced in China and Europe (OGDEN, 1999). Conversion efficiency from biomass or coal to hydrogen is ~60%-65%, and the production cost is almost twice that of steam reforming (OGDEN, 1999).  35  Electrolysis of Water In water electrolysis, electricity is passed through a conducting aqueous electrolyte, breaking down water into its constituent elements, hydrogen and oxygen via the reaction Eq. 26  2H2O → 2H2 + O2  Commercially available systems today are based on alkaline technology, but new proton exchange membrane (PEM) electrolyzers are being developed which hold the promise of low cost, quick start-up and shutdown, and the ability to handle transient operating conditions well (OGDEN, 1999). Experimental designs using solid-oxide electrolytes which can operate at 700-900oC are also under development because high temperature systems offer higher efficiency since some of the work to split the water is done by heat. “Electrolyzers are typically ~70%-85% efficient on a higher-heating-value basis [efficiency = hydrogen out (HHV)/electricity in]” (OGDEN,1999). Water electrolysis can produce clean hydrogen with no CO2 emissions if wind, solar, or hydro electricity is used, but the production cost is strongly dependent on the cost of electricity. In general electrolytic systems are competitive with steam reforming of natural gas only where lowcost ($0.01-$0.02/kWh) power is available, for example for excess hydropower like found in Brazil during off-peak hours. It might also serve as a storage buffer for highly variable sources such as wind farms in remote locations where the cost of transmission of electricity to major markets has high losses. If the environmental cost of released CO2 from methane reformers is included, it would also help the economics of green electricity and hydrogen production.  36  Summary As technology currently stands steam methane reforming (SMR) is the most cost effective, followed by biomass or coal fasification, then electrolysis of water (which can be cheaper only for very low off-peak excess hydro electricity. The aluminum system is a very distant last place and would be over 125$/GJ, compared to SMR at under 5$/GJ for large H2 plants (THOMAS, 2000). In terms of efficiency, electrolysis (70-85%) is slightly better than SMR (75-80%), and both are far better than aluminum production <25%. The same holds true for greenhouse gas emissions which can be zero for electrolysis (from wind, hydro, or solar), followed by SMR at 5.5kg CO2/kg H2, and then lastly by aluminum at over 20kg CO2/kg H2. Obviously, aluminum-produced hydrogen will never serve as a clean source of power or be practical for large energy needs such as transportation. This still leaves the area of small mobile power needs which are currently served by batteries and into which small PEM fuel cells could move if hydrogen storage issues are resolved.  2.3.2 Hydrogen Storage One of the current drivers for interest in hydrogen is its ability to store energy for use in fuel cells or direct internal combustion engines. If hydrogen is widely used as a future energy carrier, storage will be needed to meet time-varying demands for fuel and allow for mobile energy sources. If this hydrogen is produced by electrolysis from clean energy sources it promises great environmental benefits, but even if produced from wind, solar, hydro, or nuclear power there still remains the issue of transporting it to small mobile systems. Compared to gasoline, hydrogen is not a very energy dense source, with 37  approx. 7 MJ @ 21 MPa for 1 gallon compared with 132 MJ for gasoline (U.S. Dept. of Energy, Fuel Comparison Chart). For a hydrogen vehicle to have a practical driving distance it needs to carry about 4 kg of hydrogen which would occupy 49 m3 (ZHOU, 2004). Storage then becomes a huge issue and methods of interest include compression, liquefactation, physisorption, metallic hydrides, and complex hydrides (BECKER, 2001).  Pressurized Cylinders By far the most common storage solution for small systems is currently steel pressure cylinders. Compressed gas is simple to implement, refilling is almost as rapid as that for gasoline (a few minutes or less), dormancy is good, and the energy requirements for compression are modest. Ogden reports the electrical requirements for compression to high pressure are typically 5%-7% of the energy content of the hydrogen and can be lower if hydrogen is produced at high pressure. Current conventional steel pressure cylinders (20 MPa) suffer from low energy density per unit mass and volume. New lightweight composite cylinders have been developed that are able to withstand pressures up to 80 MPa, so that hydrogen can reach a volumetric density of 36 kg/m3, approximately half as much as in its liquid form at normal boiling point (ZUTTEL, 2003). The safety of pressurized cylinders is a concern, especially in highly populated regions. It is envisaged that future pressure vessels will consist of three layers: an inner polymer liner over-wrapped with a carbon-fiber composite (which is the stress-bearing component) and an outer layer of a material capable of withstanding mechanical and corrosion damage. Industry has set a target of a 110 kg, 70 MPa cylinder with a  38  gravimetric storage density of 6 mass % and a volumetric storage density of 30 kg/m3 (ZUTTEL, 2003).  Liquid Hydrogen Liquid hydrogen is attractive in that it offers low weight and volume per unit of energy when stored in cryogenic tanks at 21.2 K and ambient pressure. It must be stored in an open system since there is no liquid phase above the critical temperature (33 K) and in a closed system the pressure could increase to 104 bar at room temperature. This results in poor dormancy due to boil-off and losses while refueling. Zuttel reports that doublewalled, vacuum-insulated spherical dewars can have boil-off losses below 1% per day. The other disadvantage to liquid hydrogen is the high cost of liquefaction. The necessary theoretical energy to liquefy hydrogen from room temperature is 3.23 kWh/kg, the technical work is about 15.2 kWh/kg, almost half of the lower heating value of hydrogen combustion (ZUTTEL, 2003). The large amount of energy necessary for liquefaction and the continuous boil-off of hydrogen limit the possible use of liquid hydrogen storage systems to applications where the cost of hydrogen is not an issue and the gas is consumed in a short time, e.g. air and space applications.  Physisorption Resonant fluctuations in charge distributions, which are called dispersive or Van der Waals interactions, are the origin of the physisorption of gas molecules onto the surface of a solid. In this process, a gas molecule interacts with several atoms at the surface of a solid and forms a monolayer. Carbon is an attractive medium for hydrogen storage because it is readily available and potentially low cost. Since the amount of 39  adsorbed hydrogen is proportional to the specific surface area of the adsorbent, materials with a large specific surface area like activated or nanostructured carbon and carbon nanotubes are possible substrates for physisorption. The maximum specific surface area of carbon (1315 m2/g) has shown a measured absorption capacity of 2 mass % (ZUTTEL, 2003). These experiments were done at 77 K which is far above the critical temperature of hydrogen so the monolayer of hydrogen would not be complete. Other materials have also been studied like a zeolite (NaY) with 1.8 mass % and a micro porous metal-organic Zn4O(1,4-benzenedicarboxylate)3 with a 3.7 mass% (ZUTTEL, 2003). The big advantages of physisorption for hydrogen storage are the low operating pressure, the relatively low cost of the materials involved, and the simple design of the storage system. The rather small gravimetric and volumetric hydrogen density on carbon, together with the low temperatures necessary, are significant drawbacks.  Metal Hydrides Metal hydrides are compounds in which hydrogen is absorbed by a metal under pressure and is released when heat is applied. Metal hydrides can absorb large amounts of hydrogen at a constant pressure which does not increase with the amount of hydrogen absorbed and can result in extremely high volumetric density of hydrogen atoms present in the host lattice. Some metal hydrides absorb and desorb hydrogen at ambient temperature and close to atmospheric pressure. All of the intermetallic compounds of interest for hydrogen storage consist of an element with a high affinity to hydrogen (La, Zr, Ce, Y, Ti, or Mg) and a low affinity one which is often Ni because it is an excellent 40  catalyst for hydrogen dissociation. Metallic hydrides can reach a volumetric hydrogen density of 115 kg/m3, e.g. LaNi5 (ZUTTEL, 2003). Metal hydrides are very effective at storing large amounts of hydrogen in a safe and compact way, but all the reversible hydrides working around ambient temperature and atmospheric pressure consist of transition metals and therefore, the gravimetric hydrogen density is limited to less than 3 mass %. The systems are costly and require a relatively long recharge time (10-20 min) (OGDEN, 1999).  Complex Hydrides Group 1, 2, and 3 light metals, e.g. Li, Mg, B, and Al, give rise to a large variety of metal-hydrogen complexes. They are especially interesting because of their light weight and the number of hydrogen atoms per metal atom, which is two in many cases. The main difference between the complex and metallic hydrides is the transition to an ionic or covalent compound upon hydrogen absorption. The highest volumetric hydrogen density reported to date is 150 kg/m3 in Mg2FeH6 and Al(BH4)3. The compound with the highest gravimetric hydrogen density at RT known today is LiBH4 (18 mass%) (ZUTTEL, 2003). Again, cost and long recharge times are the significant drawbacks to these systems, and, in contrast to the metallic hydrides, hydrogen is released via cascade decompositions which lower the dynamics of the release process.  Chemical Reactions Na vs. Al Hydrogen can be generated by reacting metals and chemical compounds with water. For example Na added to water releases hydrogen spontaneously. The Na transforms to NaOH in this reaction. The reaction is not directly reversible, but NaOH 41  can be removed and reduced in a solar furnace back to metallic Na. Two Na atoms react with two H2O molecules and produce one hydrogen molecule. The hydrogen molecule produces a H2O molecule in combustion, which can be recycled to generate more hydrogen gas, but the second H2O must be supplied which lowers the gravimetric hydrogen density to 3 mass%. Using Li gives a gravimetric density of 6.3 mass%. The major challenge of this storage method is reversibility and control of the thermal reduction process. Furthermore, the final solution is very basic and would be hazardous in the event of a spill, while the unreacted metal will react spontaneously if exposed to air so must be stored carefully. Using Al has the advantage of being a safe, inert material both before and after reacting with water so storage and handling is simpler and cheaper. Because one Al molecule gives 1.5 hydrogen molecules (see Eq. 22), less water must be supplied which reduces the weight of the system giving a gravimetric density of 8.3 mass%. Energy density is over 11.6 MJ/kg including water and over 15.6 MJ/kg not counting water. Since the density of mechanically alloyed Al (65% porous) is about 1g/cc the energy density on a per liter basis is 15.6 MJ/l. In comparison, liquid H2 is about 5 MJ/l, compressed H2 is 3 MJ/l, and metal hydrides are 4 MJ/l.  Summary In summary, the most common H2 storage method is as a compressed gas which does not however provide good volumetric storage density at obtainable pressures (less than 40 kg/m3) and has the associated dangers of a high pressure system. Liquid hydrogen is a better storage system volume wise (~70 kg/m3), but suffers from boil-off, 42  energy needed to liquefy the gas, and the need to maintain very low temperatures. Metallic hydrides of heavy metals have large gravimetric densities which limit their H2 mass% to less than 3%, while light metal hydrides need high temperatures for absorption and desorption of the hydrogen. Physisorption of hydrogen on nanotubes/nanofibers or activated carbon have advantages of low operating pressure and relatively low cost of the materials involved, but suffer from rather small gravimetric densities of less than 2% mass hydrogen. From a mass% basis, some of the most promising systems are the complex hydrides. Mg2FeH6 along with Al(BH4)3 shows the highest known volumetric hydrogen density of 150 kg/m3, which is more than double that of liquid hydrogen. While LiBH4 exhibit’s the highest gravimetric hydrogen density of 18% mass (ZUTTEL, 2003). While these systems can have a better mass% density than the aluminum system which is between 5% and 11%, they suffer from being much more expensive (LI, 2003) (VIDEM, 1977) (BOGDANOVIC, 2000). For example, NaBH4 costs close to $80/kg (DICKEN, 2005) compared to Al costing $1.8/kg. Hydrogen produced from NaBH4 would cost about $630/kg (LARAMINIE, 2003) while the cost of aluminum for 1 kg of hydrogen would be about $16. When you add the mass of a reformer, tank, and solution in a NaBH4 system the total storage efficiency (% mass H2) drops to 3.35%. In the aluminum system, powder storage is simple and will not add appreciably to the system weight, but some additional mass will be lost to the reactor which mixes the powder and water as needed. From a cost perspective in automobile usage, however, none of the current systems is competitive with gasoline, unless a widespread SMR infrastructure is developed. Even using recycled aluminum carries a 2x - 3x cost premium, and that is  43  before considering the cost of the fuel cell which currently run much higher than the IC engine. There are, however, many power needs that are unsuited for IC engines and use batteries for electric power. Comparing the specific energy of various types of batteries we have in Wh/kg (LARMINIE, 2003): Lead acid - 30, NiMH - 65, Li ion - 90, NaNiCl 100, and Zn_Air - 230. Al_Air has 225 Wh/kg, but is reported to have much lower specific power than the Zn_Air batteries. Hydrogen from an aluminum water split system would give a specific energy of over 4000 Wh/kg, although some of this would be lost when the weight of the fuel cell and added water is included. In conclusion, none of the current storage systems are currently competitive with fossil fuels from a cost basis although they represent a much cleaner environmental impact. Currently the most common method is pressurized tanks because of cost and simplicity, but they suffer from the danger inherent in a high pressure system and the fact that they cannot be refilled as rapidly as liquid fuel tanks. Al powder mixtures can compete on a mass%, safety, and refueling time basis, but at a cost penalty. Al powder mixtures can also compete with batteries in situations where their greater energy density can offset the need for an extra reactor and fuel cell. They also form an environmentally less toxic alternative to current battery technology.  2.4 Review Summary As an alternative hydrogen storage system, the use of ball milled aluminum to split water has many advantages over competing technologies. The material is safe, nontoxic, inert and easily stored and transported. It only requires additional water to initiate a 44  hydrogen producing reaction which results in hydrogen and aluminum hydroxide as products. Both volumetric and gravimetric energy density is good, being almost twice that of liquid hydrogen and better than many other storage methods. The system is not pressurized and operates at reasonable temperatures (below 100oC). There are minimal dormancy issues, unlike liquid hydrogen which suffers from boil-off. The ability to produce hydrogen at very high pressures could also provide an advantage in some applications. The disadvantage is cost. Because of the high price of aluminum, hydrogen produced by the aluminum water split method is 3x-9x the cost of other systems. While the output from the reaction is pure hydrogen with some water vapor, the overall lifecycle contributes significantly more CO2 equivalents to the atmosphere than other production methods. Again, this is because of the high energy requirements for aluminum refining and the high CO2 emissions during refining. Unless major technology breakthroughs occur in aluminum production this process will be limited to smaller niche markets that can support a high energy cost in return for the inherent advantages of the aluminum water split system. The corrosion properties of aluminum in water have been studied for years, and it is accepted knowledge that aluminum forms a passivating layer of oxide or hydroxide when exposed to air or water. This layer grows until the aluminum water reaction stops. To continue hydrogen evolution in water the passivating layer must be removed by mechanical action or chemical attack such as very acidic or very basic solutions. While various mechanisms for enhancing corrosion exist such as pitting by Cl- ions, dealloying of more noble elements through mechanical stress on aluminum alloys, and weakening of 45  aluminum oxide films by increasing dislocation densities and grain boundaries, none of these mechanisms is reported to result in the release of hydrogen on a continuous basis until substantially all the aluminum has reacted in a neutral environment. The highest corrosion rates were reported for high pH solutions or high temperature water. Current methods to produce hydrogen from aluminum focus primarily on highly alkaline solutions, but alloying with other metals through mechanical milling or melting has been shown to increase corrosion by forming a galvanic cell or hindering the formation of passive films. The use of alumina to weaken the native oxide and promote pitting through a sintering process has also been reported.  46  Chapter 3 Research Objectives  The relatively high effective storage density for hydrogen produced from the corrosion of aluminum in water makes it an interesting source in the proposed hydrogen economy, and a valuable secondary use for aluminum that is otherwise not suitable for mainstream recycling. Previous work was done at UBC (CZECH, 2006) to optimize ball milling as a method for accelerating this corrosion in neutral pH environments where traditionally aluminum is considered to be highly corrosion resistant due to the formation of a passivating hydroxide layer. This thesis attempts to examine some of the phenomena observed during accelerated corrosion of Al powders in water. The changes in pH and temperature observed during the reaction of ball milled aluminum and the change in surface area produced by milling were analyzed in an attempt to determine the most important factors driving the accelerated corrosion. The addition of alumina powder to aluminum as an alternative to milling is investigated and a hypothesis given for the increased corrosion aluminum undergoes in the presence of alumina powder. More specifically, the following objectives were undertaken in an attempt to gain insight into the controlling factors governing the corrosion of ball milled aluminum powders and the corrosive effects of alumina powders on aluminum.  1) Investigate the role of temperature in this exothermic reaction with regard to rate, extent of reaction before passivation, and corrosion products. 2) Examine the effect of alkalization of the solution during the reaction. 47  3) Determine the role of the surface area of the ball milled aluminum powder after milling with alumina, on the reaction rates. 4) Study the rate and efficiency of aluminum and alumina mixed systems which have not been ball milled and their suitability as a hydrogen source. 5) Measure the hydrogen exchange current density on aluminum and platinum surfaces in contact with alumina. 6) Examine the effect of electrical contact between the aluminum and alumina particles for acceleration of corrosion and delaying of passivation. 7) Study the effect of alumina on the buildup of hydroxide on aluminum and changes in corrosion rate with temperature.  48  Chapter 4 Experimental Aspects 4.1 Materials All experiments used atomized aluminum powder (Figure 6) from Alcoa, 101 common grade (99.7% Al, main impurities: Fe (max 0.25%) and Si (max 0.15%) as well as 0.6% Al2O3 on the surface of powder particles, with a volume weighted mean particle size of 40 μm (Alcoa material spec sheet). Aluminum oxide (Figure 7), Al2O3 A16 SG (Alcoa), was used as reference additive, where the term “additive” is used to describe the material milled with Al when preparing the powders. The Al2O3 particles had a volume weighted mean particle size of 2.5 μm as measured by a Malvern instruments Mastersizer 2000 particle analyzer. Aluminum that is termed ‘washed’ refers to aluminum powder that has been milled with a water-soluble inorganic salt (WIS) such as potassium chloride, KCl (technical grade, McArthur Chemical), or sodium chloride, NaCl (99.9%, Fisher Chemicals), where the WIS was subsequently removed by a series of immersion and filtration steps. Other additives (technical grade, McArthur Chemical) used for testing and comparison purposes were quartz (SiO2), silicon carbide (SiC), magnesium oxide (MgO), and silicon nitride (Si3N4). Aluminum disks were 1.5cm diameter circles cut with a punch from aluminum foil, 1.0mm thick, annealed, 99.99% pure Al (Alfa Aesar).  49  Figure 6 Aluminum particles – as received.  Figure 7 Alumina particles – as received.  50  4.2 Sample Preparation Aluminum-additive powder mixtures were prepared by high-energy ball-milling (BM) in a SPEX 8000 shaker mill. The balls and vial inner lining were made of alumina. The grinding vial had an internal diameter of 38 mm and a length of 44 mm, corresponding to a capacity of about 55 ml. Up to 2.5 g of aluminum-additive powder mixtures were placed in the vial. The duration of milling was typically 15 minutes. In a typical experiment, a 50/50 wt% mixture of an additive and aluminum powder were loaded together with 28g (approximately 70) of alumina balls (5 ± 0.15mm diameter each), and ball-milled in air (Figure 8).  Figure 8 Al/Al2O3 mixture ball milled for 15 minutes.  The water-soluble inorganic salts were BM separately for 5 minutes to reduce their initial particle size before being added to the aluminum powder and BM together. 51  Other additives were used as-received. The mill and milling media are depicted in Figure 9 and Figure 10. For tests in an electrochemical cell the powders were pressed into pellets using a hydraulic press at pressures between 23 MPa and 69 Mpa. Loose powder in nonmilled experiments was hand mixed with additives using a stainless steel scoop in a plastic cup until the mixture was visibly uniform.  Figure 9 SPEX 8000 Shaker Mill.  52  Figure 10 Ball mill vial and alumina balls.  4.3 Hydrogen Evolution Test Apparatus For hydrogen evolution testing the solution (typically 100ml of de-ionized (DI) water (1Mohm)), was added to a 200ml Pyrex beaker and heated on an electric hot plate/stirrer (Barnstead Thermolyne). See Figure 11. Once the solution was at temperature, the powder was added and the beaker stoppered with a #12 rubber stopper with a Pyrex tube through the center. A ¼ inch inner diameter tube went from the beaker and into an inverted graduated cylinder. Hydrogen evolved from the reactions:  Eq. 27  Al + 3H2O → 1.5H2 + Al(OH)3  Eq. 28  Al + 2H2O → 1.5H2 + AlOOH  was measured by the displaced water in the cylinder. 53  Graduated Cylinder  H2  Beaker H2 Al Powder Hot Plate  Water Bath  Figure 11 Hydrogen evolution experimental setup.  4.4 Electrochemical Investigations Electrochemical tests were performed on Al disks and pressed powder pellets in a commercially available one-liter glass cell (Princeton Applied Research, Model G0096) which had been retrofitted with an outer glass jacket to allow water circulation for temperature control (Figure 12). A high-density graphite counter electrode was used with a Ag/AgCl (E=0.197 V vs. SHE) reference electrode (Fisher Scientific, Accumet 13-62053). All potentials are reported verses the standard hydrogen potential (SHE). Solutions were de-aerated with a nitrogen purge, and the working electrode was mounted in a Tefzel® Dupont fluorocarbon electrode holder (Figure 13) with a crevice-free teflon washer exposing ~0.9 cm2 of the sample. The reference electrode was kept in contact with the solution through a salt bridge terminating in a Luggin probe with a Vycor tip (porous glass) spaced within 5 mm of the working electrode surface (Figure 14). 54  Figure 12 Electrochemical cell.  Figure 13 Working electrode sample holder.  55  Figure 14 Luggin probe tip with Vycor frit and sample holder.  All electrochemical testing was done using a Princeton Applied Research Potentiostat/Gavanostat Model 273A (Figure 15).  Figure 15 Potentiostat/Gavanostat model 273A.  56  4.5 Materials Characterization The following techniques were used to characterize the incoming materials and reaction products.  4.5.1 Scanning Electron Microscopy (SEM / EDS) A Hitachi S-3500N scanning electron microscope with energy dispersive x-ray spectrometry (EDS) was used to examine particle morphology, microstructure, and composition before and after milling and reaction in water. The working distance for the EDS was 15mm and powder samples were mounted on a conductive carbon tape and sputtered with a 60%gold/40% palladium 10nm thick layer to reduce surface charging.  4.5.2 X-Ray Diffractometry (XRD) Identification of reaction products and hydroxide layers for both powder samples and Aluminum disks were done using X-ray diffraction patterns (XRD) from a Siemens D-5000 diffractometer using 40 kV / 30mA and Cu Kα radiation (λ = 1.5405 Ǻ). The diffraction patterns were recorded at an angular speed of 1.2o(2θ)/min. with a 2 s/step counting time. Data accumulation and processing was performed using Diffracplus software from Bruker Analytical X-ray Systems. Phase identifications were performed with help of PDF database and EVA V4.0 software.  4.5.3 Surface Area Measurement by BET Analysis Surface area measurements were performed with Quantachrome Autosorb-1 Surface Area Analyzer on as-received, ball-milled and washed aluminum powders.  57  Powder samples (0.5 to 8 g) were out-gassed overnight >12 hrs (150ºC, 5 mm Hg) prior to analysis. Adsorpion-desorption isotherms were measured at 77 K. The specific surface area was estimated using multi-point adsorption data from the linear segment of the N2 adsorption isotherm in the relative pressure range of 0.05 to 0.3 using Brunauer-EmmettTeller (BET) theory.  4.5.4 Transmission Electron Microscopy (TEM) A Hitachi 800-S Transition electron microscope operated at 200KeV/20mA was used to look for small particles of aluminum coating the surface of Al2O3 particles after ball milling. TEM specimens were prepared by making a suspension of BM powder in ethanol by ultrasonically agitating for 1 minute. 2-3 drops of this suspension were dropped onto clean copper grids of 3.05mm diameter, and were dried in air for 10-15 minutes.  4.6 Errors and Reproducibility There are several sources of error that are unavoidable due to the nature of the work and should be kept in mind by anyone trying to reproduce the results presented in this thesis. In electrochemical testing surface contamination can have a large effect on sample potential and therefore current densities in potentiodynamic scans such as those reported in Chapter 5 and 6. For this reason in standard testing samples are polished and cleaned prior to study. However, the pressed powder samples do not have the structural integrity to withstand polishing, so normal sample preparation could not be used and the possibility of contaminants exists. Furthermore current densities are graphed as a function  58  of sample surface area exposed to the solution. Unlike a bulk metal sample that has been polished flat, a compressed powder will have potential surface topography and porosity allowing solution penetration into the sample. For this reason, the current density is listed as apparent current with the surface area being that of the sample holders exposed area. Actual surface area may differ among samples, but it is hoped that similar powders undergoing the same process of pelletization will result in similar surfaces. Tests looking at activation energy in Section 5.2 use the same pellet at increasing temperatures which avoids the problem of surface preparation. Since the increase in rate with temperature is known to follow the Arrhenius rate equation, the experimental data can be fit with the straight line that is predicted. Tests in Section 5.2 showed good agreement with the theory and had R2 values of 0.98 - 0.99 showing good stability within a pellet and no evidence for a time-dependent surface area with solution penetration. Some repeat curves are presented in Appendix C which show almost complete overlay for potentiodynamic scans run at the same conditions for Al BM with Al2O3 or CuO. Although the effective surface area of the powder pellets does not appear to change with time, care should be taken when comparing absolute current values between blank aluminum disks and pressed powder pellets. When collecting hydrogen during evolution tests, there is a section of water filled tube that must be displaced before hydrogen can enter the cylinder. Depending on the size and length of tube used this was generally between 5 and 10 ml of hydrogen not recorded. This is a systematic error of the setup which will effect all measurements, but is obviously more significant at smaller volumes and for initial measurements. In tests involving multiple additions of powder, this error is cumulative and in the batch tests run 59  there will be hydrogen lost as the beaker is opened and closed to add aluminum. If hydrogen is evolving rapidly this loss can be significant, but is unavoidable with the current experimental setup. Hydrogen loss from the cylinder with long term tests appeared to be minimal, with reductions in the volume being less than 5ml out of 500ml over the course of days and more likely attributable to gas cooling. The reaction rates are quite sensitive to solution temperature and care should be taken to run all tests in similar sized beakers with a standard volume of water chosen. In a similar fashion ball milled powders are affected by milling conditions, time, and milling media. The number of oxide balls and amount of powder to be milled in one batch should be standardized for best reproducibility. The milling is not completely homogeneous within the alumina crucible and a white crust is often observed at the top or bottom edge which is oxide rich. This portion can be separated from the sample, resulting in the sample not having exactly the starting ratio of aluminum to oxide, or it can be used and the powder is then not homogeneously milled. This non-uniform portion is typically less than 5% of the total sample, but represents an unavoidable source of error in all results. In general, multiple runs showed very good agreement for hydrogen evolution tests. Some examples are shown in Appendix C. Five runs of 0.5g Al mixed with 2g of Al2O3 in 60oC water gave a mean evolution in one hour of 183 ml with a standard deviation of 9, although this series showed declining values as the test progressed attributed to changes in air temperature as the tests were run from day into the night. This further highlights the need to control solution and air temperature carefully.  60  Chapter 5 Corrosion of Ball Milled Aluminum This chapter covers investigations of aluminum powder that have been ball-milled and the effect that milling has on aluminum corrosion rate in water. Chaklader (2002) discovered that aluminum when ball milled with aluminum oxide powder continues to bubble hydrogen as a reaction by-product long after the reaction was expected to stop due to passivation (Figure 16). Recent work has been done to optimize the reaction rate by changing milling conditions and secondary milling particles (CZECH, 2006). 450 Al  400  BM Al  Volume of Hydrogen (ml)  350 300 250 200 150 100 50 0 0  10  20  30  40  50  Time (hours)  Figure 16 Hydrogen evolution from 0.5g of aluminum powder in 55oC DI water with as received and ball milled (50 wt% Al2O3) powders.  61  This work will try to define what role the secondary particle plays in determining the corrosion rate as well as the effect on aluminum surface area after milling. Other researchers like Fan (2007, 2008) who are studying ball milling of aluminum to generate hydrogen typically mill the aluminum with another metal such as Bi to form a galvanic cell to enhance the Al corrosion. Fan (2008) reports that no reaction occurs when Al is milled with 20 wt% NaCl, however that is not consistent with Figure 17, and in general this investigation involves oxides or salts which are not expected to form galvanic cells. When looking at the reaction of milled aluminum in a small volume of water (Figure 17) an increase in both solution temperature and pH is seen, concurrent with or preceding the evolution of hydrogen. 12  95  11 85  9  pH  8  75  7 6  65  Tem perature  5 4  55  3  H2  2  Temperature (ºC)  Volume of Hydrogen (100 ml) pH  10  45  1 0  35 0  10  20  30  40  50  60  Time (minutes)  Figure 17 1g of aluminum powder ball milled with 1g of KCl and placed in 30 mL of tap water at 55oC.  62  The exothermic nature of the reaction allows for localized temperature increases to play a role in increasing the reaction rate. Increasing pH will further increase the solubility of the reaction products in solution and therefore decrease the passivation of the aluminum surface (POURBAIX, 1966).  So to better understand the increased  corrosion rate we will look at temperature rise, pH shift, chemical composition of the additive, and surface area changes in the aluminum.  5.1 Effect of Secondary Particles When ball milling aluminum powders it is necessary to use a second powder in addition to the alumina balls used as a milling media to prevent the aluminum from galling or clumping together to form larger particles with little or no increased corrosion rate (CZECH, 2006). The effect of these secondary particles on aluminum surface area is addressed in Section 5.5. In this section the question of a chemical interaction occurring that promotes the aluminum corrosion is addressed. That the role of the secondary particles was of a chemical or compositional nature was deemed unlikely for several reasons. First, some degree of corrosion enhancement was seen over a wide range of additives (CZECH, 2006). This reduces the chance that one particular chemistry is the critical component. Second, one of the most successful secondary particles, Al2O3, is already present as a surface coating on any aluminum particle after exposure to air, and so does not represent new chemistry to the water-aluminum-oxide system. Third, aluminum powder can be milled with a water soluble salt and then the salt can be removed by a washing processes (CZECH, 2006), leaving very little of the secondary particle in the resulting powder which is still highly reactive. 63  The one common possible contaminant seen in all systems tested was the alumina used to line the milling crucible and the alumina balls used as a milling media. As mentioned previously, alumina is already found as a surface layer grown on aluminum particles by their reaction with air, but this layer would be relatively pure depending on the purity of the underlying aluminum. The incoming alumina used was 99.7% pure with the two highest impurities being Mg and Si, so tests were run in the following sections to look at the effect of increasing the level of these possible contaminants deliberately to see if a resulting increase in corrosion rate was observed.  5.1.1 Electrochemical Tests For these tests, aluminum powder was ball milled for 15 minutes at a 50% weight ratio with various secondary particles and then pressed into pellets for electrochemical testing. A 0.1M Na2SO4 solution was used to increase electrical conductivity in the cell without adding Cl- ions that are known to promote pitting corrosion in aluminum (DAVIS, 1999). As can be seen in Figure 18, all of the oxide additions (Al2O3, SiO2, and MgO) showed similar open circuit corrosion potentials (Ecorr) at around -1.4 V vs. SHE. The addition of Al2O3 and SiO2 showed similar corrosion rates with the rate for SiO2 being somewhat higher, the addition of MgO showed a decrease in the corrosion rate with a lower corrosion current in the passive region. Rates can be extrapolated by a Tafel fit to the curve to find Icorr or in the case of a passivating film a rate in the passive region used. It is important to keep in mind that when a powder has been pressed into a pellet only an approximate current density is reported because the actual surface area of the pellet and the degree of solution penetration into the pellet is unknown. In Figure 18 the 64  rate for Al/SiO2 is approximately 1 mA/cm2, but comparison should only be made between similar pressed powder samples and care taken not to equate any rates with those for solid disk samples. 0.4 0.2  50% Cu  0.0 -0.2  Potential (V)  -0.4 -0.6  50% SiO2  -0.8 -1.0  50% Al 2O3  50% MgO  -1.2 -1.4 -1.6 -1.8 -2.0 -2.2 1.0E-07  1.0E-06  1.0E-05  1.0E-04  1.0E-03  1.0E-02  1.0E-01  1.0E+00  2  Apparent Current (A/cm )  Figure 18 Potentiodynamic scans of Aluminum powder ball milled with Cu, MgO, SiO2, and Al2O3 in a 0.1M Na2SO4 solution at 25oC.  The addition of pure Cu metal showed an increase in Ecorr to -0.7 V vs. SHE as would be expected by mixed potential theory. When small additions of MgO were added (2% by weight), which is still much higher than any contamination levels that might transfer from the alumina (see Figure 19), minimal changes were seen in the potentiodynamic scans.  65  0.0 -0.2 -0.4  Potential (V)  -0.6  50%SiO2  48%Al 2O3 -2%MgO  -0.8  48%SiO2 -2%MgO  50%Al 2O3  -1.0 -1.2 -1.4 -1.6 -1.8 -2.0 -2.2 1.0E-06  1.0E-05  1.0E-04  1.0E-03  1.0E-02  1.0E-01  1.0E+00  Apparent Current (A/cm 2)  Figure 19 Potentiodynamic scans of Aluminum powder BM with Al2O3, SiO2, and 2% additions of MgO in a 0.1M Na2SO4 solution at 25oC.  In the next section hydrogen evolution rates in water for the various mixtures will be shown to correlate with the electrochemical tests, but from these results there is no evidence that chemical changes in the secondary particle are driving the increase in corrosion seen between milled and non-milled systems.  5.1.2 Hydrogen Evolution Hydrogen evolution tests (Figure 20, Figure 21) show no significant increase in corrosion rate with other oxides over the Al/Al2O3 mixture. Contrary to the results seen in the electrochemical tests (Figure 18, Figure 19) the SiO2 mixture was slightly less reactive than the Al2O3 mixture and the MgO mixture showed equal or better reaction 66  rates. While the increase seen for the MgO was not significant and would not indicate a chemical reaction is occurring that might promote its importance as a trace impurity picked up from the alumina, the difference between the hydrogen evolution results (SiO2 has a lower rate) and the electrochemical tests (where SiO2 was slightly higher) highlights one of the differences in material preparation between the tests. In the electrochemical tests the BM powders are first pressed to 69 MPa to form pellets, this will act to reduce some of the surface area that might be formed during milling as the soft aluminum will deform under pressure to form a flat surface. So the differences between the electrochemical and evolution tests support the importance of surface area to be looked at in Section 5.5, and do not contribute support to a chemical reaction explanation. 300  48% A l2 O3 +2%MgO 50% MgO 50% A l2 O3  Volume of Hydrogen (ml)  250  50% SiO2  200 150 100 50 0 0  10  20  30  40  50  60  70  80  Time (minutes)  Figure 20 Hydrogen evolution from 0.5 g of aluminum powder BM with Al2O3, SiO2, and MgO in 50ml of DI water at 70oC.  67  As expected from the electrochemical tests we see very slow (compared to other BM systems) corrosion rates for the Al/Cu mixture (Figure 21). 300  Volume of Hydrogen (ml)  250 200 150 100 50 0 0  10  20  30  40  50  60  70  Time (minutes)  Figure 21 Hydrogen evolution from 0.5 g of aluminum powder BM with 50% Cu in a 70oC 0.1 M Na2SO4 solution.  In Figure 22 some non-oxide particles are compared with the Al/Alumina mixture. We can see that while 50wt% Si3N4 is almost identical to the alumina mixture, 50wt% SiC has only half the reaction rate. This difference is explained in Section 5.5 to be a difference in the resulting aluminum surface area after milling. If a chemical reaction is occurring with milled powders, why would the reaction rates be so similar between Al2O3 and Si3N4 and yet different for SiC? Clearly another mechanism is responsible.  68  300 Si3N4 Al2O 3  Volume of Hydrogen (ml)  250  200  150  SiC  100  50  0 0  10  20  30  40  50  60  70  Time (minutes)  Figure 22 Hydrogen evolution from 0.5 g of aluminum powder BM with 0.5 g of Si3N4, Al2O3, and SiC in 100ml of DI water at 60oC.  5.2 Activation Energy The Arrhenius equation gives the dependence of the rate constant k of chemical reactions on the temperature and the activation energy ‘Ea’: Eq. 29  k = Ae-Ea/RT  By plotting the change in rate on a log scale as a function of the inverse of the temperature in Kelvin the activation energy can be found from the slope. This is very useful since the uncertainty in the area of the powder pellets makes determination of the absolute current density impossible. Since the same pellet is used at each temperature, uncertainties in the area cancel out as far as determining the activation energy. The pre69  exponential factor A can not be determined since the true Y intercept is unknown, but a change in Ea for BM samples compared to bulk aluminum would indicate a difference in the rate limiting step for the corrosion mechanism. During ball milling the aluminum powder is subject to high impact collisions producing both mechanical stress and deformation and localized heating. Damage to the native oxide coating (FAN, 2007) and retained internal strain could both affect the reaction kinetics of the subsequent corrosion. To examine this possibility reaction rates were measured as a function of temperature and Arrhenius plots made to calculate the activation energy of the rate limiting step in the corrosion of aluminum to aluminum hydroxide. Differences between bulk aluminum and ball milled powders would point to changes in the governing reaction mechanism or a lowering of energy needed to initiate reaction steps.  5.2.1 Aluminum Disk Using a blank aluminum disk held at +200mV vs. SHE, the corrosion current was measured over 12 hours as the temperature was stepped from 10oC to 85oC (see Figure 23), a 0.01M borax solution was used to approximate the pH = 9 that would be seen with BM samples later.  70  10000  Current (uA)  1000 85 oC 70 oC  100 40 oC  10 o  10 C  55 oC  25 oC  1 0.0  15.0  30.0  45.0  60.0  75.0  Time (hours)  Figure 23 Potentiostatic test of current verses time for an aluminum disk in 0.01M borax solution held at +200mV for 12 hours at each temperature step from 10oC to 85oC.  From the Arrhenius plot (Figure 24) we can determine the slope and thus calculate the activation energy for the rate limiting step in the reaction. In this case, activation energy was approximately 64 kJ/mol which is comparable to values reported by (ALEKSANDROV, 2003) and (HU, 2003) (see Table 1 page 81 for complete list of activation energies found). A marked drop in rate was seen at 85oC which is attributed to the growth of a different hydroxide at those temperatures (see Section 5.4), but rates taken after cooling to 25oC and then 10oC again fall on a straight line with similar but somewhat reduced slope.  71  8 7 Ramp from 10oC - 70oC  Rate [ln I] (uA)  6  slope = -7699  5 4 3 2 1  Ramp from 85oC - 10oC slope = -6857  0 0.002  0.0025  0.003  0.0035  0.004  1/Temperature (1/K)  Figure 24 Arrhenius plots for the aluminum disk in 0.01M Borax solution for temperature steps from 10oC to 70oC and then 85oC and back down to 10oC. Ea=64 kJ/mole.  In a similar fashion corrosion rates were determined for a disk at -200mV (see Figure 102 page 177) and the Arrhenius plot shown in Figure 25. This resulted in a higher activation energy (78.6 kJ/mole) which would be expected since the lower overpotential applied to the aluminum would force less anodic growth of the passivating film. The trend for all the values reported in Table 1 page 81 is that increases in the passivating layer decrease the activation energy. We believe this is because the overall corrosion rate becomes more dependant on diffusion of ionic species to and from the surface of the aluminum. If the disk run at +200mV is rerun a second time (Figure 103) a further decrease in activation energy is seen to 23.9 kJ/mole.  72  7 6  Rate [ln I] (uA)  5 4 3  Ramp from 10oC - 55oC  2  slope = -9452  1 0 0.0025  0.0027  0.0029  0.0031  0.0033  0.0035  0.0037  1/Temperature (1/K)  Figure 25 Arrhenius plot for aluminum disk held at -200mV in 0.01M Borax solution for temperature steps from 10oC to 55oC. Ea=76.8 kJ/mole.  In Figure 26 an aluminum and alumina mixture that was not ball milled was run through polarization scans from -2 V to +0.5 V at temperatures from 5oC to 55oC after cycling both up and down through this temperature range two times. This sample is included here as an example of a pellet that has had plenty of time to grow a protective oxide layer. Taking the current at 0V, we can form the Arrhenius plot (Figure 27) and we can calculate an activation energy of 26 kJ/mol (Table 1). This will be compared in the next section with activation energies from BM samples which tend to react quickly at higher temperatures and thus are hard to measure without having a hydroxide layer present.  73  1.0 5 ºC 0.5  15 ºC 25 ºC  Potential (V)  0.0  35 ºC 45 ºC  -0.5  55 ºC -1.0 -1.5 -2.0 -2.5 1.00E-08  1.00E-07  1.00E-06  1.00E-05  1.00E-04  1.00E-03  Apparent Current (A/cm 2)  Figure 26 Potentiodynamic scans of a non-milled pellet of 50% Al and 50% alumina in a .1M Na2SO4 solution at temperatures from 5oC to 55oC after cycling two times.  4.0 3.5 Ramp from 5oC - 55oC slope = -3118  Rate [ln I] (uA)  3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.003  0.0031  0.0032  0.0033  0.0034  0.0035  0.0036  1/Temperature (1/K)  Figure 27 Arrhenius plot for non-milled Al/alumina pellet in 0.1M NaSO4 solution.  74  0.0037  The ability of the aluminum/alumina pellet to still show the expected Arrhenius variation of rate with temperature after 2 previous cycles to high temperature also implies that it is not passivating as strongly as the aluminum pellet, which has implications that are discussed in Chapter 6. By comparison, an identical test using an aluminum pellet (Figure 28) shows that passivation has occurred, and an Arrhenius plot would give a nonsensical negative value for the activation energy.  1.0 5 ºC  0.5  15 ºC 25 ºC  Potential (V)  0.0  35 ºC  -0.5  45 ºC 55 ºC  -1.0 -1.5 -2.0 -2.5 1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  1.E-02  1.E-01  1.E+00  Apparent Current (A/cm 2)  Figure 28 Potentiodynamic scans of a non-milled pellet 100% Al in a 0.1M Na2SO4 solution at temperatures from 5oC to 55oC after cycling two times.  5.2.2 Ball Milled Powders Rate data for Arrhenius plots were best taken once the sample has had time to equilibrate to a steady state at each temperature. Even with solid aluminum disks, this was somewhat problematic as the growth of protective layers could be clearly seen at 75  certain temperatures. For BM samples the problem of rate stability was much worse. In Figure 29, the exponential decay of current with time is seen for the BM sample over the first 11 hours, with the current falling even as the temperature was stepped up from 5oC to 15oC to 25oC. When the rate finally started to stabilize, it was still much lower than initial starting values even at higher temperatures. The rate also started two orders of magnitude above that seen with the solid disk in the previous section.  4.0 3.5 Current (mA)  3.0  5o C  2.5 2.0  15o C  1.5  o  1.0  40 C 15o C  o  25 C  0.5  55o C 25o C  0.0 0.0  5.0  10.0  15.0  20.0  25.0  Time (hours)  Figure 29 Potentiostatic test for BM Al in 0.01M Borax at +200mV at various temperatures.  Using the last four temperature regions where the rate was somewhat stabilized Figure 30 was plotted and the activation energy calculated at 23.9 kJ/mol, very similar to the value found for the aluminum oxide mixture that saw a prolonged reaction.  76  7.0  Rate [ln I] (uA)  6.5  slope = -2875  6.0 5.5 5.0 4.5 4.0 0.003  0.0031  0.0032  0.0033  0.0034  0.0035  0.0036  1/Temperature (1/K)  Figure 30 Arrhenius plot for BM Al in 0.01M Borax at +200mV from 15oC to 55oC. Ea=23.9 kJ/mole.  While Figure 29 looks erratic, a similar value for activation energy was found from Tafel curves for a BM sample in 0.1M Na2SO4 solution that was first ramped to high temperature (55oC), then run again to produce Figure 31 and Figure 32. Three different samples using both milled and unmilled aluminum in different solutions all gave an activation energy of around 24 kJ/mol when the sample was immersed for long periods of time.  77  0.0 -0.1  5 ºC  -0.2  15 ºC  -0.3  25 ºC #1 25 ºC #2  -0.4  35 ºC  Potential(V)  -0.5 -0.6 -0.7 -0.8 -0.9 -1.0 -1.1 -1.2 -1.3 -1.4 1.0E-08  1.0E-07  1.0E-06  1.0E-05  1.0E-04  1.0E-03  1.0E-02  1.0E-01  Apparent Current (A/cm 2)  Figure 31 Potentiodynamic scans for BM Al/Alumina pellet in 0.1M Na2SO4 solution at temperatures from 5oC to 35oC after initial series run up to 55oC.  78  6.5  Rate [ln I] (uA)  6.0  5.5 slope = -2870 5.0  4.5 0.0032  0.0033  0.0034  0.0035  0.0036  0.0037  1/Temperature (1/K)  Figure 32 Arrhenius plot for BM Al/Alumina pellet in .1M Na2SO4 solution after initial series of runs to 55oC.  Trying to find values for BM samples before they corroded significantly was plagued with problems since the sample could not be allowed to reach a steady state. Potentiodynamic scan rates were increased from 0.5 mV/s to 5 mV/s, but to be valid the sample had to have enough time to reach the target temperature. If different samples were used then each started at room temperature before being immersed in the solution and by the time they reached the solution temperature a significant rate reduction was already being observed which would cause the corrosion rate to appear to decrease with increasing temperature. A value of 73.8 kJ/mole was calculated using only data from 5oC and 15oC, which does fall between the values of 64 kJ/mole and 79 kJ/mole found for the solid disks, but with only two points the line was of course a perfect fit. High R2 values found on the previous plots with 4 or 5 points helped validate the results as fitting with 79  theoretical curve expected. Figure 33 was finally calculated by using fast potentiodynamic curves (5 mV/s) on a sample that was catholically polarized to -2 V between each run to prevent the aluminum from corroding while the temperature was ramped to the new value. The rate at each temperature was determined by a Tafel fit to the curve. Five temperatures can be graphed (5, 15, 25, 35, and 45oC) with a R2 of 0.98 and the activation energy at 71.9 kJ/mol is very close to that found for the separate runs at 5oC and 15oC. 9  Rate [ln I] (uA)  8  7 slope = -8649 6  5  4 0.0031  0.0032  0.0033  0.0034  0.0035  0.0036  0.0037  1/Temperature (1/K)  Figure 33 Arrhenius plot for BM Al/Alumina pellet in 0.1M Na2SO4 solution during initial run from 5oC to 45oC.  There is a visible fall off in the corrosion rate seen for the higher temperatures so obviously the corrosion products are beginning to build up. If the 45oC point is dropped the R2 value gets a little better and the activation energy goes up to near what is seen for the aluminum disk held at -200 mV. Table 1 is a summary of the activation energies 80  discussed. The values found for the BM samples lie between those for the solid aluminum disk at +200 mV and -200mV which would have greater and lesser buildup of hydroxide respectively. As the samples were allowed to react longer the activation energies decreased and at the low end very similar values were seen between the BM samples after they were exposed to high temperatures and un-milled aluminum powder after several days immersed in solution being cycled through temperature changes. Table 1 Activation energies for the series of tests of Al corrosion in water.  Experiment  Slope = -Q/R  Q  q  R=8.314  [kJ/mol]  [eV/atom]  Al in 0.01M Borax -200mV 1st run  9452  78.6  0.81  BM Al in 0.1M Na2SO4 separate runs  8880  73.8  0.77  BM Al in 0.1M Na2SO4 1st run  8649  71.9  0.75  Al in Borax +200mV 1st run  7699  64.0  0.66  Al in Borax +200mV 2nd run  5564  46.3  0.48  Al + Alumina in 0.1M Na2SO4 3rd run  3118  25.9  0.27  BM Al in 0.01M Borax +200mV 1st run  2875  23.9  0.25  BM Al in 0.1M Na2SO4 2nd run  2870  23.9  0.25  In conclusion there is no clear evidence to support the idea that a different activation energy in BM systems is responsible for increased corrosion of BM samples. All activation energies found for BM samples are somewhat questionable because there was no way to let the system reach a steady state equilibrium at each temperature due to 81  the rapid nature of the reaction. However, the values that were obtained are similar to that seen for bulk aluminum, although the bulk aluminum sample had to be run in Borax so that it would not passivate too quickly.  5.3 pH Effects Hydrogen evolution is reported to alkalize a solution (CZECH, 2006), and all tests run with ball milled powders have shown some increase in pH. The reduction of water by electrons at the cathode produces hydrogen and OH- ions (Eq. 17), which can cause a local alkalinization of the solution. However, because the Al3+ ions want to form Al(OH)3 any change to the bulk solution should be transient. Another possibility is that the alumina surface will attract a surface layer of protons because it is below its iso-electric point at a pH of 7. This will leave a relative abundance of OH- in solution as a screening charge to preserve electrical neutrality which might increase solution pH. This upward shift in pH is most pronounced when the reaction occurs in a small volume of water (see Figure 17). With the solubility of aluminum hydroxide increasing at higher pH values and the increasing formation of Al(OH-)4 ions the natural passivation of aluminum in water will be decreased. The extent to which this pH shift will accelerate the corrosion of aluminum will be investigated in this section and the likelihood of high pH microclimates forming on the deformed surface discussed.  5.3.1 Electrochemical Tests Figure 34 shows a potentiodynamic scan of aluminum in 25oC water in three different forms, a solid aluminum disk, a press aluminum powder pellet, and a pellet  82  pressed from a ball milled aluminum/alumina mixture. The apparent current increases as would be expected for the more reactive BM mixture, but the current must be qualified as apparent because with the powder samples the actual surface area exposed to the solution is unknown, as is the degree to which the solution can penetrate into the porous pellet. The open circuit potential should be independent of solution contact area unless microclimates are forming in the mechanically worked BM sample or the porosity of the pressed pellets allows for localized changes near the surface of the sample where the solution is different from the bulk similar to the micro environments that are responsible for pitting corrosion. In this case, there is a change from -350mV for the Al disk to 900mV for the BM sample. Because pH is seen to shift upward during hydrogen evolution (Figure 17) a localized pH change might be one of the forms a micro climate is expressed in and detectable by buffering. 2.0 Al Powder  1.5  Potential (V)  1.0 0.5 Al Disk  0.0 -0.5  BM Al/Al2O 3  -1.0 -1.5 -2.0 -2.5 1.E-09  1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  1.E-02  Apparent Current (A/cm 2)  Figure 34 Potentiodynamic scan of an aluminum disk, an aluminum powder pellet, and a BM al/alumina pellet in water at 25oC.  83  In Figure 35, the potentiodynamic curves for an aluminum disk and a BM pellet are shown in a 0.01M Borax solution. This solution will buffer the pH to roughly 9 (CRC Handbook), which is close to the pH observed in small volume reactions run to completion. In this test it is evident that the open circuit potential of the disk (-911mV) is much closer to that of the BM sample (-1032mV) and essentially the same as that seen for the BM sample in water. This would indicate that the BM sample is experiencing a localized surface pH change in un-buffered water not seen in the bulk solution and is evidence that ball milling may produce micro climates.  0.5 0.0 Al Disk Potential (V)  -0.5  BM Al/Al2O 3  -1.0 -1.5 -2.0 -2.5 1.E-11  1.E-09  1.E-07  1.E-05  1.E-03  1.E-01  Apparent Current (A/cm 2)  Figure 35 Potentiodynamic scans of an aluminum disk and a BM al/alumina pellet in a 0.01M Borax solution at 25oC.  This does not address the question of to what extent a pH gradient can be maintained when vigorous stirring in the form of evolving hydrogen gas is present as in 84  the case of BM powders at higher temperatures. Micro climates are also discussed with regards to temperature increases in Section 5.4.  5.3.2 Hydrogen Evolution from Pure Aluminum If pH is increasing locally, the extent to which it accelerates aluminum corrosion can be estimated from the corrosion rates seen with un-milled aluminum powder which is used to avoid the temperature spikes seen in the more vigorous BM powders. The slower corrosion rate of un-milled powders allows for better control of solution temperature and pH. As seen in Figure 36, the corrosion rate for solutions at a pH of 10 is unchanged from that seen with DI water (nominally pH 7), no change was seen for pH values between 7 and 10. No experiments have seen a pH from the evolution of hydrogen over roughly 9.5, even with multiple runs in small solution volumes. With a pH of 11 the corrosion rate in the first hour is roughly doubled that seen from DI water, but still far less than observed for BM powders.  85  120  Volume of Hydrogen (ml)  100 80 60 Water Water #2  40  pH=10 pH=11  20 0 0  1  2  3  4  5  6  Time (hours)  Figure 36 H2 evolution from aluminum powder in DI water (pH 7) and NaOH solutions (pH 10 and 11).  Unless localized pH values in BM samples are approaching 12 the pH shift cannot account for the increased corrosion rate seen with BM samples. While high pH microenvironments can not be ruled out there is better evidence that high temperature microenvironments are not forming (next section). Given that the open circuit potential of BM powders seems to mirror that seen on bulk aluminum at a pH of 9, is there any evidence supporting higher local pH swings? With the finding in section 5.5 that surface area can account for the observed corrosion rate there is no pressing need to postulate high local pH shifts.  86  5.4 Temperature Effects Aluminum corrosion is an exothermic process releasing 4.3 kWh/kg of Al (WANG, 2008) of heat as it reacts and in small volumes of water can be seen to raise the temperature (Figure 17 page 62) of the solution. In Section 5.2 there was evidence presented that the surface of a ball milled powder was experiencing an elevated pH from the bulk solution, possibly in small pockets where a separate micro climate was able to form. If these micro climates were forming the question arises would temperature also be affected locally and if so to what extent would that drive the reaction rates.  5.4.1 Hydrogen Evolution for BM Powders As would be expected the rate of reaction increases with higher starting temperatures for BM powders (Figure 37) and the time to initiate the reaction decreases as the higher water temperature can more quickly warm the room temperature powder up to temperatures where the reaction begins to proceed at a noticeable pace. Higher temperatures will also drive quicker diffusion of ions through the barrier films on the aluminum surface. However, once the reaction starts to proceed vigorously there is considerable localized stirring from the evolution of hydrogen gas and the question becomes is the surface temperature of the particle different from that of the bulk solution which can be measured? One possible indication of this can be found by looking at the reaction products that form at various temperatures. By looking at the effect of temperature on solid aluminum disks, we can eliminate the possibility of small pockets formed by the mechanical deformation of milling. The reaction rates with water are slow  87  enough to minimize local heating and the temperature of the bulk solution taken as the surface temperature. 350 85oC  Volume of Hydrogen (ml)  300 250 200 150  65oC  55oC  100 50 0 0  20  40  60  80  100  Time (minutes)  Figure 37 Hydrogen evolution rate for 0.5g Al BM with 0.5g Alumina in 100mL DI water at 55oC, 65oC, and 85oC starting temperatures.  5.4.2 Electrochemical Tests Potentiodynamic scans (Figure 38) show clear increases in corrosion current as temperature is increased for aluminum powder and values taken from the scans at +200mV correlate well with currents from the potentiostatic tests on aluminum disks (Figure 23), both reading around 25uA at 25oC and 180uA at 50oC. 88  2000 1500  Potential (mV)  1000 25oC  500 0  50oC  -500 -1000 -1500 -2000 1.E-09  1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  1.E-02  2  Apparent Current (A/cm )  Figure 38 Potentiodynamic curve of aluminum powder in water at 25oC and 50oC.  1000 900 800 Current (uA)  700  70oC  600 500 400  85oC  300 200 100 0 50.0  55.0  60.0  65.0  70.0  75.0  Time (hours)  Figure 39 Potentiostatic test of aluminum disk held at +200mV in 0.01M Borax solution at 70oC and 85oC.  89  When looking at the aluminum disk held at a constant potential (Section 5.2.1) a noticeable drop in corrosion rate was seen at 85oC; an expanded graph is Figure 39 above. This type of almost exponential drop is indicative of a new film growing on the surface and passivating the aluminum. The following section verifies that this drop in rate is also seen with hydrogen evolution from powders.  5.4.3 Hydrogen Evolution Rate for Al Powders 450 400  70oC  Volume of Hydrogen (ml)  350 300 250 200  95oC  150 100 50  55oC  0 0  10  20  30  40  50  60  70  80  Time (hours)  Figure 40 Hydrogen evolution with time for aluminum powder 0.5g in DI water at 55oC, 70oC, and 95oC.  Figure 40 shows the evolution rate for hydrogen from the corrosion of aluminum powder (0.5g) at three temperatures. From 55oC to 70oC, we see the expected increase in rate with temperature. At 95oC, the rate is sharply higher initially, but starts to fall off 90  much sooner and by 6 hours the total evolved hydrogen has crossed under that seen at 70oC. When looked at as a percentage of total aluminum that corrodes (Figure 41), the amount of corrosion before passivation stops the process is less than half at 95oC as seen at 70oC. This decrease in overall corrosion above 70oC is in agreement with earlier studies (UHLIG, 1948), (VARGEL, 2004). 60  70oC  Al corroded (%)  50 40 30 95oC 20 10 55oC 0 0  20  40  60  80  Time (hours)  Figure 41 Percentage of aluminum corroded with time for 0.5g of aluminum powder in DI water at 55oC, 70oC, and 95oC.  5.4.4 XRD Characterization of Corrosion Products In Figure 42, we see the XRD results for two aluminum disks each placed in a beaker of 0.01M Borax and held at temperature for 24 hours. The sample held at 55oC shows only aluminum peaks. Any film on the surface must be less than 5% or so of the x-  91  ray sample volume which is roughly the detection limit. The sample held at 85oC is starting to show distinct AlOOH peaks and indicates a much thicker surface film.  10000 9000 8000  Al  Intensity (cps)  7000 6000 55oC  5000 4000 3000  AlOOH  2000 1000  85oC  0 0  20  40  60  80  100  2θ (degree)  Figure 42 XRD pattern for aluminum disks held in 0.01M Borax solution for 24 hours at 55oC and 85oC.  When BM powders are reacted, a much more complete reaction (over 50%) is seen, making detection of reaction by-products easier. These tests were carried out by Czech (2006) and the reaction products at 55oC showed only Al(OH)3 peaks, at 95oC only AlOOH peaks were seen, and at temperatures in between both sets of peaks visible. This indicates two important things; first, if BM samples in 55oC water are not showing AlOOH reaction products then localized micro climates in temperature must not be forming, or have only a shallow temperature gradient to the bulk solution. Since the  92  reaction of 0.5g Al in 30ml of water can raise temperatures by 30oC, the smaller volumes in proposed pockets would experience far higher temperature increases if the thermal gradient was significant. The second finding of note is that BM powders do form AlOOH reaction products at higher temperatures, which is shown to decrease the overall reaction before passivation when looking at non-milled particles, but no decrease in total corrosion was seen for BM powders at high temperature (Figure 37). Therefore, a significant portion of the corrosion on BM surfaces must be taking place before the passivating film grows thick enough to hinder the reaction, and furthermore there is a secondary mechanism which causes the rate to taper off, not linked to the reaction products which would favor the lower temperature reaction. This brings us to consider changes in surface area.  5.5 Surface Area Effects That the amount of surface exposed to solution will directly contribute to the overall corrosion rate in a linear way is obvious, but two things kept this from being the first line of investigation when looking at the success of ball milling in promoting corrosion. First, aluminum is reported to passivate in water (ALWITT, 1966), and as seen in Figure 16 very little aluminum corrodes in comparison to the BM sample so a mechanism to prevent passivation was sought. Second, it was difficult to determine what the aluminum surface area would be after ball milling with secondary particles. Milled by itself the aluminum did not break into smaller particles, but tended to clump together (CZECH, 2006). Measurement tools such as the BET would not distinguish between the 93  surface of aluminum and that of additives which needed to be added to prevent the clumping. Optical sizing techniques would not give accurate surface area measurements as much of the surface area apparently was due to surface morphology. A mathematical estimation of surface area per gram using particle size and density and assuming a spherical particle is shown in Figure 43. 0.25  2 Surface area (m /g)  0.20  0.15  0.10  0.05  0.00 1  10  100  1000  10000  Radius (um)  Figure 43 Surface area in m2 g-1 for a spherical particle of aluminum of given radius.  For an average incoming particle size of 35um one would calculate a surface area of around 0.06m2/g. However BET measurements (see Figure 48) give a value 10 times higher. This makes determining the effective surface area by SEM difficult. The development of the ‘Washed’ aluminum process was key in allowing the determination of surface area by BET after milling without any secondary milling particles present. The discovery that Al2O3 when used as secondary particle did not change its surface area on a 94  per gram basis during milling allowed for the resulting surface area of aluminum after milling to be calculated from BET results and therefore the corrosion rates to be compared as a function of surface area.  5.5.1 Changing Amount of Secondary Particles It is evident from Figure 44 that increasing the relative amount of oxide added to the mix before milling has a direct and almost linear effect on the amount of aluminum that corrodes in the first hour after immersion in water. With no compositional change occurring this is a strong argument that rates may be surface area driven.  70 25% Al  60  Was hed Al  Al corroded (%)  50 40  50% Al  30 20  75% Al  10 0 0  10  20  30  40  50  60  70  Time (minute)  Figure 44 Percentage of 0.5g of aluminum that corrode with time in 65oC water after BM with 25%, 50%, or 75% alumina by weight, or 50% KCl that is subsequently washed from the powder.  95  With sufficient oxide the rate of corrosion seen in the washed aluminum system is surpassed, although this would not be efficient from a weight perspective. 450 25% Al  Volume of Hydrogen (ml)  400 350  W ashed Al  300 50% Al  250 200 150  75% Al  100 50 0 0  10  20  30  40  50  60  70  Time (minute)  Figure 45 Volume of hydrogen gas evolved from 0.5g of aluminum in 65oC water after BM with 25%, 50%, or 75% alumina, or 50% KCl that was subsequently washed from the powder.  Figure 45 shows the same data plotted as a volume of hydrogen released instead of the percent of sample corroded to allow for comparison with Figure 51 which will graph the data scaled to a constant surface area instead of a constant mass. This will be discussed in Section 5.5.4. If the total surface area of the additive particles is more important than simply the mass of secondary particles, then a difference in corrosion rates should be seen if we mill the aluminum powder with secondary particles of different surface areas. A Si3N4 powder (8.6 m2/g) with a surface area similar to that of the alumina (9.64 m2/g) was used as well as a SiC powder (0.73 m2/g) which had a much lower surface area closer to that of the un-milled aluminum powder (0.6 m2/g). Both of these powders were ball milled with 96  aluminum using an equal weight of powder. In Figure 46 it can be seen that corrosion rates were high for the high surface area additives and significantly lower with the lower surface area SiC particles. In all three cases, lowering the mass of secondary particles relative to the amount of aluminum powder resulted in drastically lower hydrogen evolution rates.  300  50% Si 3N 4 50% Al 2O3  Volume of Hydrogen (ml)  250 200  50% SiC  150  25% Al 2O3  100  25% Si 3N 4  50 25% SiC  0 0  10  20  30  40  50  60  70  Time (minutes)  Figure 46 Volume of hydrogen gas evolved from 0.5 g of aluminum in 65oC water that was BM with either 50% or 25% by weight of Al2O3, Si3N4, or SiC.  97  5.5.2 Surface Area of Secondary Particles Using a Quantachrome Autosorb-1 Surface Area Analyzer, a nitrogen adsorption – desorption isotherm can be plotted (Figure 47) for each powder sample. 0.20 0.19 0.18  Volume (cc/g)  0.17 0.16 0.15 0.14 0.13 0.12 0.11 0.10 0.00  0.05  0.10  0.15  0.20  0.25  0.30  0.35  Relative Pressure (P/P o )  Figure 47 Isotherm plot for 8g of aluminum powder after 72 hours under vacuum at 150oC.  Using the linear portion of these plots a surface area per gram can be calculated using Brunauer-Emmett-Teller (BET) theory. For the as-received aluminum powder, Figure 48 shows the BET plot and gives a surface area of 0.59 m2/g. This can be compared to the plot for the standard BM mixture of aluminum and alumina (Figure 49) which shows a surface area of 5.3 m2/g. Similar plots were done for washed aluminum, Si3N4, SiC, Al2O3, and various BM mixtures of alumina powder. These results are summarized in Table 2, and the respective BET plots located in section A of the appendix.  98  2000.00 1800.00 1600.00  1/[W((Po/P)-1)]  1400.00 1200.00 1000.00 800.00 600.00 400.00 200.00 0.00 0.00  0.05  0.10  0.15  0.20  0.25  0.30  Relative Pressure (P/P o )  Figure 48 BET plot for 8g of Aluminum powder giving a surface area of 0.59 m2 g-1. Table 2 Surface areas for BM and as received powders from BET.  Sample  Surface Area m2/g  Aluminum powder as received  0.6  Washed Aluminum (50% KCl)  13.76  Al2O3 as received  9.64  BM Al/Al2O3 50%/50%  5.33  BM Al/Al2O3 25%/75%  7.85  BM Al/Al2O3 75%/25%  3.67  SiC  0.73  Si3N4  8.6  99  0.35  250.00  1/[W((Po/P)-1)]  200.00  150.00  100.00  50.00  0.00 0.00  0.05  0.10  0.15  0.20  0.25  0.30  0.35  Relative Pressure (P/P o )  Figure 49 BET plot for 1.97g of BM aluminum/alumina powder 50%/50% with a surface area of 5.3 m2 g-1.  5.5.3 BET for Varying Oxide Percentage Using the data from Table 2 in the previous section we can plot Figure 50 which shows that the values of surface area for the three ball milled powders using different percentages of oxide when graphed by oxide percentage, all line up with the value found for the un-milled oxide. This can either be an extraordinary confluence of changing particle sizes and surface areas between the aluminum and the oxide, or can suggest more simply that the oxide is not changing its surface area per gram when being ball milled.  100  12  2 Surface Area (m /g)  10  8  6  4  2  0 0  20  40  60  80  100  120  Oxide (%)  Figure 50 Surface area calculated from BET for aluminum BM with 25%, 50%, or 75% Al2O3 and 100% Al2O3 that was not BM.  If the alumina surface area is not changing with different BM mixtures, then Figure 50 can be interpreted in two ways. In the first case, the aluminum simply has a surface area given by the y-intercept of approximately 1.5 m2/g after BM and the remaining area is given by the oxide; or in the second case, the softer aluminum powder is smeared out during the milling process over all available oxide area or reduced to small particles that fit within the surface crevices or voids of the oxide. In the second case the measured surface area is equal to the aluminum surface area for each mixture assuming there is enough aluminum to cover the secondary particles. If the first case is true then when reacting equal masses of aluminum BM with different amounts of oxide, there would be 101  no difference in surface area of aluminum exposed so the reaction rates would be equal except for possible changes the amount of oxide might make with regards to temperature and pH, as discussed in previous sections. As was seen earlier in Figure 44 and Figure 45, it is clear that there are large differences in the corrosion rate which would indicate the surface area must be changing. If the second case is true it would explain the large differences seen with varying oxide amounts and if samples of equal surface area were reacted then similar corrosion rates would be expected.  5.5.4 Hydrogen Evolution for Constant Surface Area Using the surface areas from Table 2, the evolution of hydrogen in Figure 45 was scaled for each sample to the same 6.88 m2 seen for .5 g of washed aluminum, and plotted in Figure 51. There it is seen that BM samples with 75% aluminum and 50% aluminum have substantially equal corrosion rates to that seen from washed aluminum after 1 hour. Washed aluminum being a high surface area, but otherwise pure aluminum sample, it would have no other reactions occurring beside the aluminum-water interaction. This would indicate that substantially all of the effect of ball milling on corrosion rates comes from increasing the surface area. The decreased hydrogen evolution seen with the 25% aluminum sample can be explained by assuming there was insufficient aluminum present to completely cover the oxide. This would result in a high rate as was seen when graphed as a function of mass, but the actual surface area would be overestimated from the BET plots since the BET would measure the surface area of the oxide as well and not just the aluminum, so would show poorer results graphed as a function of surface area. 102  400  W ashed Al  75% Al  Volume of Hydrogen (ml)  350  50% Al  300 250 25% Al 200 150 100 50 0 0  10  20  30  40  50  60  70  Time (minute)  Figure 51 Volume of hydrogen gas evolved from aluminum powder in 65oC water after BM with 25%, 50%, or 75% alumina, or 50% KCl and subsequently washed, with all volumes scaled to an equal surface area of 6.88 m2.  If only surface area is affecting the rate, then a sufficient amount of un-milled powder should have a similar rate as the washed aluminum. In Figure 52, we see that the aluminum powder actually shows a much higher rate of hydrogen evolution than the washed aluminum, although the amount of aluminum needed to get this rate is much higher. Upon reflection this would also be expected, because as the initial surface of the powder is reacted the remaining surface area of the un-reacted center of the particle is almost unchanged when considering a thin initial corroded layer. For a highly deformed BM powder however, you do not have a spherical particle, much of the surface area can come from thin strips or folds that are quickly consumed in the initial moments of 103  reaction. So the effective surface area drops off as the much lower mass of BM powder reacts and the rate dies as all the aluminum is consumed, while in the case of the nonmilled powder, there is plenty of remaining aluminum and only slight decrease in surface area so only the buildup of the passivating layer slows the reaction. 1200 Al Powder - 11.3g  Volume of Hydrogen (ml)  1000 800 600 400  W ashed Al - 0.5g 200 0 0  10  20  30  40  50  60  70  Time (minute)  Figure 52 Volume of hydrogen gas evolved from aluminum powder and washed Al in 65oC water with surface area of aluminum equal to 6.88 m2.  Finally, in Figure 53, the volume of hydrogen released in each minute is plotted as a function of time for both washed aluminum and as received powder, again with a total exposed surface area of 6.88 m2. The washed aluminum has a peak rate both earlier and lower than the plain powder. The overall corrosion starts and reaches a peak earlier for the washed aluminum because there only roughly 5% of the mass needs to heat up to start reacting compared to the plain powder. The fact that the rate reaches a peak at a lower overall corrosion rate is thought to be a result of the processing done on the washed 104  aluminum. The initial milling involves high temperature collisions in an oxygen atmosphere that will tend to promote oxide growth, and the subsequent washing process to dissolve the salts from the mixture will see some growth of a hydroxide layer even in cold water. Both of these mechanisms cause the washed aluminum to start with a more protective layer in place which slows the reaction. 160  Volume of Hydrogen (ml)  140  Al Powder  120 100 80 60  W ashed Al  40 20 0 0  2  4  6  8  10  12  14  16  Time (minute)  Figure 53 Rate of hydrogen evolution per minute for equal surface area of aluminum powder and washed Al (6.88 m2) in 65oC water.  One final consideration is as follows: can the increase in surface area for a washed aluminum sample be reversed with no lasting effect from the processing involved? If washed powder is compacted in a press to form a pellet there is insufficient area to measure evolved hydrogen, but a comparison can be done with a potentiostat. A comparison between pellets made from washed aluminum (Figure 54) and aluminum  105  powder (Figure 28) shows very similar corrosion currents and open circuit potentials and little lasting difference from the ball milling process.  1.0 5 ºC  0.5  15 ºC 25 ºC  Potential (V)  0.0  35 ºC 45 ºC  -0.5  55 ºC  -1.0 -1.5 -2.0 -2.5 1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  1.E-02  1.E-01  1.E+00  Apparent Current (A/cm 2)  Figure 54 Potentiodynamic scans of a washed aluminum pellet in a 0.1M Na2SO4 solution at temperatures from 5oC to 55oC after cycling two times.  5.5.5 TEM Images of BM Aluminum A TEM was used to examine a BM Al/Alumina powder for evidence of very small high surface area particles and aluminum coating the alumina particles. The BM powder was suspended in ethanol by ultrasonically agitating for 1 minute. 2-3 drops of this suspension were dropped onto clean copper grids of 3.02mm diameter and air dried for 10-15 minutes. Bright field showed particles in the 10 – 20 nm range (see Figure 55) and dark field inspection (see Figure 56) showed at least some were a crystalline aluminum which showed up as bright compared to the amorphous appearing alumina.  106  Figure 55 Bright field of 10-20 nm particles from BM Al/Al2O3.  Figure 56 Dark field of 10-20 nm particles from BM Al/Al2O3.  When looking at clusters of larger particles it was hard to distinguish between aluminum and alumina, but under dark field a bright glow from the aluminum can be seen around the edges of many particles.  107  Figure 57 Dark field of several particle clusters.  The dark field is only responding to one orientation of the aluminum crystalline structure at a time so the observed bright areas are only a subset of the total aluminum in the picture that happens to meet the given diffraction criteria.  5.6 Summary In conclusion, ball milling with additives is an effective way of increasing the corrosion rate of aluminum in water. The use of water soluble salts is most effective from a weight efficiency standpoint because of the ability to remove the secondary particles after milling and before use in the hydrogen generating reaction. There is no evidence of a contribution from a chemical reaction with the secondary milling particles used playing a role in the subsequent corrosion rate or overall efficiency. As it is an exothermic reaction, the temperature does rise in the solution, particularly with small volumes of water relative to the amount of aluminum being reacted. This temperature increase is beneficial from a reaction rate standpoint and will accelerate the reaction as it proceeds. This eliminates the need to pre-heat completely to the desired operating temperature, but 108  does raise the issue of thermal management to reactor designs. There is a change in reaction by-product when the temperature increases above 60oC with the AlOOH hydroxide that forms at higher temperature consuming less water in the reaction while producing the same amount of hydrogen per mole of Al. Ball milled systems tend to react to completion before the buildup of a passivation layer stops the reaction so the more passivating nature of AlOOH which is seen for un-milled aluminum powder is not a problem with BM powders. The hydrogen generation seems to be alkalizing and tends to raise the pH of solution above 9 if small volumes of water are used. This pH shift will deter passivation of the powder over the long term, but again is not a significant factor in the time frame that BM systems react in. There was no evidence to support milling forming microclimates in small pockets next to the surface with respect to temperature and the expected reaction products that formed. By far the greatest effect of ball milling was on the surface area of the powder and the changes in corrosion rate and hydrogen evolution can be tied to changes in surface area. Beside the increase in surface area which both promotes rapid hydrogen evolution and allows for a greater percentage of the aluminum to corrode before passivating, the actual BM process does tend to build up a thicker oxide layer compared to the native oxide found on incoming aluminum powders. This has little practical effect on the suitability of using BM powders for hydrogen generation as the overall yield from nonmilled powders is to low for economical use as a hydrogen source. For hydrogen generation purposes it is the surface area that needs to be optimized for the time, energy, and cost of the milling. In general secondary particles with high surface area to mass  109  should be used, and for high hydrogen per gram powder production it is necessary to use secondary particles that can be removed from the powder after milling and before use.  110  Chapter 6 Enhancing Corrosion of Al without Ball-Milling In the previous chapter the mechanism behind the corrosion rate increase seen with BM aluminum powders was investigated, and the results can be used for improving hydrogen generation. However, the BM systems will continue to focus on secondary particles that can be removed from the aluminum after milling such as water soluble salts. Because of the cost of aluminum, in terms of electricity used, CO2 emitted, and value of the metal for other uses is so high, it will not compare favorably with hydrolysis or SMR as a source for hydrogen. The driving feature behind aluminum as a hydrogen source is portability which makes carrying any secondary particle expensive from a weight efficiency standpoint. When looking at washed aluminum however, there is the time and cost of the initial milling and then the cost of the subsequent wash steps to remove the secondary particles. From a technology standpoint it would be very useful to have a system for hydrogen generation that can use aluminum without the need for milling or washing, can work with 100% aluminum as the hydrogen source and still retain the benefits of using neutral pH water at moderate temperatures. Investigation of such systems is the focus of this chapter which revolves around the interaction of alumina with aluminum in non-milled systems. The ability to get high reaction rates and high hydrogen yields with the alumina serving as a reaction promoter is explored in Sections 6.1 and 6.2 and two theories are presented in Sections 6.3 and 6.4 along with tests to validate the initial hypothesis.  111  6.1 Corrosion of Aluminum Powder with added Alumina When studying the effect of BM with alumina powder, as discussed in the preceding chapter, tests were also run with an aluminum/alumina mixture that had not been BM to serve as a control for the BM samples and to compare with the straight powder samples. Because of the ductile nature of aluminum, a pure aluminum pellet was considered likely to compress into a less porous pellet during compaction and this porosity might be a difference between the pure aluminum pellet and the BM sample. It was still somewhat surprising when the non-ball milled sample showed a higher corrosion rate than the straight aluminum powder (Figure 58). With only one half the aluminum present as in the 100% sample the corrosion rate was expected to be lower for the aluminum/alumina sample. To see a higher rate implied that the oxide was providing a pathway for greater solution penetration of the pellet, exposing greater overall aluminum surface, or that the oxide was facilitating the corrosion from the available surface. This could have major implications regarding Al corrosion in various environments not related to hydrogen generation.  6.1.1 Electrochemical Tests To investigate the role the alumina might play and confirm the first results, samples of varying oxide percentage composition were prepared and tested, (Figure 59). If the oxide was only serving as a solution pathway to aluminum deeper within the pellet, then a point of diminishing returns might stabilize or reverse the apparent acceleration of corrosion seen in the first test. 112  2.5 Al/Al2O 3  2.0 1.5  Al Powder  Potential (V)  1.0 0.5 0.0 -0.5  BM Al/Al2O 3  -1.0 -1.5 -2.0 -2.5 1.E-09  1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  1.E-02  2  Apparent Current (A/cm )  Figure 58 Potentiodynamic curve of an aluminum/alumina mixture in water compared to aluminum powder and a BM aluminum/alumina mixture.  Instead, an increased corrosion rate was seen with each increase in the amount of oxide in the pellet. This suggested that the oxide was promoting the aluminum corrosion, but did not rule out that it simply allowed for more surface area of aluminum particles to be exposed to water. To test this idea, corrosion rates for loose powder in a beaker would be measured by the evolution of hydrogen, where it could be assumed that all of the available aluminum was exposed to water regardless of the presence of oxide.  113  1.2 1.0 0.8  66% Al  Potential (V)  0.6  33% Al  75% Al  50% Al  0.4 0.2 0.0 -0.2 -0.4 -0.6 1.E-09  1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  Apparent Current (A/cm 2)  Figure 59 Potentiodynamic scans for aluminum/alumina pellets in DI water with the amount of aluminum varying from 33% to 75%.  6.1.2 Hydrogen Evolution Figure 60 shows the hydrogen evolution from 0.5g of aluminum powder placed in 65oC DI water with time. There is a clear and dramatic increase in the amount of hydrogen evolved when the aluminum powder was mixed with 5g of Al2O3 powder. This increase while slower than that seen with the BM systems did result in the aluminum powder reaching corrosion levels and thus evolved hydrogen amounts equal to or greater to those seen with the BM systems.  114  600  Volume of Hydrogen (ml)  500 5g Al2O 3 + 0.5g Al 400 300 200 0.5g Al  100 0 0  50  100  150  200  250  Time (minutes)  Figure 60 Hydrogen evolution from 0.5g of aluminum powder in 75ml of 65oC DI water with 5g of alumina powder added.  As can be seen in Figure 61, the corrosion rate for a given amount of aluminum powder continues to increase although with diminishing returns as the amount of oxide added to the beaker is increased. It can also be noted that all the oxide amounts showed a much higher hydrogen evolution in the second and third hours of immersion, and the rapid fall-off in hydrogen evolution seen with the pure aluminum (zero oxide point), which is attributed to aluminum passivation, is no longer observed when alumina is present. This corrosion acceleration is not just a mechanical effect of being buried by a secondary particle. Figure 62 shows aluminum powder mixed with similar amounts of SiO2, CuO, La2O3, and SiC, none of which show noticeably more hydrogen evolution than that seen by just aluminum powder alone in water.  115  1200 1hr 2hr  Volume of Hydrogen (ml)  1000  3hr 800 600 400 200 0 0  2  4  6  8  10  Al 2O 3 Pow der (g)  Figure 61 Volume of hydrogen released in one, two, and three hours after immersion in 65oC water as a function of the number of grams of Al2O3 powder added to 1 gram of aluminum powder.  Volume of Hydrogen (ml)  350  SiC 5g  300  Al2O3 5g CuO 1g  250  w ater La2O3 5g  200  SiO2 5g  150 100 50 0 0.00  1.00  2.00  3.00  4.00  5.00  6.00  Time (hours)  Figure 62 Volume of hydrogen released with time for 0.5g Al in 70oC water with additions of Al2O3, SiC, CuO, La2O3, and SiO2.  116  6.2 Subsequent Runs While the effect of alumina on aluminum corrosion is very interesting by itself and invites explanation, if new alumina is needed with each batch of aluminum powder then it will not be a useful method for hydrogen production because of the weight inefficiency in transport of the inert alumina in addition to the aluminum powder. To be a commercially useful technology the alumina should be reusable, allowing multiple additions of aluminum powder without loss of corrosion rate or total efficiency.  800 Volume of Hydrogen (ml)  700 600 500 400  Run 1  300  Run 2  200  Run 4  100  Run 5  Run 3  Run 6  0 0  50  100  150  200  250  Time (minutes)  Figure 63 Volume of hydrogen released from 5g Al2O3 in 75ml of 65oC DI water with the addition of 0.5g Al powder at the start of each run, run 6 started one day later.  As seen in Figure 63, subsequent additions of 0.5g aluminum powder to a beaker containing 5g of Al2O3 not only maintain the rate of corrosion seen earlier, but accelerate the corrosion rate with the first 3 runs and then remain constant with a total corrosion in the first hour above that seen for the 50/50 BM mixtures. 117  A similar addition of 0.5g of aluminum, into 100ml of water containing 5g of alumina, every 30 minutes show that by the third run over 90% of the theoretical maximum hydrogen is evolving each run (Figure 64). This was maintained without pausing for 13 runs without loss of hydrogen evolution rate, and continued over the next two days for a total of 21 runs without loss of the catalytic role the alumina played. These tests were carried out on a hotplate which established the starting temperature at 65oC, although the exothermic nature of the reaction would cause the temperature to increase during the reaction. A portable system for hydrogen generation would not want to need continual heating so a test was also run with an insulated beaker.  800 Maxim un Hydrogen Evolution from 0.5 g of Alum inum  Volume of Hydrogen (ml)  700  Run 3  600  Run 2  500 400 Run 13 Run 21  300 200  Run 1  100 0 0  5  10  15  20  25  30  35  Time (minutes)  Figure 64 Volume of hydrogen evolved when 0.5g of Al powder is added every 30 minutes to 5g alumina in 75ml of 65oC DI water.  The test was started on a hotplate for the first 4 runs as in the previous test, but with the beaker insulated. At the start of run 5, the beaker was removed from the hotplate 118  and set on the counter and aluminum added every 5 minutes (Figure 65). In addition to demonstrating the ability to run without external heat, this tests shows that if a long term demand for hydrogen allows for time to stabilize a batch system, then overall hydrogen production rates can be set by the rate at which aluminum is fed into the system. A small hydrogen tank added to the system would allow for instant startup, with heat from the fuel cell starting the batch reactor with the flow of hydrogen produced controlled by the speed aluminum is added to the reaction chamber. When demand is turned off the remaining un-reacted aluminum in the chamber would finish producing hydrogen to refill the startup tank and ready the system for the next start. 700 Run 29 Run 23 Run 21  600  Volume of Hydrogen (ml)  Run 18  500  Run 17 Run 12 Run 10  400  Run 7 Run 6  300 Run 5  200  100  0 0  1  2  3  4  5  6  Time (minutes)  Figure 65 Volume of hydrogen released with the addition of 0.5g Al every 5 minutes to 5g Alumina in 75ml of DI water in insulated beaker.  119  One of the differences in the system with subsequent runs is the presence of aluminum hydroxide produced earlier. To test if this would help explain the acceleration seen with multiple runs a 0.5M boehmite solution was used to run a cathodic polarization curve on an aluminum powder pellet and this was compared with an aluminum powder pellet run in a 0.045 M KCl solution. Both solutions were at 25oC and de-aerated with N2 for 1 hour while the solution was stirred (Figure 66). -0.7 Al powder Al in boehmite 0.5M  -0.8  Potential (V)  -0.9 -1.0  Stirring Speed Increased  -1.1  Stirring Turned Off  -1.2 -1.3 1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  1.E-02  Apparent Current (A/cm 2)  Figure 66 Cathodic polarization scans of aluminum powder in 0.045 M KCl and 0.5 M boehmite solutions at 25oC.  The boehmite solution showed an increased cathodic current by ~1 order of magnitude, with the open circuit potential being almost identical to the aluminum control sample. Furthermore, the sample in the boehmite solution showed the effect of bringing fresh boehmite to the aluminum surface with the current increasing as the solution stirring was increased and then dropping when the stirring was turned off. This is typical 120  of a sample influenced by a concentration gradient. The effect of the boehmite solution was also tested on loose aluminum powder (see Figure 67). Compared to water with 5g of alumina present, the corrosion of 0.5g Al powder was much quicker in a 0.5 M boehmite solution without alumina, but the solution quickly gelled and the hydrogen evolution was hindered. The addition of alumina gave intermediate results, with the boehmite giving increased initial hydrogen evolution and the oxide preventing the rapid gelling of the solution.  300 Al2O3 5g  Volume of Hydrogen (ml)  250  0.5M boehm ite s olution 0.5M boehm ite + 5g Al2O3  200 150 100  50 0 0.00  0.20  0.40  0.60  0.80  1.00  1.20  Time (hours)  Figure 67 Hydrogen evolution with time from 0.5 g of aluminum powder in DI water with 5g of Al2O3 compared to 0.5 g of aluminum powder in a 0.5 M Boehmite solution with and without the 5 g of alumina.  121  6.3 Proton Donor Model From the tests described in the preceding two sections we can make three statements of observed facts. One, the addition of alumina (or AlOOH) powders accelerates the corrosion rate of aluminum powder. Two, the observed rate increases as the amount of alumina added increases. Three, the total percentage of aluminum that corrodes before passivation stops the reaction is higher with alumina present than observed for pure aluminum powder. To explain these observations two models or hypothesis are investigated. In this section we will look at the Proton Donor Model and in the following section the Deposition Model. For the aluminum corrosion reaction to take place, there are two reactions that occur and must be balanced. There must be a supply of OH- ions to combine with the Al3+ ions, and there must be a supply of H+ ions to carry off the excess electrons as H2 gas to maintain charge neutrality in the aluminum. Because the aluminum metal is a conductor these reactions can take place at different locations, but the aluminum corrosion to aluminum hydroxide cannot proceed faster than the hydrogen evolution reaction consumes the excess electrons and so it is reasonable to hypothesize that availability of H+ will control the reaction speed. Because the iso-electric point (IEP) of alumina is above pH 8 (BRUNELLE, 1978), when placed in a neutral pH 7 water, the surface of Al2O3 particles will acquire a positive charge of H+ ions, with total charge proportional to the surface area. In the Proton Donor model I am hypothesizing that when alumina particles come in contact with the aluminum particles this layer of protons acts to concentrate the number of  122  protons available for reduction at the aluminum surface. To be valid a model should make predictions of the system behavior under different conditions that can be tested, and help explain the observed results. In this section we will test the following hypotheses that come from the model: 1. If the interaction with aluminum is driven by a positively charged layer of protons, then we would expect particles with an IEP below 7 to have a negatively charged layer due to OH- adsorption and not enhance the corrosion rate. 2. Since the amount of adsorbed protons is a function of surface area and is thus greater in systems with high surface area, a secondary particle with low surface area should not greatly affect the aluminum corrosion rate. 3. Aluminum corrosion could also be increased by helping the diffusion of OH- ions to the aluminum surface. If it is actually the proton concentration that is involved then measuring the cathodic slopes on a potentiodynamic scan from the open circuit potential downward should show a change in the hydrogen exchange current density hio on the aluminum surface with the addition of alumina. 4. In a similar fashion, if alumina is serving as a proton donor then its addition should also increase the hydrogen exchange current density hio on platinum as well. 5. Since this model is supposing contact or close proximity to the aluminum surface to allow for transfer of electrons, then a non-conducting barrier between the aluminum and alumina should prevent the acceleration of corrosion while a conducting barrier would not.  123  6.3.1 Effect of Iso-electric Point If the proton donor model is correct then it is the layer of protons around the alumina particle that accelerate the aluminum corrosion. A particle with an iso-electric point below pH 7 would form a negative surface charge in neutral or alkaline solutions and therefore have low concentration of protons on the surface. The presence of these particles would not be expected to increase the corrosion rate. In Figure 68, there are two curves showing the evolution of hydrogen from 0.5 g of Al powder in 70oC water over 3 hours and two curves showing the evolution with the addition of 5 g of Si3N4 (IEP 6-7) (HARUTA, 2004) which would not supply protons.  100 90 W ater  Volume of Hydrogen (ml)  80 70 60 50 40 30  Si3N4 5g  20 10 0 0  0.5  1  1.5  2  2.5  3  3.5  Time (hours)  Figure 68 Volume of hydrogen evolved with time for 0.5 g of Al powder in 100ml DI water at 70oC with and without the addition of 5 g Si3N4.  124  Both tests with the Si3N4 showed good agreement with each other and a significant reduction in the evolution of hydrogen with the negative charge appearing to have a suppressing effect on the aluminum corrosion.  6.3.2 Surface Area Effect Since the electric surface charge of a particle in solution is a function of surface area, the amount of protons or OH- ions involved will also depend on particle surface area. In Figure 69 we can see that the addition of SiC (IEP 2-2.5, JACOBSON, 1991)  600 Al2O 3 5g  Volume of Hydrogen (ml)  500  400  300  200 SiC 5g 100  W ater Si3N4 5g  0 0  0.5  1  1.5  2  2.5  3  3.5  4  4.5  Time (hours)  Figure 69 Difference in hydrogen evolution from 0.5g Al powder in 70oC DI water, between high surface area particles Al2O3 and Si3N4 ( >8m2/g) and low surface area particles SiC (<0.8m2/g).  particles which have a low surface area (0.73 m2/g, see Table 2) shows a similar evolution rate as seen by just aluminum powder alone. Both Al2O3 and Si3N4 have over 125  10 times the surface area (see Table 2, page 99) and show a marked difference from plain aluminum powder, both increased (Al2O3) and decreased (Si3N4). In a similar fashion the curves for coarse La2O3 and CuO (see Figure 62) which have high reported IEP’s of 10 (BRUNELLE, 1978) and 9.5 (LEWIS, 2000) respectively show no increase in the aluminum corrosion rate because of low surface area (visibly larger particle sizes) and also smaller zeta potentials inferred from the flocculation observed which occurs for particles with zeta potentials below 0.25mV (ASTM Standard D 4187-82, 1985).  6.3.3 Cathodic Slopes A measure of how easily hydrogen evolves from a surface is the hydrogen exchange current density hio for a given material. This hio can be calculated by extrapolating the linear Tafel region of a cathodic polarization curve back to the hydrogen potential for a given solution. If the alumina is serving as a proton donor and therefore making hydrogen evolution easier, then a Tafel slope should show a higher hydrogen exchange current density. Also as the hio increases the overpotential between the open circuit potential Eoc and the hydrogen potential Eh will decrease and for materials like platinum which are commonly used as electrodes, partially because of good hydrogen evolution characteristics, almost no overpotential is seen (SAWYER, 1995). In Figure 70 a cathodic polarization scan was plotted for a pellet of aluminum powder and a pellet of a 50/50 mixture of aluminum and alumina, both in a 0.045M KCl solution at 25oC which was purged with N2 gas for one hour while stirring to remove oxygen from the solution prior to the test. The Eoc for the sample with alumina showed a decrease in the overpotential of almost 500 mV. 126  -0.3 -0.5  Al 50% + Alumina 50%  Potential (V)  -0.7 -0.9  Al powder  -1.1 -1.3 -1.5 1.E-08  1.E-07  1.E-06  1.E-05  1.E-04  1.E-03  Apparent Current (A/cm 2)  Figure 70 Cathodic polarization scan for an aluminum powder pellet and an aluminum/alumina pellet in a 0.045M KCL solution at 25oC.  A fit to the linear Tafel region is shown in Figure 71. Both fits were very good with R2 values of .998 or better. In a pH 7 solution the Nernst equation gives a hydrogen potential Eh of -0.413 V, solving for the exchange current density at this value for the aluminum powder gives a hio of 4.6E-8 which is similar to the value reported by (CONWAY, 1957) for aluminum of 1E-8. The hydrogen exchange current density of the aluminum/alumina pellet gives a much higher value of 1.1E-5 as would be predicted.  127  -0.413  -0.613 Aluminum plus Alumina Potential (V)  -0.813  hio = 1.1E-5  -1.013  -1.213 Aluminum Powder -1.413  -1.613 1.E-07  hio = 4.6E-8  1.E-06  1.E-05  1.E-04  1.E-03  1.E-02  1.E-01  2  Apparent Current (A/cm )  Figure 71 Linear fit to Tafel region of cathodic slopes for calculation of hydrogen exchange current density (hio) for aluminum powder and aluminum/alumina mixture in 0.045M KCL solution at 25oC.  6.3.4 Effect on Hydrogen Evolution off Platinum If (as the model predicts) this Al corrosion increase is due to a more abundant supply of protons near the surface and is not dependant on a chemical interaction with the aluminum, then an increase in hydrogen evolution would be predicted from a chemically inert surface like platinum. Using a strip of platinum foil as the working electrode in a 0.1 M KCl solution that had been de-aerated with N2 bubbling, a hio was calculated for the aluminum foil with and without a teabag holding 2 g of loose alumina powder around the foil strip. The addition of alumina increased the hio at both 25oC (Figure 72) and 55oC (Figure 73) by over one order of magnitude. 128  -1.20  Platinum plus Alumina  Potential (V)  -1.25  hio = 5.7E-7  -1.30 Platinum -1.35  hio = 5.6E-8  -1.40  -1.45 0.0001  0.001  0.01  Current (A/cm 2)  Figure 72 Linear fit to Tafel region for calculation of hydrogen exchange current density (hio) for platinum foil and platinum surrounded by alumina powder in a 0.1M KCl solution at 25oC.  In all cases the hydrogen exchange current density is low compared to values reported in the literature, but it has been reported that to get the highest results the surface of the platinum should be carefully cleaned and absorbed species removed (KITA, 1966). Also affecting the magnitude of the exchange current density is that these tests were not performed in a strong acid which is normally used to provide a high concentration of protons to the surface of the working electrode (KITA, 1966). However, these tests saw identical conditions between the samples with and without alumina so the relative increase shows the alumina to be serving as a proton donor at neutral pH conditions. In highly acidic solutions with a surplus of protons available there might be little additional 129  effect from the alumina. The use of alumina to assist in hydrogen evolution at the cathode in electrolysis cells still needs to be investigated.  -0.6 -0.8 Platinum plus Alumina hio = 1.75E-6  Potential (V)  -1.0 -1.2 -1.4  Platinum hio = 4.8E-8  -1.6 -1.8 -2.0 0.0001  0.001  0.01  0.1  2  Current (A/cm )  Figure 73 Linear fit to Tafel region for calculation of hydrogen exchange current density (hio) for platinum foil and platinum surrounded by alumina powder in a 0.1M KCl solution at 55oC.  6.3.5 Electrical Contact with Aluminum The theory of the alumina acting as a proton donor relies on close contact to allow electrons given up by the aluminum to be taken by protons in the layer around the alumina particles. As such, it would be expected that aluminum powder separated from the oxide by a non-conductive material would not see the acceleration expected from the 130  alumina. To test this 0.5 g of aluminum powder was placed into a small cup made of plastic and also a cup made from aluminum foil, each were then placed in beakers on top of 5 g of alumina with the water free to circulate between the aluminum and alumina (Figure 74).  700  Al in Al cup on oxide Al in plastic cup on oxide  600  Volume of Hydrogen (ml)  Al mixed with oxide Al in Al cup  500  Al cup on oxide Al cup in water  400  Al in water 300 200 100 0 0  200  400  600  800  1000  1200  1400  1600  Time (minutes)  Figure 74 Volume of hydrogen released from 0.5 g of Aluminum powder with and without contact with 5 g Alumina powder in 70oC water.  The aluminum powder held in a plastic cap that insulated it electrically from the alumina had hydrogen evolution similar to that from plain aluminum powder in water. Also very similar was the evolution from aluminum in an aluminum cup with no oxide present. However, aluminum powder in the aluminum cup resting on top of the alumina showed greatly increased hydrogen evolution much closer to that seen when aluminum 131  and oxide are mixed. A large bubble of hydrogen was also seen on the underside of the aluminum cup indicating that hydrogen is evolving near the contact with the alumina and not just from the aluminum powder. If the aluminum cup is placed in water by itself, there is only a minor amount of hydrogen evolved before the cup quickly passivates and no further hydrogen is emitted. When the aluminum cup is placed in contact with the oxide, but without aluminum powder the initial hydrogen evolution is limited, with the overall aluminum surface having a smaller area than an equivalent mass of powder. However, contact with the oxide delays passivation and hydrogen continues to evolve. This is discussed further in the next section. The amount of hydrogen seen from the aluminum cup on oxide and the aluminum powder in the cup without the oxide together is not as much as when the aluminum powder is in the cup on oxide. This is a clear indication that the oxide is able to promote the powder’s corrosion with only an electrical connection through the aluminum foil.  6.4 Surface for Hydroxide Deposition Model The proton donor model discussed in the preceding section attempts to provide a mechanism for the increased corrosion rate seen with the addition of alumina and to explain the relative increase in evolved hydrogen with increasing alumina content. However, it fails to explain why the self-passivation typical of aluminum, which was earlier shown to limit the total corrosion of the powder to around 50%, is being prevented when alumina is present which allows for continued corrosion to take place over time. In the hydroxide deposition model, we postulate the following. There is some solubility, however low, of the hydroxide products of aluminum corrosion in solution at each pH. 132  The solubility limit instead of being static is the point where equilibrium is reached between hydroxide dissolving into solution and precipitating out of solution. Because in neutral pH solutions the solubility limit for aluminum hydroxide is low (ALWITT, 1976), aluminum normally builds up a hydroxide layer on its surface during corrosion which ultimately starts to impede the further corrosion of the aluminum. Because the alumina used in this work has a surface area per gram over 10 times that of the aluminum powders (see Table 2) and is typically used in ratios of 10 to 1 by mass, there is roughly 100 times the surface area of alumina particles as aluminum particles. Since all aluminum particles come with a native oxide layer on the surface, chemically all the particles in the solution are similar. The hydroxide dissolves into solution from the aluminum particles where it is formed; it can then re-precipitate on any available oxide surface. Because statistically most of the available surface is not on the aluminum particle itself, the aluminum sees a much slower buildup of hydroxide than normal which retards its passivation. With this deposition surface hypothesis we can make several predictions: 1. It was shown earlier that at higher temperatures the switch in reaction products actually reduces the overall corrosion seen in plain aluminum (Figure 40). If the reaction products are not building up on the aluminum then higher temperatures should not limit hydrogen evolution. 2. Alumina did not accelerate the corrosion of aluminum powder that was not in electrical contact, but if the solution is in contact with aluminum and alumina then the passivation should be reduced and greater corrosion seen over time.  133  6.4.1 Effect on Hydrogen Evolution Curves The addition of 0.5 g of Al2O3 to 0.5 g of Al powder is sufficient to delay the onset of passivation seen with pure aluminum and the rate over the initial hour increases with temperature (see Figure 75). When the aluminum/alumina mixtures are left at temperature for days until the corrosion reaction is complete (Figure 76), the mixtures all show greater overall corrosion than seen from pure aluminum and the total corrosion from the mixtures at 100oC and 85oC is greater than that at 60oC. 180 100 oC with 0.5g Al2O3  160  Volume of Hydrogen (ml)  140  85 oC with 0.5g Al2O3  120  60 oC with 0.5g Al2O3  95 oC  100 80  70 oC 70 oC  60 40 20 0 0  1  2  3  4  5  6  Time (hours)  Figure 75 Volume of hydrogen evolved with time from 0.5g of Aluminum powder by itself and mixed with 0.5g of Al2O3 at temperatures from 60oC to 100oC for initial reaction.  134  This is in contrast to pure aluminum which passivates at lower total levels of evolved hydrogen for solution temperatures above 70oC, believed to be the result of a change in corrosion product from Al(OH)3 to AlOOH.  600 500 Volume of Hydrogen (ml)  85 oC with 0.5g Al2O3  100 oC with 0.5g Al2O3  60 oC with 0.5g Al2O3  400  70 oC  300 200 95 oC  100 0 0  10  20  30  40  50  60  70  80  Time (hours)  Figure 76 Volume of hydrogen evolved with time from 0.5g of aluminum powder by itself and mixed with 0.5g of Al2O3 at temperatures from 60oC to 100oC for long term immersion.  In Figure 74, it was seen that aluminum powder in a plastic cup which prevented electrical contact with the alumina did not show the accelerated corrosion seen from aluminum/alumina mixtures, however since the solution was in contact with both powders the deposition hypothesis would predict greater corrosion with time. In Figure 77, the evolution from 0.5 g of Al powder is shown when the beaker contains 5 g of Al2O3 that is not in electrical contact with the aluminum. Despite 10 times the amount of 135  oxide the hydrogen evolution rate is below all of the other oxide samples and for the first hour is below that for only aluminum powder in water. This initial decrease seen for the first hour is believed to be caused by the thermal mass of the glass cylinder used to place the powder in the beaker without coming in contact with the oxide. This cylinder was at room temperature and would delay the aluminum from warming to reaction temperatures as quickly as seen when the powder is placed by itself in the beaker of hot water.  300 70 oC, 5g Al 2O3  Volume of Hydrogen (ml)  250  70 oC, 5g Al 2O3, no contact  200 100 oC, 0.5g Al 2O3  150  85 oC, 0.5g Al 2O3 60 oC, 0.5g Al 2O3  100  95 oC 70 oC  50  0 0  1  2  3  4  5  6  Time (hours)  Figure 77 Volume of hydrogen from 0.5g Al powder at 70oC in water with 5 g of Al2O3 mixed and held separate (no contact).  After the first hour, the evolution continues at an almost linear rate long after the evolution from aluminum alone fell off in rate showing the effects of passivation by the hydroxide. Over the course of two days the presence of the alumina allowed for greater 136  total corrosion to occur before being stopped by the passivation of the aluminum surface (see Figure 78).  600 5g Al 2O3  Volume of Hydrogen (ml)  500  85 oC with 0.5g Al 2O3  100 oC with 0.5g Al 2O3  60 oC with 0.5g Al 2O3  o  70 C, 5g Al 2O3, no contact  400  70 oC  300 200 95 oC  100 0 0  10  20  30  40  50  60  70  80  Time (hours)  Figure 78 Long term effect on hydrogen evolution when 5g Al2O3 is not in contact with 0.5g Al powder in 70oC water.  To try and detect the deposition of aluminum hydroxide on surrounding particles, aluminum powder was placed in water at 70oC with Si3N4 powder for several days until all hydrogen evolution had stopped. The solution was poured off and the powder was rinsed with flowing water to try and remove any loose aluminum hydroxide that was not adhering to particles. The powder was then analyzed with EDX, and no aluminum particles found. However, all of the Si3N4 particles showed a strong Al peak with the Al appearing over the entire particle surface and not in identifiable clumps or particles. The surface appearance of the Si3N4 was altered and these pictures are taken as further 137  evidence that in the presence of other deposition surfaces the aluminum hydroxide buildup on aluminum is reduced.  Figure 79 Si3N4 particle as received showing no Al.  Figure 80 Si3N4 particle after reaction for 3 days in 70oC water with Al.  138  6.4.2 Leaching Model A common situation in hydrometallurgy involves particles under surface reaction control or “Linear Leaching”, where the particles can be modeled as shrinking spheres which develop no product layer during leaching. If the alumina is working to keep the aluminum hydroxide from building up on the aluminum particles, then the aluminum corrosion may approximate linear leaching. The common test for linear leaching is to plot 1 – (Uƒ)1/3  Eq. 30  versus time. Here Uƒ is the volume fraction of unreacted powder equivalent to the ratio of hydrogen evolved to the total hydrogen possible from a given mass (677 ml from 0.5g Al) subtracted from one. A straight line signifies linear leaching. This test can be performed on aluminum fairly easily since, unlike many other particles the evolving hydrogen gives a measure of how much of the original aluminum has corroded. 0.25  1 – (U ƒ)1/3  0.20 0.15 0.10 0.05 0.00 0  20  40  60  80  100  120  140  Time (hours)  Figure 81 The shrinking core model predicts a linear fit of 1 – (Uƒ)1/3 with time for aluminum powder in 70oC water.  139  In Figure 81 it is seen that aluminum powder in 70oC water initially has a very good fit (R2>0.999, Figure 82) for the first 5 hours before the buildup of hydroxide causes the curve to depart from a linear fit. 0.050  0.045  1 – (U ƒ)1/3  0.040  0.035  0.030  0.025  0.020 0  1  2  3  4  5  6  Time (hours)  Figure 82 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder in 70oC water.  A similar curve for aluminum in 95oC water departs from linear much faster and only the first 15 minutes of reaction (Figure 83) show a linear fit which supports the premise that AlOOH is a more highly passivating reaction product. In all cases, the fit to the model departs from linear as the reaction progresses as was seen in Figure 81. What changes is the length of time it remains linear and the extent of reaction the aluminum undergoes in this time. For this reason only the linear portion is shown in the following graphs for 140  different Al/alumina mixtures. The fit of this model is also dependent on a tight size distribution for the particles so screening the samples for size uniformity would improve overall fit by eliminating early exhaustion from smaller particles. The length of time in which the aluminum remains in the linear region is plotted for washed aluminum (Figure 84), aluminum with 0.5 g Al2O3 at 100oC (Figure 85), aluminum with 5 g Al2O3 at 60oC (Figure 86), and aluminum with 5 g of Al2O3 not in electrical contact with the aluminum in 70oC water (Figure 87). The un-reacted fraction (UF) remaining when the plot no longer fits the linear region is summarized in Figure 88 and the ability of alumina to extend this region clearly shown. 0.034 0.032  1 – (U ƒ)1/3  0.030 0.028 0.026 0.024 0.022 0.020 0  0.05  0.1  0.15  0.2  0.25  0.3  Time (hours)  Figure 83 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder in 95oC water.  141  0.12 0.10  1 – (U ƒ)1/3  0.08 0.06 0.04 0.02 0.00 0  1  2  3  4  5  6  7  Time (minutes)  Figure 84 Showing the fit to a shrinking core model with time for 0.5g of washed aluminum powder in 70oC water.  0.18 0.16  1 – (U ƒ)  1/3  0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0  0.5  1  1.5  2  2.5  Time (hours)  Figure 85 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder hand mixed with 0.5g of Al2O3 in 100oC water.  142  0.35 0.30 1 – (U ƒ)1/3  0.25 0.20 0.15 0.10 0.05 0.00 0  0.5  1  1.5  2  2.5  3  Time (hours)  Figure 86 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder hand mixed with 5g of Al2O3 in 60oC water.  0.14 0.12 1 – (U ƒ)1/3  0.10 0.08 0.06 0.04 0.02 0.00 0  1  2  3  4  5  6  7  Time (hours)  Figure 87 Showing the fit to a shrinking core model with time for 0.5g of aluminum powder not in contact with 5g of Al2O3 in 70oC water.  To some extent the degree of linearity that still represents a good fit to the model is subjective, and the more lenient the fit the further the reaction proceeds while still under surface reaction control. As shown above, using a R2 value of better than 0.999 143  gives a very good fit to all points. If the curve is extended until the R2 drops to 0.99 there is visible separation of the later points from the linear fit. The amount of aluminum still left is graphed for both cases in Figure 88 and shows a similar trend for both cases. What is seen is that the alumina is allowing a greater percentage of the aluminum to corrode before passivation begins to hinder the reaction for both Al(OH)3 and AlOOH, and that close proximity of alumina also has a similar but reduced effect. 1  UF (R2>0.999)  0.9  UF (R2>0.99)  Unreacted Fraction (Uƒ)  0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Al 70 ºC  Al 95 ºC  washed Al 70 ºC  70 ºC + 5g oxide, no contact  100 ºC + 0.5g oxide  60 ºC + 5g oxide  Figure 88 The fraction of aluminum that has not yet reacted when a linear fit to the shrinking core model (1 – (Uƒ)1/3) falls below a R2 of 0.999 and R2 of 0.99.  The results from fitting the aluminum corrosion rates to the linear leaching model support the surface deposition hypothesis that reaction byproducts are being deposited on surrounding alumina particles and thus their buildup on the aluminum particles is delayed. 144  6.5 Summary Some of the early tests run while studying the effect of ball milling on aluminum powder showed an unexpected interaction between alumina particles and aluminum. Investigation revealed the unknown phenomenon that the presence of alumina greatly increased the corrosion rate of aluminum powders. This corrosion was complete enough to replace ball milling as a necessary preparation step for hydrogen generation from aluminum. The alumina is not consumed or rendered inert by the reaction and can serve as an accelerant for multiple batches of aluminum, with the reaction rate increasing for subsequent reactions. Almost complete reaction of aluminum is seen to occur within two hours. Two theories were put forth to explain this observation, the proton donor model and the surface deposition model. The theories are complementary rather than opposing and both theories were able to make predictions that were subsequently verified. The proton donor model supposes that the corrosion of aluminum is rate limited by the availability of hydrogen ions at the aluminum surface. The surface charge that alumina acquires in a neutral pH solution results in a layer of protons on the surface. This layer of protons acts as a source to nearby aluminum particles. An outcome of this is that particles with a negative surface charge at a given pH such as silicon nitride will act to suppress the aluminum corrosion rate. In both cases a high surface area is needed to produce a significant effect. The surface deposition model is stating that the high surface area of nearby alumina particles is providing an alternative surface for the precipitation of aluminum hydroxide from solution and this tends to transfer hydroxide from the aluminum surface  145  where it is created. This delays buildup of a passivating layer and allows the corrosion to proceed longer. The resulting effect can fit to the linear leaching model of surface reaction controlled shrinking spheres used in hydrometallurgy for many common reactions. Overall this discovery provides for a technologically feasible method of providing for portable hydrogen production from aluminum without the time and energy required by ball milling or the difficulties associated with highly alkaline solutions. The two theories presented and the tests that were performed can serve as a basis for future work into the corrosion effects of high surface area secondary particles, including their possible effects on water electrolysis for H2 generation.  146  Chapter 7 Conclusions 7.1 General Conclusions Aluminum is widely used as a building material because of the resistance to corrosion it exhibits in neutral pH environments due to self passivation. This resistance shows a marked decrease for small particles primarily because of a greater surface area to volume ratio. To be advantageous as a hydrogen source it is necessary to achieve corrosion of over 50% of available aluminum to be comparable to high pressure tanks on a per weight basis. This has been achieved for ball milled systems using a variety of added secondary particles which increase the surface area of exposed aluminum. The most effective from a weight perspective involve water soluble salts which can be removed from the aluminum after milling, although this step adds to processing complexity and cost. Solution temperatures greater than 50oC are used to initiate the corrosion reaction at a sufficient rate that the exothermic nature of the reaction can then maintain the solution temperature. The reaction can increase solution temperatures to the boiling point if the ratio of aluminum to water is high enough. Above 70oC the reaction product switches from Al(OH)3 to AlOOH which forms a more passive layer and retards the long term hydrogen evolution from aluminum powders compared with lower temperature reactions. This slow down at higher temperatures is not observed for ball milled systems where high surface area allows for extensive reaction to take place before the reaction layer can grow thick enough to stop the reaction. Before and during the evolution of 147  hydrogen the solution tends to alkalize, but the pH was not observed to increase above 9.5 in the tests run, and this pH alone did not greatly increase the observed reaction rates when looking at aluminum powder in NaOH solutions. When considering ball milled aluminum a significant finding of this work was that the change in surface area can account for all of the observed increase in reaction rate. Reaction rates for ball milled samples using both alumina and washed KCl were found to be similar once adjusted for final surface area and below that of un-milled powder. Future work on ball milled systems should look at high surface area secondary particles and improved methods for removing the secondary particle after milling to reduce the weight of powder that must be carried for portable hydrogen production. The effect of alumina powder on aluminum corrosion was discovered, and shown to greatly increase the corrosion rate and extent of corrosion seen for aluminum powder in contact with alumina. The corrosion rate can be comparable to ball milled systems with no pre-processing required and over 90% of the aluminum powder corroded before the reaction stops. Repeated additions of aluminum powder show an increasing reaction rate and provide a feasible mechanism for hydrogen production from aluminum. Two models, the Proton Donor and Surface Deposition, were proposed as possible explanations and tests run to validate the predictions of those models. In the Proton Donor Model the layer of protons which will form around alumina particles in neutral and acidic solutions is proposed to function as a proton source for aluminum particles in electrical contact with the alumina. This greatly increases the corrosion rate of the aluminum by facilitating the hydrogen evolution reaction. This model correctly predicted that low surface area particles would not have the effect on 148  corrosion rates seen with high surface area particles, and that particles with iso-electric points below pH=7 would have the opposite effect and tend to suppress the corrosion rate. The model explains the rate differences seen with conductive and insulating cups separating the aluminum powder from the alumina. This model also predicts an increase in hydrogen evolution off of platinum electrodes which was observed and could be useful in the development of more efficient cathodes for electrolysis cells. The Deposition Surface model argues that any given hydroxide solubility is maintained by dissolution and precipitation being balanced at the solubility limit, increasing the available surface area for hydroxide precipitation therefore tends to redistribute the hydroxide layer responsible for aluminum passivation to the inert alumina particles. This then delays the passivation of the aluminum surface and allows for a more complete reaction. It also explains the reaction of aluminum/alumina mixtures at higher temperatures that do not experience the greater passivation seen for aluminum powder at higher temperatures. Applying a linear leaching model to the corrosion rate data showed that, in addition to increasing the rate, the addition of alumina allowed a larger percentage of the particle to react before passivation began to limit the corrosion rate. Both of these mechanisms serve as possible explanations for the increased corrosion rate seen when alumina is added to aluminum powder, but are not limited to the aluminum/alumina system and may have applicability to the corrosion processes of other powders. This understanding may gain importance as nanotechnology pushes the use of high surface area small particles into new applications in the future.  149  7.2 Importance to Industrial Applications The rights to the original Chaklader patent were acquired by GHT (Global Hydrogen Technologies) and their subsidiary HPI (Hydrogen Power Inc) to develop a portable hydrogen source based on aluminum corrosion in water. HPI funded the research here at UBC and incorporated it into their Hydrogen Now© product. The understanding gained in this work can guide future optimization of milling parameters and choice of secondary particles. The current process being used by HPI is both time and energy intensive due to the milling requirements and scaling to larger volumes of powder for commercial scale production is non-trivial. Switching to the more weight efficient washed aluminum process developed at UBC by Edith Czech adds additional time and process steps. The current technology is based on the use of aluminum powder and has not been proven scalable to the larger machine shavings available on the scrap market. The discovery presented in this work that the aluminum powder will corrode rapidly in the presence of alumina powder and that the alumina powder can be repeatedly used without loss of activity represents a great improvement in terms of processing time and energy and therefore cost since the ball-milling and washing steps are not needed. The effect shown on aluminum foil demonstrates that the corrosion increase is not limited to small particles only and the delayed passivation may allow for processing of larger sized aluminum shavings and scrap without the need for high surface area requirements of the milled system. .  150  Chapter 8 Recommendations for Future Work  Interesting areas for future work include: •  The observation that hydrogen evolution curves from aluminum closely overlay one another when corrected for surface area raises the question of reversing this measurement. The use of a BET machine for determining the surface area of particles is time consuming, expensive, and not always available. It is possible that particles of interest could be ball milled with aluminum and then reacted in water with the resulting hydrogen evolution curve compared to a standard to estimate the surface area. More work is needed to determine what type of particles might be suitable for this technique and what the expected accuracy might be, but the possibility exists of developing a quick and dirty substitute for BET.  •  The cost of hydrogen produced from aluminum powder is quite high because of the high cost of aluminum. The ability to react and fully corrode larger shavings from machine shops or chopped scrap that is not currently recycled would greatly reduce both cost and the CO2 emissions associated with the process. Work to determine the size limitations of the alumina powder process is needed before it is commercialized.  •  The models proposed do not limit the reaction mechanism to alumina particles. Other high surface area particles with both high and low iso-electric points should be studied, both to optimize the system and better understand the reaction mechanism. 151  •  As the solution becomes more alkaline the efficiency of the surface deposition model would be higher due to higher solubility of the reaction by-products, but the proton donor effect should disappear as the IEP is crossed. More study is needed to separate the effects of these two models as pH changes.  •  Initial tests using Si3N4 particles showed the presence of aluminum after reaction as would be expected from the deposition model. More carefully designed experimental setups might give more detail on the extent of coverage and distance traveled before deposition.  •  The increase in hio seen on both aluminum and platinum from contact with alumina powder invites investigation on its possible use to increase efficiency at the cathode in electrolysis cells and the development of cheaper cathodes not using platinum.  •  Effect on corrosion of other metals beside aluminum from contact with alumina should be studied.  152  References AFSETH, A., NORDLIEN, J. H., SCAMMANS, G. M., et. al., 2001. Effect of heat treatment on filiform corrosion of aluminium alloy AA3005. Corrosion Science, 43, 2093-2109. AFSETH, A., NORDLIEN, J. H., SCAMMANS, G. M., et. al., 2002. Effect of thermomechanical processing on filiform corrosion of aluminium alloy AA3005. Corrosion Science, 44, 2491-2506. ALEKSANDROV YA, TSYGANOVA EI, PISAREVUAL., 2003, Reaction of aluminum with dilute aqueous NaOH solutions. Russ J Gen Chem; 73(5):689–94. ALWITT, R. S., 1976. The Aluminum – Water System. In: M. Dekker, ed. Oxides and Oxide Films. New York, J. W. Diggle, 171-254. ANDERSEN, E.R., ANDERSEN, E.J., 2003a, Method for producing hydrogen. US Patent 6506360, 2003 . ANDERSEN, E.R., ANDERSEN, E.J., 2003b, Apparatus for producing hydrogen. US Patent 2003/0118505 A1, 2003. ANDERSEN, E.R., 2004a, Renewable energy carrier system and method. US Patent 2004/0115125 A1, 2004. ANDERSEN, E.R., ANDERSEN, E.J., 2004b. Method for Producing Hydrogen. WO patent WO 2004/052775 A1. ARMSTRONG, R.D., BRAHAM, V.J., 1996, The Mechanism of Aluminum Corrosion in Alkaline Solutions, Corrosion Science, 38 (9) 1463-1471 ASSELIN, E. 2007. High Temperature and High Pressure Corrosion of Ni-Based Alloys and Stainless Steels in Ammoniacal Sulphate Solution. Ph.D. Thesis in Dept. of Materials Engineering, Univ. of British Columbia. ASTM Standard D 4187-82, 1985, Zeta Potential of Colloids in Water and Waste Water, American Society for Testing and Materials. 1985 AUTHIER-MARTIN, M., FORTE, G., OSTAP, S., et. al., 2001. The mineralogy of bauxite for producing smelter-grade alumina, JOM, 53(12), 36-40. BAUMGARTNER, M., KAUSCHE, H., 1990. Aluminum pitting in chloride solutions: morphology and pit growth kinetics. Corrosion Science, 31, 231-236. 153  BECK, T., 1984, Salt Film Formation during Corrosion of Aluminum, Electrochemica Acta, 29 (4), 485-491 BECKER, L., 2001. Hydrogen Storage [online]. Cambridge Scientific Abstracts. Available from: http://www.csa.com/hottopics/hydrogen/oview.html [Accessed 24 January 2005]. BELITSKUS, D., 1970, Reaction of Aluminum with Sodium Hydroxide Solution as a Source of Hydrogen, J. Electrochem. Soc., 117 [8] 1097-1099 BLACK, S.A., JENKINS, J.F., 1976, Sandwiched structure for production of heat and hydrogen gas. US Patent 3942511 BRINDLEY, G.F., 1909a, Composition of matter for generating hydrogen. US Patent 909536, 1909. BRINDLEY, G.F., BENNIE, M., 1909b, Composition of matter for manufacturing hydrogen gas. US Patent 934036, 1909. BRITTON, S.C., EVANS, U.R., 1930. Journal of the Chemical Society, 1930, p. 1773. BRUNELLE, J.P., 1978, Preparation of Catalysts by Metallic Complex Adsorption on Mineral Oxides. Pure and Applied Chemistry, vol 50, pp. 1211-1229 BOGDANOVIC, B., BRAND, R.A., MARFANOVIC, A., et. al., 2000. Metal-doped sodium aluminium hydrides as potential new hydrogen storage materials. Journal of Alloys and Compounds, 302, 36-58. BUNKER, B.C., NELSON, G.C., ZAVADIL, K.R., et. al., 2002. Hydration of Passive Oxide Films on Aluminum. J. Phys. Chem., 106, 4705-4713. CHADWICK, R., et al., 1958. 4 New Extraction Processes for Metals In: C. Singer, E.J. Holmyard, A.R. Hall, T.I. Williams, eds., A History of Technology Volume V - The Late Nineteenth Century, London: Oxford University Press, 72-101. CHAKLADER, ASOK C.D., 2002. Hydrogen generation from water split reaction. US Patent 6,440,385 B1. CHAO, C.Y., LIN, L.F., MACDONALD, D.D., 1981a, A Point Defect Model for Anodic Passive Films I: Film Growth Kinetics, J. Electrochem. Soc. 128 (6), 1187-1194 CHAO, C.Y., LIN, L.F., MACDONALD, D.D., 1981b, A Point Defect Model for Anodic Passive Films II: Chemical Breakdown and Pit Initiation, J. Electrochem. Soc. 128 (6), 1194-1198  154  CHECKETTS, J.H., 1998a, Hydrogen generation pelletizaed fuel. US Patent 5728464, 1998. CHECKETTS, J.H., 1998b, Hydrogen generation system and pelletized fuel. US Patent 5817157, 1998. CHOATE, W.T., GREEN, J.A.S., 2003. U.S. Energy Requirements for Aluminum Production: Historical Perspective, Theoretical Limits and New Opportunities, U.S. Department of Energy and Renewable Energy, Washington, D.C., U.S.A. CHOATE, W.T., AZIZ, A., FRIEDMAN, R., 2005. New Technologies Will Sustain the U.S. Primary Aluminum Industry. Light Metals 2005, Halvor Kvande Ed., TMS, p. 495500. CONWAY, B.E., BOCKRIS, J., 1957, Electrolytic Hydrogen Evolution Kinetics and its Relation to the Electonic and Adsorptive Properties of the Metal. Journal of Chemical Physics, vol. 26, number 3. CZECH, E., 2006. Hydrogen Generation from Aluminum-Water Systems. Masters Thesis. Dept of Materials Engineering, UBC. DAVIS, J.R., 1999. Corrosion of aluminum and aluminum alloys. Davis & Associates DENG, Z.Y., et. al., 2001A, High Surface Area Alumina Ceramics Fabricated by the Decomposition of Al(OH)3, J. Am. Ceram. Soc., 84 [3] 485-491 DENG, Z.Y., et. al., 2001B, Bulk Alumina Support with High Tolerant Strain and Its Reinforcing Mechanisms, Acta Mater., 49 [11] 1939-1946 DENG, Z.Y., et. al., 2005A, Modification of Al Particle Surfaces by γ-Al2O3 and its Effect on the Corrosion Behavior of Al, J. Am. Ceram. Soc., 88 [4] 977-979 DENG, Z.Y., et. al., 2005B, Temperature Effect on Hydrogen Generation by the Reaction of γ-Al2O3 – Modified Aluminum Powder with Distilled Water, J. Am. Ceram. Soc., 88 [10] 2975-2977 DENG, Z.Y., et. al., 2005C, Preparation of Hydrogen-Generation Materials by the Composite Processing Method, Japan Patent 2005-162552 DENG, Z.Y., et. al., 2007, Physicochemical Mechanism for the Continuous Reaction of γ-Al2O3 – Modified Aluminum Powder with Water, J. Am. Ceram. Soc., 90 [5] 15211526 DENG, Z.Y., FERREIRA, J., SAKKA, Y., 2008. Hydrogen-Generation Materials for Portable Applications, J. Am. Ceram. Soc., 91 [12] 3825-3834 155  DIAKOV V, DIVAN M, SHAFIROVICH E, et. al., 2007, Mechanistic studies of combustion-stimulated hydrogen generation from sodium borohydride. Chem Eng Sci; 62:5586–91. DIBARI, G. A., READ, H.J., 1971. Electrochemical Behavior of High Purity Aluminum in Chloride Containing Solutions. Corrosion, 27(11), 483-493. DICKEN, C., CLOUTIER, C.R., 2005. Sodium Borohydride for Hydrogen Storage in a fuel cell car. Paper in Dept. of Materials Engineering, Univ. of British Columbia. DIGNAM, M.J., 1962, Oxide Films on Aluminum: Ionic Conduction and Structure, J. Electrochem. Soc. 109 (3), 184-191 DOWNS, A.J., 1993, Chemistry of Aluminum, Gallium, Indium and Thalium, Blackie Academic & Professional, Glasgow. EMREGUL, K. C., AKSUT, A.A., 2000. The behavior of aluminum in alkaline media. Corrosion Science, 42, 2051-2067. EMSLEY, J., 2001, Nature’s Building Blocks, Oxford University Press, Oxford. 183-191 EVANS, J.W., 2007. The Evolution of Technology for Light Metals over the Last 50 Years: Al, Mg, and Li. JOM, February, p. 30-38. FAN MQ, XU F, SUN LX., 2007a, Studies on hydrogen generation characteristics of hydrolysis of the ball milling Al-based materials in pure water. Int J Hydrogen Energy ; 32:2809–15. FAN MQ, XU F, SUN LX., 2007b, Hydrogen Generation by Hydrolysis Reaction of Ball-Milled Al-Bi Alloys, Energy & Fuels, 21, 2294-2298 FAN, M.Q., et. al., 2008, Hydrolysis of ball milling Al-Bi-hydride and Al-Bi-salt mixture for hydrogen generation, Journal of Alloys and Compounds, 460, 125-129 FRANK, W.B., HAUPIN W.E., DAWLESS R.K., et al., 1997. 21 Aluminum In: F. Habashi, ed., Volume II - Handbook of Extractive Metallurgy, Toronto: Wiley-VCH, 1039-1127. GILL, G.C., 1955, Hydrogen generator. US Patent 2721789, 1955. GITLITZ, J. 2002. Trashed Cans the Global Environmental Impacts of Aluminum Can Wasting in America, Container Recycling Institute, Arlington, Virginia, U.S.  156  GREENWOOD, N., EARNSHAW, A. (1997), Chemistry of the Elements (2nd ed.), Oxford: Butterworth-Heinemann GRJOTHEIM, K., et al., 1977. Aluminium Electrolysis: The Chemistry of the HallHéroult Process. Aluminium-Verlag GmbH·Düsseldorf. GROSJEAN MH, et. al., 2004, Effect of ball milling on the corrosion resistance of magnesium in aqueous media. Electrochim Acta, 49, 2461-2470 GROSJEAN MH, ZIDOUNE M, ROUE´ L. 2005, Hydrogen production from highly corroding Mg-based materials elaborated by ball milling. J Alloys Compd ;712–5. GROSJEAN MH, ZIDOUNE M, ROUE´ L, et. al., 2006, Hydrogen production via hydrolysis reaction from ball-milled Mg-based materials. Int J Hydrogen Energy ; 31:109–19. GUILBERT, JOHN M., PARK, C. F., The Geology of Ore Deposits, Freeman, 1986, pp. 774-795 HABASHI, F., 1997. 1 The Metal Industry In: F. Habashi, ed. Volume I - Handbook of Extractive Metallurgy, Toronto: Wiley-VCH, 1-23. HART, R.K., 1957. The formation of films on aluminium immersed in water, Trans. Faraday Soc. 53, No. 7, 1020-7 HARUTA, M., 2004, Nanoparticulate Gold Catalysts for Low-Temperature CO Oxidation, Journal of New Materials for Electrochemical Systems, vol. 7, pp 163-172 HEINE, M.A., KEIR, D.S., PRYOR, M.J. 1965. The specific effects of chloride and sulfate ions on oxide covered aluminum, Journal of the Electrochemical Society, vol. 112, 1965, p. 24-32 HETHERINGTON, L.E., BROWN, T. J., BENHAM, A. J., et al. (2007). World Mineral Production: 2001-2005 [online] British Geological Survey. Available from: http://www.mineralsuk.com/britmin/wmp_2001_2005.pdf. [Accessed 10 January 2009] HEYDORN, B., 1994. By product hydrogen sources and markets. Proc. Natl. Hydrogen Assoc. Meet., 5th, Washington, DC. Washington, DC: Natl. Hydrogen Assoc. HIRAKI T, TAKEUCHI M, HISA M, et. al., 2005, Hydrogen production from waste aluminum at different temperatures with LCA. Mater Trans ; 46(5):1052–7. HIRAKI T, et. al., 2007, Process for recycling waste aluminum with generation of highpressure hydrogen. Environ Sci Technol, 41, 4454-4457 157  HU H, et. al., 2003, Kinetics of hydrogen evolution in alkali leaching of rapidly quenched Ni-Al alloy. Appl Catal A: Gen; 252:173–83. IIYIN, A.P., et al., 2000, Reactionary ability of aluminum ultrafine powders in various oxidation processes. The 4th Russian-Korean Int Symp Sci Technol KORUS Proc; 3:299– 304. INGENITO A, BRUNO C., 2004, Using aluminum for space propulsion. J Propulsion Power; 21(6):1056–63. IVANOV V.G., et al., 1994, Combustion of mixtures of ultradisperse aluminum and gellike water. Combust Explos Shock Waves; 30(4):569–70. IVANOV V.G., et al., 2000,. Specific features of the reaction between ultrafine aluminum and water in a combustion regime. Combust Explos Shock Waves; 36(2):213– 9. IVANOV VG, SAFRONOV MN, GAVRILYUK OV., 2001, Macrokinetics of oxidation of ultradisperse aluminum by water in the liquid phase. Combust Explos Shock Waves; 37(2):173–7. IZURU, K. et. al., 2003, Process for producing hydrogen. US Patent 2003/0173229 A1 JUNG, C.R., et. al., 2008, Hydrogen from Aluminum in a Flow Reactor for Fuel Cell Applications, J. Power Sources, 175, 490-494 JACOBSON, H.W., 1991. Coated Refractory Compositions and Method for Preparing the Same. WO Patent 9118841 (A2) KATOFSKY, R.E., 1993. Production of fluid fuels from biomass. Princeton Univ. Cent. Energy Environ. Stud. Rep. No. 279. Princeton, NJ: Princeton Univ. KELLY, R.G., et al., 2003, Electrochemical techniques in corrosion science and engineering. Marcel Dekker, New York KITA, H., 1966, Periodic Variation of Exchange Current Density of Hydrogen Electrode Reaction with Atomic Number and Reaction Mechanism. Journal of the Electrochemical Society. Vol 113, No. 11 KRAVCHENKO OV, SEMENENKO KN, BULYCHEV BM, et. al., 2005, Activation of aluminum metal and its reaction with water. J Alloys Compd ; 397:58–62. KORBEL, P., NOVAK, M., 1999. The Complete Encyclopedia of Minerals, Lisse, the Netherlands: Rebo International.  158  KUZNETSOVA, A., et. al., 1998. Electrochemical Evaluation of a New Type of Corrosion Passivation layer: Artificially Produced Al2O3 Films on Aluminum. Langmuir, 14, 2502-2507. LARMINIE, J., LOWRY, J., 2003. Electric Vehicle Technology Explained. John Wiley & Sons, Ltd. LAWSON, R.T., 1974, Aluminum Corrosion Studies. Potential-pH-Temperature Diagrams for Aluminum, Aust. J. Chem. 27, 105-127 LE GALL, M., et. al., 1995, Self-diffusion in a-Al2O3 and growth rate of alumina scales formed by oxidation: effect of Y2O3 doping, J. Material Science, 30(1), 201-11 LETH-OLSEN, H., AFSETH, A., AND NISANCIOGLU, K., 1998a, Filiform Corrosion of Aluminum Sheet II. Electrochemical and Corrosion Behaviour of Bare Substrates, Corrosion Science, 40 (7) 1195-1214 LETH-OLSEN, H., AFSETH, A., AND NISANCIOGLU, K., 1998b, Filiform Corrosion of Aluminum Sheet III. Microstructure of Reactive Surfaces, Corrosion Science, 40 (12) 2051-2063 LEWIS, J.A., 2000, Colloidal Processing of Ceramics. Journal of the American Ceramic Society. Vol. 83, no. 10, pp 2341-2359 LI, Q., BJERRUM, N.J., 2002, Aluminum as Anode for Energy Storage and Conversion: A Review, J. Power Sources, 110 [1] 1-10 LI, Z.P., MORIGANZAKI, N., LIU, B.H., et. al., 2003. Preparation of sodium borohydride by the reaction of MgH2 with dehydrated borax through ball milling at room temperature. Journal of Alloys and Compounds, 349, 232-236. MCCAFFERTY, E., 2003, Sequence of Steps in the Pitting of Aluminum by Chloride Ions, Corrosion Science, 45, 1421-1438 MACDONALD, D.D., URIQUIDI-MACDONALD, M., 1990. Theory of steady-state passive films. Journal of the Electrochemical Society 137, 2395-2402 MACDONALD, D.D., 1992. The point defect model for the passive state. Journal of the Electrochemical Society 139, 3434-3449. MACDONALD, D.D., 1999. Passivity – the key to our metals-based civilization. Pure Appl. Chem. 71, 951-978. MARTINEZ, S.S., et. al., 2005, Recycling of aluminum to produce green energy. Solar Energy Mater Solar Cells, 88, 237-243 159  MARTINEZ, S.S., et. al., 2007, Coupling a PEM fuel cell and the hydrogen generation from aluminum waste cans. Int J. Hydrogen Energy, 32, 3159-3162 MEARS, R.B., BROWN, DIX, E.H. 1945. Symposium on Stress-Corrosion Cracking of Metals, (1944), p 323, American Society for Testing Materials and American Institute of Mining and Metallurgical Engineers, Philadelphia and New York, 1945. OGDEN, J.M.,1999. Prospects for building a hydrogen energy infrastructure. Annu. Rev. Energy Environ., 24, 227-279. OVARI, F., TOMCSANYI, L., TURMEZEY, T., 1988. Electrochemical study of the pitting corrosion of aluminium and its alloys – I. Determination of critical pitting and protection potentials. Electrochimica Acta., 33(3), 323-326. PALMER, D (1997). Hydrogen in the Universe [online], Available from: http://imagine.gsfc.nasa.gov/docs/ask_astro/answers/97113i.html [Accessed 10 January 2009] PLUNKERT, P.A., 2004. Bauxite and Alumina In: U.S. Geological Survey, Mineral Commodity Summaries, U.S. Department of the Interior, VA, U.S.A. Available from: http://minerals.usgs.gov/minerals/pubs/mcs/2004/mcs2004.pdf (31-32). [Accessed 10 January 2009] PLUNKERT, P.A., 2007. Bauxite and Alumina In: U.S. Geological Survey, Mineral Commodity Summaries, U.S. Department of the Interior, VA, U.S.A. Available from: http://minerals.usgs.gov/minerals/pubs/commodity/bauxite/bauximcs07.pdf [Accessed 18 April 2007] POURBAIX, M., 1966. Atlas of Electrochemical Equilibria in Aqueous Solutions. London: Pergamon Press. PYUN, S., MOON, S., 2000. Corrosion mechanism of pure aluminium in aqueous alkaline solution. J Solid State Electrochem, 4, 267-272. RICHARDS, N.E., 1994. Strategies for Decreasing the Unit Energy and Environmental Impact of Hall Heroult Cells, U. Mannwiler ed. Light Metals Warrendale, PA, U.S., 393402. RICHARDSON BS, et al., 2005, Sodium borohydride based hybrid power system. J Power Sources; 145(1):21–9. SALAZAR, J.M.G, et. al., 1999, Corrosion behaviour of AA6061 and AA7005 reinforced with Al2O3 particles in aerated 3.5% chloride solutions: potentiodynamic measurements and microstructure evaluation. Corrosion Sciences, 41: 529-545. 160  SAWYER, D.T., SOBKOWIAK, A., ROBERTS, J.L., 1995, Electrochemistry for Chemists, 2nd edition, John Wiley and Sons, Inc. SERGEV, S.S., BLACK, S.A., JENKINS, J.F., 1980, Portable diver heat generating system. US Patent 4223661 SERGEV, S.S., BLACK, S.A., JENKINS, J.F., 1981, Supercorroding galvanic cell alloys for generation of heat and gas.. US Patent 4264362 SHAFIROVICH E, DIAKOV V, VARMA A., 2006, Combustion of novel chemical mixtures for hydrogen generation. Combust Flame; 144:415–8. SHAFIROVICH E, DIAKOV V, VARMA A., 2007, Combustion-assisted hydrolysis of sodium borohydride for hydrogen generation. Int J Hydrogen Energy; 32:207–11. SHIMIZU, K., et. al., 1991, The Apparent Induction Period for γ-Al2O3 Development in Thermal Oxide Films on Aluminum, Oxidation of Metals, 36 (1/2) SHIMIZU, K., et. al., 1999, A Novel Approach for the Study of the Migration of Cl Ions in Acidic Alumina, Corrosion Science, 41, 1835-1847 SOHLBERG, K., PANTELIDES, S.T., PENNYCOOK, S.J., 2001, Surface reconstruction and the difference in surface acidity between gamma- and eta-alumina, J. American Chemical Society, 123(1), 26-9 SOLER L, MACANAS J, MUN”OZ M, et. al., 2005, Hydrogen generation from aluminum in a non-consumable potassium hydroxide solution. Proceeding International Hydrogen Energy Congress and Exhibition IHEC 2005, 13-15 July 2005, Istanbul, Turkey SOLER L, MACANAS J, MUN”OZ M, et. al., 2007a, Aluminum and aluminum alloys as sources of hydrogen for fuel cell applications. J Power Sources; 169(1):144–9. SOLER L, MACANAS J, MUN”OZ M, et. al., 2007b, Synergistic Hydrogen Generation from Aluminum, Aluminum Alloys and Sodium Borohydride in Aqueous Solutions, Int. J. Hydrogen Energy, 32, 4702-4710 STOCKBURGER D, STANNARD J.H, RAO B.M.L, et. al., 1992, On-line hydrogen generation from aluminum in an alkaline solution. Proc Symp Hydrogen Storage Mater Batteries Electrochem; 92(5):431–44. SULLIVAN, J.P., et. al., 1999, The electrical properties of native deposited thin aluminum oxide layers on aluminum: hydration effects, Electrochemical Society – Proceedings (Critical Factors in Localized Corrosion III), 111-117  161  SZKLARSKA-SMIALOWSKA, Z., 1992, Insight into the Pitting Corrosion Behavior of Aluminum Alloys, Corrosion Science, 33 (8), 1193-1202 SZKLARSKA-SMIALOWSKA, Z., 1999, Pitting Corrosion of Aluminum, Corrosion Science, 41, 1743-1767 THOMAS, C.E., BRIAN, B., LOMAX, F., et. al., 2000. Fuel options for the fuel cell vehicle: hydrogen, methanol or gasoline?. International Journal of Hydrogen Energy, 25, 551-567. TOMCSANYI, L., VARGA, K., BARTIK, I., et. al., 1989. Electrochemical study of the pitting corrosion of aluminium and its alloys – II. Study of the interaction of chloride ions with a passive film on aluminium and initiation of pitting corrosion. Electrochimica Acta, 34(6), 855-859. TROCZYNSKI, TOM, 2005. Dept. of Materials Engineering, UBC. Personal communication. UEHARA, K., TAKESHITA, H., KOTAKA, H., 2002. Hydrogen gas generation in the wet cutting of aluminum and its alloys. Journal of Materials Processing Technology, 127, 174-177. UHLIG, H. H., 1948. The Corrosion Handbook. John Wiley & Sons, Inc. U.S. Department of Energy, 2003, U.S. Climate Change Technology Program - Research and Current Activities, (DOE/PI0001). Available from: http://www.climatetechnology.gov/library/2003/currentactivities/car24nov03.pdf [Accessed 25 November 2004] U. S. Dept. of Energy, Custom Alternative Fuel Comparison Chart [online.] Available from: http://www.eere.energy.gov/afdc/progs/fuel_compare.cgi [Accessed 24 January 2005]. VARGEL, C., 2004. Corrosion of Aluminum. Elsevier VAUGHAN, J. 2007. Ionic Liquid Electrochemical Processing of Reactive Metals. Ph.D. Thesis in Dept. of Materials Engineering, Univ. of British Columbia. VIDEM, K., 1977. Electrochemical utilization of metal hydrides. Hydrides for energy storage, 2, 463-471. WANG, H.Z., et al., 2008. A review on hydrogen production using aluminum and aluminum alloys, Renew Sustain Energy Rev (2008)  162  WERTH, J., 2000. Fuel cell using an aqueous hydrogen-generating process. U.S. Patent 6,093,501. WINTER, M (2007). Hydrogen: historical information [online], Available from: http://education.jlab.org/itselemental/ele001.html [Accessed 10 January 2009] WILLIAMS, R.H., 1998. Fuel decarbonization for fuel cell applications and sequestration of the separated CO2. In Eco-Restructuring: Implications for Sustainable Development, ed. W Ayres, pp. 180-222. Tokyo: UN Univ. Press. WOODALL, J., 2007, Process Generates Hydrogen from Aluminum Alloy, American Ceramic Society Bulletin, Vol. 86, No. 7 ZHOU, L., 2004. Progress and problems in hydrogen storage methods. Renewable and Sustainable Energy Reviews, 1-14. ZHUK AZ, SHEINDLIN AE, KLEYMENOV BV., 2006, Use of low-cost aluminum in electric energy production. J Power Sources; 157:921–6. ZUTTEL, A., 2003. Materials for hydrogen storage. Materialstoday, Sept., 24-33.  163  Appendices Appendix A - BET Plots 1.8 1.6  Volume (cc/g)  1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.00  0.05  0.10  0.15  0.20  Relative Pressure (P/P o )  Figure 89 Isotherm plot for BM aluminum 50%/ Alumina 50%.  164  0.25  0.30  0.35  Figure 90 BET Plot of Al2O3 with a surface area of 9.64 m2 g-1.  165  Figure 91 BET plot for Washed Al powder with a surface area of 13.76 m2 g-1.  166  Figure 92 BET plot for as received aluminum powder with a surface area of 0.61 m2 g-1.  167  Figure 93 BET plot of BM Al/Al2O3 50%/50% mixture with a surface area of 5.33 m2 g-1.  168  Figure 94 BET plot of BM Al/Al2O3 25%/75% mixture with a surface area of 7.85 m2 g-1.  169  Figure 95 BET plot of BM Al/Al2O3 75%/25% mixture with a surface area of 3.67 m2 g-1.  170  Figure 96 BET plot of BM washed Al/KCL mixture with a surface area of 13.01 m2 g-1.  171  Figure 97 BET plot of BM washed Al/KCl mixture with a surface area of 13.76 m2 g-1.  172  Figure 98 BET plot of Al2O3 powder with a surface area of 9.64 m2 g-1.  173  Figure 99 BET plot of SiC powder with a surface area of 0.73 m2 g-1.  174  Figure 100 BET plot of Si3N4 powder with a surface area of 8.60 m2 g-1.  175  Figure 101 BET plot of NaCl with a surface area of 1.39 m2 g-1.  176  Appendix B - Activation Energy  10000  1000 Current (uA)  70o C  o  55 C  85o C  o  40 C 100 o  25 C 10  10o C 1 0  10  20  30  40  50  60  70  80  Time (hours)  Figure 102 Potentiostatic test of an aluminum disk at -200mV in 0.01M Borax held for 12 hours at each temperature from 10oC to 85oC.  177  7  Rate [ln I] (uA)  6 5 4  y = -5563.8x + 22.506  3 2 1 0 0.002  0.0025  0.003  0.0035  0.004  1/Temperature (1/K)  Figure 103 Arrhenius plot for al disk held at +200mV in 0.01M borax solution re-run a second time.  7.8 7.6  Rate [ln I] (uA)  7.4 7.2 7.0 6.8 y = -8880.1x + 38.38 6.6 6.4 6.2 0.00346  0.00348  0.0035  0.00352 0.00354  0.00356 0.00358  0.0036  0.00362  1/Temperature (1/K)  Figure 104 Arrhenius plot for BM Al in 0.1M NaSO4 at 5oC and 15oC using separate pellets with current taken from Tafel fit at Eoc.  178  Appendix C – Repeat Graphs  350  Volume of Hydrogen (ml)  300 250 200 150 KCl1  100  KCl2 Alum ina1  50  Alum ina2  0 0  10  20  30  40  50  60  70  Time (minutes)  Figure 105 Hydrogen released from 0.5 g Al when BM with KCl or alumina and put in 70oC DI water.  179  350  Volume of Hydrogen (ml)  300 250 200 150 Alum ina2 Alum ina1  100  KCl1  50  KCl2  0 0  10  20  30  40  50  60  70  Time (minutes)  Figure 106 Hydrogen released from 0.5 g Al when BM with KCl or alumina and put in 70oC tap water.  180  0.6 50%CuO Run 1  0.4  50%CuO Run 2  0.2  Potential (V)  0.0 -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 1.0E-06  1.0E-05  1.0E-04  1.0E-03  1.0E-02  1.0E-01  1.0E+00  Apparent Current (A/cm 2)  Figure 107 Potentiodynamic curve of aluminum BM with CuO in 0.1M Na2SO4 solution.  181  0.0 Run 1  -0.1  Run 2  -0.2 -0.3 -0.4  Potential (V)  -0.5 -0.6 -0.7 -0.8 -0.9 -1.0 -1.1 -1.2 -1.3 -1.4 1.0E-07  1.0E-06  1.0E-05  1.0E-04  1.0E-03  1.0E-02  1.0E-01  1.0E+00  2  Apparent Current (A/cm )  Figure 108 2 Potentiodynamic scans for aluminum BM with Al2O3 in a 0.1M Na2SO4 solution at 25oC.  182  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0067321/manifest

Comment

Related Items