<%BANNER%>

Microstructural-Based Characterization of Hydrogenation and Dehydrogenation Mechanisms in Mg Powders

Permanent Link: http://ufdc.ufl.edu/UFE0042470/00001

Material Information

Title: Microstructural-Based Characterization of Hydrogenation and Dehydrogenation Mechanisms in Mg Powders
Physical Description: 1 online resource (188 p.)
Language: english
Creator: Tien, Hung-Yu
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: dehydrogenation, hydrogen, hydrogenation, kinetics, magneisum, magnesium, microstructural
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: MICROSTRUCURAL-BASED CHARACTERIZATION OF HYDROGENATION AND DEHYDROGENATION MECHANISMS IN MG POWDERS Metal hydrides display moderately high gravimetric and volumetric capacities and are being considered as potential solid-state hydrogen storage materials for on-board fuel-cell powered vehicles. Among them, magnesium hydride has been widely studied in the past few decades owing to its abundance, low density, reversibility, and reasonable ideal capacity of 7.6wt%. However, the hydrogen capacity of magnesium is usually below its theoretical value and varies with hydrogenation condition and powder size and shape. One of the main objectives of this dissertation was to develop a mechanistic understanding of the hydrogenation process in magnesium powders based on detailed microstructural analyses in order to shed light on the parameters that control the hydrogen capacity. We have demonstrated that the hydrogenation of magnesium powders at 1Mpa pressure and in the temperature range of 150 to 300degreeC takes place in three stages. Stage 1 is the nucleation of the magnesium hydride, Stage 2 is the growth of magnesium hydride colonies up to their impingement, and Stage 3 is the further growth of the continuous hydride layer. The nucleation of the hydride was found to be limited to the powder surface, except at 150degreeC, where internal hydride formation was observed. Quantitative metallography suggests that at high temperatures the magnesium hydride nucleates at the intersection of magnesium grain boundaries with the powder surface and consequently the number of hydride nuclei remains constant during Stage 2. The activation energy associated with Stage 2 was measured to be 32.4 kJ/ mole, which is close to the diffusion of hydrogen atom through the hcp-Mg metal. Stage 3, is believed to take place by the diffusion of hydrogen atoms through the hydride layer, which is quite slow and leads to the observation of a pseudo-saturation behavior that governs the hydrogen capacity. Through different approaches it is elucidated that the reduced hydrogen capacity at low temperatures is associated with an increase in the hydride nucleation rate, which results in impingement of the hydride colonies at smaller hydride thickness values. The findings of this investigation also reveal that the powder size distribution, the average powder size and particle shape affect hydrogen capacity significantly. A broad particle size distribution was found to result in a gradual transition between Stage 2 and 3 and a less pronounced pseudo-saturation behavior. It is demonstrated that decreasing the size of the particles in the powder and increasing their aspect ratio drastically increase the hydrogen capacity. These phenomena are explained based on the nucleation and growth of the magnesium hydride. Studies near the equilibrium conditions revealed that the nucleation of the hcp-Mg influences the kinetics of dehydrogenation process noticeably. The presence of metallic magnesium phase on the surface, owing to incomplete hydrogenation, was found to enhance the kinetics and reduce the dehydrogenation temperature by eliminating the energy barrier required for the magnesium nucleation during hydrogen release process. The presence of metallic magnesium phase on the surface eliminates the energy barrier for the magnesium nucleation during hydrogen release process. This effect results in the lowering of the dehydrogenation temperature of magnesium hydride.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Hung-Yu Tien.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Ebrahimi, Fereshteh.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-06-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042470:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042470/00001

Material Information

Title: Microstructural-Based Characterization of Hydrogenation and Dehydrogenation Mechanisms in Mg Powders
Physical Description: 1 online resource (188 p.)
Language: english
Creator: Tien, Hung-Yu
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: dehydrogenation, hydrogen, hydrogenation, kinetics, magneisum, magnesium, microstructural
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: MICROSTRUCURAL-BASED CHARACTERIZATION OF HYDROGENATION AND DEHYDROGENATION MECHANISMS IN MG POWDERS Metal hydrides display moderately high gravimetric and volumetric capacities and are being considered as potential solid-state hydrogen storage materials for on-board fuel-cell powered vehicles. Among them, magnesium hydride has been widely studied in the past few decades owing to its abundance, low density, reversibility, and reasonable ideal capacity of 7.6wt%. However, the hydrogen capacity of magnesium is usually below its theoretical value and varies with hydrogenation condition and powder size and shape. One of the main objectives of this dissertation was to develop a mechanistic understanding of the hydrogenation process in magnesium powders based on detailed microstructural analyses in order to shed light on the parameters that control the hydrogen capacity. We have demonstrated that the hydrogenation of magnesium powders at 1Mpa pressure and in the temperature range of 150 to 300degreeC takes place in three stages. Stage 1 is the nucleation of the magnesium hydride, Stage 2 is the growth of magnesium hydride colonies up to their impingement, and Stage 3 is the further growth of the continuous hydride layer. The nucleation of the hydride was found to be limited to the powder surface, except at 150degreeC, where internal hydride formation was observed. Quantitative metallography suggests that at high temperatures the magnesium hydride nucleates at the intersection of magnesium grain boundaries with the powder surface and consequently the number of hydride nuclei remains constant during Stage 2. The activation energy associated with Stage 2 was measured to be 32.4 kJ/ mole, which is close to the diffusion of hydrogen atom through the hcp-Mg metal. Stage 3, is believed to take place by the diffusion of hydrogen atoms through the hydride layer, which is quite slow and leads to the observation of a pseudo-saturation behavior that governs the hydrogen capacity. Through different approaches it is elucidated that the reduced hydrogen capacity at low temperatures is associated with an increase in the hydride nucleation rate, which results in impingement of the hydride colonies at smaller hydride thickness values. The findings of this investigation also reveal that the powder size distribution, the average powder size and particle shape affect hydrogen capacity significantly. A broad particle size distribution was found to result in a gradual transition between Stage 2 and 3 and a less pronounced pseudo-saturation behavior. It is demonstrated that decreasing the size of the particles in the powder and increasing their aspect ratio drastically increase the hydrogen capacity. These phenomena are explained based on the nucleation and growth of the magnesium hydride. Studies near the equilibrium conditions revealed that the nucleation of the hcp-Mg influences the kinetics of dehydrogenation process noticeably. The presence of metallic magnesium phase on the surface, owing to incomplete hydrogenation, was found to enhance the kinetics and reduce the dehydrogenation temperature by eliminating the energy barrier required for the magnesium nucleation during hydrogen release process. The presence of metallic magnesium phase on the surface eliminates the energy barrier for the magnesium nucleation during hydrogen release process. This effect results in the lowering of the dehydrogenation temperature of magnesium hydride.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Hung-Yu Tien.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Ebrahimi, Fereshteh.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-06-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042470:00001


This item has the following downloads:


Full Text

PAGE 1

1 MICROSTRUCURAL BASED CH A RACTERIZATION OF HYDROGE NATION AND DEHYDROGENATION MECHA NISM S IN MG POWDERS By HUNG YU TIEN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Hung Yu Tien

PAGE 3

3 To my grandparents, dad, mom, sister and friends

PAGE 4

4 AKNOWLEDGEMENT S In my 3 years PhD study, my advisor Dr. Ebrahimi is the most important person who help s me a lot not only in my research also in life The attitude of learning is the major subject in her teaching. Her encourage and inspiration are the motivation for me to complete the degree. Without her painstaking teaching I would not reach this level. I would like to thank her help from my bottom of heart. I would like to thank all of my committee members, Dr. Craciun, Dr. Jones, Dr. Manuel, and Dr. Lear for their pre cious advice s toward my research. I also would like to thank Dr. Bourne for the valuable question in my qualify exam and the permission of using the polishing wheel in his lab. I am appreciated that Dr. Glicksman shares everything he knows to explain to me how to analyze my hydrogenation curves Special thanks to Dr. Wu for proving the Theta Composer for the essential coating process. I am grateful to Dr. Dempere and Dr. Wayne for their suggestion in characterizing my samples in using EBSD and SEM in MAIC I would like to thank my group member s Dr. Sankara Tartiparti, Dr. Yanli Wang, Dr. Orlando Rios, Dr. Mahesh Tanniru, Ian Liu, Sonalika Goyal Mike Kesler Glenn Bean Jr. Jon Besser Ross, David Rule, Stev en James Diego, Christian for their sincere ideas on my research and presentation To all m y friends who help me whenever I face any difficulty, I would like to thank their concern s and accompa nies This project is funded by NSF under the contract number: DMR 0605406.

PAGE 5

5 TABLE OF CONTENTS page AKNOWLEDGEMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 17 1 INTRODUCTION ................................ ................................ ................................ .... 20 2 BACKGROUND ................................ ................................ ................................ ...... 27 2 .1 The Structure of Magnesium and Magnesium Hydride ................................ ..... 27 2.2 Thermodynamic Properties of Magnesium Hydride ................................ .......... 2 7 2.2.1 Pressure Compo sition Temperature (PCT) of Magnesium ..................... 29 2.2.2 Alloying Effect ................................ ................................ .......................... 30 2.3 Kinetic Properties of Magnesium Hydride ................................ ......................... 31 2.3.1 Hydrogen Capacity of Magnesium Powders ................................ ............ 31 2.3.2 Kinetics Model of Hydrogenation of Magnesium Powders ....................... 32 2.3.3 Effect of Catalyst ................................ ................................ ..................... 36 2.4 Microstructural Analysis of Hydrogenated Mg Powders ................................ .... 38 2.5 Effect of Shape and Size on Hydrogen Capacity of Metal Hydride ................... 39 3 EXPERIMENTAL PROCEDURES ................................ ................................ .......... 46 3.1 Material Fabrication ................................ ................................ .......................... 47 3.1.1 Addition of Catalyst by Theta Composer ................................ ................. 47 3.1.2 Addition of Catalyst by Organometallic Nickel Solution ........................... 48 3.1.3 High Speed Orbital Ball Milling (HSOBM) ................................ ................ 49 3.2 Hydrogen Absorption ................................ ................................ ........................ 49 3.2.1 Hydrogena tion Setup ................................ ................................ ............... 49 3.2.2 Hydrogenation Procedure ................................ ................................ ........ 51 3.3 Hydrogen Desorption ................................ ................................ ........................ 54 3.3.1 Dehydrogenation Setup ................................ ................................ ........... 54 3.3.2 Dehydrogenation Procedure ................................ ................................ .... 54 3.3.3 Thermal Gravimetric Analyzer (TGA) ................................ ...................... 55 3.4 Analysis Methods and Characteristic Techniques ................................ ............. 57 3.4.1 Particle Size Distribution ................................ ................................ .......... 57 3.4.2 Scanning Electronic Microscope (SEM) ................................ .................. 57 3.4.3 Optical Microscope (OM) ................................ ................................ ......... 58 4 CHARACTERIZATION OF POWDERS ................................ ................................ .. 64 4.1 Characteristics of the As Received Mg powders ................................ .............. 64

PAGE 6

6 ................................ ................................ .................... 64 ................................ ................ 65 4.2 Characteristics of Mg Flakes ................................ ................................ ............. 66 4.3 Characteristics of Ni coa ted Mg Powders ................................ ......................... 67 4.3.1 Ni Coating Using Theta Composer ................................ .......................... 67 4.3.2 Ni Coating Using an Organometallic Solution ................................ .......... 67 5 EVOLUTION OF MICROSTRUCTURE DURING THE HYDROGENATION PROCESS ................................ ................................ ................................ .............. 78 5.1 Introductory Remarks ................................ ................................ ........................ 78 5.2 Microstructural Analysis ................................ ................................ .................... 79 5.3 Effect of Particle Size on Hydrogen Capacity ................................ ................... 84 5.4 Summary of Hydrogenation P rocess ................................ ................................ 86 5.5 Summary and Conclusions ................................ ................................ ............... 87 6 KINETICS OF MG HYDRIDE FORMATION ................................ ......................... 104 6.1 Introductory Remarks ................................ ................................ ...................... 104 6.2 Analysis of Kinetics of Hydride Formation by the Johnson Mehl Avrami Relationship ................................ ................................ ................................ ....... 104 6.2.1 Evaluation of Nucleation Number during Stage 2 ................................ .. 106 6.2.2 Correlation of Grain Boundary and Nuclei of Magnesium Hydride ........ 108 6.2.3 Effect of Hydrogenation Temperature ................................ .................... 109 6.2.4 Microstructural Analysis of Powders Hydrogenated at Different Temperatures ................................ ................................ .............................. 111 6.3 Effect of Nucleation Rate on Hydrogen Capacity ................................ ............ 112 6.3.1 Hydrogenation Procedures ................................ ................................ .... 112 6.3.2 Heating Procedure ................................ ................................ ................. 117 6.4 Kinetics analysis by Johnson Mehl Avrami relation for Growth Mechanism of Magnesium Hydride ................................ ................................ ....................... 120 6.5 The Fraction of Magnesi um Hydride at Impingement ................................ ..... 123 6.6 Summary and Conclusions ................................ ................................ ............. 125 7 EFFECT OF POWDER GEOMETRY and size ON THE HYDROGEN CAPACITY OF M AGNESIUM ................................ ................................ ............... 151 7.1 Introductory Remarks ................................ ................................ ...................... 151 7.2 Effect of Particle Shape on Hydrogenation Behavior of Mg Powder ............... 152 7.3 Effect of Particle Size on Hydrogen Capacity ................................ ................. 156 7.4 Mathematical Modeling ................................ ................................ ................... 157 7.5 Summary and Conclusions ................................ ................................ ............. 160 8 DEHYDROGENATION MECHANISM OF MAGNESIUM HYDRIDE .................... 168 8.1 Introductory Remarks ................................ ................................ ...................... 168 8.2 Desorption Study by Constant Heating Rate in TGA ................................ ...... 169 8.3 Isothermal Dehydrogenation Study of MgH 2 ................................ ................... 172

PAGE 7

7 8.4 Summary and Conclusions ................................ ................................ ............. 175 CONCLUSIONS AND FUTURE WORK ................................ ................................ ...... 180 LIST OF REFERENCES ................................ ................................ ............................. 182 9 BIOGRAPHIAL SKETCH ................................ ................................ ...................... 188

PAGE 8

8 LIST OF TABLES Table page 2 1 Dependence of n in the JMA equation ................................ ................................ 41 6 1 The result of the measurement of number of nucleation at different hydrogenation times ................................ ................................ ......................... 127 6 2 Hydrogen capacity at differ ent hydrogenation temperatures ............................ 127 6 3 The slope of the three stages of hydrogenation as revealed by JMA analysis in Figure 6 29. ................................ ................................ ................................ .. 127 6 4 n anf K values for Stage 2 and 3 at each temperatures ................................ .... 128 6 5 End time in minutes of each stage ................................ ................................ .... 128

PAGE 9

9 LIST OF FIGURES Figure page 1 1 The schematic hydrogenation curve showing the hydrogen capacity and the hydrogenation kinetics. ................................ ................................ ....................... 26 2 1 The crystal structure of hexag onal closed packed (HCP) Mg. ............................ 41 2 2 The crystal structure of tetragonal magnesium hydride. ................................ ..... 41 2 3 Schematic illustration of the likely epitaxial growth mode of MgH 2 on Mg on (001) Al 2 O 3 in the MgH 2 (110)[001]//Mg(001)[100] relationship. For clearness, only the Mg atoms in MgH 2 are shown. [42] ................................ ....................... 42 2 4 Mg H binary phase d iagram at 1bar (0.1MPa) hydrogen pressure. .................... 42 2 5 Energy barriers for hydrogen absorption and desorption. ................................ ... 43 2 6 PCT curves and ................ 43 2 7 The schematic growth of a product region. ................................ ......................... 43 2 8 Cross section of a m agnesium particle at two different times a curves for hydriding at A) P~P eq and B) P>>P eq ................................ ................ 44 2 9 A schematic diagram for Core Shell hydrogen absorption model. ...................... 44 2 10 Potential energy curve for hydrogen binding to a metal surface which shows: (i) physisorption for both activated and non activated processes; (ii) dissociation of hydrogen molecules and chemisorption ; (iii) surface penetration and chemisorption; and (iv) diffusion of hydrogen atoms. [73] ......... 44 2 11 Microstructural analysis of hydrogenated magnesium powders after (a) first hydrogenation, (b) a series of long term cycling test, (c) a few short term cycling tests, and (d) a series of short term cycling test. ................................ .... 45 3 1 Glove box (Vacuum Atmosphere, m odel: Nexus ) ................................ ............... 59 3 2 Monitoring the O 2 and moisture level ................................ ................................ .. 59 3 3 Theta Composer Ni coating ................................ ................................ ................ 59 3 4 Orga nometallic Ni coating. Black dot represent nano Ni particles and white one is the magnesium powder. 110 C is the operating temperature. The arrows illustrate the cycling of toluene in the beaker. ................................ ......... 60 3 5 HSOBM milling ................................ ................................ ................................ ... 60

PAGE 10

10 3 6 Schematic illustration of HSOBM milling ................................ ............................ 61 3 7 The relationship between applying voltage and mil ling speed. ........................... 61 3 8 Mounting sample cut perpendicularly to the flake surface for thickness measurement. ................................ ................................ ................................ ..... 61 3 9 H/D system actual view ................................ ................................ ...................... 62 3 10 Schematic diagram of Hydrogenation/Dehydrogenation unit. ............................. 62 3 11 The illustration of the calculation of pressure dr ops in Method I. ........................ 63 4 1 The particle size distribution of the as evaluated by laser diffraction technique ................................ ............................. 69 4 2 The SEM image of the as ......................... 69 4 3 Cross sectional optical image of the powders in the as magnesium powder. ................................ ................................ ........................... 70 4 4 The schematic illustration of polishing surfaces and calculation of aspect ratio for as ................................ ....................... 70 4 5 Particle size di stributions of the as statistic measurement in (a) number percent and (b) volume percent calculation. ................................ ................................ ................................ .......... 71 4 6 SEM images of the as der. ............................. 71 4 7 Cross sectional optical image of the as .. 72 4 8 Particle size distribution in the as on (a) number percent and (b) volume percent calculation. ............................... 72 4 9 Magnesium flake thickness distribution after 100V and 20 seconds milling. ...... 73 4 10 Magnesium flake thickness distribution after 100V and 1 minute milling. ........... 73 4 11 Magnesium flake thickness distribution after 100V and 2 minutes milling. ......... 73 4 12 Magnesium flake thickness distribution after 110V and 2 minutes milling. ......... 74 4 13 Optical microscopic analysis at cross section of magnesium flake powders. ..... 74 4 14 SEM images of as milled magnesium flake powders. ................................ ......... 74 4 15 Particle size distribution of Ni coated 44 micron magnesium particles. .............. 75

PAGE 11

11 4 16 SEM images of Ni coated the 44 micron magnesium powder after Ni coating with the Theta Composer. ................................ ................................ .................. 75 4 17 The EDS Ni mapping of Ni coated 44 micron magnesium particles. .................. 75 4 18 The X ray diffraction patterns. ................................ ................................ ............ 7 6 4 19 SEM images of Ni coated magnesium flake particles. ................................ ........ 76 4 20 The EDS Ni mapping of Ni coated magnesium flake particles. .......................... 76 4 21 SEM image of Ni coated 300 micron magnesium particles ................................ 77 5 1 Hydrogenation curve of magnesium powder sample at 210 C under 1MPa hydrogen pressure. ................................ ................................ ............................. 89 5 2 Schematic illustration of Core Shell model. ................................ ........................ 89 5 3 Hydrogenation curves of magnesium powder sample at 210C under 1MPa hydrogen pressure a t each interrupted times. ................................ .................... 90 5 4 Microstructural analysis on the cross section of magnesium powder hydrogenated for 2.5 minutes. ................................ ................................ ............ 91 5 5 Microstructural analysis on the cross section of magnesium powder hydrogenated for 7.5 minutes. ................................ ................................ ............ 92 5 6 Optical microscopic images for the investigation of the location of nucleation of magnesium hydride. ................................ ................................ ....................... 93 5 7 Microstructural analysis on the cross section of powder hydrogenated for 15 minutes. ................................ ................................ ................................ .............. 94 5 8 Microstru ctural analysis on the cross section of powder hydrogenated for 75 minutes. ................................ ................................ ................................ .............. 95 5 9 Microstructural analysis on the cross section of powder hydrogenated for 300 minutes. ................................ ................................ ................................ .............. 96 5 10 An example of different hydride fractions achieved by different sizes of particles. ................................ ................................ ................................ ............. 97 5 11 An example of the measurement of phase fraction of magnesium hydride ........ 98 5 12 Hydride fraction at different range of sizes and temperatures. ........................... 98 5 13 An illustration of hydride dist ribution and phase fraction for different size of particles at the same hydrogenation condition and time 15 minutes. ................. 99

PAGE 12

12 5 14 A comparison of hydrogenation curve with the data point estimated b y the measurement of area phase fraction. ................................ ................................ 99 5 15 Step 1: Physicsorption of hydrogen molecules. ................................ ................ 100 5 16 Step 2: The dissociatio n of hydrogen molecules. ................................ ............. 100 5 17 Step 3:The dissolving of hydrogen atoms. ................................ ........................ 101 5 18 Step 4: Chmisorption of hydrogen ions. ................................ ............................ 101 5 19 Step 5: Nucleation of magnesium hydrides. ................................ ..................... 102 5 20 Step 6: Growth of magnesium hydrides. ................................ ........................... 102 5 21 Step 7: The impingement of magnesium hydrides. ................................ ........... 103 6 1 The JMA analysis of powders hydrogenated at 210C under 1MPa. ............... 129 6 2 The measurement of perimeter for powders hydrogenated for 2.5 minutes. .... 129 6 3 The measurement of perimeter for powders hydrogenated for 7.5 minutes. .... 130 6 4 The etched cross sectional surface of magnesium powder pre heated at 210C. ................................ ................................ ................................ .............. 130 6 5 The hydrogenation curves of magnesium powde rs hydrogenated at 150, 180, 210, and 250C. ................................ ................................ ................................ 131 6 6 The relationship between hydrogen capacity and hydrogenation temperatures. ................................ ................................ ................................ ... 131 6 7 The microstructural analysis of powders hydrogenated at (a)150C, (b)180C, (c)210C, and (d)250C under 1Mpa for 5 hours. ............................ 132 6 8 The schematic illustration of proposed model for the effect of nucleation rate on hydrogenation kinetics and hydrogen capacity. ................................ ........... 133 6 9 Hydrogen absorption as a function of time for the two hydrogenation methods employed. ................................ ................................ ................................ ......... 133 6 10 The evaluation of kinetics of hydride formation by Methods I and II on the JMA relationship. ................................ ................................ .............................. 134 6 11 A comparison of the hydrogenation curve s for powders hydrogenated by Method I for 75 minutes and by Method II for 3 minutes. ................................ .. 134 6 12 SEM pictures comparing the hydride formation for powders hydrogenated by Method I for 75 minutes (a and b) and by Method II for 3 minutes (c and d);

PAGE 13

13 (a) and (C) are cross sectional view and (b) and (d) are the external view of the powders. ................................ ................................ ................................ ..... 135 6 13 The surface coverage measurement. ................................ ............................... 135 6 14 The distribution of the surface coverage of the particles hydrogenated by Method I and II. ................................ ................................ ................................ 136 6 15 SEM pictures showing cros s sectional views of the particles with high hydride coverage after 75 minutes f hydrogenation by Method I (a and b) and with low hydride coverage after 3 minutes hydrogenation by Method II (c and d). ... 136 6 16 Hydrogenation curves for 150C, 210C and the two step heating. Point A denotes the time (15 minutes) at 210C, which results in an equal amount of hydride as the powder hydrogenated at 150C for 5 hours. ............................. 137 6 17 SEM (BSE mode) images showing the hydride distribution in the powder hydrogenated at (a and c) 210C for 15 minutes, and (b and d) at 150C for 300 minutes. (a) and (b) are cross sectional images, while (c) and (d) a re external surface images. magnesium hydride appears dark because of its lower density. ................................ ................................ ................................ .... 138 6 18 SEM (BSE mode) cross sectional images showing (a) hydride distribution in powders hydrogenated at 150C for 2 minutes, (b) higher magnification of (a) showing internal nucleation of MgH 2 (c) hydride formation in powders hydrogenated by the two step heating (150C for 2 minutes and 210C for 298 minutes), and (d) hydride formation in the powders hy drogenated at 150C for 300 minutes. ................................ ................................ ..................... 139 6 19 The JMA analysis of powder hydrogenated at 250C (Sample 1). ................... 140 6 20 The JMA anal ysis of powder hydrogenated at 250C (Sample 2). ................... 140 6 21 The JMA analysis of powder hydrogenated at 210C (Sample 1). ................... 141 6 22 The JMA analysis of powder hydrogenated at 210C (Sample 2). ................... 141 6 23 The JMA analysis of powder hydrogenated at 180C (Sample 1). ................... 142 6 24 The JMA analysis of powder hydrogenated at 180C (Sample 2). ................... 142 6 25 The JMA analysis of powder hydrogenated at 150C (Sample 1). ................... 143 6 26 The JMA analysis of powder hydrogenated at 150C (Sample 2). ................... 143 6 27 The relationship between n values and temperatures at Stage 2 and 3. .......... 144 6 28 The location of transition region within the hydrogenation curves. ................... 144

PAGE 14

14 6 29 The activation energy calculated by Arrhenius plot. ................................ ......... 145 6 30 The schematic illustration of the comparison of slopes at the beginning of Stage 3 at 150C and 250C. ................................ ................................ ........... 145 6 31 The relationship between hy dride fraction and reaction rate at 250C. ............ 146 6 32 The f* calculation at 250C ................................ ................................ .............. 146 6 33 The relationship between hydride fract ion and reaction rate at 210C. ............ 147 6 34 The f* calculation at 210C. ................................ ................................ .............. 147 6 35 The relationship between hydride fraction and rea ction rate at 180C. ............ 148 6 36 The f* calculation at 180C. ................................ ................................ .............. 148 6 37 The relationship between hydride fraction and reaction rate at 150C. ............ 149 6 38 The f* calculation at 150C. ................................ ................................ .............. 149 6 39 A comparison between hydride fraction at impingement analyzed by two di fferent methods (JMA analysis and relationship between hydride fraction and reaction rate). ................................ ................................ ............................ 150 7 1 External surface images at higher magnification of (a) as magnesium spherical and (b) magnesium flake particles. ................................ 161 7 2 The hydrogenation curves of different shapes of particles. The hydrogen capacity of magnesium flakes s hows more than 5.5 H wt. % which is much highe r than the as ticles with 0.2 H wt. % capacity. ................................ ................................ ................................ ........... 161 7 3 The SEM (BSE mode) images of hydrogenated as magnesium particles after hydrogenation at 300C 1MPa for 15 hours. ........... 162 7 4 The SEM (BSE mode) images for the hydride layer measurement at higher magnification of hydrogenated as hydrogenation at 300 C 1MPa for 15 hours. ................................ .................... 162 7 5 The SEM (BSE mode) images of hydrogenated magnesium flake particles after hydrogenation at 300C 1MPa for 36 hours. (a) Thick flake particles (b) thin flake parti cles. ................................ ................................ ............................ 163 7 6 The schematics illustration of effect of shape change from sphere to flake on hydrogen capacity of magnesium particles with SEM (BSE mode) images for the comparison. ................................ ................................ ................................ 164

PAGE 15

15 7 7 The hydrogenation curves of different sizes of particles. The hydrogen icles with 0.2 H wt. % capacity. ................................ ................................ ................................ ........... 165 7 8 particles after hydrogenation at 300C 1MPa for 10 hours compared with the images of (b d) which show the cross section of hydrogenated as received ................................ ................................ ............ 166 7 9 (a) Prolate spheroid spheroid for flake ............ 166 7 10 The experimental hydrogen capacity compared with theoretical calculation by ........................... 167 8 1 (a) Hydrogenation curve of hcp Mg at 210 C under 1MPa pressure. (b) Evaluation of different stages of hydrogenation of magnesium powders based on JMA relationship. ................................ ................................ .............. 176 8 2 SEM (BSE mode) mic rographs (a) of the saturated powder revealing the impingement of magnesium hydride grains/colonies and (b) of the partially hydrogenated powder showing the presence of the hcp Mg phase on the surface. ................................ ................................ ................................ ............. 176 8 3 Dehydrogenation curves obtained using TGA of the partially and completely hydrogenated powders at 5C/minute and under He flow, revealing the difference in the hydrogen release temperature between the samples. ........... 177 8 4 SEM (BSE mode) micrographs of the partially dehydrogenated powder. (a) Lowering magnification of the particle showing the release of hydrogenation around the particle with MgH 2 inside the particle. (b) Higher mag nification image showing the extension of the hcp Mg phase to the surface of the particle. ................................ ................................ ................................ ............. 177 8 5 (a) Hydrogenation/dehydrogenation curves. The hydrogenation condition is the same for all pow der samples (At 210C under 1 MPa for 5 hours). The dehydrogenation tests were done under 0.1 MPa at 290C, 295C, 300C for 5 hours. The SEM/BSE image at point P is shown in Figure 1(b), in the vicinity near the powder surface of the partial dehydrogenate d powders after 5 minutes releasing at 300 C under 0.1MPa. ................................ .................. 178 8 6 Hydrogenation/dehydrogenation curves. The hydrogenation tests are done at 210C under 1MPa for 15 minutes and 5 hours. The d ehydrogenation tests are done under 0.1 MPa at 290C for 5 hours immediately after hydrogenation tests for both powder samples. ................................ ................. 178 8 7 The SEM/BSE images of as hydrogenated powder. In Figure 3(a), SEM/BSE images show the hydride distribution in the hydride powders hydrogenated at

PAGE 16

16 210 C under 1 MPa for 5 hours. The Sample powder exhibits pseudo saturation behavior. In Figure 3(b), the SEM/BSE images shows the hydride distribution in the hydrid e powders hydrogenated at 210C under 1 MPa for 15 minutes and the sample powder shows unsaturated behavior. ................... 179 8 8 Hydrogenation/dehydrogenation curves. The hydrogenation tests are done at 2 10C under 1MPa for 15 minutes. The dehydrogenation tests are done under 0.1 MPa at 290C for 5 hours at 275C, 280C, 290C, and 300C. ...... 179

PAGE 17

17 Abstract of Dissertation Presented to the Graduate School of the University of Fl orida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MICROSTRUCURAL BASED CH A RACTERIZATION OF HYDROGE NATION AND DEHYDROGENATION MECHA NISM S IN MG POWDERS By Hung Yu Tien December 2010 Chair: Fereshteh Ebrahimi Major: Materials Science and Engineering Metal hydrides display moderately high gravimetric and volumetric capacities and are being considered as potential solid state hydrogen storage materials for on board fuel cell powered vehicles. Among them magnesium hydride has been widely studied in the past few decades owing to its abundance, low density, reversibility, and reasonable ideal capacity of 7.6wt%. However, the hydrogen c apacity of magnesium is usually below i ts theoretical value and varies w ith hydrogenation condition and powder size and shape One of the main objectives of this dissertation was to develop a mechanistic understanding of the hydrogenation process in magnesium powders based on detailed microstructural analyses i n order to shed light on the parameters that control the hydrogen capacity We have demonstrated that the hydrogenation of m agnesium powders at 1Mpa pressure and in the temperature range of 150 to 300 C takes place in three stages. Stage 1 is the nucleation of the magnes ium hydride, Stage 2 is the growth of magnesium hydride colonies up to their impingement, and Stage 3 is the further growth of the continuous hydride layer. The nucleation of the hydride was found to be limited to

PAGE 18

18 the powder surface, except at 150 C, where internal hydride formation was observed. Quantitative metallography suggests that at high temperatures the magnesium hydride nucleates at the intersection of magnesium grain boundaries with the powder surface and consequently the number of hydride nuclei remains constant during Stage 2. The activation energy associated with Stage 2 was measured to be 32.4 kJ/ mole which is close to the diffusion of hydrogen atom through the hcp Mg metal. Stage 3, is believed to take place by the diffusion of hydrogen atom s through the hydride layer, which is quite slow and leads to the observation of a pseudo saturation behavior that governs the hydrogen capacity. Through different approaches i t is elucidated that the reduced hydrogen capacity at low temperatures is assoc iated with an increase in the hydride nucleation rate which results in impingement of the hydride colonies at smaller hydride thickness values. The findings of this investigation also reveal that the powder size distribution, the average powder size and p article shape affect hydrogen capacity significantly. A broad particle size distribution was found to result in a gradual transition between Stage 2 and 3 and a less pronounced pseudo saturation behavior. It is demonstrated that decreasing the size of the particles in the powder and increasing their aspect ratio drastically increase the hydrogen capacity. These phenomena are explained based on the nucleation and growth of the magnesium hydride. Studies near the equilibrium conditions revealed that the nucl eation of the hcp Mg influences the kinetics of dehydrogenation process noticeably. The presence of metallic magnesium phase on the surface, owing to incomplete hydrogenation, was found to enhance the kinetics and reduce the dehydrogenation temperature by eliminating the

PAGE 19

19 energy barrier required for the magnesium nucleation during hydrogen release process. The presence of metallic magnesium phase on the surface eliminates the energy barrier for the magnesium nucleation during hydrogen release process. Th is effect results in the lowering of the dehydrogenation temperature of magnesium hydride.

PAGE 20

20 CHAPTER 1 INTRODUCTION The energy supply for transportation has been mainly based on fossil fuels such as coal in the nineteenth century and gasolin e and natural gas in the twentieth century. In twenty first century, oil crisis gripped the world due to the political and religious conflicts in Middle East which has 94% known oil fields [1] It is also noted that to date no major crude oil field has been found since 1970s [2] Hubbert peak theory points that the rate of petroleum production tends to decline because of resource depletion after the point of maximum production and discovery rate [3 5] However, due to the releas e of the raising amount of carbon dioxide, the global warming and green house effects lead to a grow ing demand for the development of non polluting new energy carriers for on board auto mobile application. As for the environmental concern, zero levels of ha rmful emission should be a part of the energy cycles from all means of transportation [6 9] Hydrogen fuel, which is the key to satisfying the energy requirement and reduction of carbon dioxide emission, has a pot ential to become such a new energy carrier for the operation of fuel cells. Furthermore, hydrogen is the focus for sustainable energy source due to its abundance, easy synthesis and high efficiency [10] The volumetric hydrogen cap acity of hydrogen gas under 80Mpa is 40 kgH 2 / m 3 and that of liquid hydrogen is 71 kgH 2 / m 3 However, the safety issues and low volumetric density in compressed hydrogen storage and the technical issue of keeping the operating temperature below the boil ing point of liquid hydrogen limit their application in automotive transportation. Hydrogen storage materials in solid form play a crucial role in realization of fue l cells on board of vehicles owing to their volumetric hydrogen capacity being in the

PAGE 21

21 ran ge of 80 to 160 kgH 2 / m 3 [11] These materials, including metal and complex hydride s, are good potential storage methods due to their possibilities of high volumetric and gravimetric densities [12 16] Because of the relatively low pressures involved in the metal hydride technologies, they do not suffer the drawback as compressed and liquid hydrogen. Un like the co mplex hydrides most simple metallic hydrides, including magnesium hydride hav e almost 100% reversible capacity, which is very important for the on board automobile application. Among metal hydrides, light weight alkali metals such as l i thium and magnesium hydrides, with good hydrogen capacity, have become the first priority for th e renewable hydrogen storage application [17, 18] Although LiH has an ideal 12.5 H wt % storage capacity, under atmospheric pressure, its desorption temperature is above 450 o C due to its h ig h formation energy of 190 kJ/mole. On the other hand, magnesium hydride has a theoretical capacity of 7.6 H wt. % with an energy formation of 76kJ/ mole, which renders a lower hydrogen release t emperature of about 300 o C under atmosphere pressure [6]. However, as a n efficient hydrogen storage material this hydrogen release temperature is still too high for operation of the Proton Exchange Membrane Fuel Cells (PEMFC) [19 21] But recently, there is a trend to increase the operating temp erature of PEM fuel cell by using an electrolyte membrane of polybenzimidazole doped with phosphoric acid [22] Moreover, nano size magnesium hydride with a suitable catalyst could be used on board in a proposed two stage reservoir that enhances the kinetics of hydrogenation/dehydrogenation processes [23] In addition to the r equired low hydrogenation and de hydrogenation temperatures, the se process should take place within a reasonable time according the refueling rate and

PAGE 22

22 cycle life targets set by Department of Energy ( DOE ) [24, 25] Figure 1 1 shows a schemati c hydrogenation curve that demonstrates the pseudo saturation corresponding to the maximum capacity and the hydrogen absorption rate relating to the hydrogenation kinetics. The kinetics of h ydrogenation processes have been improved by alloying, use of catalysts on the surface and microstructural modification of the metal by processes such as ball milling [26, 27] While magnesium hydride has a reasonable ideal hyd rogen capacity, this level cannot be achieved under all hydrogenation conditions [26, 28 31] Many research groups have shown that the hydrogen capacity of Mg or Mg based alloy powder s depends on the hydrogenation temperature. In general, the hydrogen capacity decreases significantly with reducing the hydrogenation temperature [31 34] In general, two models are proposed to explain the hydrogenation kinetics in magnesium a nd its alloys. It has been argued that hydrogen absorption behavior of magnesium is better described by a three dimensional diffusion controlled shrinking core model [35, 36] This mechanism assumes that a continuo us layer of hydride forms immediately on the surface of magnesium powders upon hydrogenation. Hence, the kinetics of hydride formation is dependent on area of the interface at Mg/MgH 2 as well as the thickness of the hydride layer W ith growth of the hydrid e layer the area of the interface decreases and the hydride formation rate is reduced signi ficantly owing the reduction of hydrogen flux from the gas phase to the interface of Mg/MgH 2 As the hydride layer thickens, the diffusion length increases which in crease the time for hydrogen diffusion through the hydride layer In re semblance with porous membranes [37] a large hydrogen concentration gradient within the magnesium hydride is assumed

PAGE 23

23 to provide the necessary flux of hydrogen within the hydride. However, magnesium hydride is an ionic compound and according to the binary phase diagram of Mg H, it has almost no solubility range [38] Therefore, the assessment of the hydride as a porous layer is n ot suitable for explaining the hydrogenation behavior of magnesium. Nucleation and growth models have been suggested to describe the kinetics of magnesium hydride formation [29, 39, 40] However, t his approach does not imply a saturation level below the theoretical capacity because it is only developed for bulk materials and not for powders. I t has been proposed that the hydride grains may impinge on each other before the entire magnesium particle is transformed [28 30] According to this model, the capacity is defined as the point where the impingement is achieved. In general no microstructural analysis is reported to elucidate whether the growth mechanism is a thickening process of a conti n uous magnesium hydride layer or hydride forms as individual grains that eventually impinge on each other. Most of studies are calculated based on rate theories and there is a lack of microstructural evaluation of hydrogenation behavior o f magnesium powders. Also, the role of nucleation and growth of phases during dehydrogenation of magnesium hydride has not yet been investigated. The objective of this study is to investigate the microstructural evolution during the hydrogenation and dehy drogenation processes at different temperatures in order to develop a mechanistic understanding of the mechanisms that control the kinetics of formation of magnesium hydride and the maximum hydrogen capacity. Commercial magnesium powder was utilized in thi s investigation. Parameters such as hydrogenation

PAGE 24

24 method, powders size and size distribution as well as powder shape were studied to confirm the hypotheses developed in this study. Two different sizes of commercial magnesium particles with 44 and 300 m average size were used. For catalysis of dissociation of hydrogen molecules on the surface of powders, the powders were coated by nanocrystalline nickel through two different techniques. The hydrogen pressure employed for hydrogenation was 1MPa and for dehydrogenation was 0.1MPa hydrogen pressure. The effects of temperature on hydrogenation and dehydrogenation behaviors of magnesium powders were investigated using detailed cross sectional microscopy and quantitative metallography. A brief review of rele vant literature is presented in Chapter 2. The detailed experimental procedures are explained in Chapter 3. The results of this research are presented in five chapters. The characterizations of powders, including powder size, geometry, and also coating pro perties are given in Chapter 4. Chapter 5 describes the microstructural evolution during hydrogenation. Chapter 6 mainly focuses on the hydrogenation mechanism. The kinetics of magnesium hydride formation is discussed by the JMA (Johnson Mehl Avrami) analy sis from the hydrogenation curves. For the first time a three stage hydrogenation process is described and its validity is proven by especially designed hydrogenation tests. Chapter 7 is about the geometry and size effect of magnesium powder on the hydroge n capacity. The results presented in this chapter confirm the hypothesis developed in Chapter 5 and show that decreasing the size and changing the shape of particles from spherical to flake increases the total hydrogen capacity significantly. Chapter 8 is the investigation of dehydrogenation mechanism of magnesium hydride The dehydrogenation tests in this study reveal that

PAGE 25

25 the presence of hcp Mg phase on the surface plays a very important role for the enhancement of the dehydrogenation kinetics.

PAGE 26

26 Figure 1 1. The schematic hydrogenation curve showing the hydrogen capacity and the hydrogenation kinetics.

PAGE 27

27 CHAPTER 2 BACKGROUND 2 1 The S tructure of M agnesium and Magnesium H ydride Magnesium is a hexagonal closed packed structure with lattice constants a=3.2A and c= 5.2A, as shown in Figure 2 1. In the occurrence of hydrogen gas, magnesium absorbs and reacts with hydrogen under standard conditions. At low concentration, hydrogen atoms dissolve in the magnesium matrix, while at high con centration the compound magnesium hydride precipitates from the solution. M agnesium hydride is a like bonding between Mg and H as shown in Figure 2 2 The magnesium cation is octahedrally surrounded by H anions: the H a nion has a planar triangle of magnesium cations around it. The lattice parameters are a= 4.5A and c= 3.0 A [41, 42] Magnesium hydride forms epitaxially relative to magnesium in the thin film sample with a crystallo graphic orientation relationship to be MgH 2 (110)[001]//Mg(001)[100] on Al 2 O 3 and MgH 2 (200)[00 1]//Mg(110)[ 111] on LiGaO 2 [42] In Figure 2 3, the schematic shows the probable epitaxial growth of magnesium hydride on a thin film of magnesium. In the ball milling, a part of the magnesium hydride transform s to an [31] In this study magnesium hydride refers to a phase crystal structure since no milling is performed by using hydrogenated magnesium powders. 2 2 Thermodynamic P roperties of Magnesium H ydride At a specific temperature and pressure, the driving force of the hydride reaction (Mg + H 2 = MgH 2 entropy. f f (2 1) Figure 2 [27]

PAGE 28

28 The enthalpy of magnesium hydride f is negati ve ( 76kJ/mole) and represents the heat released during the reaction. S is mainly from the reduction of the gas phase to the chemisorbed hydrogen ( 130J/moleK) [18, 26, 43 49] A negative h formation of magnesium hydride indicates that this process is thermodynamically favorable while enthalpy formation is dominant However, at high temperatures, entropy of formation plays an important role and magnesium hydride formation be comes positive and dehydrogenation is thermodynamically favored. The binary Mg H phase diagram is shown in Figure 2 4 [38] During the magnesium and hydrogen reaction, at moderate pressure the hydride phase presenting in equilibrium with magnesium refers to magnesium hydride. The Mg H phase diagram is shown in Figure 2 4 which indicates the stable p hases. The reaction of magnesium hydride formation is shown as follows : Mg+H 2 MgH 2 For the formation reaction to proceed, the hydrogenation temperature and pressure should be by some means below the equilibrium temperature at a given pressure to prevent t he inversed reaction and confirm that each hydrogenation test was done at hydrogenation condition only. The equilibrium temperature of the hydride reaction is 287.7 C at 1 bar pressure [38] It is required to apply enough thermal energy and fast kinetics to induce magnesium hydride to be decompose d above 300 C [26, 50, 51] In practice the dehydrogenation temperature is higher than this temperature in order to provide the driving force for the nucleation of new phases. Figure 2 5 shows the schematic diagram of the energy profile of hydrogen absorption and des orption processes. Even if the hydrogenation reaction is thermodynamically favorable due to

PAGE 29

29 the negative Gibbs free energy, the reaction cannot take place with low thermal energy because the energy for hydrogen absorption needs to be overcomed. Therefore, the catalysts are necessary to lower the energy barrier of hydrogen absorption. I will talk about the catalyst in a later section. The following section will explain the alloying effect for the reduction of formation enthalpy of magnesium hydride during de hydrogenation. In Figure 2 5, the destabilization of magnesium hydride not only reduces the formation enthalpy but also lowers the energy barrier of hydrogen desorption [43] Baldi et al. reported that the thermodynamic properties were tuned by means of elastic clamping [52] The pressure composition isotherms show that the magnesium film coated by Pd or Ni exhibits a hydrogen plateau pressure more than two orders higher than bulk magnesium at the same hydrogenation condition. 2 2 .1 Pressure Composition Temperature (PCT) of M agnesium Figure 2 6 shows an example of PCT curves of magnesium First of all, hydrogen beg ins to dissolve in magnesium at interstitial sites and the solubility increase s with increasing pressure. At low hydrogen concentration, the hydrogen solubility is proportional to the square root of hydrogen pressure [53] When the concentration of hydrogen achieves the terminal solubility limit of magnesium at the plateau pressure, hydrides begin nucleating and hence growing. In the region of flat plateau, a solid solution of hydrogen in magnesium and MgH 2 phases coexist. Above the flat plateau, hydride transformation is complete and the slope is very sharp due to no solubility of H in MgH 2 The m ost important is the equilibrium plateau pressure during dissociation of H/RT S/R, P is equal to equilibrium pressure and T is constant in an isothermal process. The slope of the straight line

PAGE 30

30 represents the stability of hydrides. Thus, the stability of hydride is measured from the relationship of plateau pressure and tempera ture, lnP/(1/T)= H/R, which is the slope of the pressures. With at least two PCT curves, we can obtain a line with a slope which corresponds to the formation enthalpy of hydrides over R. 2 2 .2 Alloying E ffect Magnesium alanate, Mg(AlH 4 ) 2 with 9.3 wt% hydrogen capacity produced by chemical synthesis has been shown experimentally to release 7 wt% of hydrogen at around 135 o C [54, 55] However, theoretical calculations suggest that this alanate is metastable at room temperature and 1 bar pressure [44] Complex hydrides are usually not reversible since the final products cannot be transformed back to the original complex hydrides; hence losing the ability for cycling More recently, extensive research has been focused on destabilizing magnesium hydride, i.e. reducing its enthalpy of formation and/or increasing its entropy of formation. Song et al. assume that alloying elements will reduce the formation enthalpy based on the first principle calculation. If t he alloying element is X, metal X thus dissolves in the magnesium lattice as a solid solution in the form as (Mg, X)H 2 [56] Mechanical ball milling is one of the methods to alloy magnesium with other elements, i.e. Al [57] Ni [32, 48] Cd [32, 58] In [32] Si [18] and Sc [59, 60] which have been shown to be able to dissolve in magnesium at different solubility levels. However, the result shows that t hey do not cahnge the thermodynamic properties of the formation enthalpy of magnesium hydride

PAGE 31

31 2 3 Kinetic Properties of Magnesium Hydride 2.3.1 Hydrogen Capacity of Magnesium Powders The hydrogenation properties of commercial magnesium have been widely studied for a few decades due to its light weight, abundance, reasonable hydrogen capacity, etc. The studies go back about 50 years to the ground breaking study of Stampfer et al. [61] Magnesium has an ideal 7.6 H wt. % storage capability. However, it is difficult to fully hydrogenate magnesium powders. Later in the 1980s, detailed hydrogenat ion experiments at various temperatures and pressures were carried out by Vigeholm et al. on commercial magnesium powder [62, 63] They conclude d that any magnesium product with diameter less than 7 m will ab sorb hydrogen rapidly above 300 C under a hydrogen pressure beyond equilibrium pressure. Th e activation process was not required or needed only a few cycles of hydrogen absorption and desorpt ion. The highest rate observed was around 350C a nd absorption kinetics decreased with slightly increases in temperature due to the adverse effect. It was seen that the lower the hydrogen pressure, the closer the eventual reaction approached complete trans formation from magnesium to single phase magnesium hydride [28] In the initial absorption process at 400 o C, it was observed that the lower the hydrogen pressure, the closer the reaction approached stoichiometric hydride 7.6 H wt. %. The hydride fraction shows 97.6% at 2MPa, 95.5% at 3.2Mpa, and only 81.5% at 4.8MPa. There is an adverse effect on hydrogen capacity as the temperature increases above 350 o C owing to the competition of nucleation and growth taking place at 1.2MPa.

PAGE 32

32 2.3.2 Kinetics Model of Hydrogenation of Magnesiu m Powders 2.3.2.1 Impingement Model The dissociated h ydrogen atoms penetrate the surface layer and diffuse inside the bulk magnesium when the hydrogen concentration reaches the solubility limit of hydrogen in magnesium ; nucleation of magnesium hydride take s place [26, 30, 64] In this system the asymptotic hydrogen uptake as a function of time decreases with increasing hydrogen pressure [53] The hydride nucleation reaction rate increases with increasing applied hydrogen pressure [26, 30, 65] Theoretically, a hydr ogen atom would diffuse along the interface of m etal and hydride If somewhere below the surface hydrogen accumulated hydrides may nucleate inside the metal [26] The kinetic energy barrier can be described by Arrh e n ius relationship equation that relates to the activation energy and is a temperature dependent reaction The rate constant of most thermally act ivated processes is usually described well by the following equation [66 68] : K = K 0 exp( Q /RT) (2 2) where K 0 is the prefactor and Q is the activation energy. The Arrhenius relationship implies that a plot of mea sured rate constants as a function of the reciprocal temperature results in a straight line with a slope Q/R [66] The activation energy reported so far for hydrogenation is in the range of 66 to 308 kJ/mole. An introductory description is giv en here for the nucleation and growth theory regarding the relationship between the fraction transformed and the time in isothermal t ransformation [69] Consider first the individual transformed region. The size o f the transformed region increases continuously as a linear function of time ( is called the induction period) as shown in Figure 2 7. The reduction of reaction rate is owing to the impingement of the individual transformed volumes. I v is defined as the

PAGE 33

33 nucleation rate per unit volume. It is supposed V is t he un reacted volume between t= At t= +d some of the individual transformed regions nucleate in some arbitrary minimum size. The nucleation number of n in the time interval between t= and t= +d is I v V The extended transformed volume during this time pe riod i s dV ex =v I v V is the volume of an individual transformed phase. Thus the transformed extended follows : V ex =(4 V /3) 0 t I v G 3 (t ) 3 d (2 3a) G is the growth rate of new phase and 3D spherical growth is used in this case. However, if V ex is the extended volume of the phase the overlapped region should be taken into account. The actual V has a relationship with extended V ex : dV = (1 V /V )d V ex (2 3b) which can be re written as V ex = V ln(1 V /V ) and ap ply this equation to the previous one it can be summarized as following: ln(1 f )= (4 V /3) 0 t I v G 3 (t ) 3 d (2 3c ) f= V /V is the volume fraction of hydride. If the nucleation rate, I v and growth rate, G, are constant values, the relation can be shortene d as: f=1 exp( I v G 3 t 4 /3) (2 3d) The J ohnson Mehl Avrami equation is applied to describe reacted hydride fraction, f, as: f=1 exp( Kt n ) (2 3e) w here K is ( I v G 3 /3) which is related to the nucleation rate, growth rate and geometric factor ; and n is the reaction order and is not temperature sensitive [39, 68, 70] The way to investigate the hydrogenation kinetics is to analyze the hydrogenation curve s and obtain K and n values. Furthermore, the activation energy a nd rate limiting step can be

PAGE 34

34 realized based on the assumptions in the hydrogenation process Table 2.1 shows the summary of the dependence of n on growth dimensionality, rate limiting process, and nucleation behavior as reported by Karty et al. [29, 70] However, the listing in Table 2.1 does not exhaust all hydrogenation behaviors, but it does illustrate that in general the determination of reaction order does not provide a unique identification of the kinetics pro cess except n = 0.5 or 4. It is reported that the value of reaction order, n for hydrogen absorption of Mg 2 Cu catalyzed magnesium was equal to 0.5 and equal to 1 for hydrogen absorption for vapor deposited magnesium [70] Fernandez et al. reported that a nucleation and growth mechanism as illustrated by JMA equation with the reaction order 0.5 1 for the absor ption of hydrogen provides the best fit to the hydrogenation curves [68] Fernandez et al. investigated the hydrogen absorption kinetics by using activated magnesium powders [71] Their assumption for the hydrogenation mechanism was provided by Friedlmerier and Groll [30] A nucleation and growth model was employed in the assumption of their model. Figure 2 8 shows the isothermal absorption divided into two cases based on the pressure dependence of the applied hydrogen pressure. Figure 2 8 (a) shows absorption for an applied pressure close to the equilibrium pressure; the rate limiting step is the chemical reaction which is proportional to the interface area of Mg/MgH 2 The hydrogenation curve is sigmoidal. This situation is similar to solidification with very small under cooling which results in very limited nuclei. The reaction of hydride formation eventually leads to a large fraction of a magnesium particle An extremely slow diffusion of hydrogen is now the rate limiting step due to the very thick hydride layer [30] In Figure 2 8 (b), under a large under cooling, a large number of nuclei are formed around the surfaces as the applied

PAGE 35

35 hydroge n pressure is much smaller than the equilibrium pressure. The transformation volume of magnesium hydride is much less than in the former case. The growth rate is much slower and a thin hydride layer is formed near the surface. The chemical reaction is excl uded since no sigmoidal curve is observed. Although the initial kinetics is fast, the hydride fraction is relatively small. 2.3.2.2 Core Shell Model Martin et al. have given the following steps for the absorption of hydrogen: (1) physisorption of hydroge n molecules, (2) dissociation of hydrogen molecules and chemisorption, (3) surface penetration of hydrogen atoms, (4) diffusion of hydrogen through the hydride layer, (5) hydride formation at the Mg/MgH 2 interface [36] Chou et al. suggested that only hydrogen diffusion in hydride phase is slow; the other steps all proceed very fast [35] As seen in Figure 2 9, hydrogen first reaches the surface of the particle, is absorbed and then diffuses into hydride. When hydrogen arrives at the interface at Mg/MgH 2 a fast hydrogenation reaction takes place. The hydrogen layer becomes thicker and thicker as time increases. Finally, the hydrogenation reaction stops because the hydrogen is not able to diffuse through the thick hydride layer under that hydrogenation condition. A formula to describe the transformation fraction of absorbed hydrogen with time is shown as follows: [1 (1 f) 1/3 ] 2 = 2D H (2 4) f: hydride volume fraction. t: hydrogenation time. D H : diffusivity of hydrogen th rough magnesium hydride. 2

PAGE 36

36 R: radius of the particle. v: a coefficient that depends on substance and reaction. The relation is particle size dependent and the hy dride fraction decreases with an increase in size of particle. Therefore, unlike the nucleation and growth model, size should be taken into account in the Core Shell model. 2 .3.3 Effect of C atalyst At the beginning of hydrogenation was the physisorption pr ocess; molecular hydrogen is first physisorbed on the surface of magnesium powders The stick coefficient of hydrogen and the surface prop erties are affected by oxide, CO and other impurities [72] While the hydrogen molecule attaches to the surfac e of metals, the dissoci ation of hydrogen molecules takes place and the process is rel ative to the height of the energy barrier of hydrogen absorption [43, 73] From Zaluzka et al [26] ball milled magnesium powder is hydrogenated without catalyst at 300 o C and 400 o C. The curves show slow absorption rate of hydrogen due to the kinetics limit ation. Fast hydrogenation demands the increase of kinetics by surface treatment. In the standard condition, t he potential energy of atomic 2H is much higher than that of the gas molecule H 2 as shown in Figure 2 10 [73] When an interface is applied through a catalyst to modify the hydrogenation kinetics, the ener gy curve of atomic 2H is lowered and intersects with the curve of the potential energy of a gas hydrogen molecule at lower activation energy. This means the activation energy, Ea, is reduced. If Ea is less than zero, the interface becomes non activated. Wh en it has a non activated interface, the potential energy of 2H decreases with increasing depth from the surface and becomes major chemical form. The potential e nergy of chemisorption is moderated for hydrogen to diffuse into metals.

PAGE 37

37 A catalyst is mainly beneficial for dissociated absorption of a hydrogen molecule in hydrogenation [33, 74, 75] With a catalyst of Pd, the kinetics of absorption significantly increases by lowering the activation energy of dissociatio n of H 2 [26, 76, 77] Laing et al found that Ti is the most rapid in hydrogenation while V as catalyst in dehydrogenation is the most r apid among Ti, V, Mn, Fe and Ni [27] Gremaud et al. used Ti catalyzed material for the study of the reversibility of magnesium alanate [78] The drawback of Ti as catalyst is the formation of stabl e TiH 2 which does not allow the cycling ability [79] The dehydrogenation can be complete at 235 o C within 40 minutes using 5 at% V as catalyst but it does not change the thermodynamic properties of magnesium hydride [80] From Charlot et al [75] for dissociated absorption of a hydro gen molecule, transition metals produce a charge from TM H to weaken H H bonding and favor chemisorption. The lowing spin multiplicity might become important in chemisorption of hydrogen, and the sp orbital of Ni and the d orbital of Pd are mainly involved in the breaking of H H bonds. Beside the transition metals, metal oxides and fluoride were considered as an alternative route in the investigation of catalyst for hydrogen storage materials [81, 82] T he catalytic effect may be attributed to the electronic exchange reaction with a hydrogen molecule which accelerates the kinetics of reaction. Oelerich et al show ed that oxides display good efficiency as metal atoms having multiple valence states [83] It also benefit s from the very high defect density introduced by metal oxides at the surface. Barkhordarian et al reported that t he d esorption rate of Nb 2 O 5 is the fastest among those metal oxides, Fe 3 O 4 V 2 O 5 Mn 2 O 3 Cr 2 O 3 TiO 2 Al 2 O 3 SiO 2 [84] The dehydrogenation is complete within 140 seconds at 300 o C in vacuum, and it is much faster than pure V [80] Friedrich et al assume d at least large

PAGE 38

38 part of the reported cata lyst effects are due to the oxides present [85] We cannot prevent oxides formation during such a long milling time, so the effect of metal oxides becomes important for hydrogen sorption and desorption processes. Oxides are much cheaper than pure metals for mass production, and we can add only small amounts to have a substantial effect. W ith 0.5 at% Nb 2 O 5 addition the desorption process completes in about 90 seconds at 300 o C under vacuum which is the fastest hydrogen desorption rate so far [86] 2.4 Microstructural Analysis of Hydrogenated Mg Powders A microstructural evaluat ion of hydrogenated magnesium powders in different phases of the cycling tests is supported by Figure 2 11 [87] Figure 2 11(a) shows the polished cut image of a powder sample after first hydrogenation. The near ly spherical growing zones of magnesium hydride around the powder surface can be clearly seen and the penetration depth of hydride is not uniform. The nucleation and growth mechanism were validated by the individual growth of surrounding hydride grains nea r the powder surface. Figure 2 11(b) displays the cross sectional images of powders hydrogenated after a series of long term hydrogenations. The hydrogenation and dehydrogenation conditions are 350 o C under 25 bars and the same temperature under 10bar, resp ectively. The cycling time for long term is 6 hours. The cross sectional images exhibit a continuous hydride layer with a relatively smaller fraction of hydride in the powders. Figure 2 11 (c) reveals the cross section of hydrogenated powder after a few sh ort term hydrogenations. Short term means the cycling time is 2 hours. The hydride distribution is observed as a thinner layer than the hydrogenated powders after a series of long term cycling. In Figure 2 11 (d), after a series of short term cycling, the

PAGE 39

39 hydrogenated powder shows the thinnest hydride layer among all the cycling conditions. It can be concluded that the surface condition is very important to the hydrogenation behavior, and also the final hydrogen capacity. After a series of short term cyclin g tests, it is suggested that a large number of nucleation sites form and they favor the nucleation of magnesium hydride. 2.5 Effect of Shape and Size on Hydrogen Capacity of Metal Hydride The kinetics of hydrogenation and dehydrogenation of metal hydride s have been improved mechanically by increasing the surface to volume ratio, the grain boundaries, reducing the grain size and enhancing the hydrogen diffusion within the metal bulk which leads to the increase of hydride fraction [88] The effect of size of magnesium particles on hydrogenation kinetics has been investigated by many groups, and mechanical ball milling has been used to reduce the particle size in nano scale and also reduce the grain size of the particles. The hydrogen kinetics is relatively fast in the milled magnesium powder sample and also the hydrogen capaci ty of milled magnesium particles is higher than that of un milled ones. However, the effect of shape of magnesium particles on hydrogen capacity has not yet been investigated in depth. Mint et al. concluded that the shape variation of the powder particles can be ignored [89] hydriding by using a more complicated model than the former study. Three different shapes of sphere, cylinder, and plate are considered for their computer simulation. The Core Shell model has been used for their model development and assumption. The result reveals that the effect of change in powder geometry is not drastic [90] Jeon et al. use the High Speed Orbital Ball Milling (HSOBM) to produce the magnesium flakes [88] Their result shows that the hydrogen capacity of magnesium flakes is 4.57 H wt. %

PAGE 40

40 by using TGA for the predictio n of hydrogen absorption. Unfortunately, no hydrogenation data is obtained for the spherical magnesium particles. They have described that, for the magnesium flake, the flux area remains constant with respect to the hydrogen penetration depth but for the m agnesium spherical particle the flux of hydrogen decreases dramatically with increasing hydride thickness. This is clearly the assumption from the Core Shell model which suggests chemical reaction is the rate controlling step during the hydride growth.

PAGE 41

41 Table 2 1 Dependence of n in the JMA equation n Rate limiting step Growth dimensionality Constant nuclei number Constant nucleation rate Diffusion control 1 0.5 1.5 2 1 2 3 1.5 2.5 Interface control 1 1 2 2 2 3 3 3 4 [70] Figure 2 1. The crystal structure of hexagonal closed packed (HCP) Mg. Figure 2 2 The crystal stru cture of tetragonal magnesium hydride

PAGE 42

42 Figure 2 3 Schematic illustration of the likely epitaxial growth mode of MgH 2 on Mg on (001) Al 2 O 3 in the MgH 2 (110)[001]//Mg(001)[100] relationship. For clearness, only the Mg atoms in MgH 2 are shown. [42] Figure 2 4 Mg H binary phase diagram at 1bar (0.1MPa) hydrogen pressure.

PAGE 43

43 Figure 2 5 Energy barriers for hydrogen absorption and desorption. Figure 2 T1>T2>T3) Figure 2 7. The schematic growth of a product region.

PAGE 44

44 Figure 2 8. Cross section of a magnesium particle at two different times a nd typical curves for hydriding at A) P~P eq and B) P>>P eq Figure 2 9. A schematic diagram for Core Shell hydrogen absorption model. Figure 2 10 Potential energy curve for hydrogen binding to a metal surface which shows : (i) physisorption for both activated and non activated processes; (ii) dissociation of hydrogen molecules and chemisorption; (iii) surfac e penetration and chemisorption; and (iv) diffusion of hydrogen atoms. [73]

PAGE 45

45 Figure 2 11. Microstructural analysis of hydrogenated magnesium powders after (a) first hydrogenation, (b) a series of long term cycling test, (c) a few short term cycling tests, and (d) a series of short term cycling test.

PAGE 46

4 6 CHAPTER 3 E XPERIMENTAL PROCEDUR ES The ideal high hydrogen capacity 7.6 H wt. % of magnesium was considered the most promising material for on broad hydrogen storage material. However, high desorption temperature and low kinetics are the major drawbacks of magnesium h ydride. A fundamental model of magnesium hydride formation regarding thermodynamics and kinetics properties need s to be developed to enhance the efficiency of hydrogen absorption and desorption Magnesium powders were extremely sensitive to oxygen and mois ture; therefore all the handling and coating were conducted in an argon filled glove box or argon atmosphere environment. The glove box was maintained at very low oxygen and moisture levels to avoid any contamination during handling of powders. Figure 3 1 shows the appearance of the glove box manufactured by Vacuum Atmosphere (Model: Nexus and Part Number: 100026). Figure 3 2 displays the screen which shows the oxygen and moisture level. The oxygen concentration is maintained below 0.5ppm at all times. Moi sture level is kept below 2ppm (Dew point = 72 powders were used for various purposes of experiments i.e. effect of size and shape A catalyst was added to lower the energy barrier of the dissoc iation of hydrogen molecules to en sure that the continuous hydrogen source and other mechanism s interested were disclosed The Ni coated magnesium powders were hydrogenated using a hydrogenation part of our house built Hydrogenation/Dehydrogenation unit. D esorption of hydrogenated powders was conducted using the isothermal dehydrogenation part of the Hydrog enation/Dehydrogenation unit. The other dehydrogenation test was done by using Thermal gravimetric analyzer (TGA) which

PAGE 47

47 allows a constant heating rate du ring the hydrogen release The microstructure of the magnesium powders at different stages of hydrogenation/dehydrogenation processes was analyzed using SEM/BSE mode to study the hydride distribution by phase contrast. The material processing procedures, a nalytical testing and the sample preparation for various characterization techniques applied in this research are explained in this chapter. Surface to volume is believed to influence the hydrogenation behavior. For the effect of shape, magnesium f lake pow ders were made via HSOBM which applied shear force to the sphere particles and hence change d the shape of the magnesium powder from sphere to di s k particles 3.1 Material F abrication 3.1.1 Addition of Catalyst by Theta Composer To enhance the kinetics of hydrogen absorption and desorption magnesium powders they were coated with a catalyst. Many 3d transition elements are employed as catalysts for hydrogenation and dehydrogenation of magnesium powders. Ni is one of the effective catalysts that exhibit impro ved kinetics in these processes since a charge transfer from nickel to hydrogen weakens the bonding between H H Hence, the activation energy of hydrogen molecule dissociation is reduced. In this study, we employed a dry coating process using the Theta Com poser. 10 gram Mg/nano Ni mixture with 2 wt. % nickel nano particles (Argonide 70nm) and pure magnesium inside the glove box and entirely sealed. The Ni coating procedure was conducted under Ar atmosphere which was at the same condition in the glove box. The nickel coating was conducted at the optimum condition of 4600 rpm rotor speed, 70 rpm vessel speed and 1 mm gap between vessel and rotor The schematic of the Theta

PAGE 48

48 Composer is illustrated in Figure 3 3. Only 44 magnesium particles were employed by the Theta Co mposer Ni coating process. This material was used for the hydrogenation/dehydrogenation tests and JMA kinetics analysis in the Chapters 5, 6 and 8. 3.1.2 Addition of C atalyst by Organometallic Nickel Solution For the limited amount produced in magnesium f lakes by HSOBM, Organometallic solution was used for the Ni coating of a small amount of powder. This coating operation needs to be employed inside the glove box to avoid the exposure of magnesium powders to oxygen and moisture. An organometa l lic based Ni compound was used and the procedure was illustrated in the Patent: 4,554,152. The source of Ni is the dissociation of Bis (1,5) cyclooctadiene Ni (Strem Chemicals, 98.3% purity). A schematic illustration of the Ni coating setup is shown in Figure 3 4 The organometallic Ni and the magnesium powders were taken in 0. 1 18:1 proportion and were refluxed in a flask with toluene at 110C for 6 hours. In this study, 12 milling tests were done to collect 1.2g flakes for at least 4 hydrogenation tests including the c onsideration of reproducibility. The result will be shown in Chapter 7. For 1 .2 g magnesium flakes, 0. 1416 g organometallic Ni is used for high coating efficiency. A long jacket was attached to the top of the flask so that the toluene vapors are condensed on the walls and fall back into the flask called refluxing During the refluxing process, Ni nano particles break away from the organometallics compounds and attach onto the magnesium powder surface. After 6 hours, the Ni coated magnesium powders were colle cted by filter paper for the following hydrogenation tests In Chapter 7, 3 powder samples were used by this coating method. 1 .2 g ram of magnesium 44 300 and flakes powders was coated by organometallics Ni in the same coating condition. Size

PAGE 49

49 effect wi ll be evaluated between 44 and 300 magnesium particles and shape effect will be discussed between 300 sphere and flakes powders. 3.1.3 High Speed Orbital Ball Milling (HSOBM) Ball milling was the most common method in the mechanical processing of all oy powders with reduction of size and change in shape. Figure 3 5 shows the HSOBM. It has a sealed milling tube through which magnesium powder s are loaded inside the glove box. The milling tube has been machined to the dimensions shown, threaded on one end and sealed with an end cap. Figure 3 6 shows how the shear force is applied to the magnesium powder to change the shape of the particles from spheres to flakes. Under argon atmosphere, 0.1g ram of Alf a Aesar magnesium powder and a 6 mm in diameter 52100 c hrome steel ball with Rockwell C hardness 63 65 (Norstone Inc) we re loaded into a 316 stainless steel milling tube. The powder was then milled in the HSOBM using various combinations of starting powder size and milling speed/time. Milling speed in the HSOB M was contro lled via a variable transformer and the speed was measured using an optical tachometer. The flake powders will be mounted and the mounting was cut perpendicularly to the polishing surface, a s shown in Figure 3 7 for the thickness measurement of flakes or the hydride layer of hydrogenated flake powders. Figure 3 8 displays the relationship between the applied voltage and the milling speed in RPM, which is measured by a laser speed gun. The milling speed increases with increasing voltage. 3.2 Hydr ogen Absorption 3.2.1 Hydrogenation S etup Figure 3 9 displays the house built Hydrogenation/Dehydrogenation unit Fig ure 3 10 presents the schematic diagram of the Hydrogenation/Dehydrogenation unit with

PAGE 50

50 electrical connections. All the tubing used to compo se the system was from Swagelok and was fixed using corresponding pressure fittings and VCR connection to avoid any leakage. The hydrogenation setup is on the bottom side of the whole unit which is referred to the high and low pressure regulator with the rest of the tubing including the reservoir chamber and specimen chamber. The hydrogen absorption measurement by side of the regulator controls the hydrogen pressure at a desired constant value which means the hydrogen pressure in the specimen chamber is kept the same during the experiment and the reservoir chamber acts as a source of hydrogen tank which keeps supplying the hydrogen gas through the regulator to the sampl e powder to keep the pressure constant at the specimen chamber. The high pressure regulator is not used in this study and its highest operating pressure is 3MPa. The low pressure regulator has 0.2 5 % accuracy in the range from 0 to 1.5MPa and this regulator was used for all the hydrogenation tests in this study. A pressure transducer (PX880 1KGi) which can operate between 1 1000 psi was located at the middle of the hydrogenation step and it has very high accuracy of 0.2%. The pressure transducer is connected to a two way valve. This two way valve can be ether pointed to the specimen chamber or to the reservoir chamber. Before the data recording starts, the pressure change in the specimen chamber was measured to ensure a constant pressure. After the hydrogenat ion starts, the two way valve was pointed to the reservoir chamber to measure the dropping pressure by time. The analog signal from the voltage change was converted to a digital signal from the pressure transducer to an Omega A/D converter (DAQ 55) and sim ultaneously the pressure was also observed using the Omega

PAGE 51

51 pressure indicator (DP25B E). An Omega J type thermocouple (TJ36 ICSS 180 6) inserted into the specimen chamber was located as close as possible onto the sample surface. The thermal couple was conn ected to two places. One was to a temperature controller (Glas col heater) and the other was to the A/D converter for recording real time data. The temperature controller has the ability to use PID (Proportion; Integration; Derivative) 3 parameters to prec isely control the temperature within +/ 0.5 C in an auto tuning mode. Another therm o couple was attached on the surface of the reservoir chamber to indicate the temperature of the hydrogen gas and this was maintained at room temperature by using another te mperature controlling system (See Figure 3 8). The reservoir chamber was wrapped by a copper tube which has in/out plug onto the system to circulate water. The temperature of the reservoir chamber was maintained at 24.5 C to avoid the noise signal generate d from the room temperature change. The thermocouple attached on the reservoir chamber was also connected to the A/D converter. The pressure and temperature data were sent instantaneously from the A/D converter to the computer program LabView 8.0. 3.2.2 H ydrogenation P rocedure After the powder sample was loaded inside the glove box, the hydrogenation chamber was attached to the Hydrogenation/Dehydrogenation unit. Prior to hydrogenation of the magnesium powder, a leak test was conducted at a hydrogen pressu re of 1MPa for one h our to check the system. The powder was loaded and sealed in a stainless steel hydrogenation chamber under the argon atmosphere in a glove box. The whole setup was connected to high purity Ar gas (Airgas, 99.99999 %), high purity hydrog en gas (Airgas, 99.99999%) tanks and gas was removed via a mechanical pump. The unit was flushed three times with argon and finally with hydrogen

PAGE 52

52 to a pressure of 3 Pa. The chamber was heated using a wrapped heater rope (Omega FGR 030). Considering that th e nucleation rate of the hydride phase at a given pressure is proportional to the driving force, we designed two methods for hydrogenation. In Method I, the specimen chamber was first pressurized with hydrogen to 1 MPa and then the temperature was raised at an approximate rate of 12C/min from room temperature to 210 C. During this procedure, the pressure sensor was directed toward the hydrogenation chamber and the valve located after the pressure regulator was closed such that no hydrogen would flow from th e reservoir chamber to the specimen chamber. Upon heating, initially the gas expansion in the hydrogenation chamber resulted in an increase in the pressure and it reached a maximum value of 1.13 MPa, at which point the pressure dropped upon further hydrogen absorption. The pressure saturated at the level of 0.85 MPa after 75 min. The amount of hydrogen absorbed was estimated using the pressure decrease as calculated from the following equation: dP= P P 0 where P and P 0 are the pressure values inside the cham ber at time t obtained during testing an empty chamber and during the hydrogenation test, respectively. Figure 3 11 shows the real time calculation of dP by subtracting the pressure change during the heating of an empty chamber test. Any effect of heat rel ease owing to the exothermic nature of the hydrogenation reaction on the temperature change was ignored. However, this change is expected to be very small because the temperature registered as a function of time during hydrogenation test was similar to tha t recorded during the test with an empty chamber. The total number of moles of hydrogen atoms absorbed for Method I n, was estimated assuming the ideal gas equation.

PAGE 53

53 n = 2(V 1 dP/RT 1 + V 2 dP/RT 2 ) where V 1 is the volume of the specimen chamber which is at the hydrogenation temperature of T 1 and V 2 is the volume of the tubing between the specimen chamber and the pressure sensor which is at the room temperature of T 2 and R is the ideal gas constant. The second approach, Method II, consisted of first heating the specimen chamber under a low partial pressure of hydrogen (approximately 3 Pa) up to 210C and then increasing the pressure to 1 MPa. For this method, the reservoir was pressurized initially to 1MPa and the pressure regulator was kept open during the hydro genation process. As the sample absorbed hydrogen, the pressure regulator allowed the flow of hydrogen from the reservoir to the specimen chamber such that the pressure remained constant in the specimen chamber throughout the test. The amount of hydrogen a bsorbed was calculated from the pressure drop measured in the reservoir chamber, which was at room temperature. Again, the ideal gas equation was employed to compute the amount of hydrogen absorbed from the measured pressure changes in the reservoir. The t otal number of moles of hydrogen atoms absorbed, n, for Method II was estimated assuming the ideal gas equation : n = 2 V dP /RT V is the volume of reservoir, dP is the pressure drop during hydrogenation, and T is a constant hydrogenation temperature. Each tes t was repeated twice to ensure the consistency of the results. After hydrogenation, the samples were stored in the glove box under low oxygen pressure and moisture level. Method I was applied only in Chapter 6, the effect of change in the hydrogenation pro cedure. Other than that, Method II was mainly used in this study.

PAGE 54

54 3.3 Hydrogen D esorption 3.3.1 Dehydrogenation S etup F igure 3 10 presents the schematic diagram of the Hydrogenation/Dehydrogenation unit with electrical connections. All the tubing used to c ompose the system was from Swagelok and were fixed using the corresponding pressure fittings and VCR connection to avoid any leakages. The de hydrogenation setup is on the top side of the whole unit which is referred to the back pressure regulator with th e rest of the tubing including reservoir chamber and specimen chamber. The hydrogen de sorption measurement by time was also based on the back pressure regulator controls the hydr ogen pressure at a desired constant value 0.1MPa, which means the hydrogen pressure in the specimen chamber is kept the same during t he experiment. The excess hydrogen gas produced during dehydrogenation goes through the back pressure regulator to k eep th e pressure constant at the specimen chamber. The hydrogen pressure increases in the reservoir chamber can be calculated as the real time drawings of de hydrogen ation curves The heat transfer was considered in the setup while removing the hydrogen gas follo wing the hydrogenation test. The gas temperature reached equilibrium at room temperature rapidly owing to a temperature control (NESLAB RTE 140) system which kept the reservoir chamber at 24.5C. 3.3.2 Dehydrogenation P rocedure After the sample has been hydrogenated, a dehydrogenation test is performed immediately after the hydrogenation. Usually the dehydrogenation temperature is set t higher than the hydrogenation temperature, i.e. 210 C at hydrogenation and 300 C at dehydrogenation. The trickiest part is the pressure setting during the heating process.

PAGE 55

55 The pressure setting was managed close to the equilibrium pressure of the desired temperature to assure that neither hydrogenation nor dehydrogenation takes place during the heating process. The purpose i s only to reduce any interference and keep the same condition of hydrogenated powders during the heating process For example, the pressure setting for 300 C during the heating process is 0.07MPa lower than the equilibrium pressure of 300 C. This extremely small amount of hydride formation can be ignored in this short time which is always less than 10 minutes in this time period since the hydrogen pressure is very low The dehydrogenation starts when the dehydrogenation temperature is reached. During the de hydrogenation experiments, 0.1MPa hydrogen pressure is kept constantly in the specimen chamber the whole time by using the back pressure regulator which allows the reverse flow of hydrogen gas from the specimen chamber to the reservoir chamber. The amount of hydrogen desorbed was calculated from the pressure change measured in the reservoir chamber, which was at room temperature and whose temperature was measured continuously over time. The total number of moles of hydrogen atoms desorbed was estimated assu ming the ideal gas equation from the measured pressure increase in the reservoir chamber n = 2 V dP /RT V is the volume of reservoir, dP is the pressure increase during dehydrogenation, and T is a constant dehydrogenation temperature. 3. 3 .3 Thermal Gravimet ric Analyzer ( TGA ) In addition to the dehydrogenation study in the house built Hydrogenation/Dehydrogenation system, the kinetics study of hydrogen release experiments was carried out in TGA (Setaram setsys). The two results of experiments

PAGE 56

56 are not comparab le because the heating procedure in TGA is not isothermal and the heating rate is set to be 5C/minute. The sample powder was heated from room temperature to a maximum temperature. 350C is usually the highest temperature used for the release of magnesium hydride in TGA During the heating process the changes in the weight of the sample were observed as a function of time and temperature. The weight loss by the hydrogenated sample during the experiment confirmed the reaction was taking place in the material and the weight loss indicates the amount of hydrogen released. Unlike the in situ test in the house built system, the hydrogenated powder need s to be removed from the specimen chamber inside the glove box. The hydrogenated sample powders were brought out of the glove box using a vacuum container. The powder was removed from the vacuum chamber and was weighed precisely at 12 mg. After recording the initial weight, the sample was loaded into the TGA chamber The weight loss during the experiment corresponds to the amount of hydrogen release d from the material expressed in wt% and was calculated by using 12 mg as the initial weight of the sample. An empty crucible test wa s performed to remove the instrumental error from the measured weight loss The s ame cruci ble was used in the same conditions T he weight loss observed during this experiment was subtracted from the weight loss in the previous experiment. The weight loss subtracted after the empty crucible run indicate s the actual amount of hydrogen present in the powder. After the temperature had reached 350C the furnace was cooled rapidly to room temperature at the rate of 100C/min utes to keep the microstructure at that temperature. The corresponding SEM ( BSE mode) microstructural analysis was carried out on the partial ly or full y dehydrogenated powder samples.

PAGE 57

57 3.4 Analysis Methods and Characteristic Techniques 3.4.1 Particle S ize D istribution The particle size distribution after Ni coating was measured using a Coulter LS 13320 laser diffraction system It i s assumed that laser scattered from a spherical particle is a function of angle, size and shape of the particle ; incident wavelength ; and the refractive index of the material to light. The scattering behavior obeys the Mie model which is the solution to M axwell's equations for the interaction of spherical particles with electromagnetic plane waves. More than a hundred detector array allows equations to solve for the unknown numbers of particles for each size. A 780 nm laser diode, light source, is used for analysis in the size range from 400 nm to 2 mm. In this study, the experiment for micron sized particles within this range was conducted using this laser diode. A tungsten halogen bulb and filters assembly sizes in the range 40 to 800 nm. Isopropyl alcoho l was used for the proper mixing of the powders and also enhances the resolution of the measurement. 3.4.2 Scanning Electronic Microscope ( SEM ) For the metallographic preparation of the samples, a small amount of powder was brought outside and embedded in an epoxy resin at room temperature. Due to the reactivity of magnesium, oil was used during the mechanical polishing processes and the samp les were never exposed to water, otherwise the artificial contamination disturbs composition analysis at SEM (BSE mod e). A thin layer of carbon coating was used to increase conductivity. Microstructural evaluations were conducted using a JEOL 6400 scanning electron microscope (SEM )

PAGE 58

58 3.4.3 Optical Microscop e (OM) After the powder sample was polished it was examined under an optical microscope (Leica DM2500M ) The software attached for the image capture and various kinds of measurement is ProgRes CapturePro 2.7 The eyepiece is equipped with a digital camera (Jenoptik ProgRes C5) with a real time image displaying in the so ftware. M agnesium looks brighter and magnesium hydride appears darker under OM. M agnesium is a conductor which reflect s the incident visible light and magnesium hydride has a large band gap which absorb s the light relatively eas ily Furthermore, the interf ace at the phase contrast between magnesium and magnesium hydride can be easily seen from the cross section images. The software was also used for the measurement of the hydride thickness and the geometry of the magnesium flakes i.e. thickness

PAGE 59

59 Figure 3 1. Glove box (Vacuum Atmosphere, m odel: Nexus ) Figure 3 2. Monitoring the O 2 and moisture level Figure 3 3. Theta Composer Ni coating

PAGE 60

60 Figure 3 4. Organometallic Ni coating. Black dot represent nano Ni particles and white one is the magnesium powde r. 110 C is the operating temperature. The arrows illustrate the cycling of toluene in the beaker. Figure 3 5. HSOBM milling

PAGE 61

61 Figure 3 6. Schematic illustration of HSOBM milling Figure 3 7. The relationship between applying voltage and milling speed Figure 3 8. Mounting sample cut perpendicularly to the flake surface for thickness measurement.

PAGE 62

62 Figure 3 9. H/D system actual view Figure 3 10. Schematic diagram of Hydrogenation/Dehydrogenation unit.

PAGE 63

63 Figure 3 11. The illustration of the calcul ation of pressure drops in Method I.

PAGE 64

64 C HAPTER 4 CHARACTERIZATION OF POWDERS This chapter presents the characteristics of magnesium powders used in this research prior to hydrogenation processes. The powder morphology, phase distribution, and microstructur e have been studied and described by using optical or scanning electron microscopes. The distribution of particle size s in the powders has been analyzed via laser scattering and/or direct measurement from cross sectional images obtained by metallography. Two different sizes of powders have been used, 44 and 300 average diameter. The 44 magnesium powder was used for most experiments in this study, i.e. t he effect of temperature, study of hydrogenation mechanism and dehydrogenation study. The 300 magne sium powder was employed to investigate the effect of particle size and shape on the hydrogenation properties A Ni coating (see Chapter 3) was the c atalyst used to enhance the dissociation of hydrogen molecules on the surface of magnesium particles. The p resence and distribution of the nano size Ni particles were evaluated using the energy dispersive spectroscopy (EDS) technique in SEM. 4.1 Characteristics of the As Received Mg powders 4.1.1 44 Mg P owder Fig. 4 1 shows the particle size distribution of magnesium powder obtained by the Coulter LS 23 0 system The average diameter based on the volum etric percentage was measur ed to be 4 4.29 which is similar to the average size suggest ed by the Alfa Aesar Company for this powder in the as received condition. Fig. 4 2 presents a n SEM picture showing the size and the shape of the as received 44 magnesium powder. The shape of the particles is not spherical and is more like a prolate spheroid. Figu re 4

PAGE 65

65 3 shows the powder cross sectional images of as received 44 magnesium powders. In the schematic presented in Figure 4 4, possible cross sectional images are illustrated. The shape and the average size of the particles were characterized by measuring the aspect ratio a/b (a is the short axis and b is the long one) and the (a+b)/2 from the cross sectional images obtained from the polished samples. Obviously the cross section we see depends on how the particle has been cut. In general, the shape of the image does not represent the shape of the whole particle. Figure 4 5 exhibits the particle size distribution of the as received 44 magnesium powder based on the number frequency, Figure 4 5(a) and volume frequency, Figure 4 5(b). The average diameter from the measurement of number percent is 29 and from that of volume percent is 32 The average particle size measured from cross sectional images is smaller than the average size obtained by the laser technique. This difference may arise from the number of samplings. More samplings would help the accuracy of the particle size measurement. 4. 1.2 Characteristics of the 300 Mg P owd er Figure 4 6 presents the SEM picture showing the particle geometry of the as received 300 magnesium particles. Figure 4 7 shows the cross sectional optical image of the as received 300 magnesium powder. Quantitative analysis of the shape of the parti cles in this powder using the optical images revealed that the aspect ratio is 0.67 although the powders look spherical in Figures 4 6 and 4 7. Figure 4 8 presents the particle size distribution of 300 magnesium particles based on the number frequency, F igure 4 8(a), as well as volume fraction, Figure 4 8(b). The average particle sizes based on these analyses were 436.26 and 438.57 respectively.

PAGE 66

66 4. 2 Characteristics of Mg F lakes In order to investigate the effect of particle shape, as was described in section 4 3, flakes were produced using HSOBM. The evaluation of the particles produced under different conditions revealed that the thinnest cross sections could be obtained at the highest voltage achievable in the HSOBM system. Figure 4 9 shows the thi ckness distribution for the particles milled for only 20 seconds at 100V. Particles with a very broad thickness distribution and an average thickness of 100 were observed after the HSOBM milling. As the milling time was increased to 1 minute the distribution shifted to smaller thickness values, as shown in Figure 4 10. For the particles milled for 2 minutes, thickness was measured using more than 50 flakes and the thickness distribution is shown in Figure 4 11. The thickness of the flakes ranges from 14 to 68 m and the average thickness is 36 m. Figure 4 12 shows the thickness distribution for the powder milled at 110V for 2 minutes. The thickness distributi on shifts more to the left by 110V compared with that by 100V milling. However, this milling test was done only once since the limitation of the vibration system is no more than 100V. So for safety it was not possible to produce thinner particles unless th e design of the vibration system was changed or the loading increased. Figure 4 13 displays the cross sectional optical images of the magnesium flakes. The thickness can be measured directly by this cross section image since the flake was cut at the center and perpendicular to the flat surface as described in Chapter 3 (Figure 3 5). Quantitative evaluations showed that the aspect ratio was reduced drastically from 0.67 to about 0.05 due to the milling process. Figure 4 14 presents the SEM image of the as mi lled 300 m magnesium flakes. In comparison

PAGE 67

67 to Figure 4 6, the particles show a definite change in the shape from semi sphere to disk. 4. 3 Characteristics of Ni coated Mg P owders 4. 3 1 Ni Coating Using Theta Composer In order to enhance the dissociation of hydrogen molecules on the surface of magnesium powders, nano size nickel particles were added as a catalyst. Most of the experiments in this study were done using this material, i.e., the 44 magnesium powder coated with Ni using the Theta Composer. Figure 4 15 d isplays the particle size distribution after Ni coating with the Theta Composer. The average particle size was 43 A comparison of Figure 4 15 with Figure 4 1 indicates that Ni coating using the Theta composer does not modify the size of the magnesium pa rticles. Figure 4 16 reveals the particle geometry of the Ni coated 44 magnesium powder, which looks very similar to Figure 4 2, indicating that no shape change took place after Ni coating with the Theta Composer. Figure 4 17 shows the E nergy dispersive X ray spectroscopy (EDS) mapping of the Ni coated 44 magnesium powder which demonstrates that the nano size Ni particles are distributed uniformly over the surface of the magnesium powders. Figure 4 18 shows the XRD pattern of Ni coated 44 magnesium p owder which indicates the presences of Ni and magnesium 4. 3 2 Ni Coating Using an Organometallic Solution T he amount of magnesium flakes produced using HSOBM was very small and each milling resulted in only 0. 1g per batch. To obtain optimum milling condi tion and efficiency in the Theta Composer at least 10g powder of magnesium and nano Ni mixture is needed for each run For consistency of coating, an o rganometallic Ni coating for a small amount of magnesium flakes wa s employed Approximately 1 gram batch es

PAGE 68

68 of the as received 300 and the HSOBM treated magnesium flakes were coated under similar conditions. Figure 4 19 displays a low magnification an SEM picture of the Ni coated flake powder. In Figure 4 20, EDS mappings indicate a homogeneous distribution of nano size Ni particles on the magnesium flake powder surface. The size and the shape of the powders were found not to be affected by the organometallic Ni coating process. Figure 4 21 shows the organometallic Ni coated as received 300 magnesium parti cles which indicate no shape change after the coating process.

PAGE 69

69 Figure 4 1. The particle size distribution of the as received 44 magnesium powder as evaluated by laser diffraction technique Figure 4 2 The SEM image of the as magnesium powder.

PAGE 70

70 Figure 4 3 Cross sectional optical image of the powders in the as magnesium powder. Figure 4 4 The schematic illustration of polishing surfaces and calculation of aspect ratio for as magnesium powders.

PAGE 71

71 Figure 4 5 Particle size distribution s of the as magnesium powder by statistic measurement in (a) number percent and (b) volume percent calculation. Figure 4 6 SEM images of the as magnesium powder.

PAGE 72

72 Figure 4 7 Cross sectional op tical image of the as magnesium powder. Figure 4 8 Particle size distribution in the as magnesium powder based on (a) number percent and (b) volume percent calculation.

PAGE 73

73 Figure 4 9. M agnesium flake thickness distribution after 100V and 20 seconds milling Figure 4 1 0. M agnesium flake thickness distribution after 100V and 1 minute milling Figure 4 1 1. M agnesium flake thickness distribution after 100V and 2 minutes milling

PAGE 74

74 Figure 4 1 2. M agnesium flake thickness di stribution after 110V and 2 minutes milling Figure 4 1 3 Optical microscopic analysis at cross section of magnesium flake powders Figure 4 1 4. SEM images of as milled magnesium flake powders.

PAGE 75

75 Figure 4 1 5. Particle size distribution of Ni coated 44 micron magnesium particles. Figure 4 1 6. SEM images of Ni coated the 44 micron magnesium powder after Ni coating with the Theta Composer. Figure 4 1 7 The EDS Ni mapping of Ni coated 44 micron magnesium particles.

PAGE 76

76 Figure 4 1 8. The X ray diffraction pa tterns. Figure 4 19. SEM images of Ni coated magnesium flake particles. Figure 4 2 0. The EDS Ni mapping of Ni coated magnesium flake particles.

PAGE 77

77 Figure 4 2 1. SEM image of Ni coated 300 micron magnesium particles

PAGE 78

78 CHAPTER 5 EVOLUTION OF MICROST RUCTURE DURING THE HYDROGENA TION PROCESS 5 1 Introduct ory Remarks Previous studies have shown that the hydrogen capacity of magnesium powders depends on many parameters including temperature, pressure, processing such as ball milling, particle size and shape. In order to develop a mechanistic understanding of how these factors affect the hydrogen capacity, a microstructurally based investigation of the hydrogenation process is needed. The objective of this study was to understand the development of magnesium hydr ide as a function of time in order to elucidate the mechanism that controls the hydrogen capacity for magnesium powders. Figure 5 1 shows a typical hydrogenation curve under the conditions where nergy for the reaction is a negative value. As can be seen in Figure 5 1, initially the reaction rate for the powder tested is relatively fast, and it becomes slow er as the time increases. Eventually the reaction rate is reduced to such a low value that ps eudo saturation is reached. T he amount of hydrogen content at this pseudo saturation level is defined as the hydrogen capacity It can be noticed that the hydrogen capacity in this curve is below the theoretical value 7.6 H wt. %. As explained in Chapter 2 several models have been suggested to explain the shape of the hydrogenation curve and its dependency on parameters such as T, P, etc. The two significant models are the Core Shell model and the impingement model. The Core Shell model assumes that after the initial hydride layer is formed it grows thicker and thicker as hydrogenation time increases. As the hydride layer gets thicker, the flux of hydrogen through it becomes smaller, which results in a lowering of the reaction rate. The driving force for di ffusion of hydrogen is

PAGE 79

79 the concentration gradient of hydrogen in hydride, as shown in Figure 5 2. However, the diffusivity of hydrogen through magnesium hydride is very slow. The diffusion coefficient for hydrogen diffusing through magnesium hydride is D = 1.507 x 10 4 (m 2 s 1 ) [exp( 96,000 (J/mole)/RT] [91] In this model the high hydrogen flux r ate needed to achieve the observed hydrogenation rates is achieved by assuming that the concentration of hydrogen, rather than its activity, at the interf ace between magnesium phase and hydrogen gas phase as well as at the interface between magnesium and magnesium hydride phase is equal. This assumption is invalid because the magnesium hydride phase is an ionic compound [41] and has a negligible solubility range for hydrogen [38] .The less referenced model is based on the nucleation and growth of the magnesium hydride phase and the pseudo satura tion level during the hydrogenation process is explained by the impingement of hydride colonies/grains. According to this model hydrogen atoms diffuse through the magnesium phase rather than the hydride phase. The reaction rate decreases with time because the area of the gas/Mg phase interface shrinks as the hydride phase grows. Consequently, the flux of hydrogen decreases with time resulting in a decreased reaction rate. In this study the material used for hydrogenation was a 44 m commercial magnesium powder, which was coated with Ni nanoparticles using the Theta Composer. The hydrogenation tests were conducted for different times at a constant temperature (210 C) and a constant pressure (1Mpa). The evolution of the hydride phase was investigated to shed light on the mechanism of the hydrogenation process. 5 .2 Microstructural Analysis Magnesium powder samples were hydrogenated at 210C under 1MPa hydrogen pressure for 2.5, 7.5, 15, 75, 300 minutes. The microstructure of each hydro genated

PAGE 80

80 sample was evaluated and analyzed for the hydride evolution as a function of time Figure 5 3 presents the cur ves obtained by hydrogenation for different times overlapped on the curve obtained after 5 hours of hydrogenation There is an excellent r epeatability of the hydrogenation rate and the curves match each other well. Figure 5 4 shows the cross sectional images for the powder hydrogenated for 2.5 minutes. The hydride phase is imaged as the darker phase in the BSE mode owing to its lower aver age atomic number (Z contrast mechanism) and the ionic structure and metallic ma gnesium phase appear bright Invariably, it was found that the hydride phase nucleates and grows on the surface of the magnesium particles Usually it is assumed that the mag nesium hydride nucleates on the surface of the particles because of the highest activity of hydrogen atoms on the surface and the fact that heterogeneous nucleation is associated with a lower energy barrier than the internal homogeneous nucleation. Of cour se, in the presence of second phase particles with very high interfacial energy the internal nucleation of the magnesium hydride is perceivable. Little work has been done to microstructurally evaluate the nucleation of the magnesium hydride Antisari et al reported that the nucleation site of magnesium hydride is heterogeneously defect preferred and the magnesium hydride nuclei were found at the interface of the catalyst Ni particles and magnesium phase [92] Montone et a l demonstrated that, i n particular, magnesium hydride nucleation in magnesium is assisted by crystal defects, while free surface is the on ly nucleation site [93] Oguchi et al. suggested t hat the nucleation of magnesium hydri de is a fast heterogeneous formation of hydride followed by a drastically slow growth of the magnesium hydride

PAGE 81

81 phase [94] The growth rate was initially 1.3nm/s and the average growth rate was 0.03nm/s. As seen in Figure 5 4, the nucleation sites of the hydride units are well separated and do not touch or impinge on each ot her. The hydride units appear faceted and the interfaces between the hydride units and the magnesium phase seem to be relatively flat which suggest s that these units grow as single crystals into the magnesium matrix. Owing to the lack of further evidence, here each unit will be referred to as hydride colonie s (meaning made of several grains)/grain (meaning single crystal). The m etallic magnesium phase is still a large portion of the particles and large areas of the magnesium particles have not yet transfor med to magnesium hydride. Assuming that the hydride forms as single crystals, t he number of nodules of hydride represents the nuclei number of magnesium hydride and can be calculated from microstructural analysis. The total number of nodules was measured f rom at least 80 particles and the linear nucleation density was obtain ed by the number of nodules divided by the perimeter of the powder. Figure 5 5 shows the SEM (BSE mode) cross sectional images of particles from the powder hydrogenated for 7.5 minutes. The hydride grains/colonies are larger than the particles hydrogenated for 2.5 minutes but no drastic increase in the number of them was detecte d. Some hydrides seem to grow preferentially into the metallic magnesium phase rather than grow laterally resu lting in a hemisphere shape, a s indicated by the arr ows in Figures 5 5 (a, b, d). I n Figure 5 5(c), a hydride grain/colony can be observed inside the partially hydrogenated magnesium particle which conflicts with the assumption that the hydride nucleates e xclusively at the powder surface only. A

PAGE 82

82 serial sectioning method using an optical microscope was employed to evaluate whether this patch of hydride was observed due to the three dimensional nature of the hydride units or indeed it was associated with an i nternal nucleation event. Figure 5 6 displays the optical microscopic images after several levels of polishing. In Figure 5 6(a), it is noticed that there is an isolated hydride grains/colonies located inside of the hydrogenated particle as marked by an ar row. To investigate whether the hydride touch es the surface or not further polishing was done As evidenced in Figure 5 6 (b), interestingly the hydride patch touched the powder surface. Figure s 5 6(c, d) display s the images of further subsequent polishing. All of the images point out that the hydride patch found inside the particle in Figure 5 5(c) is due to the polishing issue and not due to the inside nucleation of magnesium hydride. As can be seen in Figure 5 5, some hydride grains/colonies have begun to impinge on each other but still a large fraction of the surface of the magnesium particles is free of hydrides and hydrogen atoms can d iffuse fast through these areas The SEM (BSE mode) images of magnesium particles hydrogenated for 15 minutes are shown in Figure 5 7. In comparison t o the powder hydrogenated for 7.5 minutes, the hydride colonies/grains have grown in size, and consequently a larger number of them have impinged on each other but the surface is still not completely covered by hydride. There is a large volume expansion (about 30% increase) associated with the transformation of magnesium to magnesium hydride. As a result, as evidenced in Figure 5 7, the hydride colonies/grains, which are brittle, crack under the volume expansion stresses. Thes e stresses can play an important role in nucleation and growth of the magnesium hydride, especially in thin films [95, 96] Also, it can be

PAGE 83

83 notice d that some colonies/grains bul ge out of the original magnesium part icle surface to accommodate the volume change. Figure 5 8 presents the microstructure of magnesium particles hydrogenated for 75 minutes. Here, the most important observation is the extent of the impingement behavior. In this powder, some of the particles have been covered completely by magnesium hydrides, and some of them still have metallic magnesium exposed on the surface. This observation shows that the complete impingement of the hydride colonies/particles does not take place in all particles simultan eously suggesting that there will be a transition period where the mechanism of the hydride growth transitions from being controlled by the diffusion of hydrogen through the metallic magnesium to a through hydride diffusion mechanism ( Core S hell model) For powders hydrogenated for 300 minutes, the transformation come s to an end and almost all of the magnesium particles are covered by a magnesium hydride layer, as shown In Figure 5 9. At this point only a small fraction of the metallic magnesium is left in the center of the particles and the growth rate of the hydride is exclusively controlled by the diffusion of hydrogen atoms through the hydride layer. Since t he diffusivity of hydrogen through the ionic magnesium hydride is very slow [91] and at 210C and 1MPa the thickness of the hydride layer at the point of impingement is relatively large, the hydride growth rate after impingement is very slow. Therefore, a pseudo saturation of hydrogen absorption, referred to as hydrogen capacity, is achie ved I t should be noted that under different hydrogenation conditions, for example low hydrogenation temperature, this pseudo saturation is not achieved. When the

PAGE 84

84 thickness of the hydride at the impingement point is small, the flux of hydrogen through a t hin hydride layer can still be significant. 5 .3 Effect of Particle Size on Hydrogen Capacity In Chapter 4, the particle size distribution of magnesium powders shows a relatively broad range in diameter. As demonstrated in Figure 5 10 the amount of metalli c magnesium left in the center of the magnesium particles after hydrogenation at 210 C for 300 minutes depends on the size of the particle. As will be shown in the forthcoming chapters, the thickness of the hydride at the point of impingement depends only on temperature and pressure of hydrogenation for a given coated magnesium powder. Therefore, as illustrated in the schematic in Figure 5 10, the larger particles contain a larger fraction of unhydrogenated magnesium hence resulting in a lower hydrogen cap acity per particle. To study the particle size distribution effect on the hydrogen capacity, the hydrogenated particles were divided into 3 groups according to their sizes, which range from 0 to 15, 15 to 30 and 30 to 60 The size range for the three gro ups is chosen arbitrarily based on the observation at low magnification (1000X) of the SEM (BSE mode) images. The method of the calculation of the phase fraction is shown in Figure 5 11. The SEM (BSE mode) image was loaded into the Microsoft PowerPoint and grids were overwritten on the images for the phase fraction measurement. The volume fraction of the hydride phase was calculated based on the point count method [97] As mentioned previously, the darker phase region is magnesium hydride and the brighter area was composed of the magnesium phase. In Figure 5 11, if the intersection located just right on the darker area, one intersection is coun ted for the magnesium hydride. The same treatment was used for the magnesium phase. For the intersection on the

PAGE 85

85 interface of Mg/MgH 2 it was divided and counted as for magnesium hydride and for magnesium phase, respectively. The total number of interse ctions was summed for each phase. The area phase fraction can be obtained by dividing the number of intersections for each phase over total number of intersections. The area phase fraction is usually assumed to be a good measure of the volume fraction. Fi gure 5 12 displays the area phase fraction of magnesium hydride for different size of particles. The hydrogen capacity varies with different sizes of magnesium powders. It is obvious that with an increase of the particle size the hydride area fraction decr eases. The hydride fraction is 0.94 for the size of magnesium powders smaller than 15 m. It is expected that the hydrogen capacity will be close to one if all the sizes of magnesium powders are below 15 m. Figure 5 13 shows examples of microstructures corr esponding to different sizes illustrating the effect of particle size distribution on the measured average hydrogen capacity. It is interesting to note that the thickness of the hydride layer is approximately the same in these particles, confirming the sug gestion that the hydride thickness at the point of impingement is controlled by the nucleation and growth rates of the hydride, hence only dependent on temperature and pressure of hydrogenation for a given coated powder. Some small size particles in the 0 15 m span reach even 100% complete phase transformation because the hydride impingement thickness is larger or comparable to the half diameter of the particles. The medium 15 30 m magnesium particles show larger areas of metallic magnesium phase left in th e particles compared to the 0 15 m particles. The large 30 60 m magnesium particles exhibit a much larger metallic magnesium area than the medium ones. The capacities of these three different sizes of hydrogenated particles contribute to the total hydrogen

PAGE 86

86 capacity for the hydrogenation curve in Figure 5 1. Figure 5 14 displays the hydrogen capacity for different sizes of magnesium particles at different hydrogenation time periods. The data points of hydrogen capacity follow the same trend as the hydrogenat ion curve. Furthermore, the hydrogenation curve is located within the capacity ranges for different size of the magnesium particles which means the area phase fraction successfully estimates the hydrogen capacity. 5 .4 Summary of Hydrogenation Process The m agnesium hydride formation has been treated as a simple layer growth which can be explained by the Core Shell model. However, based on the microstructural analysis of the hydrogenated powders, it seems that the nucleation of magnesium hydride is one of the kinetics barriers during magnesium hydride formation; individual grains/colonies of magnesium hydride have been seen in the initial stage of hydrogenation. A schematic illustration was developed to describe the hydrogenation process; F igure s 5 15 to 5 21 show the schematic diagram s at each step from the beginning to the saturation to understand step by step the magnesium hydride formation Step 1: Physicsorption of hydrogen molecules. Step 2: The dissociation of hydrogen molecules into two hydrogen atoms. Step 3: Dissolution of hydrogen atoms (Solubility of H atom in magnesium). Step 4: Chemisorption of hydrogen ions (Magnesium hydride formation). Step 5: Nucleation of magnesium hydride. Step 6: Growth of magnesium hydride. Step 7: Impingement of magnes ium hydride. Step 8: Continuous thickening of magnesium hydride layer. Step 1 for the hydride formation, the solid g as reaction is the physicsorption of hydrogen gas molecules attached on the solid material magnesium powder surface by van der Waals force Unde r the standard condition, there is an activation energy barrier

PAGE 87

87 for the hydrogen molecule to break into two hydrogen atoms. In order to eliminate this process to act as a rate limiting step during hydrogenation, nano Ni particles have been used to en hance the dissociation kinetics of hydrogen molecules. The image of EDS Ni mapping in Figure 3 5 exhibits a very homogeneous distribution of nano Ni particles on the magnesium particle. Therefor e, the dissociation of hydrogen molecule s should not be an iss ue while considering the rate limiting steps After the hydrogen molecules turn into hydrogen atoms at Step 2, the hydrogen atoms start to dissolve into the magnesium phase until reaching the solubility limit in the magnesium phase at a given temperature w hich is Step 3. At Step 4, magnesium hydride forms whereas the hydrogen concentration is oversaturated and the crystal structure transforms from hcp Mg to tetragonal magnesium hydride with about 20% volume expansions. Ideally the increase of the volume due to the phase transformation would result in the defect structure, i.e. dislocation, cracks which induce the faster kinetics. In Step 5, if the number of nuclei is higher than the critical number of nuclei, the magnesium hydride nucleates somewhere on the powder surface. Step 6 is the growth process of magnesium hydride. Step 7 is when those individual hydride grains/colonies impinge on each other and form a complete hydride layer. The last step, Step 8, is the stage that the powder surface is completely co vered by the magnesium hydride. It looks like a core shell in the cross section view. After this stage, the core shell model can be used to explain the very slow thickening process of magnesium hydride. Before this step, the nucleation and growth model has gained most agreement in describing magnesium hydride formation. 5 .5 Summary and C onclusions The hydrogenation study was performed using the hydrogenation part of a house built Hydrogenation/Dehydrogenation unit. The hydrogenation behavior was studied

PAGE 88

88 usi ng the microstructural analysis at different time periods during the hydrogenation process. The major conclusions are the following: 1. Prior to impingement of magnesium hydrides, a nucleation and growth model can best explain the mechanism of magnesium hydri de formation at hydrogenation temperature 210 C under 1MPa. However, the Core Shell model should be applied after the magnesium hydride impinges on each other (at pseudo saturation ). 2. The pseudo saturation is the hydrogen capacity at a given hydrogenation c ondition and time (6.1 H wt. % at 210 C, 1MPa for 5 hours) 3. With the decrease of the size of magnesium particles, the magnesium hydride fraction increases. 4. With the increase of the hydrogenation time, the hydride fraction increases. 5. The measurement of area phase fraction for each phase on 2 D is able to estimate the hydrogen capacity and hydrogenation curve.

PAGE 89

89 Figure 5 1. Hydrogenation curve of magnesium powder sample at 210 C under 1MPa hydrogen pressure. Figure 5 2. Schematic illustration of Core S hell model.

PAGE 90

90 Figure 5 3 Hydrogenation curves of magnesium powder sample at 210C under 1MPa hydrogen pressure at each interrupted times.

PAGE 91

91 Figure 5 4. Microstructural analysis on the cross section of magnesium powder hydrogenated for 2.5 minutes.

PAGE 92

92 Figu re 5 5. Microstructural analysis on the cross section of magnesium powder hydrogenated for 7.5 minutes.

PAGE 93

93 Figure 5 6. Optical microscopic images for the investigation of the location of nucleation of magnesium hydride.

PAGE 94

94 Figure 5 7. Microstructural analysi s on the cross section of powder hydrogenated for 15 minutes.

PAGE 95

95 Figure 5 8. Microstructural analysis on the cross section of powder hydrogenated for 75 minutes.

PAGE 96

96 Figure 5 9. Microstructural analysis on the cross section of powder hydrogenated for 300 minu tes.

PAGE 97

97 Figure 5 10. An example of different hydride fractions achieved by different sizes of particles.

PAGE 98

98 Figure 5 11 An example of the measurement of phase fraction of magnesium hydride Figure 5 12. Hydride fraction at different range of sizes and temp eratures.

PAGE 99

99 Figure 5 13. An illustration of hydride distribution and phase fraction for different size of particles at the same hydrogenation condition and time 15 minutes. Figure 5 14. A comparison of hydrogenation curve with the data point estimated b y the measurement of area phase fraction.

PAGE 100

100 Figure 5 15. Step 1: Physicsorption of hydrogen molecules. Figure 5 16. Step 2: The dissociation of hydrogen molecules.

PAGE 101

101 Figure 5 17. Step 3:The dissolving of hydrogen atoms. Figure 5 18. Step 4: Chmisorptio n of hydrogen ions.

PAGE 102

102 Figure 5 19. Step 5: Nucleation of magnesium hydrides. Figure 5 20. Step 6: Growth of magnesium hydrides.

PAGE 103

103 Figure 5 21. Step 7: The impingement of magnesium hydrides.

PAGE 104

104 C HAPTER 6 KINETICS OF MG HYDRI DE FORM ATION 6 1 Introduct ory Remarks In Chapter 5, the microstructural evolution at different time periods of hydrogenation at a given temperature and pressure was presented and discussed. The results suggested that the hydrogen capacity is controlled by the imp ingement behavior of magnesium hydride grains/colonies and the magnesium hydride formation occurs by following up the nucleation and growth model. One of the limitation s for practical application of magnesium hydride in hydrogen storage is its slow hydroge nation kinetics In this chapter, to understand the kinetics of hydrogenation at different temperatures, the nucleation and growth processes are treated by the JMA (Johnson Mehl Avrami) model [69] Magnesium powde rs were hydrogenated at different temperatures under the isothermal and isobaric hydrogenation condition s. The hydrogenation curves were measured and e mployed to identify the kinetic properties of the magnesium hydride formation. Different stages of hydrid e formation were identified based on the JMA analysis. In addition, the number of the hydride colonies/grains nuclei was measured at different temperatures in order to better understand the impingement phenomenon. The results of JMA analyses were employ ed to acquire the activation energy and identify the possible mechanism of hydride growth via constructing the 6 2 Analysis of Kinetics of Hydride Formation by the Johnson Mehl Avrami R elation ship Based on the microstructural analysis, the clearly distinguished individual grains/colonies shown in the cross section of hydrogenated magnesium powder strongly suggest that the formation behavior of magnesium hydride is best explained by a

PAGE 105

105 nucleation and growth mod el The details of the John s on Mehl Avrami (JMA) model were discussed in Chapter 2. Using this model, the relation between the hydride fraction and the hydrogenation time at a given pressure can be expressed as : f = 1 exp( Kt n ) (6 1) f: Volume fraction of magnesium hydride. K: React ion constant = K 0 exp ( Q/RT), (Q: Activation energy, R: Ideal gas constant, and T: Isothermal temperature). t: Hydrogenation time n: Reaction order. n is supposedly independent of the hydrogenation temperature only if there is no change in the nucleation mechanism. The reaction constant, K on the other hand depends on the nucleation rate growth rate and the geometry factor which are very sensitive to the hydrogenation temperature [29, 39, 69] Equation (6 1) can be rewritten as: (1 f) = exp ( Kt n ) (6 2a) ln (1 f) = Kt n (6 2b) ln{ln[1/(1 f)]} =lnK + nln(t) (6 2c) A plot of ln{ln[1/(1 f)]} vs lnt curve is shown in Figure 6 1, which is the JMA analyzed curve of Figure 5 1. The curve can be clearly divided into thr ee linear stages including a transition region between Stage 1 and Stage 3, which is contributed by different hydrogenation behaviors of the magnesium powders as discussed in Chapter 5 The time range for each stage was defined based on the best fit analys is The range was selected when the R 2 for each stage was equal or larger than 0.995 (R 2 is defined as

PAGE 106

106 the statistical measure of how well a regression line approximates real data points; an R 2 of 1 indicates a perfect fit). Based on the discussions presen ted in Chapter 5, Stage 1 is probably associated with the dissociation of hydrogen molecules and/or the nucleation of magnesium hydride. In my opinion, the energy barrier of the dissociation of hydrogen molecules is most probably reduced or eliminated in t his case with the catalyst added to the magnesium powder surface. S tage 2 is suggested to be associated with the growth of magnesium hydride assuming that the nucleation rate of magnesium hydride does not depend on time and remains constant. If the nucleat ion rate of magnesium hydride is zero, which means all the nuclei have been consumed after Stage 1, the number of magnesium hydride nuclei will be a constant value through the whole test. In this case, the controlling step of S tage 2 is the growth of magne sium hydride. T he transition region from the beginning to the end of the impingement process resides b etween Stage s 2 and 3 At the end of the transition region, the impingement is complete and further growth of the hydride takes place by the diffusion of hydrogen through the hydride layer. As discussed previously this process is very slow and therefore Stage 3 is associated with a very slow kinetics of hydride formation, which is a continuous thickening process of the hydride layer. 6 2 1 Evaluation of N uc leation Number during Stage 2 A hydrogenation test was designed to disclose the nucleation behavior in the very initial hydrogenation stage. The nuclei numbers are compared at different hydrogenation times to study whether there is a continuous nucleation of the hydride or the number of the nuclei remains constant during Stage 2. The hydrogenation times of 2. 5 and 7. 5 minutes were selected for this analysis. The microstructural analysis of these samples was conducted in detail as was shown in Chapter 5.

PAGE 107

107 Th e hydrogenation curves for 2. 5 and 7. 5 minutes, shown in Figure 5 1 exhibit a n overlapping reproducibility in the hydrogenation curve The p owder s hydrogenated for 2. 5 and 7. 5 minutes were imbedded in mounting and checked in the cross section under SEM (B SE mode). As discussed in Chapter 5 the nucleation of magnesium hydride takes place exclusively on the powder surface Therefore, linear density is sufficient for the evaluation of the number density; otherwise area density should be utilized for the nucl eation study. In this analysis it is assumed that the linear density indirectly represents the surface density and they will both show the same tendency in the calculation of nucleation number Figure 6 2 shows an example of using the ProgRes CapturePro 2. 7 software to measure the perimeter of path surrounding the area for the powder sample hydrogenated for 2.5 minutes. Since the magnification by SEM (BSE mode) is different from that by optical microscope, there is a ratio which must be applied to the calcu lation to obtain the real perimeter of the hydrogenated powder. The same magnification is used in SEM (BSE mode) for the convenience of the perimeter measurement by the software. More than 80 particles were included in this measurement The nucleation line ar density can be obtained by the following equation: D L =N L /L (6 3) D L : Linear density of nucl ei N L : Number of nucl ei on the powder surface. L: The total length of powder surface. The linear density D L calculated from equation (6 3) is 0. 2316 (#/ m) The same method was applied to the powder sample hydrogenated for 7.5 minutes. An example is shown in Figure 6 3 and about 100 particles were taken into account for this study The

PAGE 108

108 D L obtained for 7.5 minutes is 0.2347 (#/ m) which is very clo se to that for the 2.5 minutes hydrogenated powder Table 6.1 shows the result s of the mea surement of number of nuclei at different hydrogenation times. The inversed values of the linear density represent the distance between two magnesium h ydride nodules. According to these measurements, showing up and the total numbers of the nodules on a single m particle, assuming a spherical shape, would be 32 (#/particle). Based on these results, we concluded that the nuclei number is a constant value during Stage 2 and only the growth of the hydride colonies/grains that form in Stage 1 take place during Stage 2. 6.2.2 Correlation of Grain Boundary and Nuclei of Magnesium Hydride As shown in Chapter 5, the number of hydride nuclei at 210C was constant during Stage 2. One possible nucleation site for hydride is the intersection of magnesium grain boundaries with the powder surface. In order to correlate the number of intersection sites with the number of hydride nuclei, a magnesium powder s ample was heated from room temperature to 210 C under very low hydrogen pressure (~3Pa) without introducing the hydrogen gas into the specimen chamber. The powder sample was prepared for metallography using an optical microscope. Acetic glycol was used as the etching agent to reveal magnesium grain boundaries. Figure 6 4 shows the optical microscopic image of the pre heating magnesium powder sample. The grain size of magnesium is approximately 20 m. The study in our lab for the magnesium grain size measurement under TEM demonstrates that the grain size of the as received magnesium powder is approximately 2 m [98] This result indicates that the grain size has increased during the pre heating process of the magnesium powd er sample. Generally, the heterogeneous nucleation is easier than the homogeneous nucleation.

PAGE 109

109 The intersection at the powder surface becomes the nucleation site owing to its low surface energy. The linear density of intersections of the grain boundaries wi th the powder surface per unit length is measured to be 0.08(#/ m), which means about every 12.5 m an intersection at the powder surface is located. As reported in Table 6 1, the actual linear density of the hydride nuclei is about 4.3 m, which is smaller than the linear intersection density. This difference may be attributed to the lack of enough statistics as well as the fact that linear density can be different than the actual areal density. 6.2.3 Effect of Hydrogenation Temperature The objective of th is chapter is to study the hydrogenation mechanism during magnesium hydride formation. A systematic approach including reproducibility has been considered to evaluate the hydrogenation curves of magnesium powders. The majority of previously reported hydrog enation curves for magnesium have been evaluated at relatively high temperatures [31, 34, 63] In this study the kinetics of hydrogenation of magnesium C). Four hydrogenation temperatures 150C, 180C, 210C, and 250C were used for the investigation of the effect of hydrogenation temperature. Figure 6 5 shows the hydrogenation curves at these temperatures under constant hydrogen pressure of 1MPa Each test was co nducted twice in order to ensure repeatability of the tests These curves reveal that as the test temperature is decreased the hydrogen capacity is reduced. The initial hydrogenation rate, as identified by the slope of the curve in the early stages of hydr ogenation, seems to decrease with decreasing temperature up to 180 C and then becomes quite fast at 150 C. This behavior can be explained by the opposite effect of

PAGE 110

110 hydrogenation temperature on nucleation and growth of hydrides. As the hydrogenation tempera ture is reduced the hydride growth rate is expected to decrease. On the other hand, because of the increase in the driving force for the formation of the hydride phase at low temperatures, the number of hydride nuclei is much higher at low temperatures. An other interesting observation is that as the test temperature is decreased, there is less tendency to reach pseudo saturation after 5 hours of hydrogenation. As the hydrogenation temperature is decreased, owing to the large number of hydride nucleation sit es, the thickness at which impingement is achieved is reduced. Since the flux of hydrogen through a thinner hydride layer is larger, the hydrogenation rate after impingement will be faster at a lower hydrogenation temperature. The hydrogen capacities at ea ch temperature are shown i n Table 6 2 Figure 6 6 presents the relationship between the hydrogen capacities and the hydrogenation temperatures. The trend reveals that the hydrogen capacity increases with increasing temperature. It is noticed that the hydro gen capacities vary from about 3.2 to 6.9 H wt.% in this temperature span and never reach the theoretical value 7.6 H wt. % Powder s hydrogenated at 250 C show the closest values, 6.5 and 6.9 H wt. %, to the theoretical values. However, the curves exhibit no further hydrogen absorption after approximately 150 minutes during the hydrogenation tests. Powders hydrogenated at 210C show slightly lower hydrogen capacities than those at 250 C. The hydrogen capacities are 6 .1 and 6.3 H wt. %. One of the hydrogenat ion tests was interrupted at 282 minutes due to an equipment issue but luckily the curve had already reached saturation stage. As the test temperature is dropped to 180 C, the hydrogenation capacities after 5 hours of

PAGE 111

111 hydrogenation reduce to 5.4 and 5.8 H wt. %. At 150 C hydrogenation temperature the capacity drops further to 3 .2 and 3.3 H wt. % in 300 minutes 6.2.4 Microstructural A nalysis of Powders Hydrogenated at Different Temperatures A microstructure analysis of the powder s hydrogenated at th e four temperatures of 150 C, 180 C, 210 C, and 250 C after 300 minutes of hydrogenation was performed It is observed that the hydride layer is com posed of the impinged hydride grains/colonies for all temperatures, as shown in Figure 6 7. In Figure 6 7(a) the p owder sample hydrogenated at 150C shows a continuous magnesium hydride layer with a very small thickness about 2 This observation confirms that the impingement of the hydride grains/colonies takes place relatively fast and the growth of the hydride lay er takes place by the diffusion of hydrogen through the hydride layer (Core Shell model). Figure 6 7(b) shows the SEM ( BSE mode) cross sectional image of powders hydrogenated at 180C for 5 hours. Larger hydride grains/colonies are identified here with a 4 to 8 hydride layer thickness. The microstructure of powders hydrogenated at 210C is shown in 6 7(c). Some of the cracks appeared inside the hydride phase and the hydride thickness varied from 4 15 m. When the powders were hydrogenated at 250C many particl es were completely hydrogenated and as shown in Figure 6 7(d) only patches of the metallic magnesium encapsulated by the hydride growth, were left. Figure 6 8 illustrates the proposed effect of hydrogenation temperature on the microstructure of hydrogena ted particles. The driving force for hydride nucleation is relatively low near the equilibrium temperature (Fig. 6 8 a) whereas the growth rate is maximized. Under this condition a large fraction of a magnesium particle is transformed to magnesium hydride before the impingement of the hydride colonies/crystals takes

PAGE 112

112 place. On the other hand, a low hydrogenation temperature (Fig. 6 8 b) is expected to render a fast nucleation rate, which causes the impingement of the hydride colonies to be achieved at a relat ively smaller hydride thickness, i.e. at a lower volume fraction of hydride. The driving force is the temperature difference between equilibrium temperature and experimental temperature (hydrogenation temperature) T = T e q T h From the literature, the equilibrium temperature for 1MPa hydrogen pressure is 373 C [38] At 150 C the nucleation rate is relatively high as compared to other temperatures hence the impingement takes place shortly while the surface has been covered by a large number of hydride colonies/grains. We can expect that the lower the hydroge nation temperature the higher the nucleation rate of magnesium hydride. With the temperature increasing the driving force decreases since the hydrogenation temperature becomes close to the equilibrium temperature. The nucleation rate which is driving for ce decreases with increasing temperature. It should be noted that the nucleation theory only applied up to the hydrogenation temperature below the equilibrium temperature because beyond or equal to the equilibrium temperature a reversed reaction and deso rption of hydrogen happens. 6.3 Effect of Nucleation Rate on Hydrogen C apacity 6.3.1 Hydrogenation Procedures In this study the nucleation rate was changed by modifying the hydrogenation procedure. As described in Chapter 3, two methods may be used to hyd rogenate a powder. In Method I the powder was first pressurized to 1MPa with hydrogen gas. Then, the specimen is heated to the desired temperature. In Method II, which has been used throughout this dissertation, the specimen chamber is heated to the desire d temperature under vacuum, and the chamber is pressurized with hydrogen gas to

PAGE 113

113 1MPa. Fig. 6 9 displays hydrogenation curves for the methods for the hydrogenation temperature of 210 C A comparison of the hydrogenation curves by the two methods applied in this study revealed that the hydrogen capacity is strongly dependent on the sequence of pressurization and heating. It is clear that the apparent saturation level, i.e. the hydrogen capacity, is much higher for Method II (6.1 H wt. %) than that for Method I (1.5 H wt. %). To elucidate the mechanism of magnesium hydride formation, we analyzed the hydrogenation curves using the Johnson Mehl Avrami (JMA) equation The ln[ln(1/1 f )] is plotted as a function of ln ( t ) in Fig. 6 10 which demonstrates that the magnesium hydride develops in three stages as discussed in Chapter 5 The reaction rate in S tage 1, which takes place in a very short time (<2 min) is probably controlled by nucleation of magnesium hydride It is believed that the hydride phase initially nucleates on the surface of the powders. The beginning of extensive growth of the hydride corresponds to the onset of S tage 2. As the hydride colonies/grains grow, the total effective cross section area for hydrogen diffusion into the magnesium phase decre ases. The end of S tage 2 corresponds to the beginning of impingement of the hydride colonies/grains, after which the growth of hydride is limited by the diffusion of hydrogen through the hydride and maximum capacity is reached for relatively high hydrogena tion temperatures As shown in Fig. 6 10 Method II exhibited a much longer growth period, i.e., S tage 2. Table 6 3 presents the n values representing the slope of the curves shown in Fig. 6 10 for the three stages of hydrogenation. The hydrogenation metho d affected the n values significantly, suggesting that the kinetics of hydrogenation is strongly dependent on the hydrogenation method.

PAGE 114

114 In order to confirm that the difference in the hydrogenation behavior of the powders, as a result of the change in the heating and pressurizing schedule, is associated with the nucleation rate of hydride, a powder sample was hydrogenated by Method II to a hydride fraction equivalent to the saturation level of the powder hydrogenated by Method I. Fig. 6 11 compares the hydr ogenation curve by Method II for 3 min utes with that of Method I. It should be noted that the time axis for Method I starts when the heating begins and consequently a lower initial apparent kinetics is observed. As indicated on the curve, in the sampl e hyd rogenated by Method I most of the hydride w as formed prior to reaching 210 C. Fig. 6 12 shows SEM images of the hydrogenated powders with an equal amount of hydride. From the cross section and surface images, it can be seen that the samples hydrogenated t o the same level have very different distribution s of hydride colonies/grains. For the powder hydrogenated by Method I, the hydride colonies/grains are relatively small (Fig. 6 12 a) and cover the entire surface of the particle (Fig. 6 12 b). In contrast, Method II resulted in few nucleation sites for magnesium hydride as seen in the cross section micrograph (Fig. 6 12 c) and the observation of the metallic magnesium on the surface (Fig. 6 12 d). Also, the size of the hydride colonies/grains formed by Metho d II is somewhat larger compared to those observed in the powder hydrogenated by Method I. To quantify the nucleation behavior, we evaluated the apparent surface coverage fraction by magnesium hydride in both powders. As shown in Figure 6 13, the fraction of coverage was estimated from the SEM pictures of over 50 particles fo r each powder sample by: Surface Coverage = H i /( H i + M i ) (6 4)

PAGE 115

115 where H i is the length of the particle perimeter covered by the hydride, M i is the length of the particle perimeter remai ning as metallic magnesium and i denotes the particle number. Fig. 6 14 presents the fraction of particles as a function of their surface coverage. As it can be seen in this graph, Method I resulted in a large fraction of particles (0.55) with more than 9 0% coverage. On the other hand, the powders hydrogenated by Method II exhibited some powders with very low surface coverage (low number of nucleation sites) and a much smaller f r action of particles (0.26) with more than 90% coverage. Fig. 6 15 shows exampl es of high and low surface coverage in powders hydrogenated by Methods I and II, respectively. Considering the shape of the particles the cross section shape varied among the powders from circular (Fig. 6 15 a and c ) to elliptical (Fig. 6 15 b and d) sh ape as discussed in Chapter 4 In summary, b y varying the hydrogenation procedure at a given temperature, we have demonstrated that the hydrogen capacity of magnesium powder is inversely related to the nucleation rate of the hydride phase. Here, it is assu med that the magnesium hydride preferentially nucleates on the surface of magnesium particles. As the surface coverage with magnesium hydride increases with time, less area for hydrogen entry will be available, hence due to a decrease in the growth rate, the hydrogen absorption rate reduces. When the hydride colonies/crystals impinge on each other, the diffusion of hydrogen will be limited through the hydride phase. Since the diffusivity of hydrogen through the hydride phase is at least three orders of mag nitude smaller than through magnesium [91] the hydrogenation rate will be decreased significantly and a pseudo saturation is reached. This pseudo saturation level is considered as the hydrogen capacity, which is defined as the hydrogen content af ter a

PAGE 116

116 long hydrogenation period at a given temperature and pressure. The effect of temperature on hydrogen capacity can be explained by the dependency of the hydride nucleation rate, which increases with undercooling, T = T e q T h (T e is the equilibrium temperature for Mg + H 2 MgH 2 at a given pressure and T h is the hydrogenation temperature). The much smaller n value in S tage 1 for Method I in comparison to Method II can be attributed to the lower kinetics of hydroge n dissociation, which is a thermally activated process. Considering that during hydrogenation by Method I the temperature, hence the hydride growth rate, is lower, the larger slope of S tage 2 (Table 6.3 ) indicates that the hydride phase nucleated faster th an during hydrogenation by Method II. The smaller size of hydride colonies/grains and the high surface coverage of particles elucidated by microstructural analysis confirmed the rapid nucleation of the magnesium hydride under Method I. The driving force fo r nucleation is proportional to the deviation from the equilibr ium condition. At pressure of 1 MPa, the equilibrium temperature for magnesium hydride formation is 3 73 C [38] When the powder was heated after pressurization with hydrogen (Method I), the hydride formation began shortly after the heater was turned on. The large deviation from the equilib rium temperature provided a high driving force resulting in a fast nucleation rate. Consequently, the powder surface was covered with MgH 2 in a short time during hydrogenation by this method. After this point, further hydrogen absorption takes place by dif fusion of hydrogen through the magnesium hydride layer. From the literature, the diffusion coefficient of hydrogen in magnesium hydride can be expressed as D = 1.507 x 10 4 (m 2 s 1 ) [exp( 96k J/mole)/RT] [91] At the hydrogenation temperature of 21 0 C, the approximate diffusion length is

PAGE 117

117 quite small (0.35 m min 1 ) resulting in a very slow hydrogen uptake. On the other hand, in Method II, hydride nucleated and grew at 210 C, which is consistent with the lower hydride nucleation rate observed based on the microstructural investigations. The slope of S tage 3 was slightly higher for Method II, suggesting that not all particles had been completely covered by a hydride layer. The results of this study reveal that for a given hydrogenation method the powders do not display similar level of hydride formation (Fig. 6 14 ). This phenomenon is probably associated with the differences in the surface condition of the powders as well as any variation in the Ni coating among particles. The hydrogen coverage distribution curves shown in Fig. 6 14 suggest that the saturation is achieved when approximately more than 80% of the powders have a surface coverage of 80% or more. 6.3.2 Heating Procedure The objective of this study was to confirm that the pseudo saturation observed in the hydrogenation curves at relatively high tempe ratures is associated with the impingement of the hydride grains/colonies rather than the thickening of a continuous hydride layer as in the Core Shell model. In this investigation the nucleation rate was modified by applying two different temperatures dur ing a single hydrogenation experiment. Initially, the powder sample was hydrogenated at a low temperature (15 0C) to establish a high number of hydride nuclei. Then, the hydrogenation temperature was increased (210 C) to allow for faster growth rate. In ad dition, samples were hydrogenated in one step at 150 C and 210 C for comparison. Figure 6 16 presents the hydrogenation curves up to 5 h for temperatures of 150 and 210 C as well as the two step hydrogenation (150 C + 210 C). As expected, the hydrogen capa city of magnesium powder hydrogenated at 210 C (6.1 wt. % H) was much higher than that

PAGE 118

118 achieved at 150 C (3.2 wt. % H). In two step hydrogenation, the first step was conducted at 150 C for 2 min and then the hydrogenation temperature was increased to 210 C Based on the Core Sh ell model, the two step hydrogenation procedure should have resulted in hydrogen absorption kinetics close to the 210 C curve. However, interestingly, the two step hydrogenated sample showed an even lower hydrogen capacity (2.4 wt. % H) than the sample hydrogenated at 150 C. This finding confirms that the Core S hell model is not representative of hydrogenation process in magnesium powders. In order to prove the hydride nucleation concept illustrated in Figure 6 16 the microstructures of powders hydrogenated to identical quantity of hydride fraction at different temperatures were investigated. A magnesium powder sample was hydrogenated at 210 C for 15 min (point A on Fig. 6 16 ) to produce an equal amount of hydride as in the sample hydr ogenated at 150 C for 5 h. Figure 6 17 presents SEM/BSE images of cross sectioned (a and b) and whole (c and d) particles. Figure s 6 17 a and b clearly demonstrate the effect of temperature on the nucleation behavior of magnesium hydride. Only a few nucleat ion sites were activated at 210 C and the magnesium hydride phase did not cover the surface of the particles (Fig. 6 17 a). In contrast, a continuous hydride layer with much smaller hydride colonies/crystals were found in the powder hydrogenated at 150 C. T he external images (Fig. 6 17 c and d) also confirm the presence of hydride free areas on the particle surface in the 210 C hydrogenated powder, while complete coverage was achieved when powders were hydrogenated at 150 C. Figure 6 18( a ) displays the cross sectional image of powders hydrogenated at 150 C for only 2 min. As shown in the high magnification picture

PAGE 119

119 presented in Figure 6 18 b, the nucleation of magnesium hydride was not limited to the surface and indeed many fairly small hydride particles were fo und near the surface. The results presented above confirm that the hydrogen capacity is controlled by the impingement of hydride colonies/crystals. The equilibrium temperat ure at 1MPa for Mg + H 2 MgH 2 T = 223 C ), and the hydride nucleation rate is quite large and as demonstrated in Figures 6 18 b and 6 18 d the hydrogen capacity is reached at a small hydride thickness. The lower driving force for nucleation and the higher diffusion rate at 210 C ( T = 1 63 C) rend er a much smaller nucleation rate and faster growth rate, which result in extensive growth of hydrides grains/colonies before their impingement (Fig. 6 17 a). Increasing the temperature to 210 C after establishing the nucleation sites by hydrogenating at 15 0 C for 2 min only expedited the growth of the hydride phase and their impingement. As shown in Figure 6 18 b, the hydride nucleation at 150 C was not limited to the surface. The lower capacity of the two step hydrogenated sample in comparison to the 150 C hydrogenated sample (Fig. 6 18d ) can be attributed to the incomplete growth of the inner hydride nuclei in the former sample. In other words, the slower growth at 150 C allowed more growth of the inner hydride particles before the occurrence of the impinge ment. This observation suggests that when there is a large driving force for nucleation, as is the case at 150 C, hydride nucleation can take place at other defects such as grain boundaries and dislocations. Note that most of the hydride was found near the surface where the highest concentration of hydrogen is expected to exist based on the diffusion of hydrogen from the surface inward. After further hydrogenation at 210 C for 298 min, no new nucleation of the hydride phase took place and as shown

PAGE 120

120 in Figure 6 18 c the magnesium hydride particles grew and completely covered the magnesium particle. Hydrogenation at 150 C for 5h (2 + 298 min), as demonstrated in Figure 6 18 d, resulted in a similar hydride layer. This observa tion further confirms that the C ore S h ell model is not re l e v ant to the microstructural evolution during the hydrogenation of magnesium 6 .4 Kinetics analysis by John son Mehl Avrami relation for Growth Mechanism of Magnesium Hydride As mentioned in section 6 2, the magnesium hydride nucleates h eterogeneously at the intersection between grain boundary and powder surface. After t= the nucleation sites were consumed and follows the individual growth of each magnesium hydride nucleus. In this section, the objective is to investigate the growth mechanism of magnesium hydride formation. The JMA analysis for hydrogenation temperat ures of 250, 210, 180 and 150 C are presented in Figures from 6 19 to 6 26. Except for the 150 C hydrogenation temperature, all curves show the three stages as discussed in Chapter 5. The mechanism responsible for each stage can be identified by evaluatin g the activation energy. According to equation (2 2) the activation energy for Stage 2 can be calculated by considering the temperature dependency of the constant K, assuming that n is not temperature dependent. Table 6 4 summarizes the K, lnK and n values for Stage 2 and 3. It is noticed that the n values for Stage 2 decrease slightly with increasing temperature, as shown in Figure 6 27. The n values change from 0.5 to 0.68. However, n value is assumed to be unaffected by the temperature. In Table 2 1, it is noticed that for diffusion control growth, every 0.5 is taken as a unit. It is reasonable to use n=0.5 for this material. Karty et al. reported the n value for hydrogen absorption of

PAGE 121

121 Mg 2 Cu coated magnesium to be 0.5 [70] Fernendez et al. demonstrated that the hydrogenation of magnesium is controlled by nucleation and growth and reported exponent values of n =0.5 1 give the best fit to the experimental data [71] The n values from both groups provided the same value as the n value calculated in this study. Based one dimensional diffusion control growth process [29, 70] For the diffusion controlled growth process, it is shown: G=(gt) n =(gt) 1/2 (6 5) where G is the total growth at time t, g is the growth rate, and n value is 0.5. Figure 6 27 also shows the n values in Stage 3, which are much smaller than the n values for Stage 3. It should be noted that Stage 3, which is associated with the thickening of a continuous hydride layer, cannot be evaluated by JMA analysis. Table 6 5 shows the starting point and end of each stage with the corresponding hydrogenation time. The end of the impingement is the end of the transition region. It should be born in mind that the impingement only occurs in the transition region. The transiti on region is located between the two arrows in each curve in Figure 6 2 8 at each temperature. The end of the transition region represents the end of impingement. After all the surface of the magnesium powders is surrounded by hydride layers, Stage 3 occurs However, it is still unknown which phase hydrogen is diffusing through during the growth of hydride ; it can be the metallic magnesium phase or magnesium hydride phase. The following section explains the calculation of the activation energy of the growth of magnesium hydride which was compared to the result obtained by other groups. The value of activation energy would give us an idea which phase hydrogen diffuses through.

PAGE 122

122 The activation energy for Stage 2 was calculated by applying the Arrhenius plot. Th e K values for JMA analysis of powders hydrogenated at 150 C were excluded because these samples did not show a Stage 2. Only the K values of Stage 2 for hydrogenation temperatures of 180, 210, and 250 C were taken into account. Figure 6 29 presents the Ar rhenius plot where lnK is plotted versus 1/T. We can see the slope equal to Q /R is 3899.9, so the activation energy calculated here is 32.4 kJ/ mole. The value of the activation energy is very close to the activation energy of the diffusion of hydrogen through the magnesium phase latest result repor ted in the literature, which is 24kJ/mole [67, 91] The good correlation between the activation energy for diffusion of hydrogen in metallic magnesium and the activation energy for Stage 2 confirms that the magnesium hydride growth takes place by the diffusion of hydrogen through magnesium and not magnesium hydride. As the hydrogenation temperature is lowered to 150 C (Figures 6 25 and 6 26), i t is noticed that the curve exhibits only two stages during th e formation of magnes ium hydride. As demonstrated in Figure 6 18(b), because of the driving force for nucleation of magnesium hydride at this temperature, lots of surface defects were activated near the powder surface, which acted as the nucleation sites f or magnesium hydride on the surface as well as inside the powder particles. Consequently, the impingement of the hydride grains is achieved in a short time with a very thin hydride layer. It is suggested that the nucleation is not a rate controlling step f or 150 C in this case. Therefore, Stage 2, which is the growth of the hydride to impingement, is not detectable in the JMA analysis at 150 C. Because all the powders formed a thin hydride layer at the same time the transition region is also skipped. Howeve r, t here are 3 possible mechanisms which may control the Stage 3: (1)

PAGE 123

123 The dissociation of hydrogen molecules (2) the diffusi vity of hydrogen through hydride layer (3) the diffusion length of hydrogen through the hydride layer. It is believed that the rat e of the dissociation of hydrogen molecule into two hydrogen atoms increases with increasing temperature. The diffusivity of hydrogen through hydride layer is also increased with increasing temperature. Both two possible mechanisms are thermally activated which will lead to a faster reaction rate in Stage 3 at higher temperature. Figure 6 30 shows the possible slopes in the beginning of Stage 3 at 150C and 250C. If this figure is compared with Figure 6 28 it is noticed that the slope in Stage 3 at 150C is higher than that at 250C. Figure 6 7(a, d) shows the microstructural analysis of powders hydrogenated at 150 C and 250 C, respectively. The diffusion length of hydrogen through hydride layer is shorter at 150C than that at 250 C which leads to a high er slope and faster thickening process in the Stage 3. Meanwhile it also indicate s that the rate limiting step in Stage 3 is controlled by the diffusion length of hydrogen through hydride layer. 6 .5 The Fraction of Magnesium H ydride at I mpingement Friedlme ier et al. [30] based on experimental hydrogenation data, have shown empirically that for a given test temperature and pressure the df/dt (f=weight fraction of hydride formed at time t, t = time) varies linearly as a function of time. They have assumed that at relatively high temperature, from 273 to 355 C, the total hydrogen capacity can be approximated as the hydride fraction at impingement, f*. However, as demonstrated in Figures 6 31 through 6 36, the results of hydrogenation of the mag nesium powder at 1MPa and in the temperature range of 250 180 C, reveal that at low temperatures noticeable hydrogen is absorbed during Stage 3. This absorption is manifested as a drastic change in the slope of df/dt as a function of t. These figures

PAGE 124

124 demon strate that similar to the JMA results presented earlier, there is a transition from Stage 2 to Stage 3. Therefore, here, the value of f* is estimated as the intersection of the two linear sections corresponding to these two stages. The hydride fraction s f* for 250 C here are 0.55 and 0.59 for two independent tests. Figure 6 33 exhibits the curve of reaction rate, df/dt, vs the hydride fraction, f, for the hydrogenation tests completed at 210 C. In Figure 6 34, it is calculated that the f* values are 0.5 2 and 0.56. It is noticed that the f* for 210 C is smaller than those at 250 C but the difference is not very significant. For the curve analyzed for the relationship of reaction rate and hydride fraction, f, at 180 C in Figure 6 35, the f* for both curves are measured to be 0.40 and 0.47, as shown in Figure 6 36. Those f* values are clearly smaller than the f* values of 210 C and also 250 C. Figure 6 37 reveals the reaction rate against the hydride fraction, f, for powder samples hydrogenated at 150 C. The f* values for 150 C are both 0.17 as shown in Figure 6 38 and the f* values are smallest among all tests. The f* represents the hydride fraction at impingement and the trend of f* is shown to increase with increasing temperatures. Based on the theoretica l assumption of this analysis, f* should be comparable with the hydride fraction at the midpoint of the transition region from JMA analysis. Figure 6 39 compares the results from the JMA analysis with those calculated based on the linear relationship betwe en the reaction rate and the weight fraction of hydride. There is a relatively good correlation between the two approaches and the latter method results in consistently lower values. The difference may be partially associated with the mathematical methodol ogy to define f*.

PAGE 125

125 6 6 Summary and C onclusions The hydrogenation mechanism for a commercially availab le 44 m Ni coat ed magnesium powder at 1MPa and in the temperature range of 150 to 250 C was successfully evaluated and discussed in this chapter. The amount of hydrogen absorbed after 5 hours of hydrogenation was revealed to be temperature dependent and control led by the amount of hydrogen absorbed to the point of the impingement of the hydride colonies/grains (Stage 2) and the amount of hydrogen consumed in the further growth of the continuous layer of hydride after impingement (Stage 3). As the hydrogenation t emperature decreases, the contribution of Stage 3 becomes more significant and the tendency to reach a pseudo saturation level of hydrogen absorption is decreased. JMA analysis was applied to calculate the activation energy for hydride growth in Stage 2. A ccording to this calculation, the rate limiting step of magnesium hydride growth is the diffusion of hydrogen through the metallic magnesium For the study of nucleation rate on hydrogen capacity, w e investigated the hydrogenation kinetics by two hydroge nation methods and two stage heating procedures The results of both stud ies have demonstrated that the hydrogen capacity of magnesium powder is influenced drastically by the nucleation rate of the hydride. At a high nucleation rate, limited hydride growth takes place before the impingement of the hydride colonies/grains, which retards further growth of the hydride owing to the drastic reduction of hydrogen flux because of the low diffusivity of hydrogen in the magnesium hydride. Based on the findings of th is investigation, for a given particle geometry, there is an optimum hydrogenation condition that leads to the near theoretical hydrogen capacity with fast absorption rate. For example, at low hydrogenation temperatures, owing to a fast nucleation rate the capacity will be low. On the other hand, increasing

PAGE 126

126 the hydrogenation temperature will enhance the growth rate but will decrease the nucleation rate, hence hydrogenation rate. Therefore, there is an optimum hydrogenation temperature, where a good capacity is reached in a relatively fast manner. The study of hydride fraction at impingement gives us an idea how f* values vary with hydrogenation temperatures and shows the very different df/dt curves at 150 C compared with others which proves that it jumps fro m Stage 1 to Stage 3 in the JMA analysis. In conclusion: 1. The study of nucleation number in the initial hydrogenation time period suggests that the nucleation is a constant number during the growth of magnesium hydride. 2. The impingement model is best explai ned in Stage 1 and 2 (Nucleation and growth of hydride), and after impingement the Core Shell model can only explain the Stage 3 (Continuous thickening of hydride layer). 3. The microstructural analysis confirms that the hydrogenation of magnesium particles c an be explained by nucleation, growth and impingement processes before impingement of hydride. 4. The reduction of hydrogen capacity by lowering the hydrogenation temperature can be attributed to the increased nucleation rate. 5. The growth of magnesium hydride in Stage 2 is controlled by the diffusion of hydrogen through the magnesium phase during hydrogenation. The thickening of the hydride layer in Stage 3 is controlled by the diffusion of hydrogen through the magnesium hydride phase during hydrogenation.

PAGE 127

127 Tab le 6 1.The result of the measurement of number of nucleation at different hydrogenation times Hydrogenation Time (Minutes) Total length Total Number of nodules Linear density Distance between two nucleuses Number of nucleation on surface 2.5 4796 1111 0.2316 4.32 32.02 7.5 4444 1043 0.2347 4.26 32.49 Table 6 2 Hydrogen capacity at different hydrogenation temperatures Temp. 150 180 210 250 Test 1 2 1 2 1 2 1 2 Hydrogen capacity 3.2 3.3 5.8 5.4 6.1 6.3 6.5 6.9 Table 6 3. The slope of the three stages of hydrogenation as revealed by JMA analysis in Figure 6 29. Method Stage 1 Stage 2 Stage 3 I 0.27 1.60 0.14 II 0.80 0.53 0.17

PAGE 128

128 Table 6 4 n anf K values for Stage 2 and 3 at each temperatures Stage 2 Stage 3 n lnK K n lnK K 150 1 0.25 1.92 0.15 150 2 0.34 2.32 0.10 180 1 0.61 2.47 0.08 0.24 0.88 0.41 180 2 0.67 3.28 0.04 0.29 1.35 0.26 210 1 0.68 2.42 0.09 0.17 0.19 0.83 210 2 0.53 1.87 0.15 0.17 0.37 0.69 250 1 0.50 1.44 0.24 0.09 0.25 1. 28 250 2 0.55 1.84 0.16 0.17 0.28 0.76 Table 6 5. En d time in minutes of each stage Stage 1 Stage 2 Transition lnt t 1 t 1 lnt t 2 t 2 lnt t t t t 150 1 1.2 0.09 0.9 1.2 150 2 0.8 0.32 1.4 0.8 180 1 1.1 3.87 47.9 5.15 172.5 1.1 180 2 1.1 3.8 9 49.1 5.16 174.7 1.1 210 1 1.6 3.77 43.3 5.23 187.5 1.6 210 2 1.8 3.77 43.5 5.33 207.5 1.8 250 1 1.6 3.67 39.3 5.00 148.6 1.6 250 2 1.8 3.79 44.2 5.1 165.6 1.8

PAGE 129

129 Figure 6 1 The JMA analysis of powders hydrogenated at 210 C under 1MPa. Figu re 6 2 The measurement of perimeter for powders hydrogenated for 2.5 minutes.

PAGE 130

130 Figure 6 3 The measurement of perimeter for powders hydrogenated for 7.5 minutes. Figure 6 4. The etched cross sectional surface of magnesium powder pre heated at 210 C.

PAGE 131

131 Figure 6 5. The hydrogenation curves of magnesium powders hydrogenated at 150, 180, 210, and 250 C. Figure 6 6. The relationship between hydrogen capacity and hydrogenation temperatures.

PAGE 132

132 Figure 6 7. The microstructural analysis of powders hydrogenate d at (a)150 C, (b)180 C, (c)210 C, and (d)250 C under 1Mpa for 5 hours.

PAGE 133

133 Figure 6 8. The schematic illustration of proposed model for the effect of nucleation rate on hydrogenation kinetics and hydrogen capacity. Figure 6 9. Hydrogen absorption as a function of time for the two hydrogenation methods employed.

PAGE 134

134 Figure 6 10. The evaluation of kinetics of hydride formation by Methods I and II on the JMA relationship. Figure 6 11. A comparison of the hydrogenation curves for powders hydrogenated by Me thod I for 75 minutes and by Method II for 3 minutes.

PAGE 135

135 Figure 6 12. SEM pictures comparing the hydride formation for powders hydrogenated by Method I for 75 minutes (a and b) and by Method II for 3 minutes (c and d); (a) and (C) are cross sectional view and (b) and (d) are the external view of the powders. Figure 6 13. The surface coverage measurement.

PAGE 136

136 Figure 6 14. The distribution of the surface coverage of the particles hydrogenated by Method I and II. Figure 6 15. SEM pictures showing cross secti onal views of the particles with high hydride coverage after 75 minutes f hydrogenation by Method I (a and b) and with low hydride coverage after 3 minutes hydrogenation by Method II (c and d).

PAGE 137

137 Figure 6 16. Hydrogenation curves for 15 0C 210 C and the t wo step heating. Point A denotes the time (15 minutes) at 210 C, which results in an equal amount of hydride as the powder hydrogenated at 150 C for 5 hours.

PAGE 138

138 Figure 6 17. SEM (BSE mode) images showing the hydride distribution in the powder hydrogenated at (a and c) 210 C for 15 minutes, and (b and d) at 150 C for 300 minutes. (a) and (b) are cross sectional images, while (c) and (d) are external surface images. magnesium hydride appears dark because of its lower density.

PAGE 139

139 Figure 6 18. SEM (BSE mode) cr oss sectional images showing (a) hydride distribution in powders hydrogenated at 150 C for 2 minutes, (b) higher magnification of (a) showing internal nucleation of MgH 2 (c) hydride formation in powders hydrogenated by the two step heating (150 C for 2 mi nutes and 210 C for 298 minutes), and (d) hydride formation in the powders hydrogenated at 150 C for 300 minutes.

PAGE 140

140 Figure 6 19. The JMA analysis of powder hydrogenated at 250 C (Sample 1). Figure 6 20. The JMA analysis of powder hydrogenated at 250 C ( Sample 2).

PAGE 141

141 Figure 6 21. The JMA analysis of powder hydrogenated at 210 C (Sample 1). Figure 6 22. The JMA analysis of powder hydrogenated at 210 C (Sample 2).

PAGE 142

142 Figure 6 23. The JMA analysis of powder hydrogenated at 180 C (Sample 1). Figure 6 24. Th e JMA analysis of powder hydrogenated at 180 C (Sample 2).

PAGE 143

143 Figure 6 25. The JMA analysis of powder hydrogenated at 150 C (Sample 1). Figure 6 26. The JMA analysis of powder hydrogenated at 150 C (Sample 2).

PAGE 144

144 Figure 6 27. The relationship between n val ues and temperatures at Stage 2 and 3. Figure 6 28. The location of transition region within the hydrogenation curves.

PAGE 145

145 Figure 6 29. The activation energy calculated by Arrhenius plot. Figure 6 30. The schematic illustration of the comparison of slo pes at the beginning of Stage 3 at 150 C and 250 C.

PAGE 146

146 Figure 6 31. The relationship between hydride fraction and reaction rate at 250 C. Figure 6 32. The f* calculation at 250 C

PAGE 147

147 Figure 6 33. The relationship between hydride fraction and reaction rate at 210 C. Figure 6 3 4. The f* calculation at 2 1 0C

PAGE 148

148 Figure 6 35. The relationship between hydride fraction and reaction rate at 180 C. Figure 6 3 6. The f* calculation at 18 0C

PAGE 149

14 9 Figure 6 37. The relationship between hydride fraction and reaction r ate at 150 C. Figure 6 38. T he f* calculation at 15 0C

PAGE 150

150 Figure 6 39. A comparison between hydride fraction at impingement analyzed by two different methods (JMA analysis and relationship between hydride fraction and reaction rate).

PAGE 151

151 C HAPTER 7 EFFECT OF POWDER GEO METRY AND SIZE ON TH E HYDROGEN CAPACITY OF MAGNESIUM 7.1 Introduct ory Remarks The understanding of hydrogenation kinetics is crucial for energy storage applications. According to the US Department of Energy statistics, the fueling process of a hydrogen vehicle tank by 2010 must not exceed 3 minutes [90] One of the shortcomings of magnesium powder as a solid hydrogen storage material is its slow reaction rate and low hydrogen capacity. In Chapter 5 3, it was demonstrated that the particle size distribution can signif icantly affect the kinetics of hydrogen absorption in magnesium powders. In this chapter, the particle geometry as well as the average particle size impact on hydrogenation characteristics under an isothermal and isobaric hydrogenation condition will be i nvestigated. The effects of particle shape and size on the hydrogenation characteristics of magnesium powder have been studied by only a few researchers. A mathematical model using assumptions from the Core Shell model for the shape effect on hydrogenatio n kinetics including sphere, cylinder, and plate of magnesium powders has led to the conclusion that the influences of shape variation and also size distribution of the magnesium particles are not drastic [90] Mintz et al. reported that the influence is not noticeable for particle shape and size on the hydrogenation of magnesium powders by using single particle analysis (SPA) [89] However, interestingly, Jeon et al. have reported that the hydrogen capacity on magnesium flake powders is 4.57 H wt. % at 21 0C, but no other shapes of magnesium powers with the same coatin g procedure were compared in their study [88]

PAGE 152

152 Actually, in the litera ture no mathematical model and experimental data show what the role of the geometry is on the hydrogen capacity. In this study a vibrating machine called High Speed Orbital Ball Milling (HSOBM) was employed to produce flake shaped magnesium particles from a commercial powder with 300 m average particle size as described in Chapter 3. The milling creates strong shear and compression stresses and causes the flattening of the powders. The magnesium particles in the as received powder become thinner and thinne r after passing through the gap between the rolling ball and the vessel wall. For Ni coating, a Theta Composer system was not used since the necessary amount of magnesium powder for this method is about 10 gram which is not practical to obtain from HSOBM w hich produces only 0.1 gram per milling. Both as received and flake magnesium powders hence were coated by organometallic Ni nano particles for the hydrogenation tests. The effect of average particle size was evaluated by comparing the hydrogenation behavi ors of as received commercial 44 and 300 m powders. A simple mathematical model was developed to explain the observed effects of particle shape and size observed in this study. 7.2 Effect of Particle Shape on Hydrogenation Behavior of Mg Powder We hav e shown in Chapter 6 that the nucleation of magnesium hydride controls the hydrogen capacity. It is assumed that the nucleation rate depends on the driving force for nucleation of magnesium hydride, which is a function of temperature and pressure. In this section, an identical hydrogenation condition is employed to as received 30 magnesium nearly sphere and flake powders. Shape change leads to the change of surface to volume ratio (S/V), and it is well known that a spherical particle has the least surface area compared to any other shape of particles with the same

PAGE 153

153 volume. Figur e 7 1 presents the SEM (BSE mode) images for un milled and milled magnesium particles. Figure 7 1(a) shows that the shape of the as received 300 m magnesium particle is nearly spherical. Figure 7 1(b) displays the external surface image for the HSOBM mille d particles, which exhibit a disk like shape. Based on the finding in Chapter 6, the hydride thickness increases with increasing hydrogenation temperature; therefore, the hydrogenation tests were performed at 300 C to distinguish the differences in the cap acity for the as received magnesium and magnesium flake powders. Figure 7 2 shows the hydrogenation curves of the as received 300 m powder a nd flake shaped magnesium powders The as received 300 m magnesium powder sample reached saturation at about 100 minutes and end ed up with a very low hydrogen capacity of only 0. 2 H wt. % and then the test was kept recording for 10 hours On th e other hand, magnesium f la k es exhibited a much higher capacity at 5 .5 H wt. % after 33 hours. As can be seen in Figure 7 2, t he initial hydrogenation kinetics of the magnesium flake powders were also much faster than th at of the as received 300 m magnesiu m powder These findings demonstrate that the shape of the particles in a powder can play an important role in the hydrogenation behavior. The difference in hydrogen capacity can be explained by examining the microstructure of the hydrogenated powders Fi gure s 7 3 (a, b) shows the SEM (BSE mode) cross section al image s of the hydrogenated as received 300 m magnesium powder It is observ ed that the as received 300 m magnesium particle is covered by a thin nonuniform layer composed of magnesium hydride grains/colonies. In Figure 7 4, when we take a closer look at high magnification of the hydrogenated 3 00 m magnesium spherical particle s, the thickest part of hydride layer is about 26 m thick

PAGE 154

154 and the thinnest part is about 5 m thick. This variation in thickness can partially be attributed to the relative position of the cross sectioned plane and the hydride c olonies as well as any nonuniformity in the hydride nucleation process. In Figure 7 5 the SEM (BSE mode) cross section al image shows the hydride distribution in the hydrogenated magnesium flakes. The hydrogenat ed magnesium flake displays a much larger area fraction of magnesium hydride phase when compared to as received 300 m magnesium particles. However, both of the powder samples demonstrated impingement behavior by a completely covered magnesium hydride layer on the powder surface. It is indicated that that hydrogenated powders had reached saturation in both cases. Figure s 7 5 (a, c) shows that there is a metallic magnesium phase left at the center area of the powder and the surface is covered by magnesium hydride phase. From the individual growths of hydride colonies/grains it is predicted the nucleation and growth can be st describe the hydrogenation behavior of this magnesium flake particle before the impingement occurs. From now, t his hydride layer may not allow further hydrogen to diffuse through and can be treated by the core shell model Figures 7 5 (b, d) reveal the magnesium flake powders which have nearly completely hydrogenated. The thicknesses of those magnesium flake particles are thinner than the flake particles in Figures 7 5 (a, c). The thinner the flake thickness the less metallic magnesium phase was found in the center area under the same hydrogenation condition. In Figure 7 5 (c), it can be noticed the welding effect from the HSOBM test, where two magnesium flakes stack on each other owing to the high speed rolling process.

PAGE 155

155 The microstructure s of the HSOBM milled and the as received magnesium powders are expected to be identical because the hydrogenation temperature of 300C is anticipated to be sufficient for th e re crystallization of magnesium Under the same hydrogenation condition, the driving force and t hermal energy are equal and the nucleation density is expected to be similar in both sample powders. Moreover, the growth rates are anticipated to be the same since the growth mechanism is a thermally activated diffusion controlled process. Therefore, it i s suggested that an identical hydrogenation condition leads to the same area density of nucleation of magnesium hydride on the powder surface as well as equal growth rate prior to the impingement of the hydride colonies/grains. A high surface to volume (S/ V) ratio results in a relatively large number of nuclei in magnesium flake particles; therefore, the total volume of the hydride that forms per unit time is larger at the moment of impingement of hydrides. Notice here that the hydrogen capacity increases w ith increasing S/V ratio. One can consider that if a cube and a sphere have the same surface area, at the same hydrogenation condition, the hydrogenation kinetics is identical in both shapes but the hydrogen capacity will be lower in the spherical particle owing to its lower S/V ratio. Figure 7 6 shows the schematic illustration of the effect of shape on hydrogenation behavior and hydride distribution. Assuming that the hydride grains/colonies have the same size, the model presented in this figure clearly s hows the reason why the flake particles have higher hydrogen capacity than spherical particles do When energy saving is considered, low hydrogenation temperature is favored for the hydrogenation process and thin flake powders are necessary to obtain high hydrogen capacity. Based on the hydride thickness measurement in Chapter 6 at different hydrogenation

PAGE 156

156 temperatures, the lowest temperature, 150C, leads to an approximate hydride hydrogen capacity at 150C is expected to reach the theoretical value of 7.6 H wt. %. 7.3 Effect of Particle Size on Hydrogen Capacity Figure 7 7 presents, while the hydrogen capacity after 5 hours of hydrogenation for the as received 300 m magnesium powder is 0.2 H wt. %, the as received 44 m magnesium powder displays a much higher capacity of 6.5 H wt. % capacity. Schultz et al. have reported a h ydrogen absorption capacity 6.5 H wt. % for a 20 m magnesium powder hydrogenated for 25 minutes [34] Similar to the particle shape change effect, decreasing the particle size increases the surface to volume ratio, hence increasing the total number of hydride nuclei per unit weight of the material. Consequently, at a given hydride linear growth rate, the amount of hydride formed per unit weight at a given time will be increased resulting in faster observed kinetics. Also, as was demonstrated schematically in Chapter 5, the impingement of the hydride particles is achieved at a higher volume fraction of hydride in the powder with a smaller particle size. Microstructural analysis confirms the hypothesis presented in the schematic shown in Figure 5 10. Figures 7 8 (a, c) show the SEM (BSE mode) images of magnesium hydride distribution at the cross section of 44 m magnesium particles hydrogenated at 3 00 C under 1MPa hydrogen pressure for 10 hours. It can be seen that almost all the magnesium phase has transformed to magnesium hydride except a small volume of magnesium phase left at the center of the particle. Lot s of cracks are also found inside the hydrogenated particles. The hydride distribution is compared with the 300 m hydrogenated particles in Figures 7 8 (b, d). Again a continuous hydride layer indicative

PAGE 157

157 of complete impingement of the hydride colonies/grai ns is observed, but a relatively large area of un reacted metallic magnesium phase was observed in the center of the 300 m hydrogenated particles. The comparison of microstructural analysis suggests that the hydride fraction in 44 m is much larger than tha t of 300 m hydrogenated particles. This phenomenon also manifested as a variation in the volume fraction of the hydride found in the single batch of the 44 m powder sample as shown in Figure 5 10. 7.4 Mathematical Modeling In this section we attempt to pr edict the hydrogen capacity at the point of impingement, i.e. at the transition between Stages 2 and 3 as defined in Chapter 6, for different particle sizes and geometries. It is assumed that at a given temperature, T, and pressure, P, the hydride thicknes s at impingement is independent of size and geometry. This assumption is justified because the nuclei areal density and the linear hydride growth rate are expected to vary only with temperature and pressure of the hydrogenation. The shapes of the as receiv ed magnesium powder, which are approximated as prolate spheroid and the magnesium flake as oblate spheroid (disk), are shown schematically in Figure 7 9. In Chapter 4, the aspect ratio (a/b) was reported to be 0.47 and 0.67 for the as received 44 m and 30 0 m powders, respectively. Figure 7 9 (a) exhibits the schematic cross sectional image of a prolate spheroid with the labels of the equatorial radius a and polar radius b and 7 9 (b) the cross section of an oblate spheroid with the labels of polar radius a and equatorial radius b axis, respectively. The cross section looks the same but the volume calculations for those two spheroid particles are different. For the prolate, the volume is ( a 2 b ) /3 and for the

PAGE 158

158 disk, the volume is ( ab 2 ) /3 The volume fracti on of the hydride at the point of impingement at a given T and P can be calculated by knowing the average thickness of the hydride, h for the hydrogenation condition. For example for a prolate the hydride fraction, f, is given as: f = [ a 2 b )/3 a h ) 2 ( b h ) )/3 ]/ [ a 2 b )/3 ] (7 1) The volume fraction then can be changed to the wt. % of H absorbed by assuming a density of 1.74 and 1.45 (g/cm 3 ) for magnesium hydride and magnesium respectively, according to the following relationships: Wt. % of H = {1.45f /[1.74(1 f)+1.45f]}x7.6 (7 2) From Figure 7 4, the average hydride thickness is about 17.3 m for as received 300 m magnesium powders hydrogenated at 300 C under 1MPa for 15 hours. The middle thinner part of the hydride grains/colonies has not been taken i nto account because the cutting plane is not right on the center of the hydride grain/colony. Only the portion of clearly individual nodules was utilized for the thickness measurement. In Figure 7 5(c), the thickness hydride layer is easier to be distingui shed in this thicker flake particle than the thinner one in Figure 7 6(d). A measurement of hydride thickness was done at every 30 m length on those hydrogenated magnesium flakes and an average hydride thickness, 13.3 m was obtained. Unlike the way for measuring the hydride thickness in magnesium flakes, for near spherical particles, as received 44 and 300 m the hydride thickness is measured at every 20 degree from the center of the particle, 18 data will be obtained for each particle. Average thicknesses are 12.2 m and 9.9 m for as received hydrogenated 44 and 300 m magnesium particles, respectively. As the three hydride thickness es are compared, an average11.8 m and standard

PAGE 159

159 deviation 1.73 is calculated from those three hydride thicknesses. It is noticed that the three hydride thickness values are very close and of the same order. It can be concluded that the hydride thickness is similar no matter the size and shape of the particle, and at the same temperature there is a corresponding hydride thickness. The hydride thickness values at 300 C are not comparable with those of 180, 210, 250 C since the Ni coating process is different a nd hence the nucleation behavior would be dissimilar, which affects the hydrogen capacity dramatically. Figure 7 10 shows the calculated absorbed hydrogen as a function of the particle aspect ratio ( a / b ) for h =11.8 m Here both 44 m and 300 m size are considered and the calculation is done for a spherical prolate. The calculation is done as disc shape for magnesium flakes. As the aspect ratio deviates from the spherical shape ( a = b ) the hydrogen capacity for a give n hydrogenation condition, i.e. hydride thickness, increases significantly. Similarly, decreasing the volume of the particle, i.e. its size, results in an enhancement of the hydrogen absorbed at the point of impingement. The hydrogen capacity from the expe rimental data for those three powder samples are compared in Figure 7 10 with the mathematical model. The modeling predicts that 4.32 H wt. % will be obtained for magnesium flake powders at the aspect ratio = 0.05, which is very close to the experimental d ata, 5.5 H wt. %. It is also expected about 7.1 wt. % hydrogen will be absorbed at a/b = 0.47 for the 44 m prolate powder sample. The predicted value is also close to the experimental capacity 6.5 H wt. %. However, the model does not fit well for the 300 m prolate powder sample. The divergence may be caused by the deviation from the calculation for the aspect ratio for as received 300 m particles or it may come from the error of the measurement during the real time hydrogen absorption.

PAGE 160

160 7.5 Su mmary and Conclusions The aim of this chapter is to study the influence of the magnesium powder hydrogen pressure. It is concluded that the shape does affect the kinetic s and hydrogen capacity of magnesium powder particles. The flake shaped particles show impressively 25 times more hydrogen capacity than the as received sphere particles. The microstructural analysis reveals that the thinner the flakes the higher hydride f raction for the magnesium flakes. After hydrogenation, most of the flakes were completely hydrogenated at SEM (BSE mode). Some of the thick flake particles still had metal phase magnesium left in the center of the particles since the thickness of the flake is too thick to be completely hydrogenated by a given hydride thickness ~11.8 m at a given temperature 300 C and pressure 1MPa. For the effect of size, the hydrogenation behavior of 44 m magnesium powders was also compared with the 300 m magnesium particles. The result corresponded to the literature research that shows finer magnes ium particles display faster kinetics and higher hydrogen capacity. We can say that the surface to volume ratio plays an important ro le in the hydrogen capacity At a constant linear growth rate of hydride, the higher the surface to volume ratio the higher the hydrogen capacity. If the total volume of magnesium powders is fixed, larger surface area will result in faster hydrogenation kinetics. It is also concluded that a hydrogenation temperature corresponds to a specific hydride thickness no matter what th e size and shape of the magnesium particles are.

PAGE 161

161 Figure 7 1. External surface images at higher magnification of (a) as received 300 m magnesium spherical and (b) magnesium flake particles. Figure 7 2 The hydrogenation curves of different shapes of pa rticles. The hydrogen capacity of magnesium flakes s hows more than 5.5 H wt. % which is much higher than the as m magnesium par ticles with 0.2 H wt. % capacity.

PAGE 162

162 Figure 7 3. The SEM (BSE mode) images of hydrogenated as m magnesi um particles after hydrogenation at 300 C 1MPa for 15 hours. Figure 7 4. The SEM (BSE mode) images for the hydride layer measurement at higher magnification of hydrogenated as m magnesium particles after hydrogenation at 300 C 1MPa for 15 ho urs.

PAGE 163

163 Figure 7 5. The SEM (BSE mode) images of hydrogenated magnesium flake particles after hydrogenation at 300 C 1MPa for 36 hours. (a) Thick flake particles (b) thin flake particles.

PAGE 164

164 Figure 7 6. The schematics illustration of effect of shape change from sphere to flake on hydrogen capacity of magnesium particles with SEM (BSE mode) images for the comparison.

PAGE 165

165 Figure 7 7. The hydrogenation curves of different sizes of particles. The hydrogen capacity of 44 m magnesium particles shows more than 6.5 H wt. % which is much higher than the 300 m magnesium spherical particles with 0.2 H wt. % capacity.

PAGE 166

166 Figure 7 8. (a, c) The SEM (BSE mode) images of hydrogenated 44 m magnesium particles after hydrogenation at 3 00 C 1MPa for 10 hours compared with the images of (b, d) which show the cross section of hydrogenated as received m magnesium particles. Figure 7 9. (a) Prolate spheroid for AR 44 and 300 m (b) Disk spheroid for flake

PAGE 167

167 Figure 7 10. The experi mental hydrogen capacity compared with theoretical calculation by using 11.8 thick hydride layer at varying a/b aspect ratio.

PAGE 168

168 C HAPTER 8 D EHYDROGENATION MECHA NISM OF MAGNESIUM HY DRIDE 8.1 Introduct ory Remarks The industrial application of magnesium hydri de as a hydrogen storage medium has been impeded by its high thermodynamic stability which renders a high hydrogen release temperature and its dawdling kinetics. To date the hydrogen absorption and desorption kinetics can be increased by adding catalyst o n powder surface and the thermodynamic properties of magnesium hydride can be modified by the addition of alloying elements to form complex hydride [18, 56] Generally, the dehydrogenation kinetics increases signif icantly with increasing the dehydrogenation temperature [31] In the isothermal and isobaric reaction of hydride formation, Mg+H 2 MgH 2 as the hydrogenation temperature is close to the equilibrium temperature (T h eq T h is the hydrogenation temperature and T eq is the equilibrium temperature at a given pressure), it leads to large thermal energy for hydrogen diffusion and interface movement of Mg/MgH 2 in company with a small under cooling for magnesium hydride nuc leation. Furthermore, low nucleation rate and high growth rate are observed at high hydrogenation temperatures When the difference between applied dehydrogenation temperature and the equilibrium temperature is large (T dh >>T eq T dh is the desired dehydroge nation temperature), magnesium hydride is easily dissociated and rapid nucleation rate of hcp Mg with large over heating occurs upon dehydrogenation [38] In the kinetics point of view, high dehydrogenation temperature is certainly beneficial to expedite the dehydrogenation process but not favorable to the practical use because the consumption of heat reduces the efficiency of hydrogen based fuel cells

PAGE 169

169 In this chapter the effect of hcp Mg nucleation energy barrier on the dehydrogenation temperature is investigated. To investigate the role of hcp Mg nucleation on the dehydrogenation temperature two h ydrogenated powder samples were dehydrogenated both in TGA and in the house build Hydrogenation/Dehydrogenation unit One partially hydrogenated powder sample contain ed hcp Mg on the surface and the other fully hydrogenated one encompasse d no hcp Mg on it. Detailed microstructural analyses were conducted on those powder samples to evaluate the mechanism of nucleation and growth of the hcp Mg phase. In a different approach dehydrogenation tests were conducted at different temperatures and at constant 0.1MPa pressure to look for the equilibrium temperature by using uncovered hydride particle to avoid the nucleation barrier of hcp Mg upon dehydrogenation. The results were compared with the reported equilibrium temperature under 0.1MP from the literature 8 .2 Desorption Study by Constant H eating R ate in TGA The magnesium powder sample was coated with nano Ni by using organometallic solution for hydrogen absorption and desorption tests. Figure 8 1(a) illustrates the hydrogenation curve of the magnesium powder c arried out at 210 C for 5 h ours under 1 MPa pressure. After 5 h ours of hydrogenation, 6.1 wt. % of hydrogen was absorbed into the powder. As was demonstrated in Chapters 5 and 6 the hydrogen absorption takes place in three stages based on JMA analysis ln[ ln(1/1 f )] vs ln t is plotted in Figure 8 1 ( b ) which demonstrates the three reaction regimes. To elucidate the effect of the presence of hcp Mg on the surface on the dehydrogenation behavior, a partial hydrogenation test was designed, in which a powder sa mple was hydrogenated for 54 min utes corresponding to point A in Figure 8 1 ( b ) Since this point lies within S tage 2 of

PAGE 170

170 hydrogenation, the powder was expected to be partially covered with hydride. Figure 8 2 shows the SEM ( BSE mode) pictures of typical po wders after hydrogenation. Consistently, the microstructure of the powder after 5 h ours of hydrogenation, presented in Figure 8 2 ( a ) reveals that the surface of the powders is completely covered with the magnesium hydride phase, while the hydride colonies in powders hydrogenated for 54 mi nutes did not completely impinge on each other, as demonstrated in Figure 8 2 ( b ) After the hydrogenation test, the hydrogenated powder sample was weighed 12mg for the TGA analysis. Figure 8 3 illustrates the dehydrogenati on curves of the saturated and partially hydrogenated powder samples. The total amount of hydrogen released was consistent with the weight of the hydrogen calculated from the hydrogenation tests ( Fig ure 8 1 ). As speculated, a significant difference in the hydrogen release temperature was observed. The onset temperature of hydrogen release from the partially hydrogenated sample (54 min utes ) was 219 C, but for the sample hydrogenated to saturation (5 h ours ), hydrogen was released at 290 C for similar conditio ns used in the TGA. The hydrogen release rate as indicated by the slope of the weight loss vs temperature curve was found to be slightly higher in the saturated sample, owing to the higher hydrogen release temperature. The release of hydrogen from the magn esium hydride involves the breakage of the Mg H bond, the nucleation of the hcp Mg phase, the diffusion of hydrogen onto the surface and the recombination of H atoms to form a H 2 molecule. When the surface of the powders is completely covered with the hydr ide phase, the hcp Mg phase left in the core of the particles can grow upon the dehydrogenation process, but the rate will be extremely slow owing to the low diffusivity of hydrogen through the hydride phase.

PAGE 171

171 Therefore, the observable dehydrogenation proce ss most probably begins by the nucleation of the hcp Mg on the surface of the hydrogenated powders. To confirm this mechanism, microstructural analysis was carried out on the partially dehydrogenated samples. A powder sample hydrogenated to saturation (5 h ours ) was partially dehy drogenated in the TGA. This was achieved by heating the powder at the rate of 5 C/minute from room temperature to 310 C, which is just above the onset temperat ure of hydrogen release room for this powder (see Fig 8 3 ), then the samp le was cooled down at a much faste r rate of 30 C/ min utes in an attempt to retain the microstructure at 310 C. It is anticipated that the rapid cooling reduces the temperature sufficiently fast such that further hydrogen release process was stopped. Figure 8 4 illustrates the microstructure of the partially dehydrogenated powder at different magnifications. At lower magnification in Fig ure 8 4 ( a), the hcp Mg (light contrast) completely covers the surface of the particle. The darker region, which corresponds to magnesium hydride is observed in the center of the particles. The higher magnification picture Fig ure 8 4 ( b) demonstrates that the light regions extend completely to the surface. The presence of the hcp Mg on the surface shows that the dehydrogenati on started through the nucleation of the hcp Mg phase on the surface of the powders. This observation is consistent with previously reported studies. Once the hcp Mg is nucleated on the surface, the dehydrogenation continues to take place at the interface and the hydrogen diffuses through the hcp Mg phase to the surface, where H 2 molecules form. The above observations suggest that the presence of the hcp Mg phase on the surface of the partially hydrogenated powders eliminates the nucleation barrier for the formation of the hcp Mg phase from the magnesium hydride phase. However, powders hydrogenated to

PAGE 172

172 saturation require a higher thermal energy to overcome the nucleation of the hcp Mg phase; hence it exhibit s a higher dehydrogenation temperature. 8.3 Isotherm al Dehydrogenation Study of MgH 2 It was demonstrated in the previous chapters that the hydrogenation process of magnesium at 1MPa and in the temperature range of 180 to 300 C can be explained by the nucleation, growth and impingement of the magnesium hydri de [99, 100] Upon the dehydrogenation process we may perhaps treat the decomposition of magnesium hydride as the nucleation and growth behavior of the hcp Mg phase. In this study, magnesium powder samples were hy drogenated at 210C and under 1MPa pressure and then dehydrogenated at 0.1MPa in the same system, as described in Chapter 3, at several temperatures near the reported equilibrium temperature at this pressure [38] Figure 8 5( a ) shows the hydrogenation/dehydrogenation curves. The powder samples were heated at 210C under 1MPa hydrogen pressure for 5 ho urs following dehydrogenation tests at varying temperatures. As expected, the hydrogen desorption rate is rapid initially and diminishes afterward due to the reduction of hydride fraction. The increase of the temperature results in fast kinetics and large driving force for hydrogen releasing reaction. At 300C 80% of the magnesium hydride is decomposed after 15 minutes. A high dehydrogenation temperature is expected to render a fast hcp Mg nucleation rate owing to the large driving force for magnesium nucl eation and thermal energy for hydrogen diffusion, which results in numerous releasing paths of hydrogen atoms escaping through the hcp Mg phase. The microstructural evaluation of a partially dehydrogenated sample at 300C, marked as P on Figure 8 5( a ) sho wed that the dehydrogenation takes place on the surface of particles, where hydrogen can escape easily and the curvature indicates that the movement of hcp Mg/MgH 2 interface

PAGE 173

173 from the surface into the hydride phase as shown in Figure 8 1 ( b). At 290C the am ount of hydrogen release was found to be negligible. At 290C, the driving force for magnesium nucleation is small and the nucleation of hc p Mg is the rate controlling mechanism. The slow kinetics is attributed to the extraordinarily thick magnesium hydrid e layer. In this case, the rate limiting step during dehydrogenation is the diffusion of hydrogen atoms through the magnesium hydride layer. If the powder surface is covered by magnesium hydride layer, the hydride fraction remain s unchanged which shows the similar hydride distribution with as hydrogenated powders in Figure 8 5( b ) The small amoun t of hydrogen desorption at 290 C is assumed to issue from the as hydrogenated powder with very small area of hcp Mg on the powder surface. Once those powders are e ntirely dehydrogenated the hydrogen desorption process is brought to an end. In order to investigate the role of hcp Mg nucleation on the kinetics of the dehydrogenation process a sample was partially hydrogenated at 210C and then dehydrogenated at 290C The hydrogenation times are different, but dehydrogenation conditions conducted are the same for both hydrogenated samples, at 290C under 0.1MPa for 5 hours. Figure 8 6 reveals that unlike the dehydrogenation of the saturated powders, the partially hydr ogenated powder releases hydrogen easily at 290C. The desorption rate of powder sample with the existence of hcp Mg on the surface is much faster than the one without hcp Mg on the surface. This difference in behavior can be attributed to the availability of hcp Mg in the latter sample. Figure 8 7 compares the microstructures of partially hydrogenated and saturated powders. The surface of the saturated powder is completely covered with a hydride layer while the partially

PAGE 174

174 hydrogenated powder exhibits expose d hcp Mg phase. In this case, the energy required during dehydrogenation included the nucleation barrier of the hcp Mg phase has been reduced and only the growth of magnesium is happening. The availability of hcp Mg on the surface drastically increases the dehydrogenation kinetics and the nucleation of hcp Mg plays an important role in the dehydrogenation kinetics of magnesium hydride These results suggest that for the powders covered by hydride layer, the initial rate limiting step for hydrogen desorption from magnesium hydride is the nucleation of hcp Mg. In order to establish the real transformation temperature at 0.1MPa for MgH 2 hcp Mg + H 2 partially hydrogenated samples were dehydrogenated at different temperatures. Figure 8 8 illustrates different dehydrogenation temperature s varying from 275C to 300C on partially hydrogenated powder sample s The kinetics effect was in the range of 280C to 300C the diffusivity of hydrogen in hcp Mg increase s with increas ing the thermal energy hence the kinetic s increases with increasing dehydrogenation temperature. As temperature is increased, both the driving force for nucleation of hcp Mg and the diffusion rate are increased. The rate limiting step is assumed to be the diffusion of hydrogen through the magnes ium phase. At 280 C, the powder sample releases 2.5 H wt. % during dehydrogenation and the dehydrogenation kinetics is very slow. It is assumed that the equilibrium temperature under 0.1MPa is slightly below 280C. For 275C curve, thermodynamics mainly c ontrols the dehydrogenation behavior since 275C is below the equilibrium temperature of 0.1MPa no hydrogen can release at this temperature. Hence it is confirmed that the equilibrium temperature at 0.1MPa is between 275C and 280C. As the value is compar ed with the

PAGE 175

175 equilibrium temperature 287.7C reported before [38] it is about 8C lower than it. It is concluded that the reported release temperature usually needs the thermal activation for the nucleation of hcp Mg. The only reason we obtain lower release temperature than before is because we have been using material without activation barrier requ ired. To the extent that we know the difficulty of achieving equilibrium status, and the parameter for equilibrium reported so far may be a little bit off from the real case due to the unavoidable kinetics effect. 8.4 Summary and Conclusions In conclusions we have introduced a logical initiative to enhance the hydrogen desorption kinetics by preserving the hcp Mg phase on the powder surface. This consideration would expedite the dehydrogenation process via reducing the nucleation barrier of hcp Mg. The equ ilibrium temperature of 0.1MPa hydrogen pressure was experimentally tested to be the range of 275C to 280C which is about 10 C lower than the equilibrium temperature reported before. Obviously, fully covered hydride powder was used which increases the di fficulty of hydrogen release. For the study of computational simulation and phase transformation of Mg H system, this finding makes a strong point of the importance of careful consideration in kinetics while thermodynamics properties is being measured. Fro m the results of this study, it can be concluded that: 1. Dehydrogenation of magnesium hydride takes place on the surface of the powder. 2. The high er hydrogen release temperature of magnesium hydride in comparison with the equilibrium temperature based on the M g H phase diagram can be attributed t o the difficulty of hcp Mg phase nucleation. 3. The presence of hcp Mg phase in partially hydrogenated powders eliminates the nucleation barrier for formation of the hcp Mg phase, and thereby reduces the hydrogen release t emperature.

PAGE 176

176 4. Figure 8 1. (a) Hydrogenation curve of hcp Mg at 210 C under 1MPa pressure. (b) Evaluation of different stages of hydrogenation of magnesium powders based on JMA relationship. Figure 8 2. SEM (BSE mode) micrographs (a) of the saturated pow der revealing the impingement of magnesium hydride grains/colonies and (b) of the partially hydrogenated powder showing the presence of the hcp Mg phase on the surface.

PAGE 177

177 Figure 8 3. Dehydrogenation curves obtained using TGA of the partially and completely hydrogenated powders at 5 C/minute and under He flow, revealing the difference in the hydrogen release temperature between the samples. Figure 8 4. SEM (BSE mode) micrographs of the partially dehydrogenated powder. (a) Lowering magnification of the part icle showing the release of hydrogenation around the particle with MgH 2 inside the particle. (b) Higher magnification image showing the extension of the hcp Mg phase to the surface of the particle.

PAGE 178

178 Figure 8 5. (a) Hydrogenation/dehydrogenation curves. Th e hydrogenation condition is the same for all powder samples (At 210C under 1 MPa for 5 hours). The dehydrogenation tests were done under 0.1 MPa at 290C, 295C, 300C for 5 hours. The SEM/BSE image at point P is shown in Figure 1(b), in the vicinity nea r the powder surface of the partial dehydrogenated powders after 5 minutes releasing at 300 C under 0.1MPa. Figure 8 6. Hydrogenation/dehydrogenation curves. The hydrogenation tests are done at 210C under 1MPa for 15 minutes and 5 hours. The dehydrogen ation tests are done under 0.1 MPa at 290C for 5 hours immediately after hydrogenation tests for both powder samples.

PAGE 179

179 Figure 8 7. The SEM/BSE images of as hydrogenated powder. In Figure 3(a), SEM/BSE images show the hydride distribution in the hydride p owders hydrogenated at 210 C under 1 MPa for 5 hours. The Sample powder exhibits pseudo saturation behavior. In Figure 3(b), the SEM/BSE images shows the hydride distribution in the hydride powders hydrogenated at 210C under 1 MPa for 15 minutes and the sample powder shows unsaturated behavior. Figure 8 8. Hydrogenation/dehydrogenation curves. The hydrogenation tests are done at 210C under 1MPa for 15 minutes. The dehydrogenation tests are done under 0.1 MPa at 290C for 5 hours at 275C, 280C, 290C, and 300C.

PAGE 180

180 CHAPTER 9 CONCLUSIONS AND FUTU RE WORK Hydrogen is one of the most promising energy sources in the 21 st century. In this study, we demonstrated the fundamental hydrogenation behavior of Mg powders. The finding promotes the understanding of hydr ogen absorption mechanism using microstructural analysis. At temperature 180 to 250 C, the magnesium hydride formation before impingement is best explained by the impingement model, after impingement the Core Shell model is applicable for the thickening pr ocess of magnesium hydride layer. Pseudo saturation occurred at Stage 3 during the thickening process owing to the very thick hydride layer. At 150 C, it can be treated as Core Shell model only since a continuous hydride layer immediately formed and the hy dride fraction increased with increasing hydride thickness. For a given shape and size of the Mg powders there is an optimum hydrogenation temperature and pressure. By making magnesium flakes for the storage material the hydrogenation energy can be actuall y lowered to increase the efficiency of the system integration of hydrogen powdered cars. Hydrogenated magnesium powders with magnesium phase on the surface will increase the dehydrogenation kinetics by lowering the nucleation barrier of magnesium phase. T he dehydrogenation temperature can be lowered if magnesium phase presents on the hydrogenated powder surface. Suggested future work : What lack is the effect of hydrogen pressure, because not only temperature affects the capacity also the capacity is a func tion of pressure. For the hydrogenation, much more work should be emphasized on the effect of pressure. Moreover, the activation energy of hydrogen through magnesium phase was obtained; mass balance can be used to develop a mathematical model based on the hydrogen

PAGE 181

181 flux though magnesium and magnesium hydride phase. The total flux should include the hydrogen flux through both phases. The surface area of magnesium and magnesium hydride can be estimated by the growth rate of magnesium hydride in hemisphere form in Stage 2. A very thin magnesium flake should be made and thickness is suggested to smaller than 4 m if 150 C is applied for the hydrogenation temperature in order to reach 7.6 H. wt. % capacity. Magnesium coating process is suggested to be applied on th e hydrogenated magnesium powder to help to lower the nucleation barrier of magnesium and enhance and dehydrogenation kinetics during hydrogen desorption.

PAGE 182

182 LIST OF REFERENCES 1. MacKenzie, J., Oil as a finite resource. Natural Resources Research, 1998. 7 (2): p. 97 100. 2. Bockris, J.O.M., Will lack of energy lead to the demise of high technology countries in this century? International Journal of Hydrogen Energy, 2007. 32 (2): p. 153 158. 3. Hubbert, M.K. Nuclear Energy and the Fossil Fuels. Spring Meeting of the Southern District. Division of Production. American Petroleum Institute. San Antonio, Texas: Shell Development Company. 1956: p. 22 27. 4. Brandt, A.R., Testing Hubbert. Energy Policy, 2007. 35 (5): p. 3074 3088. 5. Duncan, R.C., World Energy Production, Population Growth, and the Road to the Olduvai Gorge. Population & Environment, 2001. 22 (5): p. 503 522. 6. Schlapbach, L. and A. Zuttel, Hydrogen storage materials for mobile applications. N ature, 2001. 414 (6861): p. 353 358. 7. Chandra, D., J. Reilly, and R. Chellappa, Metal hydrides for vehicular applications: The state of the art. JOM Journal of the Minerals, Metals and Materials Society, 2006. 58 (2): p. 26 32. 8. Ritter, J.A., et al., Imp lementing a hydrogen economy. Materials Today, 2003. 6 (9): p. 18 23. 9. Z ut tel, A., Materials for hydrogen storage. Materials Today, 2003. 6 (9): p. 24 33. 10. Ewan, B.C.R. and R.W.K. Allen, A figure of merit assessment of the routes to hydrogen. Internatio nal Journal of Hydrogen Energy, 2005. 30 (8): p. 809 819. 11. Fichtner, M., Nanotechnological Aspects in Materials for Hydrogen Storage. Advanced Engineering Materials, 2005. 7 (6): p. 443 455. 12. Grochala, W. and P.P. Edwards, Thermal Decomposition of the Non Interstitial Hydrides for the Storage and Production of Hydrogen. Chemical Reviews, 2004. 104 (3): p. 1283 1316. 13. Sandrock, G., A panoramic overview of hydrogen storage alloys from a gas reaction point of view. Journal of Alloys and Compounds, 1999. 293 295 : p. 877 888. 14. Sandrock, G. and R.C. Bowman, Gas based hydride applications: recent progress and future needs. Journal of Alloys and Compounds, 2003. 356 357 : p. 794 799. 15. D e Wit, M.P. and A.P.C. Faaij, Impact of hydrogen onboard storage techn ologies on the performance of hydrogen fuelled vehicles: A techno economic well to wheel assessment. International Journal of Hydrogen Energy, 2007. 32 (18): p. 4859 4870. 16. Kalisvaart W.P., N.P.H.L., Mechanical alloying and electrochemical hydrogen stora ge of Mg based systems. J. Mater. Res., 2008. 23 : p. 2179. 17. Sakintuna, B., F. Lamari Darkrim, and M. Hirscher, Metal hydride materials for solid hydrogen storage: A review. International Journal of Hydrogen Energy, 2007. 32 (9): p. 1121 1140. 18. Vajo, J .J., et al., Altering hydrogen storage properties by hydride destabilization through alloy formation: LiH and MgH2 destabilized with Si. Journal of Physical Chemistry B, 2004. 108 (37): p. 13977 13983.

PAGE 183

183 19. Steele, B.C.H. and A. Heinzel, Materials for fuel c ell technologies. Nature, 2001. 414 (6861): p. 345 352. 20. Steele, B.C.H., Material science and engineering: The enabling technology for the commercialisation of fuel cell systems. Journal of Materials Science, 2001. 36 (5): p. 1053 1068. 21. Starz, K.A., e t al., Characteristics of platinum based electrocatalysts for mobile PEMFC applications. Journal of Power Sources, 1999. 84 (2): p. 167 172. 22. Jensen, J.O., et al., 100 200[thin space]?C polymer fuel cells for use with NaAlH4. Journal of Alloys and Compou nds, 2005. 404 406 : p. 653 656. 23. Maddalena, A., et al., Study of Mg based materials to be used in a functional solid state hydrogen reservoir for vehicular applications. International Journal of Hydrogen Energy, 2006. 31 (14): p. 2097 2103. 24. http://www.nrel.gov/basic_sciences/carbon_based_hydrogen_center.cfm 25. Bouza C., R.S., Satyapal J. and Milliken, Annual DOE Hydrogen. Fuel Cell and Infrastructure Technologies Pro gram Review Meeting Philadelphia, May 25, 2004. 26. Zaluska, A., L. Zaluski, and J.O. Strom Olsen, Nanocrystalline magnesium for hydrogen storage. Journal of Alloys and Compounds, 1999. 288 (1 2): p. 217 225. 27. Liang, G., et al., Catalytic effect of trans ition metals on hydrogen sorption in nanocrystalline ball milled MgH2 Tm (Tm=Ti, V, Mn, Fe and Ni) systems. Journal of Alloys and Compounds, 1999. 292 (1 2): p. 247 252. 28. Vigeholm, B., et al., Elements of hydride formation mechanisms in nearly spherical magnesium powder particles. Journal of the Less Common Metals, 1987. 131 (1 2): p. 133 141. 29. Rudman, P.S., Hydriding and dehydriding kinetics. Journal of the Less Common Metals, 1983. 89 (1): p. 93 110. 30. Friedlmeier, G. and M. Groll, Experimental analy sis and modelling of the hydriding kinetics of Ni doped and pure Mg. Journal of Alloys and Compounds, 1997. 253 254 : p. 550 555. 31. Huot, J., et al., Structural study and hydrogen sorption kinetics of ball milled magnesium hydride. Journal of Alloys and C ompounds, 1999. 295 : p. 495 500. 32. Liang, G., Synthesis and hydrogen storage properties of Mg based alloys. Journal of Alloys and Compounds, 2004. 370 (1 2): p. 123 128. 33. Charbonnier, J., et al., Hydrogenation of transition element additives (Ti, V) du ring ball milling of magnesium hydride. Journal of Alloys and Compounds, 2004. 383 (1 2): p. 205 208. 34. Schulz, R., et al., Recent developments in the applications of nanocrystalline materials to hydrogen technologies. Materials Science and Engineering A, 1999. 267 (2): p. 240 245. 35. Chou, K. C., et al., Kinetics of absorption and desorption of hydrogen in alloy powder. International Journal of Hydrogen Energy, 2005. 30 (3): p. 301 309. 36. Martin, M., et al., Absorption and desorption kinetics of hydrogen storage alloys. Journal of Alloys and Compounds, 1996. 238 (1 2): p. 193 201. 37. Levenspiel, O., Chemical Reaction Engineering. 2nd ed. New York: John Wiley, 1999: p. 359 370.

PAGE 184

184 38. Zeng, K., et al., Critical assessment and thermodynamic modeling of the Mg H system. International Journal of Hydrogen Energy, 1999. 24 (10): p. 989 1004. 39. Rudman, P.S., Hydrogen diffusion rate limited hydriding and dehydriding kinetics. Journal of Applied Physics, 1979. 50 (11): p. 7195 7199. 40. Bazzanella N., C.R., Miotello A ., D'Acapito F., Maurizio C., Barucca G., Majni G., Mengucci P., Kinetics of hydrogen absorption and desorption in magnesium: role of the structure and of catalysts". Advanced materials for energy conversion III Warrendale,2006 TMS annual meeting", San A ntonio, Texas, 2006: p. 89 99. 41. Noritake, T., et al., Chemical bonding of hydrogen in MgH2. Applied Physics Letters, 2002. 81 (11): p. 2008 2010. 42. Kelekar, R., et al., Formation and dissociation of MgH2 in epitaxial Mg thin films. Journal of Applied P hysics, 2007. 101 (11): p. 43. Brub, V., et al., Size effects on the hydrogen storage properties of nanostructured metal hydrides: A review 2007, John Wiley & Sons, Ltd. p. 637 663. 44. Palumbo, M., et al., Thermodynamic and ab initio investigation of the Al H Mg system. Calphad, 2007. 31 (4): p. 457 467. 45. Vajo, J.J. and G.L. Olson, Hydrogen storage in destabilized chemical systems. Scripta Materialia, 2007. 56 (10): p. 829 834. 46. Zeng, K.J., et al., Thermodynamic analysis of the hydriding process o f Mg Ni alloys. Journal of Alloys and Compounds, 1999. 283 (1 2): p. 213 224. 47. Shao, H.Y., et al., Hydrogen storage properties of magnesium ultrafine particles prepared by hydrogen plasma metal reaction. Materials Science and Engineering B Solid State Ma terials for Advanced Technology, 2004. 110 (2): p. 221 226. 48. Zaluska, A., L. Zaluski, and J.O. Strom Olsen, Synergy of hydrogen sorption in ball milled hydrides of Mg and Mg2Ni. Journal of Alloys and Compounds, 1999. 289 (1 2): p. 197 206. 49. Bogdanovic, B., et al., Thermodynamic investigation of the magnesium hydrogen system. Journal of Alloys and Compounds, 1999. 282 (1 2): p. 84 92. 50. Andreasen, A., Hydrogenation properties of Mg Al alloys. International Journal of Hydrogen Energy, 2008. 33 (24): p. 74 89 7497. 51. Yermakov, A.Y., et al., Hydrogen reaction kinetics of Mg based alloys synthesized by mechanical milling. Journal of Alloys and Compounds, 2006. 425 (1 2): p. 367 372. 52. Baldi, A., et al., Destabilization of the Mg H System through Elastic Con straints. Physical Review Letters, 2009. 102 (22): p. 226102. 53. Zhdanov, V.P., A. Krozer, and B. Kasemo, Kinetics of first order phase transitions initiated by diffusion of particles from the surface into the bulk. Physical Review B, 1993. 47 (17): p. 1104 4. 54. Fossdal, A., et al., Determination of the crystal structure of Mg(AlH4)2 by combined X ray and neutron diffraction. Journal of Alloys and Compounds, 2005. 387 (1 2): p. 47 51. 55. Fichtner, M., O. Fuhr, and O. Kircher, Magnesium alanate -a material f or reversible hydrogen storage? Journal of Alloys and Compounds, 2003. 356 357 : p. 418 422.

PAGE 185

185 56. Song, Y., Z.X. Guo, and R. Yang, Influence of selected alloying elements on the stability of magnesium dihydride for hydrogen storage applications: A first prin ciples investigation. Physical Review B, 2004. 69 (9): p. 094205. 57. Shang, C.X., et al., Mechanical alloying and electronic simulations of (MgH2+M) systems (M=Al, Ti, Fe, Ni, Cu and Nb) for hydrogen storage. International Journal of Hydrogen Energy, 2004. 29 (1): p. 73 80. 58. Liang, G. and R. Schulz, The reaction of hydrogen with Mg Cd alloys prepared by mechanical alloying. Journal of Materials Science, 2004. 39 (5): p. 1557 1562. 59. Kalisvaart, W.P., R.A.H. Niessen, and P.H.L. Notten, Electrochemical hyd rogen storage in MgSc alloys: A comparative study between thin films and bulk materials. Journal of Alloys and Compounds, 2006. 417 (1 2): p. 280 291. 60. Tsuda, M., et al., Mg H dissociation of magnesium hydride MgH2 catalyzed by 3d transition metals. Thin Solid Films, 2006. 509 (1 2): p. 157 159. 61. Stapmfer J.F., H.C.E., Suttle J.F. The Magensium hydrogen system. J. Am. Chem. Soc. 1960. 64 : p. 703 7 4. 62. Vigeholm, B., J. Kjler, and B. Larsen, Magnesium for hydrogen storage. Journal of the Less Comm on Metals, 1980. 74 (2): p. 341 350. 63. Pedersen, A.S., et al., Magnesium for hydrogen storage. International Journal of Hydrogen Energy, 1983. 8 (3): p. 205 211. 64. Cerm ic J. and L. Kr ai Hydrogen diffusion in Mg H and Mg Ni H alloys. Acta Materialia, 20 08. 56 (12): p. 2677 2686. 65. Evard, E.A., I.E. Gabis, and A.P. Voyt, Study of the kinetics of hydrogen sorption and desorption from titanium. Journal of Alloys and Compounds, 2005. 404 406 : p. 335 338. 66. Andreasen, A., T. Vegge, and A.S. Pedersen, Compe nsation Effect in the Hydrogenation/Dehydrogenation Kinetics of Metal Hydrides. The Journal of Physical Chemistry B, 2005. 109 (8): p. 3340 3344. 67. Nishimura, C., M. Komaki, and M. Amano, Hydrogen permeation through magnesium. Journal of Alloys and Compou nds, 1999. 293 295 : p. 329 333. 68. Fern an dez, J.F. and C.R. S chez, Rate determining step in the absorption and desorption of hydrogen by magnesium. Journal of Alloys and Compounds, 2002. 340 (1 2): p. 189 198. 69. Christian, J.W., The theory of transformation in metals and alloys Part I London: Pergamon Press. 2002: p. 529. 70. Karty, A., J. Grunzweig Genossar, and P.S. Rudman, Hydriding and dehydriding kinetics of Mg in a Mg/Mg[sub 2]Cu eutectic alloy: Pressure sweep method. Journal of Applied Physics, 1979. 50 (11): p. 7200 7209. 71. Fernendez, J.F. and C.R. Sanchez, Rate determining step in the absorption and desorption of hydrogen by magnesium. Journal of Alloys and Compounds, 2002. 340 (1 2): p. 189 198. 72. Friedrichs, O., et al., Chemical and microstructural study of the oxygen passivation be haviour of nanocrystalline Mg and MgH2. Applied Surface Science, 2006. 252 (6): p. 2334 2345. 73. Schlapbach, L., et al., Surface effects and the formation of metal hydrides. Journal of the Less Common Metals, 1980. 73 (1): p. 145 160.

PAGE 186

186 74. Fleisch, T. and R. Abermann, On the reduction of Ag2S films by hydrogen spillover under ultra high vacuum conditions. Journal of Catalysis, 1977. 50 (2): p. 268 278. 75. Charlot, M. F. and O. Kahn, Activation de l'hydrog e mol ulaire par une petite particule de nickel. Surface Science, 1979. 81 (1): p. 90 108. 76. Higuchi, K., et al., Remarkable hydrogen storage properties in three layered Pd/Mg/Pd thin films. Journal of Alloys and Compounds, 2002. 33 0 332 : p. 526 530. 77. Janot, R., et al., Enhancement of hydrogen storage in MgNi by Pd coating. Journal of Alloys and Compounds, 2003. 356 357 : p. 438 441. 78. Gremaud, R., et al., Ti catalyzed Mg(AlH4)2: A reversible hydrogen storage material. Journal of Alloys and Compounds, 2005. 404 406 : p. 775 778. 79. Liang, G. and R. Schulz, Synthesis of Mg Ti alloy by mechanical alloying. Journal of Materials Science, 2003. 38 (6): p. 1179 1184. 80. Liang, G., et al., Hydrogen storage properties of the mechanically milled MgH2 V nanocomposite. Journal of Alloys and Compounds, 1999. 291 (1 2): p. 295 299. 81. Xie, L., et al., Superior hydrogen storage kinetics of MgH2 nanoparticles doped with TiF3. Acta Materialia, 2007. 55 (13): p. 4585 4591. 82. Aguey Zinsou, K.F., et al., Effect of Nb2O5 on MgH2 properties during mechanical milling. International Journal of Hydrogen Energy, 2007. 32 (13): p. 2400 2407. 83. Oelerich, W., T. Klassen, and R. Bormann, Comparison of the catalytic effects of V, V2O5, VN, and VC on the hydrog en sorption of nanocrystalline Mg. Journal of Alloys and Compounds, 2001. 322 (1 2): p. L5 L9. 84. Barkhordarian, G., T. Klassen, and R.i. Bormann, Fast hydrogen sorption kinetics of nanocrystalline Mg using Nb2O5 as catalyst. Scripta Materialia, 2003. 49 (3 ): p. 213 217. 85. Friedrichs, O., et al., Hydrogen sorption improvement of nanocrystalline MgH2 by Nb2O5 nanoparticles. Scripta Materialia, 2006. 54 (7): p. 1293 1297. 86. Friedrichs, O., et al., MgH2 with Nb2O5 as additive, for hydrogen storage: Chemical, structural and kinetic behavior with heating. Acta Materialia, 2006. 54 (1): p. 105 110. 87. Bogdanovic, B., et al., Ni doped versus undoped Mg MgH2 materials for high temperature heat or hydrogen storage. Journal of Alloys and Compounds, 1999. 292 (1 2): p 57 71. 88. Jeon, K. J., A. Theodore, and C. Y. Wu, Enhanced hydrogen absorption kinetics for hydrogen storage using Mg flakes as compared to conventional spherical powders. Journal of Power Sources, 2008. 183 (2): p. 693 700. 89. Mintz, M.H. and Y. Zeiri, Hydriding kinetics of powders. Journal of Alloys and Compounds, 1995. 216 (2): p. 159 175. 90. Chernov, I., et al., Influence of metal powder particle's shape on the kinetics of hydriding. International Journal of Hydrogen Energy. 35 (1): p. 253 258. 91. Ce rmak, J. and L. Kral, Hydrogen diffusion in Mg H and Mg Ni H alloys. Acta Materialia, 2008. 56 (12): p. 2677 2686. 92. Vittori Antisari, M., et al., On the nucleation step in the Mg MgH2 phase transformation. Scripta Materialia, 2009. 61 (11): p. 1064 1067.

PAGE 187

187 93. Montone, A., et al., On the barriers limiting the reaction kinetics between catalysed Mg and hydrogen. 63 (4): p. 456 459. 94. Oguchi, H., et al., In situ infrared imaging methodology for measuring heterogeneous growth process of a hydride phase. Intern ational Journal of Hydrogen Energy. 35 (3): p. 1296 1299. 95. Baldi, A., et al., Mg Ti H thin films as switchable solar absorbers. International Journal of Hydrogen Energy, 2008. 33 (12): p. 3188 3192. 96. Baldi, A., et al., Nanoscale composition modulations in MgyTi1 yHx thin film alloys for hydrogen storage. International Journal of Hydrogen Energy, 2009. 34 (3): p. 1450 1457. 97. Voort, G.V., Metallography and Microstructures. ASM Handbook, 2004. 9 : p. 1184. 98. Ebrahimi F, T.M., Tatiparti S, Tien H.Y., Jeo n K.J., Wu C.Y. A study of hydrogen absorption and desorption in magneisum. MRS Fall, 2007(Symposium S: MAterial and technolodgy for hydrogen storage). 99. Tien, H. Y., et al., Mechanism of hydrogen capacity dependence on the hydrogenation temperature. Scripta Materialia. 62 (5): p. 274 277. 100. Tien, H. Y., et al., Effect of hydride nucleation rate on the hydrogen capacity of Mg. International Journal of Hydrogen Energy, 2009. 34 (15): p. 6343 6349.

PAGE 188

188 BIOGRAPHIAL SKETCH Hung Yu Tien was b orn in 1979, in Taitung City, Taitung, Taiwan. He did his of National Tsing Hua University in Hsin study is bulk metallic glass (B MG). He has one year working experience as a 4G Flash integrate engineer in a semiconductor company, Power Chip Semiconductor Corporation in Hsin Materials Science and Engineering Department in summer 2007.