<%BANNER%>

Enhanced Hydrogen Uptake and Release Kinetics and Capacity for Magnesium Nanocatalyst Composites

Permanent Link: http://ufdc.ufl.edu/UFE0021323/00001

Material Information

Title: Enhanced Hydrogen Uptake and Release Kinetics and Capacity for Magnesium Nanocatalyst Composites
Physical Description: 1 online resource (144 p.)
Language: english
Creator: Jeon, Ki Joon
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: absorption, desorption, diffusion, hsobm, interfacial, kinetics, mg, mgh2, ni
Environmental Engineering Sciences -- Dissertations, Academic -- UF
Genre: Environmental Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The objective of this study was to investigate the hydrogen absorption kinetics and capacity of Mg-nano Ni composites. A dry particle coating technique (Theta Composer) was utilized for coating magnesium particles with nano-Ni. Subsequently, hydrogen absorption and desorption kinetics were evaluated by varying system parameters (i.e. coating time, catalyst loading, coating speed and heating rate). Hydrogen absorption curves plotted as a function of time showed that composites processed for longer periods of time exhibited significantly higher absorption rates. With increased coating time, the catalyst was more evenly distributed over the Mg surface, resulting in a product with increased hydrogen capacity and kinetics. A change in the rate limiting mechanism from the interfacial growth of MgH2 to diffusion into the Mg particle was observed. SEM/TEM characterization verified that magnesium hydride forms from the surface rather than from random nucleation and growth in Mg. An analytical model based upon the shrinking core concept was developed to evaluate the dependence of hydrogen absorption on two important rate limiting mechanisms: diffusion and the interfacial growth of MgH2. Hydrogen absorption capacity was expressed as a function of temperature, hydrogen pressure and Mg particle size at the given hydrogenation time. The analytical solution agreed very well with the experimental data for magnesium hydride formation. A high speed orbiting ball media (HSOBM) processor was developed and utilized to fabricate flake-shaped materials (thin metal flakes with large diameters). Several aspects of flake characteristics produced by the process were studied, including: flake diameter, thickness, morphology (as a function of processing time), ball media count and weight loading. Improvement in hydrogen storage kinetics and capacity of Mg by this novel approach utilizing high aspect ratio powders coated with Ni nanocatalysts was evaluated. The Mg flakes were fabricated by the HSOBM process and were coated with Ni nanocatalysts using the Theta Composer. The flakes thus produced possess more favorable hydrogen absorption/desorption characteristics and improved hydrogen storage capacity than spherical particles. Hydrogen absorption kinetics was identified to be more sensitive to variations in geometric shape as opposed to changes in grain size.
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 Ki Joon Jeon.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Wu, Chang-Yu.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021323:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021323/00001

Material Information

Title: Enhanced Hydrogen Uptake and Release Kinetics and Capacity for Magnesium Nanocatalyst Composites
Physical Description: 1 online resource (144 p.)
Language: english
Creator: Jeon, Ki Joon
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: absorption, desorption, diffusion, hsobm, interfacial, kinetics, mg, mgh2, ni
Environmental Engineering Sciences -- Dissertations, Academic -- UF
Genre: Environmental Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The objective of this study was to investigate the hydrogen absorption kinetics and capacity of Mg-nano Ni composites. A dry particle coating technique (Theta Composer) was utilized for coating magnesium particles with nano-Ni. Subsequently, hydrogen absorption and desorption kinetics were evaluated by varying system parameters (i.e. coating time, catalyst loading, coating speed and heating rate). Hydrogen absorption curves plotted as a function of time showed that composites processed for longer periods of time exhibited significantly higher absorption rates. With increased coating time, the catalyst was more evenly distributed over the Mg surface, resulting in a product with increased hydrogen capacity and kinetics. A change in the rate limiting mechanism from the interfacial growth of MgH2 to diffusion into the Mg particle was observed. SEM/TEM characterization verified that magnesium hydride forms from the surface rather than from random nucleation and growth in Mg. An analytical model based upon the shrinking core concept was developed to evaluate the dependence of hydrogen absorption on two important rate limiting mechanisms: diffusion and the interfacial growth of MgH2. Hydrogen absorption capacity was expressed as a function of temperature, hydrogen pressure and Mg particle size at the given hydrogenation time. The analytical solution agreed very well with the experimental data for magnesium hydride formation. A high speed orbiting ball media (HSOBM) processor was developed and utilized to fabricate flake-shaped materials (thin metal flakes with large diameters). Several aspects of flake characteristics produced by the process were studied, including: flake diameter, thickness, morphology (as a function of processing time), ball media count and weight loading. Improvement in hydrogen storage kinetics and capacity of Mg by this novel approach utilizing high aspect ratio powders coated with Ni nanocatalysts was evaluated. The Mg flakes were fabricated by the HSOBM process and were coated with Ni nanocatalysts using the Theta Composer. The flakes thus produced possess more favorable hydrogen absorption/desorption characteristics and improved hydrogen storage capacity than spherical particles. Hydrogen absorption kinetics was identified to be more sensitive to variations in geometric shape as opposed to changes in grain size.
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 Ki Joon Jeon.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Wu, Chang-Yu.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021323:00001


This item has the following downloads:


Full Text

PAGE 1

1 ENHANCED HYDROGEN UPTAKE AND R ELEASE KINETICS AND CAPACITY FOR MAGNESIUM NANOCATALYST COMPOSITES By KI-JOON JEON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Ki-Joon Jeon

PAGE 3

3 To my parents, Gab-Soo Jeon and Young-H ee Lee; and my younger sister, Eun-Jung

PAGE 4

4 ACKNOWLEDGMENTS I would like to express my sincere gratitude to Dr. Chang-Yu Wu for his support and guidance in my Ph.D. study. His advice and encouragement were invaluable for my graduate study. My special thanks go to my committee memb ers, Dr. Fereshteh Ebrahimi, Dr. Mei Cai, and Dr. Ben Koopman for their interest and expert advice towards my research. I am grateful to the University of Florida Alumni Fellowship and Korea Research Foundation Scholarship (M062003-000-10264-0). I would like to acknowledge Major Analytical Instrument Center (MAIC) and Particle Engineering Research Center (PE RC) at University of Florida for analytical instrumentation, as well as Dr. Satoru Watano at Osaka Prefecture University for his assistance. I am also grateful to Tokuju Corp. for the use of Theta Composer. I would like to thank our past and present group members. Special thanks go to Alex, Bob, Srikanth, Shankara and Mahesh. Finally I would like to give my sincere grat itude to my family. W ithout their care and encouragement, I could not have achieved this goal.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .........9 ABSTRACT....................................................................................................................... ............11 1 GLOBAL INTRODUCTION.................................................................................................13 Transition Metal Catalyst...................................................................................................... .14 Mechanical Ball Milling........................................................................................................ .15 Mechanism of MgH2 Formation.............................................................................................16 Models for MgH2 Formation..................................................................................................17 Objectives..................................................................................................................... ..........18 2 HYDROGEN ABSORPTION/DESORPTION KINETICS OF MAGNESIUM NANONICKEL COMPOSITES SYNTHESIZED BY DRY PARTICLE COATING TECHNIQUE...................................................................................................................... ...20 Introduction................................................................................................................... ..........20 Experimental Methodology....................................................................................................21 Dry Particle Coating........................................................................................................21 Hydrogen Absorption Test..............................................................................................22 Particle Characterization.................................................................................................23 Data Analysis.................................................................................................................. .23 Results and Discussion......................................................................................................... ..24 Particle Morphology & Texture Change.........................................................................24 Effects of Coating Time and Speed on Hydrogen Absorption........................................25 Effects of Ni Loading and Ga p Size on Hydrogen Absorption.......................................27 Effect of Heating Rate on Hydrogen Absorption............................................................28 Hydrogen Desorption Rates and Capacity......................................................................28 Summary........................................................................................................................ .........29 3 MODEL FOR HYDROGEN ABSORPTION KI NETIS AND CAPACITY IN METAL HYDRIDES....................................................................................................................... .....43 Introduction................................................................................................................... ..........43 Model Development.............................................................................................................. .45 Absorption Kinetics.........................................................................................................46 Diffusion...................................................................................................................46 Interfacial Growth....................................................................................................47 Overall Process.........................................................................................................47 Solution to Rate Limiting Condition...............................................................................48

PAGE 6

6 Interfacial Growth Limited Case..............................................................................48 Diffusion Limited Case............................................................................................49 Comparisons with Equations Reporte d in the Literature (constant CH2).................49 Overall Equation with Two Mechanisms........................................................................50 Experimental................................................................................................................... ........51 Characterization of Hydrogen Absorption Kinetics........................................................51 Transmission Electron Microscopy.................................................................................52 Presence of MgH2 Layer near Surface............................................................................52 Determination of Diffusion and Interfacial Growth Coefficients for Hydrogenation.....53 Comparative Study of Experime ntal and Calculated Data..............................................55 Effect of Temperature and Initial H2 Pressure on Hydrogen Absorption Kinetics.........55 Effect of Mg Particle Diameter on Hydrogen Absorption Kinetics................................56 Summary........................................................................................................................ .........57 Nomenclature................................................................................................................... .......58 4 FLAKE PARTICLE SYNTHESIS FROM DU CTILE METAL PARTICLES USING A NOVEL HIGH-SPEED VI BRATORY MILL.......................................................................67 Background..................................................................................................................... ........67 Mechanisms..................................................................................................................... .......69 High Speed Orbiting of Ball Media.................................................................................69 Plastic Deformation and Welding Effects.......................................................................70 Experimental................................................................................................................... ........70 Experimental System.......................................................................................................70 Product Characterization.................................................................................................71 Results and Discussions........................................................................................................ ..72 300 m Mg Particle.........................................................................................................72 140 m Fe Particle...........................................................................................................73 Effect of Ball Size...........................................................................................................75 Effect of Weight Loading................................................................................................76 Summary........................................................................................................................ .........77 5 ENHANCED HYDROGREN ABSOPRTI ON KINETICS FOR HYDROGEN STORAGE USING MG FLAKES FABRICA TED BY A HIGH SPEED ORBITING BALL MEDIA PROCESSOR................................................................................................88 Background..................................................................................................................... ........88 Experimental Methodology....................................................................................................91 Fabrication of Magnesium Fl akeNano Nickel Composites...........................................91 Hydrogen Absorption/Desorption Characterization........................................................93 Results and Discussion......................................................................................................... ..94 Determination of Minimum Ni Loading.........................................................................94 Characterization of Mg Flakes (SEM/BET/XRD)..........................................................95 Effect of Grain Size on the Hydrogen Ab sorption Kinetics without Nano Catalyst.......96 Analysis of Coating Methods and Change of Thickness.................................................97 Effect of Dispersion of Nano Cataly sts on Hydrogen Absorption Kinetics....................98 Effect of Geometric Change on Hydrogen Absorption Kinetics.....................................98

PAGE 7

7 Summary........................................................................................................................ .......100 Nomenclature................................................................................................................... .....101 6 CONCLUSIONS AND RECOMMENDATIONS...............................................................111 APPENDIX EFFECTS OF PRIMARY PARTICLE SIZE AND INITIAL GEOMETRIC STANDARD DEVIATION ON BROWNI AN COAGULATION OF FRACTAL NANOAGGLOMERATES IN THE FREE MOLECULAR REGIME...............................114 REFERENCES..................................................................................................................... .......134 BIOGRAPHICAL SKETCH.......................................................................................................144

PAGE 8

8 LIST OF TABLES Table page 2-1 Comparison of calculated hydrogen upt ake with hydrogen desorption by TGA..............31 4-1 Experimental conditions....................................................................................................79 5-1 Experimental conditions..................................................................................................102

PAGE 9

9 LIST OF FIGURES Figure page 2-1 Theta Composer............................................................................................................. ....32 2-2 Hydrogenation system.......................................................................................................33 2-3 Surface morphology change (Rotor speed: 4600 rpm, Gap Size: 1 mm, Mg-2 wt % Ni)............................................................................................................................ ..........34 2-4 Distribution of Ni nano particles on Mg (Rotor speed : 4600 rpm, Gap Size: 1 mm, Mg-2 wt. % Ni)................................................................................................................. .35 2-5 Particle texture change (Rotor speed: 4600 rpm, Gap Si ze: 1 mm, Mg-2 wt. % Ni)........36 2-6 XRD patterns for MgH2 -2 wt % Ni Composites (Ini tial hydrogen pressure: 145 psi, Hydrogenation time: 300 min, Temperature: 483K).........................................................37 2-7 Effect of coating time on H2 absorption (Rotor speed: 4600 rpm, Gap Size: 1 mm, Mg-2 wt. % Ni, Initial hydrogen pressu re: 145 psi, Hydrogenation time: 300 min, Temperature: 483K)...........................................................................................................38 2-8 Profile for hydrogenation runs with material Mg-2 wt. % Ni (90 min).............................39 2-9 Effect of Ni loading &gap size on H2 absorption..............................................................40 2-10 Effect of heating rate on H2 absorption (Rotor speed: 4600 rpm, Ni wt. %: 2 %, Coating time: 90 min, Gap size: 1 mm).............................................................................41 2-11 Hydrogen desorption analysis by TGA..............................................................................42 3-1 Schematic diagram of the shrinking core model................................................................59 3-2 Schematic diagram of the hydrogenation system..............................................................60 3-3 SEM/TEM micrographs and electron diffraction patterns obtained for the hydrogenated Mg-Ni..........................................................................................................61 3-4 Change in hydrogen pressure as a function of time (44 m Mg + 2 wt. % Ni, initial H2 pressure = 1 MPa).........................................................................................................62 3-5 Determination of hydrogen absorption co efficient following the Arrhenius form............63 3-6 Comparison of measured and mode l data (Simulation condition: 44 m Mg + 2 wt. % Ni, initial H2 pressure = 1 MPa, temperature = 483 K).....................................................64 3-7 Effect of temperature and initia l hydrogen pressure on hydrogenation.............................65

PAGE 10

10 3-8 Effect of Mg particle size on hydrogen absorption kinetics (simulation condition: Mg + 2 wt. % Ni, initial H2 pressure = 1.0 MPa, temperature = 483 K)..................................66 4-1 Schematic diagram of orbiting motion..............................................................................80 4-2 Schematic diagram of basi c orbiting media mechanisms..................................................81 4-3 Milling zone and m illing area clas sification......................................................................82 4-4 Particle size distributions for original particles.................................................................83 4-5 Particle size distributi ons for Mg particle (300 m)..........................................................84 4-6 Thickness of Mg particle as a function of milling time.....................................................85 4-7 Particle size distributi ons for iron particle (140 m) as a function of milling time...........86 4-8 Optical microscope images at 20x magnification of iron particle.....................................87 5-1 Schematic diagram of hydrogenation system..................................................................103 5-2 Surface morphology change............................................................................................104 5-3 Comparison of XRD patterns...........................................................................................105 5-4 Effect of morphological change on the hydrogen adsorpti on kinetics (w/o catalysts, experiment).................................................................................................................... ..106 5-5 Distribution of Ni nano particles on Mg..........................................................................107 5-6 Nano Ni uniformity measurements on the Mg flake at the two different coating method......................................................................................................................... .....108 5-7 Effect of dispersion of nano cat alyst on hydrogen absorption kinetics...........................109 5-8 Effects of particle geometry and co ating efficiency on the hydrogen absorption kinetics....................................................................................................................... ......110

PAGE 11

11 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ENHANCED HYDROGEN UPTAKE AND R ELEASE KINETICS AND CAPACITY FOR MAGNESIUM NANOCATALYST COMPOSITES By Ki-Joon Jeon August 2007 Chair: Chang-Yu Wu Major: Environmental Engineering Sciences The objective of this study was to investigate the hydr ogen absorption kinetics and capacity of Mg-nano Ni composites. A dry particle coating technique (Theta Composer) was utilized for coating magnesium particles with nano-Ni. Subsequently, hydrogen absorption and desorption kinetics were evalua ted by varying system parameters (i.e. coating time, catalyst loading, coating speed and heating rate). Hydrogen absorption curves plotted as a function of time showed that composites processed for longer periods of time exhibited significantly higher absorption rates. With increased coating time, th e catalyst was more evenly distributed over the Mg surface, resulting in a produc t with increased hydrogen capacity and kinetics. A change in the rate limiting mechanism from the interfacial growth of MgH2 to diffusion into the Mg particle was observed. SEM/TEM characterization verified that magne sium hydride forms from the surface rather than from random nucleation and growth in Mg. An analytical model based upon the shrinking core concept was developed to evaluate the dependence of hydrogen absorption on two important rate limiting mechanisms: diffu sion and the interfaci al growth of MgH2. Hydrogen absorption capacity was expressed as a functi on of temperature, hydrogen pressure and Mg

PAGE 12

12 particle size at the given hydrogenation time. The analytical solution agreed very well with the experimental data for magnesium hydride formation. A high speed orbiting ball media (HSOBM) pr ocessor was developed and utilized to fabricate flake-shaped materials (thin metal flakes with large diameters). Several aspects of flake characteristics produced by the process were st udied, including: flake diameter, thickness, morphology (as a function of processing time) ball media count and weight loading. Improvement in hydrogen storage kinetics and cap acity of Mg by this novel approach utilizing high aspect ratio powders coated with Ni nano catalysts was evaluated. The Mg flakes were fabricated by the HSOBM process and were co ated with Ni nanocatalysts using the Theta Composer. The flakes thus produced possess more favorable hydrogen absorption/desorption characteristics and improved hydrogen storage ca pacity than spherical particles. Hydrogen absorption kinetics was identified to be more sensitive to variations in geometric shape as opposed to changes in grain size.

PAGE 13

13 CHAPTER 1 GLOBAL INTRODUCTION Over the past decade, the looming energy crisis and the concern about the effects of fossil fuel emissions on global warming have crea ted a necessity for cleaner energy options (Dresselhaus and Thomas 2001). Am ong these, hydrogen has received particular attention as a potential solution to future energy needs (DOE 2003). The use of hydrogen as an alternative fuel in fuel cells has been of interest to many appl ications, from mobile electronics to aerospace industries (DOE 2007; NASA 2006; Schlapbach and Zttel 2001). Hydrogen fuel cells have the potential to revolutionize the way we power the world by offe ring cleaner and more-efficient alternatives to gasoline and other fossil fuels (DOE 2007a). The primary benefit of using pure hydrogen as a power source is that it produces only water vapor and waste heat as exhaust ( DOE 2007a; Schlapbach and Zttel 2001). Pure hydrogen can be obtained from the decomposition of methane, coal, liquid petroleum or biomass, via thermolysis, or from the electrolysis of water (DOE 2007a). However, providing a safe, efficient, reliable, and continuous source of hydr ogen to a fuel cell st ill remains a significant technological obstacle. For light duty automotive vehicles using fuel cells with a driving range greater than 300 miles required by DOE (DOE 2007a), on-board h ydrogen storage is essential. The primary problem with storage in pure form is that hydrogen has a very low volumetric/gravimetric storage density (Schlapbach and Z ttel 2001; Zttel 2004). If hydroge n is stored as a liquid in a cryogenic tank or as a gas in a pressurized tank to counter this problem, the amount of energy that can be stored in the available space on board is small in comparison to gasoline. The energy required to compress the gas is also a serious drawback (Zttel 2004). These issues make hydrogen cars prone to a limited driving range compared to their c onventional counterparts

PAGE 14

14 (DOE 2007a; Zttel 2004). Thus, it is essentia l that new hydrogen storage methods possess a high hydrogen storage capacity. The highest known volumetric storage densi ties of hydrogen are found in light metal hydrides and alloys (Schlapbach and Zttel 2001; Zttel 2004). Among the various light metals and alloys that are capable of absorbing/de sorbing large amounts of hydrogen, magnesium is considered a promising candidate for so lid hydrogen storage due to its high volumetric/gravimetric capacity, abundance and low cost (Chalk and Miller 2006; Schlapbach and Zttel 2001; Zttel 2004). The major disadvantages of utiliz ing pure magnesium, however, are its slow reaction kinetics and the formati on of a dense magnesium hydride which impedes prolonged reactions (Fukai 2005; Schlapbach and Ztte l 2001). Such critical issues need to be overcome before its practical ap plications can be realized. Transition Metal Catalyst Various attempts have been undertaken to chemically improve the kinetics of Mg by utilizing transition metal catalysts such as Ni, V, Fe and Ti (Holtz and Imam 1999; Huot et al. 1999; Iwakura et al. 2002; Iwakur a et al. 1999; Jeon et al. 2006; Liang et al. 1998; Varin et al. 2006; Zaluska et al. 1999). The addition of catalysts with high affinities for chemisorption results in H2 dissociation into hydrogen atoms at the in terface of the Mg-Ni composite by lowering the corresponding activation energy (S androck et al. 2002; Stillesjo et al. 1993). These hydrogen atoms can diffuse effectively along the grain boundaries of magnesium (Lueking and Yang 2004; Rozanov and Krylov 1997). By adding a small amount of Ni onto Mg surface (Holtz and Imam 1999), it has been demonstrated that the onset temperature for hydrogen absorption can simply be lowered by approximately 170C. Several factors have been identified to be influential in hydrogen absorption and desorption kinetics. Holtz et al (1999) highli ghted the importance of the effectiveness of

PAGE 15

15 dispersion of catalyst over the su rface of magnesium particles. Se veral studies have reported that a homogeneous distribution of nano catalysts on the metal drastically reduces the activation energy of hydrogen diffusion and results in more rapid formation of metal hydride even in the presence of oxides and hydroxides (Au 2005; Huot et al. 2001; Jeon et al. 2007; Seayad and Antonelli 2004; Zaluska et al 2001). Alternatively, it should be noted that when hydrogen diffusion was the rate limiting mechanism, Ni wt. % loading showed negligible impact on H2 absorption rate or capac ity (Jeon et al. 2006). The particle diameter of the catalyst has also be en identified as another influential variable. Varin et al. (2006) investigated the effect of Ni particle size (micro, submicron and nano sizes) on the performance of hydrogen ab sorption/desorption kinetics of 44 m Mg powder. It was reported that the addition of nano Ni greatly improves the hydrogen absorption/desorption rate compared to micron/submicron size catalysts. Mechanical Ball Milling Mechanical milling techniques that aim at increasing surface defects and grain boundaries of metal powders enhance diffusion of hydrogen at oms within the metal, leading to increased hydride formation (Doppiu et al. 2004; Holtz and Im am 1999; Huot et al. 1995; Huot et al. 1999; Iwakura et al. 2002; Liang et al. 1998; Revesz et al.; Tessier and Akiba 1999; Varin et al. 2006; Yu et al. 2002; Zaluska et al. 1999). A significant portion of previ ous research in this approach has focused on conventional mechanical ball milling, vibratory ball milling, and wet milling to produce changes in the microstruc ture of magnesium particles. The major disadvantages of mechanical ball milling are the prolonged milling times (3060 hrs) (Doppiu et al. 2004; Huot et al. 1995; Tessier and Akiba 1999; Yu et al. 2002) and formation of undesirable alloys. Varin et al. (2006) reported that after 20 hrs of milling, formation of Mg2Ni alloy was observed, which

PAGE 16

16 resulted in lowered hydrogen absorption capacity. In addition, it is chal lenging to control the shape and qualitative consistency of the product (Hong and Kim 2001; Hong et al. 2000; Yoshinaga et al. 1997). Therefore, the need aris es for a new milling system which is capable of reducing grain size without the formation of metal alloys and with significantly reduced operating times. Mechanism of MgH2 Formation There have been many experimental studies in recent years to understand the rate-limiting kinetics and absorption capacity of hydrogen st orage in alloys (Fuka i 2005; Holtz and Imam 1999; Liang et al. 1998; Liang et al. 1999; Orimo and Fujii 1996; Yang et al. 2000). Holtz and Imam (1999) proposed that nucleation of magne sium hydride on the surface of pure magnesium is the initial rate limiting step for absorption of H2, as initial hydrogenati on (i.e. nucleation of MgH2) takes place only at the active surface of magnesium. As hydrogenation progresses, the hydride layer (semi-impermeable to hydrogen) th ickens and expands into the bulk. Diffusion through the magnesium hydride la yer is many orders of magnitude slower than other mechanisms (i.e. nucleation of metal hydride or dissociation of hydrogen molecule into two hydrogen atoms) and subsequently becomes the ra te-limiting step in the hydrogenation process (Bloch and Mintz 1997; Friedlmeier and Groll 1 997; Sastri et al. 1998; Schweppe et al. 1997). However, these studies have not been experiment ally verified adequatel y. In fact, due to the complexity involved in the micro-structural anal ysis, very few microscopic studies have been done to investigate the formation of magnesium hydride in a particle. On the other hand, the homogeneous nucleation of magnesium hydride taking place in the bulk has been observed under optical micro-structural analys is (Douglass 1975). This contradi cts the previously described theories of hydride formation, where it is assumed that hydride forms as a continuous layer from the particle surface. Thus, th ere is a need for clarity.

PAGE 17

17 Models for MgH2 Formation Several empirical or theoretical kinetic models have also been formulated and subsequently fitted with experimental data for describing hydrogen absorption/desorption behavior of metal hydrides (Chou et al. 2005; Friedlmeier and Groll 1997; Ginstling and Brounshtein 1950; Rudman 1979; S aetre 2006). The nucleation and growth model developed by Avrami (Avrami 1939; Avrami 1940; Avrami 1941) is a frequently applied model for describing phase change in solids. In Avramis model, the rate-limiting mechanism is random nucleation and the subsequent three-dimensional growth of a new phase within the bulk. Rudman (1979) later extended Avramis model to metalhydrogen systems; the effects of temperature and hydrogen pressure on the kinetics for hydroge n diffusion-limited cases were considered, assuming a constant growth rate of magnesium hydride. The other commonly accepted model for metal hydride formation was proposed by Flanag an (1978), in which hydrogen molecules are adsorbed or weakly chemisorbed on an active metal surface where they then dissociate into hydrogen atoms. These atoms can easily diffuse between the grain boundar ies of the pure metal and then finally form metal hydride (Flanaga n 1978; von Zeppelin et al. 2002). If diffusion of hydrogen atoms occurs at a fast rate along the grain boundary (i.e. the grain size is small), a metal hydride layer is formed and thickens into a pure metal core (Zalus ka et al. 1999). This concept can be best described by the shri nking-core model (Levenspiel 1999). Several numerical studies have been c onducted based on the shrinking-core model, considering changes in growth rate for a diffusion limited case (Chou et al. 2005; Ginstli ng and Brounshtein 1950; Rudman 1979; Saetre 2006; Schweppe et al. 1997). Barkhordarian et al.(2006) reported that their experimental data of the absorption behavior of magnesium catalyzed with Nb2O5 were better describe d by a three-dimensional diffusion controlled shrinking-core model. Gabis et al.(2005) also compared Avramis model and

PAGE 18

18 the shrinking-core model with thermal desorp tion spectrometry (TDS) experiments of metal hydride, and the results showed th at the shrinking core model described the kinetics of hydrogen desorption better. However, predictions by bot h models still deviate from experimental observations, especially during th e initial period of the hydrogena tion process (Schweppe et al. 1997; Stander 1977). The reason for the deviation stems from the key assumptions of these models that the process is diffusion limited and that the change in rate limiting mechanisms during the process is not considered. Stander (1 977) divided kinetics of Mg hydride formation into two steps (nucleation and di ffusion) and likewise fitt ed his experimental data separately for each of these cases as a function of the fraction ( ) of metal hydride, and reported that the nucleation equation was valid for pur e Mg for an upper limit of 0.18 for (equivalent to a layer of 1.6 m MgH2 on a Mg particle diameter of 25 m). After the magnesium surface was utilized ( >0.18), a separate equation for diffusion was re ported to reproduce experimental data well. Objectives The objectives of this diss ertation research were to investigate hydrogen absorption mechanisms of a Mg-nano Ni composite and to assess the effects of nano catalyst and the geometry of Mg particles on enhancing hydroge n adsorption kinetics and capacity. Both experiments and theoretical modeling were carri ed out to help achieve the objectives. The dissertation is divided into four chapters (Chapt ers 2-5). The major content of each chapter is outlined in the following paragraphs: Chapter 2 The kinetics of Mg-Ni composites synthesi zed by a dry coating technique (Theta Composer) were studied. This provides insight in to the rate limiting mechanism change between interfacial growth and diffusion as well as the effects of various processing parameters on hydrogenation/desorption kinetics and capacities. The effects of heating rates on hydrogen absorption kinetics we re also explored.

PAGE 19

19 Chapter 3 The behavior of MgH2 formation in a Mg particle was experimentally identified via micro-structural analysis. A ki netic model based upon experiment al observations was developed to investigate the change in the rate-limiting mechanism and ki netics in hydrogen adsorption. The effects of magnesium partic le size, initial hydrogen pre ssure and temperature on hydrogen absorption kinetics were also assessed. Chapter 4 A novel milling processor (High Speed Orbiting Ball Milling HSOBM) was developed to efficiently prepare flake-shaped Mg particles. The milling mechanisms including plastic deformation, particle-parti cle-welding and flake deformation were examined. The effects of the processing parameters on the milling effici ency and product characteristics were studied. Chapter 5 Improvements in hydrogen storag e kinetics and capacity were evaluated through the use of high aspect rati o Mg powders (thin metal flakes with large diameters) coated with Ni nanocatalysts. The coating efficiencies of the HSOBM processo r and Theta Composer were assessed. The effects of geometry and grain size on hydrogen abso rption kinetics were analyzed.

PAGE 20

20 CHAPTER 2 HYDROGEN ABSORPTION/DESO RPTION KINETICS OF MAGNESIUM NANO-NICKEL COMPOSITES SYNTHESIZED BY DRY PARTICLE COATING TECHNIQUE Introduction Magnesium hydride has been widely studied as a means for effective hydrogen storage due to its several advantages such as light weight, high hydrogen capacity, natural abundance, low environmental impact and low cost (Schlapbach and Zttel 2001). One of the critical problems which limit the wide usage of Mg for solid stor age is its absorption and desorption kinetics, which are only feasible at high temperatures (> 650K). To overcome these limitations, past research has attempted to improve the kine tics through ball milling, hydriding combustion synthesis of the bulk magnesium metal, alloying w ith transition metal catalysts (i.e. Al, Ti, Cr, Fe, Co, Mn, Zr, Zn and Ge ) and destabilizing the magnesium-transition metal-hydrogen matrix (Gennari et al. 2002; Li et al. 2000; Liang et al. 1998; Liang et al. 1999; Raman et al. 2002; Varin et al. 2005; Yang et al. 2000). The role of the catalytic transition metal is to dissociate H2 into H atoms that can be more readily chemis orbed by magnesium (Friedlmeier and Groll 1997; Holtz and Imam 1999; von Zeppelin et al. 2002), thus improving hydrogen absorption and desorption kinetics. The distribut ion and dispersion of the catalys t over the surface of magnesium particle is a critical factor in affecting hydroge n absorption and desorption kinetics (Zaluska et al. 1999). With the help of the catalyst, the formation of MgH2 can be achieved and chemisorption becomes the rate limiting mechanism in the initial stage. However, as the reaction proceeds, an impermeable hydride layer develops and diffusi on through the layer becomes the rate limiting mechanism (Friedlmeier and Groll 1997; Holtz a nd Imam 1999). To maximize the kinetics, the duration of the first limiting mechanism s hould be kept as long as possible. In addition to adding transition metal catalys ts, the hydrogenation rate over those of the bulk material can also be improved by mechanical milling or ball milling. This changes the Mg

PAGE 21

21 microstructure by creating more defects and grai n boundaries near the surface as well as making the nanocrystalline structure amorphous (Orimo and Fujii 1996; Zaluska et al. 1999). These fissures serve as passages for hydrogen diffusi on, along with improving hydride nucleation and hydrogen atom penetration (Song 1995). Ball milling, the common technique for alloy preparation, requires processing times ranging from a few hours to 30 hours for complete milling (Charbonnier et al. 2004; Chen and Williams 1995; Gao et al. 2005; Lee et al. 2002; Spassov et al. 2005; Yoshinaga et al. 1997). Dry particle co ating can be an alternative to ball milling because of its ability to disperse nanopartic les in 60-90 minutes of time duration (Coowanitwong et al. 2003; Pfeffer et al. 2001). The objectives of this chapter were to determine the absorption/desorption characteristics and hydrogen storage capacity of Ni nanocatalysts coated magnesi um prepared by dry particle coating. The following were investigated: 1) The optimal processing conditions of the dry particle coating process (Ni loading, coating time, coating speed, gap size and heating rate) fo r hydrogen storage 2) The change in the rate-limiti ng mechanism in hydrogen absorption 3) The effects of heating rate on hydrogen absorption kinetics/capacity Experimental Methodology Dry Particle Coating Mg-Ni composites were prepared by Theta Composer (Tokuju, Corp.), a dry coating equipment. A schematic diagram of Th eta Composer is shown in Figure 2-1 Ten grams of pure Mg powder (Alfa Aesar 44 m, 325 mesh) and Ni nanoparticles (Argonide 70 nm) were introduced into the vessel while under an argon environment (4.8 grade) in a glove box (PlasLabs, Inc. Model 855-AC). Because of the potenti al reactivity of Mg w ith oxygen in air, an oxygen analyzer (Cambridge Sensotec, Ltd. Rapi dox 3000) was used to monitor the oxygen level

PAGE 22

22 in the glove box and pure argon was constantly supplied to the glove box to ensure the oxygen level was lower than 500 ppm. Two different weight % of Ni nano particles were tested with two coating times of 60 and 90 min. The coating system was tested with rotor speeds of 3600 and 4600 rpm and a fixed vessel speed of 70 rpm. Two rotor-vessel gap sizes of 1.0 mm and 0.5 mm were used. The rotor operates at high revolution while the vessel coun ter-rotates at a lower speed. Strong compression and shear forces are applied to the particles wh en they pass through the narrow gap between the rotor and vessel wall (Pfeffer et al. 2001). The roto rs and vessel are made of stainless steel in order to maintain an inert environment and prevent chemical reactions. Smaller gap sizes increase the forces on the powders and may affect the microstructu re of the powders, in addition to creating more grain boundaries and irregular surface s (Orimo and Fujii 1996; Zaluska et al. 1999). Hydrogen Absorption Test Hydrogen absorption studies were performed us ing a stainless steel chamber.0.905 g of the Mg-Ni composite was introduced in to the hydrogenation chamber (180 cm3) in a glove box. A schematic of the hydrogenation system is shown in Figure 2-2. Before introducing hydrogen into the system, the hydrogenation chamber was evacua ted (-11 psi) with a vacuum pump and argon (4.8 grade) was used to flush the chamber thr ee times. This was done in order to lower any possibility of contaminants (i.e. O2, CO2, or H2O). The chamber was then pressurized with pure hydrogen (high purity grade) to 145 psi. A bla nk test was performed using the hydrogenation chamber (without sample) to confirm that there were no leaks. The chamber was pressurized to 145 psi with hydrogen at 25 C, and then the chamber temperature was raised to 300 C; heating was then removed to allow for cooling. The initia l pressure and final pres sure, after heating and cooling, at 25 C were accurate to within 1 % of each other. This confirmed that there was no

PAGE 23

23 leakage during the hydrogenation process. Unde r normal experimental conditions, runs were controlled to reach 210 C. The temperature was measured with two thermocouples (Fisher Scientific model 15-078-39, type K) attached to the inside/outside of the chamber; a digital meter (Fisher 15-078-39) was used to display the output. The pres sure was monitored using an electronic pressure transmitter (Omega, PX880) connected to a process panel meter (Omega, DP25B-E). Once the chamber was pressurize d, heating tape (Ome ga FGS101-040, 420 Watt) was wrapped around the chamber. A DAQ (dat a acquisition module, Omega DAQ-55) and software control program was introduced to automatically acquire internal and surface temperatures as well as internal chamber pressu re data on a timely basis. External temperature was used by the software program to cont rol the internal chamber temperature. Particle Characterization The formation of MgH2 in the Mg-Ni composites before and after the hydrogen absorption test was determined by X-ray diffraction (Phill ips XRD APD 3720) with Cu K radiation. This process was controlled by PC-APD software, wh ich was used for all scans. The morphology change of the sample before and after dry co ating was investigated with Scanning Electron Microscopy (SEM, JEOL JSM 6335F). Energy di spersive spectroscopy (EDS) was used to determine the dispersion of Ni nanoparticles on the surface of the host (Mg). Desorption studies were performed using a Thermogravimetric Analyzer (TGA 2050, TA Instruments). Data Analysis During MgH2 formation, the loss in moles from gas to solid is indicated by the drop of PV/RT values, RT PV n (2-1)

PAGE 24

24 where P is pressure, T is temperature, R is the ideal gas constant, V is the volume of the reaction chamber and n is the number of H2 moles in the chamber. The PV/RT value was utilized to aggregate pressure and temperat ure effects simultaneously and to compare experimental data with the theoretical value. For an ideal gas without reaction, the n value would remain constant. When calculating the hydrogen ab sorption capacity, the initial and final pressures at room temperature were used. Experiments were perf ormed to quantify the r eaction rate during the hydrogenation process. TGA data were used as a comparison to the PV/RT data calculated. Results in Table 2-1 verify that the calculated data using measured PV/RT values were accurate to within 2 % of the experimental data from the TGA. Therefore, this method was used in subsequent analyses. Results and Discussion Particle Morphology & Texture Change Particle samples were analy zed with SEM for their morphologies. From the image of pure 44 m Mg in Figure 2-3, it was observed that Mg had an irregular sh ape and a broad size distribution. Images of 44 m Mg with nano-Ni 2 wt % after milling with Theta Composer did not show observable physical cha nges in shape or size distribu tion. However, in Figure 2-4, these coated particles showed well dispersed wh ite spots over the surface (Figure 2-4a) which were not seen in the image of pure Mg. These white spots were determined to be nano-Ni by SEM EDS mapping (Figure 2-4b). As shown in Figure 2-5, significant change s in surface morphology were observed upon comparison of the pure Mg particles only and those after hydrogenati on. After hydrogenation, a gel type texture was observed on the surface of hydrogenated particles due to the formation of MgH2. Figure 2-6 shows XRD patterns of Mg-2 wt % Ni (90 min processi ng) with different coating speeds of 4600 and 3600 rpm. Although the characteristic MgH2 peaks ( = 28 and 36)

PAGE 25

25 were present for both instances following exposure to H2, the presence of Mg verified the formation of an impermeable hydride layer a nd inability of hydrogen to permeate into the magnesium particle core (Bloch and Mintz 1997; Holtz and Imam 1999). Characteristic peaks for Ni (2 = 45, 51, 66) were also present for all of the materials processed with Ni. There was no evidence of MgxNiy alloys formed in this dry coating process; samples ball milled for long periods of time in prior studies, however, showed these alloy formations (Song 1995). Effects of Coating Time and Speed on Hydrogen Absorption The effect of processing time on hydrogen absorption was examined by hydriding composites of Mg-2 wt % Ni processed for 60 a nd 90 minutes while all other parameters were kept the same. Figure 2-7 shows the different absorption behavior (indicated by different slopes) plotted as PV/RT vs. coating time. The absorpti on rate indicated by the slope is an important parameter for application. In th e initial stage of heating, there was a diffe rence between internal and external temperature readings. The data poin ts during this transien t period (before 50 min) should therefore be ignored. It was observed that for the material processed for 90 minutes, there was a significant increase in hydride formation over those co mposites processed for only 60 minutes. A longer processing time was shown to be more effective in distributing Ni onto the Mg powder surface and resulted in an increase in the number of active absorp tion sites. A shorter processing time might not have su fficiently dispersed agglomerat es on the surface, resulting in only localized catalytic enhancement. Extende d processing in Theta Composer improved the hydrogen storage capacity and absorption rate fo r composites with the 2 wt % Ni loading. The effect of coating speed on hydrogen ab sorption was investigated with all other parameters (i.e. coating time, Ni loading, and gap size) kept constant. Pressure decay slopes as a function of time for two different coating spee ds are shown in Figure 2-8. Hydrogen absorption

PAGE 26

26 curves show two different slopes. Decaying slop es for two different processing speeds (3600 and 4600 rpm) show similar slopes in the domain of constant temperature, (0.17 and 0.20 respectively for Region I, and 0.12 and 0.10 for Re gion II), an indicative of having the same mechanisms for H2 absorption. Initially, hydrogen absorbed in Region I shows a steep slope of pressure decay, while in region II this slope becomes less steep. Th e transition of the slope from Region I to II implies a change in the rate lim iting mechanism from the interfacial growth of MgH2 to hydrogen diffusion (Bloch and Mint z 1997; Friedlmeier and Groll 1997). Higuchi et al (1999) reported that hydrogen penetrat es to a terminal depth of 30 m into the Mg surface, if sufficient time is provided. For a spherical particle this corresponds to a diameter of less than 60 m. The penetration depth of hydrogen into the Mg-Ni composite synthesized by Theta Composer was calculated using the change of H2 moles in the chamber following Equation 2-1. The loss of H2 was translated into the formation of a MgH2 film on the surface of Mg particles. The relationship between the volumes of pure Mg before and after hydrogenation can be expressed by p c PV H V ) 6 7 1 (' (2-2) where VP is the volume of pure Mg in a single particle, PV is the volume of pure Mg after hydrogenation and HC is the hydrogen capacity (wt%), and 7.6 (wt%) is the theoretical hydrogen capacity of MgH2. The penetration depth (hp) can then be determined using the following equation 3 1 '] 6 [ 2 1P p pV r h (2-3) where rp is the radius of the Mg particle (22 m in this study). For pract ical applications, it is desired to maximize the use of the interfacial growth limited Region I. Therefore, it is important

PAGE 27

27 to identify the hydride layer thic kness. Based on the hydrogen absorption capacity shown in Figure 2-8, the penetration depth into the 44 m spherical Mg particles synthesized in this study was calculated to be around 3 m for Region I. Effects of Ni Loading and Gap Size on Hydrogen Absorption The effects of Ni loading were studied with 2 and 4 wt % Ni. During the initial stage (before 200 minutes) on Figure 2-9(a) the 4 wt % Ni slope showed a steeper curve than that of the 2 wt % Ni. However, as the absorption time a pproached this point (2 00 minutes), both curves converged. In the first region where chemical ab sorption is the rate lim iting mechanism (before 200 minutes), higher Ni wt % might increase the number of active absorp tion sites so increased the reaction rate in that region. A lower Ni wt % loading might not sufficiently cover the surface, resulting in only localized catal ytic enhancement. In other words, by increasing the Ni wt % loading, the quantity of hydrogen absorbed by th e 44 m powders was initially enhanced. After the rate limiting mechanism changed from chemical absorption (Region I) to H2 diffusion into the Mg particle (Region II), however, Ni wt % loading did not affect H2 absorption capacity, indicating the limitation of the impermeable layer. In summary, chemical absorption rate was improved by the addition of Ni nanoparticles; ultimately, when hydrogen diffusion became the rate limiting mechanism, the capacity had li ttle dependence on the Ni wt % loading. The effect of rotor and vessel gap size in Theta Composer was examined by hydriding composites of Mg + 2 wt % Ni processed for 90 mi nutes. As shown in Figur e 2-9 (b), initially, the 0.5 mm gap size produced a slig htly better hydrogen absorption capacity. As chemical reaction progressed, the hydrogen absorption curves of 0.5 mm and 1.0 mm gap size came closer to each other. It implies that after the rate limiting mechanism changes from chemical absorption to H2 diffusion into the Mg particle gap size does not affect H2 absorption capacity. It should

PAGE 28

28 also be mentioned that some particles in the 0.5 mm gap case were stuck on the vessel wall due to excessive shear forces. This build-up can so metimes cause damages to the instrument and bring handling difficulties, result ing in unevenly processed partic les. Hence, the 0.5 mm gap size was not used in other experiments. Effect of Heating Rate on Hydrogen Absorption The effect of heating rate on the kinetics wa s investigated. Experi ments were performed using Mg-2 wt % Ni after 90 minutes of pro cessing. The chamber was quickly heated (<50 minutes) to 210 C, which was then maintained fo r 400 minutes. The same experiment was then repeated at a lower heating rate, which was heat ed to 210 C in 120 minutes. It can be seen in Figure 2-10 how the heating rate affected the prog ress and final extent of reaction. With a higher heating rate, higher reaction kinetic s overcomes the diffusion barrier which results in an increase in the rate of both interfacial growth of MgH2 and hydrogen diffusion. This can allow more hydrogen atoms to permeate further into the Mg core, resulting in a greater hydrogen storage capacity. The penetration depth co ndition in Region I under this c ondition was calculated to be 4.6 m, which was larger than the slower heating case (3 m).The results demonstrate the benefit of a high heating rate without raising temperature. Hydrogen Desorption Rates and Capacity In previous sections, the varying parameters experimented resulted in trends specific to each parameter. Samples with a longer coating tim e, a higher coating speed, 2 wt % Ni loading, and 1.0 mm gap size showed better absorption charac teristics. Samples with a higher heating rate also showed greatly enhanced absorption trends. TGA data were used to analyze the desorption characteristics of samples; these data are di splayed in Figure 2-11. The hydrogen capacities for normal heating rate samples showed little deviat ion among different coatin g speeds, Ni loadings, and gap sizes. However, samples processed at 3600 rpm showed an initial desorption

PAGE 29

29 temperature of approximately 220 C with a steep desorption slope. On the contrary, those samples of 4600 rpm exhibited an initial de sorption temperature of approximately 240 C, with slopes less steep than in the previous instance, indicating stronger hydr ogen bonding. Coating time was the other important parameter affecti ng desorption. A lower coating time showed a significantly lower final hydrogen capacity than a higher coat ing time. Another important parameter that affected the capacity was the heat ing rate. A higher heatin g rate resulted in a product with a much higher final hydrogen capacit y (4.82 %) than samples treated at the lower heating rate. Summary The formation of magnesium hydride on Mg-Ni composites synthesized by Theta Composer was studied. This provides the insight of the rate limiting mechanism change between interfacial growth and diffusion. The effects of various parameters on hydrogenation/desorption kinetics and capacities were investigated. Based upon the experimental re sults, a longer coating time in Theta Composer improved the hydrogen st orage capacity and abso rption/desorption rates for composite with the same Ni loading due to th e better distribution of Ni nanocatalysts. It was also observed that there was a change in the rate limiting mechanism of hydrogenation from interfacial growth of MgH2 to diffusion through the hydride layer into the Mg particle core. The penetration depth for the interfacial growth limited case was approximately 3 m at the lower heating rate (1.5 C/min). A lower Ni wt % loading did not sufficien tly cover the surface and resulted in only localized catalytic enhancement. However, afte r the rate limiting mechanism changed from the interfacial growth of MgH2 to H2 diffusion, Ni wt % loading had negligible impact on the H2 absorption rate and capacity.

PAGE 30

30 In studying the change in gap size, the 0. 5 mm gap produced a s lightly better hydrogen absorption rate. However, in this case many particles built up on the Theta Composer vessel wall due to excessive shear forces. Therefore, a 1.0 mm gap size is preferred. A higher heating rate (3.7 C/min) showed a substantial improvement in hydrogen absorption rate and storage capacity; the corresponding penetr ation depth of interfacial growth limited Region I was 4.6 m. A lower coating speed resulted in lower initial desorption temperatures and a steeper desorption rate.

PAGE 31

31 Table 2-1. Comparison of calculated hydroge n uptake with hydrogen desorption by TGA Sample (coating speed, Ni wt %, coating time, gap size) Experiment (TGA) Calculated Relative error (%) 3600x70, Ni 2%, 90 min, 1.0 mm 3.34 3.33 0.37 4600x70, Ni 2%, 90 min, 1.0 mm 3.20 3.28 2.39 4600x70, Ni 2%, 60 min, 1.0 mm 2.12 2.09 1.49 4600x70, Ni 4%, 90 min, 1.0 mm 3.35 3.33 0.67 3600x70, Ni 2%, 90 min, 0.5 mm 3.32 3.32 0.10 4600x70, Ni 2%, 90 min, 1.0 mm* 4.83 4.83 0.15 Higher heating rate

PAGE 32

32 Figure 2-1. Schematic diagram of Theta Composer

PAGE 33

33 Figure 2-2. Schematic diagra m of hydrogenation system

PAGE 34

34 Figure 2-3. Surface morphology change (Rotor speed: 4600 rpm, Gap Size: 1 mm, Mg-2 wt % Ni): (a)Pure 44 m Mg particle, (b) 44 m Mg-2 wt % Ni particle (After dry coating, Before Hydrogenation)

PAGE 35

35 Figure 2-4. Distribution of Ni nano particles on Mg (Rotor speed: 4600 rpm, Gap Size: 1 mm, Mg-2 wt % Ni): (a) SEM, (b) EDS mapping of Ni

PAGE 36

36 Figure 2-5. Particle texture cha nge (Rotor speed: 4600 rpm, Gap Si ze: 1 mm, Mg-2 wt % Ni); (a) Pure 44 m Mg only, (b) after hydrogenation (In itial hydrogen pressure: 145 psi, Hydrogenation time: 300 min, Temperature: 483K) Smooth texture Gel type texture

PAGE 37

37 Figure 2-6. XRD patterns for MgH2 -2 wt % Ni Composites (Ini tial hydrogen pressure: 145 psi, Hydrogenation time: 300 min, Temperature: 483K) a) 4,600 RPM b) 3,600 RPM

PAGE 38

38 Time, min 050100150200 PV/RT, mole 0.024 0.026 0.028 0.030 0.032 0.034 0.036 0.038 0.040 0.042 0.044 coating time : 90 min coating time : 60 min Figure 2-7. Effect of coating time on H2 absorption (Rotor speed: 4600 rpm, Gap Size: 1 mm, Mg-2 wt % Ni, Initial hydrogen pressu re: 145 psi, Hydrogenation time: 300 min, Temperature: 483K)

PAGE 39

39 Time, min 0100200300400 140 150 160 170 180 190 200 250 300 350 400 450 500 Pressure Temerature Temperature, K Pressure, Psi 280 min dP/dt= -0.17 r 2 = 0.99 dP/dt= -0.12 r 2 = 0.99 (a) 0100200300400 140 150 160 170 180 190 200 250 300 350 400 450 500 Pressure Temperature Time, minTemperature, K Pressure, Psi280 min dP/dt= -0.2 r2= 1.00 dP/dt= -0.1 r 2 = 0.99(b) Figure 2-8. Profile for hydrogenation runs with material Mg-2 wt % Ni (90 min): (a) Rotor speed: 3600 rpm, (b) Rotor speed: 4600 rp m (Initial hydrogen pressure: 145 psi, Hydrogenation time: 300 min, Temperature: 483K)

PAGE 40

40 Time, min 0100200300400 PV/RT, mole 0.024 0.026 0.028 0.030 0.032 0.034 0.036 0.038 0.040 0.042 0.044 Nickel 2 wt % Nickel 4 wt % (a) Time, min 0100200300400 PV/RT, mole 0.024 0.026 0.028 0.030 0.032 0.034 0.036 0.038 0.040 0.042 0.044 Gap size: 1.0 mm Gap size: 0.5 mm (b) Figure 2-9. Effect of Ni loading &gap size on H2 absorption: (a) Effect of Ni loading (Rotor speed: 4600 rpm, Gap Size: 1 mm, Coating time: 90 min), (b) Effect of gap size (Rotor speed: 3600 rpm, Ni wt %: 2 %, Coating time: 90 min) Hydrogenation condition (Initial hydrogen pressure: 145 psi, Hydrogenation time: 300 min, Temperature: 483K)

PAGE 41

41 Time, min 0100200300400500 PV/RT, mole 0.015 0.020 0.025 0.030 0.035 0.040 0.045 Temperature, K 250 300 350 400 450 500 PV/RT PV/RT, higher heating rate Temperature Temperature, higher heating rate Figure 2-10. Effect of heating rate on H2 absorption (Rotor speed: 4600 rpm, Ni wt %: 2 %, Coating time: 90 min, Gap size: 1 mm)

PAGE 42

42 Temperature, oC 0100200300400 Weight loss, % 94 95 96 97 98 99 100 101 3600* 70, Mg-2 wt. % Ni, 90 min,1.0 mm 4600* 70, Mg-2 wt. % Ni, 90 min,1.0 mm 4600* 70, Mg-2 wt. % Ni, 90 min, 1.0 mm, higher heating rate 4600* 70, Mg-2 wt. % Ni, 60 min, 1.0 mm 3600* 70, Mg-2 wt. % Ni, 90 min, 0.5 mm 4600* 70, Mg-4 wt. % Ni, 90 min, 1.0 mm 2.12 % 4.82 % 3.20 % 3.32 % 3.34 % 3.35 % Figure 2-11. Hydrogen desorption analysis by TGA

PAGE 43

43 CHAPTER 3 MODEL FOR HYDROGEN ABSORPTION KI NETIS AND CAPACITY IN METAL HYDRIDES Introduction Over the past decade, the use of hydrogen as an alternative fuel ha s been of particular interest to many applications, from mobile electronics to aerospace industries (DOE 2007a; NASA 2006; Schlapbach and Zttel 2001). Amo ng various available methods for hydrogen storage, solid storage using reversible metal hydr ides is considered to be one of the most promising (Biswas and Wu 2005; Schlapbach and Zttel 2001). The advantages of reversible metal hydrides include improved safety, enviro nmental friendliness, and a relatively high hydrogen storage capacity. One main limiting factor however, is the high temperature required for the formation of metal hydr ides with acceptable hydrogen uptak e and release kinetics, which makes them expensive processes (Fuka i 2005; Schlapbach and Zttel 2001). There have been many experimental studies in recent years to understand the rate-limiting kinetics and absorption capacity of hydrogen st orage in alloys (Fuka i 2005; Holtz and Imam 1999; Liang et al. 1998; Lia ng et al. 1999; Orimo and Fujii 1996; Yang et al. 2000). Holtz and Imam (1999) proposed that nucleation of metal hydr ide on the surface of pure metal is the initial rate limiting step for absorption of hydrogen, as initial hydrogenation (i.e. nucleation of metal hydride) takes place only at the metals active surface. As hydrogenation progresses, the hydride layer (impermeable to hydrogen) thickens and expands from the surface. Diffusion through the metal hydride layer is many orders of magnitude slower than other mechanisms (i.e. nucleation of metal hydride and dissociation of a hydroge n molecule into two hydrogen atoms) and subsequently becomes the rate-limiting step in the hydrogenation process (Bloch and Mintz 1997; Friedlmeier and Groll 1997; Sastri et al. 1998; Schwe ppe et al. 1997).

PAGE 44

44 Several empirical or theoretical kinetic models have also been formulated and subsequently fitted with experimental data for describing hydrogen absorption/ desorption behavior of metal hydrides (Chou et al. 2005; Friedlmeier and Groll 1997; Ginstling and Brounshtein 1950; Rudman 1979; S aetre 2006). The nucleation and growth model developed by Avrami (Avrami 1939; Avrami 1940; Avrami 1941) is a frequently applied model for describing phase change in solids. In Avramis model, random nucleation and the subsequent three dimensional growth of a new phase within the bulk is the rate-limiting step. Rudman (1979) extended Avramis model to metalhydrogen syst ems; the effects of temperature and hydrogen pressure on the kinetics for hydrogen diffusion limited cases were considered, assuming a constant growth rate of magnesium hydride. The other commonly accepted model for metal hydride formation was proposed by Flanagan (1978). Hydrogen molecules are adsorbed or weakly chemisorbed on an active metal surface where they dissociate into hydrogen atoms. These atoms can easily diffuse between the grain boundaries of the pure meta l and then finally form metal hydride (Flanagan 1978; von Zeppelin et al. 2002). If diffusi on of hydrogen atoms occurs at a fast rate along the grain boundary (i.e. the grain size is small), a metal hydr ide layer is formed and thickens into a pure metal core (Zaluska et al. 1999). This can be best described by the shrinking-core model(Levenspiel 1999). Seve ral numerical studies have been conducted based on the shrinking-core model, considering changes in gr owth rate for a diffusion limited case (Chou et al. 2005; Ginstling and Brounshtei n 1950; Rudman 1979; Saetre 2006; Schweppe et al. 1997). Barkhordarian et al.(2006) reported that their experimental data of the absorption behavior of magnesium catalyzed with Nb2O5 were better describe d by a three-dimensional diffusion controlled shrinking-core model. Gabis et al.(2005) also compared Avramis model and

PAGE 45

45 the shrinking-core model with thermal desorp tion spectrometry (TDS) experiments of metal hydride, and the results showed that the shrink ing core model better described the kinetics of hydrogen desorption. Nevertheless, predictions by both models deviate from experimental observations, especially during th e initial period of the hydrogena tion process (Schweppe et al. 1997; Stander 1977). The reason for the deviation stems from the key assumptions of these models that the process is diffusion limited and that the change in rate limiting mechanisms during the process is not considered. Stander (1977) divided the kinetics of Mg hydride formation into two mechanisms (nucleation and diffusion) and likewise fitted his experimental data separately for each of these cas es as a function of the fraction ( ) of metal hydride. He reported that the nucleation equation was valid for pure Mg for an upper limit of 0.18 for This is equivalent to a layer of 1.6 m MgH2 on a Mg particle diameter of 25 m. After the magnesium surface was utilized ( >0.18), a separate equation for diffusion was reported to reproduce experimental data well. The objectives of this chap ter were to investigate the kinetics of hydrogen absorption mechanism of a metal hydride system by consider ing multiple mechanisms in the process. A kinetic model inclusive of interfacial and diffusion controlled growth mechanisms was developed to determine the absorption rate a nd capacity for isothermal conditions. The model predictions were validated by experimental data of Mg-Ni systems to demonstrate its applicability. Model Development The mathematical model developed in this study adopted the concep t of a shrinking-core spherical particle described in pr evious literature (Levenspiel 1999). Figure 3-1 shows the conceptual schematic of the shrinking core with three sequential steps for hydrogenation considered in this study:

PAGE 46

46 Step1: Dissociation of hydrogen molecules into hydrogen atoms at the interface of the catalyst and the Mg particle surface. Step2: Diffusion of hydrogen atoms th rough the surface layer of MgH2 and into the volume of the unreacted Mg core. Step3: Interfacial growth of hydride. An issue that requires atte ntion is that the true value of hydrogen atom concentration cannot be measured in our system. Furthermor e, the system can not distinguish between interfacial growth and dissociation. To circumve nt this problem, it was assumed that dissociation of hydrogen molecules into atoms at the interface was instantaneous. Therefore, the equivalent molar concentration of hydrogen atoms (CH) chemisorbed is twice the molar concentration of hydrogen molecules (CH2) in gas. As a result, dissociati on is overlooked and only two steps are considered in our model: diffusion of the hydrogen atoms, and interfacial growth with the Mg core to form MgH2. The following sections describe the de velopment of the m odel for isothermal conditions. Absorption Kinetics Diffusion The rate of diffusion of H can be expressed as the flux across the area of unreacted core with the concentration gradient as the driving force as follows. dr C t C d D r dt dNd H c H) ) ( ( 42 (3-1)

PAGE 47

47 where CH(t) is the concentration of hydr ogen atom chemisorbed in MgH2 at time, t. This applies to an experimental situation where hydr ogen pressure in the chamber and hence CH(t) is decreases with time t. Cd is the maximum solubility of hydrog en atoms in magnesium (Callister 1997; Ovrelid et al. 1998; Ragone 1995; Ze ng et al. 1999). Compared to the hydrogen concentration in a typical hydrogenation chamber (approximately 30 ppm at 300 oC at 1.5 Mpa of H2 atmosphere) (Ovrelid et al. 1998), Cd is very small and is therefore neglected. The definition of each variable is given in the Nomenclature section. Interfacial Growth Assuming that the growth of MgH2 is interfacial controlled, the rate of MgH2 formation can be considered as a first ordered reacti on and hence proportional to the hydrogen atom concentration and the available interface area of unreacted Mg. Therefore, the loss of H per unit time due to interfacial grow th can be expressed as: ) ( 42t C k r dt dNH c H (3-2) Overall Process The consumption of two H atoms due to growth of MgH2 requires one Mg atom, i.e. H MgdN dN 2 1 (3-3) The change of number of Mg atoms can then be related to the cha nge of Mg core volume and its shrinking radius as c c Mg c Mg c Mg Mgdr r r d d dN2 34 ) 3 4 ( (3-4) In this calculation, the volume and dens ity change due to the formation of MgH2 is ignored. Incorporating these equations (3-1 ~ 3-4) gives the following overall MgH2 formation rate in terms of the change in Mg particle radius

PAGE 48

48 1 ) ( ) (2k RD r r R t C dt drc c Mg H c (3-5) Solution to Rate Limiting Condition The solution to the kinetic equation can be obtained by considering the initial condition and mass balance. The initial condi tion (for an isothermal system) is: t = 0, rc = R (3-6) The variation of H2 concentration in the gas phase with time can be related to the change of core diameter using a simple mass balance expression, )) ( ( 3 4 ) (3 3 2 0 2t r R V N t C Cc Mg H H (3-7) The results for each rate limiting case are discussed below. Interfacial Growth Limited Case When the process is interfacial growth-limite d, Equation 3-5 in a system with varying H2 concentration (Equation 3-7) can be simplified as Mg c ck dt dr Q Mr" ) ( 13 (3-8) where V N MMg3 4 and ] [3 0 2MR C QH Using Equation 3-6, the solu tion can be obtained as t b c d e g f M a k Mg )] ( ) ( ) ( 2 [ 6 12 (3-9) where a through g are expanded as follows:

PAGE 49

49 ) ) 2 ( 3 1 ( tan 3 2 ) ) ) ( 2 ( 3 1 ( tan 3 2 ) (1 1 3 / 1a a R c a a t r b M Q ac ) ln( ) ) ( ln( ) ln( ) ) ( ) ( ln(2 2 2 2a R g a t r f a Ra R e a a t r t r dc c c Diffusion Limited Case As the hydride layer grows thicker, diffusion through the layer becomes a stronger barrier that controls the rate. Equation 3-5 can then be reduced to the following )] ( [ ) (3 3 0 2 c H c c Mg cr R M C r r R DR dt dr (3-10) Using the same initial condition (Equation 3-6), the solution can be obtained as t i h a c b R e d R g f R a DMMg )] ( 2 ) ( ) ( ) ( 2 [ 6 1 (3-11) where h and i are expanded as follows: ) ln( )) ( ln(3 3MR Q i t Mr Q hc Comparisons with Equations Reporte d in the Literature (constant CH2) For a constant hydrogen concentration, the so lution to Equation 3-10 can be obtained as t RDC R Rr rMg H c c2 3 2 32 3 3 2 (3-12)

PAGE 50

50 Equation 3-12 can be further transformed us ing the following relationship between rc, and R 3 1) 1 ( R rc (3-13) The resultant equation is Dt C RMg H) ( 2 ) 1 ( ) 3 2 ( 12 2 3 / 2 (3-14) This is identical to the shrinking core mode l with three-dimensional diffusion (Ginstling and Brounshtein 1950) as expressed by: Kt 3 / 2) 1 ( ) 3 2 ( 1 (3-15) where K was called the overall coefficient as shown in earlier studies (G instling and Brounshtein 1950). Indeed, K is a function of particle size, hydrogen pressure, and the diffusion coefficient of hydrogen. Overall Equation with Two Mechanisms Equation 3-5 coupled with Equation 3-7 yields an equation that includes two mechanisms for a varying hydrogen concentration 1 ) ( ) (3k RD r r R t Mr Q dt drc c Mg c c (3-16) The solution to Equation 3-16 is

PAGE 51

51 t a k i h aR k b c d e g f DR b c d e g f DR Ma kMg ] ) ( 2 )} ( ) ( ) ( 2 { )} ( ) ( ) ( 2 [{ 6 12 2 (3-17) While Equation 3-15 provides a simple form, its applicability is limited to the assumptions of constant hydrogen concentration and a d iffusion only case. E quation 3-17 allows comparison of various rate limiting steps and hydrogen concentrations. Such comparisons have not yet been systematically reported for the entire hydrogenation process. Experimental Characterization of Hydrogen Absorption Kinetics Mg particles coated with nano Ni catalyst were prepared using a dry coating system (Theta Composer, Tokuju Corp.). Hydrogen ab sorption was studied using a custom-made hydrogenation system, which is shown in Fi gure 3-2. The methodology used in sample preparation and the elements of the hydrogenation system have been described in depth elsewhere (Jeon et al. 2006). Only the experimental conditions used in this study are outlined here. 0.907 g of Mg-Ni composite was used in the experiment and was prepared under a controlled argon environment in the glove box. The hydrogenation chamber was heated using heating tapes (Omega FGS101-040, 420 W) while in vacuum. After the chamber reached the designed temperature (463, 483, 503, and 523 K), pure hydrogen of targeted pressure (0.5, 1.0, and 2.0 MPa) was then immediately introduced. The temperatures inside and outside of the chamber were controlled using a DAQ (data acquisition module, Omega DAQ-55). Internal temperature and chamber pressure da ta were recorded automatically.

PAGE 52

52 Transmission Electron Microscopy Dual focused ion beam milling technique (F EI DB 235) with a gallium liquid metal ion source (LMIS) was used to prepare thin foil fo r transmission electron microscopy (TEM). This technique was successfully employed to prepare a 100 nm thick sample from a 44 m Mg-Ni composites hydrogenated under 1 MPa of hydrogen at 483K. The sample was imaged in the FIB system using a 5 KV beam to minimize possible damage to hydride structure.A thin layer of platinum was deposited on the su rface via an in situ needle platinum aspirator that was positioned 80 mm above the region of interest Micrographs and diffraction patterns of the sample were analyzed in JEOL TEM 200CX. Presence of MgH2 Layer near Surface Several experimental studies have b een conducted to explain hydrogen absorption behavior of metals (Friedlmeier and Groll 1997; Fukai 2005; Gabis et al 2005; Jeon et al. 2006; Liang et al. 1998; Liang et al. 1999; Orimo and Fujii 1996; Stander 1977; von Zeppelin et al. 2002; Yang et al. 2000; Zaluska et al. 1999). Howeve r, these studies did not identify the actual mechanisms of MgH2 formation in Mg particles and there is a need for clarity (Bloch and Mintz 1997). Figure 3-3(a) shows the formation of the hydrogenated layer (Region A) near the surface of the Mg-Ni composite particle. The partial cr acks observed on the same layer shown in Figure 3(b) are an indication of volume expansion when MgH2 formed. Similar observations of volume expansion have been reported in prior Mg hydrid e studies (Bloch and Mintz 1997; Friedrichs et al. 2007). The hydrogenated layer thickness varied consid erably, suggesting th at the interfacial controlled growth mechanism of hydride is a va lid assumption. When the growth is solely diffusion controlled, this layer is anticipated to have a constant thickness. The maximum depth of the reacted layer measured graphi cally in Figure 3-3(a) is 4.2 m which is capable to the

PAGE 53

53 theoretical average thickness of 5.2 m based on the observed hydrogen capacity of 4.2 wt. %. The difference between the two values likely co mes from the assumptions of sphericity and smooth surface in the equation. Figures 3-3 (c) and (d) show the electron diffraction pattern s (DP)of areas (A) and (B) marked in Figure 3-3(a).For identical magnificat ion and SAD aperture si ze, area (B) exhibited a single crystal DP while highly dens e ring DP from area (A) demonstrat es that this area consists of ultra-fine magnesium grains. The DP from th e latter area also showed the presence of MgO, which was possibly formed during sample prep aration. Although the XRD profile of the hydrogenated material showed the formation of MgH2 phase (Jeon et al. 2006), the TEM sample did not reveal the presence of this phase. Si milar observation has been made other studies (Friedrichs et al. 2007; Friedrichs et al. 2006; Shao et al. 2004). When MgH2 is exposed to the electron beams und er TEM with a high vacuum condition, the release of hydrogen gas commences within a fe w seconds (Friedrichs et al. 2007; Friedrichs et al. 2006; Shao et al. 2004) and dehydrogenated MgH2 results in the formation of very fine Mg crystals (Schober 1981). Therefore, the above resu lts indicate that area (A) had been the MgH2 layer, which was converted to ultra-fine grai ned Mg during the imaging process or possibly sample preparation. Determination of Diffusion and Interfacial Growth Coefficients for Hydrogenation A limited number of hydrogen di ffusion coefficients in MgH2 have been reported in previous studies (Bowman and Fultz 2002; Cui et al. 2001; Ni shimura et al. 1999). Ho wever, the values of diffusion coefficients obtained from these studies vary significantly (6 or ders of magnitude). One possible reason for this can be the differences in grain sizes of sample in these experiments, although such information is not available. Seve ral researchers (Callist er 1997; Higuchi et al. 1999; Revesz et al.; Varin et al. 2006; Zaluska et al. 1999) have performed hydrogen

PAGE 54

54 diffusion/nucleation experiments using singlea nd poly-crystals with c ontrolled grain size. Results showed that reduced grain size of the metal powder enhances diffusion of hydrogen atoms within the metal, leading to increased metal hydride formation. It should also be noted that the concept of interfacial controlled growth of hydride has not b een suggested prior to this study. Figure 3-4 shows the change in hydrogen pressu re as a function of hydrogenation time in a typical experiment using the hydrogenation syst em. In this study, based upon the proposed analytical model, the absorption curve is divided into four regi mes (i.e., incubation, interfacial growth limited, transition, and diffusion limited). The interfacial growth coefficient (k) was obtained from the change of hydrogen pressure in the interfacial growth limited regime (Region b in Figure 3-4) using Equation 3-9. For the diffusion coefficien t (D), the change of hydrogen flux was used to interpret the relationship using Equation 3-11, which is applicable in the diffusion-limited regime of the hydrogenation process (Figure 3-4D). Figure 3-5 shows hydrogen diffusion and interfacial growth coefficients as a function of temperature with three replicated experi ments for each condition. The hydrogen diffusion activation energy (Ea) of the Mg-2 wt. % Ni co mposites was calculated using an Arrhenius form, as shown in Figure 3-5(a). The activation en ergy was found to be 2.38 Kcal/mol, which is significantly lower than that of the bulk magnesium measured by Nishimura (1999), which was 24.1Kcal/mol. Activation energy of interfacial growth coefficient was 9.28 Kcal/mol. The activation energy of hydrogen di ffusion did not demonstrate te mperature effect on diffusion coefficient adequately. It is to be noted that ex perimental results are va lid within the range of temperatures studied. Other fundamental assumpti ons that were made are listed as follows. Mono-disperse spherical Mg particle of diameter 44 m

PAGE 55

55 Hydrogen atom concentration chemisorbed on MgH2 is half of hydrogen molecule concentration the gas phase (i.e., activation energy for vacancy formation in MgH2 was neglected). Comparative Study of Experimental and Calculated Data Using the diffusion and interfacial growth coefficients obtained, hydrogen absorption capacity can be determined usi ng Equation 3-17 and was compared to the experimental value. Figure 3-6 shows the absorption behavior of Mg-2 wt. % Ni composites as a function of hydrogenation time at 483 K. The model predictions (interfacial growth and diffusion), as shown, are almost identical to the experimental data for the entire hydrogenation process. The model prediction considering diffusion only, on th e other hand, shows a slower increase in the beginning. The close fit of the calculated and expe rimental data shown in Figure 3-6 verifies the hypothesis that there are two major rate-limiting mechanisms in the hydrogenation process: in the beginning, interfaci al limited growth of metal hydride is the rate limiting mechanism which changes to diffusion upon the thickness of the hydrid e layer. This is depict ed by the steep initial increase, followed by a gradual decrease in the ra te of change of hydrogen penetration depth with increasing hydrogenation time. Effect of Temperature and Initial H2 Pressure on Hydrogen Absorption Kinetics While the effect of temperature on hydroge n absorption capacity has been quantified within bounds of experimentation for decades (Fukai 2005; Liu et al.; Rudman 1979; Saetre 2006; Sastri et al. 1998; Schlapbach and Zuttel 2 001; Song 1995; Stander 1977), the effect of temperature on hydrogen absorption ki netics has rarely been quantit atively modeled. Figure 3-7a presents the effects of temperature on hydrogen absorption predicted by the model. Experimental data obtained for 3 temperatures at 4 hydrogenation times are also displayed. The results for the Mg-2 wt. % Ni composites tested shows linear trends for the same hydriding time within the

PAGE 56

56 range of experimental conditions. Analysis of ex perimental data reported in a recent study of a Mg-40 wt%Ti0.28Cr0.50V0.22 composites (Liu et al. 2007) also shows similar trends. The dependence of hydrogen absorp tion on initial hydrogen pressure was also evaluated. Initial hydrogen pressure was varied from 0.5 and to 2.0 MPa. Figure 3-7b shows the hydrogen absorption capacities plotted as a function of initial hydrogen pressure. Both the model prediction following Equation 3-17 and the experime ntal data are presented. As shown, the model can provide a holistic picture of expected behavior based on broad assumptions to closely simulate practical situations. Effect of Mg Particle Diameter on Hydrogen Absorption Kinetics The diameter of magnesium particles has been identified as an important variable from experimental studies for metal hydrides (Varin et al. 2006; von Zeppelin et al. 2002; Yang et al. 2000; Zaluska et al. 1999). In practical applicati ons, hydrogen refueling time at a station is one of the key technical barriers for usabi lity of metal hydrides (DOE 2007a; DOE 2007b; Schlapbach and Zuttel 2001). DOEs technical targets of refueli ng time for a storage system are on the order of 5 minutes or less by 2007 and 3 minutes by 2012 (DOE 2007b) In this study, the model was used to determine hydrogenation time required to reach the theoretical hydrogen absorption capacity of Mg () as a function of particle diamet er. Simulation runs were performed for particle size from 1.25 m to 44 m with all the other parameters kept the same. Figure 3-8a shows the results of hydrogen absorption capacity as a func tion of hydrogenation time for 5 magnesium diameters. The samples composed of smaller Mg particles tend to have thinner MgH2 layers and achieve the target hydrogen capacity much faster than their larger counterparts. In other words, smaller particles are interfacial growth limited while larger particles experience diffusion limited conditions (as exhibited by th eir slope). Figure 3-8b shows hydrogenation time as a function of magnesium particle diameter. It can be seen that the hydrogen refueling time, of

PAGE 57

57 great interest to practical application, can be dr astically reduced to meet the DOE criteria for the hydrogen refueling time by changing the diameter of Mg particle. To meet the DOE criteria set for 2012, the particle size of the Mg-2 wt % Ni used in this study should be 1.25 m ( = 2.7 min) or smaller. Nevertheless, a larger particle size can be used under higher pressure or temperature, as discussed in the previous section. Summary The kinetics of formation of magnesium hydr ide layer on Mg-Ni composites was studied. It was verified by SEM/TEM characterization that magnesium hydride layer formed from the surface rather than from random interfacial grow th and growth in Mg. The formation of the hydride layer verifies that in the beginning of hydrogenation proce ss, interfacial growth of MgH2 is the rate limiting mechanism and as tim e goes on, diffusion of hydrogen through the MgH2 layer becomes the rate limiting mechanism. Th is observation provides an insight into the partitioning of the rate limiting mechanism betw een interfacial growth and diffusion, and it forms the basis for the analytical model. The analytical model was developed for evalua ting the change of rate-limiting mechanisms of Mg-Ni composites with regards to hydr ogen absorption rate and capacity, taking hydrogenation kinetics as a func tion of hydrogenation time into acc ount. The agreement between modeled and experimentally measured data was excellent. The model reproduced the change in the rate limiting mechanism of hydrogenation fr om interfacial growth of metal hydride to diffusion through the hydride layer into the Mg particle core. Th e activation energy of diffusion coefficient (2.38 Kcal/mol) of the Mg-2 wt. % Ni composites was found to be significantly lower than that of pure Mg (24.1Kca l/mol). Activation energy of MgH2 interfacial growth coefficient, which has never been reported before, was 9.28 Kcal/mol.

PAGE 58

58 The hydrogen absorption capacity can be pred icted as a function of the temperature, hydrogen pressure and Mg particle diameter at the given hydroge nation time. The time needed to hydride a Mg particle drastically decreases as its size decreases, and it de creases as temperature and pressure increase. As demonstrated, the new model developed in this study can more accurately describe hydrogen absorption capacity and kinetics, as it accounts for the effects of all important parameters: interfacial growth of metal hydride, diffu sion of hydrogen and a variable hydrogen concentration. Nomenclature C0 H2 Initial Concentration of H2 in gas phase [mole/m3] CH2 Concentration of H2 in gas phase [mole/m3] CH Concentration of H in MgH2 [mole/m3] Cd Concentration of H dissolved in Mg [mole/m3] D Diffusion coefficient [m2/s] dp Particle diameter [m] k Hydrogen dissociation constant [m/s] k Interfacial growth coefficient [m/s] K Overall coefficient in shrink core model [1/s] N Number of Mg particles in the sample [-] NH2 Moles of hydrogen molecule in gas phase [mole] NH Moles of hydrogen atom in solid phase in the system [mole] NMg Moles of Mg particle [mole] rc Radius of Mg particle as a function of time [m] R Initial Mg particle radius [m] Rgas Ideal gas constant [cal/K.mol] T Temperature [K] t Time [sec] V Volume of hydrogenation chamber [m3] Greeks Metal hydride volume fraction [-] Fraction of the catalyst coated area [-] Mg Molar concentration of Mg particles in the system [mole/m3] The time required to reach theoretic al hydrogen absorption capacity [min] c Volume of the shrinking Mg core [m3]

PAGE 59

59 Figure 3-1. Schematic diagram of the shrinking core model

PAGE 60

60 Figure 3-2. Schematic diagram of the hydrogenation system

PAGE 61

61 Figure 3-3. SEM/TEM micrographs and elec tron diffraction patterns obtained for the hydrogenated Mg-Ni: (a) SEM picture show ing the hydrogenated layer near the surface of the Mg-Ni particle, (b) TEM micrograph showing the formation of microcracks in the hydrogenated region, (c) diffraction pattern obtained from area (A), (d) diffraction pattern obtained from area (B)

PAGE 62

62 Time, min 110100 PV/RT, mole 0.015 0.020 0.025 0.030 0.035 0.040 0.045 (a) (b) (c) (d) 483 K 573 K Figure 3-4. Change in hydrogen pre ssure as a function of time (44 m Mg + 2 wt. % Ni, initial H2 pressure = 1 MPa): (a) Incubation regime, (b) Interfacial growth limited regime, (c) Transient regime, and (d) Diffusion limited regime

PAGE 63

63 1/T 0.0022 0.0021 0.0020 0.0019 ln k" -8.4 -8.2 -8.0 -7.8 -7.6 -7.4 -7.2 -7.0 -6.8 -6.6 R 2 =0.9911 ) 10 28 9 exp( 07 8 "3T R kg (b) 1/T 0.0022 0.0021 0.0020 0.0019 ln D -27.1 -27.0 -26.9 -26.8 -26.7 R 2 =0.9952 ) 10 38 2 exp( 10 41 23 11T R Dg (a) Figure 3-5. Determination of hydrogen absorption coefficient following the Arrhenius form: a) Diffusion coefficient, b) Nucleation coefficient

PAGE 64

64 Hydrogenation time, min 050100150200250300 Hydrogen absorption capacity wt.% 0 1 2 3 4 5 6 Measured data Modeled data(Interfacial growth and Diffusion) Modeled data(Diffusion only) (c) (d) (b) (a) Figure 3-6. Comparison of measured and model data (Simulation condition: 44 m Mg + 2 wt. % Ni, initial H2 pressure = 1 MPa, temperature = 483 K) : (a) Incubation regime, (b) Interfacial growth limited regime, (c) Transient regime, and (d) Diffusion limited regime

PAGE 65

65 Temperature, K 300350400450500550 Hydrogen absorption capacity wt.% 0 1 2 3 4 5 6 Experiment, 5 sec (Liu et al., 2006) Experiment, 14 sec (Liu et al., 2006) Experiment, 30 sec (Liu et al., 2006) Experiment, 50 min (this study) Experiment, 100 min (this study) Experiment, 200 min (this study) Experiment, 300 min (this study) Model, 50 min Model, 100 min Model, 200 min Model, 300 min (a) Pressure, MPa 0.00.51.01.52.0 Hydrogen absorption capacity wt.% 0 1 2 3 4 5 6 Experiment, 50 min Experiment, 100 min Experiment, 200 min Experiment, 300 min Model, 50 min Model, 100 min Model, 200 min Model, 300 min (b) Figure 3-7. Effect of temperatur e and initial hydrogen pressure on hydrogenation: (a) effect of temperature (initial pressure = 1.0 MPa), (b) effect of initial hydrogen pressure (Temperature: 483 K)

PAGE 66

66 Hydrogenation time, min 110100100010000 Hydrogen absorption capacity, wt % 0 1 2 3 4 5 6 7 8 Mg size: 44 m Mg size: 20 m Mg size: 10 m Mg size: 5.0 m Mg size: 2.5 m Experment (44 m) Magnesium partice size, m 01020304050 min 1 10 100 1000 (a) (b) Figure 3-8. Effect of Mg partic le size on hydrogen absorption ki netics (simulation condition: Mg + 2 wt. % Ni, initial H2 pressure = 1.0 MPa, temperature = 483 K); (a) Hydrogen absorption capacity as a function of hydrogena tion time for various Mg particle sizes, (b) effect of magnesium particle size on

PAGE 67

67 CHAPTER 4 FLAKE PARTICLE SYNTHESIS FROM DUCT ILE METAL PARTICLES USING A NOVEL HIGH-SPEED VIBRATORY MILL Background A flake-shaped particle is a particle with a high diameter-to-thickness aspect ratio. Needs for flake particles vary acro ss a variety of applications such as: pigments and inks, electrochemical electrodes, fuel cell hydrogen storage devices, e xplosives, lightweight concretes and obscurences (Cashdollar 2000; Hong and Kim 2001; Hong et al. 2000; Lee et al. 2002; Trunov et al. 2005; Yoshinaga et al 2002). Properties of flake-sh aped particles have growing interests in the chemical and material indus tries, which continuously demand for increased efficiency in their processes. Flake particles ar e desirable from several viewpoints. Due to its high aspect ratio, a flake particle has a larger sp ecific surface area than a s pherical particle of the same volume, which can enhance chemical r eactivity (Cashdollar 2000; Trunov et al. 2005). High aspect ratio metal flakes, such as those ma de of aluminum, have been used to increase optical obscurence characteristics so as to mimi c a metallic look when used as pigment in paint (Smith et al. 2003). Additionally, the use of flake powders for metal-hydride electrodes in modern batteries has shown significant improvement in electrical capacity (Yoshinaga et al. 2002; Yoshinaga et al. 1997). Processes for th e production of flake-shaped particles are relatively new and have yet to be extensively studied and mastered. Conventional methods for flake particle synthesis are Attritor milling, vi bratory ball milling, ball milling and wet milling (Hong and Kim 2001; Hong et al. 2000; Kilinc et al. 2004). Attritor mills generally refer to the use of a stirring rod and pin to agitate a solution of grinding media and the target particles. During such agitation, the random collisions of balls statistically occur with the target particle in between, eventually flattening the particle after successive impactions (Kilinc et al. 2004). This me thod, however, can take at least 5 hours to

PAGE 68

68 produce consistent flake particles, wasting a larg e amount of energy to fluid friction and fruitless collisions, and also requires milling to occur in a liquid solution, which can react with the target particle (Kilinc et al. 2004). A vibratory mill oper ates under similar circumstances as that of the previously mentioned attritor mill; however, it pr oduces a high frequency of randomly colliding ball media by the means of large vibrating forces and it does not require a liquid milling solution (Dunlap et al. 2000). Though this method allows for the highest kinetic transfer of those mentioned, it is also bounded by long opera ting times (up to 8 hours; e.g. SPEX model 8000)(Dunlap et al. 2000). Ball milling methods typi cally require lubricants and anticoagulants such as stearic or oleic acid a nd/or mineral spirits to control th e shape and quality of the product flake (Hong and Kim 2001; Hong et al. 2000). These reagents are incompatible when reactive metals such as Al, Fe, Mg, and Cu must remain in their elemental form and can possibly react with the reagents during milling, especially when high energy transfers are present. Wet milling methods such as basket mills are not compa tible for similar reasons. Furthermore, these conventional methods are very time-consuming, ranging in processing time from 0.5 to 60 hours (Hong and Kim 2001; Hong et al. 2000; Kilinc et al 2004; Lee et al. 2002; Yoshinaga et al. 1997). As an example, Hong and Kim (2001) reporte d that processing scrap aluminum particles in a wet ball mill syst em with oleic acid requires 30 hours to produce flake particles of micrometer thickness. The objective of this chapter was to develop a novel high-speed vibratory mill process for the production of flake particles so as to signifi cantly reduce the milling time required to achieve micrometer thickness without the need of subsid iary milling aides. During milling operations, a cycle of particle fragmentation and reformation wa s observed. The effects of plastic deformation, particle-to-particle welding and flake fragmentat ion of this cycle on pa rticle size distribution

PAGE 69

69 were examined. Several operating parameters of the milling process were studied, including milling time, particle medium (i.e. Fe, Mg), initia l particle size, loading weight and ball media size. Mechanisms High Speed Orbiting of Ball Media The milling operation occurs within a closed milling tube, where spherical particles are subjected to strong compression and shear for ces produced by ball media that flatten the particles. The motion of ball media in the millin g tube is illustrated in Figure 4-1 (a). Compression and shear forces are produced by ba ll media that roll along the smooth inner wall of the milling tube at high speeds (several t housand rotations per minute) forming sustained orbiting regions. Figure 4-1 (b) is a photograph demonstrating the orbits created by the rolling ball media that are visible as long circular blurs functioning inside a glass test-tube. Oscillations in the xz-plane of the milling tu be produce the impetus for the media motion. Figure 4-2 illustrates the frame-by-frame move ments of the milling tube and the resulting movement of the containing media. The motion of the milling tube is indicated as going in the direction from the gray position to the black posit ion. The oscillations of the milling tube from position 1 2 3 4 repeated, cause an acceleration with a continuously changing direction away from the milling tube center indicated as a gr ay arrow. In turn this causes the internal ball media to naturally find the opposite-most point from the direction of milling tube movement. This point is indicated by the position of the dark and filled ball media in Figure 4-2. With a continuous change in the direction of the milling tube acceleration, the po sition of the ball media likewise changes continuously. Ulti mately this motion causes the in ternal media to roll along the inside edges of the milling tube at the same frequency as the milling tube oscillation.

PAGE 70

70 Plastic Deformation and Welding Effects In the context of the presented novel milli ng system, plastic deformations are shape changes occurring to the initial particle samp le during milling due to the high compression and shear forces from the orbiting ball media. Deforma tion occurs primarily in a way that results in very thin and wide flakes. Furthermore, progre ssive deformation can lead to fragmentation of flake particles, resulting in many smaller flakes. Evidence of this can be seen as an increase in the fraction of smaller particles. Due to the high frequency of orbiting media a nd chaotic swirling motion of particles inside the milling tube, random flake overlaps readily oc cur. Composite welding occurs as various flakes are joined to one another by random overl appings resulting in a single compound flake. Instances of compound flake-to-flake welding have been reported for various types of milling devices (Huang et al. 1995; Kilinc et al. 2004; Smith et al. 2003; Yoshinaga et al. 2002; Yoshinaga et al. 1997). Since there are both la rge and very small particles accumulating, the resulting compound flake may be large or not so larg e, leading to a wider di stribution of sizes for welding dominant systems. Howeve r, a compound flake must always be greater in size than its constituents, resulting in a shif t of the particle size distribution towards larger particle size ranges. Experimental Experimental System For this research, a lab scal e version of the novel vibrat ory mill process was built. A milling tube was constructed using 10 mm internal diameter aluminum tubing with a length of 50 mm. Milling tubes were mounted via plastic colla rs perpendicular to a high-speed vibrator, which provided 13,000 oscillations per minute (O PM) at 120 V. Chromium-Steel Cr-52100 balls (Norstone Inc.) with diameters of 2.0 4.0 and 6. 0 mm were used as ball media in the milling

PAGE 71

71 tube. The area between the first orbiting ball and the last orbiting ball is defined as the milling zone. Whereas, the actual area which will receive the milling forces is defined as the milling area and is the sum of the ball diameter projections on the milling tube, disregarding the inter-orbit gaps, which are considered ineffectual. This met hod of classification is illustrated in Figure 4-3. The milling area was fixed as being equal to a tota l of 12 mm for all the balls in the milling tube such that the number of balls in each tube w ould be 2, 3 and 6 for the 6.0 mm, 4.0 mm and 2.0 mm sizes, respectively. This arrange ment was chosen with the aim in mind to make the effective milling area approximately equal for all samples so as to replicate the effective milling area for all ball sizes. Two types of pa rticles were tested: 300 m ma gnesium (Alfa Aesar) and 140 m iron (Fisher Sci.). Each milling tube was sealed under ambient atmosphere with plastic end caps to contain the ball media and particles. Particles were weighed and loaded into each milling tube with the ball media in the tube. After processing, the pr oducts were separated from the ball media and then stored in labeled vials for analys is. Table 4-1 displays the parameters tested for each sample. Product Characterization Three basic characteristics of the resulting fl ake particles were analyzed, including mean particle diameter, mean particle thickness and morphology. Flake particle samples were evenly distributed on glass substrates and observed under an optical microscope (Olympus BX-60) using Spot Advanced (Diagnostic Instrument, INC) image capture software to acquire sample images. The sample images were then processed w ith Image Pro Plus (Media Cybernetic, L.P) to calculate particle number and dimensional stat istics for each sample photo batch. Martin diameter, which is defined as the length of the line parallel to a given axis that divides the particle into two equal area re gions, was measured. Approximately 4000 particles were analyzed per sample. To determine the thicknesses, flake particles were suspended and solidified in an

PAGE 72

72 epoxy resin. The dried epoxy was then cut transversall y to obtain cross sectional slices of flakes suspended in the epoxy and subsequently examined under optical microscopy. Optical microscopy was also used to study the surface mor phology of bare flake particles for evidence of increased surface cracks and multiple-flake layering from flake-to-flake welding. Because of the nature of the milling proce ss, and provided the understanding that the size distribution of particles is dependant on the fr agmentation and welding qualities of the milling parameters, its distribution does not necessarily resemble a Gaussian curve. In this study, mean particle size and standard deviat ion were evaluated side-by-side, which can effectively signify in what regime (fragmentation/welding domina nt) the sample distribution reflects. Results and Discussions 300 m Mg Particle The original particle size distributions for 300 m Mg are displayed in Figure 4-4 (a). The effects of increased milling time on particle size distribution (Martin diameter) and thickness of 300 m Mg particles were studied. The results for samples 1 and 2, after milling for 1 and 2 minutes respectively, are displaye d in Figure 4-5 (a) and (b). The resulting mean diameters for samples 1 and 2 were 372 and 442 m and the standard distributions were 195 and 386 respectively. As show n in Figure 4-5 (a) and (b), the majority of the sample remains near 400 m after both 1 and 2 minutes of milling. The peak fraction in the mean size range decreased from 1.2 in sample 1 to 0.8 in sample 2, while fractions of larger particles increased. This shows th at a significant amount of the pa rticles in the mean size range were redistributed into larger particle sizes and the size distributi on became broader with increased milling time. This growth and broadening in particle size distribution can be explained by plastic deformation and particle welding as described earlier. The lack of change in the smaller particle size regime implies that the particle fragmentation was not as important as

PAGE 73

73 welding, and that there does not seem to be a ch ange in the dominant m echanisms between 1 and 2 minutes. It can therefore be concluded that, sinc e the mean diameter is large and increasing and that the standard deviation is also broadening, plastic defo rmation and composite particle welding are the dominant mechanisms over particle fragmentation in these samples. The change in flake thickness was analyzed using microscopic images of epoxy resin slices containing the flake partic les. Figure 4-6 (a) displays the change in thicknesses in samples 1 and 2. With increased milling time, it can be seen that the flake thickness distribution decreases, similar to the flake diameter trend. The mean particle thickness decreased from 300 m of unprocessed particle s to approximately 35 m after only 1 minute. Within just 2 minutes of milling, there was an even greater decrease in flake thickness, with an average flake thickness of 12 m. Plastic deformation is responsible for this trend, since with the start of milling, flakes are continually molded thinner. It should be poi nted out that due to the nature of the milling process, not all particles are milled simultaneously During the early periods of milling, some are fully milled while others are yet to be milled. As milling time increases, the fraction of un-milled particles decreases, and the amount of milled particles ar rives at the total therefore reducing the sample discrepancy caused by the remaining un-mille d particles. This explains the decrease in the flake thickness distribution in Figure 4-6 (a). In short, the fraction of the thicker and unmilled particles greatly decreases with processing time. Figures 4-6 (b) and (c) are the microscopic images of representative flake cro ss-sections from samples 1 and 2, respectively. The results of samples 1 and 2 demonstrate the novel high-speed vibratory milling as a very effective and efficient process for the production of flake-shaped particles. 140 m Fe Particle Experiments were also carried out for Fe part icles. The original particle size distributions for 140 m Fe particles are disp layed in Figure 4-4 (b) and the results for samples 3 and 4 are

PAGE 74

74 displayed in Figure 4-7. The resulting mean diameters for samples 3 and 4 were 238 and 241 m and their standard deviations 63.4 and 89.4 respectively. It can be seen from Figures 4-7 (a) and (b) that the majority of the particle size resides in the mean range of 240 m. Similar to Figures 4-5 (a) and (b), the peak fraction decreased and larger particles developed, resul ting in a wider particle size distri bution. Similar to samples 1 and 2 of magnesium particles, the most possible explanation for the broadening particle size distribution is the effect of pl astic deformation and particle we lding consequences of the high compression forces imposed on the flake particles in side the milling tube. In the smaller particle size range, there is little change in particle si ze distribution below the mean particle size. This further affirms that samples milled up to 2 minu tes do not show significant signs of particle fragmentation dominance. Rather, because of the particle size distribution shift in the larger range, welding is the dominant mechanism that influences the particle size distribution in samples 3 and 4. Comparing the results of samples 1, 2 and 3, 4 can reveal the effects of particle material. Ductility is the primary factor that determines the samples dimensional results. Ductile and malleable materials (ones which have a low Youngs Modulus; Mg 45 MPa, Fe 211 MPa) tend to have larger particle size distribution chan ges over time, as they are more vulnerable to the compression and shear forces present during milling. In addition to size characteristics, surface qua lities of the flake particles were observed under optical microscope. Figure 4-8 shows the ac quired images of the Fe flakes. Two main characteristics of the flakes were examined: surface morphology and evidence of flake-to-flake welding events. Comparing Figures 4-8 (a) and (b), it can be seen that the surface of these iron particles became noticeably smoother with increased time.

PAGE 75

75 Additionally, evidence of flake-to-flake welding can be seen (circled in both Figures 4-8 (a) and (b)). The flake pieces circ led are likely to be in an in termediate phase of the entire process; a loose flake is first taken up by a nother and then eventually milled until the two compress into each other to the point that they are indistinguishable. Effect of Ball Size To investigate the effect of ball size on the milling operation, in regards to mean particle size and particle size distribution, two additional samples were prepared with smaller ball sizes than that of sample 2. Samples 2, 5, and 6 can be compared for the effect of the milling ball size, with all other parameters kept c onstant. Their particle size distri butions are graphed in Figure 4-5 (c) and (d). It can be seen clearly that a simple trend exists between the resulting mean particle size and the ball media size: as the ball size is reduced, the mean particle size reduces (442, 172 and 84 m for 6, 4 and 2 mm ball media, respectively). This trend can be explained by the larger ball sizes creating higher compression forces due to their greater mass, resulting in thinner and wider particles. A more detailed inspection of the particle size distributions gives clue s in regards to the welding and fragmentation characteristics of eac h sample. As mentioned earlier, fragmentation of initial flakes, by its own nature, will result in greater numbers of smaller flakes. Conversely, welding will result in generally larger flake par ticles, and depending on the extent and number of weldings, can have a broad range of particle sizes. It can be seen in Figure 4-5 (c) and (d) that there is a much higher fraction of smaller particles present in sa mple 6 (2.0 mm balls) than in sample 5 (4.0 mm balls). In other words, ther e is a broader midrange distribution and a lesser fraction of small particles in sample 5 (4.0 mm balls) than in sample 6 (2.0 mm balls). This indicates that smaller ball media tend to show a dominance of fragmentation over welding, while

PAGE 76

76 larger ball sizes tend to have a more profound we lding impact, resulting in a broader distribution of particles. This vision is al so supported by the standard devia tion values for samples 5 and 6. The low welding qualities of sample 6 can also be resorted to the low compression forces present in smaller media, which may not be sufficient to join two flake particles. Since fragmentation is readily present even in smaller ball sizes, all mill ing processes that use larger ball media have fragmentation occurrences. They differ only in th eir welding ability, which so far has been the most influential factor of the samples particle size distribution. Effect of Weight Loading The effect of weight loading of Mg in the milling tube was studied by comparing three contrasting weights. Samples 7 a nd 8 were prepared of 15 mg and 35 mg respectively, and were compared to the similar sample 2 of 25 mg. Their particle size distributio n graphs are displayed in Figure 4-5 (e) and (f). Comparing the trends of samples 7, 2, and 8, shows the effect of increased loading weight. As the loading weight increases, the size distribution can be seen to shift from smaller to larger particles. The mean diameter data for each sample also supports this observation (55, 442 and 997 m for 15, 25 and 35 mg, respectiv ely). As mentioned earlier, smaller particle sizes, like those present in samp le 7, imply a dominance of fragmentation and a deficit of welding instances. Since welding can only occur when particle overlapping and milling coincide, the frequency of welding is influenced by the probability of particle overlapping, which is directly related to the particle concentration. As the loading weight of the sample is increased, the concentration increases and th e probability of overlapping incr eases as well, resulting in a greater amount of particle-to-particle welding and ultimately a shift in the particle size distribution towards a midrange size, with a typically larger standard deviation.

PAGE 77

77 Summary Several aspects of flake characteristics pr oduced by a novel vibratory mill process were studied, including: flake diameter, thickness, a nd morphology as a function of processing time, ball media count and weight loading. The result s of these data showed similar particle size distribution trends in samples of magnesium and iron. With increased milling time, more particles in the larger size ranges developed. This expanding distribut ion can be explained largely because of the effects of plastic deform ation and composite particle welding. The more ductile magnesium particles tended to show faster changes in their particle size distributions, since they were more subjective to the forces that caused the shifti ng distribution. Flake thickness of magnesium was found to decrease as a function of time, as did the particle thickness distribution. With compression a nd shear forces caused by the or biting ball media, the flake particles were continuously milled thinner, eventually to a mean thickness of 12 m after 2 minutes of milling. Additionally, as milling proceed ed, the influence of partially milled particles reduced and welding occurrences continued thus lowered the standard deviation of flake thickness with time. The increasing weight of the milling ball media resulted in higher compression forces, and directly affected the mean particle size and its distribution based on their greater welding effects. Smaller ball media tended to show a dominance of fragmentation over welding, while larger ball sizes tended to have a more profound welding im pact, resulting in a br oader distribution of particles. Increasing the concentration of fl ake particles (weight loading) increased the probability of overlapping and ther efore resulted in a greater am ount of particle-to-particle welding. This led to a shift in the particle size distribution towards a midrange size, with a larger standard deviation.

PAGE 78

78 In summary, a novel vibratory mill consisting of ball media orbiting at high revolutions inside a milling tube produced high compression a nd shear forces capable of processing 25 mg of 300 m spherical magnesium particles in to flakes of mean diameter 442 m and thickness of 12 m in only 2 minutes. As demonstrated, the novel high-speed vibratory mill process is a very efficient and effective process for the production of flake-shaped particles from ductile metal particles.

PAGE 79

79 Table 4-1. Experimental conditions Sample Particle medium and size (m) Weight loading (mg) Ball media count and size (mm) Milling time (min) 1 Mg, 300 25 2 x 6.0 1 2 Mg, 300 25 2 x 6.0 2 3 Fe, 140 25 2 x 6.0 1 4 Fe, 140 25 2 x 6.0 2 5 Mg, 300 25 3 x 4.0 2 6 Mg, 300 25 6 x 2.0 2 7 Mg, 300 15 2 x 6.0 2 8 Mg, 300 35 2 x 6.0 2

PAGE 80

80 (a) (b) Figure 4-1. (a) Schematic diagram of orbiting mo tion; (b) Photograph of orbiting media in glass tube (the circular blurs are trails of ball movement)

PAGE 81

81 Figure 4-2. Schematic diagram of basic orbiting media mechanisms

PAGE 82

82 Figure 4-3. Milling zone and milling area classification

PAGE 83

83 Figure 4-4. Particle si ze distributions for orig inal particles: (a) 300 m Mg and (b) 140 m Fe

PAGE 84

84 Figure 4-5. Particle size distributions fo r Mg particle (300 m): (a) sample 1, (b) sample 2, (c) sample 5, (d) sample 6, (e) sample 7, and (f) sample 8

PAGE 85

85 Milling Time, min 0123 Thickness of flakes, m 5 10 15 20 25 30 35 40 45 (a) Figure 4-6. Thickness of Mg particle as a function of milling time: (a) particle thickness distribution, (b) crosssectional images for 300 m Mg particle after 1 minute and (c) 2 minutes

PAGE 86

86 Figure 4-7. Particle size distributions fo r iron particle (140 m) as a function of milling time: (a) sample 3 and (b) sample 4

PAGE 87

87 Figure 4-8. Optical microscope images at 20x magni fication of iron particle: (a) sample 3 and (b) sample 4

PAGE 88

88 CHAPTER 5 ENHANCED HYDROGREN ABSOPRTION KINETICS FOR HYDROGEN STORAGE USING MG FLAKES FABRICATED BY A HIGH SPEED ORBITING BALL MEDIA PROCESSOR Background Among the various light metals and alloys th at are capable of abso rbing/desorbing large amounts of hydrogen, magnesium has been consid ered a promising candidate for solid hydrogen storage due to its high volumetric/gravimetric cap acity, ease of availability and low cost (Chalk and Miller 2006; Schlapbach and Zttel 2001). The hydrogenation process starts from physisorption of hydrogen molecules onto the su rface of magnesium particles, followed by dissociation into hydrogen atoms. The hydrogen atoms thus obtained diffuse into magnesium lattices to initiate nuc leation and growth of magnesium hydr ide (Bloch and Mintz 1997; Conner and Falconer 1995; Fukai 2005). However, the majo r disadvantage of util izing pure magnesium is its slow absorption of hydrogen, which is du e to its low affinity for hydrogen physisorption, and the formation of a dense magnesium hydrid e layer (Atkins and DePaula 2002; Fukai 2005; Schlapbach and Zttel 2001). The hydride layer becomes a resi stive barrier for subsequent hydrogen diffusion into the bulk and therefore resu lts in limited material utilization. This ultimately becomes a crucial rate-limiting mechanism in the hydrogenation process (Bloch and Mintz 1997; Friedlmeier and Gr oll 1997; Jeon et al. 2006). Various attempts have been undertaken to chemically improve the kinetics of Mg by incorporating transition metals such as Ni, V, Fe and Ti (Holtz and Imam 1999; Huot et al. 1999; Iwakura et al. 2002; Iwakura et al. 1999; Jeon et al. 2006; Liang et al. 1998; Varin et al. 2006; Zaluska et al. 1999). The addition of transition metals with high affinities for chemisorption results in interfacial catalysis of H2 dissociation thus reducing th e activation energy required for

PAGE 89

89 dissociation of hydrogen molecules (Sandrock et al. 2002; Stillesj o et al. 1993). The resultant hydrogen atoms can more easily diffuse than hydrogen molecules along grain boundaries of magnesium (Lueking and Yang 2004; Rozanov a nd Krylov 1997). In addition, it has been demonstrated that the onset temperatur e for hydrogen absorption can be lowered by approximately 170C by adding a small amount of Ni onto Mg (Holtz and Imam 1999). Studies report that a homogeneous distri bution of nano catalysts on the metal drastically reduces the activation energy of hydrogen diffusion and results in faster formation of metal hydride even in the presence of oxides and hydroxides (Au 2005; H uot et al. 2001; Jeon et al. 2007; Seayad and Antonelli 2004; Zaluska et al. 2001). However, it should be no ted that when hydrogen diffusion was the rate limiting mechanism, Ni wt. % loading showed negligible impact on H2 absorption rate and capacity, (Jeon et al. 2006 ). In addition, the particle di ameter of the catalyst has been also recognized as an influentia l parameter. Varin et al. (2006) investigated the effect of Ni particle size (micro, submicron and nano sizes) on the hydrogen absorption/desorption kinetics of 44 m Mg powder. The addition of nano Ni wa s observed to greatly improve the hydrogen absorption/desorption rate as compared to micron/submicron sized catalysts. The kinetics of hydrogen absorption and deso rption of metal hydrides have also been improved mechanically by increasing the spec ific surface area and grain boundaries, reducing the grain size of the metal powders, and altoge ther enhancing diffusion of hydrogen atoms within the metal leading to increased hydride forma tion (Doppiu et al. 2004; Holtz and Imam 1999; Huot et al. 1995; Huot et al 1999; Iwakura et al. 2002; Liang et al. 1998; Revesz et al. 2006; Tessier and Akiba 1999; Varin et al. 2006; Yu et al 2002; Zaluska et al. 1999). Most past research studies of magnesium particles have been conducted based on powders processed by conventional ball milling due to their ability to change the microstructure of material. However,

PAGE 90

90 the major disadvantages of mechanical ball milling are the excessive milling time required (3060 hrs) (Doppiu et al. 2004; H uot et al. 1995; Tessier and Akib a 1999; Yu et al. 2002) and the formation of undesirable alloys. Varin et al. (2006) reported that af ter 20 hrs of milling, formation of Mg2Ni alloy was observed, which results in a lowered hydrogen absorption capacity. In addition, it is challenging to cont rol the shape and consis tency of the product (Hong and Kim 2001; Hong et al. 2000; Yoshinaga et al. 1997). Therefore, a new milling system is desired which has significantly reduced operatin g time and is capable of reducing grain size without the formation of metal alloys. Furthermore, a better unders tanding of the overall process of hydrogenation is needed to further develop the practicality of magnesium as a storage option. To expl ain the change of the rate limiting mechanisms in hydride formation, Je on et al. (2007) experimentally demonstrated that formation of a MgH2 layer starts from the outer layer of particle surface and grows into the bulk. Thus, in the beginning of the process, MgH2 nucleation is the rate liming mechanism; however, hydrogen diffusion through the impermeab le hydride layer eventually becomes the rate-liming mechanism once the hydride layer reaches a crit ical thickness. The maximum penetration depth for nucleation limited condition for the tested sample was calculated to be 6 m, i.e. if the layer is less than 6 m, diffusion barrier is not longer a critical problem (Jeon et al. 2006). Since a short fueling time is critical to pr actical applications, it is advantageous to accomplish the majority of hydrogen uptake while it is in the most ra pid nucleation-limited stage. In order to reduce the limiting effect of diffusion through the compact MgH2 layer and thereby extend the nucleation-limite d stage, particles with a hi gh specific surface area and a narrow thickness are highly favored as they maximize surface reaction sites and minimize the

PAGE 91

91 diffusion depth. Nano-sized Mg particles and thin Mg films have been used to minimize this problem; however, practical production costs of these materials may be a concern. Additionally, the effect of geometry on the hydrogen absorption kinetics of magnesium particles has not yet been explored in de pth. Development of new material production technologies that improve material utilization and increase produc tion efficiency is very much sought after. Recently, a high speed orbiting ball media (HSOBM) proces sor was successfully introduced as a fast and efficient method to fabr icate flake-shaped magnesium particles with a high specific surface area and a narrow thickne ss specifically for use as hydrogen storage materials (Theodore et al. 2006). The objective of this chapter was to maximize the nucleation-limited regime, to improve material utilization and to enhance hydrogen uptak e by changing either geometry or grain size of Mg. Spherical particles and flake particles were characterized to measure grain size and to evaluate its effect on hydrogenation. The HSOBM processor was also te sted for its ability to coat the nano-catalyst on magnesium flakes in orde r to evaluate its potential for combined flake/coating processor. The coating effici ency of both the HSOBM processor and Theta Composer were assessed. Finally, the effects of geometry on hydrogen absorption kinetics were analyzed from hydrogenation data. Experimental Methodology Fabrication of Magnesium Flak eNano Nickel Composites Mg flakes were prepared using a HSOBM proce ssor described in Chapter 4. The coating of nano Ni catalysts was carried out using eith er a HSOBM processor or a Theta Composer (Tokuju, Corp.) under an argon environment. In-dep th details of the mechanistic aspects of the processes have been described elsewhere (Coow anitwong et al. 2003; Jeon et al. 2006; Theodore et al. 2006), and only a brief de scription is provided here.

PAGE 92

92 The Theta Composer consists of an elliptic ro tor encased in a vessel. The rotor operates at high revolution while the vessel counter-rotates at a lower speed. During rotation it applies strong compression and shear forces to the partic les as they pass through the narrow gap between the rotor and vessel wall (Cooper et al. 2005; Coowanitwong et al. 2003; Jeon et al. 2006). The process also more evenly distributes the Ni nanoparticles onto the magnesium surface. The rotor and vessel are made of stainless steel in order to maintain an inert environment and prevent chemical reactions. The HSOBM processor uses ba ll media rolling on the walls inside a tube at high speed to create strong compression and shea r forces, flattening pa rticles as they pass between the rolling media and the vessel wall. In a likewise manor, the high speeds and large compression and shear forces are capable of breaking up nano-agglomer ates and coating the nanoparticles on the Mg flake. The basic procedure used in creating Mg-Ni flake composites is as follows: 25 mg magnesium (Fisher Scientific, 294 m) and 6.0 mm balls (Chromiu m-Steel Cr-52100, Norstone Inc.) were introduced into the HSOBM tube in a glove box (oxygen leve l was lower than 0.5 ppm). The milling system was operated with an orbiting speed of 13,000 RPM, and with 2-min milling time. After each milling process, 3 wt. % of the nano Ni (primary particle size: 72 30 nm, surface area: 6 m2/g, bulk density: 0.66 g/cm3, Argonide) was added to the Mg flakes. Then, the mixture was coated for 2 minutes with ba ll speeds of 13,000 RPM in the HSOBM (2.0 mm ball sizes) or for 90 minutes with the Th eta Composer running at 4600 70 rpm. The surface area analysis was performed on a Quanta Chrome NOVA 1200 Gas Sorption Analyzer using N2 adsorption isotherm with a multi-point Brunauer-Emmett-Teller (BET) method. Scanning Electron Microscopy (SEM JEOL JSM 6335F) coupled with Energy Dispersive Spectroscopy (EDS) was used for comparing surface morphologies on the

PAGE 93

93 magnesium powders before and after the milling process and dispersion of nanocatalysts on the surface of Mg flake. To determine the thickness of the flake particles, they were suspended and solidified in an epoxy resin. The dried epoxy wa s then sliced transversally to obtain cross sectional slices of flakes suspended in the epoxy and subsequently examined under an optical microscope (Olympus BX-60). The formation of MgH2 in the MgNi flakes composites, the orientation of crystal structure and the grain refinements and la ttice strains of flakes were determined by X-ray diffraction (Phillips XRD APD 3720) with a 2 range of 20-70 with three seconds of count time per step. Hydrogen Absorption/Desorption Characterization Hydrogen absorption experiments were ca rried out in a hydrogenation system. The hydrogenation system is shown in Figure 5-1. 0.905 g of the MgNi flake composite was introduced into the hydrogenation chamber while in a glove box. The chamber was flushed three times with argon to minimize possible contamin ation. In vacuum, th e hydrogenation chamber was heated to the target temp erature (483 K). Pure grade hydr ogen (4.8) was then instantly introduced into the hydrogenation chamber at 1 MPa and hydrogen pressure was maintained constant in the chamber. The temperature of the chamber was monitored by a thermocouple (Fisher Scientific model 15-078-39, type K) a nd the hydrogen pressure in the reservoir was monitored using an electronic pre ssure transmitter (Omega, PX880). After the hydrogenation process, the Mg-Ni flake composites were analyzed by a Thermo Gravimetric Analyzer (TGA 2050, TA Instrume nts) to determine the amount of hydrogen adsorbed in the MgNi flake composites. The PV /RT value was utilized to aggregate pressure and temperature effects simultaneously and used to compare experimental data to theoretical prediction. The PV/RT data, calcul ated from the measured pressure loss, were accurate to within 1.5 % of the experimental data from TGA (Je on et al. 2006), and were used in this study.

PAGE 94

94 Although there are many parameters that affect hydrogen absorption/desorption rate, two of the more important parameters, namely change in part icle geometry and Ni lo ading, were evaluated and other conditions were kept constant in this study. The experimental conditions are listed in Table 5-1. Results and Discussion Determination of Minimum Ni Loading Prior to experimentation, the minimum nano Ni loading for spherical and flake-shaped particles was theoretically investigated. A mini mum nano Ni wt. % is defined as the Ni wt. % required to form monolayer coverage on the surface of Mg powder. For spherical/flakes particles, the minimum Ni loading can be expressed by Equation 5-1 and Equation 5-2 respectively: 1 ) )( ( 1 1 n m n m sphd d X (5-1) 1 ) )( ) 2 ( ( 3 1 n m m n m flah d d h d X (5-2) The definition of each parameter is li sted in the Nomenclature section. The average thickness and radius of flakes, measured to be 11.73 m and 442 m respectively under optical microscopy (Olympus BX-60) (Theodore et al. 2006), were used in this calculation. For 70 nm Ni, Xfla was founded to be 3.14 wt. % for flakes while Xsph was calculated to be 2. 1 wt. % for the 44 m Mg. Thus, Nickel loading in excess of 3 wt. % was not considered. Furthermore, loss in hydrogen capacit y resulting from a heav ier weight could render

PAGE 95

95 the materials less desired for onboard storage (C halk and Miller 2006; DOE 2007; Zaluska et al. 1999). Characterization of Mg Flakes (SEM/BET/XRD) Geometric examination of the magnesium pow ders before and after the HSOBM process was conducted using SEM. By co mparing the images of pure 294 m Mg and Mg flakes in Figure 5-2 (a) and (b), it was observed that af ter milling with HSOBM, Mg showed observable physical changes (i.e. evidence of cracking or fo rmation of surface roughness) and reduction of thickness. Specific surface area of the powders be fore and after processing in the HSOBM was also measured using BET; however, the surf ace area was of too low magnitude (< 0.5 m2/g) to have reliable results. The XRD spectra of the struct ural evolution of pure Mg ar e shown in Figure 5-3 (b) and (c). After 2 min milling with HSOBM, the 1st highest peak (36.6o) for pure Mg was significantly reduced in intensity. On the ot her hand, the intensity of the 2nd highest peak for pure Mg reached a significant peak after 2 min milling using th e HSOBM processor. It was noted that the orientation of crystal structure of Mg, shown in Figure 5-3, was changed from 36.6o (orientation: 101) to 34.5o (orientation: 200). This could be due to plastic deformation resulting from shear/compression forces experienced by Mg during the HSOBM milling process (Callister 1997; Theodore et al. 2006) which w ould result in a change in the or ientation of crystal structure. The change in turn caused grain re finement of Mg flakes and cha nge of lattice stra ins. The grain refinements and lattice strains of Mg flakes for 34.5 o (orientation: 200) peaks were also determined by single line profile analysis (Skoog et al. 1998) in addition to XRD. It was observed that the average gain size of flakes (316.3 nm with 0.26 % la ttice strain) with two minutes of processing time was smaller than those of 294 m Mg powder (423.4 nm with 0.19 %

PAGE 96

96 lattice strains). The average grain size of 44 m Mg powder is 132.3 nm with 0.16 % lattice strain. The crystalline phase of Mg and Ni was preserved during processing and coating by HSOBM and Theta Composer, and no XRD peaks of MgxNiy alloy were observed, as shown in Figures 3 (a), (d), and (e). Afte r hydrogenation, the formation of MgH2 was observed in samples 1, 4, and 5. In short, the HSOBM processor is able to fabricate smaller grain size with higher lattice strains with a short processing time. Effect of Grain Size on the Hydrogen Ab sorption Kinetics without Nano Catalyst Several studies (Dornheim et al. 2007; Dor nheim et al. 2006; Huhn et al. 2005; Zaluska et al. 2001) have reported that th e reduction of grain size incr eases the number of paths for hydrogen diffusion as well as nucleation sites for reaction, which altogether enhances hydrogen absorption kinetics. Past studies have shown that particles with nano-sized grains, which have been mechanically milled for long periods (gre ater than 300 hours), have appreciable hydrogen absorption without the need for catalysts while also reacting at relatively low temperatures (Dornheim et al. 2007; Huhn et al. 2005). To inves tigate the effects of average grain size of Mg powder before and after milling on hydrogen absorption capacity, 294 m spherical Mg powders were milled for 2 minutes to form 442 m flakes (with 11.7 m thickness). Hydrogenation tests were performed for these samples that had no Ni cat alyst coating. Figure 54 shows the effects of grain size on the hydrogen absorption capacity. Although the thin flake-shaped particle had smaller grain size, as shown, there was no evidence of hydrogen ab sorption in eith er of the pure Mg samples. Therefore, it can be inferred that in the absence of Ni catalysts the reduction of grain size produced by the HSOBM processor ha s negligible effect on hydrogen absorption capacity. This confirms the precedence of the Ni catalyst for feasible hydrogen absorption on Mg particles. It also shows a devi ation from previous studies whic h indicated that powders with

PAGE 97

97 reduced grain sizes hydrogenated without the ai d of nano catalysts (Dor nheim et al. 2007; Huhn et al. 2005). The most likely reason for this devi ation is that the flake samples did not have a sufficiently small grain size (100 nm) because of the short milling time which was constrained by the mechanical endurance limitations of the device. Analysis of Coating Methods and Change of Thickness The effective dispersion of cat alysts and geometry change we re also investigated with two dry coating methods (HSOBM and Theta co mposer). Figure 5-5 shows a homogeneous distribution of nano Ni on the Mg powder/flake surface under SEM-EDS. The dispersion of nano Ni was further assessed quantitatively using Spot Advanced (Diagnostic Instrument, INC) image capture software (Coowanitwong et al. 2003; Oh et al. 2001; Theodore et al. 2006). Thirty equally divided areas were randomly selected fro m the EDS images of Ni on the Mg substrate. The number of individual Ni clusters was determin ed in each area and statistical analyses were performed. It is observed from Figure 5-6 that 2 min coating w ith HSOBM yielded Ni clusters that were larger in size but fewer in number w ith a lower mean and standard deviation (23.7 and 5.59, respectively) as compared to that of Th eta Composer with 90 min coating time (50.76 and 8.76, respectively). Figure 5-6 also shows that 44 m Mg coated with Theta Composer yielded Ni clusters with a similar mean and standard de viation (55.4 and 8.10, respectively) as compared to Mg flakes with Theta composer (50.76 and 8.76, respectively). This comparative analysis confirmed that Ni nanoparticle dispersion was more effective with Theta composer. This is mostly a result of the short operating time of the HSOBM device as well as its high coating speeds which have been known to subdue coating efficiency due to a nano-lubrication effect (Bowden and Tabor 1986; Hsu 2004). Thickness change was measured after Theta Composer and HSOBM coating under optical microscope (Olympus BX-60). The average thickness of flakes (sample 3 and sample 4)

PAGE 98

98 was 11.73 m and 11.76 m, respectively, after coating wi th HSOBM. After Theta Composer coating the average thickness of fl akes (sample 5) increased from 11.73 m to 16.87 m. Effect of Dispersion of Nano Catalysts on Hydrogen Absorption Kinetics Figure 5-7(a) plots the hydroge n absorption data for samples 1, 4 and 5, as a function of time. It can be seen that the le ast effective material was sample 4, which was a flake particle using the HSOBM coating method. Its poor perf ormance is most likely due to insufficient catalytic enhancement and is supported by the lowe r coating quality reported in section 3.4. Samples 1 and 5 were coated using the more ef ficient Theta Composer and showed significant hydrogen absorption capacities of 4.19 wt. % and 4.62 wt. % respectively, after 300 minutes. The hydrogen absorption capacities for samples 1, 4 and 5 were also confirmed using TGA analysis and can be seen in Figure 5-7(b). Th is analysis supports the trend that hydrogen absorption capacity is largely dependant on an effective Ni catalyst co ating. In order to appreciate the effect of surface area for each samples results, the specific su rface area of each sample was estimated assuming spherical partic les or smooth flakes of uniform thickness and diameter. The results gave that samples 4 and 5 sh ared relatively close spec ific surface areas to sample 1, with ratios of 1.31:1 and 0.931:1. It can therefore be r easonably inferred that difference in surface area had negligible impact on the observed trends. Effect of Geometric Change on Hydrogen Absorption Kinetics To closely observe the effect of coating effi ciency and particle geometry on kinetics, a hydrogen absorption rate vs. metal hydride volume fraction plot is rendered in Figure 5-8. The initial observations support earlier conclusions th at sample 5 shows the hi ghest initial absorption rate followed by sample 1, and the lowe st absorption rate by sample 4. When is approximately 0.28, the rate of absorption for well coated flak es was approximately four times that of poorly

PAGE 99

99 coated flakes. Coating efficiency can therefore be seen to drastically a ffect the kinetics when comparing flakes-shaped particles. The effect of particle geom etry on kinetics can also be observed between samples of spherical and flake particles which were both eff ectively coated with Ni catalyst. Compared to sample 1 of spherical particles, sample 5 clearly shows a shift towards hi gher reaction rates and capacities. This shift of the sample regression lines towards higher hydrogen absorption can be understood as an increase in the cr itical hydrogenation capacity ( ). is defined as the point at which the hydrogen absorption kinetics switch fr om the nucleation-limited to diffusion-limited regime and is estimated by the intersection of the tangential lines coming from the beginning and ends of the sample regression lines. In fact, th e change in geometry from the spherical to the flake-shaped particle extends the nucleation-limited period, which is more favorable since its kinetics is faster than that of the diffusion-li mited regime. This ultimately leads to a higher absorption capacity in a given period of time for flake-shaped particles, as observed in Figure 57. This same analysis also supports the poor results of sample 4, as its regression line is shifted towards lower hydrogen absorption area in Figure 58. This is known already to be a result of poor coating efficiency, which is evident when co mpared with sample 5 of the same geometry. The observed effects of particle geometry on hydrogen absorption can be reasonably understood based on a simple macroscopic comp arison of hydrogen diffu sion under rectangular (flake particles) vs. radial (sphe rical particles) coordi nates. For rectangular coordinate systems, as hydrogen diffuses into the Mg flake, the flux ar ea remains relatively constant with respect to the hydrogen penetration depth. However, for spherical particles unde rstood through a radial coordinate system, as the hydrogenation layer th ickens, the available flux area reduces with decreasing radius. This results in a diffusional term which reduces as a function of time and

PAGE 100

100 subsequently brings upon a diffusion-limited case earlier than in the r ectangular coordinate system. Previous research has emphasized that the grai n size of the particle is also an important factor in hydrogenation (Dornheim et al. 2007; Do rnheim et al. 2006; Zaluska et al. 2001). For sample 1, the grain size was significantly smaller than that of the flakes in sample 5, implying that the hydrogen absorption capacity and rate of the spherical 44 m particles should be much better than that of the flakes. Si nce this was not the result in th is study, it is th erefore inferred that the effect of geometry on hydrogen capacity and absorption rate is dominant over the effect of grain size. Ultimately, these results show that the flake particle which had nearly equal coating efficiency and specific surface area as that of the spherical 44 m sample exhibited improved absorption kinetics and capacitie s over spherica l particles. Summary This study aimed at improving hydrogen storag e density characteristics of magnesium so as to enhance hydrogen uptake kineti cs and storage efficiency by utilizing thin Mg flakes with large diameters coated with nano-Ni catalysts XRD analyses of the product processed by the HSOBM process showed a change in the orientati on of crystal structure of Mg and a reduction in grain size; however, this change was not la rge enough to hydrogenate pure Mg flakes. The individual crystalline phase of Mg and Ni was preserved and no XRD peaks of MgxNiy alloy were observed after processing with the HSOB M processor or when coating with Theta composer. Comparisons between samples of flakes without Ni coating, and with coating by the HSOBM processor and Theta Composer confirme d the need of well di spersed nanocatalysts particle for practical hydrogenation systems. Hydrogenation absorption trends showed that flakes had longer nucleation-limited periods than spherical particles, which lead to higher initial absorpti on rates and higher final

PAGE 101

101 absorption capacities. Geometry was identified to be a critical f actor affecting hydrogen absorption capacity and kinetics, and was do minant over the effects of grain boundary. Nomenclature dm Diameter of Mg [m] dn Diameter of Ni [m] h Initial thickness of Mg flakes [m] Xsph Minimum wt. % of catalyst for s pherical particle [wt. %] Xfla Minimum wt. % of catalyst for flake particle [wt. %] Greeks Metal hydride volume fraction [-] m Molar density of Mg particle [mole/m3] n Molar density of Ni particle [mole/m3] Critical hydrogenation capacity [-]

PAGE 102

102 Table 5-1. Experimental conditions Sample Particle size and shape Ni (wt.%) Milling time (min) Coating time (min) Predicted (wt. %) TGA Measurement (wt. %) 1 44 m Mg sphere 2 0 90** 4.19 4.2 2 294 155 m Mg sphere 0 0 0 0.00 3 442 386 m Mg flakes 0 2 0 0.01 4 438 328m Mg flakes 3 2 2* 3.31 3.32 5 427 293 m Mg flakes 3 2 90** 4.62 4.57 coating method: HSOBM ** coating method: Theta Composer

PAGE 103

103 Figure 5-1. Schematic diagra m of hydrogenation system

PAGE 104

104 Figure 5-2. Surface morphology change: (a) samp le 2 (pure 294 m Mg, 180x); (b) sample 3 (Mg flakes, 180x); (c) sample 5 after Theta Composer (294 m Mg Ni composites, 2000x );(4) sample 4 (Mg flake-Ni composites, 2000x)

PAGE 105

105 [Intensity] [Intensity] [Intensity] [Intensity] MgH2Ni Mg (d) (c) (b) (a) 203040506070 [Intensity](e) Figure 5-3. Comparison of XRD patterns: (a) sample 4 after hydrogenation; (b) sample 2 before HSOBM processing; (c) sample 3 after 2min HSOBM processing; (d) sample 1 after hydrogenation; (e) sample 5 after hydrogenation

PAGE 106

106 Hydrogenation time, min 0246810121416 PV/RT, mole 0.020 0.025 0.030 0.035 0.040 0.045 Temperature, K 250 300 350 400 450 500 PV/RT, sample 2 PV/RT, sample 3 Temperature, sample 2 Temperature, sample 3 Figure 5-4. Effect of morphological change on th e hydrogen adsorption kine tics (w/o catalysts, experiment)

PAGE 107

107 Figure 5-5. Distribution of Ni nano particles on Mg (a) SE M image (sample 1); (b) Ni mapping (sample 1); (c) SEM image(sample 4); (d) Ni mapping (sample 4); (e) SEM image (sample 5); (f) Ni mapping (sample 5)

PAGE 108

108 Number of Area 051015202530 Number of nano Ni clusters 20 40 60 80 100 Sample 1 Sample 4 Sample 5 Figure 5-6. Nano Ni uniformity measurements on the Mg flake at the two different coating method

PAGE 109

109 Hydrogenation time, min 050100150200250300 PV/RT, mole 0.015 0.020 0.025 0.030 0.035 0.040 0.045 Temperature, K 450 460 470 480 sample 1 sample 4 sample 5 Temperature 3.31 wt.% 4.19 wt.% 4.62 wt.%(a) Temperature ( o C) 050100150200250300350400 Weight Loss (%) 95 96 97 98 99 100 101 Sample 1 Sample 4 Sample 5 3.32 wt. % 4.20 wt. % 4.57 wt. %(b) Figure 5-7. Effect of disper sion of nano catalyst on hydrogen ab sorption kinetics; (a) hydrogen absorption; (b) Hydrogen de sorption analysis by TGA

PAGE 110

110 0.10.20.30.40.50.6 Hydrogen absoprtion rate, mole/min 0.0000 0.0001 0.0002 0.0003 0.0004 Sample 1 Sample 4 Sample 5 Figure 5-8 Effects of particle geometry and coating efficiency on the hydrogen absorption kinetics

PAGE 111

111 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS Over the past decade, the increasing demands for clean energy to reduce global warming and to secure the futures needs amidst the trem endous increase in energy consumption have led researchers to call for the development of new al ternative energy sources. This study was carried out to investigate Mg-nano Ni composites as an alternative for solid hyd rogen storage with high volumetric/gravimetric capacity. The kinetics and capacity of magnesium hydride formation on Mg-nano Ni composites synthesized by the Theta Composer as well as a high speed orbiting ball media (HSOBM) processor were characterized a nd evaluated for three key parameters for its practical application: dispersion of nano catalysts over Mg powder surface, the change of rate limiting mechanism, and the change of geometric shape of Mg powder. First, the formation of magnesium hydrid e on Mg-Ni composites synthesized by Theta Composer was investigated. It was observed that there was a change in the rate limiting mechanism of hydrogenation from chemical abso rption on the surface of the Mg particle to diffusion through the hydride layer into the Mg particle core. A longer coating time in Theta Composer improved the hydrogen storage capaci ty and absorption/desorption rates for composites with the same Ni loading wt % due to the improved distribution of Ni nanocatalysts. However, after the rate limiting mechanism changed from chemical absorption to H2 diffusion, Ni wt % loading had negligible impact on the H2 absorption rate and capacity. Regarding how hydride was formed, the SEM/TEM characteriza tion verified that magnesium hydride layer formed from the surface rather than from random nucleation and growth in Mg. This observation provides an insight into the pa rtitioning of the rate limiting mechanism between interfacial growth and diffusion, and it forms the basis for the analytical model.

PAGE 112

112 The newly developed analytical model considered multiple mechanisms and a variable hydrogen concentration. The agreement between m odeled and experimentally measured data was excellent, and the model was successfully demonstrated to be able to evaluate the change of ratelimiting mechanisms of Mg-Ni composites with regards to hydrogen absorption rate and capacity. To improve hydrogen storage density charac teristics and to enhance hydrogen uptake kinetics and storage efficiency thin Mg flakes with large diameters coated with nano-Ni catalysts were proposed. The Mg flakes were fabricated by a high speed orbiting ball media (HSOBM) processor and were effectively coat ed with Ni nanocatalyst using the Theta Composer. Comparisons between samples of flak es without and with Ni coating by the HSOBM processor and Theta Composer confirmed the n eed of well dispersed nanocatalyst particles. Hydrogen absorption trends showed that flakes had longer interfacial growth-limited periods than spherical particles with similar specific surface area. The difference led to higher initial absorption rates and larger final absorption capaciti es. Geometry was identified to be a critical factor affecting hydrogen absorpti on capacity and kinetics, and wa s dominant over the effects of grain boundary. Based on the knowledge learned and the e xperiences gained in this research, recommendations are made to help further adva nce the application of Mg-nano Ni composite. 1. A higher heating rate is benefici al to a substantial improvement in hydrogen absorption rate and storage capacity. The effect of higher heating rate on hydrogen absorption requires further investigation. A thermally insulating chamber and an automatic temperature control system can be used to improve uniformity of temperature profile as well as the stability of heating rate in the chamber. 2. The analytical model developed in this study accurately described hydrogen absorption capacity and kinetics, as it accounts for the e ffects of important m echanisms: interfacial growth of metal hydride and di ffusion of hydrogen. To further e xpand the applicab ility of the

PAGE 113

113 model, the effects of heating rate and ge ometric change of Mg powder on the hydrogen absorption should be included. 3. Due to the mechanical endurance limitations of the current HSOBM system, investigation of the effects of milling time couldnt be extended to more than 2 minutes. A new system that can overcome the existing barrier needs to be developed. The appl ication of the HSOBM processor can then be further explored for thi nner (< 10 m) Mg flakes or other materials. 4. Further studies to understand the forma tion of magnesium hydride layer on Mg-Ni composites fabricated by the HSOBM processor will embellish these results. 5. The microstructure of magnesium hydride on Mg-Ni composites fabricated by the HSOBM processor and Theta Composer should be furt her studied using elec tron microscopy and X ray diffraction. This kind of study will allow us to better understand the structural evolution that occurs during absorption and desorption. Su ch information will help us better understand the phase transformations and eventually propose further enhanced designs of hydrogen storage materials.

PAGE 114

114 APPENDIX EFFECTS OF PRIMARY PARTICLE SIZE AND INITIAL GEOMETRIC STANDARD DEVIATION ON BROWNIAN COAGULATION OF FRACTAL NANOAGGLOMERATES IN THE FREE MOLECULAR REGIME Introduction Coagulation is one common process encountered in the commercial pr oduction of specialty materials, such as catalysts, medicine, cosmetic s, food, plastic, as we ll as energetic and other advanced materials (Friedlander and Pui 2004; Pratsinis 1998; Stark a nd Pratsinis 2002; Yang and Biswas 1997). By nature, particles made by high temperature gasto-particle conversion followed by coagulation are agglomerates compos ed of nanoparticle chains (Pratsinis 1998; Schmidt-Ott 1988). Applications of tailored nanopa rticles are mushrooming and there have been many studies in recent years to understand the sy nthesis process and to c ontrol the properties. Experiments have been attempted to study th e characteristic beha vior of Brownian coagulation processes with agglomerates. For example, Matsoukas a nd Friedlander (1991) reported that the coagulation rate of metal oxide particles produced by flame synthesis increased with decreasing mass fractal dimension and primary pa rticle size. It should be noted here that the primary particle size of commercially available TiO2 agglomerates in diffusion and premixed flame reactors varies from a few nm to 20 nm (Degussa AG, TI 1234; Pr atsinis 1998). Akhtar et al. (1991) showed that nonspheri cal titania agglomerates grew much faster than spherical particles and they approached an asymptotic mass fractal dimension during high-temperature oxidation. Monte-Carlo simulations have also been widely used to explain the morphological evolution of agglomerates under going various processes such as diffusion (Oh et al. 2002), sintering and coagulation (Akhtar et al. 1994). Mountain et al. (1986) introduced a stochastic simulation algorithm. The trajecto ry of a primary particle was tracked by solving the Langevin equation coupled with a periodic boundary c ondition to calculate how deterministic and

PAGE 115

115 stochastic forces act on a partic le in motion. Meakin et al. (1986) proposed a computational algorithm to simulate the agglomeration pro cess. According to the proposed mechanism, morphology of agglomerates can be classified into two types: diffusion-limited and reactionlimited (Schaefer 1988). Aerosol dynamics models have also been developed to identify the effects of morphological structure on the behavi or of these agglomerates. Ther e are several types of aerosol models available to describe aerosol dyna mic processes based upon their mathematical description of size distribution function, such as moment (F renklach and Harris 1987; Lee and Chen 1984; Whitby 1979) and sectional models (Gelbard and Seinfeld 1980; Landgrebe and Pratsinis 1990). Wu and Friedlander (1993) simu lated the evolution of agglomerates starting with individual spherical partic les assuming a power law distributi on and examined the effects of fractal dimension and primary particle size. Vemu ry et al. (1994) proposed self-preserving timelag as a useful tool to express the time it take s to reach the asymptotic size distribution. Two correlation equations were devel oped for the self-preserving time-lag in the free molecular and continuum regimes as a function of initial geomet ric standard deviation. The calculations were implemented using a discrete-sectional model. In the follow-on study (Vemury and Pratsinis 1995), mass fractal dimension wa s incorporated into the m odel to examine its impact. Lee et al. (1990) derived the first analytical solution to track the si ze distribution evolution of a lognormally distributed spherical aerosol undergoing Brownian co agulation in the free molecular regime. Park and Lee (2002) recently expanded the analytical solution to include fractal agglomerates, which focused on the change of self-preserving tim e-lag as a function of mass fractal dimension. The sectional model (L andgrebe and Pratsinis 1990) and analytical solution (Park and Lee 2002) were compared for the predicted shape of the asymptotic

PAGE 116

116 agglomerate size distribution. Nu merically predicted asymptotic values by the sectional method for Df of 3.0, 2.5, and 2.0 were 1.462, 1.519, and 1.610, respectively, while the analytical solutions obtained by Park and Lees study ( 2002) were 1.355, 1.393, and 1.481, respectively A caveat of both studies is that they did not evaluate the effect of primary pa rticle size. The size of primary particles has been identified as an importa nt variable from previ ous experimental studies described in the preceding text. Further affirma tion of the importance of primary particle size was provided by Ulrich and Subramanian (1977 ) who described the behavior and physical properties of agglomerates generated from hightemperature gas phase process. They reported that the properties were highly dependent on th e number and size of primary particles and the agglomerate size As pointed out in these prior studies, the time to reach asymptotic distribution is of great interest to nanoparticle synthesis. The objective of this study, ther efore, was to develop a formula for self-preserving time-lag that incorporates all important parameters of nanoagglomerates, including primary particle size, geometric standa rd deviation and fractal dimension. This study focused on the influence of primary particle diameter, which was pointed out to affect coagulation rate although the e ffects have not been investigat ed by previous aerosol dynamic models. The study further examined the influence of the other two parame ters on the effect of primary particle size. Model Development To describe the particle size distribution undergoing Brownian coagulation, an integrodifferential equation (Friedla nder 2000) was used, which can be expressed as i i j i j v i i j i i j i jdv t v n v v t v n dv t v v n t v n v v v t t v nj 0 0) ( ) ( ) ( ) ( ) ( ) ( 2 1 ) ( (1)

PAGE 117

117 where ) ( t v nj is the particle size di stribution function at time t and ) (j iv v is the collision kennel for two particles having vi and vj volume, respectively. The collision kernel in the free molecular regime is 2 1 2 / 1 21 1 2 ) ( j i b aj ai j iv v T k d d v v (2) where ( dai+daj)2 is the collision cross section for agglomerates and ( kbT/ 2 )1/2[1 /vi+ 1 /vj]1/2 is the average relative velocity be tween colliding agglomerates. For fractal aggregates with a fractal dimension ( Df) the relationship between the primary particle diameter ( dp) and the radius of gyration of the agglomerate ( Rg) is related to the number of primary particles in the agglomerate Npp as (Wu and Friedlander 1993) fD p g ppd R k v v N ) 2 (0 (3) where 0 is the volume of the primar y particle and the prefactor k is called the structure factor (Mandelbrot 1982). If the fractal dimension of agglomerates is be tween 2 and 3, the planar projections of the agglomerate cross sections are given by 2) ( secgj giR R tion cross (4) Without a limit for Df, there would arise the anomaly wher ein the cross-section of the new agglomerate would be larger than the sum of th e cross-sections of the individual agglomerates. Therefore, a limit of Df 2 has been suggested. Moreover, it is assumed that Df is a constant, said assumption being supported by computer si mulations (Friedlander 2000). Incorporating Equations 2 and 3 with k set to 1, the collision kernel for agglomerates can be expressed by the

PAGE 118

118 gas kinetic theory for rigi d spherical particles for Df > 2 in the free molecular regime (Matsoukas and Friedlander 1991) as 2 / 1 / 1 2 11 1 ) (f fD j D i j i j iv v v v K v v (5) where f fD p D bd T k K/ 6 2 2 / 1 / 2 2 12 4 3 6 kb is the Boltzmann constant, T is the absolute temperature, and is the particle density. Due to the efficient structure and low comput ational demand, the moment method has been widely used to explain the coagulation phe nomena (Whitby and Atmospheric Research and Exposure Assessment Laboratory (U.S.) 1991; Williams and Loyalka 1991). The kth moment of the particle size distribution is written as 0, d t n Mk k (6) The moment model assuming a lognormal size di stribution was used in this study where the distribution functi on is written as ) ( ln 18 ) ( / ln exp ) ( ln 2 ) ( 3 1 ,2 2t t t t N t ng g g (7) where N(t) is the total number concentration of particles, g(t) is the geometric standard deviation and g(t) is the geometric number mean partic le volume. By incorporating Equation (1) with Equations (6) and (7), the various moment equations can then be de rived. The development of the moment method was provided in detail in Park and Lee (2001) for the derivation of the analytical solution, and thus is not repeated here. The thr ee differential equations for the moments used in this work, solved numerically using a RungeKutta 4th order method, are

PAGE 119

119 summarized in Table 1 and they are for co mparison purposes only. Once the moments are determined for each time step, the key size distri bution parameters (i.e., geometrics standard deviation and geometric mean diameter) can be then determined. In deriving the analytical solution, Park and Lee (2002) made several assumptions. These assumptions are no longer needed when solving these equations numerically. Simulation Conditions Primary particle size, mass fractal dimension a nd initial geometric standard deviation were varied to evaluate the effects of agglomerate size distribution on coagulation rate. The conditions are summarized in Table 2. The evolution of g was tracked to determine the time to reach the asymptotic state, i.e., self-preservi ng time-lag for fractal nanoagglomerates, f. Following the convention of prior studies, f is defined as the time it takes for g to approach within 1 % of the asymptotic value (Landgrebe and Pratsinis 1990; Vemury et al. 1994; Vemury and Pratsinis 1995). The simulation conditions of Set 1 were the same as in Vermury and Pratsinis (1994) and Park and Lee (2002), so that the re sults could be compared with thos e of previous studies. In Sets 2 and 3, dp and g0 were varied to investigate the dependence of f on these parameters. Results and Discussion Effects of Df and dp on coagulation rate Before the simulation was carried out, the impact of Df and dp on the dynamics of agglomerate growth was investigated. This wa s done by examining the re lationship between the collision kernels of agglomerates ( a) and spherical particles ( s) of the same particle volume. For nonspherical particles, a is expressed by Equation (5). For spherical particle s, by setting Df = 3,

PAGE 120

120 2 3 / 1 3 / 1 2 1 6 1 2 11 1 4 3 6 v u v u T kb s (8) Dividing Equation (5) by Equation (8), the ratio can be derived as 2 3 / 1 3 / 1 2 / 1 / 1 / 6 2 ) 3 2 2 (2 4 3 u v u v df f f fD D D p D s a (9) Under either a monodisperse assumption ( u = ,), or an extreme size ratio ( >> u ), the ratio of the collision kernels can be expressed as: ) 3 2 2 ( 3 2 2 / 6 2 ) 3 2 2 () 2 ( 2 4 3f f f fD p g D D p D s ad R d (10) where ) 6 ( ) 2 (3 p D p gd d R vf following Equation (3) where k is set to 1. As shown in Equation (10), agglomerates with a smaller primary partic le size and a smaller mass fractal dimension are expected to have a hi gher collision rate. Effect of Df on the asymptotic g Prior studies (Park and Lee 2002; Vemury a nd Pratsinis 1995; Wu a nd Friedlander 1993) have reported that g approaches an asymptotic value when coagulation is the dominant mechanism. Figure 1 shows the asymptotic g determined in this work as well as the analytical solution reported by Park and Lee (2002). The results obtained by th is study for fractal agglomerates, as shown, are almost identical to the analytical results (symbols overlapped in the figure) provided by Park and Lee (2002). The self -preserved geometric standard deviations for Df = 3.0, 2.5 and 2.0 were 1.355, 1.393 and 1.481 respectively. The similarity verifies that the effects of assumptions used in Park and Lee (2 002) are negligible. The results by the discretesectional methods (Vemury and Pr atsinis 1995; Wu and Friedlande r 1993) are also presented.

PAGE 121

121 Differences between the moment and the sec tional models are due to the presumed size distribution. Therefore their asym ptotic values, by na ture, are different. Differences between the volume-based and number-based model results are due to the finite sectional spacing that yields numerical diffusion (Landgrebe and Pr atsinis 1990; Wu and Biswas 1998). Effect of dp on f and g as a function of Df To investigate the effect of dp on f and g, simulation runs were performed for dp from 1 nm to 20 nm with all the other parameters ke pt constant. Figure 2a shows the results of asymptotic g as a function of Df for 3 primary particle sizes. As shown, the influence of dp is negligible (symbols overlapped) and all the asymptotic values are dependent on Df only. Different results are observed, however, for the effect on f, as shown in Figure 2b. The results are plotted in dimensionless time, f following the definition by Vemu ry and Pratsinis (1995) as t N Td kp p b f 0 6 1 2 14 3 3 (11) where t was obtained from th e simulation runs. As shown, when dp = 2 nm. f does not change much as Df varies. Besides, they agree well with those of the sectional method (Vemury and Pratsinis 1995) even th ough their asymptotic values ( g) are different. As dp and Df increase, however, the effect on f becomes more prominent. At Df = 3, 20nm is approximately 15 times larger than 2nm, even though the asymptotic g value is the same for both (Figure 2a). Agglomerates composed of smaller primary particles tend to have greater drag forces and coagulate faster than th eir larger counterparts. Hence, they can reach the asymptotic value faster.

PAGE 122

122 A statistical formula wa s then developed for f as a function of Df for 2 nm primary particles. The formula was developed using one-way non-linear regression of an exponential relationship, which has the form as follows ) 0 2 ( 8 11 1 42 exp 2734 0 12 2 f f nmD D (12) The choice of the exponent follows that implicated by Park et al. (2000). A good correlation was obtained with an R2 of 0.9963. The formula was then compared with the time lag reported by Vemury and Pratsini s (1995). As shown in Figure 2b, the results between the discrete-sectional methods (Vem ury and Pratsinis 1995) and moment method for 2 nm closely agreed. Combined effect of dp and Df on f The dependence of f on both dp and Df was further evalua ted statistically using two-way ANOVA and least square non-linear regression of an exponentia l relationship with a 95 % confidence interval. To facil itate the analysis, normalized self-preserving time lag ( n) was used in this study, which is defined in Equation (13) as the ratio of self preserving time lag of agglomerates composed of any primary particle si ze (> 2 nm) to that for 2 nm primary particle with the same Df. nm f n 2 (13) The dependence of n on dp and Df can then be expressed as ) 9938 0 ( 10 04 1 079394 12 ) 835173 2 268549 0 ( 5 R ef pD d n (14)

PAGE 123

123 The actual time lag for agglomerates composed of a given dp and Df can be derived by combining Equations (1 3) and (14) as ) 8 11 1 42 ( ) 835173 2 268549 0 ( 522734 0 / 10 04 1 079394 1f f pD D d fe e (15) Figure 3 shows the self-preserving time lag for various Df as a function of dp. As shown, with a smaller Df (e.g. 2), dp does not seem to impose any signif icant effect. Values of the self preserving time lag do not change much for the entire range of primary particle size. The effect of dp on n, however, becomes more prominent as Df increases. Integrated effect of dp, Df, and g0 on f Similar to the previous s ection, the dependence of f as a function of dp, g0 and Df was considered in the sequential st atistical evaluation us ing three-way ANOVA and least square nonlinear regression of an exponentia l relationship with a 95 % confid ence interval. Figure 4 shows f as a function of g0 for 2 nm agglomerates. Vemury et al. (1994) propose d two correlation equations to predict f, based upon g0: for g0 smaller than the asymptotic value, 3 0 2 0 0) 1 ( exp ) 1 ( g g g fC B A (16) and for g0 greater than the asymptotic value, 0 10logg fE D (17) It should be noted that th ese formulae were based on dp = 2 nm and Df = 3. Formulae for other dp and Df are of great importance as discussed earlier, The dependence of the time lag on these pa rameters for the moment model can be statistically derived. For g0 smaller than the asymptotic value, the format of Equation (16) was

PAGE 124

124 adopted, although the parameters A-C are no longer constants. For 0 greater than the asymptotic value, a new format is chosen as shown in Equation (18). ) 992 0 (R ) exp(2 0 F E Df (18) With a 95 % confidence inte rval, it was found that each of the five variables ( A, B, D, E and F ) in Euations (16) and (18) shows significant dependence on pr imary particle size and mass fractal dimension. Meanwhile, variable C is only a function of ma ss fractal dimension. The dependence of f on all these parameters was then ex amined by a 3-way ANOVA test with a 95 % confidence interval. The 3-way regression rela tionship associated with the curve fits by the R2 values has the factors (AF) as listed below, ) 251 0 337 3 exp( 10 2 1 043 35 p fd D A ) 238 0 352 3 exp( 10 8 1 132 324 p fd D B ) 10 exp( 378 2 083 61 fD C ) 033 0 exp( 549 4 ) 204 0 exp( 139 17 067 42p fd D D ) 014 0 604 0 exp( 952 18 288 2p fd D E 207 0 ) 424 3 10 41 5 exp( 043 25 p fd D F Figure 5 shows the effect of Df and g0 on f. As shown, the closer g0 is to the asymptotic value, the shorter the self preservi ng time lag. The dependence between g0 and shows an exponential relationship. Furthermore, it takes longer to reach the asym ptotic value as the agglomerates become more compact (larger Df). When the initial aero sols are monodisperse ( g0 = 1), f becomes A, which is a function of dp and Df. While the formulae developed by Vemury et al. (1994) are good for 2 nm primary particles, pa rticles produced in many industrial applications may very likely be of other sizes. With these three important parameters integrated into one equation, the new formula can be used as a conven ient tool to predict the time it takes to reach the asymptotic state.

PAGE 125

125 Implication to nanoparticle synthesis For many industrial applications, the primary particle size synthesized is usually much larger than 2 nm. For example, Degussa P25 titania is produced by flame synthesis and its primary particle is around 20 nm (Degussa, AG, TI 1234); anatase TiO2 nanoparticles synthesized by chemical vapor deposition of orga nic vapor precursors are between 12 and 23 nm (Backman et al. 2004; Miquel et al. 1993). Prior models (Park and Lee 2002; Vemury et al. 1994) are based on a primary particle size of 2 nm. The effect of dp is negligible for a small Df (e.g. 2). On the other hand, for a larger Df (= 3), as shown in Figure 4, f increases exponentially as dp increases. According to Equation (17), it take s 20 nm primary particles 7 times longer than 2 nm primary particles to reach a self preser ving distribution. The ne w formula can provide a more accurate estimate as it accounts for the effect s of all important parame ters: primary particle size, initial geometric standard deviation a nd the fractal dimension of the agglomerates. Conclusion The effects of primary particle size and initia l geometric standard de viation on the change in geometric standard deviation and self-preserv ing time-lag of fractal agglomerates in the free molecular regime were investigated. In evalua ting the effects of initial geometric standard deviation, results of the moment method were very close to thos e of the analytical solution. Meanwhile, sectional methods were not in as close an agreement in the asymptotic values due to the assumptions of the size distribu tion shape used in these models. The dependence between g0 and f showed an exponential relationship with primary particle size, initial geometric standard deviat ion, and mass fractal dimension. The asymptotic geometric standard deviation did not depend on th e primary particle size. When the mass fractal dimension was small ( Df = 2.0), the effect of primary partic le size on the self preserving time lag

PAGE 126

126 was negligible. Self preserving time lag decreases as the initial geometric standard deviation gets closer to the asymptotic value. Statistical form ula for the dependence of self preserving time lag on primary particle size, initial geometric sta ndard deviation, and ma ss fractal dimension of agglomerates were developed with high R2 values (> 0.99). These statistical formulae incorporating the three important parameters serve as a convenient tool to estimate the self preserving time lag of fractal agglomerates. Nomenclature Rg radius of gyration of agglomerates [nm] Df mass fractal dimension dp primary particle diameter [nm] dai, daj particle size of agglomerate i, j [nm] n aerosol number concentration [#/cm3] Mk kth moment of aerosol size distribution N total number concentration [#/cm3] Npp number of primary particle size [#] kb Boltzmann's constant [#/mole] T temperature [K] Vg geometric number mean volume Greek Letters Collision frequency function g G eometric standard deviation g0 Initial Geometric standard deviation Gas viscosity [Pa.s] f Self preserving time lag for fr actal nanoagglomerates [s] n Normalized self preserving time [-] Particle density [kg/m3]

PAGE 127

127 Table.1 Comparison of M0, M1, and M2 in two different regimes (Park and Lee, 2001) Moment Continuum regime Free molecular regime dt dM0 ] [/ 1 / 1 2 0f fD DM M M K f f f fD D D DM M M M M M Kb/ 2 2 / 1 / 1 ) 2 / 1 / 1 ( ) 2 / 1 / 2 ( 02 dt dM1 0 0 dt dM2 ] [ 2/ ) 1 ( / ) 1 ( 2 1f f f fD D D DM M M K 2 / 1 / 2 1 2 / 1 / 1 / 1 1 2 / 1 / 2 12 2M M M M M M Kbf f f fD D D D

PAGE 128

128 Table.2 Simulation conditions Mass fractal dimension Particle number concentration(#/cm3) Primary particle size (nm) 2 2.0/2.5/3.0 1016 1~20 1~20

PAGE 129

129 D f 2.02.53.0 g 1.3 1.4 1.5 1.6 1.7 1.8 Moment method (this work) Anaytical soultion (Park and Lee, 2002) Volume distribution (Vermury and Pratsinis ,1995) Volume distribution (Wu and Friedlander, 1993) Number distribution (Vermury and Pratsinis ,1995) Figure 1 Asymptotic g as a function of Df

PAGE 130

130 D f 2.02.53.0 g 1.34 1.36 1.38 1.40 1.42 1.44 1.46 1.48 1.50 Primary particle size :2 nm Primary particle size :10 nm Primary particle size :20 nm (a) D f 2.02.22.42.62.83.0 f 0 10 20 30 40 50 Primary particle size :2 nm Primary particle size :10 nm Primary particle size :20 nm Primary particle size :2 nm (Vermury and Pratsinis 1995) (b) Figure 2 Effect of dp as a function of Df on: (a) geometric standard de viation; (b) self preserving time lag.

PAGE 131

131 primary particle size, nm 05101520 f 0 10 20 30 40 50 Calculated Value ( D f =2.0) Calculated Value ( D f =2.5) Calculated Value ( D f =3.0) Apporiximation model ( D f =2.0) Apporiximation model ( D f =2.5) Apporiximation model ( D f =3.0) Figure 3 Effect of mass fractal dime nsion and primary particle size on f in the free molecular regime

PAGE 132

132 Figure 4 Comparison of f to attain the self preserving di stribution (dp = 2 nm and Df = 3)

PAGE 133

133 Figure 5 Effects of Df and g0 on the self preserving time lag: (a) dp = 2 nm; (b) dp = 20 nm

PAGE 134

134 REFERENCES Akhtar, M. K., Lipscomb, G. G., and Pratsinis, S. E. (1994). "M onte-Carlo simulation of particle coagulation and sintering." Aerosol Science and Technology 21(1), 83-93. Akhtar, M. K., Yun, X. O., and Pratsinis, S. E. (1991). "vapor synthesis of titania powder by titanium tetrachloride oxidation." AIChE Journal 37(10), 1561-1570. Atkins, P. W., and DePaula, J. (2002). Physical chemistry W.H. Freeman, New York. Au, M. (2005). "Hydrogen storage properties of magnesium based na nostructured composite materials." Materials Science and Engineering BSolid State Materials for Advanced Technology 117(1), 37-44. Avrami, M. (1939). "Kinetics of pha se change I General theory." Journal of Chemical Physics 7(12), 1103-1112. Avrami, M. (1940). "Kinetics of pha se change II General theory." Journal of Chemical Physics 8(12), 212-224. Avrami, M. (1941). "Granulation, phase change, and microstructure Kinetics of phase change. III." Journal of Chemical Physics 9(2), 177-184. Backman, U., Tapper, U., and Jokiniemi, J. K. (2004). "An aerosol method to synthesize supported metal catalyst nanoparticles." Synthetic Metals 142(1-3), 169-176. Barkhordarian, G., Klassen, T., and Bormann, R. (2006). "Kinetic investiga tion of the effect of milling time on the hydrogen sorption reactio n of magnesium catalyzed with different Nb2O5 contents." Journal of Alloys and Compounds 407(1-2), 249-255. Biswas, P., and Wu, C. Y. (2005). "2005 Critica l Review: Nanoparticles and the environment." Journal of the Air & Wast e Management Association 55(6), 708-746. Bloch, J., and Mintz, M. H. (1997). "Kinetics and mechanisms of metal hydrides formation A review." Journal of Alloys and Compounds 253, 529-541. Bowden, F. P., and Tabor, D. (1986). The friction and lubrication of solids Clarendon Press, Oxford. Bowman, R. C., and Fultz, B. (2002). "Metalli c hydrides I: Hydrogen storage and other gasphase applications." MRS Bulletin 27(9), 688-693. Callister, W. D. (1997). Materials science and engi neering: an introduction John Wiley & Sons, New York. Cashdollar, K. L. (2000). "Overview of dust explosibility characteristics." Journal of Loss Prevention in the Process Industries 13(3-5), 183-199.

PAGE 135

135 Chalk, S. G., and Miller, J. E. (2006). "Key challe nges and recent progress in batteries, fuel cells, and hydrogen storage for clean energy systems." Journal of Power Sources 159(1), 73-80. Charbonnier, J., de Rango, P., Fruchart, D., Miraglia, S., Pontonnier, L., Rivoirard, S., Skryabina, N., and Vulliet, P. (2004). "Hydrog enation of transition element additives (Ti, V) during ball milling of magnesium hydride." Journal of Alloys and Compounds 383(12), 205-208. Chen, Y., and Williams, J. S. (1995). "Formatio n of metal-hydrides by mechanical alloying." Journal of Alloys and Compounds 217(2), 181-184. Chou, K. C., Li, Q., Lin, Q., Jiang, L. J., and Xu, K. D. (2005). "Kinetics of absorption and desorption of hydrogen in alloy powder." International Journal of Hydrogen Energy 30(3), 301-309. Conner, W. C., and Falconer, J. L. (1995) "Spillover in heterogeneous catalysis." Chemical Reviews 95(3), 759-788. Cooper, D., Wu, C. Y., Yanskey, D., Butt, D., and Cai, M. (2005). "Hydrogen storage characteristics of nickel nanoparticle coated magnesium prepared by dry particle coating." KONA 139-151. Coowanitwong, N., Wu, C. Y., Cai, M., Ruthkosky, M., Rogers, J., Feng, L., Watano, S., and Yoshida, T. (2003). "Surface modification of Al2O3 fiber with binary nanoparticles using a dry-mechanical coating technique." Journal of Nanoparticle Research 5(3-4), 247-258. Cui, N., Luo, J. L., and Chuang, K. T. (2001) "Study of hydrogen diffusion in alphaand betaphase hydrides of Mg2Ni alloy by microelectrode technique." Journal of Electroanalytical Chemistry 503(1-2), 92-98. Degussa AG, TI 1234. Titanium dioxide as photocatalyst. Technical.serve.aerosil@degussa.com DOE. (2003). "Basic research needs for the hydrogen economy: Report of the Basic Energy Science Workshop on Hydrogen Production, Stor age and Use. D. o. Energy, Office of Science. DOE. (2007a). "Hydrogen posture plan: an integr ated research, development and demonstration Plan' http://www.hydrogen.energy.gov/news_posture.html." DOE. (2007b). "Hydrogen, fuel cells & infrastructure technologies program.2007;http://www1.eere.energy.gov/hydr ogenandfuelcells/mypp/pdfs/tech_valid.p df." Doppiu, S., Langlais, V., Sort, J., Surinach, S., Baro, M. D., Zhang, Y., Hadjipanayis, G., and Nogues, J. (2004). "Controlled reduction of NiO using reactive ball milling under hydrogen atmosphere leading to Ni-NiO nanocomposites." Chemistry of Materials 16(26), 5664-5669.

PAGE 136

136 Dornheim, M., Doppiu, S., Barkhordarian, G., Bo esenberg, U., Klassen, T., Gutfleisch, O., and Bormann, R. (2007). "Hydrogen storage in magnesium-based hydrides and hydride composites." Scripta Materialia 56(10), 841. Dornheim, M., Eigen, N., Barkhordarian, G., Kl assen, T., and Bormann, R. (2006). "Tailoring hydrogen storage materials towards application." Advanced Engineering Materials 8(5), 377-385. Douglass, D. L. (1975). "Formation and dissocia tion of magnesium alloy hydrides and their use for fuel storage in hydrogen car." Metallurgical Transactions a-Physical Metallurgy and Materials Science 6(12), 2179-2189. Dresselhaus, M. S., and Thomas, I. L. (2001). "Alternative energy technologies." Nature 414(6861), 332-337. Dunlap, R. A., Small, D. A., MacKay, G. R., O'Brien, J. W., Dahn, J. R., and Cheng, Z. H. (2000). "Materials preparation by ball milling." Canadian Journal of Physics 78(3), 211229. Flanagan, T. B. (1978). Kinetics of hydrogen absorption and desorption: Hydrides for energy storage Pergamon Press, Oxford. Fleisch, T., and Abermann, R. (1977). "Reduction of Ag2S films by hydrogen spillover under ultra high-vacuum conditions." Journal of Catalysis 50(2), 268-278. Frenklach, M., and Harris, S. J. (1987). "Aer osol dynamics modeling using the method of moments." Journal of Colloid and Interface Science 118(1), 252-261. Friedlander, S. K. (2000). Smoke, dust, and haze: fundamentals of aerosol dynamics Oxford University Press, New York. Friedlander, S. K., and Pui, D. Y. H. (2004). "Emerging issues in nanoparticle aerosol science and technology." Journal of Nanoparticle Research 6(2-3), 313-314. Friedlmeier, G., and Groll, M. (1997). "Experi mental analysis and modeling of the hydriding kinetics of Ni-doped and pure Mg." Journal of Alloys and Compounds 253, 550-555. Friedrichs, O., Kolodziejczyk, L., Sanchez-Lopez, J. C., Lopez-Cartes, C., and Fernandez, A. (2007). "Synthesis of nanocrystalline MgH2 powder by gas-phase condensation and in situ hydridation: TEM, XPS and XRD study." Journal of Alloys and Compounds 434-435, 721. Friedrichs, O., Sanchez-Lopez, J. C., Lopez-Cart es, C., Dornheim, M., Klassen, T., Bormann, R., and Fernandez, A. (2006). "Chemical and micr ostructural study of the oxygen passivation behaviour of nanocrystalline Mg and MgH2." Applied Surface Science 252(6), 2334. Fukai, Y. (2005). The metal-hydrogen system: basic bulk properties Springer, Berlin.

PAGE 137

137 Gabis, I. E., Voit, A. P., Evard, E. A., Zaika, Y. V., Chernov, I. A., and Yartys, V. A. (2005). "Kinetics of hydrogen desorption from the powders of metal hydrides." Journal of Alloys and Compounds 404, 312-316. Gao, L. H., Chen, C. P., Chen, L. X., Wang, X. H., Zhang, J. W., Mao, X. Z., and Wang, Q. D. (2005). "Hydriding/dehydriding behaviors of La1.8Ca0.2Mg14Ni3 alloy modified by mechanical ball-milling under argon." Journal of Alloys and Compounds 399(1-2), 178182. Gelbard, F., and Seinfeld, J. H. (1980). "S imulation of multicomponent aerosol dynamics." Journal of Colloid and Interface Science 78(2), 485-501. Gennari, F. C., Castro, F. J., and Gamboa, J. J. A. (2002). "Synthesis of Mg2FeH6 by reactive mechanical alloying: formation and decomposition properties." Journal of Alloys and Compounds 339(1-2), 261-267. Ginstling, A. M., and Brounshtein, B. I. (1950) "Diffuzionnoi kinetike re aktsii v sfericheskikh chastitsakh." Zhurnal Prikladnoi Khimii 23(12), 1249-1259. Higuchi, K., Kajioka, H., Toiyama, K., Fujii, H., Orimo, S., and Kikuchi, Y. (1999). "In situ study of hydriding-dehydriding properties in some Pd/Mg thin films with different degree of Mg crystallization." Journal of Alloys and Compounds 295, 484-489. Holtz, R. L., and Imam, M. A. (1999). "Hyd rogen storage characteristics of ball-milled magnesium-nickel and magnesium-iron alloys." Journal of Materials Science 34(11), 2655-2663. Hong, S. H., and Kim, B. K. (2001). "Fabrication of aluminum flake powder from foil scrap by a wet ball milling process." Materials Letters 51(2), 139-143. Hong, S. H., Lee, D. W., and Kim, B. K. ( 2000). "Manufacturing of aluminum flake powder from foil scrap by dry ball milling process." Journal of Materials Processing Technology 100(1-3), 105-109. Hsu, S. M. (2004). "Nano-lubr ication: concept and design." Tribology International 37(7), 537545. Huang, J. Y., Wu, Y. K., and Ye, H. Q. (1995). "Ball-milling of ductile metals." Materials Science and Engineering a-Structural Ma terials Properties Microstructure and Processing 199(2), 165-172. Huhn, P. A., Dornheim, M., Klassen, T., and Bormann, R. (2005). "Thermal stability of nanocrystalline magnesium for hydrogen storage." Journal of Alloys and Compounds 404406, 499. Huot, J., Akiba, E., and Takada, T. (1995). "Mechanical alloying of Mg-Ni compounds under hydrogen and inert atmosphere." Journal of Alloys and Compounds 231(1-2), 815.

PAGE 138

138 Huot, J., Liang, G., and Schulz, R. (2001). "Mechanically alloyed metal hydride systems." Applied Physics a-Materials Science & Processing 72(2), 187-195. Huot, J., Liang, G., Boily, S., Van Neste, A ., and Schulz, R. (1999). "Structural study and hydrogen sorption kinetics of ball-milled magnesium hydride." Journal of Alloys and Compounds 295, 495-500. Iwakura, C., Inoue, H., Nohara, S., Shin-ya, R., Kurosaka, S., and Miyanohara, K. (2002). "Effects of surface and bulk modifications on electrochemical and physicochemical characteristics of MgNi alloys." Journal of Alloys and Compounds 330, 636-639. Iwakura, C., Nohara, S., Zhang, S. G., and Inoue, H. (1999). "Hydriding and dehydriding characteristics of an amorphous Mg2Ni-Ni composite." Journal of Alloys and Compounds 285(1-2), 246-249. Jeon, K.-J., Fereshteh, E., and Wu, C.-Y. (2007) "A model for Hydrogen absorption kinetic and capacity in metal hydride." Journal of Applied Physics in preparation. Jeon, K.-J., Theodore, A., Wu, C.-Y., and Cai, M. (2006). "Hydrogen absorption/desorption kinetics of magnesium nano-nickel compos ites synthesized by dry particle coating technique." International Journal of Hydrogen Energy doi:10.1016/j.ijhydene.2006.09.019. Kilinc, Y., Ozturk, S., Ozturk, B., and Uslan, I. (2004). "Investigation of milling characteristics of alumina powders milled with a newly de signed vibratory horizontal attritor." Powder Technology 146(3), 200-205. Landgrebe, J. D., and Pratsinis, S. E. (1990). "A discrete-sectional model for particulate production by gas-phase chemical-reaction an d aerosol coagulation in the free-molecular regime." Journal of Colloid and Interface Science 139(1), 63-86. Lee, K. W., and Chen, H. (1984). "Coagul ation rate of polydi sperse particles." Aerosol Science and Technology 3(3), 327-334. Lee, K. W., Curtis, L. A., and Chen, H. (1990). "An analytic solution to free-molecule aerosol coagulation." Aerosol Science and Technology 12(2), 457-462. Lee, S. M., Yu, J. S., Lee, P. S., and Lee, J. Y. (2002). "Characteristics of surface-modified metal hydride electrode with flake Ni by the ball-milling process." Journal of Alloys and Compounds 330, 835-840. Levenspiel, O. (1999). Chemical reaction engineering Wiley, New York. Li, L. Q., Akiyama, T., and Yagi, J. (2000). "Hydrogen storage alloy of Mg2NiH4 hydride produced by hydriding combustion synt hesis from powder of mixture metal." Journal of Alloys and Compounds 308, 98-103.

PAGE 139

139 Liang, G., Boily, S., Huot, J., Van Neste, A., and Schulz, R. (1998). "M echanical alloying and hydrogen absorption properties of the Mg-Ni system." Journal of Alloys and Compounds 267(1-2), 302-306. Liang, G., Huot, J., Boily, S., Van Neste, A ., and Schulz, R. (1999). "Catalytic effect of transition metals on hydrogen sorption in nanocrystalline ball milled MgH2-Tm (Tm=Ti, V, Mn, Fe and Ni) systems." Journal of Alloys and Compounds 292(1-2), 247-252. Liu, X., Huang, Z., Jiang, L., and Wang, S. "Thermal stabilization and hydrogen storage properties of Mg-40 wt%Ti0.28Cr0.50V0.22 composite prepared by mechanical milling." International Journal of Hydrogen Energy doi:10.1016/j.ijhydene.2006.08.030. Lueking, A. D., and Yang, R. T. (2004). "Hydr ogen spillover to enha nce hydrogen storage study of the effect of carbon physicochemical properties." Applied Catalysis a-General 265(2), 259-268. Mandelbrot, B. B. (1982). The fractal geometry of nature W.H. Freeman, San Francisco. Matsoukas, T., and Friedlander, S. K. (1991). "Dynamics of aerosol agglomerate formation." Journal of Colloid and Interface Science 146(2), 495-506. Meakin, P., Ramanlal, P., Sander, L. M., and Ball, R. C. (1986). "Ballistic deposition on surfaces." Physical Review A 34(6), 5091-5103. Miquel, P. F., Hung, C. H., and Katz, J. L. (1993). "Formation of V2O5-based mixed oxides in flames." Journal of Materials Research 8(9), 2404-2413. Mountain, R. D., Mulholland, G. W., and Baum, H. (1986). "Simulation of aerosol agglomeration in the free molecular and continuum flow regimes." Journal of Colloid and Interface Science 114(1), 67-81. NASA. (2006). "NASA hydrogen research fo r spaceport and space based applications." CR2006-214091 Nishimura, C., Komaki, M., and Amano, M. (1999). "Hydrogen permeation through magnesium." Journal of Alloys and Compounds 295, 329-333. Oh, S. T., Lee, J. S., Sekino, T., and Niihar a, K. (2001). "Fabrication of Cu dispersed Al2O3 nanocomposites using Al2O3/CuO and Al2O3/Cu-nitrate mixtures." Scripta Materialia 44(8-9), 2117-2120. Oh, Y. W., Jeon, K. J., Jung, A. I., and Jung, Y. W. (2002). "A simulation study on the collection of submicron particles in a unipolar charged fiber." Aerosol Science and Technology 36(5), 573-582. Orimo, S., and Fujii, H. (1996). "Hydriding properties of the Mg2Ni-H system synthesized by reactive mechanical grinding." Journal of Alloys and Compounds 232(1-2), L16-L19.

PAGE 140

140 Ovrelid, E., Floistad, G. B., Rosenqvist, T., Bakke, P., and Engh, T. A. (1998). "The effect of Sr addition on the hydrogen solubility and hydrid e formation in pure Mg and the alloy AZ91." Scandinavian Journal of Metallurgy 27(3), 133-140. Park, S. H., and Lee, K. W. (2002). "Change in particle size distribution of fractal agglomerates during Brownian coagulation in the free-molecule regime." Journal of Colloid and Interface Science 246(1), 85-91. Pfeffer, R., Dave, R. N., Wei, D. G., and Ra mlakhan, M. (2001). "Synthesis of engineered particulates with tailored propert ies using dry particle coating." Powder Technology 117(1-2), 40-67. Pratsinis, S. E. (1998). "Flame aer osol synthesis of ceramic powders." Progress in Energy and Combustion Science 24(3), 197-219. Ragone, D. V. (1995). Thermodynamics of materials Wiley, New York. Raman, S. S. S., Davidson, D. J., Bobet, J. L ., and Srivastava, O. N. (2002). "Investigations on the synthesis, structural and microstruc tural characterizations of Mg-based K2PtCl6 type (Mg2FeH6) hydrogen storage material prepared by mechanical alloying." Journal of Alloys and Compounds 333(1-2), 282-290. Revesz, A., Fatay, D., and Spassov, T. "Micro structure and hydrogen sorption kinetics of Mg nanopowders with catalyst." Journal of Alloys and Compounds doi:10.1016/j.jallcom.2006.08.130. Rozanov, V. V., and Krylov, O. V. (1997). "Hydrogen spillover in heterogeneous catalysis." Uspekhi Khimii 66(2), 117-130. Rudman, P. S. (1979). "Hydrogen-diffusion-ra te-limited hydriding a nd dehydriding kinetics." Journal of Applied Physics 50(11), 7195-7199. Saetre, T. O. (2006). "A model of hydr ogen diffusion in spherical particles." International Journal of Hydrogen Energy 31(14), 2104-2111. Sastri, M. V. C., Viswanathan, B ., and Srinivasa Murthy, S. (1998). Metal hydrides: fundamentals and applications Narosa Publishing House [publisher]; Schaefer, D. W. (1988). "Fractal mode ls and the structure of materials." MRS Bulletin 13, 22-27. Schlapbach, L., and Zttel, A. (2001). "Hydrogenstorage materials for mobile applications." Nature, 414, 353-358. Schmidt-Ott, A. (1988). "In situ measurement of the fractal dimensionality of ultrafine aerosol particles." Applied Physics Letters 52, 954. Schober, T. (1981). "The Magnesium-hydrogen System Transmission electron-microscopy." Metallurgical Transactions A-Physical Metallurgy and Materials Science 12(6), 951-957.

PAGE 141

141 Schweppe, F., Martin, M., and Fromm, E. (199 7). "Model on hydride formation describing surface control, diffusion control and transition regions." Journal of Alloys and Compounds 261(1-2), 254-258. Seayad, A. M., and Antonelli, D. M. (2004). "Recent advances in hydrogen storage in metalcontaining inorganic nanostruc tures and related materials." Advanced Materials 16(9-10), 765-777. Shao, H., Wang, Y., Xu, H., and Li, X. (2004) "Hydrogen storage properties of magnesium ultrafine particles prepared by hydrogen plasma-metal reaction." Materials Science and Engineering B 110(2), 221. Skoog, D. A., Holler, F. J., and Nieman, T. A. (1998). Principles of instrumental analysis Saunders College Pub.; Harcourt Brace College Publishers, Philadelphia Orlando, Fla. Smith, G. B., Gentle, A., Swift, P. D., Earp, A ., and Mronga, N. (2003). "Coloured paints based on iron oxide and silicon oxide coated flak es of aluminium as the pigment, for energy efficient paint: optical and thermal experiments." Solar Energy Materials and Solar Cells 79(2), 179-197. Song, M. Y. (1995). "Effects of mechanical alloyi ng on the hydrogen storage characteristics of Mg-Xwt-percent Ni (X=0, 5, 10, 25 and 55) mixtures." International Journal of Hydrogen Energy 20(3), 221-227. Spassov, T., Rangelova, V., Solsona, P., Baro M. D., Zander, D., and Koster, U. (2005). "Hydriding/dehydriding prope rties of nanocrystalline Mg87Ni3Al3M7 (M = Ti, Mn, Ce, La) alloys prepared by ball milling." Journal of Alloys and Compounds 398(1-2), 139-144. Stander, C. M. (1977). "Kinetics of forma tion of magnesium hydride from magnesium and hydrogen." Zeitschrift Fur Physikalische Chemie-Frankfurt 104(4-6), 229-238. Stark, W. J., and Pratsinis, S. E. (2002). "Aerosol flame reactors for manufacture of nanoparticles." Powder Technology 126(2), 103-108. Stillesjo, F., Hjorvarsson, B., and Rodmacq, B. (1993). "Studies of interdiffusion and alloy formation in Fe/V multilayers us ing hydrogen as a local probe." Journal of Magnetism and Magnetic Materials 126, 102-104. Tessier, P., and Akiba, E. (1999). "Catalysed reactive milling." Journal of Alloys and Compounds 295, 400-402. Theodore, A., Jeon, K. J., and Wu, C. Y. (2006). "Flake particle synthesis from ductile metal particles using a novel high-speed vibratory mill." KONA 24, 83-92. Trunov, M. A., Schoenitz, M., Zhu, X. Y., and Dreizin, E. L. (2005). "Effect of polymorphic phase transformations in Al2O3 film on oxidation kinetics of aluminum powders." Combustion and Flame 140(4), 310-318.

PAGE 142

142 Ulrich, G. D., and Subramanian, N. S. (1977). "Particle growth in flames III. Coalescence as a rate-controlling process." Combustion Science and Technology 17(3-4), 119-126. Varin, R. A., Czujko, T., Chiu, C., and Wrons ki, Z. (2006). "Particle size effects on the desorption properties of nanostru ctured magnesium dihydride (MgH2) synthesized by controlled reactive mechanical milling (CRMM)." Journal of Alloys and Compounds 424(1-2), 356-364. Varin, R. A., Li, S., Chiu, C., Guo, L., Moro zova, O., Khomenko, T., and Wronski, Z. (2005). "Nanocrystalline and non-crystalline hydrid es synthesized by controlled reactive mechanical alloying/milling of Mg and Mg-X (X = Fe, Co, Mn, B) systems." Journal of Alloys and Compounds 404, 494-498. Vemury, S., and Pratsinis, S. E. (1995). "Selfpreserving size distributio ns of agglomerates." Journal of Aerosol Science 26(4), 701-701. Vemury, S., Kusters, K. A., and Pratsinis, S. E. (1994). "Time-lag for attainment of the selfpreserving particle-size di stribution by coagulation." Journal of Colloid and Interface Science 165(1), 53-59. von Zeppelin, F., Reule, H., and Hirscher, M. (2002). "Hydrogen de sorption kinetics of nanostructured MgH2 composite materials." Journal of Alloys and Compounds 330, 723726. Whitby, K. T. (1979) Lumped Mode Aerosol Growth Model Particle Technology Laboratory Publication #395, Mechanical Engineering Department, University of Minnesota, Minneapolis, MN. Whitby, E. R., McMurry, P. H., Binkowski, F., and Shankar, U. (1991) Modal aerosol dynamics modeling EPA report for contract No. 68-01-7365. Williams, M. M. R., and Loyalka, S. K. (1991). Aerosol science: theory and practice: with special applications to the nuclear industry Pergamon Press, Oxford; New York. Wu, C. Y., and Biswas, P. (1998). "Study of nume rical diffusion in a discrete-sectional model and its application to aerosol dynamics simulation." Aerosol Science and Technology 29(5), 359-378. Wu, M. K., and Friedlander, S. K. (1993). "Enh anced power-law agglomerate growth in the freemolecule regime." Journal of Aerosol Science 24(3), 273-282. Yang, G. X., and Biswas, P. (1997). "Study of th e sintering of nanosized titania agglomerates in flames using in situ light scattering measurements." Aerosol Science and Technology 27(4), 507-521. Yang, J., Ciureanu, M., and Roberge, R. (20 00). "Hydrogen storage properties of nanocomposites of Mg and Zr-Ni-Cr alloys." Materials Letters 43(5-6), 234-239.

PAGE 143

143 Yoshinaga, H., Arami, Y., Kajita, O., and Sakai, T. (2002). "Highly densed-MH electrode using flaky nickel powder and gas-atomiz ed hydrogen storage alloy powder." Journal of Alloys and Compounds 330, 846-850. Yoshinaga, H., Wada, M., Sakai, T., Miyamura, H ., Kuriyama, N., and Uehara, I. (1997). "Metal hydride electrodes with lame llar-type network structure." Journal of Alloys and Compounds 253, 665-667. Yu, Z. X., Liu, Z. Y., and Wang, E. D. (200 2). "Hydrogen storage pr operties of nanocomposite Mg-Ni-Cu-CrCl3 prepared by mechanical alloying." Materials Science and Engineering aStructural Materials Properties Microstructure and Processing 335(1-2), 43-48. Zaluska, A., Zaluski, L., and Strom-Olsen, J. O. (1999). "Nanocrystalline magnesium for hydrogen storage." Journal of Alloys and Compounds 288(1-2), 217-225. Zaluska, A., Zaluski, L., and Strom-Olsen, J. O. (2001). "Structure, catalysis and atomic reactions on the nano-scale: a systematic approach to metal hydrides for hydrogen storage." Applied Physics a-Materi als Science & Processing 72(2), 157-165. Zeng, K., Klassen, T., Oelerich, W., and Bo rmann, R. (1999). "Critical assessment and thermodynamic modeling of the Mg-H system." International Journal of Hydrogen Energy 24(10), 989-1004. Zttel, A. (2004). "Hydrogen storage methods ." Naturwissenschaften, 91(4), 157-172.

PAGE 144

144 BIOGRAPHICAL SKETCH Ki-Joon Jeon was born in Seoul, Korea, in 1973 and grew up in Seoul, Korea. He attended the In-ha University in the Envi ronmental Engineering Department. After his sophomore year in the University, he joined the Army and served at the Korean Air Force base. After three years of service in the Army, he continued his studies and received BS and MS degrees in environmental engineering at Inha Un iversity, Korea, in 2001. He then decided to pursue Ph.D. degree in the USA. In 2003, he wa s granted admission to the PhD program at the Department of Environmental Engineering Science at the University of Florida, with an Alumni fellowship from Dr. Chang-Yu Wu to Environm ental engineering and Sc iences. He received a scholarship from Korean Science & Engin eering Foundation (M06-2003-000-10264-0). His research interests include hydrogen storage and nanoparticle synthesis and his current project involves the kinetics of metal hydride formation.