USF Libraries
USF Digital Collections

Theoretical and experimental study of solid state complex borohydride hydrogen storage materials

MISSING IMAGE

Material Information

Title:
Theoretical and experimental study of solid state complex borohydride hydrogen storage materials
Physical Description:
Book
Language:
English
Creator:
Choudhury, Pabitra
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Density functional theory
Lattice dynamics
Thermodynamics
Stability
Mechano-chemical process
Nanocatalyst doping
Dissertations, Academic -- Chemical Engineering -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Materials that are light weight, low cost and have high hydrogen storage capacity are essential for on-board vehicular applications. Some reversible complex hydrides are alanates and amides but they have lower capacity than the DOE target (6.0 wt %) for 2010. High capacity, light weight, reversibility and fast kinetics at lower temperature are the primary desirable aspects for any type of hydrogen storage material. Borohydride complexes as hydrogen storage materials have recently attracted great interest. Understanding the above parameters for designing efficient complex borohydride materials requires modeling across different length and time scales. A direct method lattice dynamics approach using ab initio force constants is utilized to calculate the phonon dispersion curves. This allows us to establish stability of the crystal structure at finite temperatures. Density functional theory (DFT) is used to calculate electronic properties and the direct method lattice dynamics is used to calculate the finite temperature thermodynamic properties. These computational simulations are applied to understand the crystal structure, nature of bonding in the complex borohydrides and mechanistic studies on doping to improve the kinetics and reversibility, and to improve the hydrogen dynamics to lower the decomposition temperature. A combined theoretical and experimental approach can better lead us to designing a suitable complex material for hydrogen storage. To understand the structural, bulk properties and the role of dopants and their synergistic effects on the dehydrogenation and/or reversible rehydrogenation characteristics, these complex hydrides are also studied experimentally in this work.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2009.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Pabitra Choudhury.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 160 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002069274
oclc - 608088420
usfldc doi - E14-SFE0003164
usfldc handle - e14.3164
System ID:
SFS0027480:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200421Ka 4500
controlfield tag 001 002069274
005 20100420151306.0
007 cr mnu|||uuuuu
008 100420s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003164
035
(OCoLC)608088420
040
FHM
c FHM
049
FHMM
090
TK7885 (Online)
1 100
Choudhury, Pabitra.
0 245
Theoretical and experimental study of solid state complex borohydride hydrogen storage materials
h [electronic resource] /
by Pabitra Choudhury.
260
[Tampa, Fla] :
b University of South Florida,
2009.
500
Title from PDF of title page.
Document formatted into pages; contains 160 pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2009.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
3 520
ABSTRACT: Materials that are light weight, low cost and have high hydrogen storage capacity are essential for on-board vehicular applications. Some reversible complex hydrides are alanates and amides but they have lower capacity than the DOE target (6.0 wt %) for 2010. High capacity, light weight, reversibility and fast kinetics at lower temperature are the primary desirable aspects for any type of hydrogen storage material. Borohydride complexes as hydrogen storage materials have recently attracted great interest. Understanding the above parameters for designing efficient complex borohydride materials requires modeling across different length and time scales. A direct method lattice dynamics approach using ab initio force constants is utilized to calculate the phonon dispersion curves. This allows us to establish stability of the crystal structure at finite temperatures. Density functional theory (DFT) is used to calculate electronic properties and the direct method lattice dynamics is used to calculate the finite temperature thermodynamic properties. These computational simulations are applied to understand the crystal structure, nature of bonding in the complex borohydrides and mechanistic studies on doping to improve the kinetics and reversibility, and to improve the hydrogen dynamics to lower the decomposition temperature. A combined theoretical and experimental approach can better lead us to designing a suitable complex material for hydrogen storage. To understand the structural, bulk properties and the role of dopants and their synergistic effects on the dehydrogenation and/or reversible rehydrogenation characteristics, these complex hydrides are also studied experimentally in this work.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Co-advisor: Venkat R. Bhethanabotla, Ph.D.
Co-advisor: Elias Stefanakos, Ph.D.
653
Density functional theory
Lattice dynamics
Thermodynamics
Stability
Mechano-chemical process
Nanocatalyst doping
690
Dissertations, Academic
z USF
x Chemical Engineering
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.3164



PAGE 1

Theoretical and Experimental Study of Solid State Complex Borohydride Hydrogen Storage Materials by Pabitra Choudhury A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemical and Biomedical Engineering College of Engineering University of South Florida Co-Major Professor: Venkat R. Bhethanabotla, Ph.D. Co-Major Professor: Elia s Stefanakos, Ph.D. Yogi Goswami, Ph.D. Vinay Gupta, Ph.D. Lilia Woods, Ph.D. Sesha Srinivasan, Ph.D. Date of Approval: September 25, 2009 Keywords: Density functional theory, lat tice dynamics, thermodynamics, stability, mechano-chemical process, nanocatalyst doping Copyright 2009, Pabitra Choudhury

PAGE 2

Dedications To my parents, wife and my newly born son.

PAGE 3

Acknowledgements First, I would like to thank my wife parents and brothers, and for their unwavering love and support. Your encourag ement and loyalty mean the world to me. I am deeply indebted to my advisors, Professor Venkat R. Bhethanabotla and Professor Elias Stefanakos, fo r their excellent supervision throughout my Ph.D. study. I could not have reached this point without th eir tonic advices and patient guidance during each stage of my research work. I would also like to thank Dr. Sesha Srinivasan and Mr. Brian Smith for the help they provided during the last few years. I thank Prof. Yogi Goswami, Prof. Vinay Gupta, Prof. Lilia Woods, and Dr. Sesha Srinivasan for agreeing to be in my Ph.D. defense committee. I would also like to acknowledge the co mputational resources provided by USF Research Computing. As one of the users of their systems, an unlimited time was granted for usage and great support was provi ded for technical problems, without which, this dissertation would be possible. I am grateful for the support of my fellow group members, Subramanian Sankaranarayanan, Reetu Singh, Nianthrini Ba lakrishnan, Michael Ju rczyk and Jonathan Mbah, who have accompanied me on this journey. Thank you for the assistance you have given me and, most of all, fo r your friendship. I wish you well.

PAGE 4

i Table of Contents List of Tables ................................................................................................................ .... vii List of Figures ............................................................................................................... ... viii ABSTRACT ...................................................................................................................... xii Chapter 1 Introduction ...................................................................................................... 1 1.1. Current scenario of energy resources ............................................................... 1 1.2. Hydrogen as an energy carrier ......................................................................... 4 1.3. Hydrogen storage methods .............................................................................. 5 1.3.1. Hydrogen storage via physisorption ....................................................... 5 1.3.2. Hydrogen storage via chemisorption ...................................................... 6 1.3.2.1 Metal hydrides .............................................................................. 6 1.3.2.2 Complex hydrides ........................................................................ 7 1.4. Role of theoretical and experimen tal research on hydrogen storage materials ......................................................................................................... 8 1.5. Organization of the dissertation ..................................................................... 11 Chapter 2 Theoretical Approach..................................................................................... 13 2.1. Density functional theory ............................................................................... 13 2.1.1. Hohenberg-Kohn theorem .................................................................... 14 2.1.2. Self consistent Kohn-Sham equation .................................................... 17

PAGE 5

ii 2.1.3. Exchange and corelations f unctionals approximations ......................... 18 2.2. Solving self consistent Kohn-Sham equation ................................................ 20 2.2.1. Basis set for solution ............................................................................. 21 2.2.2. Pseudopotentials and ultra-soft pseudopotential ................................... 22 2.2.3. Projector-augmented-plane-wave method ............................................ 24 2.3. The Vienna ab-initio simulation package ...................................................... 27 2.4. Direct method lattice dynami cs and finite temperature thermodynamics ........................................................................................... 28 2.5. The PHONON package.................................................................................. 29 Chapter 3 Experimental Approach ................................................................................. 31 3.1. Experimental details ....................................................................................... 31 3.2. Thermo analytical tools/techniques ............................................................... 34 3.2.1. Differential scanning calorimetry ......................................................... 34 3.2.2. Thermo gravimetric analysis ................................................................. 35 3.2.3. Thermal programmed desorption autosorb-1 ........................................ 36 3.2.4. Pressure composition temperature apparatus ........................................ 38 3.3. Chemical analytical tools/techniques ............................................................. 39 3.3.1. Fourier transform infrared spectrometer ............................................... 39 3.3.2. X-ray diffractometer ............................................................................. 40 3.3.3. Scanning electron microscope .............................................................. 42 3.3.4. Energy dispersive X-ray spectroscopy.................................................. 43 3.3.5. Gas chromatography ............................................................................. 44 Chapter 4 First-Principles Investigation Of The Zn(BH4)2 ............................................ 45

PAGE 6

iii 4.1. Abstract .......................................................................................................... 45 4.2. Introduction .................................................................................................... 46 4.3. Computational methods ................................................................................. 49 4.3.1. Ab initio ................................................................................................ 49 4.3.2. Direct method lattice dynamics ............................................................ 50 4.4. Results and discussion ................................................................................... 51 4.4.1. Crystal structure .................................................................................... 52 4.4.2. Electronic structure ............................................................................... 58 4.4.3. Finite Temperature Reaction Enthalpy ................................................. 63 4.5. Summary ........................................................................................................ 66 Chapter 5 First-Principl es Study Of Ni-Induced Zn(BH4)2 ............................................ 68 5.1. Abstract .......................................................................................................... 68 5.2. Introduction .................................................................................................... 68 5.3. Methodology .................................................................................................. 69 5.4. Results and discussions .................................................................................. 71 5.5. Summary ........................................................................................................ 75 Chapter 6 First-Principles Investigation Of The Mn(BH4)2 ........................................... 77 6.1. Abstract .......................................................................................................... 77 6.2. Introduction .................................................................................................... 78 6.3. Simulation methods ....................................................................................... 80 6.3.1. Ab initio methods .................................................................................. 81 6.3.2. Direct method lattice dynamics ............................................................ 81 6.4. Results and discussions .................................................................................. 82

PAGE 7

iv 6.4.1. Crystal structure .................................................................................... 83 6.4.2. Electronic structure ............................................................................... 88 6.4.3. Thermodynamics................................................................................... 90 6.4.4. The enthalpy and the Gibbs energy of the reactions ............................. 94 6.5. Summary ........................................................................................................ 98 Chapter 7 First-Principles Investig ation Of Li-Mg-B-N-H System ............................. 101 7.1. Abstract ........................................................................................................ 101 7.2. Introduction .................................................................................................. 102 7.3. Computational details .................................................................................. 103 7.4. Results and discussion ................................................................................. 104 7.4.1. Structural stability ............................................................................... 105 7.4.2. Thermodynamics and reactions .......................................................... 109 7.4.3. Quantitative analysis of hydrogen release .......................................... 113 7.4.4. Reversible reaction step and van’t Hoff plot ...................................... 115 7.5. Summary ...................................................................................................... 117 Chapter 8 Experimental Study Of Li-Mn-B-H System ................................................ 119 8.1. Abstract ........................................................................................................ 119 8.2. Introduction .................................................................................................. 120 8.3. Experimental details ..................................................................................... 121 8.3.1. Materials and method .......................................................................... 121 8.3.2. X-ray diffraction ................................................................................. 122 8.3.3. Fourier transform infrared spectroscopy ............................................. 123 8.3.4. Simultaneous DSC and TGA .............................................................. 123

PAGE 8

v 8.3.5. Dehydrogenation kinetics: isotherm al volumetric measurements ...... 123 8.3.6. Temperature programmed desorption measurements ......................... 124 8.3.7. Gas chromatography analysis ............................................................. 124 8.4. Theory .......................................................................................................... 125 8.4.1. Activation energy calculations ............................................................ 125 8.5. Results and discussions ................................................................................ 126 8.5.1. Formation of complex hydride LiMn(BH4)3 –FTIR and XRD explorations ....................................................................................... 126 8.5.2. TGA, DSC and TPD studies of undoped and nanomaterials doped LiMn(BH4)3 ............................................................................ 129 8.5.3. Dehydrogenation kinetics of undoped and nano-Ni doped LiMn(BH4)3....................................................................................... 133 8.5.4. Activation energy calculations of undoped and nano-Ni doped LiMn(BH4)3....................................................................................... 134 8.5.5. GC analysis of undoped and nano-Ni doped LiMn(BH4)3 ................. 136 8.5.6. Possible mechanism of nano Ni doping on the complex hydride LiMn(BH4)3....................................................................................... 137 8.6. Summary ...................................................................................................... 138 Chapter 9 Summary, conclusi ons and recommendations ............................................. 140 9.1. Overview ...................................................................................................... 140 9.2. Zn(BH4)2 – conclusions and recommendations ........................................... 141 9.3. Mn(BH4)2 – conclusions and recommendations .......................................... 142 9.4. Li-Mg-B-N-H – conclusions and recommendations .................................... 143

PAGE 9

vi 9.5. Li-Mn-B-H – conclusions and recommendations ........................................ 144 9.6. Major contributions ...................................................................................... 145 9.7. Future work directions ................................................................................. 146 References .................................................................................................................... ... 148 About the Author....................................................................................................End Page

PAGE 10

vii List of Tables Table 1-1. World energy cons umption by fuel (2007) ....................................................... 2 Table 1-2. World primary energy production by fuel (2007) ............................................. 3 Table 4-1. Ground state energy, E in eV/mol of Zn, B and H2 from DFT calculations ..................................................................................................... 50 Table 4-2. The enthalpy of formation (in kJ/mol of H2) of Zn(BH4)2 from DFT for structures based on a similar chemical formula unit complex ........................ 53 Table 4-3. Optimized crystal structure of Zn(BH4)2. ........................................................ 56 Table 4-4. Zero point and vi brational energies of Zn(BH4)2 and its primary elements .......................................................................................................... 65 Table 6-1. The model structures (XY2Z8 type) used as input for the structural optimizations ................................................................................................... 84 Table 6-2. Optimized crys tal structure of Mn(BH4)2 ....................................................... 87 Table 7-1. Calculated ground state ener gy and zero point energy of different reactants/products: ........................................................................................ 107 Table 8-1. DSC and TGA analysis of undoped and nanoNi doped Li-Mn-B-H ............ 131

PAGE 11

viii List of Figures Figure 1-1. Schematic representation of combined theoretical and experimental approach .......................................................................................................... 10 Figure 2-1. Schematic representation of th e self-consistent loop for solution of Kohn-Sham equation18. ................................................................................... 20 Figure 2-2. Illustration of the real and pseudo wave-function and the real and pseudopotential22............................................................................................. 23 Figure 3-1. Differential scanning calo rimetry (DSC) Q10 apparatus ............................... 34 Figure 3-2. Schematic diagram of Q600 balance/furnace36 .............................................. 36 Figure 3-3. Schematic diagrams of the autosorb-1 TPD/TPR apparatus37 ....................... 37 Figure 3-4. Schematic diagrams of Si evert’s type volumetric apparatus38 ....................... 39 Figure 3-5. Perkin-Elmer spectrum one FTIR spectrometer ............................................ 40 Figure 3-6. The Philips X’Pert XRD system .................................................................... 41 Figure 3-7. Electron beam and specimen interaction signals40 ......................................... 42 Figure 3-8. Schematic working principle diagram for a SEM41 ....................................... 43 Figure 4-1. ( a-h) Phonon dispersion relations of the Pmc21, I41cd Pbca P-3, Fddd, P-1, P2/c and P-62m space group structures of Zn(BH4)2 ................... 56 Figure 4-2. Orthorhombic structure of space group Pmc21 (#26) of Zn(BH4)2 ( a ) Proposed three-dimensional crystal structure and ( b ) Projected structure along [010] plane ............................................................................. 57 Figure 4-3. Total density of phonon states g(w) of Zn(BH4)2 in Pmc21 symmetry .......... 58 Figure 4-4. DFT/LDA electronic band struct ure and total density of states (DOS) of Zn(BH4)2 relative to Fermi level ................................................................ 59

PAGE 12

ix Figure 4-5. DFT/LDA electr onic local/total density of state (DOS) relative to Fermi level for the most stable space group Pmc21 (#26) is orthorhombic structure (black line for s blue line for p and red line for d orbital) .......................................................................................................... 60 Figure 4-6. Electron localizati on function (ELF) of Zn(BH4)2 (100 plane) (Red: Zn, Blue: H and Green: B) .............................................................................. 62 Figure 4-7. Charge density of Zn(BH4)2 (100 plane) ........................................................ 62 Figure 5-1. Cohesive energy of pure and Ni-substituted Zn(BH4)2 .................................. 71 Figure 5-2. Relaxed struct ure of Ni doped Zn(BH4)2 ....................................................... 73 Figure 5-3. The total and partial phonon DOS for Zn8B16H64 (panels a d ) and Zn6Ni2B16H64 (panels e h )............................................................................... 74 Figure 6-1. Total density of phonon states g(w) of Mn(BH4)2 in I-4m2 symmetry .......... 86 Figure 6-2. Three-dimensional cr ystal structures of Mn(BH4)2 of space group I4m2 (#119) (Black (large), blue (m iddle) and green (small) spheres represent Mn, B and H atoms, respectively) ................................................... 87 Figure 6-3. DFT/GGA electronic spin polariz ed local/total density of state (DOS) relative to Fermi level for Mn(BH4)2 .............................................................. 89 Figure 6-4. ( a-b ) Electron localization function (ELF) and Charge density of Mn(BH4)2 for I-4m2 symmetry ....................................................................... 90 Figure 6-5. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for -B .................................................................. 92 Figure 6-6. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for -Mn (anti-ferromagnetic) .............................. 92 Figure 6-7. ( a-d ) Total density of phonon states g(w) of B, Mn, MnB2 and MnH2 crystal structures, respectively ........................................................................ 93 Figure 6-8. Temperature dependent r eaction enthalpy and Gibbs energy of reactions ((6-1 a )-(6-1 e )) ( a ), temperature dependent reaction enthalpy and Gibbs energy and the entropy contribution for the reaction Mn(BH4)2 = Mn + 2B+ 4H2 ( b ). ..................................................................... 98 Figure 7-1. Normalized plot of expe rimental FTIR spectra of the LiBH4-LiNH2MgH2 systems and phonon density of st ates of intermediate states ( i e (Li4BH4(NH2)3, Mg(NH2)2, LiBH4, MgH2, Li2Mg(NH)2 and LiH) during the first step hydrogen release reaction. ............................................ 106

PAGE 13

x Figure 7-2. Normalized plot of phonon density of states of Li3BN2, MgN3 and LiMgBN2 (intermediate phases of Li-M g-B-N-H system during second step hydrogen release). .................................................................................. 108 Figure 7-3. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for MgH2. ............................................................ 109 Figure 7-4. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for LiH. ............................................................... 110 Figure 7-5. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for Mg3N2. ........................................................... 110 Figure 7-6. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for LiBH4. ........................................................... 111 Figure 7-7. Temperature dependent Gibbs energy of reactions (7-1 a 7-1 b (7-2) – (7-4)) ............................................................................................................. 112 Figure 7-8. Temperature dependent reaction enthalpy of reactions ((7-2) – (7-4)) ........ 113 Figure 7-9. Normalized phase compositions of different reactants/products and also intermediate phases in different temperature ranges ............................. 115 Figure 7-10. The van’t Hoff plot derived fr om the theoretical calculations for the reaction: Mg(NH2)2 + 2LiH Li2Mg(NH)2 +2H2. ..................................... 116 Figure 8-1. Activation energy curve (a) undoped and (b) catalytic doping reactions ........................................................................................................ 126 Figure 8-2. FTIR profiles of LiBH4, MnCl2, and LiMn(BH4)3 + 2LiCl ball milled mixture representing B-H bonding bands and BH2 bending vibrations ....... 127 Figure 8-3. X-ray diffraction patterns of a pure LiBH4 and LiMn(BH4)3+2LiCl mixture obtained after milling under H2 ambient for 20 minutes ................. 128 Figure 8-4. Simultaneous DSC and TGA profiles of LiMn(BH4)3 doped with Xmol% nanoNi, (X=0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0) ball milled for 20 minutes .......................................................................................................... 130 Figure 8-5. Simultaneous DSC and TGA profiles of LiMn(BH4)3 doped with various nanocatalysts (nanoNi, nanoCo, nanoFe, nanoCu, nanoTi and nanoZn) by fixing the concentration X=1.5mol% and ball milled for 20 minutes ..................................................................................................... 131 Figure 8-6. Thermal Programmed Deso rption (TPD) profiles of undoped and Xmol% nanoNi and nanoCo doped LiMn(BH4)3; (X=1.5 mol%) ................ 132

PAGE 14

xi Figure 8-7. Dehydrogenaiton kinetics of undoped and Xmol% nanoNi (X=0.5, 1.0, 1.5, 2.0, and 2.5) doped LiMn(BH4)3 ..................................................... 133 Figure 8-8. TPD spectra of undoped and 1.5mol% nanoNi doped LiMn(BH4)3 at various ramping rates (4, 10 and 20 oC/min) ................................................ 135 Figure 8-9. Kissinger’s plot obtained from the TPD data for the undoped and 1.5 mol% nanoNi doped LiMn(BH4)3 ................................................................ 136 Figure 8-10. Gas Chromatography analysis of undoped and 1.5mol% nanoNi doped LiMn(BH4)3 ........................................................................................ 137

PAGE 15

xii Theoretical and Experimental Study of Solid State Complex Borohydride Hydrogen Storage Materials Pabitra Choudhury ABSTRACT Materials that are light we ight, low cost and have high hydrogen storage capacity are essential for on-board vehicular applicatio ns. Some reversible complex hydrides are alanates and amides but they have lower capac ity than the DOE targ et (6.0 wt %) for 2010. High capacity, light weight, reversibility and f ast kinetics at lower temperature are the primary desirable aspects for any type of hydrogen storage material. Borohydride complexes as hydrogen storage materials have recently attracted great interest. Understanding the above parameters fo r designing efficient complex borohydride materials requires modeling across different length and time scales. A direct method lattice dynamics approach using ab initio force constants is utilized to calculate the phonon dispersion curves. This a llows us to establish stability of the crystal structure at finite temperatures. Density functional theo ry (DFT) is used to calculate electronic properties and the direct method lattice dyn amics is used to calculate the finite temperature thermodynamic properties. These computational simulations are applied to understand the crystal structure, nature of bonding in the complex borohydrides and

PAGE 16

xiii mechanistic studies on doping to improve the ki netics and reversibility, and to improve the hydrogen dynamics to lower the decomposition temperature. A combined theoretical and experimental a pproach can better le ad us to designing a suitable complex material for hydrogen storag e. To understand the structural, bulk properties and the role of dopants and thei r synergistic effect s on the dehydrogenation and/or reversible rehydrogenatio n characteristics, these comp lex hydrides are also studied experimentally in this work.

PAGE 17

1 Chapter 1 Introduction Clean Energy is one of the greatest env ironmental and geopolitical challenges of recent times. Inexpensive and plentiful energy is in great demand for our current standard of living – which can at the moment only be supported by fossil fuels, which pollute the air by emitting green house gas CO2. 32.8% of the total energy is used by the transportation sector in the United Stat es according to the Energy Information Administration’s (EIA) Emissions of Gree nhouse Gases in the United States 2004 report. For transportation, petroleum products are primarily used, which cause major CO2 emission to the environment. To prevent global warming, we have to adopt new strategies to harness inexha ustible sources of energy1. 1.1. Current scenario of energy resources In general, oil, gas and coal along w ith nuclear energy a nd thermomechanical energy are considered three primary categorie s of energy sources. The first type of energy is due to breaking chemical bonds of any chemical compound such as hydrocarbon. The first type of primary ener gy sources is generally called fossil fuel. This type of energy can also come from absorbing sunlight and generating heat or electricity. The second type of energy is due to nuclear reaction from fission of heavy nuclei or fusion of light nuclei. The amount of energy generated in this reaction in the order of MeV per reaction. Th e third type of energy is thermomechanical energy which

PAGE 18

2 includes water, wind, and geothe rmal steam or hot water. The energy involved in this type is very low (meV) compare to other two types of energy sources. Global energy consumption has increased tremendously over the past few decades. The world primary energy consumption has shown in Table 1-1. The energy data are expressed in million tonnes of oil e quivalent (Mtoe) which is the amount of energy ( 42 GJ) released after burning of one tonne of crude oil. The world’s energy need is expected to (at least ) double within the next half cen tury. Essentially, the current world depends solely on first type of primar y resources, which are expected to deplete in the very near future. Overexploitation of these fossil fuels is also main cause of global warming, environmental pollution and acid rain. Table 1-1. World energy consumption by fuel (2007) Region Million Tonnes of oil equivalent (Mtoe) Oil Natural Gas Coal Nuclear Energy Hydroelectricity Total Total North America 1134.7 728.9 613.3 215.6 146.2 2838.7 Total S & C America 252.0 121.1 22.4 4.4 153.1 553.0 Total Europe & Eurasia 949.4 1040.1 533.7 275.6 188.6 2987.4 Total Middle East 293.5 269.4 6.1 5.1 574.1 Total Africa 138.2 75.2 105.9 3.0 22.2 344.5 Total Asia Pacific 1185.1 403.1 1896.2 123.4 194.0 3801.8 Total World 3952.9 2637.8 3177.6 622.0 709.2 11099.5 Source: BP Statistical review of World Energy – June 2008

PAGE 19

3 The production of energy is expected to remain adequate in coming few decades. However, imbalance of energy consumption is widely accepted around the world. Most of the developed and developing countries ne ed to consume more energy to ensure economic growth resulting very high energy co nsumption (Table 1-2). As the energy can be produced from fossil fuels via straightforward combustion process, they are relatively inexpensive and easy to transport. Howeve r, fossil fuels are essentially non-renewable energy sources in the long-term. The geologica l processes which create fossil fuels take millions of years, so they cannot be rege nerated within the timescales of human race once they have been exhausted. Table 1-2. World primary ener gy production by fuel (2007) Region Million Tonnes of oil equivalent (Mtoe) Oil Natural Gas Coal Total Total North America 643.4 706.3 629.9 1979.6 Total S & C America 332.7 135.7 55.3 523.7 Total Europe & Eurasia 860.8 968.2 445.4 2274.4 Total Middle East 1201.9 320.2 0.5 1522.6 Total Africa 488.5 171.3 154.2 814.0 Total Asia Pacific 378.7 352.3 1850.2 2581.2 Total World 3906 2654.0 3135.5 9737.6 Source: BP Statistical review of World Energy – June 2008 It can be observed from Table 1-1 that the ever-increasing (by 2.7% in 2007) consumption for energy coupled with decreasing fossil fuel resources due to

PAGE 20

4 overexploitations make the establishment of a clean and sustainable energy system a compelling need. Hydrogen-based energy systems seem to be potential solutions. Although, in the long-term, the ultimate technological challenge is large-scale production of hydrogen from renewable sources, the ch allenging issue is how to store hydrogen efficiently for onboard vehicular application powered by fuel-cell. 1.2. Hydrogen as an energy carrier Hydrogen is considered the greatest en ergy carrier amongst all sustainable fuels because of its highest heat of combustion (33.3 kwh/kg) amongst all fuels. Hydrogen is also the cleanest fuel because after combustion it produces only water which is completely harmless, and it is also easily availa ble; of course it is not freely available in nature; it can be produced easily by vari ous methods such as steam reforming, electrolysis of water, etc., a nd used as an energy carrier. Thus, hydrogen is the ideal fuel for the future since it reduces significantly the greenhouse gas emissions, reduces the global dependence on fossil fuel s, and increases the efficien cy of the energy conversion process for both internal combustion engines and proton exchange membrane (PEM) fuel cells. Hydrogen used in the fuel cell conver ts directly the chemical energy of hydrogen into water, electricity and heat2 as represented by the : H2 + O2 H2O + Electricity + Heat 1-1 From the phase diagram3 it can be observed that hydrogen is gaseous at temperatures above 32 K and H2 gas reacts, in concentrations down to 4.73%, explosively with air. Hydrogen is an extremely clean energy resource for many purposes such as turbine engines and vehicular applications ; however, it requires very high pressure for gas phase storage which is unsafe in these applications.

PAGE 21

5 1.3. Hydrogen storage methods The typical hydrogen storage options are: (i) compressed hydrogen, (ii) liquid hydrogen and (iii) solid hydrogen materials. Compressed hydrogen requires very high potential energy to store and requires bulky tanks. Liquid hydrogen has reasonably higher density than gaseous phase but it requires very low temperature (-253 oC) and also requires very high liquefaction energy. Hence, solid hydrogen storage materials are the appropriate choice for practical hydrogen storage for on-board vehicular applications. Hydrogen can be stored in the solid media via physisorption and chemisorption. The key concept of hydrogen adsorption/desorption is that hydrogen strongly affects the electronic and structural properties of the materials. In terms of electronic structure, the proton acts as an attractive potential to the host-metal electrons; electronic bands are lowered in energy and form low-lying bonding ba nds by hybridization with the hydrogen s orbital. This results in change of electronic structure and energy bandgap during hydrogen adsorption/desorption process. 1.3.1. Hydrogen storage via physisorption Nanoporous materials, which have diverse tunable phy sical properties as a function of their size and shape due to st rong quantum confinement effect and large surface to volume ratio, such as activated carbons, zeolit es, carbon nano tubes, conducting polymers and metal organic framewor ks (MOFs) can be considered to be strong candidate for hydrogen storage4-6. The main advantage of hydrogen storage via physisorption is that these materials can stor e and release hydrogen with fast kinetics and high reversibility over multiples cycles. The binding of hydrogen into the porous materials is due to the weak Van der Walls in teraction force (~10 kJ/mol) results in fast

PAGE 22

6 adsorption/desorption kinetics. But the hydrogen storage capacity for all above mentioned materials, except for some conducting polymers6, fall short of the Department of Energy (DOE) target. 1.3.2. Hydrogen storage via chemisorption Over the decades, the exploration a nd the scope of the hydrogen storage researches have been extended to hydrogen storage via chemisorption methods because of high storage capacity (gravimetric and volumetric) and tunable kinetic and thermodynamic parameters by addition of bot h catalysts and mixing of hydride phases. In chemisorption, unlike physisorption, both pr essure and temperature play an important role to change the chemical trans formation of the parent compounds via formation/breaking of chemical bonds. Th ese chemical bonds between the hydrogen atom and the metal can lead to the stronger bonded chemisorbed state, and finally the hydrogen atoms can diffuse into the host crysta l lattice. The chem isorption energy is typically in order of ~ 60 kJ/mol. Some ex amples of such materials are simple metal hydrides, hydrides in nanophase stru ctures and complex hydrides. 1.3.2.1 Metal hydrides The simplest chemical compositions of metal hydrides are called the binary hydrides, on the form MHx. Depending on the electronegativity of the metal M it forms different kinds of metal hydrides. Group I a nd II metals can be bonded to hydrogen ionic bonds. Covalent hydrides can formed with bot h non metallic (Group III – V) and light weight elements (Group I). Transition meta l hydrides are generall y form interstitial hydrides where hydrogen atoms oc cupy tetrahedral or octahedral interstitial sites, or a

PAGE 23

7 combination of both. As inte rstitial hydrides consist of relatively heavy elements therefore they have low hydrogen capacity by weight. Metal hydrid es with alloys are called intermetallic hydrides on the form AxByHz, where A typically re present a rare earth or an alkaline earth metal that tends to form stable hydrides, while B often is a transition metal which forms unstable hydrides3. The common host alloys of intermetallic hydrides are AB, AB3 and AB5 7. If this happens, the hydrogen is distributed compactly throughout the intermetallic lattice. 1.3.2.2 Complex hydrides Complex hydrides are differe nt from simple metal/intermetallic hydrides. In metal or intermetallic hydrides hydrogen atoms are encapsulated in metallic interstitial sites whereas in complex hydrides several hydrogen atoms surround other atom (such as Al, B, or N) to form complex characteristic s. Complex hydrides with light-elements have higher hydrogen coordination number than simp le metal/intermetallic hydrides and also much higher hydrogen storage capacity than transition meta l hydrides. Th eir thermal stability of the complex hydrides can be flexib le due to the co-existence of both covalent and ionic types of bonding. Pract ically it is possible to synthesize complex hydrides by mixing of existing hydrides, which allows the synthesis of materials with desired properties by tailoring the synthesis process. Among the most viable candidates, complex hydrides are considered to be th e best candidates for the onboard hydrogen storage application due to both high hyd rogen storage capacity and desirable absorption/desorption thermodynamics/kinetics.

PAGE 24

8 1.4. Role of theoretical and experimenta l research on hydrogen storage materials Critical properties of the hydrogen stor age materials to be evaluated for automotive applications are: (a) light weight, (b) cost and availability, (c) high volumetric as well as gravimetric density of hydroge n, (d) fast kinetics, (e) low dehydrogenation temperature, (f) favorable thermodynamic pr operties, (g) long cyclibility and (h) high degree of reversibility. A ll the above mentioned proper ties pose many scientific and technical challenges for the de velopment of new materials for hydrogen storage. Since a combined theoretical and experimental approach can lead to design a suitable complex material for hydrogen storage, both experimental and theoretical research is required for understanding of properties and hydroge nation-dehydrogenation mechanisms of stoichiometric complex hydrides, their synth esis and processing, and role of catalytic dopants. This requires intimat e collaboration between expe riment and theory, and also integrated efforts from physics, chem istry, material science, and engineering. The absorption/desorption energy and he nce the thermodynamic properties of the material can be tailored by manipulating the physical and electronic structures of the material. Quantum mechanical calculations ar e used to explain the nature of chemical bonding and the preferred occupancy sit es for hydrogen atoms in hydrogen storage materials, and the activation en ergy barriers, diffusion paths, and catalysis mechanisms during hydrogen uptake and release. Modeling and simulation can help in analyzing the experimental data and identifying the ke y factors for hydrogen storage capacity improvements, and also provide guidance fo r further experiments. Thus computer simulation is considered to be a powerful tool in characterizing the structure and

PAGE 25

9 hydrogen storage/diffusion properties and al so in understanding the destabilization mechanisms and various structure-properties relationships. Various methods are used to synthesize different kinds of hydrides, including solid-gas reaction, solution methods, mech ano-chemical, ion implantation, and electrochemical methods. Solution methods are usually used to synthesize complex metal hydrides. They often introduce impuriti es that are hard to remove from the materials, and result in formation of irrev ersible hydrides in many cases. Thus mechanochemical synthesis technique in a hydrogen environment is currently the most popular method to synthesize hydrogen storage materi als. Mechano-chemical milling yields samples in fine powders results in no practical growth of single crystal. These difficulties make X-ray absorption near-edge struct ure spectroscopy (XANES) and photoelectron spectroscopy (PES) hard to apply. Hydrogen at om is the lightest element in the periodic table as a result it is very difficult to identify the hydr ogen atoms using X-ray diffraction. Hydrogen shows the most significant isotope eff ect of all elements. Therefore deuterides (having better coherent scattering) can be u sed to determine the location of the hydrogen atoms easier. Sometime, Neutron powder diffr action in combination with high resolution X-ray diffraction (synchrotron) is used in th e structure determination. Inelastic neutron scattering (INS) is used to st udy interatomic interactions an d locate hydrogen positions in low concentration hydrides. Nuclear magnetic resonance (NMR), in frared (IR), Raman, are used to give information on local stru ctures and coordinated hydrogen dynamics in materials that can be compared with the theoretically predicted phonon spectra. Understanding the physico-chemical reactions an d synergistic effects of the nanocatalysts on the complex hydrides can also be predicted via different experimental techniques.

PAGE 26

10 Experimentally predictive hydrogen storag e capacity, thermodyna mic stability and kinetic enhancement due to nanocatalyst dopi ng on these complex hydrides can also be supported the theoretical calculations. Th e schematic representation of combined theoretical and experimental approach is shown below (Figure 1-1). Figure 1-1. Schematic representation of combin ed theoretical and experimental approach Sometime, materials having all the favorable properties like energetics and thermodynamics are not suitable for hydrogen storage because of the poor kinetics and irreversibility. These cannot be predicted by th eoretical simulation but with experiment, can be determined easily. Thus, both experimen tal and theoretical rese arch is needed to design suitable hydrogen storage materials.

PAGE 27

11 1.5. Organization of the dissertation This dissertation is organized as follows: € Chapter 2 provides details of the th eoretical approach implemented for fundamental study by density functional theory and finite temperature thermodynamics by direct method lattice dynamics of complex borohydrides. € Chapter 3 describes the experimental tools/techniques implemented for the analysis of complex borohydrides in our study including destructive and nondestructive methods/techniques. € Chapter 4 discuses the first principle inves tigation of stable crystal structure of zinc borohydride at finite temperature. € Chapter 5 studies the fundamental understandings of the role of Ni additives in promoting the dehydrogenation mechanism of hydrogen de-sorption in zinc borohydride. € Chapter 6 studies the crystal structure, electronic structure and dehydrogenation thermodynamics of Mn(BH4)2 and also compares with the experimental IR spectroscopy data. € Chapter 7 discuses the quantitative hydrogen storage capacity and compositions at each dehydrogenation reaction steps for the multinary complex borohydride LiMg-B-N-H system. It also discuses about the reversible hydrogen storage capacity and feasibility of application to the real fuel cell system for onboard application. € Chapter 8 presents the experimental study of Li-Mn-B-H system, member of a new class of complex borohydrides for hydrogen storage.

PAGE 28

12 € Chapter 9 summarizes the contents of this dissertation and suggests possible future research directions.

PAGE 29

13 Chapter 2 Theoretical Approach In this chapter, the basic concepts of theoretical approaches for the study of complex borohydrides are presente d. Firstly, the foundations of density functional theory calculations and the direct method la ttice dynamics calculatio ns are presented. 2.1. Density functional theory Density functional theory is a theory of correlated many-body systems. This is a remarkable theory that allows one to repl ace the complicated N-electron wave function ( x1, x2, …., xN) and the associated Schrdinger by the much simpler electron density ( r ) and its associated ca lculation scheme. Hohenberg and Kohn8 provided the basis for DFT in 1964 and proved that all electronic properties of the system coul d be uniquely defined by its ground state probability density. This provided an alternat ive to the wave functions. However, the theorems didn’t provide a practical scheme to calculate the ground state properties from electron density. The modificat ion was provided by Kohn and Sham9 (1965) who reformulated the problem of calculating the total electronic energy E as a functional of the electron density (r) to that of solving a set of single-particle Schrdinger like equations. Thus, ) ( ) ( 2 1 ) ( ) ( ] [ ] [ xc extE r d r u r d r v r T E + + + = 2-1

PAGE 30

14 T [ ], vext and u are kinetic energy, external poten tial and inter electronic repulsion, respectively. Exc is the exchange correlation functi onal which includes everything not contained in other terms. There are sever al approximations to calculate the correlation functional such as local density approximation (LDA)10, generalized gradient approximation (GGA)11 amongst several others. In LDA the spatial variations in density are ignored and potential at any point is calculated as th at for a homogeneous electron gas with the density at that point. In GGA, the exchange and corre lation energies are dependent not only on the electr on density, but also on the deri vatives of electron density. Recently, several hybrid approaches have been developed that allow even more accurate calculation of energies and structures12, 13. An extensive study by ab initio DFT calculations would allow us to develop a complete picture of the electronic structur e of complex borohydrides by (i) calculating the lattice parameter of th e crystal structure, (ii) th e systematic search for the thermodynamically stable phases of complex hydrides by calculating the enthalpy of formation of the complex hydrides, (iii) electr onic partial and total density of state (DOS) and band structure calculations of complex hydrides. 2.1.1. Hohenberg-Kohn theorem The approach of Hohenberg and Kohn8 is to formulate density functional theory as an exact theory of many-body systems The formulation applies to any system interacting particles in an external potential vext( r ), including any problem of electrons and fixed nuclei, where the Hamiltonian can be written as: U V T r r e r v m h Hj i j i ii i ext i eˆ ˆ ˆ 2 1 ) ( 2 ˆ2 2 2+ + = Š + + Š = 2-2

PAGE 31

15 The basic lemma of Ho henberg-Kohn theorem is that the ground-state density (r) of a bound system of interacting elec trons in some external potential vext( r ) determines this potential uniquely. The pr oof is reproduced as follow: Let (r) be the nondegenerate ground-state density of N electrons in the potential vext( r ), corresponding to the ground state, and the energy E Then, + + = = ) ) ( ( ) ( ) ( ) ( U T dr r r v H Eext 2-3 Where, H is the total Hamiltonian corresponding to vext( r ), and T and U are the kinetic and interaction energy operators. Now assume that there exists a second potential vext 1( r ), not equal to vext( r ) + constant, with ground state 1, necessarily ei which gives rise to the same (r). Then, + + = ) ) ( ( ) ( ) (1 1 1 1U T dr r r v Eext 2-4 Since is assumed to be nondegenerate, the Rayleigh-Ritz minimal principle for gives the inequality Š + = + + = < dr r r v r v E U T dr r r v H Eext ext ext) ( )) ( ) ( ( ) ) ( ( ) ( ) ( ) (1 1 1 1 1 1 2-5 Similarly, Š + = dr r r v r v E H Eext ext) ( )) ( ) ( ( ) (1 1 2-6 Where we use since the nondegeneracy of 1 was not assumed. Adding (2-5) and (2-6) leads to the contradiction. E + E1 < E + E1 2-7

PAGE 32

16 We conclude by reductive ad absurdum that the assumption of the existence of a second potential vext 1( r ), which is unequal to vext( r ) + constant and gives the same (r), must be wrong. The lemma is thus proved for a nondegenerate ground state. Since (r) determines both N and vext( r ) (ignoring an irrelevant additive constant) it gives us the full H and N for the electronic system. Hence (r) determines implicitly all properties derivable from H through the solution of the time-independent or time-dependent Schrdinger equation. The second lemma of the Hohenberg-Kohn theorem is that for any external potential applied to an interacting particle system, it is possible to define a universal total energy functional of the particle density, which is written as ] [ ) ( ) ( ] [ HK ext vF dr r r v E + = 2-8 Where FHK[ ] is universal functional by constru ction since the kinetic energy and interaction energy of the partic les are functionals only of th e density and is defined as ) ) ( ( ] [ + = U T FHK 2-9 dr r r T ) ( ) ( 2 1* 2-10 1 1 1 * 1) ( ) ( ) ( ) ( 1 2 1 drdr r r r r r r U Š 2-11 It follows that if the functional FHK[ ] is known, then by minimizing the total energy of the system, Equation (2-7), with respect to variations in the density function ( r ), one would find the exact ground state density an d energy. This means that the energy functional equals the ground-state energy for the correct ( r ), and has a minimum, given that the number of part icles of the system N [] ( r )d r is kept constant.

PAGE 33

17 2.1.2. Self consistent Kohn-Sham equation The Kohn-Sham approach9 (1965) is to replace the difficult interacting manybody system obeying the Hamiltonian with a differ ent auxiliary system that can be solved more easily. The ansatz of the Kohn and Sham assumes that the ground state density of the original interacting system is equal to that of some chosen non-interacting system. This leads to independent particles for the non-interacting system that can be solved with all the difficult many-body terms incorporated in to an exchange-correlation functional of the density as shown in Equation (2-1). The set of wave functions i( r ) that minimize the Kohn-Sham total energy functional are given by the self-c onsistent solutions to the Kohn-Sham shown in (2-12) and (2-13). ) ( ) ( 22r r V m hi i i eff = + Š 2-12 xc H ext effV V V V + + = 2-13 where i are the eigenvalues, i( r ) are the Kohn-Sham orbitals and Veff( r ) is the effective potential, VH is the Hartree potential given by Equation (2-14) Š = ' ) ( dr r r r VH 2-14 The Vxc, the exchange-correlation potential is the derivative of exchange-correlation energy functional with respect to the gr ound state density, given by Equation (2-15). ) ( )] ( [ ) ( r r E r Vxc xc = 2-15 By construction, the exact ground-state density, ( r ), of an N-electron system is given by

PAGE 34

18 2) ( ) (=i ir r 2-16 As can be seen, by solving the equations one can finds the ground state density and energy of original interacting system with the accuracy limited only by the approximations in the exchange-correlation functional. If the exchange-correlation potential defined in Equation (2-15) is known, the exact gr ound-state density and energy of the many-body electron problem can be found by solving the single-particle KohnSham equation. 2.1.3. Exchange and corelations f unctionals approximations The most important approximations for Exc[ ( r )] have a quasilocal form. The Exc[ ( r )] can be written in the form = dr r r r r Exc xc) ( )]) ~ ( [ ; ( )] ( [ 2-17 Where )]) ~ ( [ ; ( r rxc represents an exchange-correlation ( xc ) energy/particle at the point r which is a functional of the density distribution) ~ ( r It depends primarily on the density ) ~ ( r at points r ~ near r The simplest approximation for Exc is the so-called local-density approximation (LDA). dr r r Exc LDA xc= ) ( )) ( ( 2-18 Where xc() is the exchange-correlation energy per particle of a uniform electron gas of density. The exchange-correlation, xc() has two part exchange, x() and correlation, c() given, in atomic units, by s xr 458 0 ) (Š = 2-19

PAGE 35

19 Where rs, is the radius of a sphere containing one electron and given by 1 3 3 4Š= sr. The correlation part was first estimated by Wigner (1938)14: 8 7 44 0 ) (+ Š =s cr 2-20 and more recently with a high precision of about 1% by Ceperley (1978); Ceperley and Alder (1980) using Monte Carlo methods 15, 16. A better way to improve the LDA is to consider the exchange-correlation energy density not only depending on the density (r) but also on its gradient (r). These new expressions are called generalized gradient approximation (GGA). ] ; [ =xc GGA xcE E 2-21 In practice, there are various forms of GGA for different requirements, such as the Perdew-Becke form, the Perdew-Wang 86 form and the Langreth-Mehl-Hu form, etc. Becke introduced another approach to improve the LDA is called hybrid method: GGA xc KS x hyb xcE E E ) 1 ( Š + = 2-22 where KS xE is the exchange energy calculated with the exact KS wave functions, GGA xcE is an appropriate GGA, and is a fitting parameter17. Use of GGA’s and hybrid approximations instead of the LDA has redu ced errors of atomization energies of standard sets of small molecules, consisting of light atoms, by fact ors of typically 3–5. The improved accuracy of exchange-correlation functionals and capabili ty of DFT to deal with systems of many atoms, has, over the time DFT become the preferred method for calculating total energies and hence cohesive energies of most condensed-matter systems including solid stat e complex hydrides.

PAGE 36

20 2.2. Solving self consistent Kohn-Sham equation In Equation (2-12) the effective pote ntial depends on the electron density (r), which depends on the Kohn-Sham orbitals i( r ), which are being searched. It means that the Kohn-Sham have to be solved iteratively until a self-consistent solution being reached. The schematic diagram is shown in Figure (2-1). Figure 2-1. Schematic representation of the se lf-consistent loop for solution of Kohn-Sham equation18.

PAGE 37

21 2.2.1. Basis set for solution For a solid state system, the effective pot ential in the Kohn-Sham Equation (2-12) has the periodicity of the crystalline latt ice. Therefore, the Kohn-Sham orbitals, ) ( rn k can be written as a product of a function ) ( r un k that has the periodi city of the crystal lattice and a plane wave r ike. with k being any wave vector in the first Brillouin zone, i.e. r ik n k n ke r u r.) ( ) ( = 2-23 According to the Bloch theorem, a plane wave basis set is used, the periodic function in Equation (2-23) can be written as a sum over plane waves that have the same periodicity as it. Such plane waves are those ones correspo nding to reciprocal lat tice vectors. Thus, the expansion of ) ( rn k in this basis set is + =j r K k i n j cell n kje k C r). () ( 1 ) ( 2-24 Where cell is the volume of the primitive cell, Kj are the reciprocal lattice vectors and the parameter n is called band index. It should be poi nted out that this basis set is kdependent, i.e. the eigenstates ) ( rn k corresponding to different k vectors are represented by different basis sets. The plane wave basis set can be truncated by setting all Kj with K Kmax. This means that all reciprocal lattice ve ctors that are inside a sphere with radius Kmax are taken into the basis set. It is more common to specify the free electron energy corresponding to Kmax, which is called cut-off energy m K Ecut 2 2 max 2= 2-25 However the basis set will still be intractably large for systems that contain both valence and core electrons. If we look at the ra dial part of a conduction electron wavefunction,

PAGE 38

22 we see that close to the atom core it displa ys a nodal behavior, while further away from the core; the wavefunction is smooth and vary as plane waves. The nodal behavior ensures orthogonality of the wavefunctions. Because the radial wavefunctions vary intensely with the smallest radii, many plane waves are required to describe this part of the wavefunction, while much lesser number plane waves are required for the outermost parts. A smart way to reduce the number of plane waves needed is the use of pseudopotentials. 2.2.2. Pseudopotentials and ultrasoft pseudopotential In the previous section, many plane wa ves are needed to describe the nodal behavior close to the core region of the radi al wavefunction of a c onduction electron. If we treat the conduction electrons in the inte rstitial regions as plane waves, the energy must depend on the wavevector approximately as m kk 2 2 2= as for free electrons19. Using the Hamilton operator, one can calculate the energy of an orb ital at any point in space. This means that in the interstitial re gions, this energy will be close to the above free electron energy. The wavefunctions’ behavi or close to the core does not affect the dependency of on k much, or in other words; the outer regions are mainly responsible for the chemical reactions. This leads to th e idea of replacing the core potential with a .pseudopotential that giv es the same wavefunctio ns outside the core20, 21, but simpler ones inside the core. This is why the pseudopotentia l best fits the outer side and nearly zero fits inside the core19. An example of pseudopoten tial is shown in Figure 2-2.

PAGE 39

ou in ps de ge ro hi ca co F igu re P se udo ou termost ele in cludes s om ps eudopote nti de scri be t h em ge neral sol id ro w an d t ra ns hi gh ly l oca liz can b e mad e co nstraint l eav re 2-2 Illus trat udo potentia ls c ele ctrons an d om e of the c nti al is cal led em Altho ugh state c alc ula ns iti on m et al e liz e d p and d e softer by pu eav es little r oo trat ion of the re ps ls c an b e s o ft d requires on c ore elect ron led ultrasoft ugh the nor m ula tions, th eir al e lement s. T o rbitals w hic pu shing th e c oo m for any si 23 re al and ps eu ps eudopote nti ft or hard. A on ly a few p ron s and r equ as v ery sm c onservin g eir applicati on T he difficu lty hic h are alr ead e c utoff rad ius si gnificant im eu do wa ve fun ntial 22. A soft pse udo p lane wav es equ ires hen ce sm all numb er g pseudo p o te on is limite d f lty lies in t he ead y nodele ss. ius outwa rd 23, im proveme nt fun ction and th udo potentia l in es 23 A ha rd ce more p lan er of plane w te ntials can b d f or system s he inefficien cy ss. T he pse ud 3, 24 but th e n nt in the pro ced th e real an d l in cludes o nly rd pseudop ote lan ewaves. S w aves nee de b e used in m s containin g cy to repres en ud o wave f unc e n or m c on ser ced ur e. nly th e ote nti al S om e de d t o m an y g fir st ent t he unc tio n ser vin g

PAGE 40

24 2.2.3. Projector-augmented-plane-wave method The projector augmented wave (PAW) method25 is a general approach to solution of the electronic stru cture problem that reformulat es the orthogonalized plane waves (OPW), adapting it to modern techniques fo r calculations of total energy, forces and stresses. Like the ultrasoft pseudopotential method, it introduces projectors and auxiliary functions. The PAW approach al so defines a functional for th e total energy that involves auxiliary functions and it uses advances in algorithms for efficient solution of the generalized eigenvalue problem. The PAW method is an allelectron frozen core method, which combines the features of both the ultra-soft p seudopotentials and linear augmented plane wave (LAPW) methods. It allows an easier treatment of f irst-row and transitionmetal elements, and provides access to the full wave function. In real materials, the wave function is fairly smooth in the bonding region, whereas it va ries rapidly near the nucleus due to the large attractive potential. This is the main difficulty for electronic structure methods to describe the bonding region with accuracy as well as account for the large variations in the atomic core. The augmente d-wave methods deal with this problem by dividing the wave function into two parts, i.e., a partial-wave expansion within an atomcentered sphere and envelope functions outside the spheres. The valu e and derivative of the two parts are then matched at the sphere radius. In the PAW method, one can define a smooth part of valence wavefunction ) ( ~ rv i or an atomic orbital and a linear transformation v v ~ =that relates the set of allelectron valence functions ) ( ~ rv j to the smooth functions) ( ~ rv i The transformation is assumed to be unity except with a sphere centered on the nucleus,01 + =. For simplicity, we omit the superscript v assuming that the s are valence states, and the

PAGE 41

25 labels i j Using the Dirac notation, the expansion of each smooth function ~ in partial waves m within each sphere can be written Equation (2-26), =m m mc ~ ~ 2-26 with a the corresponding all-electron function, = =m m mc ~ 2-27 Hence the full wave function in all space can be written Š + =m m m mc } ~ { ~ 2-28 If the transformation is required to be linear, then the coefficients must be given by a projection in each sphere ~ ~ m mp c = 2-29 for some set of projection operatorsp ~ If the projection operators satisfy the biorthogonality condition, ' ~ ~ mm m mp = 2-30 Then the one-center expansion ~ ~ ~ m m mp of the smooth function ~ equals ~ itself. The requirement that the transformation in Equation (2-26) is linear leads to the following form for the operator m m m mp ~ } ~ { 1Š + = 2-31 Furthermore, the expressions apply equally well to core and valence st ates so that one can derive all-electron results by applying the ex pressions to all the el ectron states. The general form of PAW s can be cast in term s of transformation by Equation (2-31). For

PAGE 42

26 nay operator A ˆ in the original all-electron problem, one can introduce a transformed operator A ˆ that operates on the smoot h part of the wavefunctions Š + = =' ' ~ } ~ ˆ ~ ˆ { ~ ˆ ˆ ~ mm m m m m m mp A A p A A A 2-32 One can add an operator of the form in the right hand side ' ~ ~ ˆ ~ ~ ˆm m mm m mp B p B Š 2-33 with no change in the expectation values. For example, one can remove Coulomb singularity for the smooth functi on, leaving a term that can be dealt with in the radial about each nucleolus. Now the physical quantit ies can be obtained from Equations (2-31) and (2-32). For example, the charge density is given by Equation (2-34) ) ( ~ ) ( ) ( ~ ) (1 1r r r r Š + = 2-34 which can be written in terms of eigenstates labeled i with occupancy fi as =i i ir f r2) ( ~ ) ( ~ 2-35 i m m m mm m i i ir r f r ~ ~ ) ( ) ( ~ ~ ) (' ' 1 = 2-36 and i m m m mm m i i ir r f r ~ ~ ) ( ~ ) ( ~ ~ ~ ) ( ~ ' 1 = 2-37 Similarly, the energy functiona ls can be written as 1 1 ~ ~ E E E E Š + = 2-38 The last two terms 1( r ) and ) ( ~ 1rare localized around each at om and the integrals can be done in spherical co-ordinates with no problems from the string variables near the nucleus. They contain contributions from the core and smooth core states, respectively.

PAGE 43

27 In practice, the smooth core density is cons tructed instead of constructing a smooth core state for each core state individually unless one is interested in the physical properties that are related to the core states. 2.3. The Vienna ab-initio simulation package The Vienna ab-initio simulation package (VASP)18, 26, 27 is used to perform firstprinciples total-energy calculations within the density functional theory. The VASP simulates the electronic and atomic structures of molecules and crystals. It uses ultrasoft pseudopotentials or the projec tor-augmented wave method and a plane wave basis set. The expression ab initio means that only first-principles are used in the simulations, no experimental data are needed. VASP has a la rge amount of settings and specifications in the database for tailoring the calculations for different systems. The inputs to the program are text-files of whic h the most important are: INCAR POSCAR KPOINTS and POTCAR The INCAR file contains the basic instru ctions to VASP, such as which algorithm to use, what precision is required for the result, which way to relax the structure, (conjugate gradient, MD, etc.), energy cut-off for the plane waves, the EXC approximation and much more. In the POSCAR file the size of the supercell is specified along with the positions of the atoms and, if re levant, initial velocitie s and constraints on movement. The KPOINTS file controls which k -point scheme to use and finally the POTCAR file contains the atomic mass and the ultrasoft pseudopotentials or PAW potentials for the relevant atoms and the chosen EXC approximation from the database.

PAGE 44

28 2.4. Direct method lattice dynamics a nd finite temperature thermodynamics DFT is a non-empirical parameter met hod whose applications and predictive ability in different fields known for some time But, the results we obtain from DFT are always at 0 K which may not be viable in pr actical applications. Hence combination of DFT with different techniques such as linear response method28, 29 or direct methods30-32 allows us to evaluate phonon dispersion curves without empirical parameters. In the direct method33, 34, the forces are calculated via the Hellmann-Feynman theorem using DFT-derived total energies, assuming a fini te range of interaction. The phonon spectra are then derived using Newton’s equation of motion in lattice dynamics calculations. = ) ( ) ( ) ; (2 ,v m R n R E v m nj i j i 2-39 =j v m j j i iv m U v m n n F, ,) ( ) ; ( ) ( 2-40 []) ( ) 0 ( 2) ; 0 ( 1 ) ; (v m R R ik m ve v m M M v k DŠ= 2-41 ) ( ) ( ) ( ). (2j k e j k j k e k D= 2-42 Š = j kj k nd g,)) ( ( 1 ) ( 2-43 = < Š 2 2 1 0) ( x if otherwisex 2-44 Where, ) ; (,v m nj i is force constant matrix, ) ( n Fiis force due to atomic displacement jU, d is the number of degree of freedom in the unit cell, D(k;,) is called the dynamical matrix, ) (2j kis called the phonon frequency which can be

PAGE 45

29 obtained by solving the Newton and ) ( gis called the phonon density of states which is frequency distribution over normal modes. The partial phonon DOS can be calculated using the following: Š = j k i ij k j k e nd g, 2 ,)) ( ( ) ; ( 1 ) ( 2-45 Integrating phonon DOS we can calculate the vi brational internal energy and entropy as follow: ) 2 coth( ). ).( ( 2 1, 0 ,T k g d d EB i i = 2-46 = , ,) (i i vibE T E 2-47 )]} 2 exp( 1 ln[ ] 1 ) 2 )[coth( 2 ).{( ( ., 0 ,T k T k T k g d d k SB B B i B i Š Š Š = 2-48 = , ,) (i i vibS T S 2-49 Where, a contribution from any atom and degree of freedom i to the internal energy and entropy are given by Ei, and Si, respectively. 2.5. The PHONON package For building the crystal th e best is to use PHONON35. The supercell with a given space group can also be created using this. The supercell can be equal to the primitive unit cell of the crystal. Next, transfer the structural data ( POSCAR ) to an ab initio software, for example VASP, and minimize the system energy. The optimized lattice constants, and atomic positions are then re used by PHONON to creat e a large supercell, which will be used to calculate phonons sp ectra. Transfer the crystal data from

PAGE 46

30 PHONON to VASP program and optimize agai n the large supercell. Usually it is sufficient to optimize the atomic positions only (in VASP ISIF=2). One should quickly reach the minimum, since the input structural data correspond alrea dy to the minimum. In VASP create a new POSCAR with atomic positions where one atom is displaced. For each configuration with displaced atom cal culate the Hellmann-Feynman forces from a single minimization run of electronic s ubsystem only (NSW=1). The HellmannFeynman forces are then collect ed to a single file. Then import the generated HellmannFeynman file to PHONON, and calculate phonon dispersion curves, density of states, etc to calculate the finite temp erature thermodynamic properties.

PAGE 47

31 Chapter 3 Experimental Approach 3.1. Experimental details Starting precursor materials and various nano-catalytic dopa nts were obtained either from Sigma Aldrich or from Quantu mSphere Inc., CA and they were used to synthesize and characterize our desired hydroge n storage materials without any further purification. High purity H2 (99.9999%), N2 (99.99%) and He (99.99%) were procured from Airgas for the synthesis and analytical measurements. All chemical reactions and operations were performed in a nitrogen fill ed glove box. Precursor materials with different stoichiometric mole ratio were mixe d in a stainless steel bowl (80 ml) and the lid sealed with viton O-ring in the glove box. The bowl was then evacuated for 30 minutes to remove the residual oxygen and moisture down to ppm levels. A specially designed lid with an inlet and outlet valves were used for this purpose. The mechano-chemical process employing high energy milling was carried out by Fritsch pulversette planetary mono mill, P6 in an inert atmosphere. The milling parameters, ball to powder weight ratio a nd milling speed were optimized to 20:1 and 300 rpm respectively. Milling duration w as optimized for all the samples. These mechano-chemically processed complex hydrides were immediately transferred to the glove box for further characterizations. In a similar way, few mole concentrations of

PAGE 48

32 nano-catalytic dopants such as nanoNi, nanoC o, nanoFe, etc. were added during the milling process for the synthesis of nanocatalyst doped complex borohydrides. The B–H bond stretch/bending of the complex borohydrides system was measured using a Perkin-Elmer Spectrum One FTIR spectrometer. This instrument operates in a single-beam mode and is capab le of data collection over a wave number range of 370–7800 cm1 with a resolution of 0.5 cm1. The complex borohydrides samples were palletized and sealed in a specially designed KBr cell for infrared measurements. The powder X-ray diffraction of the mech ano-chemically milled complex hydride was carried out by the Philips X’pert diffractometer with CuK radiation of = 5.4060 . The incident and diffraction slit widt hs used for the measurements are 1o and 2o, respectively. The incident mask of 10 mm was used for all the samples and their XRD studies. The sample holder (zero background s ilicon disc of 32 mm diameter procured from the Gem Dugout, Pennsylvania, and USA) was covered with polyethylene tape (foil) with O-ri ng seal in a N2 filled glove box to avoid the O2/moisture pickup during the XRD measurements. Diffraction from the tape was calibrated without the actual sample and found to be occurring at the 2 angles of 22o and 24o, respectively. The XRD phase identification and particle si ze calculation was carried out using the PANalytical X’pert Highscore software with built-in Scherer calculator. The simultaneous DSC and TGA (SDT) analys is pertaining to the weight loss and the heat flow for the reaction enthalpy during thermal decomposition of undoped and nanomaterial doped complex hydrides were performed by using the TA instruments’ SDT-Q600 analytical tool. Calibration of SD T was performed in f our steps with empty

PAGE 49

33 pan and standard sapphire disc. The four calibration subr outines of TGA weight, DTA baseline, DSC heat flow and temperature were carried out before an actual measurement on the sample. A pre-weighed sample was loaded into the ceramic pan and covered with the ceramic lid inside the glove box to prev ent moisture from getting into the sample during transfer. A ramp rate of 2oC/min was used for all the measurements. TA instruments’ Universal Analysis 2000 softwa re was used to analyze the TGA and DSC profiles. The isothermal volumetric measureme nts were carried out by Hy-Energy’s PCTPro 2000 sorption equipment. This fully automated Sievert’s type instrument uses an internal PID controlled pressure regulator with maximum pressure of 170 bars. It also includes five built-in and calibra ted reservoir volumes of 4.66, 11.61, 160.11, 1021.30 and 1169.80 ml. The volume calibration without and with the sample was performed at a constant temperature with an accuracy of 1oC using a helium gas. The software subroutines for hydrogen purging cycles, leak t est, kinetics, PCT and cycling, etc. were performed by the HyDataV2.1 Lab-View progr am. The data collected from each run were analyzed using the Igor Pro 5.03 pr ogram with a built in HyAnalysis Macro. Temperature programmed desorption ( TPD) measurement was carried out by Autosorb-1 equipment of Quantachrome Inst rument. A 100 to 120 mg amount of sample was loaded in the reactor and heated, in a 25 mL/min helium flow while heating from 25 to 200 C at constant heating rate. The thermal desorption/reduction profiles were recorded and analyzed using TPRWIN software package. The gas chromatography (GC) analys is was carried out during thermal programmed desorption process for both undoped and nanocatalyst doped samples. The

PAGE 50

34 gas sample was injected (less than 50-100 mi cro liter) in to the TCD detector and recorded the GC signal over a period of retention time. The gas analysis and plotting of the curves were carried out by Sa turnview Version 5.52 software. 3.2. Thermo analytical tools/techniques 3.2.1. Differential scanning calorimetry The differential scanning calorimetry (DSC) is a thermal analysis tool and it measures the difference in the amount of heat required to increase the temperature of a sample with respect to an inert referen ce as a function of time and temperature. The specific Q10 differential scanning calorimeter manufactured by TA instruments consists of two closed pans, one is for the sample and the other is working as a reference pan (which is generally an empty) inside a helium filled chamber at 50-70 psi gas pressure. Figure 3-1 shows the DSC Q1 0 apparatus. The heat flow difference between the two pans is monitored while they are heated, or cooled, uniformly. This can be done at a constant temperature (isothermally ) or at a constant temperature rate (ramp). Figure 3-1. Differential scanning ca lorimetry (DSC) Q10 apparatus

PAGE 51

35 The basic working principle in this to ol/technique is that, when the sample undergoes a physical transformation such as pha se transitions, an extra amount of heat is required/withdrawn to/from the sample pan to maintain both the pans at the same temperature. The process can be exothermic or endothermic depending on the characteristics of the sample. Some examples of endothermic processes are glass transition, melting, evaporation, et c. and some exothermic ex amples are crystallization, oxidation, etc. 3.2.2. Thermo gravimetric analysis Thermo gravimetric analysis (TGA) is also a thermal analysis tool which uses heat as an external force to occur chemi cal reactions or/and physical changes in the materials under investigation. TGA measures the amount and rate of change in the weight of a material due to dehydration, d ecomposition, and oxidation as a function of temperature or time in a controlled atmosphere Molecular structure gives characteristic thermo gravimetric curves for specific materials due to unique sequence from physicochemical reactions occurring over spec ific temperature ranges and heating rates. The specific TGA used in our work is the Q600 series; it gives an accurate measurement of simultaneous DSC-TGA (SDT) of the same sample. A matched Platinum/PlatinumRhodium thermocouple pair embedded in the ceramic beams provides direct sample, reference, and differential temperature measu rement and a dual balance mechanism gives an accurate weight loss/gain measurement. The schematic cross-sectional diagram of Q600 SDT36 is shown in Figure 3-2.

PAGE 52

36 Figure 3-2. Schematic diagram of Q600 balance/furnace36 The balance operates based on a null-balance principle. At the zero position, equal amounts of light shine on the two photodiod es. If the balance moves out of the null position an unequal amount of light shines on th e two photodiodes. A certain amount of current is then applied to the meter to retu rn the balance to the zero position. The amount of applied current is proportional to the weight loss or gain. 3.2.3. Thermal programmed desorption autosorb-1 Temperature programmed desorption (TPD) is a technique in which the amount of desorption/reaction is monitored as a functi on of temperature. The temperature is raised in a linear fashion so that a suitabl e detection system can record a characteristic desorption/reaction profile of the sample being tested. A high sensitive thermal conductivity detector (TCD) is used to mon itor the process. The TCD signal is also proportional to the quantity of desorbed gas molecules as the thermal energy overcomes the binding energy. In Figure 3-3 shows the basic schematic diagram37 of the TPD Autosorb-1 equipment manufactured by Quantachrome Instruments.

PAGE 53

37 Figure 3-3. Schematic diagrams of the autosorb-1 TPD/TPR apparatus37 The sample between two quartz wool pieces is purged with nitrogen/He in order to avoid O2 in the sample and inside the lines, then for the TPR measurement, a mixture of 95% Ar and 5% H2 flows over the sample with a te mperature ramping, a monitoring process records the amount of hydrogen abso rbed based on the signal in the B TCD detector filament and compared with the A TCD filament. For the TPD measurement, the hydrogen mixture is replaced by an inert ga s and the TCD is changed to TPD option. An amount of 100 to 120 mg of sample was loaded in the reactor and heated, in a 25 mL/min helium flow while heating from 25 to 200 C at constant ramping and then the thermal desorption/reduction profiles were recorded and analyzed using TPRWIN software package.

PAGE 54

38 3.2.4. Pressure composition temperature apparatus The instrument used is the PCTPro-2000 manufactured by Hy-Energy, USA. This fully automated Sievert’s type instrument uses an internal PID controlled pressure regulator with maximum pressure of 170 bars. It also includes five built-in and calibrated reservoir volumes of 4.66, 11.61, 160.11, 1021.30 and 1169.80 ml. The equipment is designed to measure absorption/desorption beha vior of the complex hydrides. Important measurements can be done with this inst rument; they are pressure composition temperature (PCT), absorption/deso rption kinetics, cycle-life, etc. In the Sievert’s technique, a calibrated re ference volume at a constant temperature is filled with hydrogen gas to a measured pressure and then opened to the sample chamber; the gas uptake by the sample is calculated from th e change in the hydrogen gas pressure in the system. Hydrogen uptake is represented by the hydrogen-to host atomic ratio, H/X, by analogy with the hydrogen-to-metal ratio for a metal, H/M. Similarly, the gas release by the sample is measured from th e gas pressure difference in the system. The helium bottle (purity 99.9999%) is used fo r calibrating the volume and for purging the line from air. The Sievert’s ty pe volumetric hydrogen sorption system38 is shown in Figure 3-4, where Res 1 and Res 2 are the gas reservoirs.

PAGE 55

39 Figure 3-4. Schematic diagrams of Si evert’s type volumetric apparatus38 3.3. Chemical analytical tools/techniques 3.3.1. Fourier transform infrared spectrometer Infrared (IR) spectroscopy is a chemical analytical technique, which measures the infrared intensity versus wavelength of light. Infrared spectroscopy detects the vibration characteristics of chemical functional groups in a sample. Chemical bonds in the matter may stretch, contract and bend when an infrar ed light interacts with it. As a result, a chemical functional group tends to adsorb infrared radiation in a specific wavenumber (frequency) range regardless of the stru cture of the rest of the molecule. In a Fourier transform infrared spectrometer (FTIR), the light passes through an interferometer which creates constructive and de structive interference pattern of two light beams, the recombined beam passes through the sample. The sample absorbs all the different wavelength characte ristics of its spectrum de pending on functional groups, and this subtracts specific wavelengths from th e interferometer. The detector now detects

PAGE 56

40 variation in energy versus time fo r all wavelengths simultaneously39. A Perkin-Elmer Spectrum One FTIR Spectrometer is shown in Figure 3-5. Figure 3-5. Perkin-Elmer spectrum one FTIR spectrometer Determining these frequencies allows us to identify the chemical fingerprint of the sample, since chemical functional group s are known to absorb radiation at specific frequencies. The intensity of the absorption light is related to th e concentration of the component. Intensity and frequency of abso rption by the sample ar e depicted in a twodimensional plot called a spectru m. Intensity is generally reported in terms of percent transmittance, the amount of light that passes through it. 3.3.2. X-ray diffractometer X-ray diffraction (XRD) is an efficient an alytical technique used to characterize and identify unknown crystalline materials. Monochromatic X-rays are used to determine the interplanar distances of the unk nown materials. With this technique the samples are analyzed as powders with grains in random orientations to ensure that all crystallographic directions ar e "sampled" by the beam. The basic principle of operation

PAGE 57

41 of the XRD spectrometer is based on Bragg’ s law. When the Bragg conditions for constructive interference are obtained, a "reflection" is produced, and the relative peak height is generally proportional to the number of grains in a preferred orientation. A Philips X’Pert XRD system is shown in Figure 3-6, it consists a source of monochromatic radiation and an X-ray detector which is situated on the circumference of a graduated circle centered on the powder sam ple, a detector and sample holder are mechanically coupled with a goniometer in such a way that a rotation of the sample holder through degrees occurs in conjunction with the rotation of the detector through degrees, a fixed 1:2 ratio. Divergent slits, placed between the X-ray source and the specimen, and between the sample and the detector, limit scattered (non-diffracted) radiation, reduce background noise, and collimate the radiation. Figure 3-6. The Philips X’Pert XRD system

PAGE 58

42 3.3.3. Scanning electron microscope Scanning electron microscopy is a tec hnique by which one can get topography, morphology, compositions and crystallographic information of an object. The scanning electron microscope (SEM) uses electrons rath er than light to form an image. The Hitachi S800 scanning electron microscope used in the present study is limited to magnifications of around 3,000k. The SEM uses the secondary electrons to form the image of the surface of the specimen since they do not diffuse much inside the specimen. In Figure 3-7 shows the details of interaction between an electron beam and specimen. Figure 3-7. Electron beam and specimen interaction signals40 The schematic basic diagram of the operati on of the Hitachi S 800 SEM is show in Figure 3-8. Electrons beams come from an electron gun filament targeted at the specimen inside a vacuum chamber. That beam is collimated by electromagnetic condenser lenses, focused by an objective le ns and then swept across the specimen at high speed. The secondary electrons are de tected by a scintillation material which

PAGE 59

43 produces electron flash lights. The flash li ghts are then detected and amplified by a photomultiplier tube. Figure 3-8. Schematic working principle diagram for a SEM41 3.3.4. Energy dispersive X-ray spectroscopy Energy dispersive X-ray spectroscopy (EDS ) is a technique used to identify the elemental composition of a samp le or a section of interest within the sample for all elements with an atomic number greater than boron (B). When the electron beams of the SEM hits the surface of the sample, it produc es X-ray fluorescence from the atoms in its path. The energy of each X-ray is the charact eristic of each element which produces it. The EDS microanalysis system collects all th e X-rays, sorts and pl ots them by energy level, and automatically identifies and labels the elements which are responsible for the peaks in those energy distributions. The detector cooled via liquid nitrogen is used to capture and mapped the X-ray counts continuously.

PAGE 60

44 3.3.5. Gas chromatography Gas chromatography (GC) is a technique which is used for common chemical confirmation test. GC analysis separates all of the components in a sample and provides a representative spectral output corresponding to each component. There is a carrier gas is used in this tec hnique, which carry the component through a long column. The time elapsed between injection and elution is called the "retention time." This retention time can help to differentiate between some com pounds. By changing the carrier gas flow rate and the temperature, the retention time can also be altered.

PAGE 61

45 Chapter 4 First-Principles Investigation Of The Zn(BH4)2 4.1. Abstract First-principles calculations were performed on Zn(BH4)2 using density functional theory (DFT) within the local density appr oximation (LDA) and the projected augmented wave (PAW) method. Zn(BH4)2 is a promising candidate for hydrogen storage with a capacity of 8.5 wt%. A direct me thod lattice dynamics approach using ab initio force constants was utilized to cal culate the phonon dispersion curv es. This allowed us to establish stability of the crystal structure at finite temperatures. DFT was used to calculate electronic properties and the d irect method lattice dynamics was used to calculate the finite temperatur e thermal properties. Zn(BH4)2 is found to have an orthorhombic structure in the space group of Pmc21 with lattice parameters a = 4.118 , b = 4.864 , c = 7.916 . It is an insulating materi al having a DFT-calculated band gap of 3.529 eV. Analysis of the electroni c structure shows strong bonding between hydrogen atoms and boron in the [BH4]complex and also less polar bonding between the Zn and the hydrogen atom. The reaction enthalpy was calculated for the reaction Zn(BH4)2 = Zn + 2B + 4H2(g) to be 76.91 kJ/mol of H2 at 0 K without any zero point energy correction, and 59.90 kJ/mol of H2 including the zero point energy correction. The simulated standard enthalpy of formation for the complex Zn(BH4)2 was found to be -66.003 kJ/mol of H2 at 300 K. This suggests that the crystal structure of Zn(BH4)2 is

PAGE 62

46 stable at room temperature and this comple x hydride can thus be considered a potential candidate for hydrogen storage. 4.2. Introduction A recent challenge in hydrogen storage is to find light weight complex solid hydrides which have higher gravimetric system capacity greater than 6.5 wt % for onboard vehicular applications. The technical ch allenge is to find materials that can exhibit favorable thermodynamics and kinetics for hydrogen de-sorption and absorption, and have the ability to store a sufficient am ount of hydrogen by weight as well as by volume percent. Borohydride complexes as hydrogen storage materials have recently attracted great interest. The stabilities of borohydrides have been studied using first-principles calculations. Some researchers have reporte d that alkali borohydrides are too stable for hydrogen storage42, 43. Lithium borohydride (LiBH4) possesses a theoretical hydrogen capacity of ~18.3 wt. %, exhibiting poten tial promise for on-board applications. However, hydrogen decomposition from LiBH4 starts at an elevated temperature of 380 C and, also, this compound shows little or no reversible hydrogenation behavior. It has been reported that catal ytically doping with SiO2 lowers this temperature of hydrogen evolution to 300 C44. A systematic approach to study the phase stability of LiBH4 based on ab initio calculations has been presented and four thermodynamically stable phases have been identified, including a new phase of Cc symmetry for the first time for LiBH4 45. A correlation has been reported be tween thermodynamic stability of metal borohydrides and Pauli electronegativity of th e parent components us ing first principles calculations46.

PAGE 63

47 Most of the commonly known borohydrides are found to be unsuitable for onboard hydrogen storage applications. This is primarily attributable to the high stability resulting from very high decomposition temperatures or complete irreversibility of hydrogen desorption47. Zn(BH4)2 is considered a potential candidate for on-board applications as it has a very high theoretical hydrogen stor age capacity of 8.5 wt% and a reasonably low decom position temperature (85 oC)48 compared to complex alkali borohydrides (greater than 300 oC). Density functional theory (DFT) is at pr esent considered to be a versatile and important tool to solve innovative research pr oblems in metal-hydrogen interactions and associated mechanisms. Successful applicatio n of DFT to a materials problem involves three distinct steps; (i) tran slation of the engineering probl em to a computable atomistic model, (ii) computation of th e required physicochemical prop erties, and (iii) validation of the simulation results by comparis on with laboratory experiments49. Lattice dynamics50, 51 using a direct force constant method helps one complete the picture by allowing for calculation of the finite temperature therm odynamic properties. Recently, in hydrogen storage materials research, DFT calculations have been employed to understand and validate the catalytic behavior of T ispecies on the dehydrogenation of NaAlH4 clusters52-54. Crystal structure stability and elect ronic structure for storing high hydrogen content in light weight complex hydrides55, 56 have been reported on the basis of DFT calculations. Although Zn(BH4)2 is a potential candidate for hydrogen storage, very little is known about its structural and thermodynamic properties. In this study, we performed ab initio DFT calculations to establish the 0 K crysta l structure and electronic structure. We

PAGE 64

48 then performed lattice dynamics calculations32 to determine the finite temperature reaction enthalpy of Zn(BH4)2. We employed the criterion that 02> from the phonon frequencies to establish finite-temperature st ability of the crystal structure determined from DFT calculations. The standard enthalpy of formation is an important predictor of decomposition temperature for the decomposition reaction w ith its associated standard thermodynamic parameters. For a complex hydride, when th e material is being heated, the effect of standard entropy starts dominating the sta ndard enthalpy and at the decomposition temperature and constant pressure the standard Gibbs energy is zero. As during heating the entropy change of solid materials is very small compared to the gas, we can consider that the entropy change during the decom position reaction is primarily due to H2 evolution. Now at standard pressure and temp erature for the most simple metal hydride, ) (2H S S = 130.7 J/mol K57 which suggests that, for a simple hydride, the enthalpy of decomposition at room temperature (300 K) and 1 bar pressure is about 39 kJ/mol of H2. For complex metal hydride materials it has been suggested that if the enthalpy of reaction is between 30 and 60 kJ/mol of H2,58 we can expect the complex metal hydride to be reversible for hydrogen storage. We extended our theoretical calculations to identify the most favorable dehydrogenation reaction for Zn(BH4)2 by calculating the reaction enthalpies of different decomposition reactions. We e xpect that upon heating, Zn(BH4)2 releases hydrogen according to the following reaction during decomposition: Zn(BH4)2 Zn + 2B + 4H2(g) 4-1

PAGE 65

49 Other possible different decomposition r eactions are given in Equations (4-1 a )-(4-1 d ). Formation reactions for the decomposition products are given in Equations (4-2) and (43). Formation enthalpy is one good way to establish whether theoretically predicted phases are likely to be stable. Such results are also expected to guide us discover convenient synthesis routes. Zn(BH4)2 Zn + B2H6 + H2 4-1 a Zn(BH4)2 ZnH2 + 2B + 3H2 4-1 b Zn(BH4)2 2 1ZnH2 + 2 1Zn + 2B + 2 7H2 4-1 c Zn(BH4)2 ZnH2 + B2H6 4-1 d 2B + 3H2 B2H6 4-2 Zn + H2 ZnH2 4-3 4.3. Computational methods 4.3.1. Ab initio First-principle calculations were performed on Zn(BH4)2 using DFT within the local density approximation (LDA)59 and projected augmented wave (PAW)25, 60 method utilizing a plane wave basis set to calculat e the total energies, as implemented in the Vienna Ab initio Simulation Package (VASP)18, 26, 27. A 5 5 5 Monkhorst-Pack61 kpoint mesh was used for sampling the Brillouin zone. A kinetic energy cutoff of 312 eV was used for Zn(BH4)2 for the plane wave basis set. For Zn, B, H2 and B2H6, the same kinetic energy cutoff values as for Zn(BH4)2 were utilized, in the self consistent total energy (Table 4-1) calculations. This allows for the maximum cancellation of errors in the reaction enthalpy evaluation, which are calculated as small energy differences

PAGE 66

50 between reactants and products (Table 4-2). The criterion for se lf-consistency in the electronic structure determination was that tw o consecutive total en ergies differed by less than 0.001 meV. The atomic positions and th e lattice parameters, including the unit cell volume, were optimized by minimizing the forc es and stresses until the residual forces between the atoms were within 1.0 meV/. Table 4-1. Ground state energy, E in eV/mol of Zn, B and H2 from DFT calculations Element/ Compound Space Group (number) / type E (this work) (eV/unit-formula) E (Literature) (eV/unit-formula) Zn P 63/mm c (194) / hcp 62, 63 -0.922 -1.02 H2 P 63/mm c (194) / hcp 64, 65 -6.782 -6.792 -B R -3m (166) / trigonal 66, 67 -6.678 -6.479 4.3.2. Direct method lattice dynamics DFT is a non-empirical parameter met hod whose applications and predictive ability in different fields ar e known for some time. Combination of DFT with different techniques such as linear response method28, 29 or direct methods30-32 allows us to evaluate phonon dispersion cu rves without empirical parameters. Parlinski et al.33, 34 developed the direct method wh ere the forces are calculated via the Hellmann-Feynman theorem using DFT-derived total energies, as suming a finite range of interaction. The phonon spectra are then derived using Newt on’s equation of motion in the lattice dynamics calculations.

PAGE 67

51 PHONON code35, based on the harmonic approximation, as implemented within MOLEKEL was used to calculate the phonon spectra. Based on the optimized crystal structure, a supercell consisti ng of a large number of atoms depending on the type of unit cell was generated from the conventional cell. The interaction range is confined to the interior of the extended supe rcell and the force constants at and beyond this extended supercell can be neglected. The asymmetric atoms were displaced by +/0.02 . The dynamical matrix was obtained from the forces calculated via the Hellmann-Feynman theorem. This size of the super cell gives “exact” phonon frequencies at X, H, R Brillouin zone points in the dispersion curves. 4.4. Results and discussion Zn(BH4)2 is a potential candidate for hydroge n storage but very little is known about the thermodynamics of this material We expect that upon heating, Zn(BH4)2 releases hydrogen according to reaction (4-1) during decomposition. We calculated the enthalpy of reaction consider ing Equation (4-1) and based on the total energy calculation by combined DFT and direct method lattice dyn amics using the following basic equation: Š = productsreactantsE E H 4-4 The total energies of Zn(BH4)2, Zn, B and H2 were calculated by DFT to evaluate the enthalpy changes for reactions given above based on the following structures: different complex models listed in Table 4-2 for Zn(BH4)2, hcp for Zn ( P63/mmc ), trigonal for B ( R-3m ) and monoclinic for B2H6 ( P21/n ). We performed di mmer calculations to determine the optimum binding energy for H2 molecule, for which we used the constant velocity MD simulation method in VASP. In these calculations, the inter-ionic distance

PAGE 68

52 was increased by 0.01 per time step of 1 fs staring from 0.65 a nd ending at 0.88 . All the structures were optimized during the total energy calculations. The calculated ground state energies for each element are give n in Table 4-1. The following sections discuss the stable crystal structures, electr onic structure and finite temperature reaction enthalpy. 4.4.1. Crystal structure The structure of zinc borohydride with the lowest enthalpy of formation was found to be of the Mg(BH4)2 type which has an orthorhom bic structure in the space group of Pmc21 (#26). No literature studies are availa ble for the crystal structure of Zn(BH4)2 except a recent one by Nakamori et al .46. They have found that the crystal structure of Zn(BH4)2 is triclinic of space group P-1 (#2) They considered effective ionic radius of a [BH4] anion as 2.03 68 which is close to the ionic radii of Br 1.96 and I 2.20 . In DFT calculations, they considered structures of MX2 (X=Cl, Br and I; M=Mg, Zn, Hg and Cd) for Zn(BH4)2 to find the most stable crystal st ructure, as an ionic bonding exists between Zn++ cations and [BH4]anions in Zn(BH4)2. Most of the commonly known complex hydrogen storage materials which have the same formula unit as Zn(BH4)2 are given in Table 4-2. We starte d our calculations for the Zn(BH4)2 from these known complex models and included the zinc bor ohydride structure established by Nakamori et al.46 We optimized the atomic positions and cell parameters for the Zn(BH4)2 for each basis model and finally optimized the cell volume against the total energy. After that we calculated the enthalpy of formation using E quation (4-4) for each model. Results from these calculations are shown in Table 4-2. Negative enthalpy of formation is indicative of a thermodynamically stable structure.

PAGE 69

53 In our calculations, we found that the stable structures are Pmc21, I41cd, Pbca, P3, Fddd, P-1, P2/c and P-62m The highest negative enth alpy of formation (-76.91 kJ/mol of H2) for Zn(BH4)2 was found for the space group Pmc21 (i.e. Mg(BH4)2 model). Hence, we considered it to be th e most stable structure of Zn(BH4)2 at 0 K. We found that the enthalpy of formation at 0 K for the P-1 space group to be -16.73 kJ/mol of H2 without zero point energy (ZPE) corrections which suggests this space group is not the most stable structure. Table 4-2. The enthalpy of formation (in kJ/mol of H2) of Zn(BH4)2 from DFT for structures based on a similar chemical formula unit complex Model Z Space Group (Number) Type of unit cell Hform (0 K) kJ/mol H2 Ca(AlH4)2 8 69 Pbca (61) Orthorhombic -74.31 2 70 P21/c (14) Monoclinic 14.10 3 70 P-62m (189) Hexagonal -6.36 9 70 P-3 (147) Trigonal -68.83 Mg(AlH4)2 1 71 P-3m1 (164) Trigonal 5.27 Mg(BH4)2 2 46 P2/c (13) Monoclinic -16.63 2 56 Pmc21 (26) Orthorhombic -76.91 Zn(BH4)2 2 72 P-1 (2) Triclinic -16.73 Be(BH4)2 16 73 I41cd (110) Tetragonal -75.85 Ca(BH4)2 8 43 Fddd (70) Orthorhombic -62.98

PAGE 70

54 To establish finite temperature struct ural stability, we calculated the phonon spectra by the direct force constant lattice dynamics tec hnique. The phonon spectra are shown in Figure 1. a-h Pmc21, I41cd, Pbca, P-3 Fddd P-1, P2/c and P-62m space groups. We did not perform lattice dynamics calc ulations for space groups having positive enthalpy of formation without ZPE corrections as they are thermodynamically unstable. It can be seen from these dispersion curves that non-negative fre quency distribution is present only for the Pmc21 space group. This suggests that Pmc21 space group (orthorhombic structure) is the stable structure for Zn(BH4)2 at finite temperatures. Only for this stable structure can acoustic and optical phonons be s een clearly in the dispersion curve. The highest soft phonon mode for space group P-1 was found to be i 3.75 THz which suggests that the crystal for this phase group is uns table. This finding contradicts the results of Nakamori et al.46. They’re not incorpora ting ZPE corrections and utilization of ultra soft ps eudo potentials (versus our more accurate PAW potentials) are possible reasons for this discrepancy. The PAW potentials give better accuracy for 3d metal hydrides74. It is observed that for the struct ure with the most negative formation enthalpy, the soft phonon modes vanish. The I41cd Pbca P-3, Fddd P2/c and P-62m space groups showed the biggest soft modes of i 1.667 THz, i 3.415 THz, i 8.75 THz, i 36.063 THz, i 20.25 THz and i 12.25 THz respectively, along different high symmetry directions.

PAGE 71

55 ( a ) ( b ) ( c ) ( d ) ( e ) ( f ) G X H G R H -1 0 1 2 3 4 5 6 Wave vectorsFrequency (THz) G X H G R H -2 -1 0 1 2 3 Wave vectorsFrequency (THz) G X H G R H -4 -3 -2 -1 0 1 2 3 Wave vectorsFrequency (THz) G X H G R H -10 -8 -6 -4 -2 0 2 Wave vectorsFrequency (THz) G X H G R H -40 -30 -20 -10 0 10 20 Wave vectorsFrequency (THz) G X H G R H -5 0 5 10 15 Wave vectorsFrequency (THz)

PAGE 72

56 ( g ) ( h ) Figure 4-1. ( a-h) Phonon dispersion relations of the Pmc21, I41cd Pbca P-3, Fddd, P-1, P2/c and P-62m space group structures of Zn(BH4)2. Coordinates of high-symmetry points are (0,0,0) G(0,0,0), X(1/2,0,0), H(1/2,0,1/2), R(1/2,1/2,1/2) Table 4-3. Optimized crystal structure of Zn(BH4)2. The space group is Pmc21 (#26) with lattice parameters a = 4.118 , b = 4.864 , c = 7.916 . All the atomic positions are given by Wyckoff letter and the number of formula units in the unit cell Z =2 Element Wyckoff letter X Y Z Zn 2 a 0.00000 0.28459 0.00890 B1 2 b 0.50000 -0.06481 0.47060 B2 2 a 0.00000 -0.46194 0.24842 H1 2 b 0.50000 -0.30735 0.42745 H2 4 c 0.27449 -0.00500 -0.42983 H3 2 a 0.00000 0.39706 0.37973 H4 2 a 0.00000 0.21066 -0.24017 H5 4 c 0.25820 -0.48360 -0.32151 H6 2 b 0.50000 0.08396 0.34880 G X H G R H -30 -20 -10 0 10 Wave vectorsFrequency (THz) G X H G R H -15 -10 -5 0 5 10 Wave vectorsFrequency (THz)

PAGE 73

57 The minimum total energy for Zn(BH4)2 is -44.595 eV/mol at a unit cell volume of 150.82 3 (considering the Pmc21 to be the stable structure at finite temperatures). The optimized lattice parameters and atomic posit ions are given in Ta ble 4-3. Figure 4-2 shows the optimized orthorhombic ( Pmc21) crystal structure. Our calculations showed nearly tetrahedral shape of the BH4 complex with B-H bond lengths dB-H = 1.20-1.25 and H-B-H bond angles H-B-H = 104.16-120o. Calculations of the phonon density of states (Figure 4-3) allowed us to determine finite temperat ure thermodynamic properties. The results are described in de tail in a subsequent section. (a ) ( b ) Figure 4-2. Orthorhombic structure of space group Pmc21 (#26) of Zn(BH4)2 ( a ) Proposed threedimensional crystal structure and ( b ) Projected structure along [010] plane. (Red: Zn, Blue: H and Green: B atom) In the phonon density of states, the phonon fr equencies are classified into three groups. From the analysis of the eigenve ctors, we found eigenmodes in the regions 25.75-39.75 THz and 65.75-76.75 THz which orig inate from internal B-H bending, and stretching vibrations in the [BH4]complex anion, respectively. These results compare

PAGE 74

58 well with the experimental solid state FTIR analysis data (33.84 THz for B-H bending and 73.66 THz for B-H stretching)48. As can be seen from Figure 4-3, the librational frequencies (<22.25 THz) that originate from the displacement of Zn++ and [BH4]-, are lower than the B-H bending and stretching modes of [BH4]-. Figure 4-3. Total density of phonon states g(w) of Zn(BH4)2 in Pmc21 symmetry. Total phonon density of states is normalized as = 1 ) ( dw w g 4.4.2. Electronic structure The local and total electronic density of states (DOS) were calculated for the finite temperature stable crystal structure of Zn(BH4)2, i.e. Pmc21, and are shown in Figures 4-4 and 4-5. From these fi gures, it can be seen that Zn(BH4)2 is an insulating material with a fundamental band gap of 3.529 eV. For Mg(AlH4)2 quasi particle GW corrections to DFT calculations give th e actual excited state band gap of 6.5 eV65, which is 2.4 eV higher than DFTcalculated band gap of 4.1 eV71. Hence, we expect our calculated band gap of 3.529 eV to also be an underestimation. From the total density of 0 20 40 60 80 0 0.02 0.04 0.06 0.08 0.1 Frequency (THz)Amplitude (1/THz)

PAGE 75

59 states plot (Figure 4-4), it can be observed that there are three main valence bands and one conduction band. From the partial DOS plot (Figure 4-5), it can be seen that the lowest valence band is dominated by B 2 s electrons and H 1 s electrons with negligible contribution from Zn. The middle part of th e valence band is mainly dominated by Zn 3 d electrons. The upper valence band is occupied by B 2 p electrons and H 1 s electrons, with a small contribution from Zn 3 d electrons. The valence band is dominated by B 2 s and H 1s electrons that make a strong B-H bond. The conduction band is primarily dominated by B 2 p electrons and to a lesser extent B 2 s and Zn 4 s electrons. Although H s electrons are prominent in the valence band small st ates are also found in the conduction band, which may indicate charge transfer to H atoms. Figure 4-4. DFT/LDA electronic band structure and total density of states (DOS) of Zn(BH4)2 relative to Fermi level K points Density of states Energy (eV)

PAGE 76

60 Zn: 4s23d1 B: 3s22p1 H: 1s1 Zn(BH 4 ) 2 Figure 4-5. DFT/LDA electronic local/total density of state (DOS) relative to Fermi level for the most stable space group Pmc21 (#26) is orthorhombic structure (black line for s blue line for p and red line for d orbital) The overall features of the DOS and PD OS, are different fr om those of LiBH4 75, NaBH4 68, KBH4 68 and Mg(BH4)2 46. The main difference is that the valence band has three parts for Zn(BH4)2 whereas the above mentioned bor ohydrides have two major parts in the valence band. This ma y be due to the presence of d electrons from the Zn atom. -10 -5 0 5 10 0 5 10 15 20 25 30 35 Density of StatesEnergy (eV) -10 -5 0 5 10 0 0.5 1 1.5 2 2.5 3 Density of StatesEnergy (eV) -10 -5 0 5 10 0 0.5 1 1.5 Density of StatesEnergy (eV) -10 -5 0 5 10 0 5 10 15 20 25 30 35 40 45 Density of StatesEnergy (eV)

PAGE 77

61 However further investigations are requi red before providing more specific interpretations. The Electron Localization Function (ELF)76, 77 and charge density for the Zn(BH4)2 (100) plane were calculated for the space group Pmc21 to study the bonding between the atoms. ELF is associated wi th the probability density of finding two electrons in the same spin-state close to e ach other. It is a position-dependent function which varies between 0 and 1. An ELF value of 0.5 means a gas-electron like probability. The ELF calculated here (Figur e 4-6) shows strong and large attractors around the H atoms. Hydrogen has no core attract or, thus this indicates either a sharedelectron or a closed-shell bond. The very low ELF valu e on the B sites indicates delocalized electrons, while a spherical shell attractor is seen around Zn. The very low ELF value between Zn and [BH4]complex indicates the low ionic bond. Whereas, a very high value of ELF within the [BH4]complex indicates i onocovalent bonding. The plotted charge density (Figur e 4-7) shows large densities centered around Zn which is segregated from B and H atoms. Some de nsity is found around the H atom; however, it is significantly lower than around the Zn atom. Almost no charge density is seen around B. This also suggests that electrons are tran sferred to the H atom mo stly from the B atom resulting in a strong B-H bond in the [BH4]complex with less polar bonding between the Zn and H atoms.

PAGE 78

62 Figure 4-6. Electron localization function (ELF) of Zn(BH4)2 (100 plane) (Red: Zn, Blue: H and Green: B) Figure 4-7. Charge density of Zn(BH4)2 (100 plane)

PAGE 79

63 4.4.3. Finite Temperature Reaction Enthalpy In the thermodynamic expression for the finite temperature enthalpy PV T E T H+ =) ( ) ( 4-5 E(T) is the total internal energy which in cludes not only energies from DFT and ZPE (at 0 K) but also contributions from rotational, translational and vibrat ional (excluding ZPE) energies given in Equation (4-5 a ). We assumed that the rotational and translational energies are important for the gaseous phase only. ) ( ) ( ) ( ) 0 ( ) 0 ( T E T E T E K E K E Etran rot vib ZPE DFT+ + + + = 4-5 a The enthalpy of formation is now given by E quation (4-6), as a sum of the absolute zero temperature part and the finite temperat ure part. The first part (Equation (4-6 a )) includes the ground state energy and also the zero point vibrational energy which is independent of temperature. On the other hand, th e finite temperature part (Equation (4-6 b )) includes all the energies which are de pendent on temperature, and is calculated using statistical mechanical theory. ) ( ) 0 ( ) (0T H K H T Hform + = 4-6 Where, ) 0 ( ) 0 ( ) 0 (0K E K E K HZPE DFT + = 4-6 a Š = productsts reac DFT DFT DFTK E K E K Etan) 0 ( ) 0 ( ) 0 ( 4-6 a 1 Š = productsts reac ZPE ZPE ZPEK E K E K Etan) 0 ( ) 0 ( ) 0 ( 4-6 a 2 and ) ( ) ( ) ( ) ( ) ( PV T E T E T E T Htran g rot g vib + + + = 4-6 b

PAGE 80

64 Š = productsts reac vib vib vibT E T E T Etan) ( ) ( ) ( 4-6 b 1 Š = productsts reac g rot g rot rot gT E T E T Etan) ( ) ( ) ( 4-6 b 2 Š = productsts reac g tran g tran tran gT E T E T Etan) ( ) ( ) ( 4-6 b 3 For calculating) ( T H as an example for the reaction gi ven in Equation (4-1), we write (Equations ((4-7)-(4-9)): } ) ( ) ( ) ( { 4 ) ( 2 ) ( ) ( ) (2 2 2 2 4) (PV T E T E T E T E T E T E T HH rot H tran H vib B vib Zn vib BH Zn vib+ + + Š Š Š = 4-7 } 2 3 ) ( { 4 ) ( 2 ) ( ) ( ) (2 2 4) (T k T k T k T E T E T E T E T HB B B H vib B vib Zn vib BH Zn vib+ + + Š Š Š = 4-8 } 2 7 ) ( { 4 ) ( 2 ) ( ) ( ) (2 2 4) (T k T E T E T E T E T HB H vib B vib Zn vib BH Zn vib+ Š Š Š = 4-9 Equation (9) shows that we need to calcula te the vibrational energy of the complex material Zn(BH4)2 and its elemental components to calculate the finite temperature reaction enthalpy. Vibrationa l energies in the above E quation 9 were calculated by integrating the phonon density of states over the Brillion zone. The results for the EZPE and Evib (without ZPE) for different elemen ts/compounds are given in Table 4-4.

PAGE 81

65 Table 4-4. Zero point and vi brational energies of Zn(BH4)2 and its primary elements Element/ Compd. EZPE (kJ/f.u.) (literature) EZPE(kJ/f.u.) (calculated) Evib(300 K) (kJ/f.u.) (incl. ZPE) Evib(300 K) (kJ/f.u.) (actual) Zn 78 1.0 (approx) 1.077 7.176 6.099 -B 79 12.5 12.478 13.65 1.172 H2 65 28.29, 80 27.08, 79 25.7 28.12 28.12 0.0 B2H6 80 170.96 163.367 173.473 10.106 Zn(BH4)2 206.539 225.404 18.811 The enthalpy of formation of Zn(BH4)2 was found to be -76.91 kJ/mol of H2 at 0 K (using Equation (4-6 a 1)) without the zero point energy corrections and -59.90 kJ/mol of H2 including ZPE (using Equations ((4-6 a ) & (4-6 a 2))). We also found that the finite temperature reaction enthalpy for the dehydr ogenation reaction (4-1) by using Equations (4-6) and (4-9) was 66.003 kJ/mol H2 at 300 K or 264.012 kJ/mol of Zn(BH4)2. The ab initio theoretical formation enthalpy value of ZnH2 (reaction (4-3)) from zinc and molecular hydrogen varies from 6.464 to 31.57 kJ/mol ZnH2 81. As very little is known about ZnH2, we calculated the formation enthalpy of the ZnH2 using fluorite type structure and it was found to be 23.66 kJ/mol ZnH2 at 0 K in our calculations for the dehydrogenation reactions ((4-1 b )-(4-1 d )). For reaction (4-2), we calculated the formation enthalpy of diborane. It was found to be 25.38 kJ/mol of B2H6 at 0 K including zero point energy wher eas the experimental standard enthalpy of formation is 36.00 kJ/mol of B2H6 82. We defined the standard enthalpy of reaction, Hrxn for the reaction (4-1 a ) as given below:

PAGE 82

66 ) ) ( ( ) (2 4 0 6 2 0BH Zn H H B H Hform form rxn Š = 4-10 Similar expressions were used for other reactions. Using the standard formation enthalpies of Zn(BH4)2, B2H6 and ZnH2 we found Hrxn for the reactions ((4-1 a )-(4-1 d )) to be 300.012, 287.672, 284.440 and 323.672 kJ/mol of Zn(BH4)2, respectively. Although H2 is released from the liquid phase, we ha ve not included the heat of fusion for the dehydrogenation reaction enthalpy. So we can expect the enthalpy of formation values calculated to be overe stimated. As the enthalpy of dehydrogenation of Zn(BH4)2 for the reaction (4-1) is the lowest, this re action may be expected to start first upon heating the complex borohydride. So this reac tion can be predicted as the most favorable decomposition reaction. However, Jeon et al.48 have found some evolution diborane experimentally, which indicates that reacti on (1a) (36.00 kJ/mol higher enthalpy of dehydrogenation than reaction (4-1)) also oc curs at the same time. The exact quantification and identific ation of the decomposed gaseous components at the decomposition temperature of Zn(BH4)2 is beyond the scope of this work. The main objective of the work in this section was to find the thermodynamically most favorable decomposition reaction. 4.5. Summary In this work, we have predicted from th eory the stability of the complex zinc borohydride structure. This required invest igation of both the ground state energy by DFT and the lattice vibration energy by di rect method lattice dynamics. We have calculated the structural stability of Zn(BH4)2 at T > 0 from structure-optimization calculations starting with 10 different very similar formula units for Zn(BH4)2. These calculations establish Pmc21 symmetry to be a new thermodynamically most stable

PAGE 83

67 structure of Zn(BH4)2 at finite temperature. Our estimated value for the formation enthalpy of Zn(BH4)2 is -66.003 kJ/mol H2 at 300 K which suggest s that it should be possible to synthesize this phase. This finding suggests that the crystal structure is stable at room temperat ure and thus Zn(BH4)2 could be considered a potential candidate for hydrogen storage. The other thermodynamically st able structures in Table 4-2 which are I41cd, Pbca, P-3 and Fddd symmetries, turn out to be unsta ble at finite temperatures with respect to lattice vibrations. From electron ic structure calculati ons, ionic interaction between Zn and [BH4]and the strong ionocovalent B–H interaction within the [BH4]tetrahedral are revealed. Electr onic density of stat es studies reveal that this phase has wide-band-gap of 3.529 eV which suggests that the Zn(BH4)2 is an insulator. Standard reaction enthalpy calculations reveal that d ecomposition to primary elements is the most favorable one. The findings of our work are in qualitative agreement with experimental results. Our study has given new insights in to the thermodynamically most stable phase for the complex zinc borohydride. The methodology outlined in this work can be extended to studies on other comp lex hydrogen storage materials.

PAGE 84

68 Chapter 5 First-Principles Study Of Ni-Induced Zn(BH4)2 5.1. Abstract Fundamental understanding of the role of Ni additives in promoting the dehydrogenation mechanism of hydrogen de -sorption in zinc borohydride (Zn(BH4)2) is a key factor for using this mate rial in hydrogen storage. A sy stematic theoretical study of the energetics and hydrogen dynamics was carried out to understand this dehydrogenation mechanism. The energetic calcul ations reveal that Ni substitutes Zn in preference to B. H atoms are pulled towa rds these doped Ni atoms, which introduces instability via the breaking of multiple B-H bonds in the complex borohydride. The mechanistic understanding gained from this st udy can be applied to the design of better hydrogen storage materials. 5.2. Introduction Hydrogen is considered to be the cleane st amongst the fuels available for onboard vehicular applications. The high operational cost of using liquid hydrogen and the low explosive limit (4%) of hydrogen in the gas pha se makes its storage an important aspect for potential vehicular applica tions. The technical challenge is to find materials that exhibit favorable thermodynamics and kinetic s for hydrogen de-sorption and absorption, and have sufficient gravimetric and volumetric storage capacity. Zn(BH4)2 is considered to be a potential candidate for on-board app lications as it has very high theoretical

PAGE 85

69 hydrogen storage capacity of 8.4 wt% a nd also reasonably low decomposition temperature compared to complex alkali borohydrides47. Improvement of the hydrogen aband de-sorption kinetics in complex hydrides is essential for these materials to be good reversible hydrogen storage media in the transportation sector. It has been reported that the addition of titanium-based compounds improves the kinetics significantly for NaAlH4 52, 83, 84. The role of 3-d transition metals in improving both dehydrogenation kinetics an d thermodynamics of crystalline NaAlH4 has been clearly explained by density functional theory (DFT) calculations83, 85. Calculations reveal that dope d, destabilized metal hydride reaction systems are unstable with respect to phase separation at 0 K86. However, experimental results87, 88 suggest that only Ni doping in Zn(BH4)2 reduces the de-hydrogenation temperature by approximately 20 C, although physical unde rstanding remains elusive. Thus, theoretical studies at the atomic scale are necessary to understand the role of Ni in lowering the hydrogen de-sorption temperature. Such understanding will allow for selection of appropriate dopants for complex metal hydride systems in general, achieving superior hydrogen st orage properties. 5.3. Methodology In this chapter, we provide results of a systematic study of the role of Ni in destabilizing the crystalline structure of Zn(BH4)2. For this, we calculate the vibrational density of states of hydroge n within the tetrahedral BH4 complex nearest to the Ni dopant to show that it diminishes the hydroge n binding energy of the crystalline Zn(BH4)2 resulting in a decrease of the hydrogen decomp osition temperature. The energy cost to remove hydrogen, and the hydrogen dynamics nearest to Ni (doped) and the

PAGE 86

70 corresponding Zn (undoped) atom are calc ulated to understand the dehydrogenation mechanism in the crystalline material. We al so show that Ni weakens the multiple B-H bonds, thus allowing the hydrogen to release at lower temperature. We have utilized ab initio DFT methods to calculate th e energetics and the direct force lattice dynamics approach to model gasmetal interactions at the atomic scale to gain useful insights into the effect of additives on Zn(BH4)2 structural behavior. Energetic calculations were performed usi ng DFT within the local density approximation (LDA)59 and projected augmented wave (PAW)25, 60 method utilizing a plane wave basis set as implemented in the Vienna Ab initio Simulation Package (VASP)18, 26, 27. The atomic positions and the lattice parameters, including the unit cell volume, were optimized by minimizing the forces and stresses until the residual forces between the atoms were within 0.01 eV/. We calculated vibrational density of stat es within the harmonic approximation. Combination of DFT with different tec hniques such as li near response method28, 29 or direct methods30-32 allows evaluation of the phonon disp ersion curves without empirical parameters. Parlinski et al .33, 34 have developed the direct method where the forces are calculated via the Hellmann-Feynman theorem us ing DFT-derived total energies, assuming a finite range of interaction. In this work, the PHONON code35 was utilized to calculate the phonon spectra. Based on the op timized crystal structure, a supercell consisting of a large number of atoms depe nding on the type of unit cell was generated from the conventi onal unit cell. We calculated the cohesive energy of Ni-doped Zn(BH4)2 ( i.e. mixed complex hydrides Zn16nNinB32H128 and Zn16B32-2 nNi2 nH128, ( n =1–4)) by considering the 2x2x2

PAGE 87

71 supercell containing 16 Zn(BH4)2 formula units (176 atoms), a nd 1-4 of the units [a unit = Zn(BH4)2] Zn or B atoms were substituted with Ni atoms. The energy cost for hydrogen removal was computed by the Zn16B32H127, (Zn15Ni)B32H127 and Zn16(B30Ni2)H127 supercells. 5.4. Results and discussions Using this methodology, we first investigated the possibili ty of Ni dissolution into zinc borohydride from calculations of energetics and cohesive energy, i.e. the energy required to break the atoms of the solid into isolated atomic species. Positive cohesive energy indicates a thermodynamically favorable structure. Results are shown in Figure 5-1. We also calculated the hydr ogen atom removal energies for Zn16B32H128, (Zn15Ni)B32H128 and Zn16(B30Ni2)H128 supercells which were found to be 0.930 eV, 0.806 eV and 0.488 eV, respectively. Figure 5-1. Cohesive energy of pure and Ni-substituted Zn(BH4)2. The dotted line is for pure Zn(BH4)2, the dashed line is for the case where Ni subs titutes Zn and the solid line is for the case where Ni substitutes B 5 10 15 20 25 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 Mole fraction of Ni dopant, x%Cohesive energy eV/mol of H2

PAGE 88

72 We found that substitution by Ni doping is energetically possi ble for both cases but pure borohydride is more stab le than the doped systems. The results also suggested that the preferred substitution site for Ni is the Zn rather than the B site in the Zn(BH4)2 crystal lattice. This may be explained by the si ze effect: as Ni atomic radius (1.24 ) is closer to the Zn atomic radius (1.34 ) wh ereas the Ni atom is much bigger than the B atom (atomic radius 0.85 ). Hence, energetically, it prefers the Zn site. This size mismatch is also the cause for Ni substitution of B to be energetically less favorable. A relaxed crystal bonding anal ysis showed (Figure 5-2) that the bond lengths between Ni-H are 1.80 , 1.69 , 1.62 and 1.62 for the doped system. The corresponding Z-H bond distances are 2.0 , 1.86 , 1.84 and 1.84 for the pure system. B-H bond lengths for the pure system are 1.22 , 1.22 , 1.24 and 1.25 . The corresponding substitutional Ni-doped complex borohydride system, B-H bond distances are 1.25 , 1.25 , 1.29 and 1.32 . These results suggested that B-H bond lengths increase for substitutional Ni doped Zn(BH4)2 compared to those in pure Zn(BH4)2. The Ni atom pulls the H atoms toward s itself from the nearest B atom and helps dehydrogenation by dest abilizing the complex Zn(BH4)2. These conclusions are also supported by calculations of the energy cost for hydrogen removal. We found that the hydrogen atom removal energy for Zn16B32H128 supercell is higher than that for the (Zn15Ni)B32H128 supercell by 0.124 eV, confirming that Ni doping does indeed weaken the B-H bonds.

PAGE 89

73 Figure 5-2. Relaxed structure of Ni doped Zn(BH4)2. Red ball corresponding to hydrogen, blue ball corresponding to zinc, green ball correspondi ng to boron and purple ball corresponding to nickel, respectively Vibrational dynamical propertie s of the H atom in the BH4 complex nearest to the Ni dopant could be used as local probes to understand the destabilization mechanism. We calculated these properties for the Ni substitution of Zn model as it is energetically favorable than substitu tion of B. We calculated the vibr ational density of states using smaller supercell (88 atoms) structures than the supecells (176 atoms) used for energetic calculations. First, a (Zn3Ni)B8H32 unit cell was generated and then a 1x2x1 supercell (88 atoms) was utilized in ca lculating the vibrational density of states for this Ni-doped system. This high dopant fraction (25%) un it cell/supercell was utilized to achieve a computationally viable system, and to stay within the limitations of the PHONON code, which allows a maximum of 62 non-e quivalent displasive atoms.

PAGE 90

74 From the dynamic studies, we find that fo r Ni substitution of Zn affects both low (libration) and high frequency (mainly be nding) modes of vibrations of the BH4 complex, as seen from the phonon density of states plots in Figure 5-3. Figure 5-3. The total and partial phonon DOS for Zn8B16H64 (panels a d ) and Zn6Ni2B16H64 (panels e h ). The specific atoms/groups for which the vibrational phonon densities of state are plotted are given within the figures The small amplitude negative frequencie s observed in the to tal and partial phonon DOS of Zn6Ni2B16H64 are indications of crystal instabi lity at finite temperature as all phonon frequencies are required to be positive definite: 2 > 0 for a stable crystalline structure. The modes that soften are ma inly from both hydrogen and B displacements -20 0 20 40 60 80 Frequency (THz)Amplitude -20 0 20 40 60 80 Frequency (THz)Amplitude -20 0 20 40 60 80 Frequency (THz)Amplitude -20 0 20 40 60 80 Frequency (THz)Amplitude -20 0 20 40 60 80 Frequency (THz)Amplitude -20 0 20 40 60 80 Frequency (THz)Amplitude -20 0 20 40 60 80 Frequency (THz)Amplitude -20 0 20 40 60 80 Frequency (THz)Amplitude ( a ) Total ( b ). BH4 nearest to similar Zn atom ( c ). B ( d ). H ( e ). Total ( f ). BH 4 nearest to Ni atom ( g ). B ( h ). H

PAGE 91

75 nearest to the Ni dopant in th e complex. These soften all modes and introduce instability in the lattice, thus improvi ng dehydrogenation reactions. 5.5. Summary To summarize this work, we have studied the lattice destabilization mechanism in the Ni-substituted zinc borohydride lattice structure using DFT and lattice dynamics techniques. The ground state energy calcula tions from DFT together with the phonon spectra calculated by the direct method la ttice dynamics have provided a mechanistic understanding of the dehydrogenation in this doped system, and provide insights into how dopants affect decomposition of this complex borohydride. Cohesive energy calculations indicate that s ubstitution of the Zn atoms by Ni produces a more stable structure than that resulting from substitution of B atoms. This is expected from the comparable sizes of the Ni and Zn atomic radii, and the significantly smaller atomic radius of the B atom. Lattice dynamics indicate that Ni affects both low (librational) and high (bending) frequency modes of the BH4 complex. The modes that soften are mainly from both hydrogen and B displacements nearest to the Ni dopant in th e complex. These soften all modes and introduce instability in the lattice, thus improving dehydrogena tion reactions. From the crystal bonding analysis, we see that bond lengths between Ni -H decrease and B-H bond lengths increase for the doped complex, compared to the pure system. Also, it indica tes that Ni has an affinity towards H atom and helps the de hydrogenation mechanism by destabilizing the complex hydride. This result is also fu rther verified by the calculated hydrogen atom removal energy from the pure and doped system s. We have found that 0.124 eV/ H-atom more energy is required for the former than for the latter case. Th ese theoretical results

PAGE 92

76 are in qualitative agreement with experimental findings8, 9. The methodology outlined in this work is applicable to the study of other complex hydrogen storage materials.

PAGE 93

77 Chapter 6 First-Principles Investigation Of The Mn(BH4)2 6.1. Abstract Manganese borohydride (Mn(BH4)2) is considered to be a high-capacity ( 10 wt%) solid-state hydrogen storage candidate, but so far has not been shown to exhibit reversible hydrogenation. This study presen ts the calculated cr ystal structure and electronic structure of Mn(BH4)2 from density-functional theory (DFT) within the generalized gradient approxi mation (GGA) and thermodynamic properties from the direct method lattice dynamics. The calcula tions of phonon spectra of Mn(BH4)2 suggest that I4m2 symmetry is a stable phase at finite temperature. The formation energy of Mn(BH4)2 for the I-4m2 symmetry is -28.93 kJ/f.u. at 0 K including zero point energy (ZPE) corrections and standard state enthalpy of formation is predic ted to be -58.89 kJ/f.u. The most feasible dehydrogenation reaction is found to be Mn(BH4)2 = Mn + 2B + 4H2, which is an endothermic reaction at decomposition temperature. The spin polarized electronic density of states shows that manga nese borohydride is a half-metallic nature due to the presence of half filled 3 d electrons from Mn. The electronic st ructure analysis shows that the interaction between Mn atoms and BH4 complexes has an ionic character while the internal bonding of BH4 is essentially covalent.

PAGE 94

78 6.2. Introduction Light weight, low cost, highly reversible hydrogen storage systems are essential for realizing low temperature PEM fuel cell powered vehicles89. The breakthrough discovery of Ticatalyzed NaAlH4 90, 91 for the reversible onboa rd hydrogen storage may not be the ideal system to attain the DOE 2010 and FreedomCA R technical targets. This is because for NaAlH4, the usable hydrogen storage cap acity achievable is 5.4 wt. %, which is well below the set DOE goals of 6.0 wt. % by 2010. Hence, borohydride complexes as hydrogen storage materials have recently attracted gr eat interest. The borohydride complexes NaBH4 and LiBH4 possess high hydrogen storage capacity of 13.0 wt. % and 19.6 wt. %, respectively. However, the release of hydrogen from NaBH4 is possible only by hydrolysis (reaction with H2O), and this process is irreversible. For the case of LiBH4, the catalytic addition of SiO2, significantly enhances its thermal desorption rate at 200 C44. In general, the thermal decomposition and/or recombination of hydrogen either from NaBH4 or LiBH4 are difficult to achieve because of the thermodynamic stability due to strong B-H interactions45, 92. A systematic approach to study the phase stability of LiBH4 based on ab initio calculations has been presented and four thermodynamically stable phases have be en identified, includ ing a new phase of Cc symmetry for the first time for LiBH4 45. A correlation has been reported between thermodynamic stability of metal borohydrides and Pauli electronegativity of the parent components using first principles calculations46. Most of the commonly known borohydrid es are found to be unsuitable for onboard hydrogen storage applications. This is primarily attributable to the high stability resulting from very high decomposition temper atures or complete irreversibility of

PAGE 95

79 hydrogen desorption47. Theoretically, it was shown that Zn(BH4)2 can be stable at finite temperature and also Ni doping improves the dehydrogenation mechanism of this complex borohydride93, 94 This allows consideration fo r on-board applications as Zn(BH4)2 has a very high theoretical hydrogen stor age capacity of 8.5 wt%. However, experimentally it was found that thermal decomposition of Zn(BH4)2 comprises not only of the evolution of H2, but also production of an appr eciable amount of B–H (borane) compounds87. Lowering the decomposition temp erature by Ni doping may lead to negligible release of boranes88. Mn(BH4)2, among all other transition complex borohydrides, is stable at room tempearature47. Although Mn(BH4)2 is a potential candidate for hydr ogen storage, no literature is available about its structural and thermodynamic properties. In this study, we performed a combination of ab initio DFT and lattice dynamics calculations93 to establish the stable crystal structure and to determine the fin ite temperature thermodynamic properties of Mn(BH4)2. The enthalpy of formation is an import ant predictor of deco mposition temperature for the decomposition reaction with its asso ciated thermodynamic parameters. For a simple metal hydride, when the material is being heated, the eff ect of entropy starts dominating the enthalpy, and the Gibbs ener gy at the decompos ition temperature and pressure is zero. During heating, the entr opy change of solid materials is very small compared to the gas phase, which allows us to consider that the entropy change during the decomposition reaction to be primarily due to H2 evolution. Formation enthalpy and Gibbs energy are the best predictors of the stability of the theoretically predicted phases and such data may serve as guide for possibl e decomposition or synthe sis routes. So, our

PAGE 96

80 calculations were also extended to find the theoretically possible dehydrogenation reaction of this complex hydride based on the Gibbs energy calculations. The most favorable dehydrogenation reaction will have the lowest value of the Gibbs energy. In this study, we have considered the followi ng different possible reactions (Equations. ((61 a )-(6-1 e ))) for Mn(BH4)2 and estimated the associated Gibbs energy of dehydrogenation values: Mn(BH4)2 = Mn + 2B + 4H2 6-1 a Mn(BH4)2 = MnH2 + 2B + 3H2 6-1 b Mn(BH4)2 = MnB2 + 4H2 6-1 c Mn(BH4)2 = MnH2 + B2H6 6-1d Mn(BH4)2 = Mn + B2H6 + H2 6-1 e 6.3. Simulation methods Density functional theory (DFT) is at pr esent considered to be a versatile and important tool to so lve innovative research problems in metal-hydrogen interactions and establish associated mechanisms. Successful application of DFT to a materials problem involves three distinct steps; (i) translation of th e engineering problem to a computable atomistic model, (ii) comput ation of the required physicochemical properties, and (iii) validation of the simulation results by co mparison with laboratory experiments49. Lattice dynamics50, 51, using a direct force constant method, helps one to complete the picture by allowing for calculation of the finite temper ature thermodynamic properties. Crystal structure stability and electronic structures for complex hydrides55, 56, 93 have been reported on the basis of DFT calculations.

PAGE 97

81 6.3.1. Ab initio methods First-principle calculations were performed on Mn(BH4)2 using DFT within the generalized gradient approximation ( GGA) with the Perdew-Wang 91 (PW91) functional59 and projected augmented wave (PAW)25, 60 method utilizing a plane wave basis set to calculate the total energies, as implemented in the Vienna Ab initio Simulation Package (VASP)18, 26, 27. A 5 5 5 Monkhorst-Pack61 k-point mesh was used for sampling the Brillouin zone. A kinetic energy cutoff of 400 eV was used for Mn(BH4)2, MnH2, B2H6 and MnB2 for the plane wave basis set. For Mn, B and H2, the same kinetic energy cutoff values as for Mn(BH4)2 were utilized, in the self-consistent total energy calculations. This allows fo r the maximum cancellation of errors in the reaction enthalpy evaluations, which are calcu lated as small energy differences between reactants and products. The criterion for se lf-consistency in the electronic structure determination was that two c onsecutive total energies diffe red by less than 0.001 meV. The atomic positions and the lattice parame ters, including the unit cell volume, were optimized by minimizing the forces and stresses until the residual forces between the atoms were within 1 meV/. 6.3.2. Direct method lattice dynamics DFT is a non-empirical-parameter met hod whose applications and predictive ability in different fields ar e known for some time. Combination of DFT with different techniques such as linear response method28, 29 or direct methods30-32 allows us to evaluate phonon dispersion cu rves without empirical parameters. Parlinski et al .33, 34 developed the direct method wh ere the forces are calculated via the Hellmann-Feynman theorem using DFT-derived total energies, as suming a finite range of interaction. The

PAGE 98

82 phonon spectra are then derived using Newt on’s equation of motion in the lattice dynamics calculations. Calculations of the phonon density of st ates and the resulting vibrational contributions to the free energy were pe rformed using the PHONON code developed by Parlinski35. These calculations used the same exchange–correlation functional as our total energy calculations and an energy cutoff of 400 eV. Force cons tants were computed using displacements of asymmetr ic individual atoms by +/0.02 and the absolute force constant cutoff value utilized was 0.01 a.m.u. (THz)2. Based on the optimized crystal structure, a supercell consisti ng of a large number of atoms depending on the type of unit cell was generated from the conventional cell. It is necessary to use a relatively large supercell to avoid interacti on between images of the disp laced atom when defining the phonon density of states. 6.4. Results and discussions We calculated the Gibbs energy of reactions considering Equations ((6-1 a )–(61 e )) and based on the total energy calculation by combined DFT and direct method lattice dynamics. The total energy of Mn(BH4)2 was calculated based on the stable structure found based on positive definite frequency distribution criteria over the Brillion zone. For MnH2 and MnB2, the total energy was calculated based on fcc fluorite ( Fm-3m )95 and hcp AlB2 type ( P6/mmm )96 structures, respectively. The total energy of B was calculated based on the hexagonal ( R-3m ) structure93. We calculated the optimum binding energy for H2 molecule93, for which we used the constant velocity MD simulation method in VASP. Mn is known to be the most comp lex element in the periodic table97. Mn has 4-5 phases at different temperatures and also different type of magnetic properties at different

PAGE 99

83 conditions. From our own experime ntal results we have found Mn(BH4)2 to decompose at temperatures lower than 500 K98. Hence, in this work, Mn was considered to be in the phase (valid till 900 K). It is known that -Mn phase is paramagnetic with a Neel temperature of 95 K (bcc st ructure in the space group of I-43m ) below which it is antiferromagnetic (body centered tetragonal structure in the space group of I-42m )97. So, in our calculations, we considered -Mn as anti-ferromagnetic till 95 K and above that as a paramagnetic phase. The total energy was f ound to be -8.964 eV/f.u. which is well matched with the DFT calculated literature value of 8.97 eV/f.u.97. The following sections discuss the stable crystal structures electronic structure and finite temperature reaction enthalpy and Gibbs energy. 6.4.1. Crystal structure The structure of Mn(BH4)2 was found to be tetragonal type in the space group of I-4m2 (# 119). We started our calculations for Mn(BH4)2 from the 21 known similar complex model structures presented in Table 6-1 Using this methodology, we first calculated the cohesive energy, i.e. the energy required to br eak the atoms of the solid into isolated atomic species of Mn(BH4)2 based on these models. Then we calculated the enthalpy of formation of these structures as shown in the Table 6-1. Only five space groups showed negative enthalpy of form ation indicate possible thermodynamically favorable structures. To establish finite temp erature structural stab ility, we calculated the phonon spectra of those structures only by the direct force constant lattice dynamics technique. It was found that only I-4m2 space group shows positive phonon density of states (DOS). This suggests that I-4m2 (tetragonal) space group is the only stable phase for Mn(BH4)2 at finite temperatures.

PAGE 100

84 Table 6-1. The model structures (XY2Z8 type) used as input for the structural optimizations. The cohesive energy and enthalpy of formation (excluding ZPE) of MnB2H8 are calculated based on these model structures. Z is the number of formula units in the unit cell Model Space group no. ( Z) Cohesive energy, eV/atom Formation enthalpy, kJ/f.u. 65, 71, 99 102 MgAl2H8 1 (2) 3.0488 90.918 164 (1) 3.1739 -41.580 103 MgAl2Cl8 15 (4) 2.5658 602.257 46, 93 ZnB2H8 2 (2) 2.4123 764.771 26 (2) 3.0868 50.705 46, 104 MgB2H8 13 (2) 2.5309 639.185 119 (4) 3.2210 -91.428 69 CaAl2H8 14 (2) 2.5625 605.708 61 (8) 3.1206 14.859 147 (9) 3.1041 32.356 189 (3) 2.9232 233.814 43 CaB2H8 70 (8) 3.1765 -44.272 73 BeB2H8 110 (16) 3.1178 17.849 105, 106 ZrW2O8 19 (12) 3.1119 24.042 198 (4) 2.9679 176.554 213 (8) 1.8181 1393.734 106 ZrMo2O8 163 (6) 3.1412 -6.922 107 TiAl2Cl8 14 (2) 2.1058 1089.184 72 CuAl2Cl8 2 (2) 2.4802 692.879 108 CdAl2Cl8 7 (2) 3.057806 81.361 109 MRe2O8 (M = Mn, Co, Ni, Zn) 147 (1) 3.14069 -6.384

PAGE 101

85 In the phonon density of st ates of the stable space group (Figure 6-1), the phonon frequencies are classifi ed into three groups. From the analysis of the eigenvectors, we found eigenmodes originate from internal B-H bending in the re gion of 30.75-39.75 THz and the stretching vibrations in the [BH4]complex anion were found to be in the region of 67.25-71.75 THz for I-4m2 space group, respectively. In the 2000 2500 cm-1 (60 75 THz) region, strong bands ar e observed in the spectra of borohydride complexes that correspond to the stretching vibr ations of the B-H bonds. To compare these theoretical results with experimental data, we prepared the new transition metal based complex borohydride Mn(BH4)2 from the following stoichiometric reaction by the mechanochemical synthesis route and measur ed the solid state FTIR spectrum: 2NaBH4 + MnCl2 Mn(BH4)2 + 2NaCl Mn(BH4)2 formation was confirmed from the solid state FTIR spectrum (shown in (Figure 6-1) which indicates that the B-H stretching at 2223.9 – 2371.6 cm-1 (66.70 71.17 THz) and B-H bending at 1125.7-1213.2 cm-1 (33.78 36.40 THz) are compared well with theoretical results. It also can be seen from Figure 6-1, the librational frequencies (in regions <17.75 THz) that originate from the displacement of the heavy atom Mn and [BH4]are lower than those of the BH bending and stretching modes of [BH4]-.

PAGE 102

86 Figure 6-1. Total density of phonon states g(w) of Mn(BH4)2 in I-4m2 symmetry. Total phonon density of states is normalized as = 1 ) ( dw w g. FTIR spectrum of Mn(BH4)2 (along with those of precursors and products) prepared by ball milling of 2NaBH4 + MnCl2. The black lines represent phonon density of states of Mn(BH4)2 and blue lines represent the FTIR spectra of Mn(BH4)2, NaBH4, NaCl and MnCl2, respectively The minimum total energy and primitive cell volume for Mn(BH4)2 is -50 eV/f.u. and 333.035 3 for I-4m2 structure, respectively. The optimized lattice parameters and atomic positions are given in Table 6-2. Figure 6-2 shows the optimized tetragonal ( I4m2 ) crystal structure. The crystal structure of Mn(BH4)2 is consist of metal cation Mn2+ and nearly ideal tetrahedral [BH4]-. For t he I4 m 2 structure, it is characterized by a fourfold coordination of [BH4]ions around the Mn2+ ions. Furthermore, each Mn2+ has 8 nearest neighbor H atoms w ith the shortest Mn-H bond distance around 2.06 . The B-B nearest neighbor distance is found to be 3.68 and Mn-B distance 2.41 . The B-H bond lengths and the H-B-H bond angles within the [BH4]tetrahedral of I-4m2

PAGE 103

87 symmetry are found to be dB-H = 1.23 1.24 and H-B-H = 106 – 115, respectively. The geometry of the BH4 tetrahedron is similar to that found in isolated [BH4]ions, an ideal tetrahedral structure with a constant bond le ngth of 1.24 for B-H bonds, and tetrahedral angles of 109.5, where the B and H atoms are covalently bonded110. Table 6-2. Optimized crystal structure of Mn(BH4)2. The number of formula units in the unit cell is Z. All the atomic positions are given by Wyckoff letters Space group Element Wyckoff letter X Y Z I-4m2 (Z = 4) a = b = 8.182 and c = 9.951 Mn1 2 c 0.0000 0.5000 0.2500 Mn2 2 a 0.0000 0.0000 0.0000 B 8 i 0.2311 0.00000 0.8494 H1 8 i 0.2467 0.0000 0.7266 H2 8 i 0.6392 0.0000 0.9137 H3 16 j 0.1590 0.1270 0.8786 Figure 6-2. Three-dimensional crystal structures of Mn(BH4)2 of space group I-4m2 (#119) (Black (large), blue (middle) and green (small) sphe res represent Mn, B and H atoms, respectively)

PAGE 104

88 6.4.2. Electronic structure The local and total spin polarized elect ronic density of states (DOS) were calculated for the stable crystal structure of Mn(BH4)2 (shown in Figure 6-3). According to spin polarized total/local DOS calculations (Figure 6-3), the Mn 3 d DOS has halfmetallic character for Mn(BH4)2. The Mn 3 d DOS have a high spin-up occupation below the Fermi energy and high peak of spin-down unoccupied states in the region of 12 eV above the Fermi energy (Figure 6-3). From this point, Mn partitial 3 d DOS has metallic character for spin-down and insulating charact er for spin-up projection, are supposed to show some intermediate behavior. Other than this, there are two more valence bands and one conduction band found in the total DOS plot Both these two valence bands are well below the Fermi level. From the partial DOS pl ot (Figure 6-3), it can also be seen that the lowest valence band is dominated by B 2 s electrons and H 1 s electrons with negligible contribution from Mn. The middle part of the valence band is mainly dominated by B 3 p electrons, H 1 s electrons and smaller number of Mn 3 d electrons. The valence band well below Fermi level is dom inated by B and H electrons that make a strong B-H bond. The conduction band is primarily dominated by B 2 p electrons and to a lesser extent B 2 s and H 1 s electrons. Although H s electrons are prominent in the valence band, small number states are also found in the conduction band, which may indicate charge transfer to H atoms. The overall features of the DOS and PD OS are different fr om those of LiBH4 75, NaBH4 68, KBH4 68 Mg(BH4)2 46 and Zn(BH4)2 93. The main difference is that there is no band gap between valence and conduction bands for Mn(BH4)2, whereas, the above mentioned borohydrides have distinct, wide band gaps. This may be due to the presence

PAGE 105

89 of the half-filled d orbital of the Mn atom resulting in high peak of spin-down unoccupied states above the Fermi energy and also it s hows magnetic behavior with a total magnetic moment of 5.13 B/Mn for I-4m2 phase. However, further i nvestigations are required before providing more specific interpretations wh ich is beyond the scope of this work. ( a ) Mn: 4s23d5 ( b ) B: 2s22p1 ( c ) H: 1s1 ( d ) Mn(BH4)2 Figure 6-3. DFT/GGA electronic spin polarized lo cal/total density of state (DOS) relative to Fermi level for Mn(BH4)2. (Green line for s blue line for p and red line for d orbital) -10 -5 0 5 10 -15 -10 -5 0 5 10 15 Density of StatesEnergy (eV) -10 -5 0 5 10 -3 -2 -1 0 1 2 3 4 Density of StatesEnergy (eV) -10 -5 0 5 10 -1 -0.5 0 0.5 1 Density of StatesEnergy (eV) -10 -5 0 5 10 -50 0 50 Density of StatesEnergy (eV)

PAGE 106

90 The Electron Localization Function (ELF)76, 77 and charge density for the Mn(BH4)2 were calculated for the space group I-4m2 which is shown in Figure 6-4. The ELF calculated here shows strong and large attractors around the H atoms. The very low ELF value on the B sites indicates delocalized electrons, while a spherical shell attractor is seen around Mn. Low ELF value between Mn and [BH4]complex indicates an ionic bond. Whereas, a very high va lue of ELF within the [BH4]complex indicates covalent bonding. The plotted charge density shows th at valence charge de nsity around Mn atom is considerably low. The charge density is strongly localized around [BH4] anions. This also confirms the strong cova lent B-H bonding within the [BH4]complex. Figure 6-4. ( a-b ) Electron localization function (ELF) and Charge density of Mn(BH4)2 for I-4m2 symmetry. The plane contains both Mn and B and H atomic sites are projected on the plane 6.4.3. Thermodynamics Formation enthalpy and Gibbs energy are th e best aids to establish whether the theoretically pred icted phases are likely to be stab le or not. The thermodynamic

PAGE 107

91 expressions for the enthalpy [ H(T) ] and Gibbs energy [ G(T) ] can be obtained by the following equations: pV T E T H + = ) ( ) ( 6-2 ) ( ) 0 ( ) ( T H E T E + = 6-3 ZPE elecE E E + = ) 0 ( 6-4 ) ( ) ( ) ( ) ( T E T E T E T Htran rot vib+ + = 6-5 ) ( ) ( ) ( T TS T H T G Š = 6-6 where, Eelec is the electronic total energy of the crystal obtained from the density functional total energy calculations. Evib(T), Etran(T) and Erot(T) are the finite temperature vibrational (excluding ZPE), translational and rotational energies, respectively, and p is the pressure (1 atm), and V is the volume. The translational, rotational and pV energy terms are important for the gas phase only, an d thus are neglected for these solid-phase calculations. The EZPE for Mn was found to be 4.04 kJ/f.u.. The zero point energy of I4m2 space group of Mn(BH4)2 was calculated to be 204.03 kJ/f.u, respectively. For B and H2 gas molecule the EZPE were found to be 12.478 kJ/f.u. and 28.12 kJ/f.u.93. The calculated enthalpy [ H(T) H(298) ], entropy ( S(T) ) and Gibbs energy [-( G(T)-H(298)) ] for B and Mn are given in Figures 6-5 and 6-6, respectively. The thermodynamic functions were calcula ted on the basis of the positive phonon DOS (Figure 6-7) and resulting vibrational energy contribution to the Gibbs energy by the direct method lattice dynamics. The solid and dotted lines (Fi gures 6-5 and 6-6) represent the calculated and experimentally measured111 results, respectively. The calculated thermodynamic functions: the en thalpy and the Gibbs energy are in good

PAGE 108

92 agreement with the experimental values for B but in little less agreement (within ~20% of experimental values) for the Mn crystal lattice. Figure 6-5. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for -B. The functions including the enthalpy ( H(T)H(298) ), the entropy ( S(T) ) and the Gibbs energy [-( G(T)-H(298)) ] are presented in ( a ), ( b ) and ( c ), respectively. H(298) denotes the standard enthalpy at 298 K ( a ) ( b ) ( c ) Figure 6-6. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for -Mn (anti-ferromagnetic). The func tions including the enthalpy ( H(T)H(298) ), the entropy ( S(T) ) and the Gibbs energy [-( G(T)-H(298)) ] are presented in ( a ), ( b ) and ( c ), respectively. H(298) denotes the standard enthalpy at 298 K

PAGE 109

93 Figure 6-7. ( a-d ) Total density of phonon states g(w) of B, Mn, MnB2 and MnH2 crystal structures, respectively. Total phonon density of states is normalized as= 1 ) ( dw w g For H2 gas molecule, one needs to know the Gibbs energy at the standard pressure of 1 atm. The vibrations cannot be trea ted directly from the phonon calculations for H2 gas molecule because the phonon approach always considers the system as a crystal solid, and thus neglects the translational and rotational modes. The Gibbs energy was calculated by combining both computational a nd experimental results using Equation (67)112: ) ( ) ( ) ( ) (2 2 2 2H G pV H E H E H GT ZPE el T + + + = 6-7

PAGE 110

94 where, Eel(H2) is the electronic binding energy of a H2 molecule obtaine d from constant velocity MD simulation using VASP93. EZPE(H2) is the zero-point energy of a H2 molecule, p and V are the pressure (1 atm) and the molar volume (of the H2 ideal gas), respectively, and the last term GT(H2) is the temperature-depe ndent Gibbs energy with respect to the temperature of 0 K. As a common procedure113, the GT(H2) can be calculated using Equation (6-8). )] ( ) ( [ )] ( ) ( [ ) (2 0 2 2 0 2 2H S H S T H H H H H GT T TŠ Š Š = 6-8 where, HT(H2) and H0(H2) are the enthalpies of H2 at T and 0 K, respectively, and ST(H2) and S0(H2) are the entropies of H2 at T and 0 K, respectively. Inputting the thermochemical data114, we calculated the numerical values of GT(H2) 6.4.4. The enthalpy and the Gibbs energy of the reactions In the previous section, the calculated and experiment al thermodynamic functions of the B, Mn and Mn(BH4)2 lattices were plotted. In this section, these functions are utilized to calculate the enthalpy an d Gibbs energy of reaction. To study the reactions, one needs to calculate the temperature-dependent Gibbs energy difference (Grxn) and the entropy contribution difference (TS(T))rxn between the reactants and products using the followi ng equations (Equations ((6-9)-(6-10))): Š = productsts reac rxnT G T G T Gtan) ( ) ( ) ( 6-9 Š = productsts reac rxnT TS T TS T TStan) ( ) ( )) ( ( 6-10 The reaction enthalpy can be calculate d from Equation (6-11) as follows: rxn rxn rxnT TS T G T H )) ( ( ) ( ) ( + = 6-11

PAGE 111

95 For convenience, we denote r = reactants and p = products. A positive value of Gp-r means that the reaction is not spontane ous and thermodynamically not favorable, a negative value of Gp-r means the reaction is spontane ous and hence, thermodynamically favorable. The standard state enthal py of formation of Mn(BH4)2 was found to be -58.89 kJ/f.u. using DFT while the standard state enthalpies of formation of MnH2, MnB2 and B2H6 were found to be -12.10, -113.93 and 40. 26 kJ/mol. In comparison, literature values are -13.095, -120.0 (expt.)115/-140.0 (theory)96 and 35.3 kJ/mol, re spectively. The standard state Gibbs energies of de hydrogenation reactions (Equations ((6-1 a )-(6-1 e ))) were calculated, with roundedoff values in kJ/mol of H2 as -17, -10, -45, 25 and 13, respectively. It can be c oncluded from the calculated sta ndard state Gibbs energy of the dehydrogenation reactions th at the reactions (6-1 c ), (6-1 a ) and (6-1 b ) are the three favorable dehydrogenation reactions at standard conditions. So, it is expected that upon heating reaction (6-1 c ) will occurs first and then reactions (6-1 a ) and (6-1 b ) may or may not occur in practice. No diborane products should be produced during the dehydrogenation reaction, theoretically. The calculated Gibbs energy differences Gp-r for the reactions ((6-1 a )-(6-1 e )) for the temperature range of 0-800 K are plotted in Figure 6-8( a ). It is shown that the Gibbs energy and the enthalpy of reaction is the lo west and negative, respectively for reaction (6-1 c ) over all of this temperature range. Bu t the experimental results suggests that dehydrogenation of Mn(BH4)2 is an endothermic reaction at its decomposition temperature (413 K)98 which contradicts the theoretically predicted exothermic reaction. This contradiction may be because it is kinetically unfavorable to form the MnB2 phase at

PAGE 112

96 the lower temperature range (< 800 K) c onsidered in our calcu lations. Formation enthalpy of MnB2 has been measured by calorimetrically at very high temperatures above 2000 K115. Perhaps, kinetics prevented its forma tion for these authors to have not made their measurements at lower temperatures. So the next most favorable reaction is to be considered as reaction 6-1( a ), which is endothermic a nd also Gibbs energy become negative over all of the temperature range above 116 K. Hence equation 6-1( a ) is found to be most feasible to occu r at the experimentally determ ined reaction temperature (413 K). However, experimental i nvestigations are required to identify the phases which exist after the dehydrogenation of Mn(BH4)2 before providing more specific interpretations. The calculated Gibbs energy differences Gp-r and reaction enthalpy for the reaction (6-1 a ) for the temperature range of 0-800 K are shown in Figure 6-8( b ). The enthalpy of reaction for (6-1 a ) reaction suggests that th e dehydrogenation reaction is endothermic over all of this temperature range. It is also shown that the Gibbs energy of reaction (6-1 a ) is positive till 116 K and become negative over all of the temperature range studied above 116 K. This suggests that upon heating, Mn(BH4)2 is stable with respect to decomposition into Mn + 2B + 4H2 till the Gibbs energy become negative and then releases H2 according to the above reaction spontan eously. This is due to the large entropy contribution (TS ) from the gas molecules. Th e entropy contributions for the 4H2 gas molecules, the crystal B, crystal Mn and crystal Mn(BH4)2 lattices for the reverse formation reaction (6-1 a ) are shown in Figure 6-8( b ). High yield of gaseous products from the reactions should be favored with increasing temperature116, and this is consistent with our dehydrogenation reaction (6-1 a ). Based on standard state Gibbs energy of reaction (1 a ), and in Figure 6-8( a ), we can conclude that H2 desorption from

PAGE 113

97 Mn(BH4)2 is thermodynamically favorable essentia lly at all non-zero temperatures above 116 K. Decrease in negativ e free energies of dehydr ogenation with increase in temperature suggest that Mn(BH4)2 is metastable with respect to decomposition via reaction (6-1 a ). On the other hand, we found from our TGA experiment on Mn(BH4)2 indicates that desorption does not take plac e until the temperature is raised above 413 K98. These suggest that the relatively high temperatures needed in practice for H2 release are a consequence of poor kinetics, not unf avorable thermodynamics. It would be interesting to test whether the pred icted decomposition reaction is observed experimentally or not. The values of Hp-r are always positive above 0 K, indicating that the dehydrogenation reaction (6-1 a ) is endothermic at standard conditions. It can also be seen from Figure 6-8( b ) that the values of (TS) p-r are always positive, indicating that the entropy contribution in p is always larger than in r (as T>0 K). Therefore, it is concluded that the entropy contribution is the essent ial reason for the dehydrogenation reaction to take place. From the above figure, it can also be concluded that the net entropy contribution to the dehydrogenation reaction is mainly due to the entropy contribution from the H2 gas molecule.

PAGE 114

98 ( a ) ( b ) Figure 6-8. Temperature dependent reaction enthalpy and Gibbs energy of reactions ((6-1 a )-(61 e )) ( a ), temperature dependent reaction enthalpy and Gibbs energy and the entropy contribution for the reaction Mn(BH4)2 = Mn + 2B+ 4H2 ( b ). Energy difference between the reactant r = (Mn(BH4)2) and product p = (as an example: Mn + 2B + 4H2). Symbol Gp-r indicate the Gibbs energy difference between r and p [Gp-r = Gp – Gr]. The Gibbs energy difference Gp-r (lower palette) and the enthalpy difference Hp-r (upper palette) are shown in figures ( a ). The total entropy contribution for the reaction (6-1 a ) (TS) p-r (blue line), for H2 molecules 4TS(H2) (red line), for Mn (black line), for B (green line) and for the Mn(BH4)2 (brown line) are shown in figure ( b ). Rxns: (1-5) reactions: ((6-1 a )-(6-1 e )). 6.5. Summary In this work, we have predicted the stability and favorable dehydrogenation reactions of the complex manganese borohydride structure from theoretical calculations. This required investigation of both the gr ound state energy by DFT and the lattice

PAGE 115

99 vibration energy by the direct method lattice dynamics. Thes e calculations establish the most stable structure of Mn(BH4)2. The Gibbs energy calculations of Mn(BH4)2 at finite temperatures suggest that I-4m2 symmetry is the most stable structure at finite temperature. Our estimated value for the formation enthalpy of Mn(BH4)2 is found to be -28.93 kJ/f.u. (with ZPE) at 0 K, which suggest s that it should be pos sible to synthesize this phase. From electronic structure calcu lations, ionic interaction between Mn and [BH4]and strong covalent B–H interaction within the [BH4]tetrahedral structure are revealed. Electronic density of states studies reveal that Mn partial 3 d DOS has metallic character for spin-up and insulating charact er for spindown projection, are supposed to show some intermediate behavior which suggests that the Mn(BH4)2 is a half-metallic hydrogen storage material. The phonon dispersion relations and phonon dens ity of states of the solid phases are calculated using a direct force-constant method. On the basis of the phonon DOS and resulting vibrational energy contribution by the direct method lattice dynamics, the calculated thermodynamic functions including the enthalpy and the Gibbs energy are in good agreement with the experimental values of both B and Mn crystal lattice. Standard state enthalpy of reaction and Gibbs energy calculations reveal that reaction (6-1 a ), i.e., Mn(BH4)2 Mn + 2B + 4H2 is the most feasible dehydrogenation reaction. The calculated standard stat e enthalpy of dehydrogenation is positive, indicating that the decomposition reaction is endothermic at decomposition temperature. To identify the components produced via dehydrogenation, we have performed a computational search over 5 different de hydrogenation reactions, and identified the thermodynamically favorable reaction over the temperature range T = 0 800 K. The

PAGE 116

100 calculated Gibbs free energies of dehydrogenation of Mn(BH4)2 are negative above 116 K, indicating that Mn(BH4)2 releases H2 according to the above reaction above 116 K spontaneously. It was also found that the net entropy contribution for the dehydrogenation reaction is mainly due to the entropy c ontribution from H2 gas molecule. These findings suggest that th e crystal structure is stable at finite temperature and no diborane as a dehydrogenation product is preferred thermodynamically. Thus, Mn(BH4)2 could be considered a pot ential candidate for hydrogen storage. However, one needs to address the reversibility issue befo re its practical on-board applications. Our study has given new insights into the therm odynamically stable phase for the complex manganese borohydride. The methodology outlined in this work can be applied to studies on other complex hydrogen storage materials.

PAGE 117

101 Chapter 7 First-Principles Inve stigation Of Li-Mg-B-N-H System 7.1. Abstract A density functional theory study with th e gradient generalized approximation (GGA) and augmented plane wave method (P AW) is performed for the hydrogen storage properties of the complex multinary storage Li-Mg-B-N-H system. Using ab-initio methods, stability of the structures at finite temperatures are confirmed via. phonon spectrum calculations. Thermodynamic propertie s such as heat of reaction, and Gibbs energy for each reactant and product in the r eaction steps in different temperature zones are calculated. It is found that reversibil ity occurs in the temperature range of 160-225 oC with approximately 4.38 wt% hydrogen stor age capacity. The enthalpy of reversible re/de-hydrogenation is found to be 55.17 kJ/mol H2, which is supported by experimental data. The total hydrogen storage capacity of this material is calculated to be 8.76 wt% from the desorption behavior observed at different temperatures up to 350 oC. These theoretically esta blished reactions are validated w ith the suggested mechanism from experimental observations for the dehydrog enation reaction of this Li-Mg-B-N-H multinary system. These efforts are expected to contribute towards identification of suitable hydrogen storage materials.

PAGE 118

102 7.2. Introduction The discovery of clean, renewable and in expensive sources and carriers of energy has been the concern of researchers worldwid e. Compressed, liquid and slush have been the favored forms for mobile hydrogen storage. The low efficiency and increased risk of storage of such forms made solid stat e hydrogen the preferre d form for mobile applications. Research has gained pace to find a reversible material with efficient hydrogen storage capacities. Metal/complex hydr ides as hydrogen storage material have gained importance in the past few years due to rising fuel costs and ease of on-board storage. The individual storage propertie s of metal/complex hydr ides were found to improve in their combinations. The undesire d properties of these hydrides like ammonia liberation117 or diborane liberation87, 118, poor kinetics and irreve rsibility were overcome by their binary systems119. Many experimental and theore tical studies have been carried out to find materials with reversible propert ies, low desorption temperatures, accelerated kinetics and optimum hydrogen storage capac ity (at 298 K, the US revised DOE system target for gravimetric hydrogen stor age capacity is 5.5 wt % by 2015)93, 94, 98, 120-125. Experiments suggest that complex hydrides of Li, Na, Mg, B, Al and N have such properties119. The LiBH4LiNH2 system fails to re-hydride due to its relatively small dehydriding enthalpy, and suggests that hydrogen release from the syst em is a kinetically rather than thermodynamically hindered process79. The LiBH4-MgH2 system was found to have 8-10 wt% reversible hydrogen storage at about 350 C126. The LiNH2MgH2 system operates at 200 C with 4.5 wt% reversible hydrogen storage127. Experiments on the ternary systems of these hydrides also showed promising results and further improved storage properties. The combined LiNH2-MgH2-LiBH4 system was

PAGE 119

103 found to have lowered desorption temperat ures, increased hydroge n purity, accelerated kinetics and good reversibility. The stoichiometr y of this mixture was also a major factor for controlling the storage propert ies. It has been reported that the stoichiometric ratio of LiNH2:LiBH4:MgH2 with a 2:1:1 molar ratio gives the best hydrogen storage performance 119, 128. In this work, we tested the feasibility of the 2 LiNH2 LiBH4 MgH2 reactant mixtures using density functional theory (D FT) studies. The reac tants and products of the proposed reaction steps were checked fo r stability and then the thermodynamic properties of the compounds were found and co mpared with literature. The storage capacity of these materials was calculated from the desorption behavior observed at different temperatures. These theoretically obtained data can be used to validate the suggested mechanism from experimental ob servations for the de hydrogenation reaction. 7.3. Computational details DFT studies on hydrogen storage materi als using projecte d augmented wave (PAW)25, 60 potentials pertaining to the gene ralized gradient approximation (GGA)129 are considered the most appropriate for energy calculat ions due to their be tter accuracy than when using the local density approximation (LDA)130. The thermodynamic properties calculated using plane wave DFT and t hose found experimentally were in good agreement for a number of hydrogen storage systems131, 132. We used VASP18, 26, 27 (Vienna ab initio simulation package) to perform the DFT calculations and code PHONON35 to perform direct method lattice dynamics32-34 for finite temperature crystal stability and thermodynamics.

PAGE 120

104 The most favored structure of each compound was found from literature and its equilibrium volume was calculate d by a quadratic fit of energy. These calculations were performed using Monkhorst Pack61 meshes containing 5 x 5 x 5 k points. The energy cutoff for all the compounds was set to 400 eV. The external pressu res acting on each of these compounds were reduced by relaxing the volume and ionic positions of the structure. PAW potentials were used for these relaxations. The relaxations were continued till the forces on the atoms were reduced to about 0.1 meV/. HellmannFeynman (HF) forces were calculated using a su per cell of the relaxe d structure to avoid interaction between the atoms and th en displacing its atomic positions. These HF forces were then used to cal culate the phonon density of states (DOS) for the structure at finite temperatures, which would be positive for a stable structure. Entropy and internal energy of the structur e can be obtained from phonon DOS, which can be used to calculate th e Gibbs energy of the reactions132. The experimentally observed in situ IR spectroscopy patterns were compared with the theoretically observed ones to validate the structure of the intermediate phases such as Li4BH4(NH2)3, Mg(NH2)2, Li2Mg(NH)2 and LiMgBN2. To specify the atomic positions of the compounds with fractional occupancies such as Li2Mg(NH)2, we used the super cell approach125, 133, 134. 7.4. Results and discussion In this study, we calculated the equilibrium structures for all the compounds, namely, LiBH4, Mg(NH2)2, LiNH2, LiH, Li2Mg(NH)2, Li3BN2, MgH2, Li4BH4(NH2)3, Mg3N2, LiMgBN2 and elements like Mg, Li, B, N2 and H2, involved in the reaction mechanism (Table 1). Among all the above mentioned compounds, Mg(NH2)2 and

PAGE 121

105 Li2Mg(NH)2 are found to be com putationally intensive for phonon calculations as Mg(NH2)2 is characterized by a te tragonal unit cell belonging to the space group No. 142 with 224 atoms per unit cell122, and Li2Mg(NH)2 is characterized by a orthorhombic unit cell belonging to the space group No. 45 with Li/Mg occupancy fractions (due to ordering of atoms) on “4 b ” and “8 c ” Wyckoff sites, resulting in a big supercell for phonon calculations. Possible arrangements of Li and Mg on “4 b ” and “8 c ” Wyckoff sites and the ordering of Mg atoms on the “8 c ” and “4 b ” sites play a significant role in determining the energetics. The magnesium atoms occupy (i) either body-diagonal or face-diagonal tetrahedral sites, (ii) either face-d iagonal tetrahedral s ites, and (ii) the nearest-neighbor or body-di agonal tetrahedral sites125. The Li/Mg occupancy fractions of 0.75 0.25 and 0.50 0.50 on “4 b ” and “8 c ” sites133, respectively, are assumed in our calculations. Phonon calculations for the rema ining solid state compounds/elements were straight forward as disc ussed in our earlier study93, 94, 132. 7.4.1. Structural stability The finite temperature stability of all the compounds and elements was established by using combined DFT and dir ect method lattice dynamics approach based on positive definite frequency distribution criterion over the Brillion zone. Results for some of these compounds are shown in Figure 7-1. These stable structures were then used to calculate the finite temperature thermodynamics of the reactions132. Experimentally proposed multi-steps hydrogen release pathway for the intermediate phases Li4(NH2)3(BH4), Mg(NH2)2, Li2Mg(NH)2 and LiMgBN2 119, 128 that were formed was examined via vibrational spectroscopy, i.e., phonon density of states calculations of the intermediate phases. In line with the

PAGE 122

106 experimental FTIR study on the hydrogen reactions of LiBH4-LiNH2-MgH2 systems, our calculations showed the co-existence of B-H, N-H and NH2 stretches, as presented in Figure 7-1. The B-N stretching range f ound from the phonon density of states of Li3BN2 (formation during the exothermic reaction (4)) was 1510-1547 cm-1 and B-N stretching range found from phonon density of states of LiMgBN2 (formation during the reaction (5)) was 1700-1874 cm-1 (Figure 7-2). The experimentally determined B-N stretching range was found to be 1682 1746 cm-1 after dehydrogenation of the Li-Mg-B-N-H system at 300 oC128 which is in the high temperature range 225-325 oC. This B-N stretching information supports fo rmation of the new phase LiMgBN2 at temperatures above 300 oC. Figure 7-1. Normalized plot of experimental FTIR spectra of the LiBH4-LiNH2-MgH2 systems and phonon density of states of intermediate states ( i e (Li4BH4(NH2)3, Mg(NH2)2, LiBH4, MgH2, Li2Mg(NH)2 and LiH) during the first step hydrogen release reaction.

PAGE 123

107 Table 7-1. Calculated ground state ener gy and zero point energy of different reactants/products: Reactants/products Space group (number) Ground state energy (eV/f.u.) ZPE, kJ/f.u. (cal. value) ZPE, kJ/f.u. Ref.(lit. value) LiBH4 Monoclinic (9) -23.91 107.20 79 107.1 MgN2H4 Tetragonal (142) -35.62 132.47 122 134.8 LiNH2 Tetragonal (82) -19.05 68.26 79 69.0 LiH Cubic (225) -6.17 21.02 133 21.5 Li2MgN2H2 Orthorhombic (45) -33.09 85.95 122, 133 86.9 Li3N2B Tetragonal (141) -34.37 50.33 79 52.2 MgH2 Tetragonal (136) -8.98 39.69 122 39.10 Li4N3BH10 Cubic (199) -81.93 314.10 79 314.40 Mg3N2 Cubic (206) -25.40 33.76 122 33.20 LiMgBN2 Tetragonal (142) -31.57 45.87 NA Mg Hexagonal (194) -1.48 2.94 133 2.79 Li Cubic (229) -1.90 3.83 79 3.9 B Trigonal (166) -6.48 12.61 79 12.50 N 2 -16.66 15.66 135 14.40 H2 -6.78 28.29 135 26.3 Based on the experimental thermal program desorption (TPD) hydrogen release profile136, the reactions are categorized into four groups and they are mechanical milling (0-95 oC), first peak (95-160 oC), first shoulder peak (160-225 oC) and second peak (225325 oC). The following different r eactions in different temper ature ranges were identified based on the PXRD and in situ IR spectroscopy data119, 128. Mechanical milling (0-95 oC)

PAGE 124

108 2LiBH4 + 6LiNH2 2Li4BH4(NH2)3 7-1a 2Li4BH4(NH2)3 + 3MgH2 3Mg(NH2)2 + 2LiBH4 + 6LiH 7-1b First peak (95-160 oC) 2Li4BH4(NH2)3 + 3MgH2 3Li2Mg(NH)2 + 2LiBH4 + 6H2 7-2 Shoulder of first peak (160-225 oC) Mg(NH2)2 + 2LiH Li2Mg(NH)2 + 2H2 7-3 Second peak (225-325 oC) 3Li2Mg(NH)2 + 2LiBH4 2Li3BN2 + Mg3N2 + 2LiH + 6H2 7-4 2Li3BN2 + Mg3N2 + LiBH4 3LiMgBN2 + 4LiH 7-5 Figure 7-2. Normalized plot of phonon density of states of Li3BN2, MgN3 and LiMgBN2 (intermediate phases of Li-Mg-B-N-H syst em during second step hydrogen release).

PAGE 125

109 7.4.2. Thermodynamics and reactions The thermodynamic properties were calc ulated on the basis of the positive phonon DOS (example: Figures 7-1 and 7-2) and resulting vibrational energy contributions to the Gibbs energy by the direct method lattice dynamics. The solid and dotted lines (Figures (7-3) – (7-6)) represent the calculated and experimentally measured111 results, respectively. The calculate d thermodynamic functions: the enthalpy and the entropy are in good agreement with the experimental values for MgH2, LiH, Mg3N2 and LiBH4 (within ~8.5% of experimental values) crystal lattices. Figure 7-3. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for MgH2. The functions including the enthalpy ( H(T)H(298) ) and the entropy ( S(T) ) are presented in ( a ) and ( b ), respectively. H(298) denotes the standard enthalpy at 298 K.

PAGE 126

110 Figure 7-4. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for LiH. The functions including the enthalpy ( H(T)H(298) ) and the entropy ( S(T) ) are presented in ( a ) and ( b ), respectively. H(298) denotes the standard enthalpy at 298 K. Figure 7-5. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for Mg3N2. The functions including the enthalpy ( H(T)H(298) ) and the entropy ( S(T) ) are presented in ( a ) and ( b ), respectively. H(298) denotes the standard enthalpy at 298 K.

PAGE 127

111 Figure 7-6. Calculated (solid line) and experimental111 (dotted line) thermodynamic functions for LiBH4. The functions including the enthalpy ( H(T)H(298) ) and the entropy ( S(T) ) are presented in ( a ) and ( b ), respectively. H(298) denotes the standard enthalpy at 298 K. For gas molecules (H2 and N2), one needs to know the Gibbs energy at the standard pressure of 1 atm. The vibrati ons cannot be treated di rectly from the phonon calculations for gas molecules because the phono n approach always considers the system as a crystal solid, and thus neglects the tr anslational and rotational modes. The Gibbs energy of a gas molecule was calculate d by combining both computational and experimental results presented in our earlier work132. The Gibbs energy calculations of these react ions (Figure 7-7) s how that reactions (7-1a) and (7-1b) really do not have mu ch Gibbs energy change with respect to temperature. This suggests that these two reactions are occu rring due to the ball milling/mixing process and also do not re lease any hydrogen during the synthesis process. The first hydr ogen release occurs via reaction (7-2) above 95 oC, which is also predicted from theoretical calcu lations as shown in Figure 7-7. Similarly, the shoulder peak due to reaction (7-3) is also supported by theoretical calc ulations. The second step

PAGE 128

112 hydrogen release due to r eaction (7-4) occurs at high temperature (>225 oC) which may be due to poor kinetics, not unfavorable thermodynamics. Formation of LiMgBN2 occurs via reaction (7-5) at the end of the second step hydrogen release i.e. above 300 oC as discussed in the earlier sec tion. The reaction enthalpy calculations of the hydrogen release reactions (Figure 7-8) show that the complete fi rst step hydrogen release is endothermic and the second step is exothermic reaction. Figure 7-7. Temperature dependent Gibbs energy of reactions (7-1 a 7-1 b (7-2) – (7-4))

PAGE 129

113 Figure 7-8. Temperature dependent reacti on enthalpy of reactions ((7-2) – (7-4)) 7.4.3. Quantitative analysis of hydrogen release It is believed that during the synthesis of Li-Mg-BN-H hydrogen storage system, initially, the quaternary phase (Li4BH4(NH2)3) formation (supported by phonon frequency of N-H and B-H stretching as shown in Figure 7-1) occurs (reaction (7-1a)) by intermixing/reacting with proportionate ratio of LiNH2 and LiBH4 and then half of this quaternary phase is destabilized by proportional amount of MgH2 to produce the amide phase (reaction (7-1b)). The reactions (7-1a) and (7-1b) can be rewritten as below: LiBH4 + 2LiNH2 Li4BH4(NH2)3 + LiBH4 7-6 Li4BH4(NH2)3 + MgH2 Mg(NH2)2 + LiBH4 + LiH 7-7

PAGE 130

114 The unreacted quaternary phases then react with the residual MgH2 at higher temperature according to reaction (7-2) to re lease the first hydrogen from the system. After reaction (7-2), the residue contains LiH, Mg(NH2)2, LiBH4 and Li2Mg(NH)2. So, a hydrogen balance gives the amount of hydrogen re leased during the fi rst step from the following equation: 2LiBH4 + 8LiNH2 + 4MgH2 4LiH + 2Mg(NH2)2 + 2LiBH4 + 2Li2Mg(NH)2 + 4H2 7-8 As the starting stoichiometric ratio of LiNH2:LiBH4:MgH2 is kept constant with a 2:1:1 molar ratio, the above equation can be rewritten as follow: LiBH4 + 2LiNH2 + MgH2 LiH + Mg(NH2)2 + LiBH4 + Li2Mg(NH)2 + H2 7-9 Hence, the amount of hydrogen re leased during the first step is one mole per mole of starting materials, which is equivalent to a pproximately 2.19 wt%. Similarly, after the first shoulder peak reaction, the residue contains only Li2Mg(NH)2 and LiBH4. This information can be supported by both in situ IR spectroscopy at 200 oC128 and phonon frequency of N-H stretching from th e imide and B-H stretching from LiBH4 (Figure 7-1), respectively. The amount of hydrogen releas ed in this shoulder peak step can be calculated from: LiH + Mg(NH2)2 + LiBH4 + Li2Mg(NH)2 LiBH4 + Li2Mg(NH)2 + H2 7-10 From reaction (7-10), the amount of hydrogen which can be released another 2.19 wt%, which is similar to that from the first should er peak. The total hydrogen released in the complete first step is 4.38 wt%. Similarly, based on the second step reaction, the amount of hydrogen released is two moles of hydrogen (~ 4.38 wt%) according to Equation (710). Finally, above 300 oC, residues are likely to form a new phase LiMgBN2 according

PAGE 131

115 to reaction (7-11). The quantitative su mmary of hydrogen release at different temperatures is gi ven in Figure 7-9. LiBH4 + Li2Mg(NH)2 Li3BN2 + Mg3N2 + LiH + LiBH4 + 2H2 (10) Li3BN2 + Mg3N2 + LiBH4 + LiH 2LiH + LiMgBN2 (11) Figure 7-9. Normalized phase compositions of diffe rent reactants/products and also intermediate phases in different temperature ranges 7.4.4. Reversible reaction step and van’t Hoff plot We now discuss the thermodynamics of the reversible reaction, i.e. reaction (73)119, 121, 123, 128, 137, 138. Using the calculat ed finite temperature thermodynamic properties

PAGE 132

116 such as Gibbs energy and enthalpy of reactions (Figures 7-7 and 7-8) and entropy data as input to the van’t Hoff equation, Figure 7-10 shows a plot of the equilibrium H2 desorption pressures of the reversible reacti on as a function of temperature. From our van’t Hoff plot, the enthalpy of reversible re/de-hydroge nation of the Li-Mg-B-N-H multinary system is found to be 55.17 kJ/mol of H2 which yields moderate H2 pressures at elevated temperatures. This also conf irms that the multinary system releases the desired amount of H2 at higher temperatures than the fuel cell operati ng conditions shown in Figure 7-10. In other words, they are a little too strongl y bound for practical, reversible on-board storage purposes. Cataly tic doping may further lower the reversible dehydrogenation temperature and improve kinetics. Figure 7-10. The van’t Hoff plot derived from the theoretical calculations for the reaction: Mg(NH2)2 + 2LiH Li2Mg(NH)2 +2H2.

PAGE 133

117 7.5. Summary In this work, we have pr edicted the stabil ity and dehydrogenation reactions at different temperatures of the complex multinary borohydride from theoretical calculations. This required investigation of both the ground state energy by DFT and the lattice vibration energy by the direct method lattice dyna mics. The phonon dispersion relations and phonon density of states of the solid phases are calculated using a direct force-constant method. Using the phonon DOS and resu lting vibrational energy contribution by the direct method lattice dynamics, the calculated thermodynamic functions including the enthalpy and the Gi bbs energy are in good agreement with the experimental values for MgH2, LiH, Mg3N2 and LiBH4 crystal lattices. The stability of the structures involv ed in a hydrogen desorption pathway was evaluated. The thermodynamic properties of these compounds were calculated and the feasibility of the mechanism was validated. It is found that all samples are intimate mixtures of Li4BN3H10 with MgH2 and that no new chemical compound is formed during the ball milling. This information was veri fied from both vibrat ional spectroscopy (B-H, Mg-H and N-H bond stretches; Figure 7-1) an d Gibbs energy calculati ons (Figure 7-7). B-N stretching found at around 1700-1874 cm-1 (Figure 7-2) confirms the formation of LiMgBN2 above 300 oC. The reaction enthalpy calculations of the hy drogen release react ions (Figure 7-8) show that the complete first step hydrogen release is endothe rmic and the second step is exothermic. The total hydrogen storage cap acity of this Li-Mg-B-N-H system is calculated to be 8.76 wt% from the deso rption behavior observed at different temperatures up to 350 oC. Theoretical calculations also show that the multinary

PAGE 134

118 complex borohydride exhibits first hydrogen release at a temperature of about 95 oC and then the main reversible hydrogen adsorption/ desorption reaction (re action (7-3)) occurs at a temperature of about 225 oC, which is a little higher than the experimentally determined temperature (190 oC)136. The quantitative reve rsible hydrogen storage capacity is found to be 4.38 wt% which is also different from the experimentally determined values of 2.5 wt% at 180 oC128 (5-6 wt% at around 160 175 oC by adopting different processing r eaction pathway schemes)136. The van’t Hoff plot indicates that the en thalpy of reversible re/de-hydrogenation of Li-Mg-B-N-H multinary system to be 55.17 kJ/mol of H2 which yields moderate H2 pressure at elevated temperature only. This suggests that the multinary system release H2 at higher temperatures, which is not the best situation for ideal fuel cell operation for onboard applications. However, this multinary system is the most promising reversible candidate compare to all ot her complex hydrides studied to-date. The reversible dehydrogenation temperature, kine tics and also storage capac ity can be further improved via catalytic doping. Finally, this type of combined DFT a nd lattice dynamics study could supplement experiments in testing complex multinary hy drides for hydrogen storage properties. These calculations are helpful in identifying feasible reactions and to rule out unrealistic possibilities, and arrive at prac tical hydrogen storage materials.

PAGE 135

119 Chapter 8 Experimental Study Of Li-Mn-B-H System 8.1. Abstract In this work, we report the synthe sis and characterization of LiMn(BH4)3, member of a new class of complex borohydrides for hyd rogen storage. This new complex hydride was prepared with a 3:1 rati o of precursor materials LiBH4 and MnCl2 via the solid state mechano-chemical process. Th e B-H stretch occurrence at 2374 cm-1 in addition to two other B-H bonding bands of LiBH4 (2228 and 2297 cm-1) from the FTIR investigation confirm the formation of LiMn(BH4)3 at room temperature. Thermo gravimetric analysis (TGA) of LiMn(BH4)3 indicated that a larg e amount of hydrogen ( 8.0 wt %) can be released between 135 and 155 C in a single de hydrogenation reaction step. Reduction in the decomposition temperature was achieve d by doping this Li-Mn-B-H system with small fractions of nano-Ni. An amount of 1.5 mol% nano-Ni was estimated and found to be the optimum concentration for effective d ecomposition. Nano-Ni loading in the host hydride lowers the melting and thermal decomp osition temperatures (at least by 20 C) as evidenced from the simultaneous TGA, DSC and TPD meas urements. The doped LiMn(BH4)3 exhibits lower activati on energy (112 kJ/mole) by 20 kJ/mole as compared to the undoped sample (131 kJ/mole). Moreove r, the gas chromatography studies of the undoped and doped LiMn(BH4)3 demonstrate that the evolved gas is mainly hydrogen and does not contain members of the borane family.

PAGE 136

120 8.2. Introduction Hydrogen storage is considered to be the key component requiring a research breakthrough for streamlining hydrogenbased clean-fuel transportation89, 139-141. A recent challenge in hydrogen storage is to find light weight, low cost and high capacity hydrides with favorable hydrogen sorption kinetics and thermodynamics for on-board vehicular applications138. Complex chemical hydr ides such as alanates142-146, alanes147, 148, borohydrides126, 149, amides150, 151 and their combinations137, 152, 153 are widely investigated in recent years due to their high theoretical hydrogen capacity and tunable properties. The breakthrough discovery of Ti-doped NaAlH4 154, 155 has renewed interest in revisiting these complex hydrides for re versible hydrogen storag e. Although the Tidoped alanates show re versible hydrogen storage behavi or at moderate temperatures, these systems may not be ideal to rea lize the DOE 2010 and FreedomCAR technical targets156. This is due to the maximum usable hydrogen storage capacity of 5.4 wt% for NaAlH4, which is considered to be well below the DOE target for 2010157. On the other hand, the borohydride complexes NaBH4 and LiBH4 possess high hydrogen storage capacity of 13.0 wt% and 19.6 wt%, respectively47, 148, 158. However, the release of hydrogen from NaBH4 is possible only by hydr olysis (reaction with H2O) and this process is irreversible159. For the case of LiBH4, addition of SiO2, significantly enhances its thermal desorption160 at 200 C. In general, the dehydrogenation and/or rehydrogenation of NaBH4 or LiBH4 are difficult to achieve because of the thermodynamic stability due to strong B-H interactions45, 92. New, less stable complex borohydrides, Zn(BH4)2 have been recently reported fo r hydrogen storage (~8.2 wt.%) at temperatures below 100 oC 48, 87, 93, 161. However, it was found that thermal

PAGE 137

121 decomposition of Zn(BH4)2 comprises of not only the evolution of H2, but also of an appreciable amount of B–H (borane) com pounds. Nanomaterial doping of the Zn(BH4)2 structure not only lowers the decomposition temperature by 20 oC but also suppresses the release of boranes, as found from both experimental and theoretical studies94, 118. Bialkali162, 163 or alkali-transition metal borohydrides164 also showed potential promise for hydrogen storage due to their high hydrogen cap acity and tunable properties. In this paper, we report one such system, LiMn(BH4)3 prepared by the solid state mechanochemical process. These materials are wi dely characterized us ing X-ray diffraction, Fourier-transformed infrared spectroscopy, thermo gravimet ric analysis, differential scanning calorimetric analysis, temperatur e programmed desorption, hydrogen sorption (kinetics, PCT and cycle life) and gas chromatography. Additionally, doping of nanomaterials (e.g. nanoNi, nanoCo etc.) and the enhancement of hydrogen decomposition characteristics have been extensively studied on these new complex hydrides. 8.3. Experimental details 8.3.1. Materials and method Starting materials LiBH4 (90% purity), MnCl2 (99% purity) nano-Zn (99.99%) and nano-Ti (99%) were obtained from Sigm a Aldrich and other nano-dopants nano-Ni, nano-Co, nano-Fe, nano-Cu and nano-Pd (99. 999%) were obtained from QuantumSphere Inc., CA, which were used without further purification. High purity H2 (99.9999%), N2 (99.99%) and He (99.99%) were procured from Airgas for the synthesis and analytical measurements. All chemical reactions and oper ations were performed in a nitrogen filled glove box. LiBH4 and MnCl2 at a 3:1 molar ratio were mixed in a stainless steel bowl

PAGE 138

122 (80 ml) and the lid sealed with a viton O -ring in the glove box. The bowl was then evacuated for 30 minutes to remove residual oxygen and moisture down to ppm levels. A specially designed lid with inlet and outlet valves was used for this purpose. The mechano-chemical process was done by high energy milling using the Fritsch pulversette planetary mono mill, P6, in an in ert atmosphere. The milling parameters, ball to powder weight ratio and milling speed were optimized to 20:1 and 300 rpm, respectively. Milling duration of 20-30 minut es was maintained for all the samples. These mechano-chemically processed complex hydrides were immediately transferred to the glove box for further charac terizations. In a similar way, a few mole concentrations of nano-dopants such as nano-Ni, nano-Co, na no-Fe, etc. were added during the milling process for the synthesis of nanomaterial doped LiMn(BH4)3. 8.3.2. X-ray diffraction The powder X-ray diffraction of the mech ano-chemically milled complex hydride was carried out by the Philips X’pert diffractometer with CuK radiation of = 5.4060 . The incident and diffracti on slit widths used for the measurements are 1o and 2o, respectively. The incident mask of 10 mm was used for all the samples and their XRD studies. The sample holder (z ero background silicon disc of 32 mm diameter procured from The Gem Dugout, Pennsylvania, USA) wa s covered with polyet hylene tape (foil) with O-ring seal in a N2 filled glove box to avoid the O2/moisture pickup during the XRD measurements. Diffraction from the tape wa s calibrated without the actual sample and found to be occurring at the 2 angles of 22o and 24o, respectively. The XRD phase identification and particle size calculation wa s carried out using the PANalytical X’pert Highscore software with built-in Scherer calculator.

PAGE 139

123 8.3.3. Fourier transform in frared spectroscopy The B–H bond stretch of the Li -Mn-B-H system was measured using a PerkinElmer Spectrum One FTIR spectrometer. This instrument operates in a single-beam mode and is capable of data collection over a wave number range of 370–7800 cm1 with a resolution of 0.5 cm1. The complex borohydrides sample s were palletized and sealed in a specially designed KBr ce ll for infrared measurements. 8.3.4. Simultaneous DSC and TGA The simultaneous DSC and TGA (SDT) analys is pertaining to the weight loss and the heat flow for the reaction enthalpy during thermal decomposition of undoped and nanomaterial doped complex hydrides were performed by using the TA instrument’s SDT-Q600 analytical tool. Calibration of SD T was performed in four steps with empty pan and standard sapphire disc The four calibration subr outines of TGA weight, DTA baseline, DSC heat flow and temperature were carried out before an actual measurement on the sample. A pre-weighed sample was loaded into the ceramic pan and covered with the ceramic lid inside the glove box to prevent moisture from getting into the sample during transfer. A ramp rate of 2oC/min was used for all the measurements. TA’s Universal Analysis 2000 software was used to analyze the TGA and DSC profiles. 8.3.5. Dehydrogenation kinetics: isotherm al volumetric measurements The isothermal volumetric measurements were carried out using Hy-Energy’s PCTPro 2000 sorption equipment. This fully automated Sievert’s type instrument uses an internal PID controlled pre ssure regulator with maximum pr essure of 170 bars. It also includes five built-in and calibrated rese rvoir volumes of 4.66, 11.61, 160.11, 1021.30

PAGE 140

124 and 1169.80 ml. The volume calibration without and with the sample was performed at a constant temperature with an accuracy of 1oC using a helium ga s. The software subroutines for hydrogen purging cycles, leak te st, kinetics, PCT and cycling, etc. were performed by the HyDataV2.1 Lab-View pr ogram. The data collected from each run were analyzed using the Igor Pro 5. 03 program using HyAnalysis Macro. 8.3.6. Temperature programmed desorption measurements Temperature programmed desorption (TPD ) measurement was carried out using the Autosorb-1 equipment of Quantachrome Instrument. A 100 to 120 mg amount of sample was loaded in the reactor and heated in a 25 mL/min helium flow while heating from 25 to 200 C at 5 oC/min. The thermal desorption/reduction profiles were recorded and analyzed using TPRWIN software package. 8.3.7. Gas chromatography analysis It was observed in our previous study on Zn(BH4)2 88 that di-borane gas evolved in combination with H2 during the thermal decomposition process. This feature turns the hydrides into irreversible systems obviating practical applications In the current investigation, samples of both doped a nd undoped versions of the new complex borohydrides were subjected to gas chromatography analysis during the thermal programmed desorption process. The gas sa mple was injected (less than 50-100 micro liter) into the TCD detector and the GC signal recorded over a period of retention time. The gas analysis and plotting of the curves were carried out by Satu rnview Version 5.52 software.

PAGE 141

125 8.4. Theory 8.4.1. Activation energy calculations For any chemical reaction, activation ener gy roughly corresponds to the height of the free energy barrier. The transition state al ong a reaction coordina te is the point of maximum free energy, where bond-making and bond-breaking are balan ced. Multi-step reactions involve a number of transition states. Activation energy is also the minimum energy necessary for a specific chemical reaction to occur. The activation energy of a reaction is usually denoted by Ea, with units of kJ/mole. Activation energy can be reduced by doping of the complex hydrides as represented in Figure 8-1. The Arrhenius equation gives a quantitative basis for the relationship between the activation energy and the rate at which a reaction proceeds. From the Arrhenius equation, the basic rate equation can be expressed as ) ( ) exp( a f RT E A dt daaŠ = 8-1 where, A is the frequency factor for the reaction, R is the universal gas constant, and T is the absolute temperature. This equation sugge sts that the activation energy is dependent on temperature, in the regimes in whic h the Arrhenius equation is valid. Thus Ea can be evaluated from the rate constant at any temperature (within th e validity of the Arrhenius equation). Experimentally, we have determined th e activation energy of complex hydrides by temperature programmed desorption (TPD). Once a sample is saturated with Ar:H2 (reactive gas, 95:5%) at a fixed temperature (n ormally near ambient), a flow of inert gas

PAGE 142

126 and linear heating rate are applied to desorb previously adsorbed species (in our case it is H2). Plots of ln( /Tmax 2) vs 1/Tmax yield a straight li ne with a slope –Ea/R, where Ea is the activation energy mentioned above for the hydrogen decomposition process or the bonding strength. Activation ener gy of undoped and doped LiMn(BH4)3 were estimated according to the Kissinger’s theory165 with the data obtained from TPD measurements of the samples with the ramping rates of 4, 10 and 20 oC/min. Figure 8-1. Activation energy curve (a) undoped and (b) catalytic doping reactions 8.5. Results and discussions 8.5.1. Formation of complex hydride LiMn(BH4)3 –FTIR and XRD explorations Synthesis of the new complex hydride, LiMn(BH4)3 from the parent compounds LiBH4 and MnCl2 (3:1) was carried out in solid state employing the mechano-chemical milling process. The reactions proceed based on the stoichiometric ratio given in the Equation (8-2) below,

PAGE 143

127 3LiBH4 + MnCl2 LiMn(BH4)3 + 2LiCl 8-2 The B-H bonding environment of LiMn(BH4)3 prepared based on the above reaction was determined by the FTIR spectroscopic techniqu e as shown in Figure 8-2. The FTIR spectra of BH4 ion in LiBH4 has characteristic bands at 2224 and 2298 cm-1, whereas the LiMn(BH4)3 structural phase shows a new peak at around 2374 cm-1 from the formation of the new complex hydride, in addition to showing the parent BH stretch. The BH2 bending modes are the same for both LiMn(BH4)3 and the parent compound LiBH4. Figure 8-2. FTIR profiles of LiBH4, MnCl2, and LiMn(BH4)3 + 2LiCl ball milled mixture representing B-H bonding bands and BH2 bending vibrations Figure 8-3 represents the powder X-ray di ffraction patterns of the pristine LiBH4 and the complex mixture (3LBH4+MnCl2) mechanically milled in hydrogen ambient for 20 minutes. Bragg reflections with high crys talline phases were observed corresponding

PAGE 144

128 to the presence of pure LiBH4 for the parent compound. For the mixed complex hydrides, after the mechano-chemical reaction, the by-product consists of major LiCl peaks which agree well with react ion (8-2) above. Since LiMn(BH4)3 is not of highly crystalline nature, these peaks are not dis tinct. Nevertheless, both the FTIR and XRD spectra indirectly confirm the forma tion of the new complex hydride LiMn(BH4)3 from the reaction mixtures of 3LiBH4 and MnCl2. We have also prepared these complex borohydrides with various concentrations of nanomaterial doping for enhancing hydrogen decomposition characteristics. The succe ssful synthesis of new complex hydride, LiMn(BH4)3 and its doped counterparts were further evaluated by various other analytical studies such as TGA, DSC, TPD, PCT and GC which are elaborat ed in the following sections. Figure 8-3. X-ray diffraction patterns of a pure LiBH4 and LiMn(BH4)3+2LiCl mixture obtained after milling under H2 ambient for 20 minutes

PAGE 145

129 8.5.2. TGA, DSC and TPD studies of undope d and nanomaterials doped LiMn(BH4)3 We have demonstrated the successful pr eparation of new alkali-transition metal based complex borohydrides, LiMn(BH4)3 from the precursors of LiBH4 and MnCl2 by the mechano-chemical process. To redu ce the decomposition temperature further, we have attempted to dope the LiMn(BH4)3 with different mole concentrations of nanomaterials (nano-Ni). It was clearly noticed that doping with nanoNi lowers the decomposition temperature and at the same time enhances the kinetics of the reaction. Optimization of nano-Ni concentration is extremely important fo r favorable hydrogen storage characteristics. We have studi ed the optimization procedures for doping LiMn(BH4)3 and the hydrogen storage characteri stics of the resulting materials. Figure 8-4 represents the simultaneous thermogravimetric (TGA) and differential scanning calorimetric (DSC) profiles of undoped and doped complex borohydrides, LiMn(BH4)3. Various molar concentrations (X=0, 0.5, 1, 1.5, 2, 2.5, 3) of nano-Ni were doped with the host borohydrides matrix. From Figure 8-4, it is clea rly discernable that the doped samples exhibit at least 20-30 oC reduction in decompos ition temperature (Tdec) compared to the undoped one (see also Table 8-1). Moreover, X=1. 5 mol% of nano-Ni appears optimum in terms of the earlier on-set temperature and higher hydrogen desorption capacity. DSC profiles complimented this feature as observed from the earlier endothermic transition due to hydrogen decompos ition. Based on this optimization of the concentration analysis, we have further ca rried out studies using various nanomaterial dopants (e.g. nano-Co, nano-Fe et c.) with the X fixed at 1. 5 mol% and the results are shown in Figure 8-5.

PAGE 146

130 By fixing the optimum concentration of nano-additive (X= 1.5mol%), the TGA and DSC spectra were obtained for several additives. Nano-Ni, nano-Co, nano-Fe, nanoCu, nano-Ti, nano-Zn and nano-Pd were studied (see Figure 8-5). Sim ilar to Figure 8-4, the doped LiMn(BH4)3 materials reveal earlier decom position than the undoped samples. Among the various dopants, nano-Ni and nano-Co exhibit remarkable performance and in general the stabili ties are ordered as nanoNi < nanoCo < nanoFe < nanoTi < nanoZn < nanoCu
PAGE 147

131 Table 8-1. DSC and TGA analysis of undoped and nanoNi doped Li-Mn-B-H X-mol% nanoNi doped LiMn(BH4)3 DSC melting transition (C) TGA On-set decomposition Temperature (C) TGA Peak decomposition Temperature (C) Weight loss (%) Undoped 99 125 143 7.8 0.5 98 112 131 7.3 1.0 98 110 131 7.7 1.5 98 107 123 8.0 2.0 98 108 130 7.0 2.5 98 107 123 7.2 3.0 98 108 129 7.1 Figure 8-5. Simultaneous DSC and TGA profiles of LiMn(BH4)3 doped with various nanocatalysts (nanoNi, nanoCo, nanoFe, na noCu, nanoTi and nanoZn) by fixing the concentration X=1.5mol% and ball milled for 20 minutes

PAGE 148

132 The thermal desorption spectral (TPD) anal ysis shown in Figure 8-6 compliments the TGA measurements of Figur e 8-5. Nano-Ni doped LiMn(BH4)3 exhibits earlier decomposition temperature with larger hydr ogen release (obtaine d from the area under the curve) than the nano-Co doped one. Overall, the simultaneous TGA, DSC and TPD analyses of undoped and doped LiMn(BH4)3 reveal that nano-Ni doping with a concentration of 1.5mol% is the optimum sy stem for effective hydrogen decomposition at lower temperatures. Figure 8-6. Thermal Programmed Desorption (T PD) profiles of undoped and Xmol% nanoNi and nanoCo doped LiMn(BH4)3; (X=1.5 mol%)

PAGE 149

133 8.5.3. Dehydrogenation kinetics of undope d and nano-Ni doped LiMn(BH4)3 It is unambiguously seen from the TGA and TPD analyses th at nano-Ni doping of 1-2 mol% destabilizes the structure with at least 10-20oC reduction in hydrogen thermal decomposition temperature (Tdec). Moreover, the nano-Ni has greater effect on the hydrogen release (desorption kinetics) rate when compared to the undoped LiMn(BH4)3 as observed in Figure 8-7. The optimum c oncentration of nano-Ni was found to be 1-2 mol% which can speed up the initial decompos ition reaction at least 2 to 3 times when compared to the undoped counterpart. Figure 8-7. Dehydrogenaiton kinetics of undoped an d Xmol% nanoNi (X=0.5, 1.0, 1.5, 2.0, and 2.5) doped LiMn(BH4)3

PAGE 150

134 At lower concentration (1mol%) of nano-Ni, at least 3 to 4 fold increases in the rate of desorption is clearly seen as evid enced from Figure 8-7. A complete desorption from the nano-Ni (1 mol%) sample occurr ed within 40 minutes as compared to 70 minutes for the undoped sample. For the case of higher doping concentrations of nanocatalysts (1.5, 2 and 2.5 mol%), we see greater improvement in the desorption kinetics behavior when compared to the undoped and 1 mol% nano-Ni doped LiMn(BH4)3. It is also notew orthy to point out that the four -fold increase in the rate of desorption is still seen in these higher concentrations with the total hydrogen decomposition time found to be lowered from 40 to 20 minutes. Overall, the desorption kinetics profiles seem to agree well with the thermogravimetric analysis represented in Figures 8-4 and 8-5. Hence, it is conclude d that the optimum dopant concentrations of 1.5-2.5 mol% are necessary for desorpti on-rate enhancement and lowering of the hydrogen decomposition temperature (also, refer to Table 8-1). 8.5.4. Activation energy calculations of undoped and nano-Ni doped LiMn(BH4)3 Based on Kissinger’s theory and the appr oach discussed in Section 8.4, the thermal programmed desorption profiles of LiMn(BH4)3 at three different ramping rates of 4, 10 and 20 oC/min were collected The TPD sp ectra of undoped and nano-Ni doped LiMn(BH4)3 at these ramping temperatures are s hown in Figure 8-8. The following can be seen in Figure 8-8: (i) TPD signal strength (corres ponding to the hydrogen content) depends on the amount of sample used, (ii) uniform shift in Tmax ( temperature at the maximum desorption rate) with increasing th e ramping rate in the order of 4>10>20 oC/min, (iii) shifting of the peak position i ndependent of whether the sample is undoped

PAGE 151

135 or nano-Ni doped and (iv) reduction in the decomposition temperature at the ramping rate of 4 oC/min in comparison to the 10 and 20 oC/min rates. Figure 8-8. TPD spectra of undoped and 1.5mol% nanoNi doped LiMn(BH4)3 at various ramping rates (4, 10 and 20 oC/min) The activation energy for H2 desorption was calcula ted using the Kissinger analysis (see Figure 8-9) for bot h the undoped and nanoNi doped LiMn(BH4)3 systems from the TPD data. The slope of the straight line plot of ln( /Tmax 2) vs 1/Tmax yields the activation energy (Ea), which is a crucial parameter that needs to be optimized and investigated for efficient hydroge n storage. It is clearly se en from Figure 8-9 that the activation energy for the undoped LiMn(BH4)3 is 130.64 kJ/mol, whereas for the nanodoped samples, Ea can be lowered by at least 20 kJ/mole (for nano-Ni doped LiMn(BH4)3

PAGE 152

136 Ea = 111.55 kJ/mole). Due to this loweri ng of the activation energy of nano-doped samples, about 20-30 oC reduction in the decomposition temperature was obtained from the thermogravimetric measurements. Figure 8-9. Kissinger’s plot obtained from the TPD data for the undoped and 1.5 mol% nanoNi doped LiMn(BH4)3 8.5.5. GC analysis of undoped and nano-Ni doped LiMn(BH4)3 Figure 8-10 shows the GC analysis sp ectra of both undoped and 1.5 mol% nanoNi doped LiMn(BH4)3. From this figure, it is esti mated that no gas other than H2 desorbed upon repeated sampling. Since the GC measurements are not capable of quantifying the amount of hydrogen desorbed during TPD process, additional Mass-spec analyses are required, which ar e in progress. Unlike Zn(BH4)2, the new complex

PAGE 153

137 hydrides LiMn(BH4)3 releases a lot less or no bor ane family of gases during the decomposition process. Figure 8-10. Gas Chromatogr aphy analysis of undoped and 1.5mol% nanoNi doped LiMn(BH4)3 8.5.6. Possible mechanism of nano Ni dopi ng on the complex hydride LiMn(BH4)3 The enhancement of reaction kinetics at low temperatures and the requirement for high hydrogen storage capacity (> 6.0 wt.% ) of complex borohydrides could be made possible by either adopting destabilization strategies or nanomaterial doping. If nanostructured materials with high surface area are used as the dopants, they may offer several advantages for the physic o-chemical reactions, such as (i) surface interactions, (ii)

PAGE 154

138 adsorption in addition to bulk absorption, (i ii) rapid kinetics, (i v) low temperature sorption, (v) hydrogen atom dissociation and mo lecular diffusion via the surface catalyst. The intrinsically large surface areas and unique adsorbing properties of nanophase dopants can assist the dissociation of gase ous hydrogen molecules and the small volume of individual nanoparticles can pr oduce short diffusion paths to the materials’ interiors. The use of nanosized dopants enables a higher dispersion of the catalytically active species166 and thus facilitates mass transfer reac tions. Based on our pr evious studies on the nickel doped Zn(BH4)2 88, 94, it is easily discernible th at by nanomaterial doping, both the reduction of decomposition temperature and the cohesive energy of the complex hydrides are established. Hence, the hydroge n transfer reactions and breaking of B-H bonds have been facilitated by the Nidopant s. However, further experimental and theoretical studies to determine the exact mechanism of nanomaterials doped complex hydrides are necessary. 8.6. Summary In this work, an inexpensive mechano-chemical approach of ball milling technique was utilized to prepare a member of a new class of solvent-free, solid-state complex borohydrides (Li-Mn-B-H) for on-board hydrogen storage. It is found that the endothermic transition due to hydrogen or other gaseous decomposition from the Li-MnB-H system occurs with onset below 100oC and a complete decomposition occur at 150 oC. To reduce the decomposition temperat ure further, we attempted to dope the LiMn(BH4)3 system with different molar concentr ations of nanodopants such as nano-Ni, nano-Co, and nano-Fe. Thermo gravimetric (T GA) and desorption kinetic profiles of the undoped and nano-doped Li-Mn-B-H system show that the nanodopant materials have

PAGE 155

139 pronounced effects on the hydrogen release ki netics while lowering the decomposition temperature. Moreover, the nano-doped LiMn(BH4)3 exhibits lower activation energy (112 kJ/mol) by 20 kJ/mol in comparison to the undoped sample (131 kJ/mol). Though this Li-Mn-B-H complex borohydride exhibits high theoretical hydrogen storage capacity (8-10wt.%) at lower temperature (<150oC), the reversible hydrogenation and dehydrogenation cycling is not promising. This may be due to either the strong B-H interaction or the formation of MnB2 structure. Further invest igations are required using various destabilization mechanisms and strategi es for these new materials to evaluate the reversible hydrogen kinetics a nd storage capacity which is the further recommendation for future work.

PAGE 156

140 Chapter 9 Summary, conclu sions and recommendations 9.1. Overview The challenge is clear and fascinati ng: supplying abundant amounts of clean energy; consuming less natural resources a nd finding the appropriate solutions for any corner of the world. Both f undamental theoretical and experime ntal research is needed to understand the interactio n of hydrogen in solid-state materi als to realize the potential of these materials for hydrogen storage. In this dissertation, we have used combined-first-principles using density functional theory (DFT) and direct method lattice dynamics calculations to understand the properties for hydrogen st orage materials, mainly focused on complex borohydrides. This allowed us to establish stability of the crystal structure at fini te temperatures using the positive definite frequency distribution cr iteria of the phonon spectra over the Brillion zone. DFT was used to calculate electroni c properties and the di rect method lattice dynamics was used to calculate the finite te mperature thermal properties. The calculated thermodynamic functions: the enthalpy and the entropy are in good agreement with the experimental values for solid (within the ac ceptable limit of experimental values) crystal lattices. The Gibbs energy of crystal solid can then be predicted by usual thermodynamic relation between enthalpy, entropy and Gibbs energy. For gas molecules (H2 and N2), one needs to know the Gibbs energy at the sta ndard pressure of 1 atm. The Gibbs

PAGE 157

141 energy of a gas molecule was calculate d by combining both computational and experimental results as described earlier in Ch apter 6. Once the temperature dependent Gibbs energy and entropy are predicted fo r each element or co mpound (reactants and products), the reaction thermodynamics can be predicted easily. A number of borohydride systems have been studied using combined first principles and lattice dynamics simulations including for instance zinc borohydride, manganese borohydride and multinary borohydride system. At last, a systematic study has also been carried out for a new type of complex borohydride i.e. lithium manganese borohydride system by various experimental technique s. Nano additives were al so added to this complex borohydride to study the effect of nano additiv es and found that the nanodopant materials have destabilized the system to allow for hydrogen storage at lower temperatures and with faster kinetics. 9.2. Zn(BH4)2 – conclusions and recommendations It is found that Zn(BH4)2 is an insulating material having a wide band gap. Electronic structure calculations show strong bonding between hydrogen atoms and boron in the [BH4]complex and also less polar bonding between the Zn and the hydrogen atom. The simulated standard enthalpy of dehydrogenation of Zn(BH4)2 suggests that decomposition to primary elements is the most favorable one. The findings of our work are in qualitative agreement with e xperimental results. It is also found that Ni doping improves the dehydrogenation mech anism of this complex borohydride by reducing both the decompos ition temperature and cohesi ve energy of the complex borohydride. Ni has greater affinity towards H atoms a nd these H atoms are pulled towards these doped Ni atoms and thus affect s both low (librational) and high (bending)

PAGE 158

142 frequency modes of the nearest BH4 complex, which introduces instability via the breaking of multiple B-H bonds in the complex borohydride. However, experimentally it was reported that thermal decomposition of Zn(BH4)2 comprises not only of the evolution of H2, but also production of an appreciable amount of B–H (borane) compounds. Although Ni dopant lowers the decomposition te mperature but still lead to negligible release of boranes and thus, Zn(BH4)2 cannot be considered as suitable hydrogen storage material. 9.3. Mn(BH4)2 – conclusions and recommendations M anganese borohydride (Mn(BH4)2) is a high theoretical capacity ( 10 wt%) solid-state hydrogen storage candidate. Mn(BH4)2 is found to be tetra gonal type structure of space group I-4m2 The most feasible dehydrogenati on reaction is found to be an endothermic reaction at decomposition temperature. The spin polarized electronic density of states studies reveal that Mn partial 3 d DOS has metallic character for spin-up and insulating character for spindown pr ojection, are supposed to show some intermediate behavior, which suggests that the Mn(BH4)2 is a half-metallic hydrogen storage material. The electronic struct ure analysis shows that the interaction between Mn atoms and BH4 complexes has an ionic characte r while the internal bonding of BH4 is essentially covalent. The theoreti cal vibrational characterization of manganese borohydride is also well compared with experimental results. The dehydrogenation thermodynamics suggest that the crystal struct ure is stable at fin ite temperature and no diborane as a dehydrogenation pr oduct is preferred thermodynamically. Thus, complex borohydride of manganese could be consid ered a potential candidate for hydrogen

PAGE 159

143 storage. However, one needs to address the reversibility issue before its practical onboard applications. 9.4. Li-Mg-B-N-H – conclusions and recommendations In this work, we have pr edicted the stabil ity and dehydrogenation reactions at different temperatures of the reversib le complex multinary borohydride system from theoretical calculations. The stability of the structures involved in a hydrogen desorption pathway is validated via phonon calculations. The thermodynamic properties of these compounds are then calculated and the feasibil ity of the mechanism is validated. It is found that all samples are intimate mixtures of Li4BN3H10 with MgH2 and that no new chemical compound is formed during the ball mill ing. This information is verified from both vibrational spectroscopy (B-H, Mg-H and N-H bond stre tches; Figure 7-1) and Gibbs energy calculations (Figure 7-8). The reaction enthalpy calculations of the hy drogen release react ions (Figure 7-9) show that the complete first step hydrogen release is endothe rmic and the second step is exothermic. The total hydrogen storage cap acity of this Li-Mg-B-N-H system is calculated to be 8.76 wt% from the deso rption behavior observed at different temperatures up to 350 oC. Theoretical calculations also show that the multinary complex borohydride exhibits first hydrogen release at a temperature of about 95 oC and then the main reversible hydrogen adsorption/ desorption reaction (re action (7-3)) occurs at a temperature of about 225 oC, which is a little higher than the experimentally determined temperature (190 oC)136. The quantitative reve rsible hydrogen storage capacity is found to be 4.38 wt% which is also different from the experimentally determined values of 2.5 wt% at 180 oC128 (5-6 wt% at around 160 175 oC by adopting

PAGE 160

144 different processing reaction pathway scheme s in our clean energy research group at University of South Florida). The enthalpy of reversible re/de-hydrogenation of Li-Mg-B-N-H multinary system is found to be 55.17 kJ/mol of H2 which yields moderate H2 pressure at elevated temperature only (Figure 7-11). This sugge sts that the multinary system release H2 at higher temperatures, which is not the best situation for ideal fuel cell operation for onboard applications. However, this multinary system is the most promising reversible candidate compare to all ot her complex hydrides studied to-date. The reversible dehydrogenation temperature, kine tics and also storage capac ity can be further improved via catalytic doping, which is the further recommendation for future work. 9.5. Li-Mn-B-H – conclusions and recommendations Complex borohydride of manganese system member of a new class of complex borohydrides for hydrogen storage can be prepared with a 3:1 ratio of precursor materials LiBH4 and MnCl2 via the solid state mechano-chemical process. The occurrence of B-H stretch at three different frequency bands from the F ourier transform infrared spectrometer (FTIR) investigation confirm the formation of LiMn(BH4)3 at room temperature. It is found that the XRD an alysis of milled sample shows only LiCl (product) peaks. Since LiMn(BH4)3 is not of highly crystallin e nature, the direct XRD peaks are not distinct for Li-Mn-B-H syst em. Nevertheless, both the FTIR and XRD spectra indirectly confirm the forma tion of the new complex hydride LiMn(BH4)3 from the reaction mixtures of 3LiBH4 and MnCl2. It is also found that a large amount of hydrogen ( 8.0 wt %) can be released below 150 C in a single dehydrogenation reaction step from the thermo gravimet ric analysis (TGA) of LiMn(BH4)3. Furthermore, a

PAGE 161

145 significant reduction in the decomposition temperature can be achieved via catalytic doping of small fractions of nano-Ni to LiMn(BH4)3 system. The doped LiMn(BH4)3 exhibits lower activation ener gy by ~20 kJ/mol in compar ison to the undoped sample. Thermo gravimetric (TGA) and desorpti on kinetic profiles of the undoped and nanomaterials doped Li-Mn-B-H system show th at the nanodopant materials have pronounced effects on the hydrogen release kinetics while lowering the decomposition temperature. Moreover, the gas chromatography st udies of the undoped and doped LiMn(BH4)3 demonstrate that the evolved gas is mainly hydrogen and does not contain members of the borane family. Though this Li-Mn-B-H complex borohydride exhibits high theoretical hydrogen storage capacity (8-10 wt%) at lower temperature (<150oC), the reversible hydrogenation and dehydrogenation cycling is not promising. This may be due to either the strong B-H interaction or the formation of MnB2 structure. Further invest igations are required using various destabilization mechanisms and strategi es for these new materials to evaluate the reversible hydrogen kinetics a nd storage capacity which is the further recommendation for future work. 9.6. Major contributions The contributions of this dissertation to the field of hydrogen storage research are multifold. It develops a fundamental unders tanding of the hydrogen storage materials mainly focused on complex borohydrides. This study finds the stable crystal structure, electronic structure and nature of chem ical bondings of new borohydride complexes, identifies the favorable de hydrogenation reaction and asso ciated thermodynamics, and also unravels the local hydrogen dynamics and storage capacity of the complex

PAGE 162

146 borohydride with additive materials. Finally this type of combined DFT and lattice dynamics study could supplement experiments in testing complex multinary hydrides for hydrogen storage properties. These calculations ar e helpful in identifying feasible reactions and to rule out unreal istic possibilities, and arrive at practical hydrogen storage materials. The fundamental th eoretical simulations in this research will not only benefit the design of novel solid state borohydride mate rials, but also establish thermodynamic "guidelines" which are used to help gui de reactions and design other suitable nanostructured materials. The broader impact of the research is to prepare and gain fundamental insights for new hydrogen storage materials which can lead to the design of compact, lightweight, responsive, and affordable hydrogen st orage materials which can have 5-10 kg (depending on the size of the vehicle) of usable hydr ogen which can enable 480–km (300-miles) driving range in a single fueli ng. Improved physico-chemical reactions of hydrogen with solid state complexes will also have broad impact on our society to develop zero emission fuel cell vehicles, to mitigate the global warming effects and to supply inexpensive and plentiful clean ener gy for our current st andard of living. An improvement in the fundamental unders tanding of the therma l characteristics and adsorption/desorption dynami cs of gases on these complex hydrides would contribute to the research studies being carried in catal ysis. The modeling techniques used in the present study can be extended to a wide range of other systems. 9.7. Future work directions Kinetics is emerging as the most important challenge for theory because of the time scale of the hydrogen evolution. The general time scale for quantum mechanical

PAGE 163

147 simulations to evaluate atomic forces is in the order of picoseconds but simulations using empirical and semi-empirical potential ener gy functions can go up to the order of nanoseconds. Mesoscale simulations, such as kinetic Monte Carlo, can cover up to the time scales in seconds, and continuum met hods can cover up to the time scales from seconds to hours. As the time scales for the hydrogen evolutions is in the order of seconds to hour and thus poses serious challeng es for theory and computation. The real challenge is to integrate these approaches so that predictions of r eal materials behavior can be made with a solid phys ical and chemical basis. Hydrogen storage materials should be studied not only from the catalytic chemistry point of view, but also from the microstructural evolution point of view. For the experimental work, the preliminary results of PCT experiments for the destabilization of LiMn(BH4)3 by few mol % of MgH2 shows that the system becomes reversible at least by 3.0 wt% including the dead weight ( i.e. LiCl) at low temperature at around 100 oC. The detail study needs to be done further to understand the exact mechanism of the system and also to find the optimum quantity of MgH2 for the higher reversible capacity. Moreover, synthesis of the LiMn(BH4)3 system should be done by solution synthesis method to make single crystal to find the exact structural parameters of the complex borohydride. This method can eradicate the dead weight ( i.e. LiCl) present in the system during the synthesis of th e complex borohydride from its precursor materials using mechanochemical process. Otherwise a novel separa tion technique needs to be developed to extract the dead wei ght from the system after mechanochemical process is further recommendation for future work.

PAGE 164

148 References 1. Dresselhaus, M. S.; Thomas, I. L., Alternative energy technologies. Nature 2001, 414, 332-337. 2. Fontes, E.; Nilsson, E., Modeling the Fuel Cell. The Industrial Physicist 2001, 1417. 3. Zttel, A., Hydrogen storage methods. Naturwissenschaften 2004, 91, 157-172. 4. Dillion, A. C.; Jones, K. M.; Bekkedahl T. A.; Kiang, C. H.; Bethune, D. S.; Heben, M. J., Storage of hydrogen in single-walled carbon nanotubes. Nature 1997, 386, 377-379. 5. Rosi, N. L.; Eckert, J.; Eddaoudi, M.; V odak, D. T.; Kim, J.; O’Keeffe, M.; O.M. Yaghi, Hydrogen Storage in Micro porous Metal-Organic Frameworks. Science 2003, 300, 1127-1129. 6. Cho, S. J.; Song, K. S.; Kim, J. W.; Kim, T. H.; Choo, K., Fuel Chem Div Prepr. 2002, 47, 790. 7. Johanson, E. Synthesis and charact erisation of potential hydrogen storage materials. Ph.D., University of Uppsala, 2001. 8. Hohenberg, P.; Kohn, W., Inhomogeneous electron gas. Physical Review 1964, 136, B864–71. 9. Kohn, W.; Sham, L. J., Self-consis tent equations including exchange and correlation effects. Physical Review 1965, 140, A1133–38. 10. Jones, R. O.; Gunnarsson, O., The dens ity functional formalism, its applications and prospects. Rev. Mod. Phys. 1989, 61, (3), 689-746. 11. Perdew, J. P.; Wang, Y., Correlation hol e of the spin-polariz ed electrongas, with exact small-wave-vector and high-density scaling. Physical Review B 1991, 44, 13298–13307. 12. Becke, A. D., Density-functional th ermochemistry III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648.

PAGE 165

149 13. Burke, K.; Ernzerhof, M.; Perdew, J. P., The adiabatic connection method: a nonempirical hybrid. Chem. Phys. Lett. 1997, 265, 115-120. 14. Wigner, E. P., The transition-state method. Transactions of th e Faraday Society 1938, 34, 29–41. 15. Ceperley, D. M., Ground-state of the fermion one-component plasma montecarlo study in 2 and 3 dimensions. Physical Review B 1978, 18, 3126–3138. 16. Ceperley, D. M.; Alder, B. J., Groundstate of the electrongas by a stochastic method. Phys. Rev. Lett. 1980, 45, 566–569. 17. Becke, A. D., Density-functional thermochemistry for a new dynamical correlation functional and implica tions for exact-exchange mixing. J. Chem. Phys. 1996, 104, 1040-1046. 18. Kresse, G.; Furthmller, J., Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 1996, 6, 15. 19. Kittel, C., Introduction to Solid State Physics Wiley: 1996. 20. Phillips, J. C., Energy-band interpol ation scheme based on a pseudopotential. Phys. Rev. 1958, 1, 685–695. 21. Antoncik, E., Approximate formulation of the orthogonalized plane-wave method. J. Phys. Chem. Solids 1959, 10, 314. 22. http://en.wikipedia.org/wiki/Pseudopotential 23. Vanderbilt, D., Soft self-consistent pseudopotentials in generalized eigenvalue formalism. Phys. Rev. B 1990, 41, (11), 7892. 24. Blochl, P. E., Generalized separa ble potentials for el ectronic-structure calculations. Phys. Rev. B 1990, 41, 5414-5416. 25. Blchl, P. E., Projector augmented-wave method. Phys. Rev. B 1994, 50, (24), 17953. 26. Kresse, G.; Hafner, J., Ab initio molecular dynamics for liquid metals. Phys. Rev. B 1993, 47, (1), 558. 27. Kresse, G.; Furthmller, J., Efficient ite rative schemes for ab initio total-energy calculations using a pl ane-wave basis set. Phys. Rev. B 1996, 54, (16), 11169. 28. Baroni, S.; Gironcoli, S. d.; Giannozzi, P., Phonon dispersions in GaxAl1-xAs alloys. Phys. Rev. Lett. 1990, 65, (1), 84.

PAGE 166

150 29. Baroni, S.; Giannozzi, P.; Testa, A., Gr een's-function approach to linear response in solids. Phys. Rev. Lett. 1987, 58, (18), 1861. 30. Wei, S.; Chou, M. Y., Phonon dispersi ons of silicon and germanium from firstprinciples calculations. Phys. Rev. B 1994, 50, (4), 2221. 31. Frank, W.; Elssser, C.; Fhnle, M., Ab initio Force-Constant Method for Phonon Dispersions in Alkali Metals. Phys. Rev. Lett. 1995, 74, (10), 1791. 32. Parlinski, K.; Li, Z. Q.; Kawazoe, Y., First-Principles Determination of the Softmode in Cubic ZrO2. Phys. Rev. Lett. 1997, 78, (21), 4063. 33. Parlinski, K., Lattice dynamics of cubic BN. J. Alloys Compd. 2001, 328, (1-2), 97-99. 34. Parlinski, K.; Parlinska-Wojtan, M., Lattice dynamics of NiTi austenite, martensite, and R phase. Phys. Rev. B 2002, 66, (6), 064307. 35. Parlinski, K. PHONON 2005. 36. TA instrument website. “S DT tutorial,” online posting, www.tainstruments.com 37. Quantachrome instruments web site. “Autosorb1-C Chemisorption Physisorption Analyzer manual,” online posting, www.quantachrome.com 38. Bruce M. Clemens (2006). “GCEP t echnical report 2 006,” online posting, http://gcep.stanford.edu 39. Keck Interdisciplinary Surface Scie nce Center (2005). “Fourier Transform Infrared Spectroscopy (FT-IR),” online posting, www.nuance.northwestern.edu 40. www.infohost.nmt.edu~maximino/fesem_types_of_signals.htm 41. ETH Zrich website. “SEM technology,” online posting, www.microscopy.ethz.ch/sem.htm 42. Miwa, K.; Ohba, N.; Towata, S.-i.; Naka mori, Y.; Zttel, A.; Orimo, S.-i., Firstprinciples study on thermodynamical stab ility of metal borohydrides: Aluminum borohydride Al(BH4)3. J. Alloys Compd. 2007, 446-447, 310-314. 43. Miwa, K.; Aoki, M.; Noritake, T.; Ohba N.; Nakamori, Y.; Towata, S.; Zu¨ttel, A.; Orimo, S., Thermodynamical st ability of calcum borohydride Ca(BH4)2. Phys. Rev. B 2006, 74, 155122. 44. Zu¨ttel, A.; Rentsch, S.; Fischer, P.; Wenger, P.; Sudan, P.; Mauron, P.; Emmenegger, C., H ydrogen storage properties of LiBH4. J. Alloys Compd. 2003, 356-357, 515-520.

PAGE 167

151 45. Lodziana, Z.; Vegge, T., Structural Stability of Complex Hydrides: LiBH4 Revisited. phys. Rev. Lett. 2004, 93, (14), 145501. 46. Nakamori, Y.; Miwa, K.; Ninomiya, A.; Li, H.; Ohba, N.; Towata, S.-i.; Zttel, A.; Orimo, S.-i., Correlation between thermodynamical stabilities of metal borohydrides and cation electronegativites: First-principles calculations and experiments. Phys. Rev. B 2006, 74, 045126. 47. Grochala, W.; Edwards, P. P., Ther mal Decomposition of the Non-Interstitial Hydrides for the Storage and Production of Hydrogen. Chem. Rev. 2004, 104, (3), 1283-1315. 48. Jeon, E.; Cho, Y., Mechanochemical s ynthesis and thermal decomposition of zinc borohydride. J. Alloys Compd. 2006, 422, 273–275. 49. Hafner, J.; Wolverton, C.; Ceder, G ., Toward Computational Materials Design: The Impact of Density Functional Ca lculations on Materials Science. MRS Bulletin 2006, 31, 659. 50. Herbst, J. F.; Hector, J. L. G., El ectronic structure and energetics of the quaternary hydride Li4BN3H10. Appl. Phys. Lett. 2006, 88, (23), 231904-3. 51. Majzoub, E. H.; McCarty, K. F.; Ozolinš, V., Lattice dynamics of NaAlH4 from high-temperature single-crystal Raman scattering and ab initio calculations: Evidence of highly stable AlH4 anions. Phys. Rev. B 2005, 71, 024118. 52. iguez, J.; Yildirim, T. ; Udovic, T. J.; Sulic, M.; Je nsen, C. M., Structure and hydrogen dynamics of pure and Ti-doped sodium alanate. Phys. Rev. B 2004, 70, (6), 060101. 53. Marashdeh, A.; Olsen, R. A.; Lvvik, O. M.; Kroes, G.-J., A density functional theory study of Ti-doped NaAlH4 clusters. Chem. Phys. Lett. 2006, 426, (1-3), 180-186. 54. Vegge, T., Equilibrium structure and Ti-catalyzed H2 desorption in NaAlH4 nanoparticles from density functional theory. Phys. Chem. Chem. Phys. 2006, 8, 4853. 55. Zarkevich, N. A.; Johnson, D. D., St ructural Stability of Complex Hydrides: LiBH4 Revisited. Phys. Rev. Lett. 2006, 97, 119601. 56. Vajeeston, P.; Ravindran, P.; Kjeksh us, A.; Fjellvag, H., High hydrogen content complex hydrides: A density-functional study. Appl. Phys. Lett. 2006, 89, 071906. 57. Zeng, Q.; Su, K.; Zhang, L.; Xu, Y.; Cheng, L.; Yan, X., Evaluation of the Thermodynamic Data of CH3SiCl3 Based on Quantum Chemistry Calculations. J. Phys. Chem. Ref. Data 2006, 35, (3), 1385.

PAGE 168

152 58. Alapati, S. V.; Johnson, J. K.; Sholl, D. S., Identification of Destabilized Metal Hydrides for Hydrogen Storage Usi ng First Principles Calculations. J. Phys. Chem. B 2006, 110, 8769-8776. 59. Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C., Atoms, molecule s, solids, and surfaces: Applications of the generalized gradient approxima tion for exchange and correlation. Phys. Rev. B 1992, 46, (11), 6671. 60. Kresse, G.; Joubert, D., From ultr asoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 1999, 59, (3), 1758. 61. Monkhorst, H. J.; Pack, J. D., Special points for Brillouin-zone integrations. Phys. Rev. B 1976, 13, (12), 5188. 62. http://www.mkmc.dk/crystal/194/Mg/Zn/main.html Zn. 63. Park, C. H.; Chadi, D. J., First-princi ples study of the atomic reconstrcutions of ZnSe(100) surfaces. Phys. Rev. B 1994, 49, 16467. 64. http://www.webelements.com/webelement s/properties/text/d efinitions/crystalstructure.html 65. vanSetten, M. J.; deWijs, G. A., Ab initio study of Mg(AlH4)2. Phys. Rev. B 2005, 72, 073107. 66. http://cst-www.nrl.navy.mil/lattice/struk.xmol/alphaB.pos Boron. 67. Prasad, D. L. V. K.; Balakrishnarajan, M. M.; Jemmis, E. D., Electronic structure and bonding of -rhombohedral boron using clus ter fragment approach. Phys. Rev. B 2005, 72, 195102. 68. Pistorius, C. W. F. T., Z. Phys. Chem., Neue Folge 1974, 88, 253. 69. Lvvik, O. M., Crys tal structure of Ca(AlH4)2 predicted from density-functional band-structure calculations. Phys. Rev. B 2005, 71, 144111. 70. Wolverton, C.; Ozolins, V., Hydrogen stor age in calcium alanat e: First-principles thermodynamics and crystal structures. Phys. Rev. B 2007, 75, 064101. 71. Lvvik, O. M.; Molin, P. N., Density -functional band-struct ure calculations of magnesium alanate Mg(AlH4)2. Phys. Rev. B 2005, 72, 073201. 72. Kitajima, N.; Shimanouchi, H.; Ono, Y.; Sasada, Y., The crystal and Molecular Structure of Cu(AlCl4)2. Bull. Chem.Soc. Jpn. 1982, 55, (7), 2064 2067. 73. Marynick, D. S.; Lipsco mb, W. N., Crystal Structur e of Beryllium Borohydride. Inorg. Chem. 1972, 11, (4), 820.

PAGE 169

153 74. http://cms.mpi.univie.ac.at/vasp-workshop/ 75. Miwa, K.; Ohba, N.; Towata, S.-i.; Naka mori, Y.; Orimo, S.-i., First-principles study on lithium borohydride LiBH4. Phys. Rev. B 2004, 69, 245120. 76. Becke, A. D.; Edgecombe, K. E., A si mple measure of electron localization in atomic and molecular systems. J. Chem. Phys. 1990, 92, (9), 5397-5403. 77. Silvi, B.; Savin, A., Classificatio n of chemical bonds based on topological analysis of electron localization functions. Nature 1994, 371, (6499), 683-686. 78. Borelius, G., Zero-point Energy of El ectrons and Phonons in the Periodic System. Phys. Scr. 1978, 18, 19-12. 79. Siegel, D. J.; Wolverton, C., Reac tion energetics and cr ystal structure of Li4BN3H10 from first principles. Phys. Rev. B 2007, 75, 014101. 80. Hu, S.-W.; Wang, Y.; Wang, X.-Y.; Chu, T.-W.; Liu, X.-Q., Gas-Phase Reactions between Diborane and Carbon Mo noxide: A Theoretical Study. J. Phys. Chem. A 2003, 107, 9974-9983. 81. Greene, T. M.; Brown, W.; Andrews, L.; Downs, A. J.; Chertihin, G. V.; Runeberg, N.; Pyykko, P., Matrix Infrared Spectroscopic and ab Initio Studies of ZnH2, CdH2, and Related Metal Hydride Species. J. Phy. Chem. 1995, 99, 7925. 82. http://en.wikipedia.org/wiki/Diborane 83. Araujo, C. M.; Ahuja, R.; Guillen, J. M. O.; Jena, P., Role of titanium in hydrogen desorption in crystalline sodium alanate. Appl. Phys. Lett. 2005, 86, 251913. 84. Lee, E.-K.; Cho, Y. W.; Yoon, J. K., Ab -initio calculations of titanium solubility in NaAlH4 and Na3AlH6. J. Alloys Compd. 2006, 416, (1-2), 245-249. 85. Blomqvist, A.; Araujo, C. M.; Jena P.; Ahuja, R., Dehydrogenation from 3 d transition-metal-doped NaAlH4: Prediction of catalysts. Appl. Phys. Lett. 2007, 90, 141904. 86. Alapati, S. V.; Johnson, J. K.; Sholl, D. S., Stability analys is of doped materials for reversible hydrogen storage in destabilized metal hydrides. Phys. Rev. B 2007, 76, 104108. 87. Srinivasan, S.; Escobar, D.; Jur czyk, M.; Choudhury, P.; Goswami, Y.; Stefanakos, E., Effects of nanocatalys t doping on the hydrogen storage behavior of new complex borohydrides. IHEC '07 Proc. 2007. 88. Srinivasan, S.; Escobar, D.; Jurczyk, M.; Goswami, Y.; Stefanakos, E., Nanocatalyst doping of Zn(BH4)2 for on-board hydrogen storage. J. Alloys Compd. 2008, 462, (1-2), 294-302.

PAGE 170

154 89. Schlapbach, L.; Zuttel, A., Hydrogen-st orage materials for mobile applications. Nature 2001, 414, (6861), 353-358. 90. Bogdanovi, B.; Schwickardi, M., Tidoped alkali metal alum inium hydrides as potential novel reversible hydrogen storage materials. J. Alloys Compd. 1997, 253-254, 1-9. 91. Jensen, C. M.; Zidan, R. A., Hydrogen storage materials and. method of making by dry homogenation. U. S. Patent 2002, 6, 471935. 92. Frankcombe, T. J.; Kroes, G.-J., Qu asiharmonic approximation applied to LiBH4 and its decomposition products. Phys. Rev. B 2006, 73, 174302. 93. Choudhury, P.; Bhethanabotla, V. R.; Stef anakos, E., Identific ation of a stable phase for the high-capacity hyd rogen-storage material Zn(BH4)2 from density functional theory and lattice dynamics. Phys. Rev. B 2008, 77, (13), 134302-9. 94. Choudhury, P.; Bhethanabotla, V. R.; St efanakos, E., Ni-induced destabilization dynamics of crystalline zinc borohydride. Appl. Phys. Lett. 2008, 92, (13), 134101-3. 95. Wang, X.; Andrews, L., Matrix Infrared Spectra and Density Functional Theory Calculations of Manganese and Rhenium Hydrides. J. Phys. Chem. A 2003, 107, (20), 4081-4091. 96. Vajeeston, P.; Ravindran, P.; Ravi, C.; Asokamani, R., Electronic structure, bonding, and ground-state properties of AlB2-type transition-metal diborides. Phys. Rev. B 2001, 63, (4), 045115. 97. Hobbs, D.; Hafner, J.; Spisak, D., U nderstanding the complex metallic element Mn. I. Crystalline and noncollinear magnetic structure of alpha -Mn. Phys. Rev. B 2003, 68, (1), 014407. 98. Choudhury, P.; Srinivasan, S.; Bhethana botla, V. R.; Stefanakos, E., Transition metal complex hydrides for vehicular application. FLAVS Meeting, Orlando, FL 2008. 99. Fossdal, A.; Brinks, H. W.; Fichtner, M.; Haubacka, B. C., Determination of the crystal structure of Mg(AlH4)2 by combined X-ray and neutron diffraction. J. Alloys Compd. 2005, 387, 47-51. 100. Fichtner, M.; Engel, J.; Fuhr, O.; Gl oss, A.; Rubner, O.; Ahlrichs, R., The Structure of Magnesium Alanate. Inorg. Chem. 2003, 42, 7060 7066. 101. Fichtner, M.; Fuhr, O., S ynthesis and structures of magnesium alanate and two solvent adducts. J. Alloys Compd. 2002, 345, 286-296.

PAGE 171

155 102. Fichtner, M.; Fuhr, O.; Kircher, O ., M agnesium alanate—a material for reversible hydrogen storage? J. Alloys Compd. 2003, 356-357, 418-422. 103. Einarsrud, M. A.; Justnes, H.; Rytter, E.; Oye, H. A., Structure and Stability of Solid and Molten Complexes in the MgCl2-AlCl3 System. Polyhedron 1987, 6, (5), 975 986. 104. Ozolins, V.; Majzoub, E. H.; Wolvert on, C., First-Principl es Prediction of a Ground State Crystal Structure of Magnesium Borohydride. Phys. Rev. Lett. 2008, 100, 135501. 105. Jorgensen, J. D.; Hu, Z.; Teslic, S.; Argyr iou, D. N.; Short, S.; Evans, J. S. O.; Sleight, A. W., Pressure-induced cubi c-to-orthorhombic phase transition in ZrW2O8. Phys. Rev. B 1999, 59, (1), 215 225. 106. Andersen, A. M. K.; Carlson, S., Hi gh-pressure structre of alpha and delta ZrMo2O8. Acta. Cryst. B 2000, 57, 20 26. 107. Justnes, A.; Rytter, E.; F.Andresen, A., A Crystal structur e determination of Titanium(II) Tetrachloroaluminate by Vibrational Correlation Methods and Powder Diffration. Polyhedron 1982, 1, (4), 393 396. 108. Staffel, T.; Meyer, G., Synthesi s and Crystal Structure of Cd[AlCl4]2 and Cd2[AlCl4]2. Z. anorg. allg. Chem. 1987, 548, 45 54. 109. Butz, A.; Miehe, G.; Paulus, H.; Strauss, P.; Fuess, H., The Crystal Structure of Mn(ReO4)2 2H2O and of the Anhydrous Perrhenates M(ReO4)2 of Divalent Manganese, Cobalt, Nickel, and Zinc. J. Solid State Chem. 1998, 138, 232 237. 110. Du, A. J.; Smith, S. C.; Lu, G. Q ., Role of charge in destabilizing AlH4 and BH4 complex anions for hydrogen storage appl ications: Ab initio density functional calculations. Phys. Rev. B 2006, 74, (19), 193405-4. 111. Chase, M. W. J., NIST-JANAF Themochemical Tables, Fourth Edition, J. Phys. Chem. Ref. Data, Monograph 9,1-1951 (1998) 112. Zhang, X.-B.; Shi, S.-Q.; Ke, X.-Z .; Han, S.; Shioyama, H.; Kuriyama, N.; Kobayashi, T.; Xu, Q., Dehydrogenation Reaction for Na-O-H System: A FirstPrinciples Study. ChemPhysChem 2007, 8, (13), 1979-1987. 113. Reuter, K.; Scheffler, M., Compositi on, structure, and stability of RuO2(110) as a function of oxygen pressure. Phys. Rev. B 2001, 65, (3), 035406. 114. Stull, D. R.; Scheffler, M., JANAF Thermochemical Tables, 2nd ed. (U.S. National Bureau of Standar ds, Washington, DC, 1971)

PAGE 172

156 115. Topor, L.; Kleppa, O. J., Enthalpies of formation of first-row transition-metal diborides by a new calorimetric method. J. Chem. Thermodyn. 1985, 17, (11), 1003-1016. 116. Donald, J. S.; Wolverton, C., Reac tion energetics and crystal structure of Li4BN3H10 from first principles. Phys. Rev. B 2007, 75, 014101. 117. Liu, Y.; Hu, J.; Wu, G.; Xiong, Z.; Chen, P., Formation and Equilibrium of Ammonia in the Mg(NH2)2-2LiH Hydrogen Storage System. J. Phys. Chem. C 2008, 112, 1293-1298. 118. Srinivasan, S.; Escobar, D.; Goswami, Y.; Stefanakos, E., Effects of catalysts doping on the thermal decomposition behavior of Zn(BH4)2. Int. J. Hydrogen Energy 2008, 33, 2268-2272. 119. Sudik, A.; Yang, J.; Halliday, D.; Wolv erton, C., Hydrogen Storage Properties in (LiNH2)2-LiBH4-(MgH2)X Mixtures (X = 0.0 -1.0). J. Phys. Chem. C 2008, 112, 4384-4390. 120. Satyapal, S.; Petrovic, J.; Read, C.; Thomas, G.; Ordaz, G., The U.S. Department of Energy's National Hydrogen Storag e Project: Progress Towards Meeting Hydrogen-Powered Vehicle Requirements. Catal. Today 2006, 120, 246. 121. Wolverton, C.; Siegel, D. J.; Akbarzad eh, A. R.; Ozolins, V., Discovery of novel hydrogen storage materials: an atom ic scale computational approach. J. Phys.: Condens. Matter 2008, 20, 064228 122. Akbarzadeh, A. R.; Ozolinš, V.; Wolvert on, C., First-Principles Determination of Multicomponent Hydride Phase Diagra ms: Application to the Li-Mg-N-H System. Adv. Mater. 2007, 19, 3233-3239. 123. Chen, P.; Xiong, Z.; Yang, L.; Wu, G.; Luo, W., Mechanistic Investigations on the Heterogeneous Solid-State Reacti on of Magnesium Amides and Lithium Hydrides. J. Phys. Chem. B 2006, 110, 14221-14225. 124. Leng, H.; Ichikawa, T.; Fujii, H., H ydrogen Storage Proper ties of Li-Mg-N-H Systems with Different Ratios of LiH/Mg(NH2)2. J. Phys. Chem. B 2006, 110, 12964-12968. 125. Wang, Y.; Chou, M. Y., First-pr inciples study of cation and hydrogen arrangements in the Li-Mg-N-H hydrogen storage system. Phys. Rev. B 2007, 76, 014116. 126. Vajo, J. J.; Skeith, S. L.; Mertens, F., Reversible Storage of Hydrogen in Destabilized LiBH4. J. Phy. Chem. B 2005, 109, 3719.

PAGE 173

157 127. Luo, S.; Flanagan, T. B.; Luo, W., The effect of exposure of the H-storage system (LiNH2 + MgH2) to water-saturated air. J. Alloys Compd. 2007, 440, (1-2), L13L17. 128. Sudik, A.; Yang, J.; Siegel, D. J.; Wolv erton, C.; III., R. O. C.; Drews, A. R., Impact of Stoichiometry on the Hy drogen Storage Properties of LiNH2-LiBH4MgH2 Ternary Composites. J. Phys. Chem. C 2009, 113, 2004-2013. 129. Perdew, J. P.; Burke, K.; Ernzerhof, M., Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, (18), 3865. 130. Barrera, G. D.; Colognesi, D.; Mitchell, P. C. H.; Ramirez-Cuesta, A. J., LDA or GGA? A Combined Experimental Inelas tic Neutron Scattering and Ab Initio Lattice Dynamics Study of Alkali Metal Hydrides. Chem. Phys. 2005, 317, 119. 131. Alapati, S. V.; Johnson, J. K.; Sholl, D. S., Using First Principles Calculation to Identify New Destabilized Metal Hydr ide Reactions for Reversible Hydrogen Storage. Phys. Chem. Chem. Phys. 2007, 8, 1438. 132. Choudhury, P.; Bhethanabotla, V. R.; St efanakos, E., Manganese borohydride as a hydrogen storage candidate: First-principl es crystal structur e and thermodynamic properties. J. Phy. Chem. C 2009, 113, 13416-13424. 133. Velikokhatnyi, O. I.; Kumta, P. N., Energetics of the lithium-magnesium imide– magnesium amide and lithium hydride reacti on for hydrogen storage: An ab initio study. Mater. Sci. Eng., B 2007, 140, 114-122. 134. The Inorganic Crystal St ructure Database (ICSD), http://www.fizinformationsdienste.de/en/DB/icsd/ 135. Herbst, J. F.; L. G. Hector, J., Ener getics of the Li amide/Li imide hydrogen storage reaction. Phys. Rev. B 2005, 72, 125120. 136. Niemann, M. U.; Srinivasan, S. S.; Ku mar, A.; Stefanakos, E. K.; Goswami, Y.; McGrath, K., Processing Analysis of the Ternary LiNH2-MgH2-LiBH4 System for Hydrogen Storage. Int. J. Hydrogen Energy 2009, (Submitted). 137. Yang, J.; Sudik, A.; Siegel, D. J.; Ha lliday, D.; Drews, A.; III, R. O. C.; Wolverton, C.; Lewis, G. J.; Sachtler, J. W. A.; Low, J. J.; Faheem, S. A.; Lesch, D. A.; Ozolins, V., Hydrogen storage properties of 2LiNH2 + LiBH4 + MgH2. J. Alloys. Compd. 2007, 446-447, 345. 138. Yang, J. C., Material-based hydrogen storage. Int. J. Hydrogen Energy 2008, 33, (16), 4424-4426. 139. Crabtree, G. M.; Dresselhaus, M. S.; Buchanan, M. V., The Hydrogen Economy Physics Today 2004, 57, (12), 39-44.

PAGE 174

158 140. Edwards, P. P.; Grochala W.; Book, D.; Harris, I., Metal Hydride Synthesis and a Fuel Cell Using Metal Hydrides for Hydrogen Storage In: Office EP ed.; 2004. 141. Stefanakos, E.; Goswami, Y.; Srinivasan, S.; Wolan, J., In J. Hydrogen Energy Myer Kutz (Hrsg.) Environmentally C onscious Alternative Energy Production, John Wiley & Sons, 4, 165 (2007) 142. Bogdanovic, B.; Schwickardi, M., Tidoped alkali metal aluminium hydrides as potential novel reversible hydrogen storage materials. J. Alloys Compd. 1997, 253-254, 1-9. 143. Bogdanovic, B.; Schwickardi, M., Verfahren zur reversilben Speicherung von Wasserstoff In: Office EP ed.; 1997. 144. Graetz, J.; Lee, Y.; Reilly, J. J. ; Park, S.; Vogt, T., Structures and thermodynamics of the mixed alkali alanates. Phys. Rev. B 2005, 71, (18), 184115. 145. Jensen, C. M.; Zidan, R., Hydrogen Storage Materials and Methods of Making by Dry Hydrogenation In: Office UP ed.; 2002; Vol. 6, 471, 935 B2. USA. 146. Kang, J. K.; Lee, J. Y.; Muller, R. P. ; III, W. A. G., Hydrogen storage in LiAlH4 : Predictions of the crystal structures and reaction mech anisms of intermediate phases from quantum mechanics. J. Chem. Phys. 2004, 121, (21), 10623. 147. Brinks, H. W.; Brown, C.; Jensen, C. M. ; Graetz, J.; Reilly, J. J.; Hauback, B. C., The crystal structure of -AlD3. J. Alloys. Compd. 2007, 441, (1-2), 364-367. 148. Graetz, J.; Reilly, J. J.; Kulleck, J. G.; Bowman, R. C., Kinetics and thermodynamics of the aluminum hydride polymorphs. J. Alloys. Compd. 2007, 446-447, 271-275. 149. Au, M. Destabilized and catalyzed alkali metal borohydrides for hydrogen storage with good reversibility, U.S. Patent Appl. Publ., US 2006/0194695 A1, Aug. 31. 2006. 150. Chen, P.; Xiong, Z.; Lou, J.; Lin, J.; Tan, K. L., Interaction of hydrogen with metal nitrides and imides. Nature 2002, 420, 302. 151. Hu, Y. H.; Ruckenstein, E., H2 Storage in Li3N. Temperature-Programmed Hydrogenation and Dehydrogenation. Ind. Eng. Chem. Res. 2003, 42, (21), 51355139. 152. Lewis, J. G.; Sachtler, J. W.; Drews, A. ; Low, J. J.; Lesch, D. A.; Faheem, S. A.; Dosek, P. M.; Knight, L. M.; Halloran, L.; Jensen, C. M.; Yang, J.; Sudik, A.; Siegel, D. J.; Wolverton, C.; Ozolins, V.; Zhang, S., High throughput screening of the ternary LiNH2–MgH2–LiBH4 phase diagram. J. Alloys. Compd. 2007, 446447, 355-359.

PAGE 175

159 153. Niemann, M.; Srinivasan, S.; McGrath, K.; Kumar, A.; Goswami, Y.; Stefanakos, E., Nanocrystalline effects in complex composite hydrides. Ceramic Transactions 2009, 202, 111. 154. Bogdanovic, B.; Felderhoff, M.; Kaskel, S.; Pommerin, A.; Schlichte, K.; Schuth, F., Improved hydrogen storage properties of Ti-doped sodium aluminate using titanium nanoparticles as doping agents. Adv. Mater. 2003, 15, (12), 1012. 155. Bogdanovic, B.; Ritter, A.; Spliet hoff, B., Active MgH2-Mg systems for reversible chemical energy storage. Angew. Chem. Int. Ed. Engl 1990, 29, 223. 156. Report of the Basic Energy Science Wo rkshop on Hydrogen Production, Storage and use prepared by Argonne National Laboratory: May 13-15, 2003. 157. Satyapal, S.; Petrovic, J.; T homas, G., Gassing with Hydrogen. Scientific American 2007, 81-87. 158. Nakamori, Y.; Ninomiya, A.; Kitahara G.; Aoki, M.; Noritake, T.; Miwa, K.; Kojima, Y.; Orimo, S., Dehydriding r eactions of mixed complex hydrides. J. Power Sources 2006, 155, 47. 159. Li, Z. P.; Liu, B. H.; Arai, K.; Morigazaki, N.; Suda, S., Protide compounds in hydrogen storage systems. J. Alloys. Compd. 2003, 356-357, 469-474. 160. Zuttel, A.; Wenger, P.; Rentsch, S.; Sudan, P.; Mauron, P.; Emmeneger, C., LiBH4 a new hydrogen storage material. J. Power Sources 2003, 118, 1. 161. Narashimhan, S.; Balakumar, R., Synt hetic Applications of Zinc Borohydride. Aldrichimica Acta 1998, 31, (1), 19-26. 162. Grove, H.; Brinks, H. W.; Heyn, R. H. ; Wu, F.-J.; Opalka, S. M.; Tang, X.; Laube, B. L.; Hauback, B. C., The structure of LiMg(AlD4)3. J. Alloys. Compd. 2008, 455, 249-254. 163. Nakamura, Y.; Fossdal, A.; Brinks, H. W.; Hauback, B. C., Characterization of Al-Ti phases in cycled TiF3-enhanced Na2LiAlH6. J. Alloys. Compd. 2006, 416, (1-2), 274-278. 164. Hagemann, H.; Longhini, M.; Kaminski, J. W.; Wesolowski, T. A.; Cerny, R.; Penin, N.; Sorby, M. H.; Hauback, B. C.; Severa, G.; Jensen, C. M., LiSc(BH4)4: A Novel Salt of Li+ and Discrete Sc(BH4)4Complex Anions. J. Phys. Chem. A 2008, 112, (33), 7551-7555. 165. Kissinger, H. E., Variation of Peak Temperature With Heating Rate in Differential Thermal Analysis. Journal of Research of the National Bureau of Standards 1956, 57, 217.

PAGE 176

160 166. Joo, S. H.; Choi, S. J.; Ryoo, R., Ordered nanoporous arra ys of carbon supporting high dispersions of platinum nanoparticles. Nature 2001, 412, 169.

PAGE 177

161 About the Author Pabitra Choudhury was born in Barnia of Nadi a district in West Bengal, India. He completed both his middle a nd high school education at Be thuadahari J. C. M. High School, Nadia of West Bengal, India. Pabitra holds a BS from IIT, Bombay and an MS from IIT, Roorkee, India, both in chemical engineering. He began pursuing his Ph.D. degree in th e Chemical and Biomedical Engineering Department at University of South Florida in 2006. His di ssertation resear ch as applied to materials related to energy field resulted in publications in high quality journals such as Physical Review B, Applied Physics Lett ers, Journal of Physical Chemistry C, and International Journal of Hydrogen Energy. His research has also resulted in 15+ publications including conferen ce/symposium proceedings, asso ciated with prestigious national and international conferences.