WO1994006009A1

WO1994006009A1 - Hydrogen storage materials, their identification and utilization

Info

Publication number: WO1994006009A1
Application number: PCT/US1993/008181
Authority: WO
Inventors: Keith H. Johnson
Original assignee: Arthur D. Little, Inc.
Priority date: 1992-08-28
Filing date: 1993-08-30
Publication date: 1994-03-17

Abstract

A process is disclosed for identifying the solubilities of hydrogen or its isotopes in selected materials of single or multiple element constituents. The process tabulates the solubilities for an evaluation set having a range of material constituent concentrations. The solubility of hydrogen is established in a computer program which solves the Schrödinger wave equations using the Xα scattered wave approach to achieve plots of the electron orbital energy levels for the material and its constituents. The material constituent concentration, in an evaluation set, that has those d level orbitals that contain the Fermi level aligned with the 1s orbital of hydrogen, or its isotope, is selected as the material with the optimal solubility. Such a material has utility in hydrogen storage for fuel cells, for other applications using hydrogen from a recyclable supply, and in electrical battery technology.

Description

HYDROGEN STORAGE MATERIALS, THEIR IDENTIFICATION

AND UTILIZATION

BACKGROUND OF THE INVENTION

Storage of hydrogen (which as used herein includes its isotopes) is an important topic: 1) because hydrogen serves as a useful fuel for high efficiency, on demand electricity production in devices such as fuel cells; 2) because of its relevance to future, ecologically sensitive energy needs; and 3) because of its near term potential for the development of portable, high energy density and high power density devices such as rechargeable batteries, fuel cells and ultracapacitors.

Traditional hydrogen storage systems have used pressurized gas cylinders, chemical reaction generators, and cryogenics for liquification. The gas cylinder approach is disadvantageous because of its limited storage volume and high energy penalty in pressurizing the gas, which energy is lost as heat. The liquification systems also suffer from the loss in efficiency associated with the high power demands of the liquification step. The chemical generators suffer from the need for complex reformers and/or no reversibility.

It is known that solids including many rare earth/transition metal based materials or metal hydrides are able to store hydrogen gas atoms within the interstices of the material lattice at concentrations of hydrogen near or above the concentrations of the liquified gas.

Some materials such as LiH and CaH₂ are irreversible and possess high heats of formation. Other metal hydrides are better candidates, but no theory exists to predict those that are the more efficient storage devices. BRIEF SUMMARY OF THE INVENTION

The present invention teaches: 1) the basis for the identification of the solubility of hydrogen (which is herein understood to include its isotopes) in metals or their alloys; 2) the metal hydrides thus identified; and 3) the cells that utilize the identified material for advantageous energy generation from hydrogen as a fuel.

The invention utilizes the equivalence of the molecular orbital energy Eigenvalues calculated from a self-consistent field, Xα, scattered wave method, to orbital electronegativities. It further recognizes that the bonding of molecules or atoms is a function of the electronegativity or energy level differences of electron orbitals where bonding takes place. Ionic bonding occurs when electronegativity differences are great and covalent bonds occur with equal or nearly equal energy levels. The solubility of hydrogen in a material is seen to be a function of the difference in energy levels between the electron orbitals of hydrogen and the material in which it is desired to be soluble. In particular, a method according to the invention establishes electron orbital energy levels for a material iteratively as a series of tabulations as concentrations of the material components are changed form one iterative calculation to another. The calculations involve the self-consistent field, Xα, scattered wave solutions to the wave functions and result in a complete plot of the electron-orbital energy levels. In particular, the formulation with its energy d-band or d-orbital band containing the Fermi energy level (the top energy level at zero degrees Kelvin) centered about the Is hydrogen orbital is selected as the optimal formulation for hydrogen storage for that set of material components.

A material according to that concentration ratio is then formulated. The material thus formed is appropriately dimensioned for use as a reversible hydrogen storage element; for use as a hydrogen source wherever hydrogen as a fuel is required. DESCRIPTION OF THE DRAWINGS

These and other features of the invention are more fully described in the accompanying, solely exemplary, detailed description and accompanying drawings of which:

Fig. 1 is an exemplary energy level diagram used in practicing the invention;

Fig. 2 illustrates an exemplary material lattice with hydrogen sites shown;

Figs. 3 and 4 are diagrams illustrating nuclear bonding; Fig. 5 illustrates a method used for selecting a hydrogen material according to the invention;

Fig. 6 illustrates a method according to the invention for hydrogen storage in a material selected using features of the invention;

Fig. 7 is a representation of materials which are processed according to the invention; and

Fig. 8 is a diagram of an energy device used in practicing the invention.

DETAILED DESCRIPTION

The present invention contemplates a process for the identification and preparation of materials in which hydrogen is highly soluble and the materials thus identified. Before presenting the details of that process according to the invention, a brief presentation of the theory underlying the present invention is presented followed by a description of the computer algorithmic approach to determining electronegativities or orbital energy levels for selected materials which may be either single element materials or multi-element materials. Finally the identification of optimal storage concentrations and their utilization is presented. COMPUTATIONAL APPROACH

I. INTRODUCTION

In most solids at room temperature, the thermally induced oscillations of the nuclei have insufficient energy to excite the electrons; that is, the phonon energies are much less than the difference between the lowest unoccupied electronic state and the highest occupied electronic state. Consequently, most calculational techniques for the electronic structure of solids neglect the motion of the nuclei.

However, there are solids (generally metals) in which the highest occupied state splits into levels that have almost the same energy (i.e., are nearly degenerate). For these solids, the phonon energies can exceed the energy differences between the nearly degenerate electronic states, and nuclear oscillations can affect the electronic state. Thus, for these solids, conventional calculational techniques do not apply.

The calculational method of the present invention, the Self-Consistent-Field Xalpha Scattered-Wave (SCF-Xα-SW) technique, differs from conventional methods in that it does not consider infinite crystals and their global properties. Rather, it performs electronic wave function calculations on a small representative cluster of atoms. This cluster contains at least one atom of each element in the crystal lattice, plus (using symmetry to limit numbers) the nearest neighboring atoms in that lattice, and often next-nearest neighbors and sometimes even third-nearest neighbors.

The calculational method of the present invention considers the Coulombic interactions between all nuclei and all electrons within the cluster, solving for each electronic state the Schrödinger wave equation. It thus calculates the energy levels of all electrons within the cluster and, for all positions within the cluster, the electron wave function, which is the probability density function representing the likelihood of an electron occupying that position. In particular, by combining Slater's Xalpha (Xα) density-functional theory of electron-electron exchange-correlation (J.C. Slater, Advances in Quantum Chemistry, Academic Press, New York, 1972, p.l) with Johnson's Scattered-Wave Cluster Molecular-Orbital Method of solving, the Schrδdinger wave equation (K.H. Johnson, J . Chem . Phys . 45, 3085, 1966; K.H. Johnson, Advances in Quantum Chemistry, Academic Press, New York, 1973, p. 143), a practical and powerful new computational technique, the Self-Consistent-Field Xalpha Scattered-Wave (SCF-Xα-SW) Cluster Molecular-Orbital method of calculating, from first quantum-mechanical principles, the electronic structures and chemical bonding of polyatomic molecules and solids was developed (J.C. Slater and K.H. Johnson, Phys . Rev. B 5, 844, 1972). The computational steps for implementing this technique, using programs written by K. H. Johnson, are described in Section IV.

II. The SCF-Xα-SW Energy Eigenvalues and Orbital Electronegativity

It has been rigorously proven that the molecular-orbital energy eigenvalues calculated by the SCF-Xα-SW method are equivalent to orbital electronegativities . (K.H. Johnson, Int . J . Quantum Chem . 11S, 39, 1977; M.E. McHenry, R.C. O'Handley, and K. H. Johnson, Phys . Rev. B35, 3555, 1987). See Equation (9) below. Therefore, the greater the difference between the energy levels of two atoms or atomic clusters, the greater is the orbital electronegativity difference and tendency for ionic bonding. Conversely, the more equal the energy levels, the more covalent the bonding will be.

This principle can be illustrated for the case of hydrogen atom impurities dissolved in tetrahedral clusters of nickel, palladium, and platinum (Fig. 1). Fig. 1, first published by R.P. Messmer, D.R. Salahub, K.H. Johnson, and C.Y. Yang, Chem . Phys . Lett . 51, 84 (1977), compares the SCF-Xα-SW molecular-orbital energies (including relativistic effects) of tetrahedral nickel, palladium, and platinum clusters with and without hydrogen at the center of the tetrahedra. These clusters may be compared with the tetrahedral interstitial sites of the corresponding face-centered-cubic (fee) bulk metals, illustrated for palladium hydride in Fig. 2.

The almost perfect lining up of the hydrogen ls-orbital electronegativity with the center of the palladium d-orbital electronegativity manifold (band) in Fig. 1 indicates almost perfect covalent chemical bonding between a hydrogen atom and a surrounding palladium tetrahedral environment. In contrast, from the higher and lower positions, respectively, of the nickel and platinum d-orbital electronegativity manifolds with respect to the hydrogen ls-orbital electronegativity in Fig. 1, nickel and platinum tetrahedra are respectively electropositive and electronegative with respect to atomic hydrogen. These theoretical results were originally shown (R.P. Messmer, D.R. Salahub, K.H. Johnson, and C.Y. Yang, Chem . Phys . Lett . 51, 84, 1977) to satisfactorily explain the photoelectric emission spectra and work function trends for hydrogen chemisorbed on and dissolved in nickel, palladium, and platinum.

III. Orbital Electronegativity and H/D Solubility in Pd and Pd alloys

It is well known that atomic hydrogen is more soluble in palladium than in nickel or platinum (F.A. Lewis, The Palladium/Hydrogen System, Academic Press, New York, 1967). The strength of a heternuclear chemical bond, as originally described by Pauling (L. Pauling, The Nature of the Chemical Bond, Cornell University Press, Ithaca, I960, p.80), is the combined effect of its covalent and ionic components. It has also been established that the solubility of an impurity in a metal or alloy generally decreases with increasing electronegativity difference between solute and solvent, other factors such as atomic size factor remaining constant (W. Hume-Rothery, The Structure of Metals and Alloys , Institute of Metals, London, 1936). Thus the attainment of nearly zero net orbital electronegativity difference between hydrogen and palladium as predicted by the SCF-Xα-SW cluster molecular-orbital calculations, thereby minimizing ionic contributions to the bonding and optimizing Pd(4d)-H(1s) covalency, is consistent with the higher solubility of hydrogen in palladium as compared with nickel and platinum.

The same principles can be utilized to explain and predict the solubility of hydrogen in palladium alloys. For example, when palladium is alloyed with silver and gold, the 4d-orbital electronegativity of the alloy is increased relative to the 1s-orbital electronegativity of hydrogen, and hydrogen solubility is decreased, a well known experimental fact (F.A. Lewis, The Palladium/Hydrogen System, Academic, New York, 1967). IV. The SCF-Xα-SW Computational Procedure

To implement the SCF-Xα-SW Cluster Molecular Orbital Method of calculating orbital electronegativities and estimating relative solubilities, the following computational steps (Modules) are used.

1. HS Module: Performs SCF-Xα computations for all atoms of the Periodic Table. Outputs Atomic energy levels and charge densities for input to Cluster Molecular Potential Program, XMOLPOT. Solves Schrödinger's Equation (shown here in Rydberg atomic units, 1 Rydberg = 13 . 6 electron volts) :

iteratively for the Xalpha density-functional atomic potential V_Xα(r), where r is radial distance of an electron from the atomic nucleus, ε_n,1 are the atomic orbital energy eigenvalues, Ψ_{n, l,m}(r, θ_, Φ) ) are the atomic orbital wavefunctions, r and ø are the electronic spherical angular coordinates, and n, l,m are the atomic orbital quantum numbers. From the self-consistent wavefunctions, the spherically averaged atomic charge density:

is computed and output as input to XMOLPOT program. 2. XMOLPOT Module: Takes atomic charge densities ρ_atom (r) output from the HS Program, adds them up in each atomic sphere:

for the chosen crystal/cluster structure, and solves Poisson's Equation:

of electrostatics to determine the Coulomb contribution of the starting cluster molecular potential. The Xα exchange-correlation contribution:

is added to this. Outputs starting cluster molecular potential:

in each atomic sphere and extramolecular region for input to ESRCH Module.

3. YLM3 Module: Projects symmetrized combinations of spherical harmonics:

from symmetry group theory for the chosen cluster point group. Outputs coefficients C_yIm of symmetrized combinations for input to ESRCH and SCF Modules.

4. ESRCH Module:: Using outputs from XMOLPOT and YLM3 Programs, solves Schrδdinger'S Equation BY Scattered-Wave Method for the initial energy eigenvalues of the cluster from the secular determinantal equation:

where t_l ^j(ε) is the scattering matrix of cluster atom j for partial wave of angular momentum 1 and energy ε, and is

the Green's function "propagator" of partial waves between atoms j and j ' in the cluster. Outputs cluster potential and charge densities for input to SCF Module.

5. SCF Module: Using output from ESRCH and YLM3 Programs, solves Schrödinger's Equation iteratively by the Scattered-Wave Method, utilizing again the determinantal equation (8) and employing VGEN Subroutine to compute new cluster molecular potential of the type (6), for which Equation (8) is solved again, etc . , etc . , until self-consistency is attained. Outputs cluster molecular-orbital energy levels, charge distributions, and total energy to SUMMARY subroutine. Also outputs partial-wave coefficients of cluster molecular orbitals for input to Wavefunction Program, WAVFN.

6. SUMMARY Subroutine: Produces tabulation of cluster molecular-orbital energies, charge distributions, and total energy. The resulting energy levels ε_iXα are equivalent to orbital electroegativities X_i', as described earlier.

Fig. 1, discussed above, is a graphical comparison of the quantities in Equation (9) for the nickel, palladium, and platinum clusters with and without hydrogen. At this step, enough information is obtained to determine relative solubilities, as described in Section II. Information on the spatial character of the chemical bonding between solute and solvent is obtained by the following computational steps:

7. WAVFN Module: From the output of the SCF Program, computes cluster molecular wavefunctions on three-dimensional grid.

8. CONTOUR Program: From the output of WAVFN, converts numerical wavefunctions for input to a contour mapping program.

Fig. 5 illustrates the utilization of these computational techniques in an algorithm by which the structural characteristics of any of the elements and their combinations as molecules or alloys can be calculated according to the present invention utilizing the modules described above. As shown there, an HS module in step 4 provides the computations which can be applied to any atoms of the periodic table that generate the atomic energy levels and the corresponding charge densities. These charge densities are utilized in the XMOLPOT module in a step 8 which combines them mathematically for each atomic sphere for a desired number of atoms defining the cluster of interest. From these the Coulomb contribution of the starting cluster molecular potential is obtained from a solution to Poisson's equation and thereafter the Xα exchange correlation contribution is added and finally an output is formed of the starting cluster molecular potential in each atomic sphere and extramolecular region which in turn is provided to the ESRCH calculational module in step 10. Independently in the YLM3 calculation module of step 12 the coefficients of symmetrized combinations for the chosen cluster point group are calculated and applied to the calculation module of step 10. The module of step 10, given these two inputs, solves the Schrδdinger's wave equation, utilizing the SCF-Xα-SW method. The resulting cluster potential and charge densities calculated in the module of step 10 are applied to the SCF module, in step 14, along with the information from the module of step 12. This achieves the actual Schrödinger wave equation solution on a iterative basis utilizing the scattered wave method. The computations of the module of step 14 loop with the VGEN subroutine module in step 16 as described above for each iterative step until self-consistency occurs. The output of the module of step 14, the cluster molecular-orbital energy levels, charge distributions and total energy is applied to the SUMMARY module in rtep 18 which tabulates the calculated molecular orbital energies for the cluster, the charge distributions and the total energy. The energy levels which result from this calculation are, by definition, the orbital electronegativities and are printed in the form of Fig. 1 or in tables. A further subcalculation routine the WAVFN module in step 20 utilizes the results from the module of step 14 to provide the cluster molecular wave functions as a three dimensional plot. Additionally, the CONTOUR module in step 22 converts this three-dimensional plot into data usable in the plot module in step 23, such as by a conventional plotting.

The above described process is iteratively implemented according to the present invention as illustrated in Fig. 6 for identifying optimal metals or alloys for hydrogen solubility. The process begins with a step 24 in which the various material constituents of the alloy under question, or the individual element, under study are identified. In step 26 the electron orbitals are established, iteratively for a range of concentrations of the various material constituents of the alloy. From the plots, in step 28, those alloy combinations in.which the Is level of the hydrogen atom is best centered in the d-orbital electron band at the Fermi level is selected. In step 30 an alloy according to that prescription is formulated, and configured for use in a hydrogen storage cell for reversible hydrogen storage. In subsequent step 32 that material is placed in use in such cell for reversible hydrogen storage.

In general, a metal or other hydride 34 of Fig. 7 has its metal structure loaded with atoms of hydrogen from which hydrogen is retrievable and restorable as is known in the art. The metal or material M can be a transition metal and/or rare earth alloy including one or more constituents selected from the group Mg, Ni, La, Mm, Mn, Al, Ti, Fe, Co, Zr, Mo, Zn, Cr, V, B, Pd, alone or alloyed with P, Zr, Ag, Au, or Cu and T, alone or alloyed with Ni. For example, FeNiB in roughly equal constituency concentrations is predicted by the present invention to be a good hydrogen storage material, the more so because its glassy nature limits fractures.

The application of the present invention to the storage of hydrogen for fuel cell, internal combustion engines, gas turbines, etc. applications Is illustrated in Fig. 8. As shown there, a supply 160 of oxygen is combined with a supply 162 of hydrogen with their gaseous feeds applied to a combustion or use cell 164 as is known in the art and exemplified above. The cell 164 provides combustion of the hydrogen and oxygen with the release of an electric current or work as outputs along with water as a bi-product. Such systems are known in the art for such applications as energy sources in space. The utilization of a metal hydride as the hydrogen source is well known in the art owing to the fact that the metal hydride stores hydrogen at approximately the same, or even higher density than hydrogen is stored in a liquified state but without the utilization of cryogenics and/or high pressure vessels typically associated with liquified hydrogen. The metal in which the hydrogen may be stored is formulated according to the above-identified material designation process. It may also include materials which are selected by utilizing the above computational procedures to iteratively identify those members of the periodic table including their combinations in molecular, alloy or other structures which indicate a usable level of hydrogen storage.

It is to be noted that the supply of oxygen in Fig. 8, instead of from a containment vessel 160, may be from atmospheric or other sources, including chemical oxygen generators.

The above described embodiments of the invention are presented as exemplary only and the scope of the invention is accordingly defined solely in accordance with the following claims.

Claims

1. The method of identifying a hydrogen storage material comprising the steps of:

selecting material constituent types;

determining the molecular electron orbital Eigenvalues for a range of different concentrations of the constituents of said selected material; and

identifying the constituent concentrations for which the d-orbital having the Fermi energy level has the hydrogen Is orbital most closely centered therein.

2. The process of claim 1 further including the step of:

formulating the selected material at the identified constituent concentrations.

3. The process of claim 2 further including the step of:

placing said formulated material in a cell for reversible hydrogen storage.

4. The process of claim 3 further including the step of:

reversibly storing hydrogen in said formulated material within said storage cell.

5. Hydrides of material identified by the process of claim 1.

6. A metal hydride having the Fermi level containing molecular electron d-orbital Eigenvalues of the non-hydrogen component of the hydride centered about the Is orbital of hydrogen.

7. A metal for metal hydride application wherein the non-metal components are selected so as to have the molecular electron d-orbital Eigenvalues for the d-orbital containing the Fermi energy level centered about the Is hydrogen orbital.

8. A metal hydride having the non-hydrogen component thereof known to have its molecular electron d-orbital Eigenvalues for the d-orbital band having the Fermi energy levels centered about the hydrogen Is orbital.

9. A process for the reversible storage of hydrogen comprising the step of exposing a material of the formula FeNiB, with the elements in substantially equal concentrations, to hydrogen for the controlled storage and release thereof.