Hydrogen absorption properties of scandium and aluminium based compounds

UPTEC K10 022

Examensarbete 30 hp Augusti 2010

Hydrogen absorption properties of scandium and aluminium based compounds

Adam Sobkowiak

Teknisk- naturvetenskaplig fakultet UTH-enheten

Besöksadress:

Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0

Postadress:

Box 536 751 21 Uppsala

Telefon:

018 – 471 30 03

Telefax:

018 – 471 30 00

Hemsida:

http://www.teknat.uu.se/student

Abstract

Hydrogen absorption properties of scandium and aluminium based compounds

Adam Sobkowiak

In a time of global environmental problems due to overuse of fossil fuels, and a subsequent depletion of the supplies, hydrogen is considered as one of the most important renewable future fuels for use in clean energy systems with zero

greenhouse-gas emission. Hydrogen storage is the main issue that needs to be solved before the technology can be implemented into key areas such as transport. The high energy density, good stability and reversibility of metal hydrides make them appealing as hydrogen storage materials.

In this thesis research on synthesis and hydrogen absorption properties for intermetallic compounds based on scandium and aluminium is reported. The compounds were synthesized by arc melting or induction melting and exposed to hydrogen in a high pressure furnace. Desorption investigations were performed by thermal desorption spectroscopy. The samples were analyzed by x-ray powder diffraction and electron microscopy.

ScAlNi, crystallizing in the MgZn2-type structure (space group: P63/mmc; a = 5.1434(1) Å, c = 8.1820(2) Å), was found to absorb hydrogen by two different mechanisms at different temperature regions. At ~120 °C hydrogen was absorbed by solid solution formation with estimated compositions up to ScAlNiH0.5. At ~500 °C hydrogen was absorbed by disproportionation of ScAlNi into ScH2 and AlNi. The reaction was found to be fully reversible due to destabilization effects which lowered the decomposition temperature of ScH2 by ~460 °C.

ISSN: 1650-8297, UPTEC K10 022 Examinator: Gunnar Westin Ämnesgranskare: Annika Pohl

Handledare: Yvonne Brandt Andersson

Sammanfattning på svenska

Den globala uppvärmningen och dess effekter på vår miljö har blivit ett allt hetare diskussionsämne under senare år. De stora frågorna kretsar kring hur vi ska minska våra utsläpp av växthusgaser, framförallt koldioxid (CO2), och hur vi kan ersätta fossila bränslen som huvudsaklig energikälla.

En av de enskilt största miljöbovarna idag är transportsektorn, dvs. bilismen, flyget osv.

Stora satsningar har således gjorts på utveckling av miljövänlig, eller s.k. ”grön”,

energiteknik inom detta område. För att signifikant minska utsläppen måste man gå över till en annan sorts bränsle än de som normalt används idag. Vätgas (H2) har visat sig vara en högintressant kandidat för detta ändamål. Energiutvinningen ur vätgas vid reaktion med syrgas (O₂) ger praktiskt taget endast vatten som biprodukt vilket gör processen fullständigt miljövänlig.

H

+ ½O

→ H

O + energi

Reaktionen kan utnyttjas på två olika sätt i ett tänkbart scenario under huven på en bil.

Antingen i en förbränningsprocess i en klassisk förbränningsmotor, eller i en

elektrokemisk process som genererar elektricitet till driften av en elektrisk motor. Vätgas har stora energetiska fördelar med tre gånger mer energi per viktenhet jämfört med fossila bränslen. Förbränningen av vätgas med syre leder dessutom till mindre värmeförluster än förbränningen av bensin, dvs. effektiviteten är större.

Fördelarna med vätgas som bränsle är uppenbara, men det finns ett antal faktorer som förhindrar en kommersialisering av vätebaserade energisystem inom en nära framtid. Ett av de största problemen idag är vätelagringen. Det finns ännu ingen tillräckligt bra lagringsteknik som skulle göra vätedrivna bilar till ett konkurrenskraftigt alternativ på marknaden. Vätelagringstekniken måste bl.a. uppfylla stränga säkerhetskrav, ha låg produktionskostnad och framförallt erbjuda attraktiva färdsträckor jämfört med

existerande alternativ. I dagens vätedrivna prototypbilar sker lagringen i tryckbehållare, antingen i gasfas eller i nerkyld vätskefas. Dessa tekniker anses dock ej vara tillräckligt effektiva med för låga lagringskapaciteter i förhållande till systemens vikt och volym.

Säkerheten anses även vara otillräcklig pga explosionsrisken vid eventuella trafikolyckor.

Tryckbehållare anses således inte vara en lösning för framtiden.

Under de senaste årtiondena har forskningen på nya lagringstekniker där vätet binds kemiskt i fasta material ökat stadigt. En typ av material man tror mycket på i dessa sammanhang är metallhydrider som bildas då en metall eller metallegering reagerar med vätgas under specifika temperatur- och tryckförhållanden. Det är stabila föreningar där väteatomer lagras genom att ockupera hålrummen mellan metallatomerna i materialet (se figur 2, sid. 8, för schematisk representation). De största fördelarna med metallhydrider är potentiellt stora lagringskapaciteter, god kemisk stabilitet samt god reversibilitet.

Problemen man ställs inför i dagens metallhydridforskning är att hitta en metallförening med de rätta egenskaperna. Materialet måste kunna lagra relativt stora mängder väte i förhållande till sin vikt och volym. Vätelagringen måste även vara reversibel under milda förhållanden, dvs man måste kunna få in och få ur vätet utan att behöva applicera

orealistiskt höga temperaturer och tryck. Samtidigt måste materialet ha en lång livslängd och lågt pris. Att hitta en metallförening med den perfekta kombinationen av dessa egenskaper har visat sig vara en svår uppgift.

Målet med detta examensarbete var att undersöka vätelagringsegenskaperna hos olika skandium- och aluminiumbaserade metallegeringar som aldrig tidigare studerats i detta syfte. Legeringen ScAlNi, en förening mellan lika stora mängder av metallerna

skandium, aluminium och nickel, visade sig besitta intressanta egenskaper. Resultaten visade bl.a. att materialet absorberar och lagrar vätgas genom två distinkt skilda

mekanismer vid olika temperaturer. Vid lägre temparturer, med start vid c:a 120 °C, kan väte lagras i mindre mängder genom fast lösning i metallstrukturen. Vid högre

temperaturer ökar lagringskapaciteten fyrfaldigt genom att materialet sönderfaller i två nya faser, ScH₂ och AlNi, där den förstnämnda innehåller höga halter väte. Denna reaktion kan beskrivas med följande formel:

ScAlNi + H

→ ScH

+ AlNi

Reaktionen visade sig vara fullständigt reversibel vilket är sällsynt för skandiumbaserade legeringar, i synnerhet för den stabila skandiumhydriden (ScH₂). Normalt krävs

temperaturer upp emot 1000 °C för att skandiumhydriden ska släppa ifrån sig vätet. I det hydrerade tillståndet av ScAlNi visade det sig att desorptionstemperaturen kunde sänkas med 500 °C genom destabiliseringseffekter från den samexisterande AlNi-fasen.

Trots vissa intressanta egenskaper har ScAlNi inte tillräckligt bra prestanda för att kunna användas i praktiska tillämpningar. Absorptions- och desorptionstemperaturerna är fortfarande för höga och lagringskapaciteten är för låg. Därtill tillkommer höga materialkostnader, främst pga det höga priset på skandium. Resultaten i denna studie bidrar dock till allmänt ökad kunskap om hur metallegeringar baserade på skandium och aluminium reagerar med vätgas under olika temperatur- och tryckförhållanden. Studien visar också att reaktionshastigheterna för stabila föreningar som skandiumhydrid är högst påverkningsbara samt att man kan finna alternativa vätelagringsmekanismer i

metallhydrider. Denna information är värdefull för att i framtiden bättre kunna skräddarsy nya materials vätelagringsegenskaper.

Table of content

1. Introduction ... 5

2. Metal hydrides ... 7

2.1 Overview ... 7

2.2 Hydrogen absorption mechanisms ... 8

2.3 Thermodynamics... 9

2.4 Alloys with scandium and aluminium ... 10

3. Aim ... 11

4. Experimental ... 12

4.1 Synthesis ... 12

4.2 Hydrogen absorption/desorption ... 13

4.3 Thermal desorption spectroscopy (TDS) ... 13

4.4 Characterization ... 14

4.4.1 X-ray diffraction (XRD) ... 14

4.4.2 Scanning electron microscopy (SEM) ... 15

5. Results & discussion ... 16

5.1 XRD characterization of synthesized samples ... 16

5.1.1 ScAlNi ... 16

5.1.2 ScAl1-xSix ... 17

5.2 Hydrogenation of ScAlNi ... 19

5.2.1 Hydrogenation at lower temperatures ... 19

5.2.2 Hydrogenation at higher temperatures ... 21

5.3 Hydrogen desorption of ScAlNi-Hx ... 23

5.3.1 Hydrogen desorption and cycling of α-phase hydride ... 23

5.3.2 Thermal desorption spectroscopy of disproportionated ScAlNi ... 24

5.4 Scanning electron microscopy of ScAlNi ... 27

6. Summary ... 28

7. Conclusions ... 28

8. Acknowledgements ... 29

9. References ... 30

1. Introduction

The environmental awareness has increased significantly in the society over the last years and both climate and energy issues have become very important and engaging political topics. The major questions that are being discussed deals with how we can reduce the emission of greenhouse-gases into the atmosphere and, in the long run, how we can reduce our dependence on fossil fuels which is the most widely used energy source today [1]. The fact that the fossil fuel supplies are depleting while the future energy demands are steadily increasing makes the research and development of clean energy systems that involves renewable energy more important and of bigger interest than ever before [2].

Transport is one of the biggest contributors to massive air pollutions and thus much effort has been devoted to develop clean energy systems for the vehicular field. To reach this goal the liquid hydrocarbons used today must be replaced with a new energy carrying substance and hydrogen gas is considered to be one of the most promising candidates.

There are essentially two ways to run a vehicle on hydrogen; either by burning it in oxygen from air in a combustion engine, or by reacting it with oxygen electrochemically in a fuel cell or battery to run electrical engines. In either case energy is gained from the reaction while water is produced as waste product, which makes it a virtually pollution- free reaction.

H

+ ½O

→ H

O + energy

The specific chemical energy of molecular hydrogen,142 MJ/kg, is roughly three times higher compared to other common chemical fuels (the corresponding value for liquid hydrocarbons is 50 MJ/kg) [3]. In combustion engines much of the energy is lost during the transformation of chemical, through thermal, to mechanical energy, as described by the Carnot efficiency equation. The combustion of hydrogen-air mixtures is however about 25 % more efficient than for hydrocarbon-air mixtures. In the electrochemical reaction the electron transfer process is not limited by the Carnot efficiency and can be about 50-60% more efficient than the thermal combustion [3, 4].

Hydrogen is the most abundant element in universe but less than 1% is present as

molecular H2. In the case of future hydrogen-driven cars the production of hydrogen gas has to be increased drastically. There are already different established techniques

available. However, most hydrogen produced today comes from hydrocarbon fuels [5].

To consider hydrogen based energy systems as totally clean the hydrogen should be produced from clean and renewable energy sources. Such production technologies do exist today and are in constant development. Most promising are different water splitting techniques, e.g. photoelectrolysis where sunlight is used to directly decompose water into hydrogen and oxygen over a catalyst. Another interesting production method is to use microorganisms, such as green algae, for production of hydrogen under anaerobic conditions [6].

The advantages of using hydrogen as a clean energy carrier are obvious. Scientists, engineers, companies and agencies are convinced that it will play an important role in the future. However, there are still some major barriers that have to be overcome before it can be used commercially. One of the biggest issues today is the storage. There is yet no way of storing hydrogen onboard a vehicle that fulfills the numerous demands set by the industry regarding economical, safety, and driving range aspects. In todays prototype cars hydrogen is typically stored in pressurized tanks, either as gas at ambient temperatures or as cryogenic liquid. However, both systems suffer from deficient safety and insufficient energy density and are not considered as satisfying future solutions [7].

To meet the demands set on hydrogen storage new solutions must be considered. Many promising alternative storage technologies based on mechanisms where hydrogen is interacting chemically and/or physically with solid materials have been under investigation during recent years. Their main advantage is the dense hydrogen

compaction, i.e. high energy densities. The materials considered for hydrogen storage are commonly categorized into three major groups; metal hydrides, complex hydrides and hydrogen adsorbents:

(i) Metal hydrides are formed when elemental metals or alloys absorb hydrogen interstitially. The materials have gained huge interest among researchers due to their great stability and generally good reversibility under mild conditions.

(ii) Complex hydrides are ionic compounds and can be written on the general formula A^x+[BH_z]^y-, where A is a group I or II element and B usually is a transition metal, but can also be a non-metallic element. Some well known examples are NaAlH4 and LiBH4 and their major advantage is the high energy density.

(iii) Hydrogen adsorbents are low weight, mainly carbon based, nano-designed materials with large surface areas that physically adsorb hydrogen. Their low weight and high capacity are promising properties for hydrogen storage materials [8].

Fig. 1. Categorization of different hydrogen storage technologies.

Even though these types of materials have promising attributes few have reached a stage of practical usage. They often suffer from one or several drawbacks, whether it is low capacity, poor reversibility, high operation temperatures or slow kinetics. However, the fields are still far from fully explored which is why further research on new hydrogen storage materials is of huge interest and great importance.

Hydrogen storage technologies

Metal hydrides

Cryogenic liquid Complex hydrides Hydrogen adsorbents

Pressurized gas

2. Metal hydrides

The ability of metals to absorb hydrogen has been known since the mid-nineteenth century with the discovery of palladium hydride (PdHx) [9]. In 1969 researchers at the Philips Eindhoven Laboratories in the Netherlands accidently discovered SmCo5, the first known alloy capable of absorbing hydrogen reversibly at low temperatures [10]. This became the starting point of extensive metal hydride research which soon lead to the discovery of the classical LaNi5 alloy [11]. Today its derivatives, i.e. different AB5

compounds, are widely used in Ni-MH batteries. However, while these compounds fulfill all criteria‟s in terms of reversibility, stability and kinetics, the hydrogen storage capacity is only 1.49 wt.% which is too low for automobile applications [3].

2.1 Overview

There are many different metallic systems that can accommodate hydrogen. Numerous elemental metals, especially the d- and f- shell transition metals, are able to form stable binary hydrides such as Pd (PdH0,6), Mg (MgH2), Sc (ScH2) and Y (YH2, YH3) [12].

However, due to their stability they often suffer from high desorption/decomposition temperatures and slow kinetics not suitable for practical applications. This problem can be solved by alloying with other elements to gain e.g. catalytic or destabilizing effects as seen in Mg2Cu/MgCu2 [13] and Mg24Y5 [14] where kinetics were improved compared to pure MgH₂. The hydrogen absorbing properties of intermetallic compounds very much depend on the elemental composition. Generally the compound must contain at least one element which itself is capable of forming hydride phases. The role of the other

components is to beneficially influence the hydrogen absorption properties and hydride stability. Various metals such as Al, Ni, Cu, Ti, Fe as well as some p- elements have been extensively used as constituents in promising hydride systems.

The crystal structure also plays a critical role for the absorption properties. Stability of hydrides has been correlated with crystallographic factors such as the size of the interstices in the host lattice which must be 0.4 Å or larger [15]. The distance between hydrogen atoms in stable hydrides is always more than 2.1 Å and they preferentially occupy the interstitial sites closest to the hydride forming element [16]. Beside the mentioned AB5 compounds there are numerous structures of binary, pseudo-binary and ternary crystals which have shown to be suitable for hydrogen absorbtion, such as AB₂, AB, A2B, AB1-xCx and ABC. Some particular examples are ZrV2 (ZrV2H6.5), FeTi (FeTiH2), Mg2Ni (Mg2NiH4), ScAl0.8Mg0.2 (ScAl0.8Mg0.2H2) and TbNiAl (TbNiAlH1.4) respectively [12, 26, 24]. However, with increasing number of components the systems gets more complex and difficult to interpret. It is thus difficult to predict how a given multi-component compound will react with hydrogen without performing experimental studies. Practically all metal hydrides have unique properties and thus every system has to be individually investigated and optimized after its own conditions.

2.2 Hydrogen absorption mechanisms

Fig. 2. Schematic representation of hydrogen absorption in metal hydrides. Hydrogen gas dissociates at the metallic surface and diffuses into the material in an atomic state. At low pressures a solid solution α–

phase is formed while at higher pressures a more ordered β-phase may be obtained.

The hydrogen absorption mechanism in metals and alloys is schematically illustrated in figure 2. Initially molecular hydrogen is physically adsorbed onto the surface of the material by van der Waals forces, as seen on the shaded area in the figure. The metallic surface catalyzes the dissociation of hydrogen molecules which in turn diffuses into the material in an atomic state. Hydride formation is commonly described as a two step process:

(i) At low concentrations (low pressures) the hydrogen is dissolved in the crystal as a solid solution (α-phase). In this state the atoms occupy interstitial positions randomly which result in a small expansion of the original host lattice. The expansion is locally somewhat larger over a certain volume from each hydrogen atom and makes the nearest vacant sites more favorable for other hydrogen atoms to occupy.

(ii) As the hydrogen pressure, and consequently concentration, increases the interaction between hydrogen and metal atoms becomes more important and a more ordered phase starts to grow (β-phase). The formation of β-phases generally increases the host lattice substantially and the metal atoms often undergo major rearrangements which in some cases also may lead to stabilization of new crystal structures. However, the β-phase does not necessary appear in every hydrogen accommodating material. Its existence depends on each unique system, i.e. the combination and arrangement of its component elements [17,18].

H

gas α-phase β-phase

2.3 Thermodynamics

Fig. 3. Schematic image of a pressure-concentration-temperature plot (left) and a van’t Hoff plot (right) describing the thermodynamic aspects of hydride formation.

The hydrogen absorption process can be described from a thermodynamic point of view by a pressure-concentration-temperature (PCT) and van't Hoff plot (figure 3). The PCT plot shows how the absorption is related to the applied hydrogen pressure at isothermal conditions. In the single phase regions large pressure changes are needed to significantly increase the hydrogen concentration. However, at a specific pressure an equilibrium state between α-and β-phases is obtained, seen as a flat plateau, where large amounts of hydrogen is absorbed with small pressure variations. The equilibrium pressure, Peq, is dependent on the temperature, T, through the van‟t Hoff equation where P0 is the atmospheric pressure, ΔH and ΔS are the changes in enthalpy and entropy respectively and R is the gas constant:

ln(P

/P

) = ΔH/(RT) – ΔS/R

The reaction of hydrogen with metals is usually exothermic, i.e. the heat of formation (ΔH) is negative, and thus the plateau pressure increases with temperature. As the temperature increases the absorption capacity decreases and ultimately drops to

theoretical minimums above a critical temperature Tc where the transition between α- and β-phase becomes continuous. The entropy of formation mainly corresponds to the change from molecular hydrogen gas to dissolved hydrogen atoms and is roughly 130 JK^-1mol^-1 for all metal-hydrogen systems. The enthalpy term characterizes the stability of the metal-hydrogen bond and is an important factor to take into consideration when

evaluating whether or not a specific compound is suitable as a storage material [3]. The hydride should be stable enough to be formed, but labile enough to desorb hydrogen at moderate conditions. To find materials with these properties still remains as one of the most challenging tasks in present hydride research.

1/Temperature

log(H2 equilibrium pressure)

Hydrogen/Metal

log(H2 equilibrium pressure)

T3

T2

T1

TC

α-phase α + β-phase

β-phase

T3 > T2 > T1

2.4 Alloys with scandium and aluminium

Scandium is highly capable of accommodating hydrogen, either in its elemental state by the formation of ScH₂, or as a component in intermetallic compounds such as ScFe₂ (ScFe2H2) and Sc2Ni (Sc2NiH5) [12, 19, 20, 36]. Elemental aluminium does not react with hydrogen, although amorphous AlH3 can be produced by reaction between LiAlH4

and AlCl₃ in ether [21]. Aluminium has however shown to be beneficial as a component in intermetallic compounds with high hydrogen affinity, such as REAlNi, RE3Ni8Al and RE3Ni6Al2 (RE = rare earth metal) [26].

The usability of Sc-Al alloys as hydrogen storage materials has been discussed in several papers with divergent opinions. Burkhanov et al. have showed that Sc2Al tend to

decompose into ScH2 and Al during hydrogenation rather than forming a hydride as a single phase compound. This is due to the relatively low heat of formation of Sc₂Al, 7 kcal/mol, compared to ScH2, 50 kcal/mol, which is favoring the decomposition process [20,22]. The reaction was not reversible at temperatures below 1200 °C. They claim that this prevents the compound from being used for hydrogen storage applications [20].

However, Sahlberg et al. have showed that the decomposition temperature of ScH2

gained from hydrogenations of ScAl1-xMgx-compounds can be lowered to 500 °C by destabilization effects with full hydrogen reversibility for several cycles [23,24].

Antonova et al. have also showed that all binary phases in the Sc-Al system are able to absorb hydrogen and form single phase hydrides to different extent [25].

3. Aim

The aim of this project has been to investigate the hydrogen absorption properties of two different scandium- and aluminium based alloys, ScAlNi and ScAl1-xSix, which have not been studied for this purpose before to the best of my knowledge. The choice of these specific alloys was based on promising studies reported on similar systems.

REAlNi compounds with the Fe2P-type structure have shown to absorb substantial amounts of hydrogen (1-1.4 H/f.u.) at ambient pressures for RE = Y, Gd, Tb, Dy, Er, Lu and Zr, [26, 27]. Scandium, the lightest 3d transition metal, forms stable dihydride with 4.4 wt % but have not been investigated in this particular intermetallic system.

Preliminary results have shown that the ternary ScAlNi could have interesting hydrogen absorption properties. The aim was to perform more thorough investigations of the compound and its hydrogen absorption properties.

ScAl with the CsCl-type structure have shown to absorb fairly high amounts of hydrogen at pressures of 2.5 MPa [25]. Silicon and phosphorus additions to intermetallic

compounds have shown to influence the hydrogen absorption properties by lowering the absorption equilibrium pressure. The aim was to investigate possible effects on hydrogen absorption properties in ScAl by additions of Si (ScAl1-xSix).

4. Experimental

4.1 Synthesis

All samples were prepared by using direct reactions between appropriate amounts of the elements (Sc: 99.99 %, Stanford Materials Corporation, Batch no. Qx-05; Al: 99.999%, Gränges Essem; Ni: 99.999 %, Alfa Aesar, Batch no. Nm11879; Si: 99.999 % single crystal, Highway). Two different high temperature synthesis techniques were used, arc melting and induction melting.

Arc melting is a common technique used for alloy synthesis. The samples are melted in an inert atmosphere of argon by direct contact with a plasma arc. The arc is produced by striking a current from a charged electrode to a grounded metallic plate.

The arc furnace used in this project was equipped with a tungsten electrode and water cooled copper heart. The reaction chamber was initially flushed with high purity argon several times and held at a constant pressure of 50 bar during the synthesis. To react possible oxygen residuals a piece of titanium metal was melted as a getter for 5 minutes prior to the sample melting. All samples were re-melted five times to ensure complete melting and good homogeneity.

Induction melting is another clean, frequently used, technique for alloy synthesis. The samples are placed in the middle of a solenoid through which an alternating current is passed at a suitable frequency for the specific material type, sample size and desired temperature. Eddy currents are generated within the sample through electromagnetic induction which heats the material. The melting is performed in vacuum.

To ensure a clean reaction environment the samples in this project were sealed into tantalum metal ampoules in argon atmosphere. The samples were heated to and kept at the desired temperature for several minutes to ensure complete melting. The temperature was measured by a pyrometer on the surface of the ampoule.

All samples were crushed and grounded to a fine powder. Some samples were

additionally heat treated in evacuated silica ampoules at temperatures of 800-1000 °C for several days.

Fig. 4. From the left: pieces of scandium, aluminium, nickel, silicon and arc melted ScAlNi.

4.2 Hydrogen absorption/desorption

Hydrogen absorption and desorption reactions were performed in a high pressure furnace.

The grounded sample was placed in an alumina crucible which in turn was placed in a high pressure resistant metal casing connected to a hydrogen gas source. The metal casing was sinkable into a pit furnace for temperature control of the system. The system was thoroughly flushed with hydrogen prior to experiments. Hydrogen absorptions were performed at different temperatures and pressures ranging between 20 to 500 °C and 10 to 50 bar respectively. Desorptions were performed at similar temperatures in vacuum (7*10^-2 mbar). By monitoring the hydrogen pressure as a function of temperature the absorption/desorption temperature could be determined. The absolute amount of absorbed/desorbed hydrogen was estimated by the total pressure drop in the system in combination with gravimetric measurements. Hydrogen content in some samples was also estimated by Peisl‟s relationship between hydrogen absorption and unit cell expansion for α-phase hydrides (1H/unit cell expansion of 2.9 ų) [32]

4.3 Thermal desorption spectroscopy (TDS)

In thermal desorption spectroscopy partial pressures of gaseous compounds are measured as a function of temperature. Detection is done by a mass spectrometer which enables for qualitative pressure determination of several substances simultaneously. In metal hydride research TDS measurements are often used in combination with Kissinger analysis to evaluate the activation energy for the desorption reaction. The temperature of maximum desorption, T_max, depends on the heating rate, β, of the sample. The activation energy, E_a, is determined according to the Kissinger equation:

E

/R = -dln(β/T

2

)/d(1/T

)

In this project the TDS-measurements were performed in an ultra high vacuum (UHV) furnace at pressures of 10^-12 bar. A Microvision Plus residual gas analyzer/mass spectrometer was used for detection.

4.4 Characterization

4.4.1 X-ray diffraction (XRD)

X-ray diffraction (XRD) is a very powerful and commonly used analytical technique for characterization of crystalline materials. The technique relies on coherent elastic

scattering, termed diffraction, of x-rays in crystal structures with long range atomic order.

The conditions under which diffraction occurs are described by Bragg´s law nλ = 2dsin(θ)

where n is an integer, λ is the wavelength of the radiation, d is the spacing between atomic planes and θ is the diffraction angle (figure 5). The diffraction gives rise to a set of well defined beams, referred to as reflections, whose intensities and positions in space are dependent on the atomic numbers and arrangement, i.e. the crystal structure. Thus each unique material gives rise to a unique set of reflections. The information gained from x-ray diffraction measurements is visualized in diffractograms where the reflection intensities are presented as a function of the diffraction angle.

Since the reflection intensities are approximately directly proportional to the atomic number hydrogen gives the weakest signals of all elements. Thus, in metal hydride research XRD is used for observing changes of the host lattice rather than locate the positions of hydrogen atoms.

The X-ray powder diffraction measurements in this project were performed on a Bruker D8 diffractometer equipped with a Våntec position sensitive detector (4° opening) using CuKα1-radiation in a standard θ-2θ locked-couple setup. Diffraction data was recorded for 2θ = 10-90 degrees. Unit cell parameters were refined with the program Unitcell.

Fig. 5. The geometry of Bragg’s law for the diffraction of x-rays from a set of crystal planes (hkl) with

dhkl

2θ θ

Diffracted beam

Undiffracted beam X-ray beam

Atom planes (hkl)

4.4.2 Scanning electron microscopy (SEM)

Scanning electron microscopy is a technique used for generating highly magnified images with high resolutions and large depths of field. It is commonly used for

microstructural analysis. A focused electron beam is scanned over the sample surface in vacuum. The electrons interact with near-surface atoms in the sample whereupon detectable signals such as backscatter electrons (BSE), secondary electrons (SE), characteristic x-rays and auger electrons are emitted. For image creation usually the secondary electrons are detected while backscattered electrons are mainly used for chemical information. Each “scanning point” on the sample corresponds to a pixel in the generated image. The resulting image is a distribution map of the intensity of the signals being emitted from the scanned area. The obtained contrast in secondary electron

imaging (SEI) is strongly related to the surface topography but is also dependent on factors such as the atomic numbers in analysis of multi-elemental samples. In this project a scanning electron microscope of model LEO 1550 was used.

5. Results & discussion

5.1 XRD characterization of synthesized samples 5.1.1 ScAlNi

Fig. 6. XRD patterns of arc melted ScAlNi (upper pattern) and additionally annealed sample at 800°C for three days (lower pattern). The x-labeled tics marks the expected Bragg positions received from unit cell refinements.

The arc melted ScAlNi was single phase according to the X-ray diffraction (XRD) patterns. Unit cell refinements were performed on the annealed sample and the crystal structure was identified to the MgZn2-type (C14 Laves phase), space group P63/mmc, with cell parameters a = 5.1434(1) Å, c = 8.1820(2) Å. The x-labeled tics in figure 6 marks the expected Bragg positions for this structure.

The fast cooling rate in the arc furnace induced internal stresses in the material which could be seen as fairly broad diffraction peaks (upper pattern in figure 6). A certain asymmetry in the peaks was also observed which probably originate from an inhomogeneous crystallization leading to a composition gradient in the sample.

Additional annealing at 800 °C for 3 days in evacuated silica ampoules allowed for thermal relaxation and a more ordered sample was obtained, seen as more well-defined peaks (lower pattern in figure 6). However, after annealing a few unidentified peaks appeared in the diffraction pattern; the largest at 2θ ≈ 11.5° and 2θ ≈ 25° (lower pattern, figure 6). Initial hydrogenation experiments revealed that non-annealed samples had significantly better hydrogen absorption properties compared to annealed samples where

Intensity (a.u.)



11 20 30 40 50 60 70 80 90

the new peaks originate from oxides formed on the material surface, acting as a physical barrier. The distinct differences in absorption properties could also be a result of

enhanced diffusion in the non-annealed and more disordered samples due to higher defect concentrations. Due to the deteriorating absorption properties of additionally annealed samples all hydrogenation experiments presented in this thesis were performed on non- annealed samples.

ScAlNi did not crystallize in the Fe₂P-type structure which has shown to have high hydrogen affinity for most REAlNi compounds [26, 27]. The MgZn2-type structure has however also attracted attention as a beneficial structure for hydrogen absorption as seen in different binary AB₂-compounds such as ZrMn₂ [28, 29].

5.1.2 ScAl_1-xSi_x

Fig. 7. XRD patterns of arc melted ScAl1-xSix-samples with additional annealing at 900 °C for three days.

Upper and lower pattern shows ScAl1-xSix, x = 0.1 and x = 0 respectively. The x- and ●-labeled tics marks the Bragg positions of orthorhombic ScAl and cubic ScAl2 respectively.

The XRD patterns of arc melted and additionally annealed ScAl_1-xSi_x revealed multiphase samples. Two phases, orthorhombic ScAl [30] and cubic, MgCu₂-type, ScAl₂ [31] could be identified from the diffraction patterns for ScAl1-xSix, x = 0 and x = 0.1, upper and lower pattern respectively in figure 7. The additional peaks could not be identified.

Intensity (a.u.)



20 30 40 50 60 70 80

Fig. 8. XRD patterns of induction melted ScAl1-xSix. Upper and lower pattern shows ScAl1-xSix, x = 0.1 and x = 0 respectively. x- and ●-labeled tics marks the Bragg positions of cubic ScAl and cubic ScAl2

respectively.

Induction melted and additionally annealed ScAl_1-xSi_x, x= 0, crystallized in the CsCl-type structure according to the XRD pattern (lower pattern in figure 8). The additional peaks originated from small amounts of cubic, MgCu2-type, ScAl2 [31].

The diffraction pattern of induction melted and annealed ScAl1-xSix, x= 0.1, revealed that multiple phases were obtained (upper pattern in figure 8). Cubic ScAl with CsCl-type structure was identified as the main phase together with substantial amounts of cubic, MgCu2-type, ScAl2 [31]. The additional peaks could not be identified.

The results indicated that Si and Al are not interchangeable in this ternary system. Since the desired pseudo-binary CsCl-type ScAl_1-xSi_x could not be obtained the effects of Si additions on hydrogen absorption properties in ScAl would not be possible to determine.

No further hydrogenation investigations were thus performed on these compounds.

Intensity (a.u.)



20 30 40 50 60 70 80

5.2 Hydrogenation of ScAlNi

ScAlNi was found to absorb hydrogen by two distinctly different mechanisms at different temperature regions. The results are therefore divided into different temperature and phase sub-categories.

5.2.1 Hydrogenation at lower temperatures

Fig. 9. A typical hydrogenation plot of ScAlNi at lower temperatures. The sample was hydrogenated at

~10.2 bar at a heating rate of 5 °C/min. A clear pressure drop could be seen starting at ~120 °C and continuing to ~160 °C. The box in the lower right corner shows a scaled down background reference measurement performed at the same pressure and temperature conditions for comparison.

ScAlNi was found to absorb hydrogen at temperatures of ~120 °C. A typical

hydrogenation plot is showed in figure 9. Hydrogen was absorbed between ~120 °C and

~160 °C during heating from 22 °C to 280 °C at a rate of 5 °C/min.

The final pressure difference, combined with gravimetrical measurements, revealed hydrogen uptakes up to ~0.24 wt%, or ~0.31 H/f.u. ScAlNi at pressures of 50 bar.

Examples of estimated hydrogen content based on experimental results from hydrogenations at 10 and 50 bar are presented in table 1.

9,8 10 10,2 10,4 10,6 10,8 11 11,2 11,4

0 50 100 150 200 250 300

Temperature (°C) H2 pressure (bar)

Fig. 10. Comparison between XRD patterns of pure ScAlNi (blue) and hydrogenated samples at 10 bar (red) and 50 bar (green). Significant unit cell expansions were observed.

Hydrogen absorption was confirmed by XRD measurements where significant unit cell expansions were observed (figure 10). Crystallographic parameters of pure ScAlNi and hydrogenated samples at different pressures are presented in table 1.

The relatively small hydrogen capacities and no phase transitions during hydrogenation indicated the formation of solid solution α-phase hydrides at the investigated

temperatures and pressures. The hydrogen content was additionally estimated based on Peisl‟s finding that the solid solution of hydrogen in a variety of metals and alloys causes an volume expansion, ΔV, averaging 2.9 ų per hydrogen atom [32].

Table 1. Chrystallographic characteristics of ScAlNi and its hydrides. Estimated hydrogen content based on hydrogenation pressure changes and unit cell expansions.

ScAlNi samples a-axis (Å)

c-axis (Å)

Unit cell volume (ų)

Hydrogen content (wt%; H/f.u.)*

Hydrogen content (wt%; H/f.u.)**

Mother compound 5.1434(1) 8.1820(2) 216.45 - -

Hydrogenated at 10 bar 5.1645(2) 8.2859(8) 221.00 0.15; 0.20 0.30; 0.39 Hydrogenated at 50 bar 5.1763(1) 8.3049(4) 222.52 0.24; 0.31 0.40; 0.52

* Estimated hydrogen content based on experimental results

** Estimated hydrogen content according to Peisl‟s relationship between hydrogen absorption and unit cell expansion [32]

Intensity (a.u.)



20 30 40 50 60 70

5.2.2 Hydrogenation at higher temperatures

Fig. 11. Typical hydrogenation plot of ScAlNi at higher temperatures. The sample was hydrogenated at

~50 bar at a temperature of 485 °C. An initial small pressure drop was seen at ~120 °C. At temperatures above 300 °C a second, much larger, pressure drop was observed.

ScAlNi was found to absorb substantial amounts of hydrogen at higher temperatures.

During temperature increase the material started to absorb hydrogen at ~120 °C which corresponded to the formation of the α-phase. At temperatures above 300 °C and up to 485 °C a much higher absorption occurred. The final pressure difference at room temperature was roughly six times larger compared to hydrogenations at lower temperatures.

48 49 50 51 52 53 54 55 56 57 58

0 100 200 300 400 500 600

Temperature (°C)

Hydrogen pressure (bar)

Fig. 12. XRD patterns of (a) synthesized ScAlNi and hydrogenated samples at (b) 10 bar, (c) 30 bar, and (d) 50 bar. All samples were heated to and dwelled at 485 °C for ~6 hours. (x)-, (●)- and (^▼)- labeled tics marks the Bragg positions of ScAlNi, ScH2 and AlNi respectively.

XRD analysis of hydrogenated ScAlNi at 50 bar and 485 °C revealed a virtually

complete phase transition of the initial hexagonal phase into cubic ScH₂ and AlNi. Only traces of the two main ScAlNi-peaks at 2θ ≈ 41.5° and 42.2° were detected. The broad diffraction peaks of the new phases indicated that the grain sizes were very small. The results indicate that the hydrogen absorption occurred in a one-step reaction according to the scheme:

ScAlNi + H2 → ScH2 + AlNi

ScH2 crystallizes in the CaF2-type structure (cubic, Fm-3m) with the cell constant a = 4.7832(4) Å [33]. Hydrogen occupies the tetrahedral sites (¹/4, ¹/4,¹/4) in the cubic close packed structure of Sc. AlNi crystallizes in the CsCl-type structure (cubic, Pm-3m) with the cell parameter a = 2.880 Å [34].

The complete consumption of ScAlNi and formation of ScH₂, as seen in the upper diffractogram in figure 12, gives the material a theoretical hydrogen storage capacity of 1.53 wt% (2 H/f.u.). This capacity was confirmed by gravimetric measurements.

(a) (b) (c) (d)

Intensity (a.u.)



20 30 40 50 60 70

5.3 Hydrogen desorption of ScAlNi-Hx

5.3.1 Hydrogen desorption and cycling of α-phase hydride

Fig. 13. Cycling of ScAlNi α-phase hydride; (a) hydrogenation, cycle 1, (b) hydrogen desorption, cycle 1, (c) hydrogenation, cycle 2, (d) hydrogen desorption, cycle 2, (e) hydrogenation, cycle 3, (f) hydrogen desorption, cycle 3.

Cycling of ScAlNi at ~10.3 bar and ~300 °C showed to improve the

absorption/desorption kinetics (figure 13). In the first cycle hydrogen absorption occurred in a narrow and distinct temperature region between ~120 °C and ~160 °C, while in the following cycles absorption occured continuously throughout the entire heating process and started at lower temperatures. Desorption temperatures were clearly decreased from

~170 °C in the first cycle to ~100 °C in the third cycle.

10 10,2 10,4 10,6 10,8 11 11,2 11,4

0 50 100 150 200 250 300

Temperature (°C) H2 pressure (bar)

10 10,2 10,4 10,6 10,8 11 11,2 11,4

0 50 100 150 200 250 300

Temperature (°C) H2 pressure (bar)

0 0,05 0,1 0,15 0,2 0,25

0 50 100 150 200 250 300

Temperature (°C) H2 pressure (bar)

10 10,2 10,4 10,6 10,8 11 11,2 11,4

0 50 100 150 200 250 300

Temperature (C)

H2 pressure (bar)

0 0,05 0,1 0,15 0,2 0,25

0 50 100 150 200 250 300

Temperature (°C) H2 pressure (bar)

0 0,05 0,1 0,15 0,2 0,25

0 50 100 150 200 250 300

Temperature (°C) H2 pressure (bar)

a b

c d

e f

Table 2. Hydogen absorption during cycling showing the capacity loss.

Cycle (nr)

Hydrogen absorption (wt%; H/f.u)

1 0.14; 0.18

2 0.11; 0.14

3 0.10; 0.13

The cycling also showed that the α-phase hydride was partially reversible. The storage capacity decreased to ~78 % and ~72 % of the initial value after the first and second cycle respectively (showed in table 2) due to incomplete desorption under the

investigated conditions. Based on the cycling data the material seemed to approach a constant reversible capacity after the first cycle where the major capacity loss occurred.

5.3.2 Thermal desorption spectroscopy of disproportionated ScAlNi

Fig. 14. XRD patterns of the (a) synthesized, (b) hydrogenated and (c) desorbed sample showing the reversibility.

Thermal desorption spectroscopy (TDS) measurements were performed on the

disproportionated ScAlNi, i.e. ScH2 + AlNi, to evaluate the desorption temperature and corresponding activation energy. XRD analysis of desorbed samples revealed that the mother compound was recombined after dehydrogenation (figure 14). The material was thus shown to be fully reversible with respect to hydrogen storage and to possess the hydrogenation-disproportionation-desorption-recombination behaviour (HDDR) commonly seen in Nd-Fe-B permanent magnets [35].

Intensity (a.u.)



20 30 40 50 60 70

(c)

(b)

(a)

Fig. 15. TDS spectra showing hydrogen desorption peaks at different heating rates.

TDS- measurements (figure 15) showed that hydrogen desorption started at ~300 °C in small amounts and then occurred in two distinct steps between ~450 °C and ~550 °C. The main peak of maximum desorption was observed at a temperature of ~500 °C and did most probably originate from the recombination reaction further described below. The origin of the secondary peak was not investigated but one possible explanation of its presence is that some hydrogen was dissolved in the recombined ScAlNi crystal which lead to a retarded desorption.

The observed desorption temperature of ~500 °C is remarkably low as pure ScH2 has a reported decomposition temperature of 960 °C [36]. Lowering of the decomposition temperature for ScH2 has however been observed for HDDR-behaving ScAlxMg1-x- compounds where it was described as a destabilization by the co-existing AlMg-phase [24]. The destabilization phenomenon was supported by theoretical calculations [23]. It was suggested that the small grain size in the disproportionated sample also may have played an important role for the reversibility. Adopting this model for ScAlNi imply that the ScH2 was destabilized due to the presence of the AlNi-phase.

The results observed for ScAlNi and reported for ScAlxMg1-x [23,24] are in contrast to what have been observed for Sc2Al which was not reversible at temperatures below 1200

°C after disproportionation into ScH₂ and Al.

300 400 500 600

Temperature (°C) H2 partial pressure (a.u.)

Fig. 16. Kissinger plot for the main desorption peak of ScAlNi β-phase hydride.

Kissinger analysis was used to determine the activation energy for the recombination reaction by using the temperatures of the main desorption peaks in the TDS spectra. The recombination reaction is defined as the reverse absorption reaction, i.e. desorption of hydrogen gas and reformation of ScAlNi from ScH2 and AlNi.

ScH₂ + AlNi → ScAlNi + H₂

The activation energy for the desorption reaction was determined to 182 kJ/mol (figure 16). This is very similar to the activation energy for recombination of hydrogenated ScAl0.8Mg0.2 compounds, 185 kJ/mol. The corresponding value for pure ScH2 has not been found in the literature, but the decomposition of pure YH2 has a reported activation energy of 272 kJ/mol [37] as a comparison.

-13,5 -13 -12,5 -12 -11,5 -11 -10,5

1,23 1,24 1,25 1,26 1,27 1,28 1,29 1,3 1,31 1,32 1,33 1,34 1000/Tm (K^-1)

ln(β/Tm2 )

5.4 Scanning electron microscopy of ScAlNi

Fig. 17. SEM pictures showing the particle size of ScAlNi at two different magnifications after different treatments. Picture (a) and (b) shows pure ScAlNi sample. Picture (c) and (d) shows disproportionated sample of ScAlNi into ScH2 and AlNi. Picture (e) and (f) shows recombined ScAlNi after complete desorption.

Scanning electron microscopy (SEM) investigation of pure, disproportionated and recombined ScAlNi showed no change in particle size during the first

absorption/desorption-cycle. The result was unanticipated since the HDDR behavior has shown to reduce the particle size during hydrogen absorption and desorption leading to improved kinetics [35]. The wide size distribution in the pure sample complicated the determination of particle size changes to some extent. More cycles might also be necessary to see significant changes of the particle size.

a b

c d

e f

100μm

100μm

100μm

10μm

10μm

10μm

6. Summary

Some hydrogen absorption properties of arc melted ScAlNi, crystallizing in the MgZn2- type structure (C14 Laves phase), have been investigated. The compound was found to absorb hydrogen by two distinctly different mechanisms at different temperatures.

At low temperatures, ~120 °C, hydrogen was dissolved in a solid solution α-phase causing significant unit cell expansions. The hydrogen absorption capacity after

hydrogenations at 50 bar was estimated to ~0.24 wt% (ScAlNiH_0.31). A second estimation based on the unit cell expansion suggested a hydrogen capacity of ~0.40 wt%

(ScAlNiH0.52). Cycling showed to significantly improve the kinetics but did also reveal capacity losses up to 30 % after 3 cycles.

At higher temperatures, ~500 °C, the compound absorbed hydrogen with a storage capacity of 1.53 wt% by disproprtionation into ScH₂ and AlNi. The reaction was shown to be fully reversible and the material showed to possess the HDDR behavior. The decomposition temperature of ScH2 was lowered by 460 °C due to destabilization effects by the coexisting AlNi phase. The activation energy of the desorption reaction was determined to 182 kJ/mol by TDS measurements. The particle size was not affected by the first absorption/desorption cycle.

Attempts of synthesizing pseudo-binary ScAl_1-xSi_x, 0 < x < 0.1, for hydrogen absorption investigations has been performed. Both arc melting and induction melting resulted in multiphase compounds due to poor exchangeability between Si and Al in this specific ternary system. No further investigations of the effects of Si-additions to ScAl on hydrogen absorption properties could thus be done.

7. Conclusions

The investigations showed that ScAlNi does not crystallize in the Fe2P-type structure as most other RENiAl intermetallic compounds which previously has shown to have high hydrogen affinities. The obtained MgZn2-type structure (C14 Laves phase) of ScAlNi was found to absorb relatively small amounts of hydrogen by solid solution. The study also confirmed previously reported results which states that scandium and aluminium based compounds tend to decompose into the stable ScH₂ at higher temperatures and hydrogen pressures. However, in contrast to some other compounds such as Sc2Al, the disproportionation of ScAlNi into ScH2 and AlNi was found to be fully reversible at moderate conditions which show the possibility of finding new materials for hydrogen storage with alternative storage mechanism. The hydrogen absorption properties of ScAlNi are not good enough for practical usage and the price for scandium is too high to be interesting in commercial applications. However, the results in this thesis contribute to a better understanding of some important hydrogen absorption properties of scandium and aluminium based compounds which with further research can lead to modifications of the material that releases hydrogen at lower temperatures. The results may also provide

8. Acknowledgements

I would like to thank my supervisors Yvonne Andersson and Martin Sahlberg for the opportunity to work on this project and for the help and support during the journey. It has been six rewarding and incredibly fun months. I am also deeply grateful for the

opportunity to attend at the truly inspiring hydrogen storage conferences in Vilnius and at Svalbard. Financial support from Nordic Energy Research Project is gratefully

acknowledged.

I would like to thank Mikael Ottoson for the support in XRD and TDS measurements, and Anders Lund for the help with solving various problems in the basement laboratory.

Thanks to reviewer Annika Pohl and examiner Gunnar Westin. Special thanks go to Fredrik for many fun moments at the office, on strolls to Möbius and lunches at Rullan.

Last but not least, many thanks also go to all friends and co-workers for making my stay at the department of Materials chemistry enjoyable and memorable.

9. References

[1] International Energy Agency (IEA) – Key World Energy Statistics, (2009).

[2] European Commission, EUR 20719 EN – Hydrogen Energy and Fuel Cells – A vision of our future, (2003).

[3] L. Schlapbach, A. Zuttel, Hydrogen-storage materials for mobile applications, Nature 414 (2001) 553-558.

[4] Lovins, A. B. Twenty Hydrogen Myths; Rocky Mountain Institute: Snowmass, (2003).

[5] J. D. Holladay, J. Hu, D. L. King, Y. Wang, An overview of hydrogen production technologies, (2009).

[6] M. L. Ghirardi, L. Zhang, J. W. Lee, T. Flynn, M. Siebert, E. Greenbaum, A. Melis, Microalgae: a green source of renwable H2, Trends in biotechnology, 18 (2000) 506- 511.

[7 ] D. Mori, K. Hirose, Recent challenges of hydrogen storage technologies for fuel cell vehicles, International journal of hydrogen energy, 34 (2009) 4569-4574.

[8] Hui-Ming Cheng, Quan-Hong Yang, Chang Liu, Hydrogen storage in carbon nanotubes, Carbon 39 (2001) 1447-1454.

[9] T. Graham, Journal of the chemical society 20 (1867) 235.

[10] J. H. N. van Vucht, F.A. Kuijpers, H. C. A. M. Bruning, Reversible room- temperature absorption of large quantities of hydrogen by intermetallic compounds, Philips. Res. Repts. 25 (1970) 133-140.

[11] J. H. N. Van Vucht, F.A. Kuijpers, H.C.A.M. Bruning, Philips research reports, 25 (1970) 133.

[12] G. Sandrock, G. Thomas, The IEA/DOE/SNL on-line hydride databases, Appl. Phys.

A 72 (2001) 153–155.

[13] A. Andreasen, M. B. Sorensen, R. Burkarl, B. Moller, A. M. Molenbroek, A. S.

Pedersen, T. Vegge, T. R. Jensen, Dehydrogenation kinetics of air-exposed

MgH2/Mg2Cu and MgH2/MgCu2 studied with in-situ X-ray powder diffraction, Appl.

Phys. A 82 (2006) 515-521.

[14] C. Zlotea, M. Sahlberg, S. Özbilen, P Moretto, Y Andersson, Hydrogen desorption studies of the Mg24Y5-H system: Formation of Mg tubes, kinetics and cycling effects, Acta Materialia 56 (2008) 2421-2428.

[15] C. E. Lundin, F. E. Lynch, C. B. Magee, J. Less-Common Met. 56 (1977) 19.

[16] A. C. Switendick, Z. Phys. Chem. N. F. 117 (1979) 89.

[17] G. K. Shenoy, B. D. Dunlap, P. J. Viccaro, D. Niarchos, Hydrogen absorption mechanism and location in intermetallic compounds, Hyperfine interactions 9 (1981) 531-546.

[18] S. Rundqvist, R. Tellgren, Y. Andersson, Hydrogen and deuterium in transition metal-p element compounds: crystal chemical aspects of interstitial solid solunility and hydride phase formation, J. Less-Common metals, 101 (1984) 145-168.

[19] P. H. Smith, K. H. J. Buschow, ⁵⁷Fe Mössbauer effect in ScFe2 and ScFe2H2, Phys.

Rev. B 21 (1980) 3839.

[20] G. S. Burkhanov, N. B. Kol‟chugina, O. D. Chistyakov, V. N. Verbetsky, A. A.

Salamova, E. Yu. Andreeva, E. S. Volkova, Interaction of hydrogen with internetallic compounds Sc₂Al and Sc₂Ni, Inorganic materials 42 (2006) 491-495.

[21] F. M. Brower, N. E. Matzek, P. F. Reigler, H. W. Rinn, C. B. Roberts, D. L.

Schmidt, J. A. Snover and K. Terada, Preparation and properties of aluminium hydride, J. Am. Chem. Soc. 98 (1976) 2450–2453.

[22] C. Colinet, The thermodynamic properties of rare earth metallic systems, J. Alloys and Compounds 225 (1995) 409-422.

[23] M. Sahlberg, P. Beran, T. Kollin Nielsen, Y. Cerenius, K. Kadas, M. P. J.

Punkkinen, L. Vitos, O. Eriksson, T. R. Jensen, Y. Andersson, A new material for hydrogen storage, ScAl0.8Mg0.2, Accepted for publication in Journal of Solid State Chemistry.

[24] M. Sahlberg, C. Zlotea, M. Latroche, Y. Andersson, Fully reversible hydrogen absorption and desorption reactions with Sc(Al1-xMgx), x = 0.0, 0.15, 0.2, manuscript published in „Light-metal hydrides for hydrogen storage‟ by Martin Sahlberg.

[25] M. M. Antonova, V. B. Chernogorenko, Reststance to hydrogen of Al-Sc alloys, Russian J. Appl. Chem. 74 (2001) 396-399.

[26] A.V. Kolomiets, L. Havela, V. A Yartys. A.V. Andreev, Hydrogen absorptio- desorption, crystal structure and magnetism in RENiAl intermetallic compounds and their hydrides, Journal of Alloys and Compounds, 253-254 (1997) 343-346.

[27] M. Yoshida, E. Akiba, Y. Shimojo, Y. Morii, F. Izumi, Hexagonal ZrNiAl and its hydride (deuteride) with the Fe2P-type structure, Journal of Alloys and compounds, 231 (1995) 755-759.

[28] G. Sandrock, S. Suda, L. Schlapbach, in L. Schlapbach (ed.), Hydrogen in Intermetallic Compounds II, Springer-Verlag, (1992) 197.

[29] H, Fujii, M. Saga, T. Okamoto, Magnetic, crystallographic and hydrogen absorption properties of YMn₂ and ZrMn₂ hydrides, J. less-common metals 130 (1987) 25-31.

[30] J. Schuster, J. Bauer, The ternary systems Sc-Al-N and Y-Al-N, J. of the less common metals 109 (1985) 345-350.

[31] Reckkin et al. Sov. Phys. Crystallogr. (Engl. Transl.) 9 (1964) 325.

[32] H. Peisl, G. Alefeld, J. Völkl, Hydrogen in metals I, Top. Appl. Phys. 28 (1978) 53.

[33] J. C. McGuire, C. P. Kempter, J. Chem. Phys. 33 (1960) 1584.

[34] Imperial Chemical Industries, Northwich, England, UK, Private Communication [35] I. R. Harris, P. J. McGuiness, Hydrogen: its use in processing of NdFeB-type magnets, J. Less-Common metals, 172-174 (1991) 1273-1284.

[36] A. G. Aleksanyan, S. K. Dolukhanyan, V. Sh. Shekhtman, Kh. S. Harutyunyan, K.

A. Abrahamyan, N. L. Mnatsakanyan, The thermal radiation process in the Sc-H system, J. Alloys and Compounds 356-357 (2003) 562-565.

[37] J. Jung, Nuclear Mat. 49 (1974) 299.