of 112 /112
Characterisation and Optimisation of Blowing Agent for Making Improved Metal Foams vorgelegt von Diplom-Ingenieurin Metallurgie Biljana Matijašević-Lux aus Belgrad, Serbien und Montenegro von der Fakultät III - Prozesswissenschaften der Technischen Universität Berlin zur Erlangung des akademischen Grades Doktor der Ingenieurwissenschaften - Dr. - Ing. - genehmigte Dissertation Promotionsausschuss: Vorsitzender: Prof. Dr. F. Behrendt Berichter: Prof. Dr. J. Banhart Berichter: Prof. Dr. H. Schubert Tag der wissenschaftliche Aussprache: 24. Februar 2006 Berlin 2006 D 83

Characterisation and Optimisation of Blowing Agent for ... · Characterisation and Optimisation of Blowing Agent for Making Improved Metal Foams vorgelegt von Diplom-Ingenieurin Metallurgie

Embed Size (px)

Text of Characterisation and Optimisation of Blowing Agent for ... · Characterisation and Optimisation of...

  • Characterisation and Optimisation of Blowing Agent for Making

    Improved Metal Foams

    vorgelegt von Diplom-Ingenieurin Metallurgie

    Biljana Matijaevi-Lux aus Belgrad, Serbien und Montenegro

    von der Fakultt III - Prozesswissenschaften der Technischen Universitt Berlin

    zur Erlangung des akademischen Grades

    Doktor der Ingenieurwissenschaften - Dr. - Ing. -

    genehmigte Dissertation

    Promotionsausschuss: Vorsitzender: Prof. Dr. F. Behrendt Berichter: Prof. Dr. J. Banhart Berichter: Prof. Dr. H. Schubert Tag der wissenschaftliche Aussprache: 24. Februar 2006

    Berlin 2006

    D 83

  • CONTENTS

    i

    1 Introduction......................................................................................................................................... 1

    2 Metal foams ......................................................................................................................................... 3

    2.1 Definition of solid foam................................................................................................................ 3 2.2 Aluminium foams ......................................................................................................................... 4

    2.2.1 Factors influencing aluminium foam quality ........................................................................ 6 2.3 Blowing agents (hydrides) .......................................................................................................... 10

    2.3.1 Covalent hydrides ............................................................................................................... 10 2.3.2 Saline hydrides.................................................................................................................... 11 2.3.3 Metallic hydrides................................................................................................................. 11

    2.3.3.1 Titanium-hydrogen phase system ................................................................................... 12

    3 Experimental procedures ................................................................................................................. 15

    3.1 Materials...................................................................................................................................... 15

    3.1.1 Powder specification ........................................................................................................... 15 3.1.2 Preparation of blowing agent powders................................................................................ 16 3.1.3 Preparation of precursors .................................................................................................... 17 3.1.4 Preparation of metallic foams ............................................................................................. 19

    3.2 Characterisation methods............................................................................................................ 20

    3.2.1 Powder analysis................................................................................................................... 20 3.2.1.1 Oxygen content analysis ................................................................................................. 20 3.2.1.2 Particle size analysis ....................................................................................................... 20

    3.2.2 Microscopy analysis............................................................................................................ 21 3.2.2.1 Light microscopy (LM) analysis..................................................................................... 21 3.2.2.2 Scanning electron microscopy (SEM) analysis .............................................................. 21 3.2.2.3 Transmission electron microscopy (TEM) analysis........................................................ 21

    3.2.3 Thermal analysis ................................................................................................................. 22 3.2.3.1 Thermogravimetric (TG) measurements coupled with mass spectrometry (MS) and differential thermal analysis (DTA)................................................................................ 22

    3.2.4 X-ray diffraction experiments............................................................................................. 23 3.2.5 In-situ X-ray diffraction experiments ................................................................................. 24 3.2.6 Expandometer experiments................................................................................................. 26

    4 Results: Powders ............................................................................................................................... 27

    4.1 Powder characterisation .............................................................................................................. 27 4.2 Characterisation of TiH2 powder ................................................................................................ 34

    4.2.1 Decomposition behaviour of untreated TiH2 powder ......................................................... 35 4.2.2 Oxidation behaviour of untreated TiH2 powder.................................................................. 38 4.2.3 Isothermal pre-treatment of TiH2 powder........................................................................... 39

    4.2.3.1 Thermal treatment under argon....................................................................................... 39 4.2.3.2 Thermal treatment in air.................................................................................................. 41

    4.2.4 Decomposition of pre-treated TiH2 powders ..................................................................... 42 4.2.4.1 Decomposition of pre-treated TiH2 powders in argon .................................................... 42

  • CONTENTS

    ii

    4.2.4.2 Decomposition of pre-treated TiH2 powders in air .........................................................43 4.2.5 Isothermal experiments at 600C of TiH2 powders (after constant heating at 20 K/min) ..45 4.2.6 X-ray diffraction investigations...........................................................................................47

    4.2.6.1 Ex-situ X-ray diffraction of TiH2 powder .......................................................................47 4.2.6.2 In-situ X-ray diffraction of TiH2 powder ........................................................................48

    4.2.7 TEM investigations .............................................................................................................52 4.3 Characterisation of ZrH2 and HfH2 powder (as a blowing agent)...............................................55

    4.3.1 Decomposition of pre-treated ZrH2 and HfH2 powders ......................................................55 4.3.1.1 Decomposition of pre-treated ZrH2 powders in argon ....................................................55 4.3.1.2 Decomposition of pre-treated ZrH2 powders in air .........................................................56 4.3.1.3 Decomposition behaviour of HfH2 powder.....................................................................57

    5 Results: Foams...................................................................................................................................59

    5.1 Isothermal experiments at 600C of AlSi6Cu4 precursor with TiH2 powders (after constant heating at 20 K/min)....................................................................................................................59 5.2 TEM investigations of Al precursor ............................................................................................60 5.3 Impact of pre-treatment on foaming behaviour...........................................................................60

    5.3.1 Precursors with TiH2 powders.............................................................................................61 5.3.2 Precursors with ZrH2 and HfH2 powders ............................................................................70 5.3.3 Expandometry AlSi6Cu4 precursor with TiH2 powder ......................................................72

    6 Discussion ...........................................................................................................................................79

    7 Summary ............................................................................................................................................97

    8 References ..........................................................................................................................................99

    9 Acknowledgement ...........................................................................................................................108

  • 1 Introduction

    1

    1 Introduction

    Aluminium foams are porous metallic structures which combine typical properties of cellular materials

    with those of metals [1]. The high stiffness - to-mass ratio and good crash energy dissipation ability has

    led to a variety of applications [2] especially in automotive industry [3,4].

    The properties of metal foams depend on many morphological features such as pore size distribution, cell

    wall curvature, defects etc. [5]. Although the exact interrelationship between properties and structure is

    not yet sufficiently known, one usually assumes that a uniform distribution of convex pores free of

    defects is highly desirable. Therefore, the task for the experimentalist is to produce such structures. A

    short look at existing foams shows that there is still much potential for development since these often tend

    to be very irregular [6].

    Metals can be foamed in various ways [7,8]. One very promising method is the powder compact foaming

    process: a powder mixture of an alloy and a blowing agent is compacted to a dense precursor and then

    heated to above the melting point of the alloy [9]. Above the decomposition threshold of the blowing

    agent hydrogen is released which blows bubbles in the melting alloy. Titanium hydride (TiH2) was shown

    to be a suitable blowing agent for aluminium alloys [10] although other hydrides can also be used [11].

    Hydrogen release from TiH2 starts around 400C, which is markedly below the melting point of most

    commercial aluminium alloys [9]. This difference between decomposition and melting temperature

    causes the formation of irregular, crack-like pores in early expansion stages which then can lead to

    irregularities in the final product [12]. To minimize this temperature mismatch the melting point of the

    alloy can be lowered by further alloying [13] with possible unwanted side effects or the

    decomposition threshold of the blowing agent can be raised by thermal pre-treatment [14]. If TiH2

    powder is pre-heated in air an oxide layer is formed on the surface of the particles. This layer delays gas

    release from the particles, so that ideally hydrogen is released during foaming only after the melting

    temperature of the alloy has been reached [14]. On the other hand, pre-treatments lead to hydrogen losses

    lowering foaming power. The task therefore is to find treatments which form a sufficiently thick oxide

    layer while a minimum of hydrogen is lost. Some pre-treatments have been suggested in metal foam

    literature [1421]. However, as the mechanisms leading to a retardation of gas evolution are not yet

    known, these parameters were found in a mostly empirical way.

  • 1 Introduction

    2

    It is well known that the mechanical properties of metallic foams show a large knockdown compared to

    theoretical models of uniform, regular foams, because of the presence of imperfections in the cell

    structure. The origin of the imperfections lies in the physics of the foaming process, which is poorly

    understood compared to the foaming process of polymer foams. Polymer foams have properties that

    follow the theoretical models of perfect foams [22]. Many commercially available types of foams like

    aluminium foams are manufactured by the decomposition of TiH2 in the molten (melt-route) or semi-solid

    powder metallurgical (PM) route.

    Therefore, the aims of the present thesis are:

    to investigate in detail the hydrogen release from untreated TiH2 powder;

    to investigate the effect of pre-heat treatment of TiH2 on its hydrogen release characteristics;

    to characterise gas release from TiH2 in temperature ranges relevant for Al foaming;

    to clarify the influence of oxide layers on the structure of powders and gas release;

    to clarify the difference between TiH2, ZrH2 and HfH2 powder;

    to show how powder treatments influence foaming of various alloys.

    Chapter 2 gives a general information about metal foams. An overview of the current state of the art with

    regard to the PM production route of the porous metallic materials is given in this chapter as well.

    Experimental procedures used in the present work for foaming agent preparation, foam production and

    the used experimental methods are described in Chapter 3. The experimental results are presented in

    Chapters 4 and 5. The results are discussed in Chapter 6 and summarised in Chapter 7.

  • 2 Metal foams

    3

    2 Metal foams

    2.1 Definition of solid foam

    The term "foam" can be explained and defined as shown in Fig. 2.1. Fig. 2.1 lists the designations for all

    possible dispersions of one phase in a second one and it shows that foams are uniform dispersions of a

    gaseous phase in either a liquid or a solid [23]. The single gas inclusions are separated from each other by

    portions of the liquid or solid, respectively. The term "foam" in its original sense is reserved for a

    dispersion of gas bubbles in a liquid. The morphology of such foams, however, can be preserved during

    the solidification, thus obtaining what is called "solid foam". Speaking about "metallic foams" generally

    means solid foam, which is more commonly called a "cellular solid" [9].

    Fig. 2.1. Dispersions of one phase into a second one. Each phase can be in one of the three states of

    matter [23].

    Currently there are two main methods to produce metallic foams, as shown in Table 2.1. The first one is

    the direct foaming method where the gas is injected continuously from an external source into a specially

  • 2 Metal foams

    4

    prepared molten metal containing uniformly dispersed non-metallic particles to which gas bubbles are

    added to create foams [24] or where a gas releasing chemical agent is added. The indirect foaming

    method start from a solid precursor which consists of metallic matrix containing uniformly dispersed

    blowing agent particles [9]. For aluminium based alloys the hydrides of transition elements like titanium

    or zirconium hydride powder are mostly used as blowing agent powders [11]. Above melting temperature

    this precursor expands and forms a foam. In the present work the indirect foaming method is used.

    Table 2.1. Basic foaming routes for metal foams and current manufacturers for aluminium-based foams

    and their products [25].

    direct foaming melt alloy indirect foaming prepare foamable precursor

    make alloy foamable remelt precursor

    create gas bubbles create foam

    collect foam solidify foam

    solidify foam

    manufacturers Cymat, Canada (SAF) manufacturers alm, Germany (AFS)

    (products) Foamtech, Korea (Lasom) Alulight, Austria (alulight)

    Htte Kleinreichenbach (HKB),

    Austria (Metcomb)

    Gleich-IWE, Germany

    Schunk, Germany

    Shinko-Wire, Japan (Alporas)

    (Distributor: Gleich, Germany)

    2.2 Aluminium foams

    Over the past 15 years metal foams have become one of the most challenging materials for scientific and

    industrial investigations. Many metals can be foamed but foams from aluminium proved to be a serious

    candidate for industrial applications especially in automotive industry [3,4].

    In 1943 Benjamin Sosnick [26] tried to foam aluminium with the aid of Hg in a closed chamber under

    high pressure. With the release of the pressure the vaporisation of the Hg appears at the melting

    temperature of aluminium and leads to the formation of a foam. The fundamental ideas for the indirect

    foaming process were developed in the late 1950s by Benjamin Allen [27]. In 1990, the method was

    rediscovered at the Fraunhofer Laboratory in Bremen, Germany [28] and brought to a considerable level

    of sophistication for manufacturing foams and foamed components.

  • 2 Metal foams

    5

    Using the powder metallurgical route aluminium, zinc [9] and lead [29] foams can be prepared in a cost

    effective manner. This process begins with the mixing of a metal powder (pure metal or suitable alloy)

    with a blowing agent (TiH2, ZrH2, CaCO3, etc.) [30], after which the mixed powders are compacted e.g.

    by hot pressing, extrusion or powder rolling to a dense foamable precursor [9]. In principle, the

    compaction can be done by any technique that ensures that the blowing agent is embedded into the metal

    matrix without any notable open porosity [31]. Heating of the precursor above the melting temperature of

    the metal or alloy is the so-called foaming process. During this process gas evolves from the foaming

    agent blowing bubbles in the melting alloy. The released gas forces the semi - solid compacted precursor

    material to expand. The result is a product with a highly porous structure called metal foam [1], a material

    that offers manufacturers a significant potential for lightweight structures, for energy absorption and for

    thermal management [7]. Therefore, it can be concluded that the temperature profile during foaming

    consists of a heating stage, isothermal holding stage and a cooling stage as it is shown in Fig. 2.2. The

    pores continuously grow during foaming due to the gas supplied by the blowing agent and thermal

    expansion. Expansion is controlled by a continuous decomposition of the blowing agent. Therefore,

    foaming already starts during heating of the precursor and the gas supply depends on the decomposition

    kinetics of the blowing agent. As it is mentioned in Ref. [32] complex relationships exist between the

    decomposition kinetics of the blowing agent and the melting of the alloy, which has a beneficial influence

    on metal foam quality. It the next chapter the factors influencing aluminium foam quality are explained in

    more detail.

  • 2 Metal foams

    6

    Fig. 2.2. Schematic representation of foam evolution by foaming precursors containing blowing agents

    [32].

    2.2.1 Factors influencing aluminium foam quality

    Metal foams made by first compressing mixtures of aluminium alloy powders and TiH2 and then melting

    and foaming the resulting densified compact [9,33] are now being exploited commercially under trade

    names such as "Alulight" or "Foaminal" [34]. Production volumes are still low owing to both the still

    relatively high costs and to the properties of foamed aluminium alloys which are not always sufficient for

    a given application. For this reason, many research activities are directed towards improving the

    properties of such foams and to make their manufacture more reliable, reproducible and, as a

    consequence, less expensive.

    Intuitively one tends to assume that metal foam properties are improved when all the individual cells of a

    foam have a similar size and a spherical shape. This, however, has not really been verified

    experimentally. There is no doubt that both the density of a metal foam and the properties of the matrix

    alloy influence, e.g. modulus and strength of the foam [1,35,36]. A clear influence of cell size distribution

    and morphological parameters on foam properties, however, has not yet been established. The reason for

  • 2 Metal foams

    7

    this is that it has not yet been possible to control these parameters during foaming. Therefore, studies of

    the interdependence of morphology and properties of foamed metals have in the past either been purely

    theoretical or have been concentrated on more simple systems such as open cell structures or honeycombs

    which can be manufactured in a more controlled way. Still, there are various opinions on what the ideal

    morphology of a closed - cell foam yielding the highest stiffness or strength at a given weight should be.

    Some authors find an insensitivity of strength on cell size distribution [37,38], others claim that a bimodal

    distribution of cell sizes is especially favourable [39]. Irregular foams were found to have a higher tangent

    modulus at low strains, whereas regular foams with a more unique cell size were stronger at higher strains

    [40]. Size effects were found to be an important issue because a certain number of cells across a sample is

    necessary to ensure improved mechanical properties (see e.g. Ref. [41]) which makes foams with small

    pores look more favourable. Importance has also been ascribed to the shape of cell walls. Especially

    corrugated cell walls seem to have a strong detrimental influence on foam properties [42]. Foams with

    missing cell walls or cell walls containing holes or cracks were also studied and found to be mechanically

    weaker than perfect foams (see Ref. [43] for an overview). The variability of mechanical properties in a

    group of specimens of identical nominal dimensions and densities has been analysed statistically and was

    found to be very large even for the most regular foams available at present [44]. It is likely that more

    regular foams, even if they do not show better mechanical properties, lead to a lower variability of these

    properties, which would also be a very valuable feature.

    Despite the unclear situation researchers are striving to make more uniform foams, assuming that their

    production method is the most valuable one (see e.g. Ref. [45]). As the foaming process comprises

    various steps [32] and each step can be influenced by many parameters the properties of the final foamed

    part can vary very much if production conditions are changed. Especially in the early days of metal

    foaming the search for appropriate foaming parameters was difficult because the mechanisms of metal

    foaming were unknown and the complex interdependence between parameters was not understood.

    Another problem is that there is no simple measure for the quality of a foam. Usually one assesses the

    uniformity of cells in a qualitative way by looking at sections through foamed specimens. Whenever a

    foam possesses many cells of a similar size it is often considered "good", if there are many obvious

    defects it is called "inferior". Fig. 2.3 shows four samples as taken from different batches of aluminium

    foam sandwich (AFS) panels made by the company Applied Lightweight Materials (Saarbrcken,

    Germany) which clearly show a difference in foam quality.

  • 2 Metal foams

    8

    Fig. 2.3. Sections of aluminium foam sandwich materials as obtained in industrial production [46]. From

    bottom to top: improving quality (Courtesy: H.-W. Seeliger).

    The problem with such qualitative measures is that important morphological features such as the

    asphericity of cells or crack alignments in three dimensions might not be detectable this way. Therefore,

    tomography and 3D image analysis were proposed [12] and successfully applied [47]. However, for the

    purposes of parameter optimisation this method is often too elaborate. Another problem encountered is

    the very restricted significance of individual foaming trials. Systematic variations in foam quality

    associated with the change of a processing parameter are not easy to detect and require a large number of

    experiments. Finally, there also seems to be a size effect in foaming. The smaller the samples the more

    uniform the pore structure. The size of the sample makes a comparison of different results difficult.

    The current state of knowledge suggests that the following rules have to be obeyed to obtain good

    aluminium alloy foams via the powder compact foaming technique:

    The melting behaviour of the alloy system and the decomposition characteristics of the blowing agent

    have to be coordinated. Ideally, gas evolution is suppressed below the solidus temperature of the alloy to

  • 2 Metal foams

    9

    avoid formation of cracks before melting. After partial melting the gas released by the blowing agent then

    leads to the formation of spherical pores provided that the liquid fraction is sufficiently high. As TiH2 has

    a very low decomposition temperature starting at about 400C for untreated hydride and commercial

    aluminium alloys have solidus temperatures above 525C there is an obvious gap. It is clear that untreated

    TiH2 does not fit within the melting range of any of commercial aluminium alloy. The first foaming

    experiments [48] were carried out with pure aluminium or AlCu4 and often yielded inferior results. A

    first improvement was the replacement of these wrought alloys by casting alloys such as AlSi7 and

    AlSi12. Later the alloy AlSi6Cu4 was proposed and successfully foamed [35]. Therefore, in this work the

    detail investigation was carried out on AlSi6Cu4 alloys and how these pre-treated TiH2 powders influence

    the foaming process. In the next step the variation of the copper content was applied in order to produce

    relatively high fraction of liquid which is necessary for good foaming.

    Furthermore, heating conditions during foaming are important. When foaming parts, especially large and

    flat foam panels, differences in temperature over the entire extension of the part lead to non-uniform

    foaming. This does not only make it difficult to fully expand the entire component without any collapse

    taking place in the region of the highest temperature, but also leads to an inferior pore structure due to the

    formation of cracks in the foam [49]. The heating profile during foaming is also important. In pre-heated

    furnaces the temperature usually exponentially approaches the final temperature. In regulated furnaces

    almost arbitrary heating profiles can be realised. It was found that the temperature profile around the

    melting point of the precursor significantly influences foam evolution [50]. Overheating can have a

    detrimental effect on foaming and should be avoided.

    A further factor which influences foam uniformity is the cleanliness of powder processing. The use of

    impure powders or the presence of adsorbed water, dirt or gases entrapped in the precursor during

    compaction seem to have an adverse effect on foaming, in the sense that these impurities can act as nuclei

    for big voids in early stages of gas evolution from the blowing agent. The voids are then thought to grow

    to large pores [51]. These effects are rather spurious and have not been quantified but are based on the

    experience of a number of researchers including the present author.

    Other conditions during powder processing are also important. Overheating of the powder during

    extrusion was found to be detrimental to foam expansion [52]. The compaction temperature during hot

    pressing has also to be selected carefully since even slightly elevated temperatures can lead to a reduced

    foam expansion [53, 54].

  • 2 Metal foams

    10

    In conclusion, the best recipe for a good foam, i.e. a foam with a high expansion factor developing a

    uniform cell size distribution and fairly smooth convex cells seems to be:

    use a low - melting aluminium alloy,

    use TiH2 as blowing agent which has been adapted to the alloy by pre - treatment,

    process powders under clean and reproducible conditions,

    take care to avoid overheating during powder compaction,

    carefully select the temperature profile during foaming: high heating rates without overheating

    and uniform temperature distributions are preferable.

    Tailoring the decomposition characteristics of TiH2 powder is the main topic of this work. Therefore, in

    the next section some general information about the hydrides is given.

    2.3 Blowing agents (hydrides)

    The term "hydride" can be described as a binary combination of hydrogen and a metal or metalloid [55].

    By the nature of the hydrogen bond, hydrides are classified into three principal categories [56]: covalent

    or volatile, saline or ionic and metallic.

    2.3.1 Covalent hydrides

    Covalent hydrides may be solid (usually polymeric), liquid, or gaseous. They have a very big similarity in

    their properties. The bond between hydrogen and the element is of the nonpolar electron sharing type

    where valence electrons are shared on a fairly equal basis between the elements held by bond. Molecules

    of covalent hydrides are not strongly attracted to each other, and this absence of strong intermolecular

    forces results in the high degree of volatility and low melting points of the covalent hydrides. They are

    thermally unstable with increasing atomic weight of the parent element. They burn readily in air or in

    oxygen with liberation of considerable quantities of heat. Typical covalent hydrides are: aluminium

    hydride, tin hydride, boron hydrides and germanium hydrides.

  • 2 Metal foams

    11

    2.3.2 Saline hydrides

    Saline hydrides are formed by the reaction of the strongly electro-positive alkali metals with hydrogen

    like MgH2, CaH2 and SrH2. The bond of saline hydrides results from the strong electrostatic forces

    existing between the dissimilar electric charges of the two ions. Therefore, they are highly polar. These

    hydrides are crystalline, exhibit high heats of formation and high melting points and are electrical

    conductors in the molten state.

    2.3.3 Metallic hydrides

    Metallic hydrides have metallic properties in the accepted sense: these include high thermal conductivity

    and electrical resistivity, hardness and in some instances useful mechanical properties. The difference

    between metals and hydrides is that they are usually quite brittle. There is a wide range of possible

    applications. Hydrides of transition elements were shown to be used as blowing agents for the foaming of

    metals [11]. TiH2 powders are also interesting for hydrogen storage applications or as getter material in

    vacuum systems [57-60]. Metal hydrides possess properties that make them desirable for nuclear

    application, for the preparation of powders and very pure metals, for chemical reducing agents, for

    deoxidation and desulfurization of molten ferrous alloys and for use as high energy fuels. Some of the

    hydrides can be used as portable sources of hydrogen. It is interesting to mention that zirconium hydride

    contains about twice as much hydrogen atoms per unit volume then liquid hydrogen [61]. Metal hydrides

    can also be used for preparing a coating on metals. Titanium, thorium and zirconium hydrides are used

    effectively for metal-nonmetal bonds.

  • 2 Metal foams

    12

    0

    2.3.3.1 Titanium-hydrogen phase system

    A relatively large amount of information is available about the Ti-H system. Despite such a large amount

    of information this system is still not well investigated. The phase diagram used for explanation of the Ti-

    H system is published by San-Martin and Manchester [62]. Therefore, investigations for this work mostly

    used this source for the details presented below.

    Fig. 2.4 shows an equilibrium titanium - hydrogen phase diagram. The diagram is one of the eutectoid

    types and consists of the following phases: (1) cph ; (2) bcc ; (3) , a fcc hydride; (4) fct , a

    tetragonally distorted hydride with axial ratio c / a

  • 2 Metal foams

    13

    The hydrogen atoms in the hydride at room temperature occupy the tetrahedral interstitial site CaF2 type

    structure [66] as shown in Fig. 2.5. There are different opinions within the literature about the occupation

    of hydrogen atoms within TiH2 lattice. Some of the authors assume that hydrogen lies on the tetrahedral

    interstices in an fcc metal lattice [64]. They claim that it is possible at the same time to have hydrogen

    occupancy of the octahedral interstices at room temperature [64] as well.

    Fig. 2.5. The tetrahedral lattice sites of H atoms in the TiHx lattice on the room temperature. Ti-red, H-

    dark blue.

    Some authors preformed several x-ray diffraction measurements [65-67] in order to determine the system

    phase boundaries at lower temperatures. There are still some discrepancies on this in the literature, arising

    mainly because of unreliable characterisation of the Ti-H system namely in the concentration and the

    influence of the impurities.

    It is an interesting observation that at the concentration of 66% of hydrogen at a certain temperature the

    fcc hydride starts to be transformed to fct hydride. When the fcc hydride starts to be transformed to

    fct hydride the temperature was observed around 25C [68].

  • 2 Metal foams

    14

    Data which illustrates this effect is shown in Fig. 2.6. The transformation is manifested by splitting and

    shifting of the diffraction peaks of the original cubic phase as the temperature is lowered.

    Fig. 2.6. Cell dimensions of titanium hydride as a function of temperature [68].

    This transformation is quite rapid and reversible. By cycling in the temperature range from 20 to 30C

    the repeated transformation from cubic to tetragonal structure and vice versa was observed by Crane [68].

    Yakel [69] and Takasaki [70] reported a different critical transformation temperature namely at 37C.

    Important facts like purity and the hydride preparation methods as reported from [65] and as pointed out

    by San-Martin and Manchester [62], that a critical concentration specification is necessary miss a more

    specific temperature reference than "room temperature".

  • 3 Experimental procedures

    15

    3 Experimental procedures

    3.1 Materials

    3.1.1 Powder specification

    TiH2 powder supplied by three different manufacturers was used in this study. As it is shown in Table 3.1

    there is a slight variance in the purity of these powders and particle sizes as reported from the

    manufacturer. Detail investigations were carried out on TiH2 powder supplied from the company

    Chemetall GmbH [71]. ZrH2 and HfH2 powders supplied by Chempur GmbH were used in this study as

    well. The powder specifications given by the manufacturer are shown in the Table 3.1. Metallic powders

    used for making aluminium alloys are specified in Table 3.2.

    Table 3.1. Specification of the used metal hydride powders (specifications by manufacturer).

    Powder Producer Purity

    [%]

    Density

    [g cm-3]

    Particle size

    [m]

    Impurities [wt.%] H [wt.%]

    TiH2 Chemetall GmbH, Frankfurt 98.8 3.76 < 63 Si

  • 3 Experimental procedures

    16

    3.1.2 Preparation of blowing agent powders

    The powders were characterised in the "untreated" state and after several pre-treatments. TiH2 known as

    the hydride with the best foaming potential, was investigated in more detail. ZrH2 powder and HfH2

    powder were shown for foaming experiments.

    Pre-treatments of the powders were carried out isothermally at various temperatures and times both in an

    argon atmosphere (purified by a zirconium getter to 10-12 ppm) inside a tube furnace and under air in a

    chamber furnace. In the former case the powder was filled into flat ceramic crucibles (powder height 2

    mm) which were then placed into a horizontal quartz tube. After repeated evacuation and filling with

    argon a furnace was slid over the tube for the time required. In the latter case the ceramic crucibles were

    simply placed into a large volume chamber furnace and were left there for the time specified. After heat

    treatment the powders were gently homogenised by tumbling in a container. Two further TiH2 samples

    were obtained by using initially untreated samples which had then been subjected to an isothermal heat

    treatment for 180 minutes at 520C inside the mass spectrometer both under argon (Fig. 4.11) flow and

    air flow (Fig. 4.12) to study the influence of continuous hydrogen removal during treatment. Pre-

    treatment parameters for each powder were chosen as shown in Table 3.3. The powders could be treated

    without individual particles sticking together (this became a problem at T 540C). It is worth

    mentioning that the pre-treatment parameters are applicable in the serial production of the metal foam

    precursors.

  • 3 Experimental procedures

    17

    Table 3.3. Parameters of the pre-treatment of the blowing agent powders (furnace).

    Air Argon Powder

    T [C] t [min] T [C] t [min]

    440 180 440 180

    480 180 460 180

    500 180 480 180

    520 90 520 90

    520 180 520 180

    TiH2, Chemetall GmbH

    520 360 - -

    TiH2, SeJong Materials GmbH 480 180 - -

    TiH2, GFE GmbH 480 180 - -

    400 180 400 180

    480 180 480 180

    520 90 520 90

    ZrH2, Chempur GmbH

    520 180 520 180

    HfH2, Chempur GmbH 480 180 - -

    To obtain bulk samples suitable for TEM investigations the TiH2 powders were cold pressed to small

    pellets of 6 mm diameter and 3 mm height prior to pre-treatment in order to avoid deformation after heat

    treatments which are known to modify the decomposition characteristics [11]. For the in-situ diffraction

    experiments the samples were prepared the same way.

    3.1.3 Preparation of precursors

    Foamable precursors were prepared by mixing aluminium powder, silicon powder and copper powder in

    fractions that led to different commercial aluminium alloys. First pure Al powder was used, then AlSi6,

    AlSi6Cu4 and AlSi6Cu10. In these alloys 0.5 wt.% TiH2 was added either in the untreated state or after

    one of the various pre-treatments under air. 0.5 wt.% ZrH2 and 0.5 wt.% HfH2 powder were added in

    AlSi6Cu4 alloy as shown in Fig. 3.1. The current PM route used for metallic foam production in this

    work is shown in Fig. 3.2.

  • 3 Experimental procedures

    18

    Fig. 3.1. Overview of precursor preparation.

    blowing agent

    mixing

    metal powder(s)

    uni-axial pressing

    foamableprecursor

    foaming

    blowing agentblowing agent

    mixingmixing

    metal powder(s)metal powder(s)

    uni-axial pressing

    foamableprecursor

    foaming

    Fig. 3.2. Production steps of PM foams [72].

    Homogenization of powder mixture is done for a certain period of time in the appropriate powder

    blenders. Depending on the application there is a different kind of blenders for dry mixing of powders, as

    shown by Hausner [73]. The powders for metal foam production were mixed for 90 minutes in a tumbling

    mixer to produce a homogeneous distribution of all components. The powder mixture prepared for uni-

    axial pressing is shaped in special tools on a hydraulic press.

  • 3 Experimental procedures

    19

    The requirements put to these pressing tools are very important as the pressures applied in pressing are

    around 200 MPa and the tool is simultaneously heated to the desired temperature. Pressing die and the

    punches must be of hardened steel or hard metal. The pressing tool was made of steel 1.2344. The powder

    mixtures were first pre-pressed at room temperature in a cylindrical die of 36 mm diameter, after which

    the die containing the powder mixtures was heated up to 450C and held there for temperature

    equalisation. Subsequent hot pressing at about 200 MPa for 30 min yielded tablets with more than 99%

    density. The density achieved during pressing depends on the compressibility of the powders and the

    pressing force. To prevent the sticking of the compressed sample, the pressing tool was spray coated

    before each pressing with a molykote spray, which is a molybdenum disulfide aerosol.

    Hot pressing is a kind of heat treatment known as sintering in powder metallurgy. It is very important that

    during simultaneous heating and pressing the particles are brought close to one another, as the atom

    mobility is increased forming a new atomic network. This occurs on the contact particle spots and enables

    forming of contact spots on which the powder particles are firmly interconnected. With this the sintering

    is started. It continues with further atom exchanges in the interior of a particle. The pressing procedure is

    explained in more detail in Refs. [74, 75].

    As AlSi6Cu4 with TiH2 powder as a blowing agent is one of the common aluminium alloys for metal

    foams production it was chosen for further investigations on the hydrogen release from the precursor. To

    monitor hydrogen release from the precursor during foaming small cubes (l 3 mm) of precursor material

    were formed. For the foaming experiments thin slices of 36 diameters and 4 mm thickness were cut off

    the pressed tablets by electric discharge machining (EDM).

    For the TEM observations of TiH2 particles embedded in an aluminium matrix the procedure described

    above was applied with the only difference that no Si and Cu was added.

    3.1.4 Preparation of metallic foams

    The foamable precursor was converted into foam by heating it above its melting temperature. The

    different types of precursors used are listed in Fig. 3.1. For foaming a sample was placed on a steel block

    which was pre-heated to the foaming temperature inside a furnace and held at the same temperature. A

    thermocouple was in contact with the samples at all times so that it was possible to monitor the

    temperature inside the foam during the entire process. Each foaming experiment was stopped as soon as a

    sample reached a given porosity, and the foaming time tf and final foaming temperature Tf was recorded.

  • 3 Experimental procedures

    20

    All samples were tried to be foamed to the same final volume corresponding to about 60% porosity. As it

    was not possible to have visual control of the foaming experiment several samples of each composition

    with slightly varying foaming times were prepared and selected. The ones which had porosity in the

    desired range were chosen for further examination. To foam Al precursors the foaming temperature inside

    a furnace was 750C. All alloys were foamed at 650C. The preparation technique is explained in more

    detail in Refs. [19, 74, 76]. Samples were retrieved after cooling, sectioned (using a microtome or EDM)

    and photographed with a Digicam.

    3.2 Characterisation methods

    3.2.1 Powder analysis

    3.2.1.1 Oxygen content analysis

    The oxygen content was determined by carrier gas hot extraction (CGHE) [77] using the nitrogen/oxygen

    analyser Horiba EMGA 620 WC. The procedure of this method comprises the quantitative reduction of

    oxygen in the sample, the release of gaseous reaction products and their selective and quantitative

    detection. About 5 mg of powder were first filled into a tin, then into a nickel capsule. This package was

    dropped into a graphite crucible which then was electrically heated up to about 2500C for 40 s. The

    capsules act as metal additives to obtain a homogenous reduction-assisting melt. At these high

    temperatures the oxygen in the sample reacts with carbon forming CO which was detected by infrared

    absorption analysis.

    3.2.1.2 Particle size analysis

    Particle size distributions (PSD) of all powders were measured using a Sympatec Helos Vectra particle

    analyser. Powders were dispersed in water assisted by ultrasonic treatment to destroy agglomerates. In the

    applied laser diffraction spectrometry technique (LDS), the angular distribution of light is measured after

    passing through an optically dilute dispersion of the suspended particles. The Frauenhofer approximation

    was the optical model for the analysis of diffraction spectra. So the particle size distribution was then

    derived from the patterns of a diffracted laser beam.

  • 3 Experimental procedures

    21

    3.2.2 Microscopy analysis

    3.2.2.1 Light microscopy (LM) analysis

    Samples for LM were prepared by using standard metallographic preparation techniques: embedding the

    cut samples into embedding resin (Kulzer 4071 or 5071), grinding with a silicon carbide paper 1200-

    4000, rough polishing with the 3m diamond grinding medium at MD-Dac (Struers). Final polishing was

    done with 1m diamond at MD-Dac (Struers). End polishing was done using OPU-Enpolitur (SiO2) at

    MD-Chem (Struers). Microstructure was observed on light microscope Zeiss Axioplan 2, supported with

    a software Zeiss Axiovison 4.

    3.2.2.2 Scanning electron microscopy (SEM) analysis

    The morphology of all powders was characterised by SEM using a Philips XL 20 operated at 10 to 20 kV.

    Structure analysis of the powder particle is very important to evaluate and optimise the process of the

    powder production which affects the properties of the final product. This technique can visualize the

    morphology of the surface, the presence of satellite, impurities and agglomerates as well as the

    deformation or fracture of powder particles [78,79]. EDX analysis was used in order to get detail

    information about the content of oxygen around individual TiH2 powder particles after the pre-treatment

    in air at 480 and 520C for 180 min.

    3.2.2.3 Transmission electron microscopy (TEM) analysis

    TiH2 powders were characterized by TEM employing a Philips CM30 operated at 300 kV. Cold pressed

    pellets were cut into sizes suitable for TEM preparation and were then mechanically polished and

    subsequently ion-thinned in an ion beam milling machine. Cold pressing does not lead to a merger of

    individual particles so that these could still be separated in the TEM images. For studying TiH2 particles

    embedded in an aluminium matrix after compaction, i.e. in the foamable precursor, focused ion beam

    (FIB) milling was used to cut a thin foil out of a bulk sample including an interface between the Al matrix

    and a TiH2 particle. After this a micromanipulator was used to transfer the foil from the original sample to

    a TEM grid coated with a porous carbon support film.

  • 3 Experimental procedures

    22

    3.2.3 Thermal analysis

    3.2.3.1 Thermogravimetric (TG) measurements coupled with mass spectrometry (MS) and differential thermal analysis (DTA)

    To study the decomposition of blowing agent powders as a function of temperature or time a Netzsch

    simultaneous thermal analyser STA 409C was used which allowed simultaneous differential thermal

    analysis (DTA), thermo-gravimetric analysis (TGA) and mass spectrometry (MS) of both hydrogen and

    water. The analyser was equipped with a silicon carbide furnace connected to a quadrupole mass

    spectrometer via a skimmer coupling system. The distance between skimmer and sample holder

    amounted to 12 mm. In all the measurements a constant amount of 150 mg of either the untreated or pre-

    treated blowing agent powders or the foamable precursor were filled into the alumina crucible which was

    placed on top of a ceramic sample carrier, as shown in Fig. 3.3. The first pressure reduction step located

    in front of the skimmer and between the sample holder, realized by a ceramic shield with an orifice,

    reduces the pressure to 10-1

    mbar. The skimmer orifice further reduces the pressure down to 10-5 mbar

    where the gases arrive at the analyzer ion source at the head of the quadrupole mass spectrometer,

    positioned in cross beam geometry. A vertical gas flow of an inert gas (argon) through the furnace core

    and the short distance between sample and skimmer guarantee the transport of even heavy molecules to

    the analyzer ion source. To obtain an optimum sensitivity the mass spectrometer was operated with a

    secondary electron multiplier.

    After the positioning of the crucible the thermo-balance was evacuated three times and back-filled with

    argon prior to heating the sample under a continuous flow of argon (flow rate 150 cm3 min-1).

  • 3 Experimental procedures

    23

    Fig. 3.3. Simultaneous thermal analyser - Mass spectrometry with skimmer coupling system (Courtesy:

    Netzsch Company).

    3.2.4 X-ray diffraction experiments

    A Bruker - AXS Advance diffractometer model D8 equipped with a graphite monochromator and a

    scintillation counter was used to determine the phases present in the TiH2 powders before and after pre-

    treatments in air at 520C. The diffraction patterns were measured for 2 ranging from 15 to 90 at steps

    of 0.02. An acquisition time of 5 s per step using CuK radiation was sufficient for a successful phase

    analysis. The XRD patterns were evaluated using evaluation program (EVA) from Bruker - AXS and

    compared to a JCPC standard diffraction database.

  • 3 Experimental procedures

    24

    3.2.5 In-situ X-ray diffraction experiments

    The in-situ diffraction experiment was performed at the KMC2 beamline at the BESSY synchrotron light

    source using a photon energy of 8 keV. The use of an area sensitive detector (HiStar, Bruker AXS) to

    simultaneously acquire the photons scattered in 2 angle between 30 and 60 degrees combined with the

    intense synchrotron beam allowed to acquire a complete scattering pattern every 10 seconds.

    (a)

    (b)

  • 3 Experimental procedures

    25

    (c)

    Fig. 3.4. Experimental set up for the in-situ diffraction measurements; a) sample holder, b) high

    temperature chamber, c) diffraction geometry.

    The sample was mounted in a high temperature chamber on a Huber goniometer as shown in Fig. 3.4. The

    chamber was equipped with a beryllium dome, which is transparent for the photon energy used. This

    allows heating the sample up to 1000C in vacuum or in controlled atmosphere and simultaneously to

    perform scattering experiments in a wide range of incident and diffracted beam directions. For current

    experiment the chamber was flushed with He to accelerate the H2 emission from the sample.

    Since the area detector can span only 30 degrees in 2 (Fig.3.4c), the range between 30 and 60 degrees,

    was chosen. In this range all major peaks of all expected phases appear. Samples were heated from room

    temperature up to 780C with a heating rate of 10 K/min. Starting from 200C an XRD pattern was

    recorded every 10 sec. Temperature was controlled using a Eurotherm process controller and a

    thermocouple was placed in the vicinity of the sample. It was not possible to mount the thermocouple

    directly on the sample because of a possible interference with the scattering experiment.

  • 3 Experimental procedures

    26

    3.2.6 Expandometer experiments

    The precursor tablets of AlSi6Cu4 prepared with untreated and differently pre-treated TiH2 powders were

    foamed in a so-called "expandometer". This dilatometer was specially constructed for metal foam

    characterisation and allows to heat up a piece of precursor in a quartz glass tube using lamps with a given

    heating profile while recording sample temperature and vertical expansion. Expansion can be interrupted

    by turning off the lamps and starting to cool with pressurised air [80]. The used experimental procedure is

    explained in more detail in Refs. [54, 81]. The cooling rates achieved are about 2 K s-1.

    Two types of experiments were carried out:

    1. Experiments in which the precursor sample was foamed without interruption under the influence

    of a given temperature profile, and

    2. Experiments in which the foaming process was interrupted to reach a pre-selected expansion.

    In the former case the temperature profile chosen included a constant heating rate of 1 K s-1 up to 650C

    and a constant temperature with a given dwell time. In the latter case the final temperature was kept

    slightly lower at 630C to slow down the foaming process. Samples were retrieved after cooling,

    sectioned (using a microtome or EDM), photographed with a Digicam and further investigated by SEM.

    In order to check the reproducibility of the expandometer measurements, about five experiments were

    carried out for each set of parameters. As the resulting curves showed some variation a measurement was

    accepted whenever three curves were identical within an error limit given by the precision required for

    the analysis.

  • 4 Results: Powders

    27

    4 Results: Powders

    4.1 Powder characterisation

    In Fig. 4.1 the morphology of both untreated TiH2 powder and TiH2 pre-treated under air is made visible

    by SEM. The powders are angular shaped, the particle size is variable and the powder surface is irregular.

    Obviously, pre-treatment in air does not significantly influence the morphology of the powders. Fig. 4.2

    shows the EDX analysis of pre-treated TiH2 in air at 480 and 520C for 180 min obtained with an energy

    of 10 kV. As oxygen is a light element it is difficult to measure quantitatively. Qualitatively, EDX

    analysis shows that the pre-treatment in air at 520C for 180 min produces a higher oxygen content on the

    surface of the powder particle.

    (a)

  • 4 Results: Powders

    28

    (b)

    Fig. 4.1. SEM images of TiH2 powder particles; a) untreated, b) pre-treated in air at 480C for 180 min.

    (a)

  • 4 Results: Powders

    29

    (b)

    Fig. 4.2. EDX analysis of TiH2 powder particles pre-treated in air; a) 480C for 180 min, b) 520C for

    180 min.

    The oxygen content of pre-treated TiH2 powders under conditions given in Table 3.3 was determined by

    hot gas extraction as shown in Fig. 4.3. Untreated TiH2 powder contains about 1 wt.% oxygen. Treatment

    in air leads to a continuous increase up to 7 wt.% for a treatment at 520C for 360 minutes. Treatment

    under argon leads to a slight decrease of the oxygen content down to 0.65 wt.%. Changes in the colour of

    the TiH2 powder were observed during heat treatment in air. The initially dark grey powder turned blue at

    480C, green at 500C and light brownish to dark brown at 520C, depending on annealing time, as

    shown in Fig. 4.4. Apparently the heat treatment in air does not change the morphology of the powder but

    due to the light interference the optical properties changed as shown in Fig. 4.4.

  • 4 Results: Powders

    30

    0 100 200 300 400 500

    0

    1

    2

    3

    4

    5

    6

    7

    8ox

    ygen

    con

    tent

    [wt.%

    ]

    annealing temperature [C]

    Ar 180 min 90 min 360 min

    450 475 500 5250,6

    0,7

    0,8

    0,9

    1,0Air

    180 min 90 min 360 min

    Fig. 4.3. Oxygen content obtained by CGHE of untreated TiH2 powder and pre-treated at various

    temperatures and times under air and argon.

    a) b) c)

  • 4 Results: Powders

    31

    d) e) f)

    Fig. 4.4. TiH2 oxidised at different conditions. The colour of the powder depends on the thickness of the

    oxygen layer a) unterated b) 480C for 180 min c) 500C for 180 min d) 520C for 90 min e) 520C for

    180 min f) 520C for 360 min.

    Fig. 4.5 shows the morphologies of untreated ZrH2 and HfH2 powder obtained by SEM. Evidently the

    powder morphologies are comparable with TiH2 powder, as well. They are angular shaped and the

    particle size is variable. The particle size distribution and oxygen content analysis of used metallic

    hydrides are given in Table 4.1. Analysis of the various metal powders used for metal foam precursor

    preparation yielded to the values as shown in Table 4.1. The morphologies of the used powders as

    obtained by SEM are shown in Fig. 4.6.

    (a)

  • 4 Results: Powders

    32

    (b) Fig. 4.5. SEM images of untreated powder particles; a) ZrH2, b) HfH2.

    Table 4.1. Physical and chemical properties of the ZrH2, HfH2, TiH2, Al, Cu and Si powders used.

    Particle size distribution [m] Material

    Oxygen analysis [wt.%]

    D(v.90) D(v.50) D(v.10)

    ZrH2 Chempur GmbH 0.58 21.14 8.66 2.62

    HfH2 Chempur GmbH 0.19 70.91 21.42 2.39

    TiH2 Chemetall GmbH 1.00 30.79 14.14 3.18

    TiH2 SeJong Material 0.51 25.32 13.46 2.82

    Blowing

    agents

    TiH2 GFE GmbH 0.88 23.90 9.01 1.92

    Aluminium Eckart 0.57 150.00 66.50 19.70

    Copper Chempur GmbH 0.46 121.00 84.00 51.00

    Metal

    powders Silicon lschlger 0.53 76.10 27.80 4.70

  • 4 Results: Powders

    33

    (a)

    (b)

    (c)

    Fig. 4.6. SEM images of metal powders with obtained particle shapes a) aluminium, irregular, b) copper,

    spherical c) silicon, angular.

  • 4 Results: Powders

    34

    4.2 Characterisation of TiH2 powder

    The onset and peak temperatures of mass spectrometric (MS) or thermoanalytic data are specified in this

    and in the next chapter. The peak is defined as the location of the maximum value of the (MS) curve

    under consideration, while the onset is defined as the point where the MS curve has risen from the

    baseline to 1% of the peak value. Only in Fig. 4.16 this criterion had to be changed to 2.5%, while the

    sharp edge in Fig. 5.1 allowed to use this point to determine the onset point. These definitions are adapted

    to the quality of the data available. The onset temperatures obtained this way are a bit arbitrary and do not

    reflect any physical transition in the material but they allow to accurately identify changes due to pre-

    treatments.

    Most measurements were repeated two or three times to be able to assess statistical errors. An excellent

    reproducibility was found and therefore only one curve is shown with very small error bars (look Fig.

    6.7). In Fig. 6.7 most of the results are summarized and error bars are given to demonstrate accuracy.

    As TiH2 leads to the best foaming characteristics this powder was investigated in more detail. Therefore,

    an extensive thermoanalytical programme is carried out. The experiments are organised as outlined in

    Fig. 4.7. In Chapter 4.2.1 and 4.2.2 untreated TiH2 powder is characterised by thermal analysis in air and

    in an argon atmosphere to obtain a first overview of the concurring processes "hydrogen release" and

    "oxidation". Isothermal experiments at 520C which is a typical pre-treatment temperature are carried

    out in Chapter 4.2.3. In a second step in Chapter 4.2.4 characterisation of TiH2 powders pre-treated in air

    or argon prior to thermal analysis are shown. Again in Chapter 4.2.5 isothermal experiments of untreated

    and pre-treated TiH2 powders in air are performed this time at 600C which is a typical foaming

    temperature for aluminium alloys. Finally, in Chapter 5.1 hydrogen release from the foamable aluminium

    precursor is characterised under the same isothermal conditions as shown in Chapter 4.2.5.

    The phases present in TiH2 before and after pre-treatment in air are determined by XRD. The oxide layers

    around individual TiH2 particles are investigated with TEM. Finally, in Chapter 5.3 it is shown that pre-

    treatment actually improves the structure of aluminium foams.

  • 4 Results: Powders

    35

    Fig. 4.7. Overview of all the thermo-analytical experiments performed.

    4.2.1 Decomposition behaviour of untreated TiH2 powder

    Fig. 4.8a shows the release of hydrogen and the decomposition behaviour of untreated TiH2 powder as a

    function of time and temperature. An argon gas flow and a heating rate of 10 K/min were applied. The

    temperature profile was linear throughout the entire range so that T(t) is not shown explicitly but can be

    read from the upper axis. As no oxygen is present the decrease in mass has to be ascribed to the release of

    hydrogen. The mass curve (TG) shows distinct changes in slope which become visible when showing the

    derivative (DTG). Two peaks and one shoulder can be observed in the DTG curve as well as in the mass

    spectroscopic (MS) curve. The first step of hydrogen release as identified by mass spectrometry applying

    the 1% criterion starts at 415C (marked as "1"), the peak position is located at 524C. A second

    dehydrogenation step starts at about 540C after deconvolution (Fig.4.8b) and reaches the maximum at

    626C. The shoulder around 800C indicated a third step which, however, is not further investigated since

    such high temperatures are never reached during foaming of aluminium. Mass spectrometry of hydrogen

    (MS) and the derivative of the mass (-DTG) agree well even though they are slightly shifted against each

    other. The DSC tends to a very similar result. Because of the higher sensitivity of the DSC measurement

  • 4 Results: Powders

    36

    the starting point of decomposition becomes even visible at a lower temperature of around 390C ("2").

    The negative maximum of the DSC curve indicates an endothermal reaction. Mass reduction continues up

    to temperatures of around 920C ("3"). The total mass loss up to this temperature is 3.78%, which is close

    to the calculated value of 4% (TiH2Ti corresponds to 5048 = -4%). The areas of the first and the

    other peaks are consistently related as 1:3.21, 1:3.22 and 1:3.01 for the MS, DTG and DSC curve,

    respectively. Hence, about 25% of the hydrogen gas is released in the first reaction. Some water is

    detected during heating of the powder over a wide temperature range, the water ion signal is about 30

    times smaller than that of hydrogen. The MS curve of hydrogen in Fig. 4.8b is shown after deconvolution

    and shows three Gaussian fits of hydrogen release.

    0 20 40 60 80 100 1200.0

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2

    97

    98

    99

    100

    101200 400 600 800 1000 1200

    0.00

    0.05

    0.10

    0.15

    0.20

    0.25

    13

    TG

    MS hydrogen

    - DTG

    temperature [C]

    MS

    ion

    curr

    ent [

    10-9A]

    time [min]

    MS (H2)

    TG

    DTG

    [%/m

    in]

    TG [%

    ]

    200 400 600 800-2,0

    -1,5

    -1,0

    -0,5

    0,0

    2DS

    C [

    W/m

    g]

    DSC

    temperature [C]

    DTG

    (a)

  • 4 Results: Powders

    37

    200 400 600 800 1000 12000.0

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2 MS (H2) Gauss fit MS (H2) Gauss fit peak 1 Gauss fit peak 2 Gauss fit peak 3

    MS

    (H2)

    ion

    curr

    ent [

    10-9A

    ]

    temperature [C]

    (b)

    Fig. 4.8. Analysis of untreated TiH2 powder heated from 30C to 1200C at 10 K/min under argon. (a)

    Calorimetric (DSC), mass (TG), mass change (-DTG=-d/dt(TG)) and mass spectrometric (MS) signals for

    H2 are shown. The numbers in circles mark specific temperatures discussed in the text. At (b) MS (H2)

    curve after deconvolution and respective Gaussian fits are shown.

    Fig. 4.9 shows the decomposition behaviour of untreated TiH2 powders produced from various

    manufacturers. Powder specifications from the manufacturer, pre-treatment parameters and powder

    characterisation are given in Tables 3.1, 3.3 and 4.1, respectively. Applying the 1% criterion, it can be

    seen that hydrogen first starts to be released from GFE powder at 360C. Powder from SeJong Materials

    releases hydrogen at 365C applying the same criterion. 415C is the onset temperature of gas release

    from TiH2 powder supplied by Chemetall. Therefore, the variance in the onset temperatures is 55 K. Gas

    release ends at 900C for GFE powder and 950C for Sejong Materials as well as Chemetall TiH2

    powder. The peak maximum shifts to +18 K for SeJong Material powder and to -17 K for GFE powder

    compared to TiH2 powder from Chemetall.

  • 4 Results: Powders

    38

    400 600 800 10000.0

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2peak positions: 609

    644

    626

    3

    2

    1

    Chemetall SeJong

    MS

    ion

    curr

    ent [

    10-9A

    ]

    temperature [C]

    GFE

    ig. 4.9. Mass spectrometric analysis of untreated TiH2 powders supplied from different manufacturer

    4.2.2 Oxidation behaviour of untreated TiH2 powder

    ig. 4.10 shows the behaviour of untreated TiH2 powder heated at a rate of 10 K/min in air. Oxidation

    F

    during heating from 30C to 1200C at 10 K/min in an argon atmosphere (low temperature range

    omitted). Curve for untreated powder from Chemetall is already given in Fig. 4.8 and is shown for

    comparison. Peak positions are specified in the diagram (in C). The numbers in the circles mark specific

    temperatures discussed in the text.

    F

    starts at 250C ("1") and leads to a continuous mass increase. Due to strong exothermal effects the

    temperature curve departs from the linearity at 600C and leads to a peak at 680C ("2") associated with a

    sudden increase of the slope of the TG curve. Both, hydrogen and water evolution was observed.

    Hydrogen is released above 380C ("3"), while water evolution sets in at 510C ("4"). The total mass gain

    from room temperature to T=1200C is 59.7%, which fits very well to the calculated increase of 60%

    (TiH2TiO2 corresponds to 5080 = +60%).

  • 4 Results: Powders

    39

    0 20 40 60 80 100

    200

    400

    600

    800

    1000

    1200

    4

    3

    2

    1MS hydrogen

    MS water

    tempe

    rature

    MS

    ion

    curr

    ent [

    10-9A

    ]

    TG [%

    ]

    time [min]

    temperature

    tem

    pera

    ture

    [C

    ]

    time [min]

    0 20 40 60 80 100

    100

    110

    120

    130

    140

    150

    160

    TG

    0.0

    0.5

    1.0

    1.5

    2.0

    2.5

    MS (H2) MS (H2O)

    ig. 4.10. Analysis of untreated TiH2 powder heated from 30C to 1200C at 10 K/min in air. Mass (TG)

    4.2.3 Isothermal pre-treatment of TiH2 powder

    4.2.3.1 Thermal treatment under argon

    ig. 4.11 shows experiments on untreated TiH2 performed under argon applying a heating rate of 10

    F

    and mass spectrometric (MS) signals for H2 and H2O are shown. The numbers in circles mark specific

    temperatures discussed in the text.

    F

    K/min from room temperature up to 520C followed by an isothermal treatment at this temperature for

    180 min. Different regimes of reaction can clearly be recognized in the MS, TG, -DTG and DTA curves.

    The first regime ("A") is the temperature range from 400 to 520C (40 to 60 min) and is associated with a

    loss of weight and strong evolution of hydrogen. The maximum of gas release is observed after 50

  • 4 Results: Powders

    40

    minutes at 500C shortly before isothermal conditions have been reached. The negative maximum of the

    DTA curve indicates an endothermal reaction. The second regime ("B") of hydrogen release becomes

    dominant under isothermal conditions in the time range from 60-100 min and constant temperature. The

    MS, TG, DTG and DTA curves notably change their slopes and develop a shoulder. In the third regime

    ("C") the release of hydrogen levels off quickly and tends to zero. Obviously, the TG, MS and DTA

    curves are coupled. The transition from the second to the third regime can best be detected from the

    changing slope of the MS-curve. The shoulder of the MS curve matches with the shoulder of the DTA

    curve (see inset in Fig. 4.11).

    50 100 150 200

    97.5

    98.0

    98.5

    99.0

    99.5

    100.0

    100.550 100 150 200

    100

    200

    300

    400

    500

    600

    -0.20

    -0.15

    -0.10

    -0.05

    0.00

    0.05

    0.00

    0.05

    0.10

    0.15

    0.20

    0.25

    0.30

    0.35A B C

    MS

    ion

    curr

    ent [

    10-9

    A]

    tem

    pera

    ture

    [C

    ]

    DTG

    [%/m

    in]

    TG [%

    ]

    time [min]

    time [min]

    10 K

    /min

    MS hydrogen TG

    - DTGtemperature

    50 100 150 200

    -0.4

    -0.3

    -0.2

    -0.1

    0.0

    t

    TDTA

    CBA

    DTAT

    100200300400500600

    ig. 4.11. Analysis of untreated TiH2 powder heated up to 520C and held there under Ar flow. F

    Calorimetric (DTA), mass (TG), mass change (DTG) and mass spectrometric signals (MS) for hydrogen

    are given. Letters (numbers) in circles denote reaction ranges (temperatures) as discussed in the text.

  • 4 Results: Powders

    41

    4.2.3.2 Thermal treatment in air

    Fig. 4.12 shows the results of an analogous isothermal experiment under air. The increasing TG curve

    indicates significant oxidation above 250C (TG curve) ("1") after a small initial mass change and reveals

    that the highest oxidation rate is found in the range between 460 and 520C. Hydrogen release as given

    by the MS curve gets notable at approximately 350C ("2") and reaches a maximum at 507C after 49

    min. This is slightly earlier than in Fig. 4.10 reflecting inaccuracies in the starting conditions of each

    experiment. After some time at 520C hydrogen evolution rapidly drops and reaches a low level after

    about 100 min ("3"). Associated with H2 evolution, the formation of water was observed from the

    corresponding mass spectrometric signal. If the isothermal holding temperature is 480C instead of 520C

    oxidation proceeds much slower which can be seen from the corresponding TG curve included in Fig.

    4.12.

    0 50 100 150 200100

    101

    102

    103

    104

    105

    106

    time [min]

    32

    1

    TG (480C)

    calc. TG

    TG

    MS water

    MS hydrogen

    10 K

    /min

    temperature

    MS

    ion

    curr

    ent [

    10-9A

    ]

    tem

    pera

    ture

    [C

    ]

    TG [%

    ]

    time [min]

    0 50 100 150 200

    100

    200

    300

    400

    500

    600

    0.00

    0.05

    0.10

    0.15

    0.20

    Fig. 4.12. Analysis of untreated TiH2 powder heated up to 520C and held there under air. Mass (TG) and

    mass spectrometric signals (MS) for hydrogen and water are given. Letters (numbers) in circles denote

    reaction ranges (temperatures) as discussed in the text. The dash-dotted line is a TG curve obtained at

    480C, dotted lines are fit curves calculated using Eqs. 2 and 3 (p. 85).

  • 4 Results: Powders

    42

    4.2.4 Decomposition of pre-treated TiH2 powders

    The following experiments are based on powders pre-treated in air or argon prior to characterisation. At

    this stage no further oxidation occurs since thermal analysis is carried out under argon. TiH2 powder

    solely releases hydrogen which is also the case when metals are foamed.

    4.2.4.1 Decomposition of pre-treated TiH2 powders in argon

    Fig. 4.13 shows the decomposition behaviour of TiH2 powders previously pre-treated under argon.

    Starting from the curve for untreated powder which was already described in Fig. 4.8, it can be seen that

    heat treatment at 440C for 180 min eliminates the first decomposition stage and leaves only a single

    stage. The single peak structure remains the same even for treatments at higher temperatures or for longer

    durations. The position of the remaining peak is slightly shifted to lower temperatures compared to that of

    the second peak of untreated TiH2. The maximum peak shift is -12 K. Another observation is that the

    onset of gas evolution at 415C is not changed significantly by pre-treatments under argon, with all onset

    temperatures being between 413C and 415C. Gas release levels off for temperatures beyond the peak

    and ends between 950C for untreated powders or powders pre-treated at 440 and 460C, and 1100C

    for powders pre-treated at 480 and 520C.

    The decomposition curve for the powder pre-treated under the conditions described by Fig. 4.11 (curve

    marked with an asterisk in Fig. 4.13) looks very different to the corresponding curve of the powder pre-

    treated in a closed argon-filled container for almost the same time and temperature. While onset and peak

    position are found at the same temperature, much less hydrogen is detected.

  • 4 Results: Powders

    43

    400 600 800 10000.0

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2

    614*

    untreated

    pre-treated in argon at:

    440C,180 min 460C,180 min 480C,180 min 520C,90 min 520C,180 min

    *pre-treated in argon (in TG): 520C,180 min

    peak positions:

    614

    621623

    624

    625

    626

    524

    MS

    ion

    curr

    ent [

    10-9A]

    temperature [C]

    Fig. 4.13. Mass spectrometric analysis of pre-treated TiH2 powders in Ar during heating from 30C to

    1200C at 10 K/min in an argon atmosphere (low temperature range omitted). Curve for untreated powder

    as already given in Fig. 4.8, shown for comparison. Peak positions are specified in the diagram (in C).

    4.2.4.2 Decomposition of pre-treated TiH2 powders in air

    The powders pre-treated in air prior to thermal analysis show a different decomposition behaviour as

    demonstrated in Fig. 4.14. As it was already observed for pre-treatment under argon, heat treatment for

    180 min at 440C in air eliminates the first decomposition stage. Unlike treatment under argon, however,

    treatment in air significantly changes the temperature of both the onset of gas evolution and the

    temperature at which maximum gas expansion occurs. Treatments up to 360 min or treatments at higher

    temperatures up to 520C shift the onset up to +170 K and peak positions up to +52 K. For all treatments

    gas release is smaller than that for untreated powders at T750C and ceases earlier, namely at 900C

    as compared to 950C for untreated powder.

  • 4 Results: Powders

    44

    400 600 800 10000.0

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2

    677*

    peak positions:

    untreated

    pre-treated in air: 440C, 180 min 480C, 180 min 500C, 180 min 520C, 90 min 520C, 180 min 520C, 360 min

    *pre-treated in air (in TG): 520C, 180 min

    657643

    677678

    673663

    626

    524

    MS

    ion

    curr

    ent [

    10-9

    A]

    temperature [C]

    Fig. 4.14. Mass spectrometric analysis of pre-treated TiH2 powders in air during heating from 30C to

    1200C at 10 K/min in an argon atmosphere (low temperature range omitted). Curve for untreated powder

    as already given in Fig. 4.8, is shown for comparison. Peak positions are specified in the diagram (in C).

    The decomposition curve for the powder exposed to the temperature profile given in Fig. 4.12 prior to

    characterisation (180 minutes at 520C after a heating ramp, curve marked with an asterisk in Fig. 4.14)

    shows the same onset and peak temperature but less hydrogen. Compared to the analogous effect under

    argon the reduction in hydrogen evolution is much smaller.

    Mass spectrometric curves of hydrogen release from pre-treated TiH2 powder in air supplied from various

    manufacturers are shown in Fig. 4.15. It can be seen that in all three cases the same pre-treatment

    parameter (480C for 180 min) as shown in Table 3.3 was used. The maximum peak shift is -9 K and -14

    K for GFE and SeJong Materials, respectively in relation to TiH2 powder supplied from Chemetall

    Company. Applying the criterion of 1% the onset temperature of H2 release is 495C, 510C and 528C

    for the GFE, SeJong Materials and Chemetall producer, respectively. It shows that the maximum

    difference in the onset temperature is 33 K after pre-treatment in air under the same conditions.

  • 4 Results: Powders

    45

    400 600 800 10000.0

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2

    Chemetall

    MS

    ion

    curr

    ent [

    10-9A]

    temperature [C]

    peak positions:

    648

    643657

    GFE SeJong

    Fig. 4.15. Mass spectrometric analysis of TiH2 powders pre-treated in air at 480C for 180 min supplied

    from different manufacture during heating from 30C to 1200C at 10 K/min in an argon atmosphere (low

    temperature range omitted). Curve of TiH2 powder produced from Chemetall is already given in Fig. 4.14

    and is shown for comparison. Peak positions are specified in the diagram (in C).

    4.2.5 Isothermal experiments at 600C of TiH2 powders (after constant heating at 20 K/min)

    Isothermal experiments at 600C with a heating rate increased to 20 K/min intended to simulate the

    temperature profile which occurs during foaming of common aluminium alloys. First experiments were

    carried out with TiH2 powder both in the untreated state and after heat treatments in air at various

    temperatures (440, 460, 480, 520C) and times (90 min, 180 min). Fig. 4.16 shows a first release of

    hydrogen gas from the untreated powder after 21 min at 420C ("1") applying the 2.5% criterion (see

    Chapter 4.2). A gas release peak appears after 34 minutes. At this time the maximum temperature has

    been reached.

  • 4 Results: Powders

    46

    The curve resembles the one shown in Fig. 4.12 although there the heating rate and isothermal

    temperature were lower. Employing oxidised powders, the onset of hydrogen evolution is shifted to

    between 510C ("2") and 595C ("3") depending on the pre-treatment in analogy to the constant heating

    rate experiments in Fig. 4.14. The maximum of gas release is also shifted to later times for pre-treated

    powders.

    0 10 20 30 40 50 60 70 130 140 1500.0

    0.2

    0.4

    0.6

    0.8

    1.0

    0 10 20 30 40 50 60 70 130 140 150time [min]

    peak positions:

    tem

    pera

    ture

    [C

    ]

    20 K

    /min

    temperature

    43

    41

    3836

    3534

    untreated 440C, 180 min 460C, 180 min 480C, 180 min 520C, 90 min 520C, 180 minM

    S io

    n cu

    rrent

    [10-

    9 A]

    time [min]

    100

    200

    300

    400

    500

    6003

    2

    1

    Fig. 4.16. Mass spectrometric analysis of loose TiH2 powders pre-treated under air (analysis is carried out

    under Ar). Isothermal conditions at 600C apply after heating at 20 K/min. Numbers denote times at

    which H2 evolution peaks (minutes) occur.

  • 4 Results: Powders

    47

    4.2.6 X-ray diffraction investigations

    4.2.6.1 Ex-situ X-ray diffraction of TiH2 powder

    XRD patterns for TiH2 powder pre-treated in air at 520C for 90, 180 and 360 min are shown in Fig. 4.17

    and are compared with those for the untreated powder. The unmarked peaks correspond to TiH2 or sub-

    stoichiometric TiHx compounds. The untreated powder was found to be single-phased. Its composition

    was identified as TiH1.924 (3.89 wt.% hydrogen) with a fcc crystal structure (a=0.4448 nm) using powder

    diffraction files. No signs of a tetragonal phase were found. After annealing in air the peaks

    corresponding to TiHx are shifted to higher angles (see Chapter 3.2.4) corresponding to a shift in lattice

    constant to 0.4392, 0.4385 and 0.4378 nm for 90, 180 and 360 min at 520C, respectively. In addition,

    peaks corresponding to Ti3O and TiO2 (rutile) are detected, corresponding to a hexagonal and a tetragonal

    crystal structure, respectively.

    d )

    c )

    b )

    a )

    O

    O

    OO

    O

    OO

    O

    O

    O

    O

    O

    O

    XXXX

    XX

    XX

    X

    X

    XX

    X

    X

    2 th e ta [d e g re e s ]

    O

    2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0

    O

    XX

    X X

    Fig. 4.17. XRD patterns for TiH2 powder; a) untreated, b-d) pre-treated in air at 520C for 90, 180 and

    360 min; x: TiO2, o: Ti3O.

  • 4 Results: Powders

    48

    4.2.6.2 In-situ X-ray diffraction of TiH2 powder

    Fig. 4.18 shows selected XRD patterns during an in-situ experiment at the current temperatures during

    heating of untreated TiH2 powder. There are three different kinds of diffraction peaks which occur during

    the measurement. The present phases are marked on the diagram. As it is shown in Fig. 4.18, at room

    temperature the -TiH2 phase is present. With the heating to 548C a high temperature -Ti phase is

    present and some presence of -Ti is observed, as well. At the temperature of 740C the presence of -Ti

    is detected.

    30 35 40 45 50 55

    Ti Ti

    Ti

    Ti

    Ti

    Ti Ti

    TiHx

    TiHx

    Ti

    T=740C

    T=548C

    T=200C

    inte

    nsity

    [a.u

    .]

    2 theta

    Fig. 4.18. Scattering curves corresponding to the phases present during heating of untreated TiH2 powder

    Fig. 4.19 shows the phase change during heating of untreated TiH2 powder. Three phases are recognised

    in this diagram. -TiH2 phase is present from room temperature up to 540C, -Ti phase increase at

    520C and after 550C it slowly decreases. Already at around 520C -Ti appears and the presence of this

    phase increases until the end of the experiment. Around 420C the lattice constant of -TiH2 phase

  • 4 Results: Powders

    49

    decreases. Obviously, this change is caused by the emission of hydrogen due to the decreasing hydrogen

    amount in the sample.

    200 300 400 500 600 7000

    20

    40

    60

    80

    100

    Ti Ti TiH2ph

    ase

    com

    posi

    tion

    [vol

    . %]

    temperature [C]

    Fig. 4.19. Changes of phases during heating of untreated TiH2 powder from 200C to 780C at 10 K/min

    in He atmosphere.

    Fig. 4.20 shows the scattered intensities during in-situ heating of untreated TiH2. Three phases are

    recognised in this diagram: -TiH2 from room temperature to 540C, -Ti phase appears at around 520C,

    and -Ti phase at around 520C as well. From 750C the whole sample transforms to -Ti phase. Before

    first hydrogen emission takes place the peak sifting to lower angles is observed. This happens due to the

    thermal expansion of the sample during the heating from room temperature. Important detail is the bend

    (marked with an arrow on Fig. 4.20) to larger scattering angles before sample reaches the first transition

    temperature. This bend denotes a large emission of hydrogen. It is assumed that the change in the lattice

    constant in pure -phase is proportional to the emission of hydrogen.

  • 4 Results: Powders

    50

    Fig. 4.20. Scattered intensities are plotted in grey scale. Dark regions are high, and light regions are low

    intensities of untreated TiH2 powder during heating from 200C to 780C at 10 K/min.

    Fig. 4.21 shows the change of the lattice constant of -TiH2 with temperature. The thermal expansion (red

    curve) can be estimated from the slope below 400C. Since it is known that at 520C the phase boundary

    is reached it is possible to calculate the hydrogen concentration in the sample.

  • 4 Results: Powders

    51

    200 300 400 500 600 7000.980

    0.985

    0.990

    0.995

    1.000

    1.005

    1.010

    1.015

    1.020a/

    a 300

    temperature ]C]

    Fig. 4.21. Change of lattice constant (a) relative to the lattice constant at T=300C (a300) of untreated TiH2

    powder during heating from 200C to 780C at 10 K/min.

    Fig. 4.22 shows the phase change during heating of TiH2 powder which was before pre-treated in air at

    480C for 180 min. There is an obvious similarity to the experiments shown in Fig. 4.19. Three phases

    are recognised in this diagram as well. -TiH2 phase is present from room temperature up to 550C. -Ti

    phase increases suddenly at 550C and after 565C it slowly decreases. At room temperature the presence

    of -Ti is found. At approximately 520C there is a sudden increase in the amount of -Ti and it is

    increasing until 100% and holds it to the end of the experiment.

  • 4 Results: Powders

    52

    200 300 400 500 600 7000

    20

    40

    60

    80

    100

    Ti Ti TiH2ph

    ase

    com

    posi

    tion

    [vol

    . %]

    temperature [C]

    Fig. 4.22. Changes of phases during heating of pre-treated TiH2 powder in air at 480C for 180 min from

    200C to 780C at 10 K/min in He atmosphere.

    4.2.7 TEM investigations

    Fig. 4.23 (left) shows a bright field (BF) TEM micrograph of an untreated TiH2 particle. The surface of

    the particle does not show any signs of oxide layers in this image. In contrast, TiH2 particles pre-treated in

    air at 480C for 180 min exhibit a clearly visible surface layer with a rather constant thickness of about

    100 nm (Fig. 4.23, right). In the example given the TiH2 fragment was so thin that oxide layers can be

    identified on both sides. The oxide layers are polycrystalline as verified by the selected area electron

    diffraction pattern (SAED) which is given in Fig. 4.23 together with the pattern corresponding to the non-

    oxidised TiH2 core. From the position of the diffraction peaks a number of interplanar spacings dexp can be

    derived. These are given in the first column of Table 4.2.

  • 4 Results: Powders

    53

    A qualitative analysis of the oxygen content inside the oxide layer is also carried out. The EDX

    measurements were performed along a line diagonally across the oxide layer. As a quantitative

    determination of light elements is difficult without precise stand