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

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

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 600°C 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 600°C 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 400°C, 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 [14–21]. 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)

Hütte 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 1950’s 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 (Saarbrücken,

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 400°C for untreated hydride and commercial

aluminium alloys have solidus temperatures above 525°C 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 α; (2) bcc β; (3) δ, a fcc hydride; (4) fct ε, a

tetragonally distorted hydride with axial ratio c / a <1.

0 10 20 30 40 50 60 7

-200

0

200

400

600

800

1000

ε

δ

βTiαTi

tem

pera

ture

[°C

]

hydrogen (atomic %)

Fig. 2.4. Equilibrium phase diagram of titanium-hydrogen system [62].

2 Metal foams

13

The hydrogen atoms in the δ 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 δ hydride starts to be transformed to fct ε hydride. When the fcc δ hydride starts to be transformed to

fct ε hydride the temperature was observed around 25°C [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 30°C

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 37°C.

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


3.1 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

[µm]

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

TiH2 Chemetall GmbH, Frankfurt 98.8 3.76 < 63 Si<0.15, Fe<0.09,

Mg<0.04 min 3.8

TiH2 GFE GmbH, Frankfurt 99.7 3.76 < 45-500 Si<0.3, Fe<0.07,

C<0.05 -

TiH2 SeJong Materials GmbH, Corea 99.5 3.76 < 44 Si<0.05, Fe<0.08,

Mg<0.02 min 3.8

ZrH2 Chempur GmbH, Karlsruhe 99.7 5.47 < 45 Fe<0.1, Ca<0.01,

Hf<0.012 -

HfH2 Chempur GmbH, Karlsruhe 99.5 11.4 - Si<0.05, Fe<0.01,

Al<0.05 <1

Table 3.2. Specification of metallic powders used for making metal foams.

Powder

Producer

Purity

[%]

Density

[g cm-3]

Colour

Particle size

[µm]

Aluminium Eckart Werke, Ranshofen, Austria 98.8 2.7 light grey < 160

Copper Chempur GmbH, Karlsruhe 99.7 8.93 red > 250

Silicon Ölschläger 99.5 2.33 dark grey 60-100


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 520°C 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 ≥ 540°C). It is worth

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

precursors.


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.


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.


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 450°C 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.


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 750°C. All alloys were foamed at 650°C. 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 2500°C 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.


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 3µm diamond grinding medium at MD-Dac (Struers). Final polishing was

done with 1µm 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 520°C 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.


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).


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 520°C. The diffraction patterns were measured for 2Θ ranging from 15° to 90°

at steps

of 0.02°. An acquisition time of 5 s per step using CuKα 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.


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 2Θ angle between 30 and 60 degrees combined with the

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

(a)

(b)


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 1000°C 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 2Θ (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 780°C with a heating rate of 10 K/min. Starting from 200°C 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.


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 650°C

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

slightly lower at 630°C 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 520°C 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 520°C 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 480°C for 180 min.

(a)

4 Results: Powders

29

(b)

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

480°C, green at 500°C and light brownish to dark brown at 520°C, 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) 480°C for 180 min c) 500°C for 180 min d) 520°C for 90 min e) 520°C for

180 min f) 520°C 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 [µm] 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 Ölschläger 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 520°C – 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 600°C 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 415°C (marked as "1"), the peak position is located at 524°C. A second

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

626°C. The shoulder around 800°C 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 390°C ("2").

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

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

to the calculated value of 4% (TiH2→Ti corresponds to 50→48 = -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 [µ

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) i

on c

urre

nt [1

0-9A

]

temperature [°C]

(b)

Fig. 4.8. Analysis of untreated TiH2 powder heated from 30°C to 1200°C 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 360°C. Powder from SeJong Materials

releases hydrogen at 365°C applying the same criterion. 415°C 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 ≈ 900°C for GFE powder and ≈ 950°C 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 30°C to 1200°C 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 ≈ 250°C ("1") and leads to a continuous mass increase. Due to strong exothermal effects the

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

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

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

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

(TiH2→TiO2 corresponds to 50→80 = +60%).

4 Results: Powders

39

0 20 40 60 80 100

200

400

600

800

1000

1200

4

3

2

1MS hydrogen

MS water

temperature

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 30°C to 1200°C 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 520°C 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 520°C (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 500°C 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 520°C 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 250°C (TG curve) ("1") after a small initial mass change and reveals

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

by the MS curve gets notable at approximately 350°C ("2") and reaches a maximum at 507°C 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 520°C 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 480°C instead of 520°C

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 (480°C)

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 520°C 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

480°C, 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 440°C 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 415°C is not changed significantly by pre-treatments under argon, with all onset

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

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

for powders pre-treated at 480 and 520°C.

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:

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

*pre-treated in argon (in TG): 520°C,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 30°C to

1200°C 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 440°C 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 520°C 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 T≥750°C and ceases earlier, namely at ≈ 900°C

as compared to ≈ 950°C 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: 440°C, 180 min 480°C, 180 min 500°C, 180 min 520°C, 90 min 520°C, 180 min 520°C, 360 min

*pre-treated in air (in TG): 520°C, 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 30°C to


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 520°C 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 (480°C 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 495°C, 510°C and 528°C

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 480°C for 180 min supplied

from different manufacture during heating from 30°C to 1200°C 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 600°C of TiH2 powders (after constant heating at 20 K/min)

Isothermal experiments at 600°C 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, 520°C) and times (90 min, 180 min). Fig. 4.16 shows a first release of

hydrogen gas from the untreated powder after 21 min at 420°C ("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 510°C ("2") and 595°C ("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 440°C, 180 min 460°C, 180 min 480°C, 180 min 520°C, 90 min 520°C, 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 600°C 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 520°C 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 520°C, 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 520°C 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 δ-TiH2 phase is present. With the heating to 548°C a high temperature ß-Ti phase is

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

is detected.

30 35 40 45 50 55

β Tiβ Ti

β Ti

α Ti

α Ti

α Ti α Ti

δ TiHx

δ TiHx

β Ti

T=740°C

T=548°C

T=200°C

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. δ-TiH2 phase is present from room temperature up to 540°C, ß-Ti phase increase at

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

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

α Ti β Ti TiH2ph

ase

com

posi

tion

[vol

. %]

temperature [°C]

Fig. 4.19. Changes of phases during heating of untreated TiH2 powder from 200°C to 780°C 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: δ-TiH2 from room temperature to 540°C, ß-Ti phase appears at around 520°C,

and α-Ti phase at around 520°C as well. From 750°C the whole sample transforms to α-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 δ-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 200°C to 780°C at 10 K/min.

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

curve) can be estimated from the slope below 400°C. Since it is known that at 520°C 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=300°C (a300) of untreated TiH2

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

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

480°C 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. δ-TiH2 phase is present from room temperature up to 550°C. ß-Ti

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

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

α Ti β 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 480°C for 180 min from

200°C to 780°C 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 480°C 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 standards only the relative change can be

given. It was found that the oxygen content drops to about half the value at the outer side when crossing

the layer.

Fig. 4.23. Bright field TEM images of, left: untreated TiH2 particle, right: particle pre-treated in air at

480°C for 180 min. The corresponding electron diffraction patterns within the [110]-zone axis of the TiH2

particle as well as the oxide layer around the core (marked "A") are shown in the middle.

In Fig. 4.24 a HRTEM image of this polycrystalline oxide layer around the core of a TiH2 particle is

displayed. Small crystals between 3 and 10 nm in diameter oriented in various directions are observed.

4 Results: Powders

54

Table 4.2. Interplanar spacings dexp in the oxide layer around TiH2 particles annealed at 480°C for 180

min as measured by SAED. The 2nd and 4th column contain values calculated from lattice parameters for

TiO2 and Ti3O taken from literature [68] assuming one of the crystallographic directions given in columns

3 and 5.

Calculated d

For TiO2 For Ti3O

Experimental d

from SAED

dexp (Ǻ) d (Ǻ) [h,k,l] d (Ǻ) [h,k,l]

4.6 - - 4.7 [003]

3.2 3.2 [110] 3.2 [103]

2.8 - - 2.8 [005]

2.3 2.29 [200] 2.3 [006]

2.2 2.18 [111] 2.2 [113]

2.1 2.05 [210] 2.1 [202]

1.9 - - 1.9 [115]

1.6 1.6 [211] 1.6 [210]

Fig. 4.24. HRTEM image of polycrystalline oxide layer around the core of TiH2 particle pre-treated in air

at 480°C for 180 min.

4 Results: Powders

55

4.3 Characterisation of ZrH2 and HfH2 powder (as a blowing agent)

As it is shown in Chapter 4.2.4 the pre-treatment of TiH2 powder changes the decomposition properties of

this blowing agent. It makes sense to have a short look at the hydrogen release of other blowing agents

like ZrH2 and HfH2. Therefore, in this chapter the influence of pre-treatment in argon or air prior to the

thermal analysis on hydrogen release of the blowing agents ZrH2 and HfH2 powder is investigated.

4.3.1 Decomposition of pre-treated ZrH2 and HfH2 powders

Powders pre-treated in air or argon prior to characterisation are being presented in the following chapter.

At this stage no further oxidation occurs since thermal analysis is carried out under argon. ZrH2 and HfH2

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

4.3.1.1 Decomposition of pre-treated ZrH2 powders in argon

Fig. 4.25 shows the decomposition behaviour of untreated ZrH2 powder as well as ZrH2 previously pre-

treated under argon. Pre-treatment parameters are specified in Table 3.3. Argon flow and a heating rate of

5 K/min were applied. Three peaks can be observed in the mass spectroscopic (MS) curve of untreated

powder. The first step of hydrogen release as identified by mass spectrometry applying the 1% criterion

starts at 205°C, a first peak is located at 436°C. A second dehydrogenation step starts at about 480°C and

reaches a maximum at 552°C. The third dehydrogenation step starts at about 580°C with a small shoulder

and reaches the maximum at 752°C.

Heat treatment at 400°C for 180 min eliminates the first decomposition stage and leaves two peaks only.

Hydrogen release in two stages remains the same even for treatments at higher temperatures (520°C) or

for longer annealing times (180 min). The position of the remaining two peaks is shifted to lower

temperatures compared to that of untreated ZrH2. The maximum peak shift is -120 K. Another

observation is that the onset of gas evolution at 205°C is not changed significantly by pre-treatments

under argon. For all pre-treatments the onset temperatures are between 205°C and 211°C. Gas release

levels off for temperatures beyond the peak and ends between ≈ 950°C for untreated powder or powders

pre-treated at 400°C and ≈ 900°C for powders pre-treated at 480 and 520°C.

4 Results: Powders

56

200 400 600 800 1000

0.1

0.2

0.3

0.4

0.5

0.6peak positions:

632656

663729

752

untreated

pre-treated in argon a t:

400°C , 180 m in 480°C , 180 m in 520°C , 90 m in 520°C , 180 m in

MS

ion

curr

ent [

10-9A

]

tem perature [°C ]

Fig. 4.25. Mass spectrometric analysis of untreated and pre-treated ZrH2 powders under Ar during heating

from 30°C to 1100°C at 5 K/min in an argon atmosphere (low temperature range omitted). Peak positions

are specified in the diagram (in °C).

4.3.1.2 Decomposition of pre-treated ZrH2 powders in air

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

shown in Fig. 4.26. As it was already observed for pre-treatment under argon, heat treatment at 400°C for

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

in air significantly changes the temperature of the onset of gas evolution. Treatments up to 180 min or

treatments at higher temperatures up to 520°C shift the onset up to +258 K, while peak positions decrease

up to -103 K. For all treatments gas release is smaller than that for untreated powders and appear earlier,

namely at ≈ 850°C for treatment at 400°C, and ≈ 800°C for treatment at 480°C, 180 min and 520°C, 90

min, and ≈ 750°C for treatment at 520°C, 180 min as compared to ≈ 950°C for untreated powder.

4 Results: Powders

57

200 400 600 800 10000.0

0.1

0.2

0.3

0.4

0.5

0.6peak positions:

649

678682

698

752 untreated

pre-treatm ent in a ir a t:

400°C , 180 m in 480°C , 180 m in 520°C , 90 m in 520°C , 180 m in

MS

ion

curr

ent [

10-9A

]

tem perature [°C ]

Fig. 4.26. Mass spectrometric analysis of pre-treated ZrH2 powders in air during heating from 30°C to


as already given in Fig. 4.25 is shown for comparison. Peak positions are specified in the diagram (in °C).

4.3.1.3 Decomposition behaviour of HfH2 powder

Fig. 4.27 shows the decomposition behaviour from untreated and pre-treated HfH2 powder in air at 480°C

for 180 min prior to thermal analysis. Applying the 1% criterion untreated powder releases hydrogen at

273°C (marked as "1"). Pre-treatment in air shifts the hydrogen release to 460°C (marked as "2"). After

the pre-treatment in air HfH2 releases five times less hydrogen then compared to untreated powder. Peak

maximum shifts up to +209 K for pre-treated powder.

4 Results: Powders

58

200 400 600 800 10000.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40peak positions:

749

540 untreated

pre-treated in air at:

480°C, 180 min

MS

ion

curr

ent [

10-9A

]

temperature [°C]

12

Fig. 4.27. Analysis of untreated and pre-treated HfH2 powder in air heated from 30°C to 1200°C at 10

K/min in an argon atmosphere. Mass spectrometric (MS) signals for H2 are shown. Peak positions are

specified in the diagram (in °C) and the numbers in circles mark specific temperatures discussed in the

text.

5 Results: Foams

59

5 Results: Foams

5.1 Isothermal experiments at 600°C of AlSi6Cu4 precursor with TiH2 powders (after constant

heating at 20 K/min)

The analogous isothermal experiments on foamable precursors employing the same parameters than for

loose powders can be seen in Fig. 5.1. Here the TiH2 particles are embedded in a dense aluminium alloy

matrix. The ion current measured has a magnitude of about 1/20 of the current in Fig. 4.16 reflecting the

blowing agent content of 0.5 wt.% in the precursor. Hydrogen release from the untreated powder begins

at ≈ 370°C ("1") for this case. As the MS curve shows a pronounced edge this point was used as the onset

temperature of hydrogen release. For precursors made from pre-oxidised powders the hydrogen release

occurs between ≈ 450°C ("2") and ≈ 570°C ("3") depending on pre-treatment parameters. Compared to

the experiments on loose powders as shown in Fig. 4.16 the hydrogen release from the precursor is shifted

to an earlier moment for each of the pre-treatment states.

0 10 20 30 40 50 60 70 130 140 150

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

100

200

300

400

500

600

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

20 K

/min

3

2

1

temperaturepeak positions:

3837

3635

32

31

28

tem

pera

ture

[°C

] untreated 440°C, 180 min 460°C, 180 min 480°C, 180 min 520°C, 90 min 500°C, 180 min 520°C - 180 minM

S io

n cu

rren

t [10

-11 A

]

time [min]

Fig. 5.1. Mass spectrometric analysis of untreated and pre-treated TiH2 powders under air in foamable

precursor (analysis is carried out under Ar). Isothermal conditions at 600°C apply after heating at 20

K/min. Numbers denote times at which H2 evolution peaks (minutes).

5 Results: Foams

60

5.2 TEM investigations of Al precursor

Fig. 5.2 shows a BF TEM image of a pre-treated TiH2 particle embedded in compacted aluminium. The

TiH2 particle and the Al matrix are marked in this figure. Magnified views of the oxide layer on both sides

of the TiH2 particle are given. The layer shown on the left side splits into two sub-layers. Thickness and

structure of the layers appear to be uniform. The magnification on the right side shows further variations.

Coming from right to left the double layer structure vanishes and turns into a thick zone containing oxide

and aluminium. Possibly, the uniform structure of the oxide layer has been disrupted here by the

mechanical forces occurring during compaction.

Fig. 5.2. Bright field TEM image of TiH2 particle pre-treated in air at 480°C for 180 min and embedded

in an Al matrix by powder pressing. Enlarged views of image in the centre are shown on the left and

right.

5.3 Impact of pre-treatment on foaming behaviour

In this chapter the difference of using pre-treated blowing agent powders and its influence on the foaming

behaviour and a foam quality is shown. The types of prepared precursors for each blowing agent are

already given in Fig. 3.1.

Fig. 5.3 shows an optical image of the microstructure of foamable precursor material after hot pressing.

This picture shows homogeneously distributed powder particles after compaction. Light grey TiH2

particles, darker angular shaped silicon particles, and red copper particles can be seen in the aluminium

metallic matrix.

5 Results: Foams

61

Fig. 5.3. Foamable AlSi6Cu10 precursor material containing 1 wt% TiH2 [74].

5.3.1 Precursors with TiH2 powders

Fig. 5.4 shows the interdependence between foaming time and the final temperature at which experiments

were stopped in order to reach a desired expansion of about 60%. The data were measured on samples

containing differently pre-treated TiH2. The heating conditions were identical in all cases, except for

foaming Al precursor. Here, the temperature of the furnace was increased to 750°C and therefore this

curve cannot be directly compared.

In all cases, the final temperature Tf increases steadily with foaming time tf. This dependence reflects the

temperature profile in the sample during heating and is approximated by an exponential in all cases as

shown in the diagrammes in Fig. 5.4. The main observation is that a precursor which contains a blowing

agent that has been pre-treated more heavily (at a higher temperature-520°C or for a longer time-180 min)

takes longer (and correspondingly a higher final foaming temperature) to expand to a given volume. The

Cu

Si

TiH2

Al

5 Results: Foams

62

final foaming temperature Tf and the foaming time tf when the expansion of about 60% is reached

decrease with the higher copper content alloys.

400 450 500 550670

680

690

700

710 520°C-180 min

520°C-90 min

480°C-180 min

440°C-180 min

untreated

T f [°C

]

tf [s]

(a)

5 Results: Foams

63

240 280 320 360 400 440 480600

610

620

630

640

650

660

520°C-180 min

520°C-90 min

480°C-180 min

440°C-180 min

untreated

T f [°C

]

tf [s]

(b)

150 200 250 300 350 400 450580

590

600

610

620

630 520°C-180 min

520°C-90 min

480°C-180 min

440°C-180 min

untreated

T f [°C

]

tf [s]

(c)

5 Results: Foams

64

100 150 200 250 300

540

550

560

570

580

590

600

610520°C -180 m in

520°C -90 m in

480°C -180 m in

440°C -180 m in

untreated

T f [°C

]

t f [s ]

(d)

Fig. 5.4. Final sample temperature Tf vs. foaming time tf for aluminium samples foamed up to a given

porosity level of ≈ 60% with TiH2 pre-treated in different ways as specified in the diagrammes: a) Al

precursor, b) AlSi6 precursor, c) AlSi6Cu4 precursor, d) AlSi6Cu10 precursor. For Al precursor, furnace

temperature was 750°C. For the experiments under b, c and d the temperature in the furnace was 650°C.

The influence of the oxidation level of the blowing agent on the shape of the pores created during

foaming is demonstrated in Fig. 5.5. The metal foam quality in relation to the pore structure is explained

in Chapter 2.2.1.

Untreated TiH2 produced pores with jagged boundaries and a non-uniform size distribution. Associated

with an increasing oxidation level of the hydride changes in pore shape are notable. Pores become more

spherical and have smooth surfaces even for the shorter pre-treatments. Moreover, there is a positive

influence on the pore size distribution. The powder pre-treatment leads to a more homogeneous pore size

distribution with equal pore morphology. In all cases powders pre-treated at 520°C produce more uniform

foams than powders pre-treated at lower temperatures of 440°C. An extended pre-treatment for a very

long time leads to increasingly coarser foams.

5 Results: Foams

65

The alloys with the increasing copper content have smoother and more uniform pore shapes. There is an

improvement of the pore structure when copper is added. A higher copper content provides a larger

amount of melt in the alloy. In combination with pre-treated powders pores are formed with the tighter

pore size distribution. Therefore, there is a significant improvement of the pores homogeneity.

5 Results: Foams

66

i) Al foamed using a) untreated TiH2 (porosity P=58%), b) pre-treated TiH2, 440°C, 180 min (P=56%), c)

480°C, 180 min (P=61%), d) 520°C, 90 min (P=60%), e) 520°C, 180 min (P=62%).

5 Results: Foams

67

ii) AlSi6 foamed using a) untreated TiH2 (porosity P=62%), b) pre-treated TiH2, 440°C, 180 min

(P=60%), c) 480°C, 180 min (P=63%), d) 520°C, 90 min (P=63%), e) 520°C, 180 min (P=58%).

5 Results: Foams

68

iii) AlSi6Cu4 foamed using a) untreated TiH2 (porosity P=53%), b) pre-treated TiH2, 440°C, 180 min


5 Results: Foams

69

iiii) AlSi6Cu10 foamed using a) untreated TiH2 (porosity P=61%), b) pre-treated TiH2, 440°C, 180 min


Fig. 5.5. Foaming of aluminium precursors using differently pre-treated TiH2 powders. i) Al, ii) AlSi6, iii)

AlSi6Cu4, iiii) AlSi6Cu10. The direction of expansion was from bottom to top.

5 Results: Foams

70

5.3.2 Precursors with ZrH2 and HfH2 powders

Fig. 5.6 shows the same type of the experiment as explained in Chapter 5.3.1. Experiments were stopped

in order to reach an expansion of about 60%. This time the data were measured only on AlSi6Cu4

samples containing differently pre-treated ZrH2 and HfH2 powders. The heating conditions were identical

in both cases and the furnace temperature was 650°C. Obviously the precursor which contains a blowing

agent that has been more heavily pre-treated (at a higher temperature or for a longer time) needs more

time to expand to a given volume. It can be seen that the precursors with HfH2 powder are foamed

slightly shorter compared with ZrH2 precursors as shown in Fig. 5.6. The results of HfH2 powder are

linear connected. As it is shown in Fig. 5.7 and 5.8 there is an obvious improvement in the pore structure

when using pre-treated ZrH2 and HfH2 powders, respectively. The pores become more spherical after pre-

treatment and the distribution is more homogeneous (see Chapter 2.2.1). Compared to Fig. 5.5. iii) it can

be seen that the foaming starts earlier with these two powders compared to precursors with TiH2 powder,

indicating differences in H2 releases from theses powders (see Fig. 6.1).

100 150 200 250 300 350

580

590

600

610

620

630

480°C-180 min

untreated

ZrH2 precursor 520°C-180 min

520°C-90 min

480°C-180 min

400°C-180 min

untreated

T f [°C

]

tf [s]

HfH2 precursor

Fig. 5.6. Final sample temperature Tf vs. foaming time tf for aluminium samples AlSi6Cu4 foamed up to a

given porosity level of ≈60% with ZrH2 and HfH2 pre-treated in air on different ways as specified.

Furnace temperature was 650°C.

5 Results: Foams

71

Fig. 5.7. AlSi6Cu4 foamed using a) untreated ZrH2 (porosity P=58%), b) pre-treated ZrH2 in air, 400°C,

180 min (P=50%), c) 480°C, 180 min (P=50%), d) 520°C, 90 min (P=53%), e) 520°C, 180 min (P=52%).

The direction of expansion was from the bottom to the top.

5 Results: Foams

72

Fig. 5.8. AlSi6Cu4 foamed using a) untreated HfH2 (porosity P=63%), b) pre-treated HfH2 in air, 480°C,

180 min (P=63%). The direction of expansion was from the bottom to the top.

5.3.3 Expandometry AlSi6Cu4 precursor with TiH2 powder

Fig. 5.9 shows results of non-interrupted expansion experiments for the six different modifications of

TiH2. For all the samples an onset of expansion after about 600 s followed by continuous foaming up to

expansion factors above 4.5 was observed. The foaming process can be divided into three stages for all

the experiments:

(i) a small initial expansion after which there is a short pause, followed by,

(ii) a stage with a high expansion rate which turns into,

(iii) an almost linear expansion regime with a lower expansion rate. Expansion comes to an end

rather abruptly, after which the volume is constant.

5 Results: Foams

73

0 500 1000 1500 20001.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

iii

ii

(5)(2)

untreated (1) 440°C-180 min (2) 460°C-180 min (3) 480°C-180 min (4) 520°C- 90 min (5) 520°C-180 min (6)

expa

nsio

n η=

h/h 0

time [s]

T

η

(1)

(3)(4)(6)

i

0

100

200

300

400

500

600

700

T [°

C]

Fig. 5.9. Expansion curves of precursors based on six different modifications of TiH2. Volume expansion

η = V/V0 = h/h0 is given for all samples; the temperature curve is shown for one sample only. The various

expansion stages discussed are marked by (i), (ii), and (iii).

Defining the onset of expansion as η = 1.01, i.e. 1% expansion, one can read the time and temperature

from Fig. 5.9 and display it as shown in Fig. 5.10. The points determined must all lie on the global

temperature profile imposed on the sample by the furnace (broken lines). That this is indeed the case

shows the good control of temperature in the experiment. In an analogous way points corresponding to

volume expansions of η = 2 and 2.5 (open symbols) were determined.

5 Results: Foams

74

600 650 700 750 800 850 900560

580

600

620

640

660

4x2.5x2x1.01x

interruptedfoaming

continuousfoaming

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

tem

pera

ture

[°C

]

time [s]

Fig. 5.10. Sample temperatures Tf measured in foaming experiments in the moment tf when a given

volume expansion η was reached. Open symbols: values for continuous foaming experiments as read

from Fig. 5.9, full symbols: values for interrupted foaming experiments at the moment the lamps were

turned off and air cooling was started. Broken lines are the overall temperature profiles applied.

Interrupted foaming leads to samples in different expansion stages which can be analysed after cooling.

Expansion factors of η = 2, 2.5, and 4 were chosen. Foaming was interrupted for a short time before the

desired volume was reached to take account of the thermal inertia of the expandometer which leads to a

certain after-foaming. As a rule, the volume still increased for 45 s after interrupting and the

corresponding volume expansion was about 0.5 on the η-scale in Fig. 5.9. The corresponding foaming

times and temperatures in the moment of initiation of cooling are given in Fig. 5.10 as solid symbols.

They lie on the horizontal line at 630°C.

Some of the corresponding foams are shown in Fig. 5.11. The morphology of foams in different

expansion stages blown with untreated (left) and pre-treated (right column) TiH2 can be compared.

Images corresponding to other modifications of TiH2 were also obtained and evaluated but are not shown

here.

5 Results: Foams

75

Fig. 5.11. Aluminium foams after expansion to a height of approx. 2, 2.5 and 4 times the height of the

original precursor (from top to bottom). Foamed with, left column: untreated TiH2, right: TiH2 treated at

520°C for 180 min. Sample width is 36 mm.

5 Results: Foams

76

SEM images provided further insight into the samples. Figs. 5.12 and 5.13 show low-magnification

images of foams blown with both untreated and pre-treated TiH2.

(a)

5 Results: Foams

77

(b)

Fig. 5.12. SEM images of selected features of a foam expanded with untreated TiH2 to 4 times the height

of the original precursor. a) pore shape (top), and ruptured cell wall (bottom) b) enlarged views of image

at upper left hand side.

5 Results: Foams

78

The evacuation of the SEM chamber took four times longer for the foams blown with untreated TiH2 (as

shown in Fig. 5.12) than for the foams blown with pre-treated TiH2 such as the one in Fig. 5.13,

indicating a different gas permeability of the two foam types.

Fig. 5.13. Enlarged views of SEM image (upper left hand side) of a foam expanded with TiH2 pre-treated

at 520°C for 180 min to 4 times the initial precursor height.

6 Discussion

79

6 Discussion

All three methods applied for the characterisation of powders, namely DSC (or DTA), TG and MS yield

comparable results. The slight shift between the peak positions of DTG and MS is possibly caused by the

time delay which can be explained by the necessary transport of the released H2 from the furnace to the

ionisation chamber of the quadrupole mass spectrometer. Obviously DSC is the most sensitive method

since the onset of hydrogen release is observed at slightly lower temperatures as compared to TG and MS.

To get an overview of the differences of hydrogen release from the used blowing agents, MS curves of

hydrogen release for all three hydrides are plotted in Fig. 6.1. Hydrogen release as identified by mass

spectrometry applying the 1% criterion starts at 273°C (marked as "1") for HfH2 powder. ZrH2 start to

loose hydrogen at 295°C (marked as "2") applying the 1% criterion. In Chapter 4.3.1 lower heating rate

was used in order to reach better resolution of peak maximum therefore the onset temperature appear

earlier. TiH2 release hydrogen at 415°C as already mentioned in Chapter 4.2.1.

TiH2 shift the onset by +120 K compared with ZrH2 powder and +142 K compared with HfH2 powder.

The maximum of the peak position for TiH2 is moved by up to +8 K and +86 K in relation to ZrH2 and

HfH2, respectively. Therefore, TiH2 and ZrH2 powders were selected for detailed investigations as shown

in the results chapter. Later the main focus was put on TiH2 powder, as this powder gives the onset

temperature in the range which is necessary for the production of the metal foamed parts.

6 Discussion

80

200 400 600 800 1000 12000.0

0.2

0.4

0.6

0.8

1.0

1.2

31

2

untreated powders:540

peak positions: 618 626

TiH2

MS

ion

curr

ent [

10-9A

]

temperature [°C]

HfH2

ZrH2

Fig. 6.1. Analysis of untreated TiH2, HfH2 and ZrH2 powders heated from 30°C to 1200°C at 10 K/min, in

an argon atmosphere. Mass spectrometric (MS) signals are shown for H2. The numbers in circles mark

specific temperatures discussed in the text. Curves for TiH2 and HfH2 powders are already given in

Figures 4.8 and 4.27, respectively and are just shown for comparison.

The ZrH2 powder releases hydrogen in three well defined stages. Three steps of hydrogen release from

untreated ZrH2 powder were observed by other authors as well [82]. They reported that the decomposition

of ZrH2 is accompanied by the phase conversions ε → δ + β → α.

Untreated TiH2 was found to have an fcc-based CaF2-structure which corresponds to the δ-phase of the Ti

H system [62]. The tetragonal ε-phase reported for temperatures below ≈25°C [68] is not observed, most

likely because the laboratory temperature exceeded this value. Thermal characterisation of the untreated

TiH2 is leading to results which are in good agreement with measurements given in the literature, see e.g.

Refs. [9,20].

6 Discussion

81

Accordingly, hydrogen starts to be released from TiH2 at about 360-420°C with some variations between

powders of different origin. Integrating the curves in Fig. 4.9, it is observed that all powders release

almost the same amount of hydrogen, 2.2×10−6 As, 2.04×10-6 As and 2.00×10−6 As detected ions for

Chemetall, SeJong and GFE TiH2 powders, respectively. Table 4.1 shows that the TiH2 powder from

Chemetall is the coarsest powder compared with SeJong Materials and GFE powder. Bach et. al [83]

found a similar behaviour. They used TiH2 powder classified by different grades and observed that the

coarsest powder releases gas more delayed. Obviously the origin of the untreated TiH2 powder plays an

important roll on the hydrogen release. This statement has to be taken into account when the similar

experiments are compared to those of other authors. Gergely found an onset at 465°C, Kennedy and

Lopez at 440°C, but at higher heating rates [16,18] and possibly using a coarser powder and slightly

different definitions of the onset temperature.

Three decomposition peaks from untreated TiH2 powder are visible on the MS curve in Fig. 4.8b after the

deconvolution, suggesting that the decomposition process occurs in three steps. 25% of the hydrogen is

released in the first stage in accordance with literature [18]. The appearance of the shoulder around 800°C

indicates a third reaction which, however, is not relevant for the foaming process of aluminium alloys and

it will not be considered here. The existence of first two stages of hydrogen release from TiH2 is not fully

understood up to now. There are different opinions within literature about these phenomena. Some of the

authors claim that H-atom positions are on different places in the TiH2 lattice [64, 84-86], i.e. octahedral

and tetrahedral holes in the δ-TiH2. An interesting observation is that partial transition of hydrogen atoms

from tetra into octahedral pores was observed in titanium hydrides at high temperatures [87]. Therefore,

they claim that there is a tendency for hydrogen migration to octahedral sites where the hydrogen bonding

is weaker than in the tetrahedral sites [87, 88]. Therefore during heating up the hydrogen atoms from the

octahedral positions would first be released. The hydrogen release from the tetrahedral positions takes

place later. It is assumed that the first interpretation responds to the first peak of the hydrogen release and

the second one to the second peak of the hydrogen release [11]. Neutron experiments at room temperature

have shown that hydrogen in titanium is always on the tetrahedral lattice sites, never on octahedral ones

[63]. The most probable reason for double peak stages could be a phase transformations as the hydrogen

concentration decreases [16]. Kennedy [18] claims that first the reduction in the stoichiometry of the

hydride to δ-TiH1.5, and secondly, the progressive heating, leads to the decomposition of the hydride and

the formation of the ß- and α-Ti phases. Schoenfelder stated that after the hydrogen release the particles

behave as described like in the core-shell model [58]. These theoretical considerations were

experimentally not yet verified.

6 Discussion

82

0 10 20 30 40 50 60 70

-200

0

200

400

600

800

100096.5 97.0 97.5 98.0 98.5 99.0 99.5 100.0

-200

0

200

400

600

800

1000

ε

δ

βTiαTi

TG [%]

tem

pera

ture

[°C

]

tem

pera

ture

[°C

]

hydrogen (atomic %)

TG curve of untreated TiH2 powder

Fig. 6.2. TG curve (shown in Fig. 4.8) of untreated TiH2 plotted in Ti-H phase diagram.

Fig. 6.2 shows the TG curve plotted in the Ti-H phase diagram. As it can be seen on Fig. 6.2 the presence

of α phase would be never observed in the in-situ XRD experiments, if the sample would be

homogeneous. Fig. 6.3 shows the enlarged view of Ti-H phase diagram with the H2 composition as

calculated from (i) in-situ XRD measurements (see Chapter 4.2.6.2) and (ii) TG experiment. Curves

behave similar before the first phase transformation as shown in Fig. 6.3.

6 Discussion

83

56 58 60 62 64 66300

350

400

450

500

550

60099.4 99.5 99.6 99.7 99.8 99.9 100.0

300

350

400

450

500

550

600

β+δ phase boundary phase composition curve from in-situ XRD exp.

β+δ

δ-TiH2

TG [%]

tem

pera

ture

[°C

]

tem

pera

ture

[°C

]

hydrogen (atomic %)

TG curve of untreated TiH2 powder

Fig. 6.3. Enlarge view of Ti-H phase diagram with the change in the phase composition curve from in-situ

XRD experiment of untreated TiH2 as well as TG curve of untreated TiH2 powder.

However, after phase transformation (Fig. 4.19) in situ measurements around 520°C show that the α and ß

phase appears simultaneously although the δ phase is still present. Since, there is no state in the phase

diagram where all three phases are present at the same time, it is obvious that the particles behave

inhomogeneously. This can be explained with the core-shell model [58], where the core is the mixture of

δ and ß phases and the shell is α phase with less then 8% hydrogen. Diffusion coefficient of hydrogen

across α-Ti at the temperatures above 500°C is 1.6*10-3 cm2s-1, and pure ß-Ti is 5.39*10-3 cm2s-1, as

reported from Landolt-Börnstein [89]. Since the diffusion of hydrogen through α-Ti is slower than in ß-Ti

this could explain the slowing down of hydrogen release.

In the MS experiments, when the heating of the sample is stopped at 520°C and isothermal conditions are

applied (see Fig. 4.11), the gas evolution falls to well below the level seen when heating is continued. In

the isothermal experiment a sharp peak is observed after 50 minutes shortly before the final temperature

6 Discussion

84

of 520°C has been reached. It corresponds to the first peak in Fig. 4.8. After the peak, a drop to much

lower released rates is measured, whereas for continued heating only a small dip is observed before the

occurrence of the second peak. The shoulder in the isothermal peak between 60 and 100 minutes indicates

that the temperature is high enough to ensure a continuous gas evolution. It is plausible to assume that the

hydrogen remaining in the powder, after passing the first decomposition stage, is released in a second

peak (Fig. 4.8) during continuous heating while it is slowly released at a constant temperature over a

longer period in the isothermal experiment (Fig. 4.11). Accepting this, the shoulder in regime "B" in Fig.

4.11 would mainly correspond to the second decomposition stage. The total accumulated mass loss after

180 min at 520°C is around 2.5%. Therefore, still 1.3% hydrogen is present in the hydride after that time.

It is interesting to compare the MS curves in Fig. 4.11 and in Fig. 4.16 corresponding to the untreated

powder. Due to the higher heating rate in Fig. 4.16 the first and the second peaks fall together. The

shoulder presented in Fig. 4.11 is missing and the entire hydrogen evolution has merged into one broad

peak.

Fig. 4.3 reveals that the oxygen content drops during annealing under argon. This effect is not visible in

the mass curve in Fig. 4.11 since the mass change due to hydrogen evolution is dominating. It can be

assumed that oxygen present in the untreated powder reacts with some of the hydrogen and escapes from

the powder as water vapour. Indeed, a very small water peak was observed with the mass spectrometer

(not shown).

The shapes of the TG curves in Fig. 4.10 and Fig. 4.12 reflect an increase of mass due to the oxidation of

the TiH2 powder. The exothermal reaction at 680°C observed in Fig. 4.10 coincides well with water

formation indicating that it is caused by a reaction between hydrogen and oxygen. A similar effect was

reported by Gergely [16] who observed a well-defined exothermal peak at 685°C in a DTA curve. Mass

accumulation due to oxidation rapidly increases above about 500°C. In the isothermal experiments at

480°C and 520°C the oxidation rate levels off as soon as the temperature has reached a constant value

after 50 minutes (Fig. 4.12), whereas the mass gain continues up to complete oxidation if the temperature

increases further. The reason for the levelling off is that at constant temperatures the rate of reaction

between oxygen and titanium is constant, but kinetics are determined by the time delay caused by the

diffusion of oxygen through increasingly thick oxide layers. Assuming a simple model of spherical

titanium particles which form a growing oxide shell of the same specific volume, the reaction rate is

limited by the diffusion of oxygen through the already reacted layer. Diffusion through the layer with

thickness d = (R0 − R(t)) is dominated by tDd ∆= 6 , where R0 is the original radius before oxidation,

6 Discussion

85

D the diffusion constant of oxygen in TiO2 and ∆t the time needed to deliver the gas to the unreacted Ti

core. Thus, the radius R of the residual unreacted core can be represented by:

,)(

62

0 RRDk

tk

dtdR

−−=

∆−= (1)

where k contains all the reaction coefficients. The solution of this differential equation is simply:

,)18( 3/10 kDtRR −= (2)

The mass increase is

),)((3

42

330 TiTiORRm ρρπ

−−=∆ (3)

where ρ are the densities of the oxide and the metal. This function can be fitted to the measured mass

increase by adjusting kD (see Fig. 4.12). With kD found this way, R/R0 was 0.978 and 0.989 after 180

minutes at 520°C and 480°C, implying that the oxide films on an average particle of 14 µm diameter are

150 µm and 75 µm thick, respectively. These values agree qualitatively with the results of the TEM

investigations (see Fig. 4.23). Obviously the agreement is not perfect but shows that the model of

continuous oxidation is plausible. In reality the particles are angular which could enhance oxidation in

early stages and the oxide layers have a larger volume than the unreacted metal which slows down

diffusion in later stages. Fig. 4.12 also shows an isothermal oxidation experiment at 480°C (only TG). It

can be seen that the oxidation behaviour is very similar while the oxidation rate is smaller. The oxydation

levels of 3.5% and 6% after 180 min at 480°C and 520°C given in Fig. 4.3, respectively, are in perfect

agreement with those given in Fig. 4.12. The oxydation levels measured by Gergely (Fig. 6.6 of [16]),

however, are much lower. A mass increase of just 1% after 150 min at 550°C and of 6% at 650°C was

observed. This could be due to a coarser powder that might have been used.

6 Discussion

86

Oxidation also influences the shape of the decomposition peak observed in the isothermal experiments

(Fig. 4.12). The peak gets much broader as an oxide layer builds up in Fig. 4.12, whereas in the absence

of oxidation the decomposition takes place in a much narrower temperature interval (see Fig. 4.11).

Integration of the MS curves in Fig. 4.11 and 4.12 yields 4.7×10−6 As and 2.3×10−6 As detected ions for

samples pre-treated under argon and air, respectively. Therefore, oxidation effectively reduces

dehydrogenation.

The changes of the decomposition characteristics of TiH2 powders by prior thermal and oxidising pre-

treatments are the main concern of the current work. Fig. 4.13 and 4.14 prove the importance of oxidation

as the annealing treatments were identical in the two cases with the exception of the type of ambient

atmosphere. Whereas pre-treatments under air and Ar both remove the first decomposition peak, only

annealing under air shifts both the onset of gas evolution and the maximum of gas evolution to higher

temperatures. Thermal treatment without oxidation therefore merely reduces the amount of hydrogen and

lowers the decomposition peak (see Fig. 4.13). Oxidation as seen in Fig. 4.14 leads to the formation of

oxide layers around the core of individual titanium hydride particles which then act as diffusion barriers

delaying the evolution of hydrogen. Hydrogen does not only have to diffuse through the bulk titanium

matrix from the inner region of the particle towards the surface but also has to overcome the oxide barrier

which is an additional kinetic hindrance. This effect is already known and has been exploited to control

hydrogen evolution from TiH2 in metal foam manufacture [14,24]. These results allow us to derive that

the beneficial effect of pre-treatment is based on the formation of oxide layers and not on the removal of

the first decomposition peak. Moreover, the results enable to predict "good" pre-treatment parameters for

making foams as it can be tried to optimise the decomposition characteristics with the melting properties

of the respective alloys.

Results obtained by integration of the curves for the decomposition of pre-treated TiH2 powder supplied

from different manufactures are shown in Fig. 4.15. It is observed that after the heat treatment in air the

similar amount of hydrogen is released from hydrides supplied from different companies. 1.47×10−6 As,

1.42×10-6 As and 1.39×10−6 detected ions from Chemetall, SeJong Materials and GFE TiH2 powders,

respectively, are calculated. It can be seen that there is only a minor difference in the amount of released

hydrogen and the onset temperature as well. Obviously the source of the powder does not play a big

difference on the H2 release after pre-treatment in air.

6 Discussion

87

Han et al. find similar results for the peak of hydrogen evolution for pre-treatments for 60 and 120 hours

at 400°C [15] using a degassing apparatus. Lehmhus et al. use DSC to study the influence of various pre-

treatments under air and argon [20]. They find the same basic phenomena, namely a shift of both onset

and peak temperature for pre-treatments under air but the peak positions determined are significantly

lower than the ones shown in Fig. 4.14. For powders pre-treated under argon a weak positive dependence

of peak position on pre-treatment time is found which could not be observed here. Kennedy carried out

DSC studies on TiH2 powders pre-treated in air under non-isothermal conditions [18]. Although his

conditions are somewhat different, the results are in accordance with measurements of this work. Gergely

finds a similar behaviour for pre-oxidised TiH2, namely a shift of both onset and peak temperatures using

DTA and TG [16].

The isothermal experiments at 600°C ("foaming simulations") shown in Fig. 4.16 confirm this

interpretation. Both the onset and peak times are shifted to higher values by pre-treatment in air. Onset

and peak positions found in both isochronal and isothermal experiments are displayed and compared in

Fig. 6.4 (remember that the onset temperature (or time) was defined as the point at which 1 or 2.5% of the

peak height were exceeded, see Chapter 4.2). Analysis of the curves in Fig. 4.14 yields temperatures, that

of Fig. 4.16 yields times. The lower part of Fig. 6.4 corresponds to the range of constant heating for both

the isothermal and isochronal experiments (however, at different heating rates), while in the upper part

the temperature profile for both experiments is different. For this reason isothermal and isochronal

experiments can only be compared qualitatively. The onset of hydrogen evolution is shifted to higher

temperatures (and longer times) with more intense pre-treatment reflecting the fact that oxidation is

increasing more effectively at higher temperatures and longer times. The two curves for the isothermal

and isochronal experiment are relatively similar despite the different heating rate showing that the onset

of gas evolution is not very rate sensitive. The peak positions (upper curve) also move to higher

temperatures but the dependence on pre-treatment temperature is less pronounced.

6 Discussion

88

0 100 200 300 400 500

400

450

500

550

600

600

620

640

660

680

diffe

rent

pre

-trea

tmen

ts a

t 520

°C

T [°

C]

isoc

hron

al

Tpretreatment [°C]

20222426283032

t [m

in]

isot

herm

al

onset

Fig.4.14, 180 min Fig.4.14, 90,360 min

T [°

C]

isoc

hron

al

peak

283032343638404244

t [m

in]

isot

herm

al

Fig.4.16, 180 min Fig.4.16, 90 min Fig.5.1, 180 min Fig.5.1, 90 min

Fig. 6.4. Position of onset (lower) and peak (upper) of gas evolution derived from the results shown in

Fig. 4.14 and Fig. 4.16 and 5.1. Temperatures are given in the former, times in the latter case. Open

symbols correspond to 520°C but have been moved slightly to the right for better visibility.

In Fig. 4.26 it can be seen that the onset temperatures are shifted to higher values by pre-treatment in air,

unlike after the pre-treatment under argon (Fig. 4.25) where the onset temperatures are almost unchanged.

Onset and peak positions found in Fig. 4.26 are presented in Fig.6.5. The onset of hydrogen evolution

from ZrH2 powder is shifted to higher temperatures (and longer times), like observed for TiH2 as well.

The opposite behaviour of the peak maximum for ZrH2 powder after annealing in air is observed. The

peak positions (upper curve) move to lower temperatures.

6 Discussion

89

0 100 200 300 400 500

200

300

400

500

600

700

800peak

onset

peak T (400, 480, 520°C; 180 min) onset T (400, 480, 520°C; 180 min) peak T (520°, 90 min) onset T (520°C, 90 min)T

[°C

]is

ochr

onal

Tpre-treatment [°C]

Fig. 6.5. Position of onset (lower) and peak (upper) of gas evolution for ZrH2 powder derived from the

results shown in Fig. 4.26.

The substantial differences with respect to physical behaviour between these two powders are

investigated by [90,91]. It is clear that in both cases after the pre-treatment in air an oxide layer is formed

on each powder particle. At first, TiO2 and ZrO2 appear to be similar. However, there are substantial

differences with respect to the physical behaviour. By reheating the oxidised powders the onset

temperature of hydrogen release is delayed in both cases. The maximum of the peak of TiH2 after pre-

treatment in air is moved to higher temperatures, but the maximum of the peak of ZrH2 shows an opposite

behaviour. It is moved to the lower temperatures. TEM images (Fig. 4.23) have shown, that TiH2 is

primarily covered with Ti3O followed by a TiO2 coating. The layer is crystallized in the rutile structure.

Rutile is known to be a n-type semiconductor. The structure is tetragonal with an oxygen pair on the basal

plane. It might be stated that the crystallographic structure has no close similarity to the TiH2 structure. In

contrast zirconia is known to crystallize also in the cubic CaF2-structure. The Zr- ion site again occurs in

the fcc arrangement with the oxygen being located on the tetrahedron sites, i.e. the same location as the

protons in the hydride structure. Hence, zirconia is an oxygen ionic conductor. The conductivity needs

6 Discussion

90

anion vacancies for fast transport. For tetragonal zirconia which is a slightly modified derivative of the

fluoride structure proton conduction was observed using ERDA [90]. Summarizing one can state that the

Ti-oxides show a quite different crystalline structure compared to the hydride structure, whereas the

similarity between the zirconia and the hydride lattice would indicate a faster penetration of hydrogen

through the layer. This would be the most plausible explanation for different hydrogen release rates.

0 100 200 300 400 500

100

200

300

400

500

600

ZrH2 powder

Fig.4.26, 180 min Fig.4.26, 90 min

Tpre-treatment [°C]

∆T [K

]

TiH2 powder

Fig.4.14, 180 min Fig.4.14, 90,360 min

Fig. 6.6. Difference between onset and peak temperature ∆T = Tpeak − Tonset is shown for the isochronal

measurement of Fig. 4.14 and Fig. 4.26.

As the variation of the onset temperature from Figs. 6.4 and 6.5 is larger, the difference between peak and

onset temperature drops with increasing pre-treatment temperature as shown in Fig. 6.6 (see data derived

from Figs. 4.14 and 4.26 shown in Fig. 6.6). This is important as this means that the range of strong

hydrogen evolution for TiH2 powder narrows from 211 K to 102 K when replacing untreated by pre-

treated TiH2 powder. For ZrH2 powder the range of strong hydrogen evolution narrows from 547 K to

186 K when replacing untreated by pre-treated ZrH2 powder.

6 Discussion

91

In a technological sense the results shown in Fig. 4.14 and Fig. 4.26 for TiH2 and ZrH2 powder

respectively, is the most important one since it allows us to tailor the decomposition characteristics of the

blowing agent. Higher annealing temperatures reduce the total amount of hydrogen available but shift the

decomposition range to higher temperatures. As one would like to lose as little hydrogen as possible

while shifting the temperatures to the highest values, a compromise has to be sought. It is worth to

mention that the shifting of the peak maximum for ZrH2 powder to the left is exactly in the temperature

range relevant for foaming many aluminium alloys.

In the next step it is necessary to see the effect of pre-treatment of TiH2 powder on the expansion of

aluminium alloy and their cell morphology after solidification. Foam expansion starts at about 590–640°C

depending on the pre-treatment of TiH2 as seen in Fig. 5.9 and even more clearly in Fig. 5.10. These

temperatures should not be taken too literally because there might be a delay between the measured

temperature and the true sample temperature due to the limited heat flow within the expandometer as the

thermocouple is fixed in the middle of the 3 mm thick substrate supporting the sample. It is obvious that

pre-treatments of the blowing agent delay the expansion process. Not only the onset of expansion is

observed at higher temperatures but also a given volume expansion is reached at a later time as shown for

three expansion factors in Fig. 5.10.

The time required for reaching a given foam expansion slightly varies between the two foaming

experiments which is understandable since there are two differences: first, the foaming temperature is 20

K higher in the continuous foaming experiment which leads to faster foaming. Second, however, foaming

was interrupted prior to reaching the desired volume in the interrupted foaming experiments to account

for after-expansion. Therefore, the true foaming time corresponding to the expansion value given would

be about 45s higher. The result is a near compensation of the two effects, leading to the slight difference

observed.

In general, the time delay during foaming is larger for blowing agents subjected to a higher pre-treatment

temperature or pre-treated for a longer time. The shift of the onset of foaming to higher temperatures is

beneficial for pore formation since the fraction of liquid present when gas starts to evolve is higher and

correspondingly the partially molten metallic matrix can accommodate the emerging gas in bubbles. As a

consequence, a difference in morphology between the various foams as demonstrated in Fig. 5.11 for two

different TiH2 powders was observed. As untreated powder leads to early gas evolution already in the

solid state and an earlier onset of expansion the corresponding foam structure (see left column) is less

uniform and the individual pores are jagged and more irregular than for the foam blown with a pre-treated

TiH2 powder (right column). During foaming the alloy melts in both cases and some of the imperfections

6 Discussion

92

created in early stages disappear, but some of the non-uniformities survive even in the fully expanded

foam.

The SEM images confirm this picture. Fig. 5.12 shows irregular pores and cell walls which are partially

punctured or ruptured in the foams made using untreated TiH2. In contrast, foams made using pre-treated

TiH2 exhibit round cells and smooth, non-ruptured and regular cell walls (Fig. 5.13). The observed longer

evacuation time in SEM investigations of foam blown with untreated TiH2 is an indication that many cells

are interconnected by microscopic cracks or holes and that gas continues to flow for a long time when the

sample is placed into a vacuum.

Another effect of TiH2 pre-treatment is that it increases the value of maximum expansion as seen in Fig.

5.9. Untreated TiH2 leads to expansions around 4.5, while pre-treated powders lead to values up to 5.5.

The increase is correlated with the temperature and time of pre-treatment in a sense that higher pre-

treatment temperatures and longer times improve foam expansion within the parameter range

investigated. This result is surprising since the total amount of hydrogen available in the blowing agent is

reduced by a pre-treatment. Isothermal measurements at 600°C revealed that the amount available for

foaming drops to about 50% when TiH2 is pre-treated at 520°C for 180 min, Fig. 4.16. The explanation

for this apparent contradiction is that hydrogen is released from pre-treated blowing agents at higher

temperatures and in a narrower temperature range than from untreated powder, Fig. 6.6. Therefore,

untreated blowing agent releases gas in early stages where it can escape through residual pores and the

gas supply runs out in later stages. Pre-treated blowing agent, in contrast, still provides blowing gas in

late stages and, therefore, leads to a higher final expansion.

A related study was carried out by Kennedy [17]. He subjected TiH2 to heat treatments in air at 400 –

550°C for 15 min and produced pure aluminium foams in free foaming experiments. There was a slight

increase of 13% in maximum foaming height when untreated powder was replaced by powder pre-treated

at 500°C. The time to maximum foaming was also slightly delayed in accordance with our findings.

Fig. 6.7 gives the total amount of hydrogen available in TiH2 powder after isothermal heating as

determined by integration of the individual curves in Figs. 4.13, 4.14 and 4.16 and normalising to the gas

release from the untreated powder. Despite the scatter it can be concluded that around 65-75% of the

hydrogen is still available up to 480°C annealing temperature after which this fraction drops considerably

(one point at 480-180 giving a very high value seems to be an artefact).

6 Discussion

93

Fig. 4.11 allows to determine how much of the available hydrogen is driven out during pre-treatment at

520°C for 90 and 180 minutes, which is 60 and 68%, respectively. The amount of hydrogen therefore

remaining in the hydride, 40 and 32%, has been added to Fig. 6.6 and should correspond to the integrals

obtained from Fig. 4.13. However, the resulting values are lower than expected, as indicated by arrow

"A". Moreover, for the powders pre-treated in the thermobalance under a stream of Ar, only about 10% of

the original amount of hydrogen is found (arrow "B"). The reason for the loss of hydrogen obviously is

that during pre-treatment in the thermobalance (Fig. 4.11) the stream of argon carries away the evolving

hydrogen, thus facilitating decomposition. In contrast, in the experiments in Figs. 4.13 and 4.16 (except

for the curves marked with an asterisk) pre-treatment was carried out in a closed tube furnace. A build up

of a hydrogen partial pressure can be expected here, thus retarding decomposition. A similar although

much smaller effect is found for air. The air stream during pre-treatment in the thermobalance leads to a

reduction of the hydrogen content compared to pre-treatment in resting air (arrow "C"). In any case

precise conditions for pre-treatment have to be defined in order to obtain meaningful results. It is worth to

mention that the same results were calculated by reading the mass loss of the TG curves for the TiH2

powders under treatments like shown in Figs. 4.13 and 4.14.

6 Discussion

94

440-180 460-180 480-180 500-180 520-90 520-180 520-3600

10

20

30

40

50

60

70

80

90

100

data read from

integrated data

Fig.4.13

hydr

ogen

rele

ase

from

pow

ders

(% o

f unt

reat

ed)

pre-treatment temperature [°C] - time [min]

Fig.4.14

Fig.4.16

Fig.5.1

C

B

A

* Fig.4.11

Fig. 6.7. Total amount of hydrogen released from pre-treated TiH2 powders during isochronal or

isothermal heating as a function of pre-treatment parameters. Loose powders and precursors are

considered. All values were either obtained by integration of the curves shown in the figures specified or

by reading a value directly from Fig. 4.11. The asterisk indicates powders pre-treated in the TG furnace

under a stream of Ar or air. A mass current offset of 3×10−13 A was taken into account in Fig. 5.1. Values

for pre-treated powders are normalised to untreated powder and expressed in %. Error bars were derived

from various measurements corresponding to the same parameter set. Arrows compare the hydrogen

contents caused by different heat treatments as discussed in the text.

There is plenty of direct evidence for the existence of oxide layers. Fig. 4.2 shows an EDX analysis

obtained by SEM. As oxygen is a light element it is difficult to measure quantitatively. But qualitative

proof shows that the higher temperature and time of annealing produce thicker oxygen layers on the

surface of the powder particles. From Fig. 4.1 it becomes evident that oxidation of the powder does not

change the powder morphology. The only change easily visible is the colour of the powder (Fig. 4.4). As

thin films build up the optical properties change and the powder goes through a series of colours [19,88].

TEM observation clearly reveals these oxide films (Fig. 4.23).

6 Discussion

95

They are about 100 nm thick after a heat treatment at 480°C for 180 min and appear to be dense.

Comparable thicknesses were also found by elipsometry [92] yielding values between 70 and 170 nm for

the oxidation stages corresponding to the results shown in Fig. 4.4. Electron microdiffraction yields

various values for interplanar spacings given in Table 4.2. This can be explained by assuming that the

layer consists of two different oxides, namely TiO2 (rutile) and Ti3O. This finding is in accordance with

Ref. [18]. Qualitative EDX profiles confirm that the oxygen content drops to about half the value at the

outer surface of the oxidised TiH2 particle when crossing the oxide layer. Thus it can be concluded that

rutile forms an outer shell and Ti3O is located between the rutile layer and the TiH2 core. The XRD

measurements confirm this picture (Fig. 4.17). The change of structure associated with hydrogen loss and

the formation of the two types of oxides is observed, again in agreement with Ref. [18]. In the literature

further transient oxides such as TiO, Ti2O3, Ti3O5 [88, 93], Ti2O, Ti6O [94] and oxihydrides TiOxHy [95]

have been mentioned. None of these have been observed in these experiments.

TiH2 particles embedded in an aluminium powder compact also show an oxide layer which is well

discernible from the aluminium matrix and the TiH2 core (see Fig. 5.2). In this case a division into two

sub-layers is visible. The magnification on the left side shows that these layers are very uniform, appear

dense and are about 100 nm thick. The magnified view on the right side shows that the oxide layers could

be distorted in some areas. Coming from the lower right in this image one observes a double oxide layer

structure which then ends in an area where the oxides could have piled up during the mechanical

deformations associated with pressing. Such distorted areas could facilitate the escape of hydrogen from a

TiH2 particle.

Gas release from the precursor as given in Fig. 5.1 follows the same rules as from the loose powder. Pre-

treatments lead to a delay of hydrogen evolution. Some of the hydrogen evolving during foaming is

caught in the emerging pores but as samples are so small most of it reaches the surface of the sample and

can eventually be detected by the mass spectrometer. In Fig. 6.4 the positions of the onset and maximum

of hydrogen release from the TiH2 powder embedded in the precursor can be compared to the analogous

data for the loose powder. Gas evolution from the precursor peaks earlier than for the loose powder

although one might have expected the opposite arguing, e.g., that the entrapped hydrogen is held back by

the metallic matrix or that decomposition of TiH2 is delayed by the hydrogen surrounding the blowing

agent in the metal. The effect of earlier hydrogen release has already been observed previously [11,20]

and was explained by a possible partial fracture of the oxide layers on TiH2 particles during compaction

[11].

6 Discussion

96

The evidence given by one of the TEM images discussed above, showing such a fractured oxide layer,

could support this hypothesis but a definite proof is still lacking. Fig. 6.7 shows that the amount of

hydrogen available in the precursors roughly equals the amount in loose powders.

Pre-treated ZrH2 powder delays foaming and increases temperatures at which the foam evolves by up to

30 K for AlSi6Cu4 alloy. Precursors with HfH2 powder are foamed 25 sec earlier on slightly lower

temperatures than with ZrH2 powder. Foaming the selected aluminium alloys with pre-treated TiH2

powder the foaming is delayed while temperatures at which the foam evolves increase up to 35 K, 40 K,

45 K and 55 K for Al, AlSi6, AlSi6Cu4 and AlSi6Cu10, respectively as seen in Fig. 5.4. This is exactly

the effect sought: That the gas evolution at higher temperatures is expected to be beneficial because the

formation of cracks [10, 12] in the solid state is avoided and rounder pores are blown in the liquid or

semi-solid state. Additionally, the higher copper content provides a larger amount of liquid fraction which

leads to the formation of spherical pores as well. The images shown in Figs. 5.5, 5.7 and 5.8 underline

these explanations. Foams blown with pre-treated powders do have rounder and less jagged pores.

Uniformity of cell size is also improved with longer pre-treatment times as shown in Figs. 5.5, 5.7 and

5.8. The differences shown in these figures apply to a fairly early foaming stage, namely an expansion by

a factor of less than 2.5. Fully expanded foams with TiH2 can have five times the original precursor

volume. Such foams were also produced and it was found that some differences are still visible but they

are less important than the differences in the early stage. Apparently there is some self-healing effects

associated with the metal flow during foaming reducing some of the flaws. From a technological point of

view, however, even small improvements of foam quality are important since they are expected to

improve foam properties and process stability. Therefore, gaining control over the release of hydrogen

from the blowing agent powders is an important issue and leads to improved cell structures and as a

consequence better foam quality.

7 Summary

97

7 Summary

There are many factors influencing aluminium foam evolution. A key point is the use of an appropriate

blowing agent which releases gas at the right temperature to ensure high expansion and the formation of a

uniform porosity. Several investigations have been carried out to evaluate the behaviours of blowing

agent during foaming:

• Onset temperature of hydrogen release from titanium hydride is in the range which is necessary for

the production of the metal foamed parts.

• Untreated TiH2 powder supplied by different manufactures release hydrogen on various temperatures.

After heat treatment in air the temperature difference is decreased and there is no difference in the

amount of released hydrogen from untreated or pre-treated TiH2 powder in air.

• If heating of TiH2 powder is carried out under air an oxide layer grows which is roughly 100 nm thick

after 180 minutes at 480°C and contains an outer shell of TiO2 and an inner shell of Ti3O.

• A clear three - stage decomposition is only observed for untreated TiH2 and ZrH2 powder.

• Pre-treatments under argon reduce the amount of hydrogen stored in the hydride but do not

significantly change the onset and peak temperatures.

• Pre-treatments of TiH2 powder under air reduce the amount of hydrogen and shift both onset and peak

of decomposition by higher temperatures (by +170 K and +52 K, respectively). The temperature range

of decomposition is compressed by pre-treatments.

• Pre-treatments of ZrH2 powder under air reduce the amount of hydrogen and shift onset of

decomposition to higher, the peak to lower temperatures (by +258 K and -103 K, respectively). The

temperature range of decomposition is compressed by pre-treatments.

• Gas flow during pre-treatment under argon and air as opposed to pre-treatment under a resting

atmosphere does not change the position of decomposition onset and the decomposition peaks but

decreases the amount of hydrogen finally stored in the powders.

• TiH2 decomposition is shifted to lower values if it is contained in densified precursors.

By pre-treating TiH2 powder in air at a certain temperature and for a given expandometer experiments

showed:

• The onset of foaming is delayed by pre-treatment. In the current configuration the delay can be up to

45s, corresponding to a temperature difference of 45K.

• Final expansion of the foam can have five times the original precursor volume.

• Uniformity of the emerging foam is improved and the pores are rounder, the cell walls smoother.

7 Summary

98

• Using pre-treated TiH2 for foaming Al, AlSi6, AlSi6Cu4 and AlSi6Cu10 alloys and ZrH2 for foaming

AlSi6Cu4 alloy delays foaming and leads to a more uniform distribution of rounder pores. The best

parameters with both powders are found at 480°C or 520°C for 180 minutes.

• The results given here are of a general validity a fine-tuning of process parameters is recommended

for anyone attempting to set up industrial production of Al foams.

• Gaining control over the release of hydrogen from the blowing agent powders is an important issue

and leads to improved cell structures and, therefore, better foam quality.

8 References

99

8 References

[1] M.F. Ashby, A. Evans, N.A. Fleck, L.J. Gibson, J.W. Hutchinson, and H.N.G. Wadley. Metal

foams — a design guide. Butterworth-Heinemann, London, 2000.

[2] J. Banhart. Functional applications. In: Handbook of Cellular Metals, Editors: H.-P. Degischer, B.

Kriszt, 313-320, Wiley-VCH, Weinheim, 2002.

[3] J. Banhart. Aluminium foams for lighter vehicles. International Journal of Vehicle Design,

37:114–125, 2005.

[4] H.-W. Seeliger. New aspects of getting aluminium-foam sandwich (AFS) parts into volume

production. In J. Banhart, M.F. Ashby, and N.A. Fleck, editors, Cellular Metals and Metal

Foaming Technology, pages 5–16. MIT-Verlag Berlin, Germany, 2001.

[5] T. Daxner, H.J. Böhm, M. Seitzberger, and F.G. Rammerstorfer. Modelling of cellular metals. In

H.-P. Degischer and B. Kriszt, editors, Handbook of Cellular Metals, pages 245–280. Wiley-

VCH, Weinheim, Germany, 2002.

[6] H.P. Degischer and B. Kriszt. Handbook of Cellular Metals. Wiley-VCH, Weinheim, 2002.

[7] J. Banhart. Manufacturing routes for metal foams. Journal of Metals 52, 22-27, 2000.

[8] J. Banhart. Manufacture, characterisation and application of cellular metals and metallic foams.

Progress in Materials Science, 46:559–632, 2001.

[9] F. Baumgärtner, I. Duarte, and J. Banhart. Industrialisation of P/M foaming process. Advanced

Engineering Materials, 2:168–174, 2000.

[10] M. Weber. Herstellung von Metallschäumen und Beschreibung der Werkstoffeigenschaften. PhD

thesis, Universität Clausthal, Clausthal-Zellerfeld, Germany, 1995.

[11] F. von Zeppelin, M. Hirscher, H. Stanzick, and J. Banhart. Desorption of hydrogen from blowing

agents used for foaming metals. Composite Science and Technology, 63:2293–2300, 2003.

8 References

100

[12] L. Helfen, T. Baumbach, H. Stanzick, J. Banhart, A. Elmoutaoaukkil, and P. Cloetens. Viewing

the early stage of metal foam formation by computed tomography using synchrotron radiation.

Advanced Engineering Materials, 4:808–813, 2002.

[13] HM. Helwig. Untersuchung der Einflüsse von Legierungszusammensetzung und

Temperaturführung auf das Aufschäumverhalten von Aluminiumschäumen. PhD thesis,

Technische Universität Berlin, in preparation.

[14] S.E. Speed. Foaming of metal by the catalysed and controlled decomposition of zirconium

hydride and titanium hydride. US Patent 3,981,720, 1976.

[15] F. Han, Z. Zhu, J. Gao, and W. Song. Effect of oxidation treatment and surface filming on

hydrogen degassing from TiH2. Metallurgical Transactions B, 29:1315–1319, 1998.

[16] V. Gergely. Metal route processing for production of metallic foams. PhD thesis, University of

Cambridge, Cambridge, UK, 2000.

[17] A.R. Kennedy. The effect of TiH2 heat treatment on gas release and foaming in Al–TiH2 preforms.

Scripta Materialia, 47:763–767, 2002.

[18] A.R. Kennedy and V.H. Lopez. The decomposition behaviour of as-received and oxidised TiH2

foaming agent powder. Materials Science and Engineering, A357:258–263, 2003.

[19] B. Matijasevic, S. Fiechter, I. Zizak, O. Görke, N. Wanderka, P. Schubert-Bischoff, and J.

Banhart. Decomposition behaviour of as-received and oxidised TiH2 powder. In H. Danninger and

R. Ratzi, editors, Proceedings PM 2004 Powder Metallurgy World Congress Vienna, volume 4,

pages 149–155. European Powder Metallurgy Association, Shrewsbury, UK, 2004.

[20] D. Lehmhus and G. Rausch. Tailoring titanium hydride decomposition kinetics by annealing in

various atmospheres. Advanced Engineering Materials, 6:313–330, 2004.

[21] B. Matijasevic, J. Banhart, S. Fiechter, O. Görke, N. Wanderka. Modification of titanium hydride

for improved aluminium foam manufacture. Acta Materialia, accepted for publication, 2005.

8 References

101

[22] J.K.W. Sandler, V. Altstädt. Physics of polymer foams. In: R.F. Singer, C. Körner, V. Altstädt, H.

Münstedt, editors. Cellular Metals and Polymers 2004. Uetikon-Zurich (CH): Transtech

Publications, p. 195-200, 2005.

[23] J. Banhart. Foam Metal: the recipe. Europhys News, 30:17-20, 1999.

[24] V. Gergely, TW. Clyne. The FORMGRIP process: foaming of reinforced metals by gas release in

precursors. Advanced Engineering Materials, 2:168-74, 2000.

[25] J. Banhart. Metal Foams: the mystery of stabilisation. Proceedings MetFoam 2005 World Congress,

Kyoto, 21-23 September 2005.

[26] B. Sosnick. US Patent 2, 434, 775, filed 1943.

[27] BC. Allen, MW. Mote, AM. Sabroff. US Patent 3, 087, 807, 1963.

[28] J. Baumeister. Verfahren zur Hehrstellung poröser Metallkörper. German Patent 4, 018, 360,

1990.

[29] A. Irretier, J. Banhart. Lead and lead alloy foams. Acta Materialia, 53, 4903-4917, 2005.

[30] V. Gergely, D.C. Curran, T.W. Clyne, The FOAMCARP process: foaming of aluminium MMCs

by the chalk-aluminium reaction in precursors. Composites Science and Technology 63, 2301 –

2310, 2003.

[31] A.R. Kennedy, Effect of compaction density on foamability of Al-TiH2 powder compact. Powder

metallurgy 45,75,2002.

[32] N. Babcsán, J. Banhart. Metal foams towards high-temperature colloid chemistry. In: B.P. Binks

and T.S. Horozov, editors. Colloidal Particles at Liquid Interfaces. Cambridge University Press

2006, in press.

[33] I. Duarte and J. Banhart. A study of aluminium foam formation-kinetics and microstructure. Acta

Materialia, 48, 2349-2362, 2000.

8 References

102

[34] Trademarks by Alulight International, Ranshofen and Fraunhofer-Gesellschaft, München.

[35] J. Banhart, J. Baumeister. Deformation characteristics of metal foams. Journal of Materials

Science, 33:1431-40, 1998.

[36] D. Lehmhus, J. Banhart. Properties of heat treated aluminium foams. Materials Science and

Engineering A, 349:98–110. 2003.

[37] P. Kenesei, Cs. Kadar, Zs. Rajkovits, J. Lendvai. The influence of cell-size distribution on the

plastic deformation in metal foams. Scripta Materialia; 50: 295–300, 2004.

[38] TG. Nieh, K. Higashi, J. Wadsworth. Effect of cell morphology on the compressive properties of

open-cell aluminium foams. Materials Science and Engineering A; 283: 105–10, 2000.

[39] F. Han, H. Cheng, J. Wang, Q. Wang. Effect of pore combination on the mechanical properties of

an open cell aluminium foam. Scripta Materialia, 50:13–7, 2004.

[40] HX. Zhu, AH. Windle. Effects of cell irregularity on the high strain compression of open-cell

foams. Acta Materialia, 50:1041–52, 2002.

[41] JF. Rakow, AM. Waas. Size effects and the sheer response of aluminium foam. Mech Mater,

37:69–82, 2004.

[42] J. Grenestedt. Influence of wavy imperfections in cell walls on elastic stiffness of cellular solids. J

Mech Phys Solids, 46:29–50, 1998.

[43] T. Daxner, HJ. Böhm, M. Seitzberger, FG. Rammerstorfer. Modelling of cellular metals. In: H-P.

Degischer, B. Kristz, editors. Handbook of cellular metals. Weinheim, Germany: Wiley-VCH; p.

245–80, 2002.

[44] U. Ramamurty, A. Paul. Variability in mechanical properties of a metal foam. Acta Materialia,

52:869–76, 2004.

8 References

103

[45] D. Leitlmeier, H-P. Degischer, HJ. Flankl. Development of a foaming process for particulate-

reinforced aluminium metals. Advanced Engineering Materials, 4:735–40, 2002.

[46] H-W. Seeliger. Aluminium foam sandwich (AFS)—ready for market introduction. In: J. Banhart,

NA. Fleck, A. Mortensen, editors. Cellular metals—manufacture, properties, applications.

Germany: MIT-Verlag Berlin; p. 5–12, 2003.

[47] A. Rack. Charakterisierung komplexer Materialsysteme mittels Synchrotron-Tomographie und

3D-Bildanalyse. PhD Thesis, TU Berlin in preparation.

[48] H-D. Kunze, J. Baumeister, J. Banhart, Weber M. P/M technology for the production of metal

foams. Powder metallurgy international, 25:182–5, 1993.

[49] H-W. Seeliger. private communication.

[50] F. Garcia-Moreno, N. Babcsan, J. Banhart. X-ray radioscopy of liquid metal foams: influence of

heating profile, atmosphere and pressure. Colloids Surf A: Physicochem Aspects, 263:290–4,

2005.

[51] H. Stanzick, M. Wichmann, J. Weise, J. Banhart, L. Helfen, T. Baumbach. Process control in

aluminium foam production using real-time X-ray radioscopy. Advanced Engineering Materials,

4:814–23, 2002.

[52] Fraunhofer IFAM Bremen J. Banhart, unpublished work.

[53] S.H. Hiller, HM. Helwig, J. Banhart. Influence of compaction parameters on foaming behaviour of

a Al alloy, in preparation.

[54] S.H. Hiller. Untersuchung des Einflusses verschiedener Prozessparamenter auf die Schäumbarkeit

von Aluminium. Studien Arbeit, Technische Universität Berlin, in preparation.

[55] J. P. Blackledge, An Introduction to the Nature and Technology of Hydrides, Metal Hydrides, W.

M. Müller, J. P. Blackledge, and G. G. Libowitz (eds.) Academic Press, New York, 1-20, 1968.

8 References

104

[56] A. J. Maeland. Survey of the Different Types of Hydrides. Hydrides for energy storage, Kjeller,

Norway, A.F. Andresen and A. J. Maeland (eds) Pergamon Press, Oxfore, UK, 19-32, 1977.

[57] R. J. Wasilewski, G.L. Kehl, Metallurgia 1954, 49, 225.

[58] C.W. Schoenfelder, J.H. Swisher, Kinetics of thermal decomposition of TiH2. J.Vac.Sci.Technol.

10,862,1973.

[59] V.A. Lavrenko, V.Z. Shemet, S.K. Dolukhanyan, A.M. Kalinichenko, I.N. Frantsevich, Doklady

Akadmii Nauk SSR, 262,136, 1982.

[60] V.Z. Shemet, A.P. Pomytkin, V.A. Lavrenko, V.Z. Ratushnaya, presented at the World Hydrogen

Energy Conference Hawaii, 1990, July 22nd-27th.

[61] M. J. Trzeciak, D. F. Dilthey and M. W. Mallett, Study of Hydrides, USAEC Report BMI-1112,

Battelle Memorial Insitute, July 1956.

[62] A. San Martin and F.D. Manchester. The H-Ti (hydrogen-titanium) system. Bulletin of Alloy

Phase Diagrams, 8:30–42, 1987.

[63] J. Banhart, B. Matijasevic and C. Ritter. Unpublished results.

[64] V.I. Trefilov, I.I. Timofeeva, L.I. Klochkov, I.A. Morozov and R.A. Morozova. Effectc of

temperature change and hydrogen content on titanium hydride crystal lattice volume. Int. J.

Hydrogen Energy, Vol.21, No. 11/12, pp.1101-1103, 1996.

[65] Z.M. Azarkh and P.I. Gavrilov. Structural Changes in Titanium Hydride at Large Hydrogen

Concentrations, Kristallografiya, 15, 275-279, 1970.

[66] H.D. Bale and S.B. Peterson. X-Ray Diffraction Study of the Structural Transformations on TiH2,

Solid State Commun, 11, 1143-1145, 1972.

[67] P.E. Irving and C.J. Beevers. Some Metallographic and Lattice Parameter Observations on

Titanium Hydride, Metall. Trans, 2, 613-615, 1971.

8 References

105

[68] R.L. Crane, S.C. Chattoray, and M.B. Strope. A room-temperature polymorphic transition of

titanium hydride. Journal of Less-common Metals, 25:225–227, 1971.

[69] H.L. Yakel, Jr. Thermocrystallography of Higher Hydrides of Titanium and Zirconium, Acta

Cristallogr. 11, 46-51, 1958.

[70] A. Takasaki, Y. Furuya, K. Ojima, Y. Teneda. Hydride dissociation and hydrogen evolution

behaviour of electrochemically charged pure titanium. Journal pf Alloy and Compounds 224, 269-

273, 1995.

[71] R. Kresse. Characteristics and handling of titanium hydride. In J. Banhart, M.F. Ashby, and N.A.

Fleck, editors, Metal Foams and Porous Metal Structures, pages 109–110. MIT-Verlag Berlin,

Germany, 1999.

[72] J. Baumeister and H. Schrader, US Pat. 5 151 246, 1992.

[73] H. H. Hausner. Handbook of Powder Metallurgy, Chemical Publishing. Co. Inc. New York

(1973), p.68.

[74] A. Bütow. Strukturuntersuchungen an Metallschäumen in verschiedenen Entwicklungsstadien.

Master’s thesis, TU Berlin, 2004.

[75] B. Matijasevic-Lux, A. Rack, A. Haibel, J. Banhart . Influence of powder pre-treatments on metal

foam pore structure. In: Singer RF, Körner C, Altstädt V, Münstedt H, editors. Cellular metals and

polymers. Uetikon-Zurich (CH): Transtech Publications; 2005. p. 57.

[76] J. Banhart, D. Bellmann, and H. Clemens. Investigation of metal foam formation by microscopy

and ultra small-angle neutron scattering. Acta Materialia, 49:3409–3420, 2001.

[77] W. Gruner. Determination of oxygen in oxides by carrier gas hot extraction analysis with

simultaneous COX detection. Journal of Analytical Chemistry, 365: 597-603, 1999.

[78] R. German, Powder Metallurgy Science: Powder Characterization, MPIF, Princenton, NJ, USA,

1994.

8 References

106

[79] M. Mitkov, D. Božić, Z. Vujović, Metalurgija Praha, publisher BMG Beograd 1998.

[80] Expandometer provided by the Slovak Academy of Sciences, Bratislava.

[81] A. Albou. Comparison between Aluminium foams issued by different powder metallurgical

processes. Master’s thesis, University of Toulouse, 2005.

[82] V.Zh. Shemet, A.P. Pomytkin, V.A. Lavrenko and V.Zh. Ratushnaya. Decomposition of metal

hydrides in low temperatures and in high-temperature oxidation. Int. J. Hydrogen Energy, 18, 511-

516, 1993.

[83] Fr.-W. Bach, D. Bormann, P. Wilk, R. Kucharski. Detachable fasteners for aluminium foams. In

In: Singer RF, Körner C, Altstädt V, Münstedt H, editors. Cellular metals and polymers. Uetikon-

Zurich (CH): Transtech Publications; p. 181-184, 2005.

[84] V.A. Lavrenko, V.Zh. Shemet, S.K. Dolukhanyan. Dokl. Akad.Nauk SSSR, 1982, 262(1), 136.

[85] V.I. Trefilov, V.A. Lavrenko, G.G. Momot, et.al. Dokl. Akad.Nauk SSSR, 1987, 293(2), 404.

[86] V.I. Trefilov, V.A. Lavrenko, G. G. Momot, et.al. Dokl. Akad.Nauk SSSR, 1987, 295(2), 431.

[87] Ch.D. Goldshmidt, Interstitial Alloys (in Rus.), M. Mir Press (1971).

[88] V.A. Lavrenko, V.Z. Shemet, L.A. Petrov, O.A. Teplov, and S.K. Dolukhanyan. High temperature

oxidation of titanium hydride powders. Oxidation of Metals, 33:177–189, 1990.

[89] Landolt-Börnstein. Diffusion in Solid Metals and Alloys, Group III: Crystal and Solid State

Physics, volume 26. Editor: H. Mehrer, Springer-Verlag, Berlin, 1990.

[90] O. Kruse, H.D. Carstanjen, P.W. Kountouros, H. Schubert and G. Petzow. Characterization of

H20-Aged TZP by Elastic Recoil Detection Analysis (ERDA). Science and Technology of

Zirconia V. Edit by S.P.S. Badwal, M.J. Bannister, R.H.J, Hannink, Technomic Publishing

Company, Inc., Lancaster, Pennsylvania, USA, 1993.

8 References

107

[91] H. Schubert, P.W. Kountouros, J. Pross. G. Petzow and A.O. Boschi. Sintering of Large TZP

Components. Science and Technology of Zirconia V. Edit by S.P.S. Badwal, M.J. Bannister,

R.H.J, Hannink, Technomic Publishing Company, Inc., Lancaster, Pennsylvania, USA, 1993.

[92] T. Fukuzuka, K. Shimogori, H. Satoh, and F. Kamikubo. On the beneficial effect of the titanium

oxide film formed by thermal oxidation. In H. Kimura and O. Izumi, editors, Titanium 80, pages

2783–2792. AIMM and Petroleum Engineers, 1980.

[93] O. Kubaschevski and C.B. Alcock, Metallurgical Thermo-Chemistry Fifth Edition, Pergamon

Press, 1983.

[94] V.N. Fokin, Y.I. Malov, E.E. Fokina, S.L. Troitskaya, and S.P. Shilkin. Investigation of

interactions in the TiH2-O2 system. International Journal of Hydrogen Energy, 20:387–389,

1995.

[95] A.R. Gromov, N.N. Kouznetsova, S.L. Yudina, and V.V. Lunin. The investigation of titanium

hydride oxidation process. Journal of Alloys and Compounds, 261:269–272, 1997.

9 Acknowledgement

108

9 Acknowledgement

I would like to thank gratefully those who contributed to this work, especially:

Prof. Dr. J. Banhart for the opportunity to succeed this PhD work under his guidance and for his extended

support and help!

Prof. Dr. Helmuth Schubert for being the reviewer part, for creative discussions and possibility to work in

his group. Therefore, I would like to use this opportunity to say thank you to the complete staff of

Ceramics Institute at TU Berlin for scientific and technical support, and for a very nice working

atmosphere.

Special thanks go to Dr. Ing. Oliver Görke for constructive ideas, discussions and support during my

experimental work, writing and proofreading of the thesis.

I am grateful to Dr. I. Zizak for the constructive discussions and support during in-situ X-ray diffraction

experiments, Dr. S. Fiechter for support during thermo-analytical experiments, Dr. N. Wanderka for TEM

investigations and Urlich Gernert of Zelmi at TU Berlin. Help by Dr. W. Seeliger is gratefully

acknowledged.

Furthermore I would like to thank P. Schubert-Bischoff, H. Kropf, C. Förster, W. Rönfeld, C. Leistner, J.

Bajorat, Y. Herzog and H. Krzyzagorski for the technical support.

I would like to say thank you to all colleagues of the SF 3 Department in Hahn-Meitner Institute!

At the end I would like to thank my family for being always there for me!

This work was financially supported by Deutsche Forschungsgemeinschaft (DFG), grant Ba1170/4-2.

Documents

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