An investigation of metallic glass as binder phase in hard metal $(function(){PrimeFaces.cw("Tooltip","widget_formSmash_items_resultList_12_j_idt799_0_j_idt801",{id:"formSmash:items:resultList:12:j_idt799:0:j_idt801",widgetVar:"widget_formSmash_items_resultList_12_j_idt799_0_j_idt801",showEffect:"fade",hideEffect:"fade",target:"formSmash:items:resultList:12:j_idt799:0:fullText"});});

An investigation of metallic glass as binder phase in hard metal $(function(){PrimeFaces.cw("Tooltip","widget_formSmash_items_resultList_12_j_idt799_0_j_idt801",{id:"formSmash:items:resultList:12:j_idt799:0:j_idt801",widgetVar:"widget_formSmash_items_resultList_12_j_idt799_0_j_idt801",showEffect:"fade",hideEffect:"fade",target:"formSmash:items:resultList:12:j_idt799:0:fullText"});});
Department of Physics, Chemistry and Biology
Master’s Thesis
An investigation of metallic glass as binder phase in
hard metal
Malin Leijon Lind
LiTH-IFM-A-EX–15/2978–SE
Department of Physics, Chemistry and Biology
Linköpings universitet, SE-581 83 Linköping, Sweden
Master’s Thesis
LiTH-IFM-A-EX–15/2978–SE
An investigation of metallic glass as binder phase in
hard metal
Malin Leijon Lind
Adviser:
Per Eklund
IFM
Erik Holmström
Sandvik Coromant
Examiner:
Björn Alling
IFM
Linköping, 16 January, 2015
Avdelning, Institution
Division, Department
Datum
Date
Thin Film Physics Division
Department of Physics, Chemistry and Biology
Linköpings universitet, SE-581 83 Linköping, Sweden
2015-01-16
Språk
Language
Rapporttyp
Report category
ISBN
Svenska/Swedish
Licentiatavhandling
ISRN
Engelska/English
Examensarbete
C-uppsats
D-uppsats
Övrig rapport
—
LiTH-IFM-A-EX–15/2978–SE
Serietitel och serienummer ISSN
Title of series, numbering
—
URL för elektronisk version
Titel
Title
En studie om metalliskt glas som bindefas i hårdmetall
An investigation of metallic glass as binder phase in hard metal
Författare Malin Leijon Lind
Author
Sammanfattning
Abstract
In this study, the possibilities to produce metallic glass as binder phase in hard
metal by means of powder metallurgical methods have been investigated. The aim
of the study was to do an initial investigation about metallic glass as alternative
binder phase to cobalt in hard metal. Production of samples with metallic glass
forming alloys and an amorphous powder as binder phase in hard metal by means
of quenching and hot pressing have been performed. Moreover, mechanical alloying
of metallic glass forming powder to achieve amorphicity has been performed.
The samples and powders were analyzed by means of XRD, LOM, STA, SEM
and EDS. The results showed that no glass formation of the binder phase was
achieved by quenching, hot pressing or mechanical alloying. However, interesting information about glass formation by means of metallurgical methods was
obtained.
The main conclusion was that production of metallic glass by means of metallurgical methods is complicated due to changes in the binder phase composition
throughout the production process as well as requirements of high cooling rates
when quenching and high pressures when hot pressing.
Nyckelord Alternative binder, bulk metallic glass, cemented carbide, metallic glass matrix,
Keywords
tungsten carbide
Abstract
In this study, the possibilities to produce metallic glass as binder phase in hard
metal by means of powder metallurgical methods have been investigated. The aim
of the study was to do an initial investigation about metallic glass as alternative
binder phase to cobalt in hard metal. Production of samples with metallic glass
forming alloys and an amorphous powder as binder phase in hard metal by means
of quenching and hot pressing have been performed. Moreover, mechanical alloying
of metallic glass forming powder to achieve amorphicity has been performed.
The samples and powders were analyzed by means of XRD, LOM, STA, SEM
and EDS. The results showed that no glass formation of the binder phase was
achieved by quenching, hot pressing or mechanical alloying. However, interesting information about glass formation by means of metallurgical methods was
obtained.
The main conclusion was that production of metallic glass by means of metallurgical methods is complicated due to changes in the binder phase composition
throughout the production process as well as requirements of high cooling rates
when quenching and high pressures when hot pressing.
v
Acknowledgements
First of all, I would like to thank my supervisors at Sandvik; Erik Holmström,
for the time and effort you put into this project, and Susanne Norgren, for your
expertise and the possibilities you create.
Thank you to all the people at Sandvik, who have helped me with my strange
samples and experiments.
Finally I would like to thank my supervisor Per Eklund and examiner Björn
Alling at Linköping University for the ease in collaboration and for helping me to
improve my work and report further.
This work was performed within the framework of VINN Excellence Center
Hero -m.
vii
Contents
1 Introduction
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 Aims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1
2
2 Theoretical background and literature review
2.1 Cemented carbide . . . . . . . . . . . . . . . . .
2.2 Metallic glass . . . . . . . . . . . . . . . . . . .
2.2.1 Background . . . . . . . . . . . . . . . .
2.2.2 Thermodynamics and kinetics . . . . . .
2.2.3 Glass transition temperature . . . . . .
2.2.4 Glass forming ability . . . . . . . . . . .
2.2.5 Mechanical properties of metallic glass .
2.2.6 Metallic glass fabrication methods . . .
2.2.7 Effect of reinforcements in metallic glass
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
matrix composites
3
3
3
3
4
7
8
10
12
13
3 Experimental methods
3.1 Production methods . . . . . . . . . . . . .
3.1.1 Milling . . . . . . . . . . . . . . . . .
3.1.2 Drying . . . . . . . . . . . . . . . . .
3.1.3 Pressing . . . . . . . . . . . . . . . .
3.1.4 Sintering . . . . . . . . . . . . . . .
3.1.5 Quenching . . . . . . . . . . . . . . .
3.1.6 Arc melting . . . . . . . . . . . . . .
3.2 Computational thermodynamics . . . . . .
3.3 Characterization methods . . . . . . . . . .
3.3.1 Vickers hardness test . . . . . . . . .
3.3.2 Fracture toughness . . . . . . . . . .
3.3.3 Simultaneous thermal analysis . . .
3.3.4 Carrier gas hot extraction . . . . . .
3.3.5 Light optical microscopy . . . . . . .
3.3.6 X-ray diffraction . . . . . . . . . . .
3.3.7 Scanning electron microscopy . . . .
3.3.8 Energy dispersive x-ray spectroscopy
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
15
15
15
16
16
16
18
18
18
19
19
19
19
20
20
20
21
21
ix
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
x
4 Results
4.1 Glass
4.1.1
4.1.2
4.1.3
4.1.4
4.2 Glass
4.2.1
4.2.2
Contents
formation by rapid solidification . .
Binder alloy 1 . . . . . . . . . . . .
Binder alloy 2 . . . . . . . . . . . .
Binder alloy 3 . . . . . . . . . . . .
Binder alloy 4 . . . . . . . . . . . .
formation by solid state processing .
Mechanical alloying . . . . . . . .
Hot pressing . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
23
23
24
32
32
35
38
38
39
5 Conclusions
43
6 Future work
45
Bibliography
47
A Thermocalc
53
B Tables
59
C XRD
63
D STA
71
E EDS
73
F LOM
79
Chapter 1
Introduction
1.1
Background
Cemented carbide
Cemented carbide is a successful composite engineering material used in for instance metal cutting tools. It is a versatile material with properties such as high
hardness, strength and toughness, which can easily be tuned to fit different applications by means of precisely controlled powder metallurgical methods. As a
leading company in tools and tooling systems for metal cutting and mining, many
of Sandvik’s products are made of cemented carbide, with tungsten carbide (WC)
and cobalt (Co) as the main constituent elements. By the years, hazardous effects of cobalt to both human and environment have been reported [1] [2] [3], and
with the new findings of cobalt metal as carcinogenic by NTP (National toxicology
program) [4], it is urgent to find a replacement.
Much effort in finding alternative binder phases is ongoing. However, the excellent properties of cobalt, such as good wettability to WC and a rare combination of
high hardness and high toughness [5], have been hard to recreate. A binder phase
of iron-manganese (Fe-Mn) have been considered, due to its similar characteristics
to Co regarding melting temperature and crystal structure. However, as many
alternative binders that have been investigated, it has low wettability to WC, and
moreover its toughness is too low [6]. Nickel-based binders have good wettability
and high toughness, however, their hardness is low and the production process is
complicated by high sintering temperatures [7]. It is clear that the work of finding
a replacement for cobalt is problematic, and since conventional alternative binders
all seem to have their problems, new concepts need to be investigated.
Metallic glass
Since the first discovery of metallic glass formation in Au-Si in 1960 [8], metallic
glass have been widely investigated. Their unique properties, such as high hardness
and good corrosion resistance [9], make them interesting materials. To date, bulk
metallic glass (BMG) is used in various applications such as sport goods, medical
1
2
Introduction
devices and optics [10], but due to the challenge with producing large pieces, their
usability is still limited.
The limitation in size of the produced metallic glass pieces comes from the
need of a high cooling rate when producing BMGs. More heat has to be removed
during the cooling stage for larger samples and this lowers the cooling rate. The
achieved cooling rate is hence determined by an interplay between the volume,
heat capacity (Cv ) and thermal conductivity (κ). In research, cooling methods are
based on cooling from the outside of the sample, most commonly by casting. In
casting, a melt is cast on a cooled copper plate or in a cooled copper crucible. The
large contact area of the liquid to the copper plate, and the high heat conductivity
of copper means that a large cooling rate is achieved. However, the thermal
conductivity of metallic glass is reported to be lower than that of its crystalline
counterpart [11][12], and moreover tend to decrease with decreasing temperature
[13]. The possibility of extra cooling effect of reinforcing WC grains, that results
in a much higher heat conductivity of the composite than conventional metallic
glasses, in BMGs has not been considered. Thus, the glass forming ability (GFA)
of BMGs with a high fraction of WC-reinforcement is unknown.
It has been shown that many properties of BMGs improve when being reinforced with particles, such as WC, silicon carbide (SiC) and zirconium carbide
(ZrC), however, in research, a maximum of 50 Vol% of reinforcement have been
used, due to limitations in the processing methods. The behavior of BMGs as
composite matrix to a higher fraction of reinforcement is unknown.
1.2
Aims
The long term goal of Sandvik Coromant is to determine whether BMGs can be
used as replacement to cobalt as binder phase in hard metal products. This study
was the starting point of this effort and had the following aims:
• To investigate the possibility to achieve metal amorphous phases with powder
metallurgical methods and equipment available to Sandvik Coromant.
• To perform fundamental research on BMG composite materials.
Various kinds of BMGs as binder phase replacement were investigated and
attempts to produce samples with fully amorphous binder phase were made.
In this report, the expression ”bulk metallic glass” will be used for metallic
glasses that are possible to produce in larger pieces, even if the size in the application as binder phase will only be micrometer thin layers in between the WC
grains.
Chapter 2
Theoretical background and
literature review
2.1
Cemented carbide
Cemented carbide is a group of hard materials consisting of carbide particles in
a metallic matrix called binder phase. Cemented carbide is used in a variety
of machining tools, e.g. for rock drilling and metal cutting. It has a unique
combination of high hardness and high toughness. The amount of carbide phase,
varies from 70 to 97 wt.% with particles size ranging from 0.4 to 10 µm. One of
the most successful combinations is WC particles with Co as binder phase, but
types with varying fraction of titanium carbide (TiC), tantalum carbide (TaC)
and niobium carbide (NbC) are also used. Also, cobalt can be exchanged partly
or totally by other metals, such as iron (Fe), chromium (Cr), nickel (Ni) and
molybdenum (Mo). Cemented carbide is typically produced by mixing WC- and
Co powders with ethanol and a pressing agent. After drying, the composition is
pressed to a body with the same shape as the desired product. The pressed piece
is sintered, where the pressing agent evaporates and the Co melts and wets to the
WC grains, to the finished product [14].
2.2
2.2.1
Metallic glass
Background
An amorphous solid, also called non-crystalline solid, is a material that lacks the
long range order characteristic of a crystal, meaning that the positions of the
atoms no longer match with any crystalline lattice [15]. Instead, it has random
structure, where only short range order can be found, as shown in Figure 2.1. A
typical amorphous material is soda-lime-silica glass, used in window panes and
glass containers. A less known category of glass is the metallic glass, consisting
only or partly of metallic elements.
3
4
Theoretical background and literature review
Figure 2.1. Structure of a crystalline material (left) and the random packing in an
amorphous material (right).
In 1960, the first formation of metallic glass of Au75 Si25 was reported by Duwez
at Caltech, USA. Duwez used a quenching technique that allowed cooling at rates
of 105 − 106 K/s [8]. Since then, formation, structure and property investigations
of metallic glasses have attracted increasing attention. However, the required high
cooling rates limited the formation of metallic glasses to forms of ribbons and thin
sheets for a long time [9]. In 1969, Chen and Turnbull succeeded in formation of
0.5 mm amorphous Pd-Cu-Si at a cooling rate of 100-1000 K/s [16], and in 1974,
Chen broke the millimeter scale barrier of the bulk metallic glass when producing
1 mm diameter rods of the same composition using suction-casting method with
a cooling rate of 1000 K/s [17]. Since 1980, a number of strong glass forming
multi-component systems with critical cooling rates, sometimes as low as 1-100
K/s, have been developed. In the beginning they were based on expensive Pd, Pt
and Au, followed by less expensive Zr-, Ti- and Ni-based systems [18]. To date, it
is possible to produce bulk metallic glasses in up to 10 cm pieces [19].
Theoretically, any material can be obtained in an amorphous state just by
cooling it from its liquid state with a sufficiently high cooling rate. However, the
fast diffusion in most materials makes it impossible to reach ”sufficiently” high
cooling rates for this to occur. To understand why some materials have higher
tendency to form glass than other, understanding of the driving forces behind
crystallization and how it can be avoided is of great importance.
2.2.2
Thermodynamics and kinetics
The stability of a system is determined by its Gibbs free energy G
G = H − TS
(2.1)
where H is the enthalpy, T is the absolute temperature and S is the entropy of
2.2 Metallic glass
5
the system. When having the lowest possible value of Gibbs free energy, a system
is in equilibrium and will not transform into any other phase.
The enthalpy is derived as
H = U + pV
(2.2)
where U is the internal energy, p is the pressure and V is the volume of the
system. A solid, which has a small volume and strong bonding and thereby low
mobility of atoms, will yield a low enthalpy. At low temperatures, this term will
be dominating in the equation of Gibbs energy, hence the solid phase will be
thermodynamically favorable. When raising the temperature, the −T S term will
dominate and allow stable phases with more freedom of atomic movement, i.e.
liquids and gases.
In a metastable, or non-equilibrium, phase, a system has a higher Gibbs free
energy than in its equilibrium, but is prevented to access equilibrium by a potential
energy barrier [20] as illustrated in Figure 2.2.
Figure 2.2. Unstable, metastable and stable phases in the energy landscape. ri is a
3N-dimensional coordinate vector that represents a particular set of N atomic positions.
One such metastable phase is the supercooled liquid region (SLR), seen in Figure 2.3. When cooling a liquid below melting temperature (Tm ), a transformation
into the crystalline phase occurs as the system lowers its enthalpy and thereby the
Gibbs free energy. However, for a crystal to form, a seed crystal or nucleus of regularly arranged atoms must exist so that a crystal can form around that nucleus
6
Theoretical background and literature review
(heterogeneous nucleation). If such a nucleus is lacking, an even lower temperature is necessary for crystallization to occur, then by homogeneous nucleation
where a few atoms spontaneously arrange to a formation that can act as nucleus
for crystallization. Thus, the liquid phase can be extended to reach below melting
point and the liquid will enter the so-called supercooled liquid region. When the
liquid is in the SLR it can at any time crystallize by homogeneous nucleation.
Figure 2.3. Cooling process of a liquid. If heterogeneous nucleation (A) into crystalline
phase can be bypassed, the liquid will enter the supercooled liquid region (SLR), from
where it can either crystallize by homogeneous nucleation (B) or transform into glass.
By continuous cooling, the organization of atoms into the crystalline structure
will be slower and with a sufficiently high cooling rate, crystallization will be totally avoided and the liquid will solidify into its current structure; an amorphous
structure is achieved. It has been found that all materials exhibit a SLR. However, this region is mostly very narrow and most materials are unstable here and
will solidify into crystalline phase. The cooling rate required for metallic glasses
to bypass crystallization depends on its composition and varies from 10−1 K/s, in
glass forming Pd-based systems [21], to millions of K/s for elemental metals.
It was expected that reinforced particles in a glass forming liquid would act as
nucleation points and make the melt crystallize. However, this is not necessarily
true, as will be explained in Section 2.2.7.
2.2 Metallic glass
2.2.3
7
Glass transition temperature
The liquid-glass transformation implies no structure change and cannot be considered a real phase transition. Therefore no well defined liquid-glass transformation
temperature, often called glass transition temperature (Tg ), exists. In practice the
transformation temperature into glass is rather a temperature interval between
the SLR and the glassy state. Over the years, different conventional Tg have been
developed. One well known convention was suggested in 1948 by Kauzmann [22]
as
Tg =
2Tm
3
(2.3)
Tg is here related to the melting temperature, however, the lower the cooling
rate is through the SLR, the more time the atoms have to arrange at any given
temperature and a lower temperature for ”freezing” the liquid into glass is required.
Thus, the glass transition temperature interval increases with the cooling rate, and
a simple correlation to the melting temperature is misleading, although it gives a
hint about a reasonable range of the glass transition temperature.
Element
Iron
Copper
Nickel
Steel
Tg = 2Tm /3 (◦ C)
1023
721
969
901
Table 2.1. Glass transition temperatures for some element and alloys, calculated using
the Kauzmann equation. Melting temperatures taken from [23].
Another, more convenient, conventional definition of Tg is the temperature at
which the shear viscosity reaches 1013 P1 [24]. However, different behavior of
the viscosity of liquids have given raise to a specific classification of glass forming
liquids into ”strong” and ”fragile”, to reflect their dependence on temperature
[25]. Strong glass formers are those by which the viscosity obeys the Arrhenius
behavior, see equation 2.4. Fragile liquids are non-Arrhenius in their behavior.
k = Ae−Ea /kB T
(2.4)
The Arrhenius’ equation describes the dependence of the rate constant k of a
chemical reaction on the absolute temperature T , where A is a constant, Ea is
the activation energy and kB is the Boltzmann constant. It is an empirical rule,
and the physics behind the deviation in fragile liquids from this behavior is to
date unknown and of great interest in the research of glass forming liquids. A
representation of the ”fragility” in liquids is seen in an Angell plot, which shows
the deviation of fragile liquids from the Arrhenius behavior of strong liquids, see
Figure 2.4
1 poise,
unit of dynamic viscosity in the CGS system. 1P = 0.1Pa · s.
8
Theoretical background and literature review
Figure 2.4. Angell plot of the dependence of viscosity of ”strong” and ”fragile” liquids
on temperature. When cooling from T = ∞ (Tg /T = 0) to T = Tg (Tg /T = 1), the
logarithm of the viscosity (log η) of a ”strong” liquid will increase linearly, i.e. obey
an Arrhenius behavior. ”Fragile” liquids will deviate from this curve with their nonArrhenius behavior.
2.2.4
Glass forming ability
The classification of fragility has become an indicator of the glass forming ability in a liquid. At any given temperature in the SLR, a strong liquid has higher
viscosity than a fragile liquid, hence the atoms are less mobile and the ability to
arrange in crystal structure lower. Thus, a strong liquid is an easier glass former
than a fragile liquid.
How to predict the fragility and glass forming ability in a liquid is still not
totally known. However, during the search for new glass formers, the knowledge
of how to compose good glass formers has improved. While the first discovered
glass forming alloys were binary or ternary systems [8][17][26], the good glass
formers that are seen today usually consist of several elements. The vast amount
of work that has been performed since then has led to a few rules of thumb that
are based on experience. In 2000, Inoue et al. condensed these experiences into
three empirical rules that good glass formers obey [27]:
1. They are multi-component systems, consisting of three or more elements.
2. They have a significant atomic size ratio above 12 % among the three main
elements.
3. They have negative heats of mixing among the three main elements.
Thermodynamically, these rules prevent segregation in the liquid mixture and
assure mixing of the composition while cooling, as the mixing entropy of the multicomponent systems, Smix , is large and will favor randomness over organization.
2.2 Metallic glass
9
The rules also ensure a slow-down of the kinetics when supercooling and thereby
crystallization is prevented. Large negative heats of mixing mean strong bonding
between the elements and formation of local atomic pairs in the supercooled liquid
will complicate nucleation. Moreover, the more elements and the larger atomic size
difference, the more complex are the interactions and the more complicated will
the diffusion be.
By adding small amounts of additional alloys to the base alloy, these effects
can be enhanced. The additional elements increase the atomic size difference and
the complexity of interactions, which in turn complicate diffusion and stabilize
the liquid phase. Given that the concentrations of the additions are small enough
to not change the original competing crystalline phases, the added atoms will
destabilize the competing crystalline phase by positioning themself in or between
crystalline lattice points. This will introduce large strains in, and thus destabilize,
the crystalline lattice [28].
Additions can also contribute to elimination of oxygen impurities. Oxygen
impurities in metallic glass are shown to have a deteriorating effect on the GFA,
by reducing the supercooled liquid region [29] and stabilizing quasi-crystalline
phases [30]. The strong affinity of rare earth elements, such as yttrium (Y) and
neodymium (Nd), with oxygen can be used to prevent oxygen impurities by formation of oxide inclusions in the BMG. It has also been shown that additions
can modulate the microstructure at an atomic scale, thus improving properties as
thermal stability, strength, plasticity, magnetic properties and corrosion resistance
[28].
However, what might be the optimal addition for one system, can ruin another.
For example, adding 1 % of boron to Ni-Nb-Sn alloys has proven to improve the
GFA significantly [31], but the same amount deteriorates the GFA in a Zr-based
alloy [32]. Hence, the amount of additions has to be controlled to a very high
degree of accuracy.
In iron based glasses, small metalloid atoms, such as C, B, Si and P are frequently used as alloying additions together with rare-earth elements as Y and Tm
and/or special alloying elements such as Ln, Ga, Zr, Nb and Mo having significant atomic size difference and large negative heat of mixing between Fe and the
metalloid.
Glass forming criteria
Although there are rules to follow when composing glass forming alloys, the outcome of an additional element or modified properties in an alloy is hard to predict
and usually first known after the material is experimentally produced. GFA is
directly related to the critical cooling rate (Rc ) and maximum attainable size
(Dmax ) of the material. The smaller Rc or the larger Dmax , the higher is the GFA
of the system. However, Rc is hard to measure experimentally and Dmax depends
strongly on the fabrication method used [33]. Instead, different glass formation
criteria have been developed trying to predict the GFA in different systems. However, a correct description of the GFA must consider thermodynamic, kinetic and
structural constraints of the glass formation, which has proven to be difficult.
10
Theoretical background and literature review
One of the first glass forming criteria was the reduced glass transition ratio
proposed by Turnbull in 1960 [34] as the ratio of the glass transition temperature
to the melt temperature
Trg = Tg /Tm
(2.5)
The higher the value of the reduced glass transition ratio, the better the GFA
in the composition, which agrees well with the fragility classification of liquids as
strong liquids will yield a higher Trg than fragile liquids. However, the Turnbull
criterion was developed for monoatomic systems, which limits its applicability
in more complex multi-component BGMs [33]. Over the years, further criteria
based on the characteristic temperatures and different versions of the reduced
glass transition temperature have been developed, with the aim to improve the
usability in multi-component system [35][36][37]. Examples of such criteria are
γc = (3Tx −2Tg )/Tl [38], ω = Tg /Tx −s·Tg /(Tg +Tl ) [39], and β = Tg /Tx −Tg /Tl ·η
[40], where Tx is the crystallization temperature and Tl is the liquidus temperature.
However, criteria based on characteristic temperatures are to a high degree
empirical and for many BMGs mostly just a simple relation between Dmax and
the measured characteristic temperatures [41]. Moreover, the temperature based
criteria consider neither kinetics nor structural descriptions of the physics behind
the glass formation.
Other criteria, based on e.g. thermodynamical calculations of formation energies [42][43][44], formulas of kinetic factors of phase transformation or the size
difference of constituent elements in the glass forming alloy have also been developed. However, a universal glass forming criterion is still lacking as most of them
are just empirical rules based on experimental data from existing glass forming
liquids, and do not succeed in combining kinetic, thermodynamic and structural
description of the driving forces behind glass formation [41].
2.2.5
Mechanical properties of metallic glass
Metallic glasses exhibit unique properties, such as high strength and hardness and
good corrosion and wear resistance. However, they are in general also very brittle,
i.e. they undergo catastrophic fracture with very little or no plastic deformation
beforehand.
In crystals, plastic deformation usually occurs around and in presence of dislocations, which are defects in the crystal structure. Such defects can be so called
edge-dislocations, where an extra half-plane is introduced in the middle of the
crystal, or screw-dislocations, where parts of a plane is displaced and replaced by
its neighboring plane. When the crystal deforms, shear bands can be seen in areas
where the shear stress is maximized, as can be seen in Figure 2.5. The more shear
bands the sample generate, the larger is the plasticity of the material [45]. After
deformation, the crystallographic structure is preserved; hence, the material can
undergo further deformation without breaking.
2.2 Metallic glass
11
Figure 2.5. Visible shear bands after compression in z-direction of a cylindrical sample.
Illustrative example taken from [46].
In glass, the lack of long range order means that no dislocations exist. Plastic
deformation occurs by movement of local clusters of atoms to lower the stress,
as illustrated in Figure 2.6. However, the movement of the clusters is limited by
surrounding atoms, thus it cannot propagate through the material as in a crystal.
Also the movement creates local clusters of free volume, which are preferential
sites for breakage. With continued stress, local cracking will occur in the local
clusters of free volume that may propagate until complete fracture occurs. This is
expressed in the sample by very narrow shear bands and a low degree of plasticity
before fracturing [45].
Figure 2.6. Plastic deformation in an amorphous structure by movement of local clusters
of atoms, after which local clusters of free volume are created.
The limit in plastic deformation due to uninhibited propagation of shear bands
makes the metallic glass exhibit very high yield strength compared to the corresponding crystalline composition.
12
Theoretical background and literature review
Increased temperatures
Properties as high hardness and strength are valid for metallic glass at low temperatures. However, at temperatures close to the glass transformation temperature,
the deformation mechanism changes. The increased mobility of the atoms makes
it easier for them to rearrange and counteract the free volume in the local clusters
of atoms, making the material increasingly plastic as the temperature is increased.
Corrosion and wear resistance
Corrosion is a common problem for crystalline materials, usually occurring in grain
boundaries. However, the lack of grain boundaries in amorphous materials gives
the metallic glass high corrosion resistance. The combination of high corrosion
resistance, good mechanical strength and high elasticity make the glass wear resistant [45].
2.2.6
Metallic glass fabrication methods
The fabrication methods to produce glass can be divided into two categories; techniques that are based on rapid solidification of a melt and solid state processing.
Rapid solidification
Rapid solidification can be achieved by quenching a melt in a quenching media,
such as water, liquid nitrogen or air. The composition is usually heated in a
quartz tube above melting temperature, after which the quartz tube is quenched
in the quenching media. Quenching is a simple method, however, the cooling
rate is relatively slow, up to 103 K/s depending on the quenching media and the
composition of the material. The cooling rate of water is much higher than of air
and liquid nitrogen, due to the high heat capacity and thermal conductivity in
water, which makes it a preferable quenching media [47].
Another method based on rapid solidification, by which higher cooling rates
can be achieved, is casting, or injection molding. Here, a melt is poured into a
cooled mold, usually made of copper, where it solidifies extremely rapidly due to
the very high heat capacity and thermal conductivity of the mold. Over the years,
different casting techniques have been developed, e.g. die casting, in which high
pressure is used to force the melt into the mold, suction casting and ingot casting.
When producing a metallic glass matrix with particle-reinforcement by rapid
solidification, the glass matrix is heated to a melt before cooling, while the reinforced particles are still solid.
Solid state processing
An amorphous phase can also be achieved without heating the material above
melting temperature. By mechanical alloying, an amorphous powder can be produced, and by hot pressing the powder a bulk metallic glass is created. When hot
2.2 Metallic glass
13
pressing, the amorphous powder is heated to the SLR and with a pressure of at
least 1 GPa a dense body can be formed [48]. As the method only implies solid
state transformations, it is called solid state processing.
2.2.7
Effect of reinforcements in metallic glass matrix composites
By creating a composite material with hard reinforcing particles in a metallic glass
matrix, the mechanical properties of the glass can be improved.
In a study in 1997, pre-alloyed metallic glass was produced with 5-30 Vol%
reinforced SiC, WC, TiC, W and Ta by means of copper mold casting [49]. It was
noticed that the casting method set a maximum limit of the volume fraction of
reinforcements to 30 Vol%. The reason was that when the fraction of reinforced
particles increased, the viscosity of the liquid alloy increased, which in turn decreased the possible injection speed and thereby the cooling rate. Thus, a too high
volume fraction of particles in the glass matrix would hinder the glass formation
and make the liquid crystallize.
Analysis of the WC-reinforced metallic glass composites showed that the reinforcement did not affect the GFA of the metallic glass. This was surprising, as earlier studies have predicted that crystalline additions to glass forming melts would
act as catalytic sites for heterogeneous nucleation [26], as described in Section 2.2.2.
The unaltered GFA was ascribed to the relatively low processing temperatures at
which the glass matrix was produced, which would prevent excessive reactions
between the particles and the glass. Also the characteristic temperatures; Tg and
Tx , are reported to be unchanged with reinforcement, and the reinforcements are
being homogeneously distributed in the metallic glass [49][50]. When increasing
the volume fraction of reinforcement, an increase in Young’s modulus and Vickers
hardness are observed [50]. Also compressive fracture strength and compressive
plastic elongation improves significantly with reinforcements in the metallic glass.
In 1999, Yim et al. observed improved compressive strain to failure and energy
to break of reinforced metallic glass [51] and ascribed the increase in toughness
to the formation of multiple shear bands. As the shear stress build around the
particles in the glass under compression, the propagating shear bands slow down
and the surface area over which the fracture occur increases, explaining the improvement in the plasticity of the material.
In order to overcome the limitation in volume fraction of reinforcements, a
study of metallic glass with up to 50 Vol% reinforcement produced by means of
water quenching was performed in 2002 [52]. It was seen that crystals formed near
the interface between the matrix and the particles. A dramatic increase in strain
to failure was seen when increasing the volume fraction of particles, indicating
that the trend holds even for higher volume fractions of reinforcement.
To date, successful experiments with reinforcements in different types of metallic glass matrices produced by casting [53] [50] [49] [51] [54], quenching [52] and
mechanical alloying [55] have been made. The outcome from these studies can be
summarized as:
14
Theoretical background and literature review
• The reinforcement does not seem to change the GFA of the metallic glass
matrix.
• The reinforcement does not seem to change the characteristic temperatures
(Tg , Tx and Tm ) of the metallic glass matrix.
• Reinforced particles tend to distribute homogeneously in the glass matrix.
• Mechanical properties of the metallic glass, such as ductility, elongation and
toughness seem to improve with increasing volume fraction of reinforcement.
• A crystalline layer forms on the grain boundaries of the reinforcements.
For casting, the results above only apply to a limited volume fraction of reinforcements, due to limitations in the processing method. Moreover, no study has
been found using more than 50 Vol% reinforcements, hence no conclusion of the
behavior of metallic glass matrices with higher amounts of reinforcements can be
drawn.
Chapter 3
Experimental methods
3.1
3.1.1
Production methods
Milling
Mechanical milling
Mechanical milling (MM) is a milling technique used in order to reduce particles
size and gain homogeneity in a powder mixture. The powder mixture, together
with a grinding media, typically cylindrical or spherical milling bodies made of
stainless steel or cemented carbide, milling liquid and eventual additives are put
in a vial. While milling, collisions between the milling bodies will crack the powder
and cause reduction of its grain size [20]. The milling liquid ease the grinding and
prevents agglomeration and oxidation during milling.
Longer milling time gives smaller particle size, however the material will oxidize. As the oxygen will react with carbon during sintering, extra carbon is often
added to the powder in the milling process. During the milling process, test samples can be taken from the mill, which is called pulp sampling [14].
In this project, milling was performed in a low energetic mill. The raw materials
were milled together with WC-Co cylpebs (milling bodies) and ethanol in 0.25 l
jars.
Mechanical alloying
Mechanical alloying (MA) is a high energy milling process that, in difference to
conventional milling, result in particle fracture and heavy plastic deformation and
particle rewelding. By agitating the powder and milling bodies at a high speed,
the powder particles get crushed and flat, and together they form a thin layered
structure. MA is often a lengthy process with processing times from a few to
hundreds of hours. MA can be used to synthesize a variety of equilibrium and
non-equilibrium materials at room temperature [20]. Mechanical alloying was
performed within this project using an attritor mill.
15
16
3.1.2
Experimental methods
Drying
To separate a powder from a liquid, e.g. after milling, different drying methods
can be used. The easiest method is pan drying, where the mixture is put in a
drying cabinet to make the liquid evaporate. However, by spray drying a more
homogeneous powder with consistent particle size distribution can be achieved. In
spray drying, the liquid mixture is sprayed into a gas filled chamber through a
nozzle that distributes it into small drops. The drops dry to fine granules and are
channeled into a container where they are collected. When spray drying, much
of the powder gets stuck on the walls inside the equipment and is wasted. The
process is very time consuming for small amounts of powder compared to pan
drying [14].
3.1.3
Pressing
Pressing of powder is used in variety of applications. For example, compression
of milled and dried powder to a so called green body is done before sintering in
production of cemented carbide. Such pressing is usually performed at room temperature with pressures around 100 MPa. PEG (Poly ethylene glycol), a pressing
agent, which is added to the powder before milling, keeps the piece together [14].
In hot pressing, higher temperatures are used to cause a transformation in the material. When synthesizing diamond, hot pressing with temperatures around 2000
K and pressures from 5-10 GPa are used, so called HPHT (High Pressure/High
temperature) pressing [56].
3.1.4
Sintering
Sintering is a heat process in which a green body is ”baked” at a high temperature.
The process aims to reduce porosity and get a solid pore free body called ”blank”,
with desirable shape and composition and intended micro structure and property
profile. There are different sintering methods, e.g. spark plasma sintering and
liquid phase sintering. In this project vacuum sintering (denoted DA) and gas
pressure sintering (denoted GPS) are used. The sintering process is described
below.
Sintering Process
The sintering process includes four main steps, illustrated in Figure 3.1.
3.1 Production methods
17
Figure 3.1. Main steps in the sintering process. The full line shows the temperature in
◦
C and the dashed line shows the pressure in mbar.
1. Debinding in flowing H2 -gas from room temperature to 450◦ C. Removal of
PEG. The PEG molecule decomposes and reacts with H2 and forms CO,
CO2 , H2 O, CH4 and Cx Hy Oz .
2. Vacuum heating to sintering temperature to reduce oxygen that might originate from the raw material or the milling process or from aging of powder or
pressed green bodies. Together with carbon from the sample, CO is formed.
Thus the carbon level will be affected and additional carbon is often added
to the powder before milling. Vacuum pumps are used to pump away CO.
3. Liquid state sintering at 1410-1520◦ C in controlled vacuum at low pressure.
Binder phase melts which causes several reactions:
• Pore closure, which leads to shrinkage of body. This starts already
during the solid state sintering.
• Binder phase redistribution. The binder phase distribution is uneven,
however, when increasing the temperature it gets homogenized.
• WC-grain growth, where small WC grains dissolve and precipitate on
the bigger ones. Homogenization of γ phase, if added, i.e. the carbides
in the raw material will form a stable phase consisting of cubic carbides
with W-solutions.
4. Cooling in Ar-gas, where the binder phase solidifies. When sintering of WCCo cemented carbide, a thin surface of Co forms and shape distortions may
appear.
18
Experimental methods
Gas pressure sintering
By introducing high pressure towards the end of the normal sintering holding time,
i.e. after all porosities are closed, structural defects as porosity and binder lakes
are minimized. This is called gas pressure sintering or HIP (Hot Isostatic Pressing)
sintering [57].
Porosity reduction
When the binder phase spreads and forms a solid shell around the piece (see step
four in the sintering process) so called closed porosity is achieved. Thus, when
applying pressure on the piece, the pores in it will be reduced in size as the system
tries to reach equilibrium between the internal and applied pressure. However, if
this shell is not solid, the system will be open and internal and applied pressure
equal. Thus, the pores will not close, no matter how high the applied pressure is.
This is called open porosity [14].
3.1.5
Quenching
Quenching is a rapid solidification method that can be used when producing metallic glasses. By quenching a melt in a liquid, such as water, tin bath or liquid
nitrogen, the melt will undergo a rapid solidification and form glass phases. The
cooling rate by water quenching is about 102 − 103 K/s [9], which is much less
than the cooling rates achieved by newer techniques, such as casting. Thus, water
quenching is a technique that is less used in experimental research today.
3.1.6
Arc melting
In arc melting, a sample or powder is placed on a cooled copper plate and melted
by a DC current. The method is used in metallurgical industry and implies several
advantages over melting in open furnace; gases in the sample do not dissolve in the
liquid metal but escape to the vacuum chamber, centerline porosity and segregation
are prevented and the solidification rate can be tightly controlled.
3.2
Computational thermodynamics
By computational thermodynamics the equilibrium phases for different compositions can be calculated. In this project, the computational thermodynamical
software Thermocalc was used to calculate the weight fraction of the elements in a
composition to get the desired amount of binder phase, the sintering temperature
needed for the binder phase to melt and the optimal amount of carbon to avoid
graphite or unwanted phases [58].
3.3 Characterization methods
3.3
3.3.1
19
Characterization methods
Vickers hardness test
The Vickers hardness test is a method to measure the hardness of a material. With
a pyramid shaped indent of diamond, indentations in the material are performed
with desired force. The hardness is determined by the force applied to the sample
to the surface area of the resulting indentation (F/A [kgf/mm2 ]) and its unit is
called Vickers Pyramid Number (HV). For a cemented carbide with 80 Vol% WC
and 20 Vol% Co, the hardness is approximately 1300 HV30, where 30 is the applied
load in kgf. The hardness increase with decreasing volume fraction of binder phase
in the sample.
3.3.2
Fracture toughness
Fracture toughness (KIc ) measures the resistance to propagation of cracks in a
material in units of MPa · m1/2 . The toughness in cemented carbides with 80
Vol% WC and 20 Vol% Co varies from 15 to 20 MPa · m1/2 .
3.3.3
Simultaneous thermal analysis
Simultaneous thermal analysis (STA) is a thermoanalytical instrument that allows
simultaneous application of thermogravimetric (TG) analysis together with differential scanning calorimetry (DSC) or differential thermal analysis (DTA). The
STA can also be coupled with a mass spectrometer. In this project a STA409CD
of model Netzsch was used to perform DCS - TG analysis. It was also coupled
with a mass spectrometer. Mass spectroscopy was performed using a Quadropole
Mass Analyzer (QMS) of model Pfeiffer ThermoStar.
Differential scanning calorimetry
In DSC, caloric reactions are determined by measurement of the heat difference
between the sample and a reference during heating. The technique can be used to
study phase transitions and chemical reactions, such as oxidation, in a material.
In amorphous materials, DSC is used to detect the glass transition temperature
Tg and crystallization temperature Tx .
Differential thermal analysis
DTA measures the temperature difference between the sample and a reference
during heating. This will give information about phase transitions in the sample.
It is also possible to run simultaneously with a mass spectrometer.
Mass spectrometry analysis
In mass spectrometry analysis (MSA), in which a mass spectrometer is connected
while running DTA, the amount of evaporating atoms or molecules can be esti-
20
Experimental methods
mated. The vapor will affect ions and by measuring the ion charge a mass number
can be obtained and matched with relevant atom or molecules.
Thermogravimetric analysis
By a TG analysis the mass of the sample during heating is measured. A mass loss
usually occurs due to vaporizing of a constituent element or molecule within the
sample.
3.3.4
Carrier gas hot extraction
Carrier gas hot extraction (CGHE) is a chemical analysis method, by which nonmetals, e.g. oxygen (O), nitrogen and hydrogen can be detected in a material.
For oxygen measurement, the oxygen in the sample is reduced by formation of
carbon monoxide (CO) at temperatures below melting point. A carrier gas, often
helium, sweeps out the CO, which is directly, or after conversion to carbon dioxide
(CO2 ), detected by infrared absorption.[59]
3.3.5
Light optical microscopy
In light optical microscopy (LOM) an optical microscope using visible light and a
system of lenses, is used to magnify images of samples. By focusing light from a
light source that is being either transmitted through or reflected from the sample,
a magnified image of the surface can be created. LOM is a cheap and easy method
to analyze the surface of a sample. The limitations are its low magnification (100x1200x) and the ability to only image dark or strongly refracting objects [60].
In LOM, the porosity of the sample can be classified with a code, according
to ISO 4505 standard. The code consists of a letter from A to C, which is an
estimation of the pore size, and a number from 00 to 08, which is an estimation
of the pore frequency. For example, a porosity of A02 means small and few pores,
and a porosity of C08 means many large pores.
Additionally, various types of etching techniques may be used to improve the
contrast of different phases, e.g. η phase, for analysis in the LOM.
3.3.6
X-ray diffraction
In x-ray diffraction (XRD), x-rays are sent onto a surface, after which the reflected
beams are being analyzed. In a sample, different crystal planes will cause the
incident x-rays to diffract into specific directions that depend on the atom types,
interplanar distances and crystal symmetries. By measuring the intensity and
angles of the outgoing, diffracted beams, a mapping over the different crystalline
elements and lattice structure in the sample can be created.
In non-crystalline materials, the pattern will be continuous in appearance and
a small ”halo” can be seen, arising from the short-range order existing in noncrystalline materials. When no crystalline signal is visible in XRD, the sample is
called x-ray amorphous [61]. An x-ray amorphous XRD result may also be due to
very small crystalline grain size in a material, as the detection limit of crystalline
3.3 Characterization methods
21
grain size in XRD is roughly 30Å.
In this project, XRD was performed using PanAlytical CubiX 3 Minerals and
a Bruker Discover-GADDS D8, both with a Cu − Kα radiation source. XRD was
mainly used to compare differences and eventual disappearance of crystalline peaks
in the XRD diffractogram before and after quenching or arc melting of a sample.
3.3.7
Scanning electron microscopy
In a scanning electron microscope (SEM), information about the topology and
composition of a sample can be achieved. The SEM scans a sample with a focused
beam of electrons, which by interaction with the atoms in the sample are being
scattered in different directions. By measurement of secondary electrons, i.e. electrons that are being emitted by atoms excited by the electron beam, a topological
image of the surface of the sample is created.
If instead measuring the backscattered electrons, i.e. electrons from the primary beam, which, after reacting with the atoms in the sample, are being sent
back with little energy loss, compositional changes in the sample can be imaged.
The resolution of SEM is down to 1 nm, and it is an important instrument for
characterization of materials. However, it requires the samples to be conducting
or prepared with a conducting coating [62].
3.3.8
Energy dispersive x-ray spectroscopy
In energy dispersive x-ray spectroscopy (EDS), x-rays produced by emitted electron from atoms in the samples are detected. The x-rays are characteristic for
the different elements and therefore a chemical mapping over the sample can be
created. An EDS is typically integrated into a SEM. However, too light elements
cannot be detected and detection with good precision of some elements is complicated due to overlapping x-ray signals [62]. Within the project, SEM and EDS
analysis were carried out using a Zeiss, Supra 40 and/or Supra 55V in order to
determine the homogeneity and composition of the binder phase.
22
Experimental methods
Chapter 4
Results
4.1
Glass formation by rapid solidification
Different glass forming alloys were produced as binder phase with WC as hard
phase. Hereafter, the expression ”binder alloy x” refers to the glass forming alloy
x. Furthermore, the expression ”sample x-y” refers to the sample number y in a
set where binder alloy x is used as binder phase. All samples were produced with
approximately 80 Vol% WC and 20 Vol% binder phase. Some of the samples were
quenched. The quenched samples will be denoted with a ”q” behind the sample
name, i.e. sample 1-1q refers to sample 1-1 after quenching.
Four different glass forming alloys were selected and used as binder phase. The
alloys are listed in Table 4.1:
Alloy no.
1
2
3
4
Binder alloy (at.%)
Fe66.7 P8.7 C7 B5.5 Mo4.5 Si3.3 Cr2.3 Al2
Fe73 P8.7 C7.0 B5.0 Si3.3 Mo3.0
Fe41 C15 Cr15 Mo14 Co7 B6 Y2
Fe64 C15 Mo14 B7
Dmax (mm)
6
7
16
2.5
Reference
[63][64]
[65][66][67]
[68]
[69]
Table 4.1. Glass forming alloys that were used as binder phase, and their reported
maximum obtained glass forming thickness (Dmax ).
All binder alloys together with tungsten carbide were simulated using Thermocalc to achieve approximately 20 Vol% binder phase. The carbon level was used
as variable to avoid stable phases with high melting points. Relevant data and
diagrams from the Thermocalc simulations can be found in Appendix A.
All in all, 19 samples were produced using any of the binder alloys as binder
phase. All data about drying methods, milling times and sintering processes for
the different samples are summarized in Table B.1. WC with the grain size of
approximately 5 µm was used in all samples. A selected number of the sintered
23
24
Results
pieces were heat treated and subsequently quenched in ice water. The heat treatment of the samples was performed in a tube furnace and in an induction furnace.
One sample was treated by arc melting.
4.1.1
Binder alloy 1
The powder of binder alloy 1 was an amorphous powder produced by means of gas
atomization in an earlier Sandvik project [70]. A chemical analysis on the powder
within that project showed that the powder composition was
Fe77 P5.1 C1.73 B1.5 Mo9 Si2.5 Cr2.6 Al0.86 Mn0.63 Ti0.5 Ni0.13 N0.002 (wt.%). This composition was used when calculating the carbon window for the samples with this
powder as binder phase.
Sample 1-1
Sample 1-1 was produced using all grain sizes of the amorphous binder alloy 1
powder. It was milled for 3 hours, pan dried and vacuum sintered in 1450 ◦ C.
The short milling time was chosen to not risk damaging the powder configuration
and glass structure. From LOM, a very high and open porosity of the sample was
detected (Figure 4.1).
Figure 4.1. LOM of sample 1-1, in which a very high porosity is seen.
To find the cause to the high porosity, the amorphous binder alloy 1 powder
4.1 Glass formation by rapid solidification
25
was sieved; to grains smaller than 20 µm, between 20 and 125 µm and larger than
125 µm. The smallest grains were clay like powder and the 20-125 µm grains
were fine spherical granules. The largest grains, i.e. what was left, consisted of
large granules, metallic flakes and whiskers with the shape of dust bunnies. XRD
of the different powders (see Figure 4.2) shows a small halo, characteristic for
an amorphous structure, in all three diffractogram, but no significant difference
between the grain sizes.
Figure 4.2. XRD diffratogram of amorphous binder alloy 1 powder sieved to grains
smaller than 20 µm (uppermost), 20-125 µm (middle) and larger than 125 µm (downmost).
It was decided to only use the powder with grain sizes smaller than 20 µm in
binder alloy 1, as a homogeneous powder was desired and the smallest grains were
most likely to be fully amorphous. Hereafter, all experiments with binder alloy 1
have been performed with the sieved powder with grains smaller than 20 µm if
nothing else is said.
Sample 1-2 and 1-3
In sample 1-2 and 1-3 sieved binder alloy 1 powder was used. The samples were
produced using different drying methods; sample 1-2 was dried by spray drying,
while sample 1-3 was dried by oven drying.
In LOM of the sintered samples (see Figure 4.3) a somewhat lower porosity
than in sample 1-1 was seen, which was explained by a more homogeneous powder
and binder phase distribution due to smaller grains. However, the porosity was
still open. It was also seen that the different drying methods had no significant
effect on the porosity. Therefore, spray drying, as it is a very time and powder
consuming method, was not used as drying method for the following experiments.
26
Results
Figure 4.3. LOM of sample 1-2 (left) and sample 1-3 (right). No significant difference
is seen between the samples, which were dried with different drying methods.
As the porosity was still very high, CGHE was performed on the amorphous
binder alloy 1 powder to see if a high oxygen content could be the cause to the
high porosity. The powder was also analyzed in SEM. The CGHE analysis showed
an oxygen content of 0.036 wt.%, which is low and considered not high enough to
cause the porosity seen in the samples. In SEM (see Figure 4.4) a similar surface
structure and morphology are seen in all grains, which indicate that they have the
same composition. The different geometrical shapes of the grains can be explained
by the rapid cooling rate in the gas atomization process.
Figure 4.4. SEM image of amorphous binder alloy 1 powder, in which similar surface
structure and morphology are seen in all grains.
4.1 Glass formation by rapid solidification
27
Sample 1-4, 1-5, 1-6 and 1-7
Sample 1-4, 1-5, 1-6 and 1-7 were produced with the extended milling time of 30
hours with the aim to get a more homogeneous power mixture and thereby try to
lower the porosity. The samples were produced with different sintering methods
(see Table 4.2) to investigate the effects of sintering pressure and temperature on
the porosity.
Sample
1-4
1-5
1-6
1-7
Sintering method
Vacuum (DA)
Vacuum (DA)
Press (GPS)
Press (GPS)
Sintering temperature (◦ C)
1410
1500
1410
1500
Table 4.2. Sintering methods for samples 1-4, 1-5, 1-6 and 1-7.
In LOM (Figure 4.5) a lower porosity in the sintered samples than in previously
samples was seen, however, still high. No significant difference in porosity was seen
after the different sintering processes, indicating that the porosity was still open.
Figure 4.5. LOM on sample 1-4 (upper left), 1-5 (upper right), 1-6 (bottom left) and
1-7 (bottom right). No significant difference in porosity is seen in the samples, which
were produced with different sintering methods.
28
Results
The amorphous binder alloy 1 powder was analyzed in STA to determine if
gassing of any element within the sample could be the cause to the high porosity.
From DSC and TG (Figure D.1) a mass loss was detected, starting at approximately 1300◦ C. Figure 4.6 shows DTA and MSA on the same powder where a
vapor with the mass number 28 can be seen. The mass number matches with Si.
The conclusion was that gassing of Si was the cause to the open porosity in the
samples. CO too matches with the mass number 28, however, is unlikely to be
gassing at that high temperature.
Figure 4.6. DTA, MSA and TG on the amorphous binder alloy 1 powder. Mass curve
of mass number 28 is plotted, which corresponds to evaporation of Si.
In addition, oxygen analysis was performed on powder of 80 Vol% WC and 20
Vol% binder alloy 1, milled for three hours. The measurement showed a oxygen
content of 0.096 wt.%, which is in the range of normal oxygen content for other
hard metal powder. Longer milling time was considered not able to cause an
unmanageable oxygen content.
Sample 1-8
In order to avoid evaporation of Si, which according to Figure D.1 started at
1300◦ C, sample 1-8 was sintered at 1250◦ C with a holding time of one hour. This
did however worsen the porosity, as was seen in LOM (Figure 4.7). This was
probably due to lower degree of pore closure as a consequence of the low sintering
temperature.
4.1 Glass formation by rapid solidification
29
Figure 4.7. LOM on sample 1-8. A high porosity is seen.
As the problem with gassing of Si could not be overcome by means of relevant
methods available in this project, no further experiments were done using binder
alloy 1.
Quenching of sample 1-1, 1-2 and 1-3
As the binder phase was expected to crystallize during the slow cooling after sintering, quenching was performed in order to retain the amorphous phase. Pieces of
sample 1-1, 1-2 and 1-3, with the approximate size of 12 · 6 · 5mm3 , were quenched
after heat treatment in a tube furnace. The samples were put in a molybdenum
crucible and placed in the middle of a Sarlin tube furnace at a temperature of
1400 ◦ C. The heat treatment was performed in an Ar fluxed environment. After a
holding time of 40 minutes, the samples were taken out and quenched in a bucket
with ice water. The time from the sample being in the oven to water was approximately 1 minute. Quenching data are summarized in Table B.4.
From EDS mapping on sample 1-2q (see Figure 4.8) it was seen that Mo has
dissolved into WC. Also, areas with a high concentration of Ti in the binder phase
were seen as well as areas where P is missing in the binder phase. All EDS mapping
images can be found in Appendix E. The changes in the binder phase composition
did most likely occur during the sintering process.
30
Results
Figure 4.8. EDS mapping on 1-2q. In the uppermost images, it can be seen that Mo
has dissolved into WC. The images in the middle show regions with high concentration
of Ti and the downmost images show regions where P is missing in the binder phase.
4.1 Glass formation by rapid solidification
31
The samples were analyzed with XRD before and after quenching. In all diffractogram, crystalline peaks from WC were seen. In the range 37-48◦ , crystalline
peaks from the binder phase were seen in all samples. Some peaks had decreased
or disappeared after quenching. However, the binder phase in all samples were still
crystalline. Diffractogram from a full scan of sample 1-3 and 1-3q and enlarged
diffractogram in the range 37-48◦ from all samples can be found in Appendix C.
Enlarged XRD diffractogram from sample 1-3 and 1-3q are seen in Figure 4.9.
Figure 4.9. Enlargement of XRD in region 37-48 ◦ on sample 1-3 (uppermost) and 13q (downmost). It can be seen that quenching changed the crystalline structure of the
binder phase, however, amorphization did not occur.
The crystalline peaks in the x-ray diffractogram indicate that the amorphous
phase in binder alloy 1 crystallized during cooling after sintering and was not retrieved by quenching. This could be explained by changes in the binder phase
composition during sintering as seen from EDS mapping. Another reason could
be a too low cooling rate as the samples were allowed to cool in air for quite
long time before they were quenched in water. This time could not be decreased
without compromising with the safety during the experiment. Instead, alternative
quenching methods were investigated for the following experiments.
It was clear that usage of binder alloy 1 powder as binder phase caused problems
with porosity, due to gassing of Si during sintering. This might been overcome by
a specially designed sintering process, however, this was considered to be too time
consuming for this study.
32
4.1.2
Results
Binder alloy 2
One sample was produced with binder alloy 2 as binder phase. Industrial raw
materials, mostly carbides, were used and can be found in Table B.2. Stones
with the reported chemical composition Fe72.61 P27.23 Si0.16 (wt.%) were used as
phosphorus source. The stones were pulverized and sieved to 180 µm and will be
referred to as ”FeP-stones”.
Sample 2-1 was produced according to data in Table B.1. LOM on sample 2-1
(Figure 4.10) shows a high porosity and large binder lakes along the edges of the
sample.
Figure 4.10. LOM on sample 2-1. In the rightmost picture, one of several binder lakes
is seen.
CGHE analysis on the pulverized FeP-stones showed a varying oxygen content
between 15 and 34.6 wt.%. Thus, the binder phase had been pressed towards the
edges by evaporating oxygen during sintering, which explained the binder lakes.
To avoid the high oxygen level, drying of the FeP-stones was considered. However, the stones would still bind oxygen in the milling process and the problem
would remain. Moreover, reduction of this much oxygen during sintering is not
feasible. As no reasonable solution to get rid of the oxygen and no alternative
phosphorus source were available, the usability of binder alloy 2 as binder phase
was considered as not feasible and no further steps were performed.
4.1.3
Binder alloy 3
Sample 3-1, 3-2, 3-3, 3-4 and 3-5 were produced using industrial raw materials,
mostly carbides. The raw materials can be found in Table B.2. The samples were
produced with different sintering methods according to Table 4.3. Due to the
absence of Si in the binder phase, which caused high porosity in sample 1-1 to
1-8 due to gassing during sintering, the porosity was expected to be lower than in
those samples.
4.1 Glass formation by rapid solidification
Sample
3-1
3-2
3-3
3-4
3-5
Sintering method
Vacuum (DA)
Vacuum (DA)
Press (GPS)
Press (GPS)
Press (GPS)
33
Sintering temperature (◦ C)
1410
1500
1410
1500
1520
Table 4.3. Sintering methods for the samples 3-1, 3-2, 3-3, 3-4 and 3-5.
From the LOM analysis (Figure 4.11), a significantly decrease in porosity compared to sample 1-2 to 1-8 is seen. The porosity was estimated to A08 and B06 in
sample 3-4 and A06 and B06 in sample 3-5, i.e. the effects of the changes in the
sintering temperature and pressure were not significant. In LOM image of sample
3-2 regions of η phase are seen, which were observed in all samples and confirmed
by etching. LOM images of etched samples can be found in Appendix F. The η
phase formation was explained by a too low carbon content.
Figure 4.11. LOM on samples 3-1 (upper left) 3-2 (upper middle) 3-3 (upper right) 3-4
(bottom left) and 3-5 (bottom right).
From SEM and EDS analysis, it was seen that Mo had dissolved into WC and
that yttrium oxide was formed (Figure 4.12). However, in the reference, from
where the glass forming composition of binder alloy 3 was chosen [68], it is unclear
whether yttrium is added to eliminate oxygen impurities by formation of yttrium
oxide (as discussed in Section 2.2.4) or to be part of the final glass composition.
34
Results
Figure 4.12. EDS on sample 3-2, where formation of yttrium oxide is confirmed.
Hardness test and KIc were performed on sample 3-5, with four indentations.
The average value for hardness was 1932 HV30 and for KIc 6.74 MPa · m1/2 . The
high hardness was explained by a lower fraction of binder phase than calculated
due to formation of η phase. Data from the hardness test and KIc for all four
indentations can be found in Table B.3.
Quenching of sample 3-1 and 3-3
Sample 3-1 and 3-3 were cut into pieces with approximate sizes of 12 · 5 · 1mm3
and quenched after heat treatment in an induction oven. The samples were put in
a molybdenum crucible in a platinum crucible inside an induction coil and heated
to approximately 1400◦ C, after which the samples were immediately quenched in
a bucket with ice water. The holding time was 15 minutes for sample 3-1q and
approximately 3 minutes for sample 3-3q. The long holding time for sample 3-1q
lead to formation of a surface layer.
XRD analysis was performed on the samples before and after quenching. In the
diffractogram of sample 3-3 and 3-3q in the range 37-48◦ (see Figure 4.13), it can
be seen that signals from crystalline binder phase have changed during quenching,
but have not disappeared which means that the binder phase in both samples are
crystalline after quenching. In the diffractogram of sample 3-1 and 3-1q, that can
be found in Appendix C, appearance of new peaks are seen in sample 3-1q, which
was due to signals from the surface layer.
4.1 Glass formation by rapid solidification
35
Figure 4.13. XRD diffractogram of sample 3-3 and 3-3q in the range 37-47◦ . No significant change in the crystalline structure of the binder phase can be seen after quenching.
A reason why amorphization did not occur might be that the cooling rate was
too low. However, the cooling rate should according to literature be high enough
for this alloy to form glass. Another reason may be that the production process
changed the composition of the binder phase so that the new composition had a
lower GFA. Furthermore, rapid cooling may have occured in wrong temperature
window, below the SLR, i.e. when the sample have already crystallized. If so, a
higher temperature during heat treatment is needed, however, no higher temperature was possible when using the induction oven.
Arc melting of sample 3-4
Two pieces of sample 3-4 with the approximate size of 7 · 6 · 5mm3 were treated by
arc melting, in a Mini Arc Melter, from Edmund Bühler. The pieces are denoted
3-4ac1 and 3-4ac2. The pieces were heat treated for approximately 5 seconds (34ac1) and 30 seconds (3-4ac2) and cooled on a copper plate. In XRD analysis (see
Appendix C) crystalline peaks from the binder phase are seen in both samples.
4.1.4
Binder alloy 4
Sample 4-1, 4-2, 4-3 and 4-4 were produced with binder alloy 4 as binder phase.
Industrial raw materials were used and can be found in Table B.2. The samples
were produced with different sintering methods according to Table 4.4.
36
Results
Sample
4-1
4-2
4-3
4-4
Sintering method
Vacuum (DA)
Vacuum (DA)
Press (GPS)
Press (GPS)
Sintering temperature (◦ C)
1410
1500
1410
1500
Table 4.4. Sintering methods for sample 4-1, 4-2, 4-3 and 4-4.
From LOM (Figure 4.14) a low porosity was seen. In sample 4-4, the porosity
was estimated to A00. This was explained by the fact that binder alloy 4 had a
simple composition with convenient constituents. η phase formation was seen in
all samples and confirmed by etching. LOM images of the etched samples can be
found in Appendix F.
Figure 4.14. LOM on samples 4-1 (upper left), 4-2 (upper right), 4-3 (bottom left) and
4-4 (bottom right).In all samples, η-phase formation is seen.
Measurement of hardness and KIc was performed on sample 4-4. From five
indentations, the average value for hardness was 1668 HV30 and for KIc 8.52
MPa · m1/2 .
4.1 Glass formation by rapid solidification
37
Quenching of sample 4-3
Sample 4-1 and 4-3 were quenched with the same procedure as sample 3-1 and
3-3 (see Section 4.1.3). The size of the pieces was approximately 12 · 5 · 2mm3 .
Holding times of heat treatment were 15 minutes for sample 4-1 and approximately 3 minutes for sample 4-3. XRD analysis on the samples before and after
quenching was performed. In the diffractogram in the range 37-47◦ , crystalline
peaks from the binder phase were seen in the samples before and after quenching.
X-ray diffractogram from sample 4-1 and 4-1q can be found in Appendix C. X-ray
diffractogram from sample 4-3 and 4-3q are shown in Figure 4.15 below.
Figure 4.15. Enlargement on XRD on sample 4-3 and 4-3q in region 37-47 ◦ . It can
be seen that quenching changed the crystalline structure of the binder phase. However,
amorphization did not occur.
The porosity in the samples were significantly lower than in the previous samples and, with some improvements in the production process, in reach of the porosity values needed to be usable as working material. However, the low porosity was
explained by the simple composition of binder alloy 4, which also implied a lower
GFA.
Reasons why amorphization did not occur, as previously discussed in Section
4.1.3, might be too low cooling rate, decrease in GFA of the binder phase due to
changes in composition during production or cooling in wrong temperature window. As binder alloy 4 has lower GFA than binder alloy 3, the lack of amorphous
phase was not surprising.
38
4.2
Results
Glass formation by solid state processing
Attempts to produce amorphous powder and BMGs were performed with different
methods, starting from different powders. The powders and the used production
methods are summarized in Table 4.5.
Powder
5
6
Alloy composition (at.%)
Ti50 Cu28 Ni15 Sn7
Fe66.7 P8.7 C7 B5.5 Mo4.5 Si3.3 Cr2.3 Al2
Production method
Mechanical alloying
Hot pressing with low pressure
Table 4.5. Experimental set, glass forming alloys and the production methods that were
investigated.
4.2.1
Mechanical alloying
The aim of this part of the investigation was to find out if it was possible to
produce amorphous powders in an easy way with the equipment at hand. As a
first test, reproduction of a work where an amorphous powder was prepared by
mechanical alloying [71][72] was performed.
Powder 5 was milled in a high energetic attritor mill with WC milling bodies
and ethanol at a speed of 700 rpm in an N2 fluxed environment. The raw materials
that were used can be found in Table B.2. Pulp samples were taken regularly,
according to Table 4.6.
Pulp sample
5-1
5-2
5-3
5-4
5-5
5-6
5-7
Milling time (h:m)
1
2
3
4
5
5:20
5:33
Table 4.6. Pulp samples and their milling times from mechanical alloying of alloy 5.
After 5 hours and 33 minutes, the experiment was interrupted due to significant
wear on the attritor beaker that was close to failure. The powders were dried in a
drying cabinet at 80 ◦ C. All samples were analyzed with XRD (see Figure 4.16).
In the diffractogram, it can be seen that crystalline signals that appear in the pulp
samples with shortest milling time have decreased and some have disappeared in
the pulp samples with longer milling time. However, there are still crystalline
peaks in the powder with longest milling time which means that the powder is not
amorphous.
4.2 Glass formation by solid state processing
39
Figure 4.16. XRD on pulp samples from mechanical alloying. In all diffractogram,
signals from the holder are seen. Uppermost diffractogram is from the holder only,
followed by pulp samples after 1, 3, 5 and 5:33 hours.
A reason why the powder with the longest milling time is still crystalline might
be that higher milling speed was needed. In the reference, in which amorphicity
was achieved by mechanical alloying of powder, a shaker mill, which can reach
higher intensity than the attritor mill, was used. Moreover, that experiment had
milling times up to seven hours. However, the usability of a shaker mill is limited
as it can only hold a small amount of powder. The attritor mill that was used
in this project was run at maximum speed, and the wear of the milling beaker,
due to the high speed, made it impossible to run for longer time than what was
completed. Also, the procedure was very time consuming as, when running at
maximum speed, the attritor mill had to be constantly monitored.
Even if higher speed and longer milling time might result in x-ray amorphous
powder, this can be due to extremely small grains, and a thorough transmission
electron microscopy analysis must be conducted to ensure that amorphization has
really occurred.
Due to the complications mentioned above, it was decided not to investigate
the method of mechanical alloying further.
4.2.2
Hot pressing
Starting with an amorphous powder, BMGs can be produced by hot pressing the
powder into solid pieces. This is performed with high pressure and a temperature
close to Tg .
In this project, powder 6, which was an amorphous powder, was hot pressed in
order to achieve a BMG. More about the powder, which was the same as binder
alloy 1, can be found in Section 4.1.1.
40
Results
The powder was pressed in a KCE SAK180 press with pressures and temperatures as given in Table 4.7. From DCS on the region below crystallization
(see Figure 4.17), Tg can be estimated to 530 ◦ C. Pressing started below this
temperature and continued with increasing temperature trying to reach optimal
temperature window. The pressures was the highest possible with the equipment
available.
Sample
6-1
6-2
6-3
6-4
Temperature (◦ C)
450
550
650
1000
Pressure (MPa)
38
38
38
127
Note
Below SLR
In SLR
Above SLR
Melting point
Table 4.7. Pressing temperatures of samples when hot pressing at low pressures.
Figure 4.17. Enlarged DSC on powder 6/binder alloy 1 in the region around crystallization. From the graph, temperatures of the SLR and Tg can be estimated.
Pressing of sample 6-1 did not result in a solid piece, but just powder. Pressing
of sample 6-2 and 6-3 resulted in fragile pieces and powder. XRD analysis on
the pieces of sample 6-2 and 6-3 (See Figure 4.18) shows crystalline peaks. The
appearance of new crystalline peaks in sample 6-3 that are not seen in sample 6-2
indicates that the pressing temperature (650◦ C) was too high, i.e. above the SLR.
Pressing of powder sample 6-4 was performed with a higher pressure, however,
this caused breakage of the holder and overflow of the melt. The outcome was
solid metallic pieces, however, not glass like.
4.2 Glass formation by solid state processing
41
These results are similar to the results in a reference [70], where the same
powder was hot pressed with a pressure of 150 MPa.
Figure 4.18. XRD on Powder sample 6-2 (upper) and 6-3 (lower) after low pressure
hot pressing. In powder sample 6-3, crystalline peaks that are not seen in powder sample
6-2 have appeared.
In an experiment where a Ti-based amorphous powder was hot pressed to
a BMG it was reported that the minimum pressure that was needed to avoid
crystallization was 0.96 GPa [48]. The maximum pressure that could be obtained
in this project, without breakage of the equipment was only 38 MPa. Thus, it was
concluded that the pressure that was required to compress the amorphous powder
into a BGM was too high to be achievable at the lab in Västberga. However,
the necessary pressure range is possible to reach by using equipment and methods
used in diamond synthesis process. This would make it possible to control the
composition in the binder phase. However, the possibility to press pieces with
high amount of WC is not known and unexpected hinder might appear.
Such a test was prepared and started in this study, however, the results were
delayed and are not included in this report.
42
Results
Chapter 5
Conclusions
The conclusions from this study are listed below. Although no metallic glass
as binder phase in hard metal was produced, valuable information about glass
formation of metals in industrial environment were discovered.
Glass formation by rapid solidification
• The use of binder alloy 1 and binder alloy 2 as binder phase resulted in
samples with very high porosity, which was due to evaporation of Si and O
during the sintering process.
• The use of binder alloy 3 and binder alloy 4 as binder phase resulted in
samples with a relatively low porosity. The lowest porosity was found in the
samples with binder alloy 4 as binder phase. This was ascribed to the simple
configuration and well known elements in the composition.
• No amorphous phase was achieved after quenching or arc melting of any
sample.
• In all samples, the composition changed during the production process, e.g.
Mo dissolved into WC and Y2 O3 was formed.
• Amorphous powder did not generate an amorphous bulk after sintering.
In general, metallic glass that have high GFA consist to high degree of several
elements, some of which are not commonly used in industrial powder metallurgical
production. Moreover, metallic glasses that have simple configurations and consist
of elements, with well known behavior, generally have low GFA. The sensitivity
in GFA to changes in the alloy composition in metallic glass makes it crucial to
be able to control the alloy composition throughout the production very precisely
to ensure that glass formation is possible in the produced binder phase. Powder
metallurgical methods, that involve milling and sintering, make accuracy in alloy
composition of the metallic glass difficult to ensure, since evaporation and WCbinder phase intermixing will change the glass forming composition of the binder
43
44
Conclusions
phase. This means that metallic glass with high GFA will be harder to produce,
however, the metallic glass that might be possible to produce have in general lower
GFA.
Glass formation by solid state processing
• Mechanical alloying of a glass forming composition in an attritor mill did not
result in amorphous powder. Amorphicity in powder by mechanical alloying
in an attritor mill may be possible in small scale, however, milling times are
long and even if x-ray amorphicity is achieved, this might be due to defects
in the crystal structure after milling.
• Hot pressing of amorphous powder with or without WC into a BMG is not
possible at low pressures, as the pieces will not totally solidify and the powder
will crystallize.
• Production of hard metal with high fraction of WC and metallic glass as
binder phase by hot pressing at high pressures may be possible. However,
no studies of hot pressing of metallic glass powder with higher fraction of
reinforcement than 12 Vol% have been found. Thus, further investigations
must be conducted.
A literature study shows that most metallic glass have low glass transition temperature and crystallization temperature, around 700-900◦ C, which makes them
inappropriate as binder phase in cutting and mining equipment where the working
temperature is around 900◦ C. However, there might be other applications if a solid
piece of WC-BMG is possible to obtain.
Confirmation of amorphous phase in the binder phase is hard to perform in a
simple way. The high amount of WC dominates the XRD diffractogram, which
is the most simple method to confirm amorphous phase. Moreover, an x-ray
amorphous result might be due to small grain sizes or defects in the crystalline
structure. Thus, to confirm if a material is truly amorphous, TEM, which is a
more expensive method, must be conducted.
Chapter 6
Future work
Glass formation by rapid solidification
As the samples with binder alloy 3 and binder alloy 4 had relatively low porosity,
these compositions might be interesting to investigate further by
• Investigate the effect of different carbon content. η phase formation was seen
in some of the samples which indicates that more carbon is needed. Also, in
the samples where the added carbon content according to Thermocalc would
give graphite, no graphite was seen. This was probably due to carbon loss
during sintering.
• Investigate the effect of different fraction of binder phase. As areas without
binder phase was seen in the produced samples, a higher fraction of binder
phase might be of advantage. Moreover, crystallization will most likely occur
close to the WC boundary according to literature, thus, larger areas of binder
phase might be needed to achieve glass formation. With more binder phase,
it might also be easier to analyze the samples with XRD, as the crystalline
peaks from WC would be less dominating. However, by increasing the binder
phase fraction, the expected cooling from WC might be lost.
• Modify the binder alloy compositions so that the final compositions after
sintering correspond to the desired glass forming compositions, e.g. by calculating how much Mo that dissolve into WC and adding extra Mo to the
binder alloy. For binder alloy 3, it must be known if the formation of yttrium
oxide is a desired effect or not.
• Modify the sintering process to create less possible impact on the binder
phase composition and to avoid gassing during sintering.
Moreover, the behavior and properties of other glass forming metallic alloys,
e.g. Pd- or Zr-based glass might be interesting to investigate.
45
46
Future work
Glass formation by hot pressing
No solid pieces were achieved by hot pressing with low pressure. However, it
might be interesting to perform hot pressing with high pressures, above 1GPa, on
amorphous powder with and without WC reinforcement, and moreover investigate
the effect of different volume fraction of WC reinforcement. Such experiments were
started in this project, however, the results were delayed and are not included in
the report.
Bibliography
[1] A. A. Jensen and F. Tuchsen. Cobalt exposure and cancer risk. Critical
Reviews in Toxicology, 20:427–437, 1990.
[2] O. Karovic, I. Tonazzini, N. Rebola, E. Edström, C. Lövdahl, B. B. Fredholm,
and E. Dare. Toxic effect of cobalt in primary cultures of mouse astrocytes:
Similarities with hypoxia and role of HIF-1alfa. Biochemical pharmacology,
73:694–708, 2007.
[3] L. O. Simonsen, H. Harbak, and P. Bennekou. Cobalt metabolism and toxicology - a brief update. Science of the Total Environment, 432:210–215, 2012.
[4] National Toxocology Program. Toxicology studies of cobalt metal in F344/N
rats and B6C3F1 mice and toxicology and carcinogenesis studies of cobalt
metal in F344/NTac rats and B6C3F1/N mice. Technical Report NTH Publication No. 14-5923, National Institute of Health, Public Health service, U.S.
Department of Health and Human Services, 2013.
[5] J. R. Davis. Tool materials. ASM international, 1995.
[6] C. Hanyaloglu, B. Aksakal, and J. D. Bolton. Production and indentation
analysis of WC/Fe-Mn as an alternative to cobalt-bonded hardmetals. Materials Characterization, 47:315–322, 2001.
[7] H. Rong, Z. Peng, X. Ren, Y. Peng, C. Wang, Z. Fu, and H. Miao L. Qi.
Ultrafine WC-Ni cemented carbides fabricated by spark plasma sintering.
Materials Science and Engineering:A, 532:543–547, 2012.
[8] W. Klement Jun., R. H. Willens, and P. Duwez. Non-crystalline structure in
solidified gold-silicon alloys. Letters to nature, 187:869–870, 1960.
[9] C. Suryanarayana and A. Inoue. Bulk metallic glasses. CRC Press, 2011.
[10] http://liquidmetal.com/our-products/who-we-work-with/. Accessed: 201409-03.
[11] M. Yamasaki, S. Kagao, Y. Kawamura, and K. Yoshimura. Thermal diffusivity and conductivity of supercooled liquid in Zr41Ti14Cu12Ni10Be23 metallic
glass. Applied Physics Letters, 84:4653–4655, 2004.
47
48
Bibliography
[12] U. Harms, T. D. Shen, and R. B. Schwarz. Thermal conductivity of Pd40Ni40xCuxP20 metallic glass. Scripta Materialia, 47:411–414, 2002.
[13] C. L. Choy, W. P. Leung, and Y. K. Ng. Thermal conductivity of metallic
glasses. Journal of Applied Physics, 66:5335–5339, 1989.
[14] Sandvik Coromant. Materials for machining. Basic course, 2013.
[15] S. R. Elliott. Defects and disorder in crystalline and amorphous solids.
Springer Netherlands, 1994.
[16] H. S. Chen and D. Turnbull. Formation, stability and structure of palladiumsilicon based alloys glasses. Acta Metallurgica, 17:1021–1031, 1969.
[17] H. S. Chen. Thermodynamic considerations on the formation and stability of
metallic glasses. Acta Metallurgica, 22:1501–1511, 1974.
[18] W. H. Wang, C. Dong, and C. H. Shek. Bulk metallic glasses. Materials
Science and Engineering, 44:45–89, 2004.
[19] M. Telford. The case for bulk metallic glass. Materials today, 7:36–43, 2004.
[20] C. Suryanarayana. Mechanical Alloying and Milling. CRC Press, 2004.
[21] A. Inoue, N. Nishiyama, and H. Kimura. Preparation and thermal stability
of bulk amorphous Pd40Cu30Ni10P20 alloy cylinder of 72 mm in diameter.
Materials Transactions, JIM, 38:179–183, 1997.
[22] W. Kauzmann. The nature of the glassy state and the behavior of liquids at
low temperatures. Chemical reviews, 43:219, 1948.
[23] Carl Nordling and Jonny Österman. Physics handbook for science and engineering. Studentlitteratur, 8:4 edition, 2006.
[24] P. G. Debenedetti and F. H. Stillinger. Supercooled liquids and the glass
transition. Nature, 410:259–267, 2001.
[25] C. A. Angell. Formation of glasses from liquids and biopolymers. Science,
31:1924–1935, 1995.
[26] D. Turnbull. Under what conditions can a glass be formed? Contemporary
Physics, 10:473–488, 1969.
[27] A. Inoue. Stabilization of metallic supercooled liquids and bulk amorphous
alloys. Acta materialia, 48:279–306, 2000.
[28] N. Chen, L. Martin, D. V. Luzguine-Luzgin, and A. Inoue. Role of alloying
additions in glass formation and properties of bulk metallic glasses. Materials,
3:5320–5339, 2010.
[29] A. Gebert, J. Eckert, and L. Schultz. Effect of oxygen on phase formation and
thermal stability of slowly cooled Zr65Al7.5Cu17.5Ni10 metallic glass. Acta
Materialia, 46:5475–5482, 1998.
Bibliography
49
[30] B. S. Murty, D. H. Ping, K. Hono, and A. Inoue. Direct evidence for oxygen
stabilization of icosahedral phase during crystallization of Zr65Cu27.5Al7.5
metallic glass,. Applied Physics Letters, 76:55–57, 2000.
[31] W. H. Wang, Z. Bian, P. Wen, Y. Zhang, M. X. Pan, and D. Q. Zhao. Role
of addition in formation and properties of Zr-based bulk metallic glasses.
Intermetallics, 10:1249–1257, 2002.
[32] H. Choi-Yim and W. L. Johnson D. Xu. Ni-based bulk metallic glass formation
in the Ni-Nb-Sn and Ni-Nb-Sn-X (X=B,Fe,Cu) alloy systems. Applied Physics
Letters, 82:1030–1032, 2003.
[33] M. K. Miller and P. Liw. Bulk metallic glasses: an overview. Spinger, 2008.
[34] D. Turnbull and M. H. Cohen. Free-volume model of the amorphous phase:
glass transition. Journal of Chemical Physics, 34:120, 1961.
[35] Y. Li, S. C. Ng, C. K. Ong, H. H. Hng, and T. T. Goh. Glass forming ability
of bulk glass forming alloys. Scripta Materialia, 36:783–787, 1997.
[36] Z. P. Lu, Y. Li, and S. C. Ng. Reduced glass transition temperature and
glass forming ability of bulk glass forming alloys. Journal of Non-Crystalline
Solids, 270:103–114, 2000.
[37] X. H. Du and J. C. Huang. New criterion in predicting glass forming ability
of various glass-forming systems. Chinese Physics B, 17:No.1, 2008.
[38] S. Guo and C. T. Liu. New glass forming ability criterion derived from cooling
consideration. Intermetallics, 18:2065–2068, 2010.
[39] Z. Long, H. Wei, Y. Ding, P. Zhang, G. Xie, and A. Inoue. A new criterion for
predicting the glass-forming ability of bulk metallic glass. Journal of alloys
and compounds, 475:207–219, 2009.
[40] B-S. Dong, S-X. Zhou, D-R. Li, C-W. L, F. Guo, X. Ni, and Z. Lu. A
new criterion for predicting glass forming ability of bulk metallic glasses and
some critical discussions. Progress in natural science: Materials international,
21:164–172, 2011.
[41] C. Chattopadhyay, S. Sangal, and K. Mondal. On the unavailability of universal glass forming ability criterion. Transactions of the Indian Institute of
Metals, 67:451–458, 2014.
[42] L. J. Gallego, J. A. Somoza, and J. A. Alonso. Glass formation in ternary
transition metal alloys. Journal of Physics: Condensed Matter, 2:1990, 62456250.
[43] J. Basu, B. S. Murty, and S. Ranganathan. Glass forming ability: Miedema
approach to (Zr, Ti, Hf)-(Cu, Ni) binary and ternary alloys. Journal of Alloys
and Compounds, 465:163–172, 2008.
50
Bibliography
[44] H. Li, X. Lu, Y. Liu, R. Wu, H. Tin, M. Han, and G. Chen. Thermodynamic
calculation of glass formation for Co-ETM alloys based on Miedema’s model.
Physica B: Condensed matter, 413:24–30, 2013.
[45] M. M. Trexler and N. N. Thadhini. Mechanical properties of bulk metallic
glasses. Progress in Materials Science, 55:759–839, 2010.
[46] http://en.wikipedia.org/wiki/Shear band. Accessed: 2014-12-05.
[47] Y. Zhang Z. Xu. Quench rates in air, water, and liquid nitrogen, and interference of temperature in volcanic eruption columns. Earth and planetary
science letters, 200:315–330, 2002.
[48] C. C. Wang, C. K. Lin, Y. L. Lin, J. S. Chen, R. R. Jen, and P. Y. Lee. CuZr-Ti bulk metallic glass composites produced by mechanical alloying and
vacuum hot-pressing. Materials Science Forum, 475-479:3443–3450, 2005.
[49] H. Choi-Yim and W. L. Johnson. Bulk metallic glass matrix composites.
Applied Physics Letters, 71:3808–3810, 1997.
[50] H. Kato and A. Inoue. Synthesis and mechanical properties of bulk amorphous
Zr-Al-Ni-Cu alloys containing ZrC particles. Materials Transactions, JIM,
38:793–800, 1997.
[51] R. D. Connor, H. Choi-Yim, and W. L. Johnson. Mechanical properties
of Zr57Nb5Al10Cu15.4Ni12.6 metallic glass matrix particulate composites.
Journal of Materials Research, 14:3292–3297, 1999.
[52] H. Choi-Yim, R.D. Connor, F. Szuecs, and W. L. Johnson.
Processing, microstructure and properties of ductile particulate reinforced
Zr57Nb5Al10Cu15.4Ni12.6 bulk metallic glass composite. Acta Materialia,
50:2737–2745, 2002.
[53] H. K. Lim, E. S. Park, and J. S. Park. Fabrication and mechanical properties of WC particulate reinforced Cu47Ti33Zr11Ni6Sn2Si1 bulk metallic glass
matrix composites. Journal of Materials Science, 40:6127–6130, 2005.
[54] P. Shen, X-H. Zheng, H-J. Liu, and Q-C. Jiang. Wetting of WC by a Zr-base
metallic glass-forming alloy. Materials Chemistry and Physics, 139:646–653,
2013.
[55] I-K. Jeng and P-Y. Lee. Synthesis of Ti-based bulk metallic glass containing
WC particles. Materials Transactions, 46:2963–2967, 2005.
[56] I. A. Dobrinets, V. G. Vins, and A. M. Zaitsev. HPHT-treated diamonds.
Springer Science and Business Media, 2013.
[57] Sandvik Coromant. Sintering technology for cutting tools. Basic course, 2013.
[58] http://www.thermocalc.com/products-services/software/thermo-calc/.
cessed: 2014-11-04.
Ac-
Bibliography
51
[59] Thomas R. Dulski. A manual for the chemical analysis of metals. ASTM
International, 1996.
[60] D. B. Murphy and M. W. Davidson. Fundamentals of light microscopy and
electronic imaging. John Wiley and Sons, 2nd edition, 2013.
[61] A. Guinier. X-ray diffraction in crystals, imperfect crystals, and amorphous
bodies. Courier Dover Publications, 1994.
[62] L. Reimer. Scanning electron microscopy. Springer science and business
media, 2nd edition, 1998.
[63] H. X. Li, K. B. Kim, and S. Yi. Enhanced glass-forming ability of Fe-based
bulk metallic glasses prepared using hot metal and commercial raw material
through the optimization of Mo content. Scripta Materialia, 56:1035–1038,
2007.
[64] C. K. Lin, C. C. Hsu, R. R. Jeng, Y. L. Lin, C. H. Yeh, and P. Y. Lee.
Glass forming ability in amophous Ti50Cu35-xNi15Snx alloys prepared by
mechanical alloying. Materials Science Forum, 475-479:3451–3458, 2005.
[65] H. X. Li, J. E. Gao, Z. B. Jiao, Y. Wu, and Z. P. Lu. Glass-forming ability
enhanced by proper additions of oxygen in a fe-based bulk metallic glass.
Applied Physics Letters, 95:161905, 2009.
[66] Z. B. Jiao, H. X. Li, J. E. Gao, Y. Wu, and Z. P. Lu. Effects of alloying
elements on glass formation, mechanical and soft-magnetic properties of Febased metallic glass. Intermetallics, 19:1502–1508, 2011.
[67] J. E. Gao, H. X. Li, F. Jiang, B. Winiarski, P. J. Withers, P. K. Liaw, and
Z. P. Lu. Effects of cooling rates on glass formation and magnetic behavior
for the Fe73C7.0Si3.3B5.0P8.7Mo3.0 bulk metallic glass. Metallurgical and
Materials Transactions A, 44A:2004–2009, 2013.
[68] J. Shen, Q. Chen, J. Sun, H. Fan, and G. Wang. Exceptionally high glassforming ability of an FeCoCrMoCBY. Applied Physics Letters, 86:151907,
2005.
[69] V. Ponnambalam and S. Poon. Synthesis of iron-based bulk metallic glasses as
nonferromagnetic amorphous steel alloys. Applied Physics Letters, 83:1131–
1133, 2003.
[70] Hans Söderberg. Preparation of an amorphous alloy through gas atomization.
Technical Report, AB Sandvik Materials Technology, 2013.
[71] C. K. Lin, C. C. Hsu, R. R. Jeng, Y. L. Lin, C. H. Yeh, and P. Y. Lee.
Glass forming ability in amophous Ti50Cu35-xNi15Snx alloys prepared by
mechanical alloying. Materials Science Forum, 475-479:3451–3458, 2005.
[72] H. M Lin, C. K. Lin, R. R. Jeng, H. Y, Bor, and P. Y. Lee. Preparation and
properties of Ti50Cu28Ni15Sn7 bulk metallic glass by vacuum hot pressing.
Metallurgical and Materials Transactions A, 39A:1857–1861, 2008.
52
Bibliography
Appendix A
Thermocalc
In this section, phase diagram calculated in Thermocalc for the different binder
alloys are shown. The phase diagram was used to calculate what amount of the different elements would give the right volume fraction and composition. The carbon
content was used as variable to prevent compositions that would give unwanted
carbon phases or graphite. In the phase diagram, phases of significant volume
fractions are seen. The melt of the binder phase is denoted ”Binder”.
53
54
Thermocalc
Figure A.1. Phase diagram for approximately 80 Vol% WC and 20 Vol% binder phase
of binder alloy 1. Carbon content in the produced sample was chosen as 0.065g, which
lies in the graphite window, to ensure that no unwanted carbon phases would occur.
55
Figure A.2. Phase diagram for approximately 80 Vol% WC and 20 Vol% binder phase
of binder alloy 2. Carbon content in the produced sample was chosen as 0.43 at.%.
56
Thermocalc
Figure A.3. Phase diagram for approximately 80 Vol% WC and 20 Vol% binder phase
of binder alloy 3. Carbon content in the produced sample was chosen as 0.46 at.%.
57
Figure A.4. Phase diagram for approximately 80 Vol% WC and 20 Vol% binder phase
of binder alloy 4. Carbon content in the produced sample was chosen as 0.44 at.%.
58
Thermocalc
Appendix B
Tables
59
60
Tables
Sample
1-1
1-2
1-3
1-4
1-5
1-6
1-7
1-8
2-1
3-1
3-2
3-3
3-4
3-5
4-1
4-2
4-3
4-4
Binder
alloy
1
1
1
1
1
1
1
1
2
3
3
3
3
3
4
4
4
4
Milling
time (h)
3
3
3
24
24
24
24
24
9
30
30
30
30
30
30
30
30
30
Drying
method
Pan
Spray
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Pan
Sintering method
Vacuum (DA)
Vacuum (DA)
Vacuum (DA)
Vacuum (DA)
Vacuum (DA)
Press (GPS)
Press (GPS)
Vacuum (DA)
Vacuum (DA)
Vacuum (DA)
Vacuum (DA)
Press (GPS)
Press (GPS)
Press (GPS)
Vacuum (DA)
Vacuum (DA)
Press (GPS)
Press (GPS)
Sintering temperature (◦ C)
1450
1410
1410
1410
1500
1410
1500
1250
1450
1410
1500
1410
1500
1520
1410
1500
1410
1500
Table B.1. Milling, drying and sintering data for the produced samples.
Samples
1-1
1-2
6-1
2-1
3-1
4-1
to 1-8
to 6-4
to 3-5
to 4-4
Raw materials (at.%, if nothing else written)
Fe77 P5.1 C1.73 B1.5 Mo9 Si2.5 Cr2.6 Al0.86 Mn0.63 Ti0.5 Ni0.13 N0.002
(wt.%) (all grain sizes), WC
Fe77 P5.1 C1.73 B1.5 Mo9 Si2.5 Cr2.6 Al0.86 Mn0.63 Ti0.5 Ni0.13 N0.002
(wt.%) (-20 µm), WC
B4 C, C, Fe, Fe72.61 P27.2 Si0.16 (”FeP-stones”), Mo2 C, SiC, WC
B, C, Co, Cr3 C2 , Fe, Mo2 C, WC, Y
B4 C, C, Fe, Mo2 C, WC
Table B.2. Raw material used in the different samples.
61
Sample
3-5
3-5
3-5
3-5
4-4
4-4
4-4
4-4
4-4
Indentation
1
2
2
4
1
2
3
4
5
HV30 (kgf/mm2 )
1902
1938
1930
1956
1700
1649
1663
1649
1681
KIc (MPa · m1/2 )
6.71
6.78
6.62
6.84
8.55
8.39
8.55
8.47
8.65
Table B.3. Hardness and fracture toughness of sample 3-5 and 4-4.
Sample
2-1q
1-1q
1-2q
1-3q
3-1q
3-3q
4-1q
4-3q
Quenched sample
2-1
1-1
1-2
1-3
3-1
3-3
4-1
4-3
Heating method
Tube furnace
Tube furnace
Tube furnace
Tube furnace
Induction furnace
Induction furnace
Induction furnace
Induction furnace
Table B.4. Methods of heat treatment for the quenched pieces.
62
Tables
Appendix C
XRD
In this section XRD diffractogram of a selected number of the samples are shown.
XRD analysis was mainly used to determine differences in the binder phase before
and after quenching. In several samples, crystalline peaks from the binder phase
are not visible in the normal sized diffractogram but in enlargements of specific
regions.
63
64
XRD
Figure C.1. XRD diffractogram after full scan on sample 1-3 and 1-3q. The visible
peaks are from WC.
65
Figure C.2. Enlarged XRD diffractogram on sample 1-1 and 1-1q in range 36-48◦ .
Here, peaks that do not originate from WC are seen. It can be seen that some peaks
have decreased after quenching.
66
XRD
Figure C.3. Enlarged XRD diffractogram on sample 1-2 and 1-2q in range 36-48◦ .
Here, peaks that do not originate from WC are seen. It can be seen that some peaks
have decreased or disappeared after quenching.
67
Figure C.4. Enlarged XRD diffractogram on sample 1-3 and 1-3q in range 36-48◦ .
Here, peaks that do not originate from WC are seen. It can be seen that some peaks
have decreased or disappeared after quenching.
68
XRD
Figure C.5. Enlarged XRD diffractogram on sample 3-1 and 3-1q in range 36-48◦ . Here,
peaks that do not originate from WC are seen. It can be seen that the crystalline peaks
from the binder phase have increased and new peaks have appeared. This was due to
signals from the surface layer that was formed during heat treatment.
69
Figure C.6. Enlarged XRD diffractogram on sample 3-4ac1 and 3-4ac2 in range 36-48◦ .
The samples was heat treated by arc melting for different holding times. In both samples,
crystalline peaks from the binder phase are seen. The surface of sample 3-4ac2, on which
XRD was run, was rough and uneven, which might have caused the high peaks seen in
the diffractogram.
70
XRD
Figure C.7. Enlarged XRD diffractogram on sample 4-1 and 4-1q in range 36-48◦ . No
significant change in the crystal structure in the binder phase is seen after quenching.
Appendix D
STA
Figure D.1. DSC and TG on binder alloy 1 powder. In the TG curve a mass loss
can be detected above 1300 ◦ C. The linear increase in the TG curve is due to drift in
the system. In the DSC curve, the peak around 600 ◦ C shows the crystallization of the
amorphous powder. The dip around 1000 ◦ C shows the solid-liquid transformation. The
wide dip below 200 ◦ C is due to an intern apparatus start up effect.
71
72
STA
Appendix E
EDS
73
74
EDS
Figure E.1. EDS mapping set 1 of sample 2-1q. From the mapping of Mo (bottom
left), it is seen that Mo has dissolved into WC.
75
Figure E.2. EDS mapping set 2 of sample 2-1q. From the mapping of Ti (upper right)
regions with high concentration of Ti in the binder phase are seen.
76
EDS
Figure E.3. EDS mapping set 3 of sample 2-1q. From the mapping of P (upper right)
regions where P is missing in the binder phase are seen. In the mapping of Si (bottom
left), is is seen that those regions have an increased concentration of Si.
77
Figure E.4. EDS of sample 3-2. Formation of yttrium oxide can be seen.
78
EDS
Appendix F
LOM
79
80
LOM
Figure F.1. LOM of sample 3-2 (uppermost) and 3-4 (downmost) after η-phase etching.
81
Figure F.2. LOM of sample 4-2 (uppermost) and 4-4 (downmost) after η-phase etching.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement