1users.telenet.be/erasmus-wolverhampton/Thesis.doc · Web viewA die to produce cylindrical samples...
Transcript of 1users.telenet.be/erasmus-wolverhampton/Thesis.doc · Web viewA die to produce cylindrical samples...
PREPARATION AND CHARACTERISATION OF
SUBMICRON/NANO STRUCTURED POWDERS FROM TUNGSTEN
CARBIDE –COBALT/ALTERNATIVE BINDERS HARDMETALS
Peter ADRIAENSEN en Raf MOORS
Afstudeerwerk ingediend tot het behalen van het diploma vanindustrieel ingenieur in elektromechanica optie automatisering
master in de industriële wetenschappen: elektromechanica
Promotoren: dr. ir. T. Laoui (University of Wolverhampton) dr. ir. A. Van Bael (XIOS Hogeschool Limburg)
XIOS HOGESCHOOL LIMBURGDEPARTEMENT INDUSTRIELE WETENSCHAPPEN EN TECHNOLOGIE
A. Academiejaar 2004 - 2005
-Abstract- III
Abstract
Cobalt has been the most suitable and most commonly used binder for tungsten carbide
based hardmetals. The most important factor in favour of cobalt (Co) is its excellent
wetting behaviour for tungsten-carbide (WC).
Due to the poor corrosion resistance of Co, its high cost and environmental toxicity,
substantial research has been devoted to find suitable alternative binders for WC systems.
The aim is to reduce the amount of Co, or possibly, to completely replace Co binder. Two
promising alternatives are described and utilised in this project, the first one is a mixture
of iron (Fe), nickel (Ni) and cobalt (Co) and the second alternative is composed of iron
(Fe) and manganese (Mn). Compared to cobalt, Fe and Mn are very cheap and non toxic.
A literature review was performed on different relevant aspects covering the field of
hardmetals, powder preparation methods, powder metallurgy and nanomaterials. The
submicron/nano-structured composite powders were prepared by the mechanical alloying
method using both planetary ball and high-energy ball milling processes.
A series of experiments were performed with the planetary ball mill by varying milling
time (2.5, 5, 10 hrs) and rotation speed (250, 400rpm) parameters to process WC-10wt
%Co, WC-10%FeNiCo and WC-10%FeMn. It was noticed that as the milling time
increased (above 2.5 hours for 150rpm) the amount of elements (Fe, Cr) picked up from
the stainless steel vial inner wall increased. The contamination level increased further at a
rotation speed of 400rpm. This indicates that both speed and time should be kept low to
minimise contamination or a hard steel vial should be utilised. For that, additional powders
were prepared using the high-energy ball mill.
The grain size of WC phase was calculated using the Scherer equation and the
corresponding X-ray diffraction peaks while the WC particle size was evaluated using
Raf Moors – Peter Adriaensen
-Abstract- IV
scanning electron microscopy images. Composite powders were successfully made in
which fine WC particles (submicron down to about 200nm size) were distributed within
the matrix (Co, FeNiCo or FeMn).
The next step would be to compact such powder for a subsequent sintering process. For
that appropriate compaction dies were designed using Inventor CAD software. A die to
produce cylindrical samples for microstructural and hardness analyses was designed as
well as another die to produce samples for 3-point bending tests. Both dies were designed
according to ASTM standards.
Raf Moors – Peter Adriaensen
-Acknowledgements- V
-
Acknowledgements
The aim of our final thesis project was to prepare submicron/nano-structured powders from
WC-Co system and replace Co with suitable alternatives. This project was accomplished at
the University of Wolverhampton (UK) in line with our Master Degree Industrial Sciences.
First of all, we would like to thank everybody who helped to bring our final thesis to a
good end. A special word of thanks goes to our supervisor Dr. ir. T. Laoui and to S.
Hewitt, of the University of Wolverhampton, for enriching us with the knowledge they
have and the daily good care for us.
Further, we would like to thank Dr. ir. A. Van Bael, of the XIOS Hogeschool Limburg, for
allowing us the opportunity to accomplish our training in Wolverhampton and for reading
our final thesis project and Ms. Bauwens for helping us arrange the paperwork involving
our stay in Wolverhampton.
We would also like to thank our parents, for giving us the opportunity to do our thesis
project abroad.
Last word of thanks to everybody, especially our parents and girlfriends, for supporting us
in the difficult times we sometimes had.
Our stay at the UK was part of a project in the Erasmus framework.
Raf Moors – Peter Adriaensen
-Table of contents- 1
Table of contents
Abstract................................................................................................................................III
Acknowledgements...............................................................................................................V
Table of contents....................................................................................................................1
List of figures.........................................................................................................................6
List of tables...........................................................................................................................9
List of symbols.......................................................................................................................9
1 Introduction and project objectives..............................................................................10
1.1 Introduction..........................................................................................................10
1.2 Objectives of the thesis........................................................................................12
1.2.1 Looking for alternative binders to Co for WC particles..............................12
1.2.2 Powder Processing by Mechanical Alloying (MA).....................................12
1.2.2.1 Searching for best fit parameters for planetary ball milling....................13
1.2.2.2 Horizontal high energy simoloyer............................................................14
1.2.3 Powder characterization...............................................................................14
2 Literature review..........................................................................................................15
2.1 Hard metals..........................................................................................................15
2.1.1 Introduction..................................................................................................15
2.1.2 Powder production.......................................................................................15
2.1.3 Powder production techniques.....................................................................16
2.1.3.1 Atomization..............................................................................................16
Raf Moors – Peter Adriaensen
-Table of contents- 2
2.1.3.2 Gas- and water atomization......................................................................16
2.1.3.3 Centrifugal process...................................................................................18
2.1.3.4 Chemical processes..................................................................................18
2.1.3.5 Electrolysis...............................................................................................19
2.1.4 WC-Co.........................................................................................................19
2.1.5 Alternative binders to Co for WC................................................................20
2.1.5.1 Fe-Mn as alternative binder to Co for WC...............................................20
2.1.5.2 Fe/Ni/Co as alternative binder to Co for WC...........................................21
2.1.6 Grain growth................................................................................................21
2.1.6.1 Grain growth inhibitor..............................................................................22
2.1.6.1.1 The effect of V8C7 and Cr2C2 additives on the sintering of WC-Co. .23
2.1.6.1.2 Effect of V8C7 and Cr3C2 additions on WC-Co grain growth and
mechanical properties...........................................................................................23
2.2 Powder metallurgy...............................................................................................25
2.2.1 The process...................................................................................................25
2.2.1.1 Mix the powder with a suitable lubricant.................................................25
2.2.1.2 Powder compaction..................................................................................25
2.2.1.3 Sintering...................................................................................................25
2.2.2 Reasons for using PM..................................................................................26
2.2.3 Applications of PM......................................................................................27
2.2.3.1 Self-lubricating bearings..........................................................................27
2.2.3.2 Hard metals..............................................................................................28
2.2.3.3 Friction materials.....................................................................................28
2.2.4 The future of PM..........................................................................................29
2.3 Nanostructural materials......................................................................................29
2.3.1 What are nanostructured materials...............................................................29
2.3.2 Synthesis......................................................................................................30
2.3.2.1 Mechanical alloying.................................................................................31
2.3.2.1.1 Mechanism of alloying.......................................................................32
2.3.2.1.2 Types of mills.....................................................................................36
Raf Moors – Peter Adriaensen
-Table of contents- 3
2.3.2.1.2.1 Planetary ball mills......................................................................37
2.3.2.1.2.2 High energy ball milling.............................................................39
2.3.2.1.2.3 Other types of mills.....................................................................41
2.3.2.1.3 Process variables................................................................................43
2.3.2.1.3.1 Milling container.........................................................................43
2.3.2.1.3.2 Milling speed...............................................................................44
2.3.2.1.3.3 Milling time.................................................................................45
2.3.2.1.3.4 Grinding medium........................................................................45
2.3.2.1.3.5 Ball-to-powder weight ratio........................................................46
2.3.2.1.3.6 Extent of filling the vial..............................................................47
2.3.2.2 Liquid phase techniques...........................................................................47
2.3.2.3 Vapour phase techniques..........................................................................47
2.3.2.4 Plasma heating.........................................................................................48
2.3.2.5 Solid phase techniques.............................................................................49
2.3.2.6 Equal channel Angular Extrusion............................................................49
2.3.2.6.1 Simple shear concept..........................................................................51
2.3.2.6.2 Inhomogeneous deformation..............................................................52
2.3.3 Properties......................................................................................................53
2.3.4 WC-Co particles...........................................................................................53
2.4 Crystal structures and Point Defects....................................................................54
2.4.1 The Body-Centered-Cubic (BCC) structure.................................................54
2.4.2 The Hexagonal-Close-Packed (HCP) structure...........................................55
2.4.3 Miller indices – Cubic Crystals....................................................................56
2.4.4 Close Packed planes.....................................................................................57
2.5 Grain measurement of WC...................................................................................58
2.5.1 BET Surface Area........................................................................................59
2.5.2 X-ray sedigraph............................................................................................59
2.5.3 Laser Diffraction..........................................................................................60
2.5.4 Ultracentrifuge.............................................................................................60
2.5.5 Photon correlation spectrography................................................................61
Raf Moors – Peter Adriaensen
-Table of contents- 4
2.5.6 Microscopical image analysis, SEM, TEM..................................................61
2.5.7 X-ray line broadening..................................................................................62
2.5.8 Chemical reaction.........................................................................................62
3 Experimental procedure...............................................................................................63
3.1 Description of the powders..................................................................................63
3.1.1 Tungsten carbide (WC)................................................................................63
3.1.1.1 Tungsten carbide < 20µm........................................................................63
3.1.1.2 Tungsten carbide < 4.3µm.......................................................................64
3.1.2 Cobalt (Co)...................................................................................................64
3.1.2.1 Cobalt < 20µm.........................................................................................64
3.1.2.2 Cobalt < 4,3µm........................................................................................65
3.1.3 Iron (Fe).......................................................................................................65
3.1.4 Nickel (Ni)...................................................................................................65
3.1.5 Manganese (Mn)..........................................................................................65
3.1.6 Vanadium Carbide (VC)..............................................................................65
3.2 Preparation of the powders...................................................................................66
3.3 Milling process.....................................................................................................66
3.3.1 Planetary ball mill........................................................................................66
3.3.2 Horizontally high energy mill......................................................................67
3.3.3 Development of dies for compaction...........................................................68
3.3.3.1 Compaction die........................................................................................68
3.3.3.1.1 The die for compaction......................................................................68
3.3.3.1.2 The upper punch.................................................................................69
3.3.3.1.3 The lower punch.................................................................................69
3.3.3.1.4 The die for the Charpy test.................................................................70
3.3.3.1.5 Die for 3 point bending test................................................................70
3.4 Analysis methods.................................................................................................71
3.4.1 Sieving..........................................................................................................71
3.4.2 X-ray diffraction...........................................................................................71
Raf Moors – Peter Adriaensen
-Table of contents- 5
3.4.2.1 Formulae used in the X-ray diffraction....................................................72
3.4.2.2 Example: unit cell size from Diffraction data..........................................73
3.4.2.3 Instrumentation........................................................................................74
3.4.3 XRF..............................................................................................................75
3.4.3.1 Description of the machine......................................................................75
3.4.3.2 Preparation of the samples.......................................................................76
3.4.4 Scanning electron microscopy.....................................................................76
3.4.4.1 Description of the machine......................................................................76
3.4.4.2 Preparation of the samples.......................................................................77
3.4.5 Optical microscopy......................................................................................77
4 Results and discussion..................................................................................................78
4.1 Results from planetary ball mill...........................................................................78
4.1.1 Getting started..............................................................................................78
4.1.2 Reference sample.........................................................................................78
4.1.2.1 XRD.........................................................................................................79
4.1.2.2 XRF..........................................................................................................80
4.1.2.3 Calculated grain size of the starting WC particles...................................81
4.1.3 XRD results of the planetary ball milled samples at 250 rpm.....................82
4.1.3.1 The 2,5 hours milled sample....................................................................82
4.1.3.2 The 5 hours milled sample.......................................................................84
4.1.3.3 The 10 hours milled sample.....................................................................85
4.1.4 The SEM pictures of the 250 rpm samples..................................................86
4.1.4.1.1 The 2,5h milled sample......................................................................86
4.1.5 XRD results of the planetary ball milled samples at 400 rpm.....................87
4.1.5.1 The 5 hours milled sample.......................................................................87
4.1.5.2 The 10 hours milled sample.....................................................................88
4.1.6 The SEM pictures of the 400 rpm samples..................................................89
4.1.6.1.1 The 10h milled sample.......................................................................89
4.1.7 Contamination level.....................................................................................89
Raf Moors – Peter Adriaensen
-Table of contents- 6
4.1.8 The grain size calculation.............................................................................91
4.1.8.1 The Scherer equation................................................................................91
4.1.8.1.1 The planetary ball milled samples at 250 rpm...................................91
4.1.8.1.2 The planetary ball milled samples at 400 rpm...................................92
4.1.8.1.3 Conclusions of the Scherer equation..................................................92
4.1.9 Strain evaluation from peak shift.................................................................95
4.2 Results from the high energy mill........................................................................95
4.2.1 The estimation of the contamination level of the milled powder.................96
4.2.2 SEM pictures from samples milled with the high energy mill.....................97
4.2.2.1 SEM pictures from Fe/Mn as alternative binder......................................97
4.2.2.2 SEM pictures from Fe/Ni/Co as alternative binder..................................98
5 Conclusions and suggestions for further work...........................................................100
5.1 Introduction........................................................................................................100
5.2 Conclusions........................................................................................................101
5.2.1 The planetary ball mill...............................................................................101
5.2.2 The horizontally high energy mill..............................................................101
5.2.3 The alternative binders...............................................................................102
5.3 Further work.......................................................................................................102
6 References..................................................................................................................104
Appendix A (Website for the project)...............................................................................A-1
Appendix B (Videoconferencing facilities at UoW)..........................................................B-1
Appendix C (Technical drawings).....................................................................................C-1
Appendix D (Dutch Summary).......................................................................................D-1
Raf Moors – Peter Adriaensen
-List of figures- 7
List of figures
Figure 1: Gas atomization, EPMA, 2004..........................................................................................................18
Figure 2: Water atomization, EPMA, 2004.......................................................................................................18
Figure 3: Particle shape with gas atomization, EPMA, 2004...........................................................................18
Figure 4: Particle shape with water atomization, EPMA, 2004.......................................................................19
Figure 5: Comparison of the effect of grain refiner content on WC grain size, LUYCKX, S., 2001................25
Figure 6: Venn-diagram, GERMAN, R.M.........................................................................................................27
Figure 7: Self -lubricating bearing, EPMA, 2004.............................................................................................29
Figure 8: Friction materials, EPMA, 2004.......................................................................................................29
Figure 9: Ball-powder-ball collision of powder mixture during MA, SURYANARAYANA, C., 2001..............33
Figure 10: Narrow particle size distribution caused by tendency of small particles to weld together and
large particles to weld together and large particles to fracture, SURYANARAYANA, C., 2001......................35
Figure 11: Refinement of particle size and grain sizes with milling time, SURYANARANA,C., 2001.............36
Figure 12: Schematics of microstructural evolution during milling of a ductile-brittle combination of
powders, SURYANARAYANA, C., 2001............................................................................................................37
Figure 13: Fritch Pulverisette 5 planetary ball mill.........................................................................................38
Figure 14: Movement of balls in the planetary ball mill, Zoz...........................................................................39
Figure 15: Movement of balls in the planetary ball mill during milling, Zoz...................................................39
Figure 16: Schematic depicting the ball motion inside the ball mill, Courtesy of Gilson Company................40
Figure 17: High-energy mill, Metal-Powder....................................................................................................41
Figure 18: Running high-energy mill, Zoz........................................................................................................41
Figure 19: Inside a high-energy mill, Zoz.........................................................................................................42
Figure 20: SPEX 8000 mills in the assembled condition, SURYANARAYANA,C., 2001..................................42
Figure 21: Attritor, IMP....................................................................................................................................43
Figure 22: Working principle of the attritor, SAAR-HARDMETAL.................................................................43
Figure 23: Grinding medium, NARANG ENTERPRICES................................................................................47
Figure 24: Principle of plasma heating, Max Planck Institute.........................................................................50
Figure 25: Principle scheme of the ECAE........................................................................................................51
Figure 26: Sketch of a BCC structure, ALCAN.................................................................................................55
Figure 27: The atoms in a BCC structure, ALCAN..........................................................................................55
Figure 28: Sketch of a HCP structure, ALCAN................................................................................................56
Figure 29: Atoms in a HCP structure, ALCAN.................................................................................................56
Figure 30: Miller indices – Cubic crystals, ALCAN.........................................................................................57
Raf Moors – Peter Adriaensen
-List of figures- 8
Figure 31: Negative Miller index, ALCAN........................................................................................................57
Figure 32: Cubic lattices only a direction will be perpendicular to the plane, ALCAN...................................58
Figure 33: First plane of FCC..........................................................................................................................58
Figure 34: The second plane of FCC................................................................................................................59
Figure 35: The third layer of FCC....................................................................................................................59
Figure 36: BET surface area.............................................................................................................................60
Figure 37: X-ray sedigraph...............................................................................................................................60
Figure 38: Laser diffraction..............................................................................................................................61
Figure 39: Ultracentrifuge................................................................................................................................61
Figure 40: Photon correlation spectography....................................................................................................62
Figure 41: The beam of the photon correlation spectography..........................................................................62
Figure 42: The range were the different techniques are valid..........................................................................63
Figure 43: SEM picture of WC particles < 20 micron......................................................................................65
Figure 44: Fritsh Pulverisette 5........................................................................................................................68
Figure 45: The Maltoz operation software for the high energy mill.................................................................68
Figure 46: Assembly of the compaction die......................................................................................................69
Figure 47: Assembly of the upper punch...........................................................................................................70
Figure 48: Assembly of the lower punch...........................................................................................................70
Figure 49: Assembly of the Charpy test die......................................................................................................71
Figure 50: Assembly of the 3 point bending die................................................................................................71
Figure 51: Fritsh Sieve......................................................................................................................................72
Figure 52: Philips XRD.....................................................................................................................................73
Figure 53: Reflection of X-rays from 2 planes of atoms in a solid...................................................................74
Figure 54: X-ray diffraction pattern.................................................................................................................75
Figure 55: Schematic of an X-ray powder diffractimeter.................................................................................75
Figure 56: Spectro Xepo XRF machine.............................................................................................................76
Figure 57: ZEIS SEM........................................................................................................................................77
Figure 58: The optical microscope...................................................................................................................78
Figure 59: XRD pattern of the unmilled WC-Co reference powder sample.....................................................80
Figure 60: Graph of the contributing elements in the WC-Co reference sample.............................................81
Figure 61: SEM picture of the starting WC material < 20 micron...................................................................83
Figure 62: XRD pattern of the 2.5h ball-milled WC-Co sample @ 250 rot/min..............................................84
Figure 63: XRD pattern of the 5h ball milled WC-Co sample @ 250 rot/min.................................................85
Figure 64: XRD pattern of the 10 h ball milled WC-Co sample @ 250 rot/min..............................................86
Figure 65: SEM picture of the 2,5h planetary ball milled sample....................................................................87
Figure 66: XRD pattern of the 5h ball milled WC-Co sample @ 400 rot/min.................................................88
Raf Moors – Peter Adriaensen
-List of figures- 9
Figure 67: XRD pattern of the 10h ball milled WC-Co sample @ 400 rot/min...............................................89
Figure 68: SEM picture of the planetary ball milled sample at 400 rpm.........................................................90
Figure 69: Estimation of the contributing elements in milled powder..............................................................91
Figure 70: Contamination level in function of the milling time........................................................................91
Figure 71: The excel counting sheet for the average grain size calculation with the Scherer equation (250
rpm milled sample)............................................................................................................................................92
Figure 72: The excel counting sheet for the average grain size calculation with the Scherer equation (400
rpm milled sample)............................................................................................................................................93
Figure 73: Fourier analysis of the un-milled WC powder < 20 micron...........................................................94
Figure 74: Results of the Winfit! V1.2 on the <20 micron WC particles........................................................95
Figure 75: Grain size from the <20 micron WC particles with the Winfit! V1.2 software...............................95
Figure 76: Peak shift of WC with 250 rpm milling speed.................................................................................96
Figure 77: Estimation of the contributing elements in milled powder..............................................................97
Figure 78: contamination after milling.............................................................................................................98
Figure 79: SEM picture of the 1h horizontal energy milled sample WC-Fe/Mn..............................................99
Figure 80: SEM picture of the 1h horizontal energy milled sample WC-Fe/Ni/Co........................................100
Figure 81: Homepage of the website...............................................................................................................A-4
Figure 82: Personal data page........................................................................................................................A-4
Figure 83: Navigation page of the website......................................................................................................A-5
Figure 84: Movies page...................................................................................................................................A-6
Figure 85: The link page.................................................................................................................................A-6
Figure 86: Comparison in cost between communication possibilities............................................................B-4
Figure 87: View over the Access Grid video conference at RIATec office Wolverhampton...........................B-5
Figure 88: The micros used for the video conference.....................................................................................B-7
Figure 89: Audio component with the high-end echo canceller......................................................................B-7
Raf Moors – Peter Adriaensen
-List of tables- 10
List of tables
Table 1: The composition of the powders..........................................................................................................67
Table 2: Used compositions for the binders......................................................................................................67
List of symbols
BCC Body-Centered-Cubic
BETBrunauer, Emmett, and Teller (the three scientists that optimized the theory for measuring surface area)
Co CobaltFe Iron
H2O WaterHCl HydrochlorateHCP Hexagonal-Close-Packed
HNO3 Nitryl hydroxideMA Mechanical AlloyingMn ManganeseNi NickelPM Powder Metallurgy
SEM Scanning Elektron microscopy
TEMTransmission Electron Microscopy
UoW University of WolverhamptonVC Vanadium CarbideWC Tungsten CarbideXRD X-ray diffractionXRF X- Ray Fluorescence
Raf Moors – Peter Adriaensen
-Introduction and project objectives- 11
1 Introduction and project objectives
1.1 Introduction
Metal powders are used in industry for a diverse range of products. Some of these products
include welding electrodes, paints, printing inks and explosives. In all these products the
particles retain their identities. Traditional powder metallurgy is a process whereby a solid
metal, alloy or ceramic in a form of a mass of dry particles, normally less than 150 µm in
maximum diameter, is converted into an enginering component of a predetermined shape
possessing properties which allow it to be used in most cases without further processing.
The basic steps in traditional powder production are:
Powder production
Compaction of the powder
Sintering which involves heating the preform to a
temperature below the melting point of the major
constituent, at which point the powder particles lose their
identities through inter-diffusion processes and required
properties are developed.
Since early civilisations, there can be no doubt that certain metals in the form of powders,
and consolidated by sintering, have been forced to meet man’s needs. The iron pins keying
the marble blocks of the Parthanon in Athens are believed to have been made in this way
whilst the Delhi Pillar, dating from 300 A.D., is made of more than 6 tons of sponge iron.
With the advent of melting practices from Iron and Copper, powder metallurgy remained
unchanged for many centuries.
By the mid-19th century, a new process was developed based on the principle of fusion.
Further major industrial development in powder metallurgy had to await the advent of the
Raf Moors – Peter Adriaensen
-Introduction and project objectives- 12
electric lamp and the search for a filament material of high melting point, low evaporation
rate and adequate conductivity. Tungsten met these requirements but the sintered material
was brittle. The difficulty was resolved in 1910 when it was discovered that tungsten bars,
produced by sintering at 3000°C, developed enough ductility to permit continued working
at progressively lower temperatures. The required filament diameter could be cold drawn.
The process is still the standard method used all over the world for the production of
incandescent lamp filaments.
The mid-1920’s saw the emergence of two powder metallurgy products. That set the
patterns for the future industrial development of the technology. The first of these was a
hard, wear-resistant product known as cemented carbide. Cemented carbide is produced by
compacting and sintering a mixture of tungsten carbide and cobalt. Developed originally
for wire-drawing dies, cemented carbides have found extensive application in metal
cutting, rock drilling and hot working dies. Later on, this principle was applied to the
heavy metals.
In the late 1920’s, a second significant development was invented. It was the porous
bronze self-lubricating bearing. The powder mixture is pressed and compact sintered in a
sizing die after which it is impregnated with oil.
The first sintered iron structural parts were produced towards the end of the 1930’s. At this
point, it was recognized that the process for producing the porous bearing was a basis for
other applications. It was an important development because up until then the powder
metallurgy process had only been used to make products which could not be made by any
other technique. Now it was possible to produce new components at a relatively low cost.
This was the main reason for the success. During the past twenty years; the growth in this
industrial sector of the industry has been substantial. One of the reasons was the
contribution from the design engineer who has responded to an increasingly effective
dialogue with the powder metallurgy technologist.
Raf Moors – Peter Adriaensen
-Introduction and project objectives- 13
The powder metallurgy process is neither energy nor labour intensive; it conserves
material, is ecologically clean, and it produces components of high quality with
homogeneous and reproducible properties. In recent years, these attributes have attracted
studies into the application of the technique to the development of materials for high
technology uses. Until now, the scientific input into the process and its products had been
confined largely to the specialist areas of the refractory metals and nuclear materials.
Recently, there has been a lot of scientific research in the traditional powder metallurgy
industry, especially in the process and product developments.
A process designated “mechanical alloying”(MA) has been developed in which metal
powders are subjected to high energy milling. Carbides and other non-metallic additives
become coated with the softer metal, allowing the particles to become redistributed as a
very fine dispersion through repeated fracture and re-welding of the composite powder
particles. Originally this process was developed for alloys used in high temperature service
but lately the process has been extended to light alloys for aerospace applications and other
engineering materials [Jenkins 1993, pp. 1-5].
1.2 Objectives of the thesis
1.2.1 Looking for alternative binders to Co for WC particles
For the production of hard metals, Tungsten Carbide is the most commonly used hard
compound. During processing there has to be one compound that will turn into a liquid
phase to form a solid alloy. This compound is called the binder. The most effective binder
for most applications is cobalt. WC-Co cermets or hard metals are composites with more
than 60 vol% WC bound by a metallic phase (Co), which exhibit attractive mechanical
properties making them indispensable in a variety of industrial applications [Santhanam et
al. 1990]. To overcome the shortcomings of cobalt, namely poor corrosion resistance, high
cost and environmental toxicity [Gonzalez et al. 1995], substantial research has been
carried out to find other materials that can replace it. Investigations of binder substitutes
Raf Moors – Peter Adriaensen
-Introduction and project objectives- 14
that can replace or reduce the Co phase will be based on Fe/Ni/Co alloys. Our first mission
is to find the appropriate amounts of each element in the alloy. Other alternative binders
will be investigated, including Fe/Cu/Co which were studied by Laoui, Froyen and Kruth
[1999].
1.2.2 Powder Processing by Mechanical Alloying (MA)
The hard metal powder particles as well as the binder particles are captured in a vial
together with a number of metal balls. As the vial starts rotating, the balls collide and some
amount of powder is trapped in between them. During this process, the powder particles
are repeatedly flattened, cold welded, fractured and rewelded.
The powder particles will be plastically deformed by the force of the impact. This leads to
work hardening and fracture. The new surfaces created enable the particles to weld
together, leading to an increase in particle size in an early stage of the process. During
deformation, the particles become work hardened. These particles will fracture by a fatigue
failure mechanism. It is also possible that they fracture by the fragmentation of fragile
flakes. In the absence of strong agglomerating forces, fragments generated by this
mechanism may continue to reduce in size. At this stage, the inclination to fracture prevails
over cold welding. Due to the continued impact of grinding balls, the structure of the
particles will steadily refine.
Alloy powder compositions will be based on WC-10wt.% (Fe/Ni/Co) and WC-10wt.%
(Fe/Mn) with the addition of a grain growth inhibitor, VC at a level of 1wt.%. A reference
sample of WC-10wt% Co will be utilized.
For mechanical alloying, two different techniques are used. The first one was the planetary
ball milling whereas the second involved milling with a horizontal high energy simoloyer.
Raf Moors – Peter Adriaensen
-Introduction and project objectives- 15
1.2.2.1 Searching for best fit parameters for planetary ball
milling
The various powder combinations will be mixed with ethanol and sealed in a stainless steel
vial with stainless steel balls. Powder milling will take place in a planetary ball mill. The
influence of rotation speed and milling time will be investigated. After milling, the mixture
will be dried in an oven.
Structural analysis of the powders will involve by X-ray diffraction and optical or
Scanning Electron Microscope (SEM).
1.2.2.2 Horizontal high energy simoloyer
Because of the long milling times to achieve nanostructured grains with the planetary ball
mill, an other type of mechanical alloying is preferred. Therefore the horizontal high
energy simoloyer is recommended. With this type, nanostructures are obtained after a few
hours milling. The influence on longer milling times is investigated as well. The process is
carried out under an inert argon atmosphere.
1.2.3 Powder characterization
A very important factor for hard metals is the grain size of the hard phase. There are many
possible techniques to determine this grain size. It was our aim to find the most suitable
technique for grain size determination in our application. The first technique we have used
was X-ray diffraction (XRD) combined with several analytical analysis techniques, for
example the Sherer equation. After we have discovered that XRD is only useful for grain
size determination in a range between 0,1 and 10µm we had to search for other techniques.
These techniques are described in chapter 4.5. In our further investigation, for grains in
between that range, Scanning Elektron Microscopy (SEM) was the easiest solution.
Raf Moors – Peter Adriaensen
-Literature review- 16
2 Literature review
Before the practical experiments could take place, a thorough literature review was needed.
This chapter must be seen as a guide to place our study in the field of hard metals.
2.1 Hard metals
Hard metals are an important subject in our daily live. Therefore is interesting to get a wide
spectrum to place our study well.
2.1.1 Introduction
In many areas of engineering especially aerospace and car manufacture, demands for
higher functionality and reliability are leading to the increasing use of materials with
specially tailored properties. Examples of these properties include low wear under
exposure to corrosive media, high heat resistance, increased mechanical strength and low
specific gravity. Materials which can be mentioned in this context are sintered, hard, wear-
resisting materials based on the carbides of one or more of the elements tungsten, tantalum,
titanium, molybdenum, niobium and vanadium, bonded with a metal of lower melting
point usually cobalt. Tungsten carbide is the most widely used.
2.1.2 Powder production
The first step in powder technology is the production of powders. For example, if there is a
need for an alloy of tungsten and carbide, tungsten powder and carbon (graphite) have to
be mixed. This mixture is tungsten-carbide (WC).
In mechanical alloying a suitable binder in powder form is necessary. The most common
used binder for WC is cobalt (Co).
Raf Moors – Peter Adriaensen
-Literature review- 17
2.1.3 Powder production techniques
Many different techniques are used to produce powders. A brief description of some
common used production techniques will follow below. For further reading we recommend
specific literature.
2.1.3.1 Atomization
In this process molten metal is broken up into small droplets and rapidly frozen before the
drops come into contact with each other or with a solid surface.
The principal method is to disintegrate a thin stream of molten metal by using the impact
of high energy jets of gas or liquid.
Air, nitrogen and argon are commonly used gases, and water is the most widely used
liquid.
Atomization is particularly useful for the production of alloys in powder form, since the
metals are fully alloyed in the molten state. That way each powder particle has the same
chemical composition [EPMA 2004].
2.1.3.2 Gas- and water atomization
The particle size can be controlled by changing several parameters:
design and configurations of the jets,
pressure and volume of the atomizing fluid,
thickness of the stream of metal etc.
Raf Moors – Peter Adriaensen
-Literature review- 18
A schematic of the vertical gas atomization and water atomization technique are shown in
figures 1 and 2.
Figure 1: Gas atomization, EPMA, 2004
Figure 2: Water atomization, EPMA, 2004
The rate of solidification mainly determines the particle shape. This varies from spherical,
if a low heat capacity gas is employed, to highly irregular if water is used. In principle the
technique is applicable to all metals that can be melted. Results of gas and water
atomization are displayed in figures 3 and 4 [EPMA 2004].
Raf Moors – Peter Adriaensen
-Literature review- 19
Figure 3: Particle shape with gas atomization, EPMA, 2004
Figure 4: Particle shape with water atomization, EPMA, 2004
2.1.3.3 Centrifugal process
Rotation is used to accelerate and disintegrate the melt by centrifugal atomization. There
are several different variants based on this principle. The most important difference
between them is the choice of vacuum or protective atmosphere. A vacuum atmosphere
limits the methods of heating to electron-beam melting, whilst using argon or helium
permits heating by arc or plasma. Further on, the flight path required to complete
solidification is much longer in vacuum than in an inert atmosphere, which requires special
equipment considerations [EPMA 2004].
Raf Moors – Peter Adriaensen
-Literature review- 20
2.1.3.4 Chemical processes
Thermal decomposition of a chemical compound is used in some cases, mainly in nickel
carbonyl.
In the beginning, it was a process developed for refining nickel and raw metals which
caused a selective reaction between carbon monoxide and nickel or another raw material
under pressure, resulting in the formation of carbonyl. Carbonyl is a gaseous substance
produced at the reaction temperature. This gas decomposes at a higher temperature and a
lower pressure.
The same process is used for iron, and carbonyl iron powder finds small scale application
where its very high purity is useful.
Typically the particle size of carbonyl iron powder is 1 - 5 µm, but, as in the case of nickel,
it can be tailored to suit particular requirements.
In the Sherritt-Gordon process, nickel powder is made by hydrogen reduction of a solution
of a nickel salt under pressure [EPMA 2004].
2.1.3.5 Electrolysis
By choosing suitable conditions, composition, strength of the electrolyte, temperature,
current density, etc., many metals can be deposited in a spongy or powdery state. However
it is possible that some extensive washing, drying, reducing, annealing and crushing may
be required.
Copper is the main metal to be produced in this way but chromium and manganese
powders can also be produced by electrolysis with these powders. However, a dense and
normally brittle deposit is formed and requires crushing from powders.
Raf Moors – Peter Adriaensen
-Literature review- 21
Electrolytic iron was at one time produced on a substantial scale but it has been largely
superseeded by powders made by less costly processes. Very high purity and high density
are two distinguishing features of this process [EPMA 2004].
2.1.4 WC-Co
For the production of hard metals, tungsten carbide (WC) is the basic and most widely
used hard compound, whereas cobalt was found to be the best binder material for most
applications. Tungsten alloys and hard metals are essentially composite materials
consisting of a refractory or hard phase bound by a more ductile matrix [Nishiyama 1977].
The reason for selecting the WC-Co system is twofold: it is a classic system, exhibiting
excellent wetting properties between the two phases and it has attractive features for
applications (prototyping of cutting tools, mould inserts and cavities) [Honeycombe 1981,
pp.35-42]. However there are also disadvantages with cobalt as the binder, namely the high
cost, poor corrosion resistance and environmental toxicity [Gonzalez et al., 1995].
Therefore significant effort has been done to replace cobalt as a binder in the system.
2.1.5 Alternative binders to Co for WC
There are different reasons (cost, toxicity and corrosion resistance) to search for alternative
binders to Co for WC and in this part a few alternative binders are given.
2.1.5.1 Fe-Mn as alternative binder to Co for WC
The properties of WC-hard metals are determined by the carbides, as well as the binder
metals and can be widely varied through WC-content, WC-grain size and alloying
additions. The carbides are responsible for properties such as hardness and wear resistance,
but need to be bonded by metals or alloys or provide toughness and strength in the
composite material.
Raf Moors – Peter Adriaensen
-Literature review- 22
Iron–manganese alloys show characteristics similar to Co with regard to melting
temperature, crystal structure, and γ(fcc) → ε(hcp) phase transformation on cooling and
because of the high wear resistance of Fe-Mn steel, it was felt that Fe-Mn alloys could
provide a high wear resistance when used as a binder for WC-based hard metals
[Hanayaloglu et al. 2001, pp. 315-322].
Mn, like nickel (Ni), is an austenite stabilizer, but when alloyed with Fe, it is
approximately twice as effective as Ni in stabilizing the austenite phase down to room
temperature. Fully or partially austenitic structures can be retained at room temperature
with Fe–Mn alloys containing 13.5 wt % Mn, whereas a Fe-Ni alloy would require 30 wt
% Ni [Shimizu and Tanaka 1978, pp. 685-693 – Schumann 1967, pp. 275-283].
Within the intermediate (10–15 wt.%) Mn range, the Fe-Mn alloy system also exhibits
transformation characteristics of pure Co. That is, the alloys undergo a γ(fcc) → ε(hcp)
martensitic phase transformation on cooling that may well create stacking faults by a
deformation-induced transformation similar to that which is claimed to occur in Co
[Nishiyama 1977].
An additional attraction of Fe-Mn alloys is the possibility of creating a binder phase with
its own inherent wear resistance due to carbon being dissolved from the WC into the
binder phase during the sintering [Hanayaloglu et al. 2001, pp. 315-322].
2.1.5.2 Fe/Ni/Co as alternative binder to Co for WC
In the early days of hard metal history, especially in the German hard metal industry during
and after the Second World War some efforts were made to produce hard metals with Fe-
based binders.
Raf Moors – Peter Adriaensen
-Literature review- 23
These hard metals never entered the market successfully and disappeared quickly due to
detrimental properties or difficulties in producing stable qualities. In the late 1970’s,
Prakash made the first, well-founded investigations on a broad range of (Fe/Ni/Co)-alloys
and showed that hard metals with such Fe-rich binders have improved properties such as
higher hardness, abrasive wear resistance, toughness and strength compared to Co bonded
hardmetals [Prakash 1993, pp. 80-109 – Prakash 1979].Binders based on Fe-Ni/Ni-Co and
Fe-Cu-Co alloys are reported to be good Co substitutes for cemented carbides and diamond
tools [Gonzalez et al., 1995] [Gonzalez et al., 1998].
From previous investigations, the binder composition 9 wt% (75wt%Fe//15wt%Ni/10wt
%Co) gives results similar to the Co binder [Gille et al. 1999]. However the porosity
increases a little bit when a mixed binder is used instead of a pre-alloyed binder and the
Vickers hardness increases in comparison of the Co binder while the density stays the
same.
Another composition that has been investigated by Laoui, Froyen and Kruth [1999] for the
effect on selective laser sintering process, was the binder composition of 9,5 wt%(54wt
%Fe/28wt%Ni/18wt%Co).
2.1.6 Grain growth
One of the most important factors is grain growth in the hardmetal powders and can be
explained as the phenomenon of increasing grain size of the hard metal particles (WC in
this case) after sintering. This has a negative effect because the purpose of mechanical
alloying is to achieve a homogeneous structure. That’s why inhibitors are usually added to
decrease the size of the binder grains.
The purity of the starting materials is also a very important factor that has to be known
before any experiments can commence.
Raf Moors – Peter Adriaensen
-Literature review- 24
2.1.6.1 Grain growth inhibitor
There are different kinds of grain inhibitors and the efficiency of the inhibitor depends on
the binder that is used. Therefore, grain inhibitors are discussed for a range of possible
binders and different preparations of the powders, starting with cobalt.
There has to be a difference between high-carbon alloys and low-carbon alloys because the
grain growth inhibitor has a different efficiency in these systems [Brookes K.J.A. 1998, pp.
76-78].
After the WC powder has been compounded, the total composition of the powder has to be
made. The powder that is used in this thesis follows the formula (90 – x) wt% WC, 10%
Co and x wt% grain inhibitor. From previous works, it is already known that 1 wt% grain
inhibitor gives a good effect so x equals 1 in the following experiments [Brookes K.J.A.
1998, pp. 76-78].
Before high energy milling can begin, a decision has to be made whether dry milling or
wet milling should be used. The pros and cons will be explained briefly in the following
points [Brookes K.J.A. 1998, pp. 76-78].
There is a spectrum of possibilities in the sort of grain refiner but the most effective grain
growth inhibitors for the WC-Co system have been found to be V8C7 and Cr2C2 [Brookes
K.J.A. 1998, pp. 76-78].
Raf Moors – Peter Adriaensen
-Literature review- 25
2.1.6.1.1 The effect of V8C7 and Cr2C2 additives on the sintering of
WC-Co
There is still a debate on whether it is more effective to add V8C7 and or Cr3C2 powders to
WC and Co powders at the milling stage or to add the oxides of vanadium or chromium to
W before carburization. The later option has the advantage that oxides are cheaper than
carbides and that the V8C7 and Cr3C2 produced during carburization together with WC are
better dispersed [Brookes K.J.A. 1998, pp. 76-78]. However, nowadays most producers do
not produce their own WC and their only option is to add V8C7 and/or Cr3C2 at the milling
stage.
Both carbides are added to WC-Co in very small amounts, typically not exceeding 1 wt%.
These small amounts dissolve totally in the cobalt during sintering, but it has been found
that during cooling V8C7 precipitates as nanoparticles of (V,W)C [Gille et al. 1999] while
Cr3C2 remains in solution in the cobalt, although it tends to diffuse towards the WC-Co
bounderies [Okada et al. n.d.]
Both V8C7 and Cr2C3 have been found to lower the onset temperature of the eutectic
reaction in WC-Co. In the case of V8C7 it was found to decrease by 25-30°C [Lucyckx and
Alli 2000, pp. 507-510] and in the case of Cr3C2 by approximately 40°C [Luyckx et al.
1996, pp. 39-41].
2.1.6.1.2 Effect of V8C7 and Cr3C2 additions on WC-Co grain growth
and mechanical properties
It is generally agreed that V8C7 is the most effective grain growth inhibitor, since smaller
amounts are required to achieve equal inhibiting effects. Figure 5 shows a comparison
between the grain size obtained in sintered WC-10 wt% Co using equal amounts of V8C7
and Cr3C2.
Raf Moors – Peter Adriaensen
-Literature review- 26
However, there does not appear to be an agreement on the effect of the two inhibitors on
the properties of the sintered material. Some authors claim that additions of V8C7 harden
the material more than additions of Cr3C2 , while other authors claim the opposite
[Lucyckx and Alli 2000, pp. 507-510].
Figure 5: Comparison of the effect of grain refiner content on WC grain size, LUYCKX, S., 2001
As far as resistance to fracture is concerned, the general belief among hardmetal producers
is that V8C7 embrittles WC-Co more than Cr3C2 does, and that the embrittlement is due to
the intrinsic brittleness of V8C7. The parameter controlling crack resistance is the binder
mean free path, which, at constant Co content, is proportional to the WC grain size. Thus
for equal inhibitor additions V8C7 lowers the crack resistance more than Cr3C2 because it
produces finer grain sizes and narrower mean-free paths [Lucyckx and Alli 2000, pp. 507-
510].
Since Cr3C2 remains in solution while V8C7 precipitates during cooling, Cr3C2 increases the
total binder content and hardens the binder. On account of this, alloys containing Cr3C2
Raf Moors – Peter Adriaensen
-Literature review- 27
reach the same hardness at larger WC grain sizes and for equal hardness have higher crack
resistance than alloys containing V8C7 [Lucyckx and Alli 2000, pp. 507-510].
The main advantage of adding V8C7 as a grain refiner, however, is that it can aid the
development of extremely fine grain sizes, which can be obtained only by adding much
larger amounts of Cr3C2 [Luyckx and Alli 2000, pp. 507-510].
2.2 Powder metallurgy
Among the various metalworking technologies, powder metallurgy is the most diverse
manufacturing approach. One attraction to powder metallurgy (PM) is its ability to
fabricate high quality, complex parts to close tolerances in an economical manner [German
n.d., pp. 6-7].
2.2.1 The process
The process can be seen as three different steps and they will be explained in the following
points.
2.2.1.1 Mix the powder with a suitable lubricant
In this step, the powder will be mixed with a suitable lubricant, e.g. zinc stearate. The only
reason why a lubricant is added is to reduce friction during the compaction process.
The powders could be produced by the processes described in 4.1.2 earlier in this thesis.
2.2.1.2 Powder compaction
Raf Moors – Peter Adriaensen
-Literature review- 28
After applying pressure on the milled powder, a compacted part is produced. This part
requires only sufficient cohesion to enable it to be handled safely and transferred to the
next stage. Such compacts are referred to as green which means unsintered. The terms
green density and green strength are used to describe the green state.
2.2.1.3 Sintering
Sintering takes place in a protective atmosphere under vacuum. The temperature will be
controlled below the melting point of the main constituent so that the powder particles
weld together and confer sufficient strength to the object for its intended use.
This process is called sintering. In certain cases a minor constituent becomes molten at the
sintering temperature in which case the process is referred to as liquid phase sintering. The
constituent that becomes molten is referred to as the binder.
The amount of liquid phase must be limited so that the part (green state) retains its shape.
In certain special cases compaction and sintering are combined at an elevated temperature
such that sintering occurs during the process. This is referred to as hot pressing or pressure
sintering [EPMA 2004].
2.2.2 Reasons for using PM
Three overlapping categories provide an introductory concept for the reasons for using
PM. Figure 6 is a Venn diagram showing how the applications for PM can be categorized.
First are the many applications which rely on the low cost production of complex parts.
Components for the automotive industry represent good examples of this area and their
production is a large PM activity. Within the area of economical parts production come
concerns with productivity, tolerances and automation.
Raf Moors – Peter Adriaensen
-Literature review- 29
Figure 6: Venn-diagram, GERMAN, R.M.
Contrasting the powder route with fusion metallurgy (casting), both the precision and cost
are very attractive. Furthermore, with casting there are problems and costs associated with
segregation, machining and final tolerances which can be avoided with metal powder-
based approaches.
As figure 6 shows, there are also unique properties or unique microstructure justification
for using PM approaches. Some examples include porous metals, oxide dispersion
strengthened alloys, cermets and cemented carbides.
The inability to fabricate these unique microstructures by alternative techniques has
contributed a large part to the growth of PM.
The final circle shown on the Venn-diagram corresponds to captive applications. These are
the materials which are quite difficult to process by any other techniques. Ideal examples
are the reactive and refractory metals where fusion techniques are not practical. Another
rapidly emerging group of materials are the amorphous or glassy metals. In many cases, it
Raf Moors – Peter Adriaensen
-Literature review- 30
is desirable to form a powder and develop low temperature processing to avoid the
microstructural damage accompanying elevated temperatures. PM techniques are attractive
since all of the processing can be performed in the solid state.
Usually elements from all three categories shown in figure 6 exists in most practical PM
applications. Indeed the major growth and expansion will most likely com from further
combinations of these three elements in forming unique, low cost, high quality products.
2.2.3 Applications of PM
In this section, a few applications of PM are given.
2.2.3.1 Self-lubricating bearings
Self-lubricating bearings are typically made of bronze. The starting materials may be
mixed from elemental powders of copper and tin, fully pre-alloyed bronze powder, or
mixtures of the three [EPMA 2004].
These kinds of bearings are largely used in the automotive industry. Figure 7 gives a
picture from self-lubricating bearings.
Figure 7: Self -lubricating bearing, EPMA, 2004
Raf Moors – Peter Adriaensen
-Literature review- 31
2.2.3.2 Hard metals
Hard metals are widely used as high precision cutting tools because of their good wear-
resistance.
2.2.3.3 Friction materials
Sintered metal friction components are particularly useful for heavy-duty applications, e.g.
aircraft breaks, heavy machinery clutch and brake linings etc.
They consist essentially of a continuous metal matrix, into which varying amounts of non-
metallic friction generators, such as silica and emery are bonded [EPMA 2004].
A picture of friction materials is given in figure 8.
Figure 8: Friction materials, EPMA, 2004
Raf Moors – Peter Adriaensen
-Literature review- 32
2.2.4 The future of PM
The past successes of PM have derived from the economic benefits. In the more recent
times, the unique/difficult to process materials have contributed tot the expansion in the
technological base. These same attributes are expected to continue to come together to
form new applications for PM. Five areas appear to hold the necessary ingredients for
continued growth:
1. High volume production of precise, high quality structural parts, typically from
ferrous based alloys.
2. Difficult-to-process materials, where fully dense high performance alloys can be
fabricated with uniform microstructures.
3. Specially alloys, typically composites containing mixed phases.
4. The nonequilibrium materials such as amorphous, microcrystalline or metastable
alloys.
5. The complex parts possessing unique and uncommon shapes or ingredients.
The future promises more challenges with the combination of cost savings and factors such
as reliability, quality, strength, dimensional control and unique shaping capabilities. The
appreciation of these advantages will provide both economic and technological growth
opportunities [German n.d., p. 7].
2.3 Nanostructural materials
2.3.1 What are nanostructured materials
o a broad class of materials, with microstructures modulated in zero to three
dimensions on length scales less than 100 nm.
Raf Moors – Peter Adriaensen
-Literature review- 33
o materials with atoms arranged in nanosized clusters, which become the
constituent grains or building blocks of the material
o any material with at least one dimension in the 1-100nm range
Conventional materials have grains sizes ranging from microns to several millimeters and
contain several billion atoms each. Nanometer sized grains contain only about 900 atoms
each. As the grain size decreases, there is a significant increase in the volume fraction of
grain boundaries or interfaces. This characteristic strongly influences the chemical and
physical properties of the material. For example, nanostructured ceramics are tougher and
stronger than the coarser grained ceramics. Nanophase metals exhibit significant increases
in yield strength and elastic modulus.
The growing demand for nanopowders arises from the change in physical, chemical and
electrical properties exhibited by particles when their size falls below about 100 nm. The
laws of quantum physics, rather than the laws of classical physics, come into play at these
small particle sizes and the behaviour of the surfaces start to dominate the bulk behaviour
of the material. For example, materials that would normally be conductors of electricity
can become insulators at the nanoscale, or vice versa [Pritchard 2004].
Using a variety of synthesis methods, it is possible to produce nanostructured materials in
the following forms: thin films, coatings, powders and as a bulk material [Nanostructured
Materials 2004].
2.3.2 Synthesis
There are two basic approaches to the production of nanomaterials [Nanostructured
Materials 2004]. The first is the ‘top-down’ approach, which involves the breaking down
of the bulk material into nanosized structures or particles. These techniques are an
extension of those that have been used for producing micron-sized particles. An example
Raf Moors – Peter Adriaensen
-Literature review- 34
of this technique is mechanical alloying. The alternative approach, which has the potential
of creating less waste and is more economical, is the ‘bottom-up’ approach. In this
approach individual atoms or molecules are built up to form the require nanostructure or
nanoparticles. Many of these techniques are still under development or are just beginning
to be used for commercial production of nanopowders. The major technical difficulties to
overcome in developing a successful bottom-up approach is controlling the growth of the
particles and then stopping the newly formed particles from agglomerating. Other technical
issues are ensuring the reactions are complete so that no unwanted reactant is left on the
product and completely removing any growth aids that may have been used in the process.
Bottom-up approaches, in current use or at an advanced stage of development, can be
classified into liquid, vapour or solid phase techniques [AP Materials Inc n.d.].
2.3.2.1 Mechanical alloying
Pure powders with particle sizes in the range of 1 to 200 μm, the so called raw materials,
are widely commercially available. The only consideration that has to be made is that the
powder particle size is smaller than the grinding ball size, so the starting size of the powder
isn’t important. The explanation therefore is that the powder particle size decreases
exponentially with time, so it will reach a small value of a few microns after only a few
minutes of milling.
In the early days of MA, the powder charge consisted of at least 15 vol% of a ductile,
compressible and deformable metal powder to act as a binder. However mixtures of fully
brittle materials have been milled successfully to form an alloy [Koch 1991, pp. 193-245].
Therefore, it is no longer necessary to have a ductile metal powder during milling [Ivanov
1992, pp. 475-480].
As a result, for the production of novel alloys ductile-ductile, ductile-brittle, and brittle-
brittle powder mixtures are milled. Mixtures of solid powder particles and liquids have also
Raf Moors – Peter Adriaensen
-Literature review- 35
been milled in recent times. Milling metal powders with a liquid medium is termed wet
grinding. When there is no liquid involved in the milling process it is referred to as dry
grinding [Okada et al. 1992, pp. 862-864 - Bellosi et al. 1997, pp. 255-260].
Wet grinding has been reported as a more suitable method to obtain finer-ground products
than dry grinding because the solvent molecules are adsorbed on the newly formed
surfaces of the particles and lower their surface energy. Another advantage is the less-
agglomerated condition of the powder particles in the wet condition. It has been reported
that the rate of amorphization is faster during wet grinding than during dry grinding
[Dolgin et al. 1986, pp. 281-289]. A disadvantage of the wet grinding, however, is the
increased contamination of the powder.
Raf Moors – Peter Adriaensen
-Literature review- 36
2.3.2.1.1 Mechanism of alloying
During high-energy milling, the powder particles are repeatedly flattened, cold welded,
fractured and rewelded. Whenever two steel balls collide, some amount of powder is
trapped in between them. During each collision, there are around 1000 particles with an
aggregate weight of 0,2 mg trapped [Suryanarayana 2001d, p.32] (figure 9).
Figure 9: Ball-powder-ball collision of powder mixture during MA, SURYANARAYANA, C., 2001
The powder particles will be plastically deformed by the force of the impact. This leads to
work hardening and fracture. The new surfaces created enable the particles to weld
together, leading to an increase in particle size. In the early stages of milling, the particles
are soft and as a result, their tendency to weld together and form large particles is high.
This statement is only valid when using either a ductile-ductile or ductile-brittle material
combination. A broad range of particle sizes develops. Some of them are as large as three
times the starting particles. The composite particles at this stage have a characteristic
layered structure. This structure consists of various combinations from the starting
constituents. During deformation, the particles become work hardened. These particles will
fracture by a fatigue failure mechanism. It is also possible that they fracture by the
fragmentation of fragile flakes. In the absence of strong agglomerating forces, fragments
generated by this mechanism may continue to reduce in size. At this stage, the inclination
Raf Moors – Peter Adriaensen
-Literature review- 37
to fracture prevails over cold welding. Due to the continued impact of grinding balls, the
structure of the particles will steadily refine. However, the particle size remains to be the
same. As a consequence, the inter-layer spacing decreases and the number of layers in a
particle increase.
Nevertheless, it should be remembered that the efficiency of particle size reduction is very
low. The efficiency in a conventional ball mill is about 0.1%. It is possible that in high-
energy ball milling processes the efficiency may be somewhat higher, but it is still less
than 1% [Suryanarayana 2001c, pp. 32-33].
The remaining energy is lost, mostly in the form of heat. However, a small amount is also
utilized in the elastic and plastic deformation of the powder particles.
Steady-state equilibrium is attained when a balance is achieved between the rate of
welding and the rate of fracturing after milling for a certain length of time. The rate of
welding tends to increase the average particle size [Benjamin 1990, pp. 122-127] while the
rate of fracturing tends to decrease the average composite particle size. Smaller particles
are able to resist deformation without fracturing. They also tend to be welded into larger
pieces. The overall tendency is to drive both, the very fine and the very large particles,
towards an intermediate size. At this stage, each particle contains essentially all of the
starting ingredients together. Due to the accumulation of strain energy, the particles reach
their saturation hardness. The particle size distribution at this stage is narrow. The reason is
that particles larger than average are reduced in size at the same rate. These particles split
smaller than the average grow trough agglomeration of smaller particles [Lee et al. 1998,
pp. 235-239]. From the forgoing, it is clear that during MA heavy deformation is
introduced into the particles.
This is shown by the attendance of a variety of crystal defects such as:
dislocations,
Raf Moors – Peter Adriaensen
-Literature review- 38
vacancies,
stacking faults,
and increased number of grain boundaries.
Figure 10: Narrow particle size distribution caused by tendency of small particles to weld together and large
particles to weld together and large particles to fracture, SURYANARAYANA, C., 2001
The presence of this defect structure enhances the diffusivity of solute elements into the
matrix. Further, the refined microstructural features decrease the diffusion distances. The
slight rise in temperature during milling further aids the diffusion behaviour.
Consequently, true alloying takes place amongst the constituent elements. While this type
of alloying generally takes place nominally at room temperature, it may sometimes be
necessary to anneal the mechanically alloyed powder at an elevated temperature for
alloying to be achieved. This is particularly true when formation of intermetallics is
desired.
To develop a given structure in any system, you require specific times. These times would
be a function of the initial particle size and characteristics of the ingredients as well as the
specific equipment used for conducting the MA operation and the operating parameters of
the equipment. In most cases, the rate of refinement of the internal structure (particle size,
Raf Moors – Peter Adriaensen
-Literature review- 39
crystallite size, lamellar spacing, etc.) is roughly logarithmic with processing time and
therefore the size of the starting particles is relatively unimportant. Lamellar spacing
usually becomes small and the crystallite (or grain) size is refined to nanometer
dimensions. These dimensions can be acquired in a few minutes to an hour. That way, the
production of nanostructured materials is easy. This is one reason why MA has been
extensively employed to produce nanocrystalline materials [Koch 1993, pp. 109-129 –
Suryanarayna 1995b, pp. 41-64].
As mentioned above, it is possible to conduct MA of three different combinations of
metals and alloys:
ductile-ductile,
ductile-brittle,
brittle-brittle systems.
The ductile-brittle combinations are the most important ones because WC-Co and the
possible alternative binders for Co belongs to this combination. This combination is going
to be investigated.
More information for the other two combinations can be found in:
BENJAMIN, J.S. (1990) Metal Powder Rep. 45, pp. 122-127.
GILMAN, P.S. and BENJAMIN, J.S. (1983) Annu Rev Mater Sci. 13, pp.279-300.
DAVIS, R.M. and KOCH, C.C. (1987) Scripta Metall. 21, pp. 305-310.
DAVIS, R.M., McDERMOTT, B. and KOCH, C.C. (1988) Metall Trans. A19, pp.
2867-2874.
Raf Moors – Peter Adriaensen
-Literature review- 40
Figure 11: Refinement of particle size and grain sizes with milling time, SURYANARANA,C., 2001
The traditional ODS (oxide dispersion strengthened) alloys belong to this category because
the brittle oxide particles are dispersed in a ductile matrix. The microstructural evolution in
this type of system was also described by Benjamin and others [1983, pp. 279-300]. In the
initial stages of milling, the ductile metal powder particles get flattened by the ball-
powder-ball collisions, while the brittle oxide or intermetallic particles get
fragmented/comminuted (Fig. 12a). These fragmented brittle particles tend to become
occluded by the ductile constituents and trapped in the ductile particles. The brittle
constituent is closely spaced along the interlamellar spacings (Fig. 12a).
With further milling, the ductile powder particles become work hardened and the lamellae
get convoluted and refined (Fig.12b). The composition of the individual particles
converges toward the overall composition of the starting powder blend. With continued
milling, the lamellae get further refined, the interlamellar spacing decreases, and the brittle
particles get uniformly dispersed, if they are insoluble as in an ODS alloy (Fig. 12c).
On the other hand, if the brittle phase is soluble, alloying also occurs between the ductile
and brittle components also. That way chemical homogeneity is achieved. Whether
alloying occurs or not in a ductile-brittle system also depends on the solid solubility of the
brittle component in the ductile matrix. If a component has a negligible solid solubility
Raf Moors – Peter Adriaensen
-Literature review- 41
then alloying is unlikely to occur, e.g., boron in iron. Thus, alloying of ductile-brittle
components during MA requires not only fragmentation of brittle particles to facilitate
short-range diffusion, but also reasonable solid solubility in the ductile matrix component.
Figure 12: Schematics of microstructural evolution during milling of a ductile-brittle combination of
powders, SURYANARAYANA, C., 2001
2.3.2.1.2 Types of mills
Different types of milling equipment are used to produce mechanically alloyed powders
[Suryanarayama 2001a, pp. 1-184].There is a big difference between laboratory equipment
and milling devices for commercial use.
Raf Moors – Peter Adriaensen
-Literature review- 42
They differ in their:
capacity,
efficiency of milling
additional arrangements for cooling,
heating
2.3.2.1.2.1 Planetary ball mills
A popular mill for MA is the planetary ball mill as shown in figure 13.
Figure 13: Fritch Pulverisette 5 planetary ball mill
Its name is derived from the movement of the vials that describe a planet-like movement.
While the support disk performs a circular movement, a special drive mechanism causes
the vials to rotate around their own axes. The centrifugal forces caused by the rotating
movements will act on the contents of the vials. Since the vials and the supporting disk
Raf Moors – Peter Adriaensen
-Literature review- 43
rotate in opposite directions, the centrifugal forces alternately act in like and opposite
directions as shown in figure 13. This causes the grinding balls to run down the inside wall
of the vial which is called the friction effect. Hereby the material will be ground; the
grinding balls will lift off and travel freely through the inner chamber of the vial. Their
flight will be ended as they bump against the opposing inside wall, termed as the impact
effect.
The Fritch Pulverisette has the possibility of changing the movement direction during the
milling process.
Figure 14: Movement of balls in the planetary ball mill, Zoz
Figure 15: Movement of balls in the planetary ball mill during milling, Zoz
Raf Moors – Peter Adriaensen
-Literature review- 44
Figure 14 and 15 displays the movement of the balls inside of the vials when the planetary
ball mill is working.
Grinding vials and balls are available in different materials, including tungsten-carbide,
agate, silicon-nitride, sintered corundum, zirconia, chrome steel, Cr-Ni steel, and plastic
polyamide [Suryanarayama 2001a, pp. 1-184].
Figure 16: Schematic depicting the ball motion inside the ball mill, Courtesy of Gilson Company
2.3.2.1.2.2 High energy ball milling
An example of the top-down technique is high-energy ball milling. Often an abrasive is
added to the process to aid the milling process and milling times can vary from several
hours up to many days. In general, increased energies and milling times result in decreased
particle sizes. An alternative technique is cryogenic milling, where the material is first
cooled to a low temperature to make it more brittle and easier to break down by milling.
The difficulty with top-down approaches is ensuring all the particles are broken down to
the required particle size. Milling typically results in a Gaussian size distribution with a
long ‘tail’ representing the un-milled product. Furthermore, longer milling times will result
in more milling impurities, which together with any milling aids, if used, can be difficult to
remove. The Simoloyer is a horizontal high-energy ball mill and is known from academic
Raf Moors – Peter Adriaensen
-Literature review- 45
as well as industrial applications in mechanical alloying, high-energy milling and reactive
milling.
These devices supply the highest relative velocity of grinding media, a high level of kinetic
energy transfer, an intensive grinding effect and short processing times. The contamination
of the processed powders by the milling tools is naturally lower since the process is based
on the collision of grinding media.
Since a horizontally arranged rotor inside the grinding vessel accelerates the grinding
media, these devices do not have to move unnecessarily any large masses like the entire
chamber/mill in case of vibration or planetary ballmills.
Large-scale systems of several hundred litres volume and are economically and
ecologically favourable.
Figure 17: High-energy mill, Metal-Powder
Atmosphere and cooling seems non-problematic as these mills can be operated, loaded and
unloaded under vacuum or inert gas and are equipped with efficient cooling or cooling and
heating systems. Figure 17 shows a real image of a horizontal high-energy ball mill in
operation.
Raf Moors – Peter Adriaensen
-Literature review- 46
The effect of collision in MM is visualised in figures 18 and 19 and the working principle
can be imagined where the rotor is the tool to transfer the kinetic energy into the grinding
media and the grinding media transfers it into the powder material.
The system is operated with water cooling or heating at rotation frequencies up to 1800
rpm.
Figure 18: Running high-energy mill, Zoz
Figure 19: Inside a high-energy mill, Zoz
2.3.2.1.2.3 Other types of mills
Many other types of milling devices are available. Shaker mills such as SPEX mills (figure
20), which mill about 10 to 20 g of powder at a time, are most commonly used for
laboratory investigations and for alloy screening purposes.
Raf Moors – Peter Adriaensen
-Literature review- 47
Figure 20: SPEX 8000 mills in the assembled condition, SURYANARAYANA,C., 2001
The most common type of this mill contains only one vial. The sample and the grinding
balls are put together in the vial, secured in the clamp and swung energetically backwords
and forwards several thousand times each minute. This shaking motion is combined with
lateral movements of the ends of the vial. This causes the appearance of a figure of eight
movement of the vial. The milling and mixing of the sample is caused by the impact of the
balls against the sample with each swing of the vial. As a result, of the amplitude (about 5
cm) and speed (about 1200 rpm) of the clamp motion, the ball velocities are high (on the
order of 5 m/s). This results in an unusually large force of the ball’s impact. Therefore,
these mills can be classified as a high-energy type [Suryanarayama 2001a, pp. 1-184].
An attritor mill (figure 21) consists of a vertical drum with a series of impellers inside it.
Powder size reduction is caused by the impact between balls, between balls and container
wall, a gitator shaft, and impellers. A powerful motor rotates the impellers, which in turn
agitate the steel balls in the drum [Suryanarayama 2001a, pp. 1-184].
Raf Moors – Peter Adriaensen
-Literature review- 48
Figure 21: Attritor, IMP
Attritor mills allow milling of large amounts of powder at a time (from about 0,5 to 40
kg). The velocity of the grinding medium is much lower (about 0,5 m/s) than in Fritsch or
SPEX mills and consequently the energy of the attritors is low. The working principle is
shown in figure 22.
Figure 22: Working principle of the attritor, SAAR-HARDMETAL
Commercial mills for MA are much larger in size than the mills described above and can
process over hundred pounds at a time. Mechanical alloying for commercial production is
Raf Moors – Peter Adriaensen
-Literature review- 49
carried out in ball mills of up to about 1250 kg capacity [Suryanarayama 2001a, pp. 1-
184].
2.3.2.1.3 Process variables
Due to the complexity of the MA process optimization involves a number of parameters to
achieve the desired product phase and/or microstructure. The most important variables that
have an effect on the final product are:
type of mill,
milling container,*
milling speed,*
milling time,*
type, size, and size distribution of the grinding medium,*
ball-to-powder weight ratio,*
extent of filling the vial,*
milling atmosphere,
process control agent,
temperature of milling.
(* these are the parameters which will be explored and explained. An explanation of the
other terms can be found in C. Suryanarayana [2001b, pp. 21-29]
Some of these parameters are linked with each other. For example, the optimum milling
time depends on the type of mill, size of the grinding medium, temperature of milling, ball-
to-powder ratio, etc.
One important conclusion from the above parameters is that the material of the vial, the
balls and one of the powders has to be the same to decrease the contamination.
2.3.2.1.3.1 Milling container
Raf Moors – Peter Adriaensen
-Literature review- 50
Due to impact of the grinding medium on the inner walls of the container, some material
will be dislodged and get incorporated into the powder. This leads to contamination of the
powder or alterations in the chemistry of the powder. The powder may be contaminated
with the grinding vial material if the material of the grinding vessel is different from that of
the powder. If the two materials are the same, the chemistry can be changed. Therefore
there have to be taken precautions against the additional amount of the element
incorporated into the powder.
Hardened steel, tool steel, hardened chromium steel, tempered steel, stainless steel, WC-
Co, WC-lined steel [Ivanov et al. 1999, pp. 377-383], and bearing steel are the most
common types of materials used for the grinding vessels.
2.3.2.1.3.2 Milling speed
The faster the mill rotates, the higher the energy input into the powder. However there are
limitations to the increase in speed. For example, increasing the speed of rotation will
increase the speed with which the balls move. Above a critical speed, the balls will be
pinned to the inner walls of the vial and will not fall down to exert any impact force, so it
is important that the maximum speed stays below this critical value.
The high amount of energy will cause an increase in temperature, so temperature will be
another limitation for the maximum speed. In some cases, higher temperatures can be an
advantage, especially when diffusion is required to promote homogenization and alloying
in the powders. There is also a disadvantage; the increase in temperature accelerates the
transformation process and results in the decomposition of supersaturated solid solutions or
other metastable phases formed during milling [Kaloshkin et al. 1997, pp. 565-570].
Summary: the maximum temperature is influenced by many parameters and these values
vary widely (also dependent on the type of mill).
Raf Moors – Peter Adriaensen
-Literature review- 51
Research has been carried out looking at the difference of the final powder constitution
when vanadium and carbon powders were milled together at different energy levels [Calka
A. et al. 1993, pp. 189-195]. For example, at very low milling energy (or speed), the
powder consisted of nano-sized grains of vanadium and amorphous carbon, which on
annealing formed either V2C or a mixture of V and VC. At an intermediate energy level,
the milled powder contained a nanostructure. This powder will be transferred to VC when
it’s annealed. At the highest energy level, the VC formed directly on milling.
2.3.2.1.3.3 Milling time
A very important parameter is the milling time. The milling time should be chosen to
achieve a steady state between fracturing and cold welding of the powder. The type of mill
has an influence on the times required.
By choosing the appropriate time, one should also consider the type of mill used. Also the
combination of powders is a parameter which should be considered. As the milling time
increases, the level of contamination will also increase. Therefore it is desirable that the
powder is milled just for the required time and no longer [Suryanarayana 2001a, pp. 1-
184].
2.3.2.1.3.4 Grinding medium
The most common steels used for the grinding medium (milling-balls) are: hardened steel,
tool steel, hardened chromium steel, tempered steel, stainless steel, WC-Co and bearing
steel. The impact force on the powder created by the balls is determined by the density of
the grinding medium. Therefore the density must be high enough.
The milling efficiency is also dependant on the size of the grinding medium, a larger
medium has higher impact on the powder particles. Therefore efficiency will increase with
a larger medium.
Raf Moors – Peter Adriaensen
-Literature review- 52
Also the final constitution is dependent upon the size of the grinding medium. For
example, it has been reported that, when balls of 15 mm diameter were used to mill the
blended elemental Ti-Al powder mixture, a solid solution of aluminium in titanium was
formed. On the other hand, use of 20 and 25 mm diameter balls resulted in a mixture of
only the titanium and aluminium phases, even after a long milling duration [Lai and Lu
1998]. Figure 23 shows a figure of the grinding medium.
Generally the sizes that are used in an investigation are limited to one ball diameter.
Although there have been studies where different sized balls were used [Atzom 1990, pp.
487-490 and Gavrilov et al. 1995, p. 11].
Figure 23: Grinding medium, NARANG ENTERPRICES
It has been predicted that the highest collision energy can be obtained if balls with different
diameters are used.
During the first time period of the milling process, the powder becomes coated onto the
surface of the milling balls. Hereby high wear of the grinding medium is prevented. Also
contamination of the powder due to wear of the milling balls will be avoided.
However, to become a homogeneous final product, the thickness of this layer must be kept
to a minimum. This powder coating on the grinding balls is also difficult to remove.
The amount of cold welding and the amount of powder coated onto the surface of the balls
will be reduced by using a combination of large and small sized balls during milling
[Takacs et al. 1994, pp. 5864-5866]. Using balls of the same size in diameter causes the
Raf Moors – Peter Adriaensen
-Literature review- 53
balls to follow certain tracks and the balls roll along a well defined trajectory. Therefore it
is useful to combine small and larger balls in order to obtain a randomized motion [Takacs
1996, pp. 453-464].
2.3.2.1.3.5 Ball-to-powder weight ratio
The ratio of the weight of the balls to the powder (BPR), sometimes referred to as charge
ratio (CR), is an important variable in the milling process. Investigations have been done
with BPR varied from 1:1 [Chin and Perng 1997, pp. 235-238 and pp. 121-126] to 220:1
[Kis and Beke 1996, pp. 465-470]. The most common ratio used in a small capacity mill is
10:1. When milling is preformed in a large capacity mill, a higher BPR of up to 50:1 or
even 100:1 is used.
The higher the BPR, the greater the number of collisions per minute, due to higher weight
proportion of the balls, leading to more energy being transferred to the powder particles
and so alloying takes place faster.
There is a possibility that due to the higher energy, more heat is generated. This could have
an effect on the final constitution of the powder. The BPR may also effect the phase
formed after the milling process.
2.3.2.1.3.6 Extent of filling the vial
In order to allow the balls to move so they can produce impact forces on the particles, it is
necessary that there is enough space in the vial. Therefore, the extent of filling the vial with
the powder and the balls is important. If the quantity of the balls and the powder is very
small, then the production rate is also very small. On the other hand, if the quantity is
large, then there is not enough space for the balls to move around and so the energy of the
impact is less. Therefore it is important not to overfill the vial. Generally 50% of the vial
space is left empty.
Raf Moors – Peter Adriaensen
-Literature review- 54
2.3.2.2 Liquid phase techniques
Wet chemistry techniques that have been used for producing larger particles are being
adapted to produce nanosized material, the product crystallising or precipitating out of the
solution. In order to achieve the reduction in particle size the reaction chemistry must be
chosen to provide a fast spontaneous reaction and also be able to limit subsequent growth
of the particles after nucleation. The particle size can be controlled by the use of various
polymers, gels or micro-emulsions to constrain the growth of the particles [AP Materials
Inc. n.d.].
2.3.2.3 Vapour phase techniques
There are a number of techniques that can be classed under this category, but they all
involve two basic steps. The first step is the vaporisation of the material followed by a
rapid controlled condensation to produce the required size of particle. In some techniques
the powder formed is the same composition as the starting material, while others rely on
decomposition occurring during the vaporisation step to produce the desired product.
Production of carbon black, obtained by collecting the soot produced by the burning of
natural gas or other hydrocarbons, is probably the earliest example of a vapour phase
technique. Silica (silicon dioxide) and titanium dioxide are also produced in bulk by a
combustion (flame synthesis) process. A modification of the flame synthesis, known as
sodium flame and encapsulation (SFE) technology [Argonide 2004], is currently under
development, which should also allow the bulk production of non-oxide powders of
ceramics, metals and composites. In SFE the particles are encapsulated as they formed by a
protective layer, which prevents oxidation and inhibits agglomeration.
Other energy sources besides combustion are now being used to vaporise the material. For
example metal nanopowders can be produced by an exploding wire technique. This
technique is applicable to any metal that is available as a continuous ductile wire[Shenzen
n.d.]. Diamond nanopowders are being produced by a detonation process [Tetronics ltd.
Raf Moors – Peter Adriaensen
-Literature review- 55
n.d.]. Metal, metal oxide and ceramic nanopowders are being produced using a plasma arc
to generate the vapour [Qinetiq Nanomaterials Inc. 2003 – Office of technology Transfer
n.d]. An electromagnetic vaporisation process (EVP) [Epma 2004a], using an alternating
magnetic field to generate fine droplets of molten metal, is being used to produce metal
nanopowders. Yet another variant is laser ablation [Okada et al. 1992, pp. 862-864]. In this
technique a high-energy laser beam is used to evaporate a compressed precursor powder.
The evaporated material is then condensed into nanoparticles by a process of collision and
growth. Development work continues on all of these techniques to increase their yields and
reduce the production costs.
2.3.2.4 Plasma heating
Plasma heating provides a clean, directional, controllable, high intensity, localised source
of contaminant free heat. Due to the extreme high temperatures in the plasma zone (up to
20 000 degrees Kelvin) a range of novel materials can be produced. These often take the
form of spherical nanometric powders with diameters in the range 15 - 300 nm, with
exceptionally high specific surface area/mass ratios, up to >200m²/g. These materials have
generated a considerable level of commercial interest in the electronic, ceramic,explosive
enhancement and catalytic sectors [AP Materials Inc n.d.]. Figure 24 shows the principle of
plasma heating.
Raf Moors – Peter Adriaensen
-Literature review- 56
Figure 24: Principle of plasma heating, Max Planck Institute
2.3.2.5 Solid phase techniques
An example of this technique is mechanochemical processing (MCP) that is being used to
produce metal oxide nanopowders. Dry milling is used to induce chemical reactions
through ball-powder collisions that result in nanoparticles being formed within a salt
matrix. The particle size produced depends on the chemistry of the reactant mixture,
milling and heat treatment conditions. Particle agglomeration is minimised by the salt
matrix, which is then removed by a simple washing procedure.
2.3.2.6 Equal channel Angular Extrusion
Equal Channel Angular Extrusion (ECAE) was invented in the former Soviet Union by
Vladimir Segal in 1977 [Tamu n.d.].
ECAE is an innovative process capable of producing uniform plastic deformation in a
variety of materials, without causing significant change in geometric shape or cross
Raf Moors – Peter Adriaensen
-Literature review- 57
section. Multiple extrusions of billets by ECAE permits severe plastic deformation in bulk
materials. By changing the orientation of the billet between successive extrusions, complex
microstructures and textures can be developed. Changing the chosen billet orientation after
each pass, five fundamental equal-channel angular extrusion routes are defined and utilized
to obtain different textures and microstructures.
Research has concentrated in two areas:
development of theoretical and practical knowledge of the mechanics of ECAE (mathematically modelling, etc.)
investigation of the various technological advantages of ECAE. Many advantages were found with ECAE. ECAE has also been found to be an excellent method for powder consolidation.
The process involves extruding a metal billet through a die consisting of two equal
channels intersecting at an angle (see figure 25). After the billet has passed through the exit
channel, it has experienced severe plastic deformation, primarily in the form of shear
strain, yet no change in cross section. Therefore, one can re-insert the billet into the entry
die and impose more severe plastic strain. It is not uncommon for the ECAE processing
procedure to involve up to ~ 16 – 20 passes through the die. Different processing routes are
specified by the sequences of twists about the billet long axis imposed after each pass and
prior to re-insertion.
Raf Moors – Peter Adriaensen
-Literature review- 58
Figure 25: Principle scheme of the ECAE
ECAE is a process involving both SPD and strain path changes. With each pass through
the die, one can impose from approximately 83% up to 100% shear strain as the die angle
decreases from 135° to 90° [Segal 1995, pp. 57-164 – Iwahashi et al. 1996, pp. 143-146].
The severe plastic deformation generated in one pass alone is sufficient to generate
subgrain boundaries, the preliminary stages of eventual grain refinement. Reinsertion of
the billet changes the strain path for all routes, including route A, which involves no twists
about the billet axis. Thus, by altering the sequences of twists from pass to pass, one
effectively alters the strain path imposed, which leads to changes in microstructural
development. The combined influence of SPD and strain path changes can potentially lead
to the unique microstructures responsible for their superior strength and ductility.
Numerous ECAE studies in recent years have shown that the relationships between the
resulting microstructure (e.g. grain size, boundary misorientation, texture), material (e.g.
crystal structure, alloying, initial microstructure), and deformation history (die geometry,
route, number of passes, friction) are indeed complex ones. As one can imagine, there is a
great amount of flexibility in the design of the ECAE process. Material specifications for
these advanced nanomaterials will likely depend on application, some wanting a specific
ultrafine grain size, maximum strength and ductility, or a minimum degree of anisotropy
Raf Moors – Peter Adriaensen
-Literature review- 59
and inhomogeneity in properties, to name a few. The challenge in advancing ECAE to the
marketplace will be in developing predictive capabilities for designing nanomaterials to
specification and optimizing the process for fewer passes and larger specimens. Achieving
such a task will involve multiscale modeling from the nanoscale level up to the continuum
level (at least 7 orders of magnitude difference).
2.3.2.6.1 Simple shear concept
Naturally some of the first ECAE modeling efforts focused on the largest scale; that is, the
continuum level, modeling the plastic flow of the billet during one pass. The simplest way
is to classify it as a homogeneous deformation of simple shear localized along the
intersection plane of the two channels (figure 25) The applied shear γ as a function of Φ was determined by Segal [1995, pp. 57-164].
Once a strain state is defined, it is common to reduce it to a scalar measure, typically the
von Mises equivalent strain εvM. This scalar measure of strain is considered to be additive
from pass to pass; for N passes the accumulated strain is NεvM. Due to the strain path
changes, no point in the material experiences continuous monotonic simple shear in
multipass ECAE [Beyerlain et al. 2003, pp. 122-138 –Vogel et al. 2003, p. 2661].
2.3.2.6.2 Inhomogeneous deformation
In reality, ECAE deformation can neither be generally described as a simple shear along a
single plane nor as homogeneous. Finite element models (FEMs), both 2-D and 3-D [Park
and Suh 2001, pp. 3007-3014 – Kim et al. 2000, pp. 86-90 – Li et al. 2000 – Bowen et al.
2000, pp. 87-99 – Kim 2002, pp. 172-179 – Kim et al. 2001, pp. 856-864 – Budilov et al.
Raf Moors – Peter Adriaensen
-Literature review- 60
2004, pp. 189-198 – Beyerlein et al. 2004, pp. 185- 192] and slip line theory [Segal 2003,
pp. 36-46 – Stoica and Liaw 2003, pp. 119-133], have exposed much more regarding the
internal plastic flow characteristics and the conditions leading to non-uniform plastic flow
in a single pass [Park and Suh 2001, pp. 3007-3014 – Kim et al. 2000, pp. 86-90 – Li et al.
2000 – Bowen et al. 2000, pp. 87-99 – Kim 2002, pp. 172-179 – Kim et al. 2001, pp. 856-
864 – Budilov et al. 2004, pp. 189-198 – Beyerlein et al. 2004, pp. 185- 192 – Semiatin et
al. 2000, pp.1841-1851 – Segal 2003, pp. 36-46 – Stoica and Liaw 2003, pp. 119-133]. In
most cases, plastic deformation takes place in a broad zone whose shape depends
sensitively on factors such as contact friction, material flow response, backpressure,
pressing rate, and die design. Upon passing through such broad plastic deformation zones
(PDZs), the sample is left in an inhomogeneous deformation state, which can vary
significantly from top to bottom. It appears that the ideal simple shear viewpoint is
approached when the inner and outer corners are sharp, frictionless interfacial conditions
exist, backpressure is applied, and the material is rigid plastic [Park and Suh 2001, pp.
3007-3014 - Stoica and Liaw 2003, pp. 119-133].
Inhomogeneous deformation during each ECAE pass will most certainly lead to
heterogeneity in the final microstructure, texture, and mechanical properties across the
sample of the material [Beyerlain et al. 2003, pp. 122-138 and Li et al. n.d.].
Raf Moors – Peter Adriaensen
-Literature review- 61
2.3.3 Properties
The material properties of the nanostructured materials show remarkable improvement or
deviation from the properties exhibited by the coarser grained material. These unique
properties are attributed to the significant increase in grain boundary area due to the small
grain size. In terms of the mechanical properties, nanostructured metals have shown
increases in hardness values (ultimate and yield strengths). Specifically, pure nanophase
metals have shown a clear increase in hardness with decreasing grain size, following the
well known Hall-Petch equation. Nanostructured ceramics have exhibited superplastic
properties at low temperatures. This is significant as ceramics are conventionally brittle
materials [Nanostructured Materials 2004].
2.3.4 WC-Co particles
WC-Co hardmetals finds wide range of industrial applications. They are for example being
widely used in metal cutting tools because of their high hardness, good wear-resistance,
good fracture resistance and high temperature strength [Zhang et al. 2003 pp. 1123-1128].
Mechanical properties of hardmetals are strongly dependent on the microstructure of the
WC-Co hardmetal, and additionally affected by the microstructure of WC powders before
sintering. An important feature is that the toughness and hardness increase simultaneously
with the refining of WC. Therefore significant effort in research for development of
nanostructured WC-Co hard metals is done.
Nanometre materials, characterized by an ultrafine grain size, were prepared by vapor
condensation [Gleiter and Maquardt 1984, p. 263] until 1988 when Shingu et al. [1998, p.
29] found that nanocrystalline Al-Fe alloy can also be obtained by mechanical alloying.
Due to its simplicity and relatively inexpensive equipment high energy ball milling is a
simple and efficient way of manufacturing the fine powder with nanostructure. Another
advantage is its easy way to upscale the process. The problem of contamination
Raf Moors – Peter Adriaensen
-Literature review- 62
[Suryanarayana 2001, p. 1] can be eliminated by using a milling media that is made of
WC-Co hardmetal.
After a certain period of high energy milling the grain size of WC can be reduced to
nanoscale and become embedded in the cobalt particles together with a considerable
internal strain [Zhang et al. 2003, pp. 1123-1128].
2.4 Crystal structures and Point Defects
2.4.1 The Body-Centered-Cubic (BCC) structure
Figure 26: Sketch of a BCC structure, ALCAN
The BCC structure has one atom located at each of the cube corners and one atom at the
cube centre. Each atom has 8 neighbours i.e. coordination no. = 8. Examples of metals with
BCC structures include Cr, Fe, Mo [Cottrel 1995].
Raf Moors – Peter Adriaensen
-Literature review- 63
Figure 27: The atoms in a BCC structure, ALCAN
2.4.2 The Hexagonal-Close-Packed (HCP) structure
The HCP structure also has a coordination no. of 12. In this case two lattice parameters are
required, denoted a and c. The ideal HCP crystal has the closest packing of all and it is
possible to calculate the ideal c/a ratio = . Real metals deviate from this ratio
e.g., Mg has c/a = 1,623 and Zn has c/a = 1,856 [Cottrel 1995].
Figure 28: Sketch of a HCP structure, ALCAN
Raf Moors – Peter Adriaensen
-Literature review- 64
Figure 29: Atoms in a HCP structure, ALCAN
2.4.3 Miller indices – Cubic Crystals
It is often necessary to refer to crystal planes and directions for example to describe the
orientation of a particular precipitate type. The Miller indexing system provides a method
of achieving this using the three axis x, y, z system.
In the Miller system, however, the indices are used are inverse indices. Thus the direction
from the origin through the atom position x = 1, y = ½, z = 0 has the
Miller indices [1, 2, 0] [Cottrel 1995].
Directions are always denoted using [] square brackets.
Raf Moors – Peter Adriaensen
-Literature review- 65
Figure 30: Miller indices – Cubic crystals, ALCAN
Where the intercept on any axis is negative the corresponding Miller index will also be
negative. This negative index is represented using a bar over the integer, e.g. ( 00).
Figure 31: Negative Miller index, ALCAN
It is sometimes necessary to refer to all similar planes or directions in a crystal. This is
done by using different brackets, to describe a family of planes <uvw> to describe a
family of directions [Cottrel 1995].
For example, the cube faces have the indeces (100), (010), (001), ( 00), (0 0), (00 ). The
notation to describe cube faces would be . Similarly the notation to describe the cube
edges would be <100>.
Raf Moors – Peter Adriaensen
-Literature review- 66
For cubic lattices only a direction will be perpendicular to the plane which has the same
indices.
Figure 32: Cubic lattices only a direction will be perpendicular to the plane, ALCAN
2.4.4 Close Packed planes
Arrangement of atoms in a close packed plane. This is the arrangement of atoms in the
basal plane of an HCP structure or the (111) plane of FCC
Figure 33: First plane of FCC
The next plane is positioned over the spaces in the first plane to form the stacking sequence
A-B.
Raf Moors – Peter Adriaensen
-Literature review- 67
Figure 34: The second plane of FCC
The third layer can then be positioned over the A atoms. This forms an HCP sequence
ABABAB. If the third layer is positioned over the remaining un-occupied spaces this
forms the FCC lattice with the sequence ABCABC [Cottrel 1995].
Figure 35: The third layer of FCC
2.5 Grain measurement of WC
To get a good characterisation of the powder, it is needed to have a descent grain
measurement. There are different possibilities to measure them and every possibility has its
own range.
2.5.1 BET Surface Area
Raf Moors – Peter Adriaensen
-Literature review- 68
This is generally a research laboratory method for the powders that only gives a value for
the mean grain size; no information on grain size distribution is obtained. The surface
roughness of grains may affect the result and it is a slow method. For WC in the range 0.2
– 1000 m²g-1, d (diameter) is given by 6/ρSp where ρ = density and Sp = surface area/unit
mass [1970, pp. 429-448 – Friederich and Exner 1984, pp. 334-341 – Fischmeister et al.
1966, pp. 106-124 – Le Roux and King 1987, pp.243-248]. Figure 36 shows a picture of
this apparatus.
Figure 36: BET surface area
2.5.2 X-ray sedigraph
It gives information on grain size distribution as well as mean value, presented as wt
fraction cumulative probabilities. It is used for powders with mean grain sizes between the
1 – 8 μm grain size ranges. The measurement time is about 1h. There is a need to
deagglomerate and disperse the powders [1970, pp. 429-448 – Friederich and Exner 1984,
pp. 334-341 – Fischmeister et al. 1966, pp. 106-124 – Le Roux and King 1987, pp.243-
248]. Figure 37 shows a X-ray sedigraph apparatus.
Raf Moors – Peter Adriaensen
-Literature review- 69
Figure 37: X-ray sedigraph
2.5.3 Laser Diffraction
It is intended only for WC-Co agglomerated ready-to-press powders. It can give
information on grain size distribution in the range 2 – 1000 μm [1970, pp. 429-448 –
Friederich and Exner 1984, pp. 334-341 – Fischmeister et al. 1966, pp. 106-124 – Le Roux
and King 1987, pp.243-248]. The measurement time is about 10 min. Figure 38 shows a
laser diffraction machine.
Figure 38: Laser diffraction
2.5.4 Ultracentrifuge
The method can give grain size distributions, especially in the range < 0.5 μm where the X-
ray sedigraph and laser diffraction methods have limitations [1970, pp. 429-448 –
Friederich and Exner 1984, pp. 334-341 – Fischmeister et al. 1966, pp. 106-124 – Le Roux
and King 1987, pp.243-248]. Figure 39 shows the machine.
Raf Moors – Peter Adriaensen
-Literature review- 70
Figure 39: Ultracentrifuge
2.5.5 Photon correlation spectrography
This is a relatively expensive method, used for correlation with the ultracentrifuge for grain
sizes in the range 2 nm – 3 μm. One measurement takes 2h [Exner 1970, pp. 429-448 –
Friederich and Exner 1984, pp. 334-341 – Fischmeister et al. 1966, pp. 106-124 – Le Roux
and King 1987, pp.243-248]. Figure 40 and figure 41 are examples from photon correlation
spectography.
Figure 40: Photon correlation spectography
Raf Moors – Peter Adriaensen
-Literature review- 71
Figure 41: The beam of the photon correlation spectography
2.5.6 Microscopical image analysis, SEM, TEM
These are very direct methods: they are widely used in research. SEM in particular is used
to look at heterogeneity, i.e. large grains in a fine matrix. For mean grain sizes less then
about 0.5 μm a good resolution SEM is required, typically, with a field emission electron
source. SEM is better than TEM for morphological studies. However, polycrystallinity of
grains is easier to see in TEM [Anders 1992, pp. 195-204]. TEM suffers from poor
statistical accuracy for size measurements but is useful for very small grains less then 50
nm in size.
2.5.7 X-ray line broadening
A research method for ultrafine or nanograin powders. It only gives a mean grain size
value, not grain size distribution. Careful analysis and interpretation of data is required,
since grains may have internal structure. X-ray diffraction is insensitive in the range
between 0,1 and 10 µm [WHISTON C. 1987, p. 92]
2.5.8 Chemical reaction
Raf Moors – Peter Adriaensen
-Literature review- 72
Dissolution rates of WC in an iodide solution can be used to measure grain size (Krupp –
Widia) or alternatively in 10 ml HCl (32%), 10 ml HNO3 (conc), 50 ml H2O. Interrupted
experiments to measure the amount of unreacted WC gives a measure of the WC grain
size, since coarser grains dissolve more slowly than finer grains [Roebuck et al. 1999, pp.
47-54].
Summary of the most frequently used grain size measurement methods
In figure 42, a view over the different ranges of the discussed grain size measurement
methods is given.
Figure 42: The range were the different techniques are valid
Raf Moors – Peter Adriaensen
-Experimental procedure- 73
3 Experimental procedure
The different steps in the procedure are been spoken here.
3.1 Description of the powders
To become familiar with the powders and the equipment, starting experiments included the
use of spare WC/Co powder samples with unknown sizes, purity and composition.
Therefore the powders were sieved with a 20 µm sieve so the ball milling could take place
with more accuracy. After sieving, an X-ray analysis took place to investigate the sizes,
purity and composition of the powders.
These powders are used to investigate the parameters for planetary ball milling, powder
compression and sintering temperatures parameters. Different samples were prepared with
the same composition (90 wt% WC a 10 wt% Co) and were milled at different conditions.
For research on the influence of alternative binders and inhibitors we’ve used new powders
with known compositions and sizes.
3.1.1 Tungsten carbide (WC)
Experiments took place with two kinds of tungsten carbide powders: tungsten carbide
powder with measurements less then 20 micron and measurements less then 4.3 micron.
3.1.1.1 Tungsten carbide < 20µm
Company: unknown
Purity: XRF results:
Raf Moors – Peter Adriaensen
-Experimental procedure- 74
Sizes: sieved with 20µm sieve
An SEM picture of pure WC <20µm is given in fig 43:
Figure 43: SEM picture of WC particles < 20 micron
3.1.1.2 Tungsten carbide < 4.3µm
Company: William Rowland Limited
Purity: > 99.999%
Sizes: 3.48µm
3.1.2 Cobalt (Co)
Experiments took place with two kinds of cobalt powders: cobalt powder with measures
less then 20 micron and measures less then 4.3 micron.
Raf Moors – Peter Adriaensen
-Experimental procedure- 75
3.1.2.1 Cobalt < 20µm
Company: unknown
Purity: XRF results:
Sizes: sieved with 20µm sieve
3.1.2.2 Cobalt < 4,3µm
Company: William Rowland Limited
Purity: 99.87%
Sizes: 1.5µm
3.1.3 Iron (Fe)
Company: William Rowland Limited
Purity: > 99.8%
Sizes: 5.2 – 6.4µm
3.1.4 Nickel (Ni)
Company: William Rowland Limited
Purity: > 99.8%
Sizes: 2.9µm
3.1.5 Manganese (Mn)
Company: William Rowland Limited
Purity: > 99.10%
Sizes: 90.40% < 325 mesh
Raf Moors – Peter Adriaensen
-Experimental procedure- 76
3.1.6 Vanadium Carbide (VC)
Company: NewMet Koch
Purity: 99.8%
Sizes: 10µm
Raf Moors – Peter Adriaensen
-Experimental procedure- 77
3.2 Preparation of the powders
Following powder compositions were prepared:
Name Wt% WC Wt% Binder Wt% inhibitor
A WC-Co 90 wt% WC 10 wt% Co
B WC-Co with inhibitor 89 wt% WC 10 wt% Co 1 wt% VC
C WC- Fe/Mn 90 wt% WC 10 wt% Fe/Mn
D WC- Fe/Mn with inhibitor 89 wt% WC 10 wt% Fe/Mn 1 wt% VC
E WC - Fe/Ni/Co 90 wt% WC 10 wt% Fe/Ni/Co
Table 1: The composition of the powders
Compositions for the binders:
Binder Wt% Fe Wt% Ni Wt% Co Wt% Mn
1 Fe/Ni/Co 75 15 10 /
2 Fe/Mn 86.5 / / 13.5
Table 2: Used compositions for the binders
The powders are weighted with a Fischerbrand PS- 200 balance, with an accuracy of
0.1mg.
3.3 Milling process
There were two kinds of milling processes; first the planetary ball mill and secondary the
horizontally ball mill. Both of them are discussed in this chapter.
3.3.1 Planetary ball mill
The powders are milled in a Fritch pulveristette 5 planetary ball mill. To find the optimal
milling parameters, the reference sample has been milled for 2.5 ; 5 and 10 hours at
velocities of 250 and 400 rpm. We have used stainless steel vials and grinding balls. The
Raf Moors – Peter Adriaensen
-Experimental procedure- 78
used ball to powder ratio was 15:1. The vial was filled with 125 ml ethanol. After milling
the vials were placed in a fumeboard until the ethanol was vaporized (figure 44).
Figure 44: Fritsh Pulverisette 5
3.3.2 Horizontally high energy mill
A ZoZ high energy simoloyer (figure 45) has been used for milling. The milling process
was programmed with the software Maltoz. The first step was mixing of the powders,
involving 5 minutes rotating with a speed of 200 rpm. After this a 30 minutes programmed
routine was executed. This routine consisted of 12 repetitions of 1minute milling at 600
rpm and 4 minutes at 1000 rpm. The emptying procedure was 20 minutes milling at a
velocity of 1200 rpm. We used a ball to powder ratio of 1:10 with stainless steel balls.
Because the small particles were hard to remove, this high speed was necessary.
Raf Moors – Peter Adriaensen
-Experimental procedure- 79
Figure 45: The Maltoz operation software for the high energy mill
The milling proceeded under an inert argon atmosphere. Therefore it was necessary to
create a vacuum. With the combination of an airlock system and an Edwards vacuum
pomp, this was easy to achieve.
3.3.3 Development of dies for compaction
To further research the characteristics of the used powder combination, various dies have
been made to enable tests such as compaction and Charpy impact to be carried out.
The dies that are used in these experiments follow the ASTM standards. The 3D pictures of
the different dies are displayed on the following pages and they are all designed with the
Autodesk Inventor 5.0. The standards can be seen on www.astm.org.. The specific
drawings can be found in appendix C.
3.3.3.1 Compaction die
Raf Moors – Peter Adriaensen
-Experimental procedure- 80
First of all, there is a die in which powder compaction takes place, incorporating two
punches, an upper punch and a lower punch. These punches apply pressure on to the
powder mixture. At the end of the compaction phase, the lower punch will push the green
powder compact out of the die.
3.3.3.1.1 The die for compaction
The die consists of an inner die and an outer die. The inner die gives the green its specific
form. An assembly of the die is shown in figure 46.
Figure 46: Assembly of the compaction die
3.3.3.1.2 The upper punch
In the assembly shown in figure 47, the upper punch is secured to the test machine via 2
M16 screws. The upper punch is mounted on an additional mounting plate and 1 M8
screw.
Raf Moors – Peter Adriaensen
-Experimental procedure- 81
Figure 47: Assembly of the upper punch
3.3.3.1.3 The lower punch
The lower punch (figure 48) has the same geometrical form as the upper punch with an
overall length of 90 mm compared to 60 mm for the upper punch.
Figure 48: Assembly of the lower punch
3.3.3.1.4 The die for the Charpy test
Raf Moors – Peter Adriaensen
-Experimental procedure- 82
The die for the Charpy test is designed in accordance with the ASTM-standards. The die
also consists of an inner die and an outer die. Figure 49 shows an assembly of the die.
Figure 49: Assembly of the Charpy test die
3.3.3.1.5 Die for 3 point bending test
The die for 3 point bending test is designed in accordance with the ASTM-standards. The
die also consists of an inner die and an outer die. Figure 50 shows an assembly of the die
Figure 50: Assembly of the 3 point bending die
Raf Moors – Peter Adriaensen
-Experimental procedure- 83
3.4 Analysis methods
During the thesis there were different kinds of analysis methods that took place. They are
discussed in this chapter.
3.4.1 Sieving
For a rough determination of particle sizes we’ve used a Fritch analysette 3 pro sieve with
20µ mesh sieves. The sieve was programmed for a cycle of 20 minutes whereby the
amplitude alternated between 0.5 and 1 mm (figure 51).
Figure 51: Fritsh Sieve
3.4.2 X-ray diffraction
X-rays are electromagnetic radiation of wavelength about 1 Å (10-10 m), which is about the
same size as an atom. They occur in that portion of the electromagnetic spectrum between
gamma-rays and the ultraviolet. The discovery of X-rays enabled scientists to probe
crystalline structure at the atomic level. X-ray diffraction has been in use in two main
areas, for the fingerprint characterization of crystalline materials and the determination of
Raf Moors – Peter Adriaensen
-Experimental procedure- 84
their structure. Each crystalline solid has its unique characteristic X-ray powder pattern
which may be used as a "fingerprint" for its identification. Once the material has been
identified, X-ray crystallography may be used to determine its structure, i.e. how the atoms
pack together in the crystalline state and what the interatomic distance and angle are etc.
X-ray diffraction is one of the most important characterization tools used in solid state
chemistry and materials science. The technique is the only one available for phase
identification, and the associated mini-computer holds 60.000 inorganic and 20.000
organic reference patterns. This allows fast automatic identification of several phases.
The size and the shape of the unit cell for any compound can be determined easily using
the diffraction of X-rays.
A Philips PW 1729 X-ray generator with a Cu, shown in figure 52, anode has been used to
measure the X-ray diffraction patterns. The machine was operated at 40 kV and 10 mA.
The software used with the machine is X’pert plus.
Figure 52: Philips XRD
3.4.2.1 Formulae used in the X-ray diffraction
Raf Moors – Peter Adriaensen
-Experimental procedure- 85
The path difference between two waves:
2 x wavelength= 2dsin(θ)
For constructive interference between these waves, the path difference must be an integral
number of wavelengths:
n x wavelength= 2x
This leads to the Bragg equation:
n x wavelength = 2dsin(θ)
Figure 53 shows the x-ray diffraction pattern from a single crystal of layered clay. Strong
intensities can be seen for a number of values of n; from each of these lines the value of d
can be calculated meaning the interplanar spacing between the atoms in the crystal
[Institute for materials research 2004].
Figure 53: Reflection of X-rays from 2 planes of atoms in a solid
Raf Moors – Peter Adriaensen
-Experimental procedure- 86
3.4.2.2 Example: unit cell size from Diffraction data
The diffraction pattern of copper metal was measured with x-ray radiation of wavelength
1.315 Å. The first order Bragg diffraction peak was found at an angle 2θ of 50.5 degrees.
Calculate the spacing between the diffracting planes in the copper metal [Institute for
materials research 2004].
The Bragg equation is:
n x wavelength = 2dsin(θ)
Rearrange this equation for the unknown spacing d:
d = n x wavelength/2sin(θ).
=> theta is 25,25 degrees, n =1, and wavelength = 1.315Å, and therefore
d= 1 x 1,315/(2 x 0,4266) = 1,541 Å
Figure 54: X-ray diffraction pattern
3.4.2.3 Instrumentation
The X-ray diffraction experiment requires an X-ray source, the sample under investigation
and a detector to pick up the diffracted X-rays. Fig 55 is a schematic diagram of a powder
X-ray diffractometer [Institute for materials research 2004].
Raf Moors – Peter Adriaensen
-Experimental procedure- 87
Figure 55: Schematic of an X-ray powder diffractimeter
The X-ray radiation most commonly used is that emitted by copper, whose characteristic
wavelength for the K radiation is =1.542Å. When the incident beam strikes a powder
sample, diffraction occurs in every possible orientation of 2θ. The diffracted beam may be
detected by using a moveable detector such as a Geiger counter, which is connected to a
chart recorder. In normal use, the counter is set to scan over a range of 2θ values at a
constant angular velocity. Routinely, a 2θ range of 5 to 70 degrees is sufficient to cover the
most useful part of the powder pattern. The scanning speed of the counter is usually 2θ of 2
degrees min-1 and therefore, about 30 minutes are needed to obtain a trace [Institute for
materials research 2004].
3.4.3 XRF
3.4.3.1 Description of the machine
For making an estimation of the contamination level, we’ve used a Spectro Xepo XRF
machine shown in figure 56. We can only speak of an estimation because the sample used
for this analyze has a diameter of 25 mm. A small sample as this isn’t representative for a
large amount of powder.
Raf Moors – Peter Adriaensen
-Experimental procedure- 88
Figure 56: Spectro Xepo XRF machine
By using this machine the composition off the investigated powder will be given.
Unfortunately this machine only detects elements heavier than sodium. This means that the
level of carbon can’t be measured.
3.4.3.2 Preparation of the samples
Preparation of the samples involved mixing 85 wt% powder with 15 wt% wax. After the
powder and wax are mixed, the mixture is compressed with a pressure of 3 bar.
3.4.4 Scanning electron microscopy
The SEM was used to take magnified pictures so we could see if the WC was embedded in
the Co. Another reason to use the SEM was to measure the size of the grains.
Raf Moors – Peter Adriaensen
-Experimental procedure- 89
3.4.4.1 Description of the machine
To get a high magnification of the samples, so the WC embedded in Co became visible,
we’ve used a ZEIS Scanning Electron Microscope (SEM), equipped with a backscattered
detector at 20 kV (figure 57).
Figure 57: ZEIS SEM
3.4.4.2 Preparation of the samples
Powders are mixed with non-conductive bakelite. Hot mounting was applied at a
temperature of 170°C for 7 minutes. The mixture is compressed is compressed with a
pressure of 3.2bar. After a 3 minutes cooling time, the samples were ready for grinding and
polishing. The surfaces were then ground 600 and 1200 grinding papers and a final polish
was applied using first a 6µm cloth followed by a 1µm cloth.
Raf Moors – Peter Adriaensen
-Experimental procedure- 90
After polishing, optical microscopy was applied to see if the surface was good enough for
SEM.
3.4.5 Optical microscopy
The samples for SEM were first inspected with a regular optical microscope, shown in
figure 58, to make sure that the surfaces were well polished.
Figure 58: The optical microscope
Raf Moors – Peter Adriaensen
-Results and discussion- 91
4 Results and discussion
4.1 Results from planetary ball mill
4.1.1 Getting started
To become familiar with the powders and the equipment, starting experiments included the
use of WC/Co powder samples with unknown sizes, purity and composition. Therefore the
powders were sieved with a 20 µm sieve so the ball milling could take place with more
accuracy. After sieving, an X-ray analysis took place to investigate the sizes, purity and
composition of the powders.
As a starting sample, a WC-Co sample was made with a composition of 90 wt% WC and
10 wt% Co; no inhibitors where added and this sample will be referred
The parameters that were changed for the ball milling were the velocity (rpm) of the
milling machine and the milling time. To investigate those parameters, different samples
were prepared with the same composition (90 wt% WC a 10 wt% Co) and were milled at
different conditions so the made composition was milled at a rotational speed of 250 rpm
and 400 rpm and this for 2.5h; 5h and 10h. After milling, the contamination level was
estimated with an XRF test and the average particle sizes and the contributing elements
were determined with the help of an XRD print and the equation of Scherer.
4.1.2 Reference sample
As a reference sample, a mixture of 90 wt% WC and 10 wt% Co (no inhibitors were
added) was made. This is the reference sample and it won’t be milled.
Raf Moors – Peter Adriaensen
-Results and discussion- 92
4.1.2.1 XRD
Figure 59 is giving the XRD pattern of the not milled reference sample. The particles used
for the XRD are the composite particles because in those particles the WC is embedded
best into the Co. The composite particles are used for every experiment following.
With this graph, the starting elements and their average size could be measured. That way,
it was easier to search for the links between the velocity, the milling time and their effect
on the grain size of the particles.
Figure 59: XRD pattern of the unmilled WC-Co reference powder sample
The WC peaks at 2θ = 31.6° and 2θ = 35.9° together with the Scherer equation were used
to determine the average grain size of the WC particles. During the milling process, the
Raf Moors – Peter Adriaensen
-Results and discussion- 93
particles need to be broken down into smaller particles and need to be embedded into the
Co (MA).
These 2 peaks are chosen, and will be used throughout the whole research, because the
possible contamination elements, during milling, coming from the used stainless steel vials
and balls will not interfere these peaks. That way, these peaks are 2 representative peaks
for the WC.
4.1.2.2 XRF
To determine the contributing elements in the powder and to make a quick estimation of
the quantities, an XRF was used. In the figure 60, the most important elements were named
in function of their estimated quantities.
0
10
20
30
40
50
60
wt%
W Ni Co Fe Mn Cr
Elements
Estimation of the quantities of the contributing elements the not-milled powder
Figure 60: Graph of the contributing elements in the WC-Co reference sample
Raf Moors – Peter Adriaensen
-Results and discussion- 94
The level of W is only about 60 wt% because the XRF machine can not detect elements
with an atomic number below 11 and C has an atomic number of 6. The quantity of Co is
estimated at 10 wt%. There are smaller quantities of Ni, Fe, Mn and Cr and they are
ascribed to the powder production process and the purity of the starting elements used to
make the powders.
4.1.2.3 Calculated grain size of the starting WC particles
The Scherer equation was used to calculate the grain size from the starting WC particles:
κ = 1
λ = X-ray wavelength, λCu = 1,5418
β = FWHM of the diffraction line
θ = diffraction angle
The values of β and θ can be found in the XRD data because β is the value, in radians, of
the WC peak situated at 2θ and θ (in radians) is half the value of 2θ, the position of the
peak. The formula was used on both of the WC peaks (2θ = 31.6° and 2θ = 35.6°), and
afterwards there was made an average of the 2 sizes. That way, a more accurate result for
the average grain size was achieved.
In the starting WC particles, the FWHM of the two WC peaks are both 0.00635 radians.
The diffraction angles are respectively for the 2 WC peaks 0.275 and 0.310 radians. When
these values are put into the Scherer formula, the result is respectively for the 2 WC peaks:
Raf Moors – Peter Adriaensen
-Results and discussion- 95
From these results, the grain size is calculated as the average of the results of the Scherer
equation:
This was an unexpected result because the size of the starting WC material was sieved with
a 20 micron mesh and couldn’t be in the nanoscale. Therefore, a SEM picture of this
material was shot, so that a quick estimation of the sizes of the starting WC particles could
be made. Figure 61 shows this SEM picture.
Figure 61: SEM picture of the starting WC material < 20 micron
Out of the SEM picture, the grain size of the starting WC material was estimated
approximately 10 μm. The reason for this difference is because X-ray diffraction is
insensitive in the range between 0.1 and 10 µm [WHISTON C. 1987, p. 92]. Because
Raf Moors – Peter Adriaensen
-Results and discussion- 96
we’ve spend a lot of time using the XRD machine for grainsize determination and to find
out why the results didn’t fulfil our expectations, the results from XRD are included in this
work
4.1.3 XRD results of the planetary ball milled samples at 250 rpm
4.1.3.1 The 2,5 hours milled sample
The 2.5 hours ball milled sample was wet-grinded with stainless steel vials and balls with a
diameter of 5 mm while the used liquid was ethanol. Every half an hour, there was a pause
of 10 min, to let the vials, balls and powder cool down.
After milling, an XRD test was done on the sample. Figure 62 gives the XRD pattern of
the sample.
Raf Moors – Peter Adriaensen
-Results and discussion- 97
Figure 62: XRD pattern of the 2.5h ball-milled WC-Co sample @ 250 rot/min
The first thing what was special on this graph, is that the peaks are lightly shifted to the
right because the values of the same WC peaks have now a 2θ value of 31.8° and 35.9°
what can be seen visually. This means that there are tensions in the powder so they will not
be broken down, but will be pressed together. Another option to see that there is strain in
the particles, is calculating the lattice parameters and the result of these parameters will be
different then those of the not milled reference sample.
The FWHM (= Full Width Half Maximum) of the WC peaks at the given 2θ angles are
respectively 0.00714 and 0.00714 radians. The FWHM is bigger then with the reference
sample what means that there is a reduction in grain size of WC.
Raf Moors – Peter Adriaensen
-Results and discussion- 98
4.1.3.2 The 5 hours milled sample
To make a good representative calculation for the best fit parameters of the ball-milling
machine, the same quantities of WC and Co, the same milling cycle, material of the vials
and bowls as well as the ethanol liquid was used. There was made an XRD test on the
sample too with pattern given in figure 63.
Figure 63: XRD pattern of the 5h ball milled WC-Co sample @ 250 rot/min
Another thing that was conspicuous was the fact that there was no clear line broadening
what means that the particles of WC stay more or less the same. The FWHM (= Full Width
Half Maximum) of the WC peaks at the given 2θ angles are respectively 0.00555 and
0.000555 radians. The FWHM of these peaks are smaller as those of the reference sample
what means that the particles are bigger as those of the starting material.
Raf Moors – Peter Adriaensen
-Results and discussion- 99
Again, the peaks moved lightly to 2θ = 31.7° and 2θ = 35.9°. That means that there are
tensions in the powder particles what means that they are not broken down but pressed
together.
4.1.3.3 The 10 hours milled sample
The same parameters as described for the 5 hours milled sample were used. There was
made a XRD on the sample to given in figure 64.
Figure 64: XRD pattern of the 10 h ball milled WC-Co sample @ 250 rot/min
The 2θ angles of the same WC peaks are now approximately 31,818° and 35,773°. Again
there are tensions and the peaks are shifted to the right.
Another thing that was conspicuous was the fact that there was no clear line broadening
what means that the particles of WC stay more or less the same. The FWHM (= Full Width
Half Maximum) of the WC peaks at the given 2θ angles are respectively 0.00748 and
Raf Moors – Peter Adriaensen
-Results and discussion- 100
0.00834 radians. The FWHM of these peaks are smaller as those of the reference sample
what means that the particles are bigger as those of the starting material.
Again, the peaks moved lightly to 2θ = 31.8° and 2θ = 35.8°. That means that there are
tensions in the powder particles what means that they are not broken down but pressed
together.
4.1.4 The SEM pictures of the 250 rpm samples
The purpose of the SEM pictures was twofold. The first reason was to look if the Co is
embedded with WC. The second reason is to get a view on the size of the WC particles so
that these sizes can be compared to these of the Scherer equation, discussed later. It is also
important to say that for these experiments the composite particles are used.
In the pictures, the little light spots are WC and the grey matrix represents Co phase. The
black areas are bakelite used to hold the powders.
4.1.4.1.1 The 2,5h milled sample
Figure 65 shows the SEM view over the 2.5h planetary ball milled sample.
Raf Moors – Peter Adriaensen
-Results and discussion- 101
Figure 65: SEM picture of the 2,5h planetary ball milled sample
It is clear that WC particles are embedded within the Co matrix and the size of the largest
WC particles is 0.8 micron.
Raf Moors – Peter Adriaensen
WCCo
-Results and discussion- 102
4.1.5 XRD results of the planetary ball milled samples at 400 rpm
4.1.5.1 The 5 hours milled sample
As well as the 250 rot/min, the 5 hours ball milled sample was wet-grinded with stainless
steel vials and balls with a diameter of 5 mm while the used liquid was ethanol. Every half
an hour, there was a pause of 20 min, to let the vials, balls and powder cool down.
After milling, an XRD test was done on the sample. Figure 66 gives the XRD pattern of
the sample.
Figure 66: XRD pattern of the 5h ball milled WC-Co sample @ 400 rot/min
The first thing what was special on this graph, is that the peaks are lightly shifted to the
right because the values of the same WC peaks have now a 2θ value of 31.8° and 35.8°
what can be seen visually. This means that there are tensions in the powder so they will not
Raf Moors – Peter Adriaensen
-Results and discussion- 103
be broken down, but will be pressed together. Another option to see that there is strain in
the particles, is calculating the lattice parameters and the result of these parameters will be
different then those of the not milled reference sample.
The FWHM (= Full Width Half Maximum) of the WC peaks at the given 2θ angles are
respectively 0.00748 and 0.00831 radians. The FWHM is bigger then with the reference
sample what means that there is a reduction in grain size of WC.
4.1.5.2 The 10 hours milled sample
The 2θ angles of the same WC peaks are now approximately 31.7° and 35.9°. Again there
are tensions and the peaks are shifted to the right. Figure 67 gives the XRD pattern of the
10 hours milled sample.
The FWHM (= Full Width Half Maximum) of the WC peaks at the given 2θ angles are
respectively 0.00807 and 0.00879 radians. The FWHM of these peaks are bigger as those
of the reference sample what means that the particles are bigger as those of the starting
material. This is different then the 250 rot/min samples and follows our expectations.
Again, the peaks moved lightly to 2θ = 31.8° and 2θ = 35.7°. That means that there are
tensions in the powder particles what means that they are not broken down but pressed
together.
Raf Moors – Peter Adriaensen
-Results and discussion- 104
Figure 67: XRD pattern of the 10h ball milled WC-Co sample @ 400 rot/min
4.1.6 The SEM pictures of the 400 rpm samples
4.1.6.1.1 The 10h milled sample
Figure 68 shows the SEM picture of the 10h planetary ball milled sample at 400 rpm. The
cursor bars shows that the sizes of the embedded WC particles are between 2.273 μm and
239.2 nm. This is quite a large range, but on such pictures it is not clear how deep the WC
particles are embedded, so the grains size will be short to the average of these sizes.
Another thing that is clear, is that the WC particles are embedded in the Co.
Raf Moors – Peter Adriaensen
-Results and discussion- 105
Figure 68: SEM picture of the planetary ball milled sample at 400 rpm
4.1.7 Contamination level
Because the vials of the planetary ball mill are stainless steel, and the main content of the
powder is WC, which is very hard, there is a natural wear of the vials. This causes other
elements to contaminate the powder. The main contaminating elements are Fe, Cr, Ni and
Mn.
The level of contamination is investigated by XRF. This is not a very accurate way to find
the precise amount of elements in weight percent. The beam of the XRF covers only the
surface of the prepared sample and the infiltrating depth is unknown and depends on the
material. Therefore different samples of the same powder may give different results, but
nevertheless, it is a good estimation.
Figure 69 shows how the contaminating elements increase in weight percent with
increasing milling times and speed. It is obviously that with stainless steel vials a milling
time of 10h at 400 rpm is to long. Both time and speed must be high enough to reduce the
particle size and get the WC embedded in the Co particles.
Raf Moors – Peter Adriaensen
Co
WC
-Results and discussion- 106
0
10
20
30
40
50
60
wt%
W Ni Co Fe Mn Cr
Elements
Estimation of the contributing elements in milled powder
not milled5h@250rpm10h@250rpm5h@400rpm10h@400rpm
Figure 69: Estimation of the contributing elements in milled powder
Contamination after milling
0,3545
34,737
47,026
0,3545
60,353
75,071
010203040
50607080
0 2 4 6 8 10
milling time (h)
Cont
amin
atio
n (w
t%)
250rpm
400rpm
Figure 70: Contamination level in function of the milling time
In figure 70 shows the contamination in function of the time. It is clear that the longer is
milled, the more contamination appears. Another thing what is important, is that the
contamination level, with the same milling time, of the 400 rotational speeds is higher then
the contamination of the 250 rotational speeds.
4.1.8 The grain size calculation
Raf Moors – Peter Adriaensen
-Results and discussion- 107
To determine the best fit parameters and to use the Taguchi matrix, it is necessary to
determine the grain sizes of the WC particles after milling. The first method is the Scherer
equation.
4.1.8.1 The Scherer equation
4.1.8.1.1 The planetary ball milled samples at 250 rpm
The grain sizes of the milled samples are calculated with the help of an excel counting
sheet and the Scherer formula, discussed in 6.1.2.3 Figure 71 gives an overview of this
sheet and the out coming results for the 250 rpm milled samples.
Grain size of WC for the 2,5h milled powder at 250rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00714 0.277 224.509 0.02241.542 0.00714 0.313 226.989 0.0227 average 225.749 Angström 0.0226 μm
Grain size of WC for the 5h milled powder at 250rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00555 0.277 288.611 0.02891.542 0.00555 0.314 291.860 0.0292 average 290.235 Angström 0.0290 μm
Grain size of WC for the 10h milled powder at 250rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00555 0.278 288.677 0.02891.542 0.00552 0.312 293.565 0.293 average 291.121 Angström 0.0291 μm
Figure 71: The excel counting sheet for the average grain size calculation with the Scherer equation (250
rpm milled sample)
Raf Moors – Peter Adriaensen
-Results and discussion- 108
4.1.8.1.2 The planetary ball milled samples at 400 rpm
The grain sizes of the milled samples are calculated with the help of an excel counting
sheet and the Scherer formula, discussed in 6.1.2.3. Figure 72 gives an overview of this
sheet and the out coming results for the 400 rpm milled samples.
Grain size of WC for the 5h milled powder at 400rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00748 0.277 214.280 0.0214
1.542 0.00831 0.312 194.940 0.0194
average 204.610 Angström 0.0204 μm
Grain size of WC for the 10h milled powder at 400rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00807 0.277 198.516 0.0198
1.542 0.00879 0.313 184.453 0.0184
average 191.484 Angström 0.0191 μm
Figure 72: The excel counting sheet for the average grain size calculation with the Scherer equation (400
rpm milled sample)
4.1.8.1.3 Conclusions of the Scherer equation
Because these results are not compatible with the results of the SEM, another method for
calculating grain size of the WC was needed. A possible reason why the Scherer equation
was not effective for the grain size calculation, was that there where lattice strains in the
WC grains coming from the ball milling [Segmüller et al. 1989, pp. 21-66]. Thus, a
formula that included a parameter for these strains was searched. The first person’s dealing
with these strains was Stokes and Wilson. Because this was still a single line method (less
Raf Moors – Peter Adriaensen
-Results and discussion- 109
accurate), another (multi line) method was needed, coming from the Fourier convolution
method to obtain the purely physical broadened line profiles and the Warren-Averbach
formula is such a formula.
The main disadvantages of these multi-line analysis are:
They are not useful for large line overlapping
Not useful for weak structural broadening
Stokes can not be applied without severe errors
The main reason why integral breadth fittings became more attractive, was because of the
mathematical process involved. Nevertheless, a full powder pattern fitting was used. This
is the moment that Voigt and pseudo Voigt were introduced into the field of x-ray powder
diffraction.
There are a lot of software packets dealing with line-broadening analysis and for our
purpose we used the Winfit! V1.2 program from S. Krumm.
Figure 73 shows the Fourier analysis of the XRD pattern from the starting WC particles <
20 μm for the two discussed WC peaks.
Raf Moors – Peter Adriaensen
-Results and discussion- 110
Figure 73: Fourier analysis of the un-milled WC powder < 20 micron
Figure 74 shows the results of the program.
Raf Moors – Peter Adriaensen
-Results and discussion- 111
Figure 74: Results of the Winfit! V1.2 on the <20 micron WC particles
Figure 75 shows the grain size calculation for the WC grains. Again, there is a big
difference between the results from Winfit! V1.2 software and the SEM pictures.
Figure 75: Grain size from the <20 micron WC particles with the Winfit! V1.2 software
Raf Moors – Peter Adriaensen
-Results and discussion- 112
4.1.9 Strain evaluation from peak shift
As already discussed, there are strains in the grains due to the milling. In this part, a clear
view is given over the peak shifts. The calculations of the lattice strains are not given in
this thesis.
In figure 76 a graph is given of the unmilled reference powder and the 10 h ball-milled
powder at 250 rpm.
Figure 76: Peak shift of WC with 250 rpm milling speed
It is clear that there is a peak shift and it can be seen by the space between the 2 red
vertical markers. Grain size determination from XRD results is not possible because XRD
line broadening analysis is insensitive in the range between 0,1 and 10µm (WHISTON C.
1987, p. 92).
4.2 Results from the high energy mill
Raf Moors – Peter Adriaensen
-Results and discussion- 113
In this part, a good powder characterisation will be made for the two considered alternative
binders. First, an estimation of the contamination level was made on a powder composition
of WC-10 Co. This powder was milled for half an hour, 1 hour and 3 hours.
Because XRD is not usable in our situation, we didn’t continue with this method and used
SEM to determine the grain sizes.
The high energy mill was programmed for 10 minutes of mixing the powders at a low
rotational speed of 500 rpm followed by a cycle of 15 minutes milling at a speed of 1000
rpm and 5 minutes break. This cycle was repeated 3 times thus in total there was 1h of
milling at high rotational speed. Thereafter, the powder was discharged for approximately
15 minutes.
4.2.1 The estimation of the contamination level of the milled powder
With the horizontal high energy mill the same problem occurs as with the planetary ball
mill namely contamination. This contamination in the milled powder comes from the wear
from the stainless steel grinding unit and balls. The main contaminating elements are Fe,
Cr, Ni and Mn.
It has to be considered that XRF gives only an estimation of the weight percentages of the
contributing elements of the powder but XRF is reliable enough to determine the trend of
contamination level with higher milling times.
In figure 77 an estimation of the most important contamination elements is shown. It is
clear that the longer the milling times are, the more contaminating elements are in the
powder. As with the planetary ball milled powder, the level of contamination increases
with the milling time.
Raf Moors – Peter Adriaensen
-Results and discussion- 114
0
10
20
30
40
50
60
wt%
W Ni Co Fe Mn Cr
elements
Estimation of the contributing elements in milled powder
unmilled30min high energy1hour high energy
Figure 77: Estimation of the contributing elements in milled powder
Figure 78 shows that the contamination in weight percent doubles when the milling time
doubles.
contamination after milling
0,3545
1,4105
3,683
0
0,5
1
1,52
2,5
3
3,5
4
0 0,5 1
milling time (h)
cont
amin
atio
n le
vel (
wt%
)
estimation of thecontamination level
Figure 78: contamination after milling
Raf Moors – Peter Adriaensen
-Results and discussion- 115
4.2.2 SEM pictures from samples milled with the high energy mill
The purpose of the SEM pictures was twofold. The first reason was to look if the
alternative binder is embedded with WC. The second reason is to get a view on the size of
the WC particles. It is also important to say that for these experiments the composite
particles are used.
In the pictures, the little light spots are WC and the grey spots are coming from the
alternative binder. The black spots are coming from the bakelite used to make the samples.
4.2.2.1 SEM pictures from Fe/Mn as alternative binder
Figure 79 shows the SEM picture of the 1h milled sample. The cursor bars shows that the
sizes of the embedded WC particles are between 1,273 μm and 39,2 nm. This is quite a
large range, but on such pictures it is not clear how deep the WC particles are embedded,
so the grains size will be short to the average of these sizes.
Another thing that is clear, is that the WC particles are embedded in the Fe/Mn binder. Our
first conclusions are that this is a good alternative to Co but further investigations
involving the mechanical properties are necessary to know if it is a good alternative.
Raf Moors – Peter Adriaensen
-Results and discussion- 116
Figure 79: SEM picture of the 1h horizontal energy milled sample WC-Fe/Mn
4.2.2.2 SEM pictures from Fe/Ni/Co as alternative binder
Figure 80 shows the SEM picture of the 1h milled sample with Fe/Ni/Co as alternative
binder to Co. The cursor bar below the picture shows that the sizes of the embedded WC
particles are between 1,273 μm and 39,2 nm. This is quite a large range, but on such
pictures it is not clear how deep the WC particles are embedded, so the grains size will be
short to the average of these sizes.
Another thing that is clear, is that the WC particles are embedded in the Fe/Ni/Co binder.
Our first conclusions are that this is a good alternative to Co but further investigations
involving the mechanical properties are necessary to know if it is a good alternative.
Raf Moors – Peter Adriaensen
Fe/MnWC
-Results and discussion- 117
Figure 80: SEM picture of the 1h horizontal energy milled sample WC-Fe/Ni/Co
Raf Moors – Peter Adriaensen
WC
Fe/Ni/Co
-Conclusions and suggestions for further work- 118
5 Conclusions and suggestions for further work
5.1 Introduction
Metal powders are used in industry for a diverse range of products. Some of these products
include welding electrodes, paints, printing inks and explosives. In all these products the
particles retain their identities. Traditional powder metallurgy is a process whereby a solid
metal, alloy or ceramic in a form of a mass of dry particles, normally less than 150 µm in
maximum diameter, is converted into an enginering component of a predetermined shape
possessing properties which allow it to be used in most cases without further processing.
The basic steps in traditional powder production are:
Powder production
Compaction of the powder
Sintering which involves heating the preform to a
temperature below the melting point of the major
constituent, at which point the powder particles lose their
identities through inter-diffusion processes and required
properties are developed.
These 3 phases are described in the literature review and the powder production phase has
been carried out practical.
Due to the poor corrosion resistance of Co, its high cost and environmental toxicity,
substantial research has been devoted to find suitable alternative binders for WC systems.
The aim is to reduce the amount of Co, or possibly, to completely replace Co binder. Two
promising alternatives are described and utilised in this project, the first one is a mixture
of iron (Fe), nickel (Ni) and cobalt (Co) and the second alternative is composed of iron
(Fe) and manganese (Mn). Compared to cobalt, Fe and Mn are very cheap and non toxic
Raf Moors – Peter Adriaensen
-Conclusions and suggestions for further work- 119
The first step in powder metallurgy is the powder production. The next step would be to
compact such powder for a subsequent sintering process. For that appropriate compaction
dies were designed using Inventor CAD software, a die to produce cylindrical samples for
microstructural and hardness analyses and another die to produce samples for 3-point
bending tests. Both dies were designed according to ASTM standards.
5.2 Conclusions
5.2.1 The planetary ball mill
A series of experiments were performed with the planetary ball mill by varying milling
time (2.5, 5, 10 hrs) and rotation speed (250, 400rpm) parameters to process WC-10wt
%Co, WC-10%FeNiCo and WC-10%FeMn. It was noticed that as the milling time
increased (above 2.5 hours for 150rpm) the amount of elements (Fe, Cr) picked up from
the stainless steel vial inner wall increased. The contamination level increased further at a
rotation speed of 400rpm. This indicates that both speed and time should be kept low to
minimise contamination or a hard steel vial should be utilised.
The results from the XRF tests show that the contamination level, using the 400 rpm
rotational speed, almost doubles when the milling time is kept constant. SEM micrographs
indicate that further particle size reduction was not achieved with increasing the milling
time at this rotation speed. It can be concluded that using a milling speed higher than 200
rpm is not beneficial with the stainless steel vial.
XRD technique was utilised to determine the grain size of WC phase (average of 24.1 nm).
The technique is also useful to determine the lattice strains exerted on the WC phase
(which can be determined from of the XRD peak shifts) due to the mechanical milling
process. SEM was used to evaluate the particle size of WC (average of 1.103 μm).
Raf Moors – Peter Adriaensen
-Conclusions and suggestions for further work- 120
5.2.2 The horizontally high energy mill
To eliminate the problem of a high contamination level found in the mechanically milled
powder, a hard steel vial was utilised. For that, additional powders were prepared using the
horizontal high-energy ball mill. Furthermore, this equipment allowed achieving a fine
structure with shorter milling time due its much higher rotation/milling speeds capabilities.
From the results of the first test, it was obvious that this method allowed a much quicker
way to reach the nanosize structure of the powders. This is due to the high kinetic energy
of this process, offering milling speeds over 1000rpm, a serious milling time reduction was
possible. Furthermore, compared to the planetary ball milling equipment, the
contamination level was fairly low (1 wt% Fe by with the horizontally high energy mill to
25.1 wt% Fe with the planetary ball milled sample, milled for 5h at 250 rot/min) as shown
by the XRF analysis. An other advantage of the horizontally high energy mill is the large
capacity of the container (2 litters). Hereby it is possible to produce a relatively larger
quantity of powder in one single cycle.
Considering these points we can conclude that the horizontally high energy mill has a
higher efficiency than the planetary ball mill.
5.2.3 The alternative binders
Two alternative binders for Co have been investigated and mixed with WC. The first one
was Fe/Mn as alternative. Out of the SEM picture, it was clear that WC particles were
nicely embedded in Fe/Mn binder and the size of WC particles was reduced by the
mechanical alloying process. That was a promising result. The compaction phase and the
sintering phase will be carried out in a later phase.
Raf Moors – Peter Adriaensen
-Conclusions and suggestions for further work- 121
Similarly for the second alternative binder Fe/Ni/Co, SEM pictures revealed that WC
particles were reduced in size and distributed within the binder matrix. Again, the
compaction and sintering phase will be carried out in a later phase.
5.3 Further work
Further work includes powder compaction of the different powder compositions prepared
by mechanical alloying process followed by sintering the compacted samples. Parameters
related to both the compaction and sintering experiments should be optimised for each
powder system (WC-Co, WC-FeNiCo, WC-FeMn) in order to prevent excessive grain
growth of the WC submicron/nano-structure achieved in the powders.
For the sintered samples, the physical and mechanical properties, such as density;
hardness/microhardness and 3-point bending fracture strength, will be measured as well as
microstructural analysis, evaluation of WC grain size and distribution, will be performed.
Raf Moors – Peter Adriaensen
-References- 122
6 References
ADVANCED POWDER TECHNOLOGY website n.d. Manufacturers of the worlds
finest powder. < http://www.apt-powders.com/ > (last accessed on 1octobre 2004).
AGNEW, S.R. et al. (1998) Proceedings of the 19th Risø International Symposium on
Materials Science, Roskilde, p. 201.
ALMOND, E.A. and ROEBUCK, B. (1988) Mat. Sci. and Engng, A105/106, p.237.
ANDERS NORDGREN (1992) Microstructural characterisation of cemented carbides
containing tailored structure variations using SEM-based automatic image analysis, J.
Hard Materials, 3(2), pp. 195 – 204.
AP Materials Inc website. Missouri n.d. < http://www.apmaterials.com > (accessed on
25september 2004).
ARGONIDE (2004) Nanomaterial Technologies. <http://www.argonide.com>(last
updated 2004 accessed on 24 September 2004).
ASTM: B 312 -96 (1996) Standard Test Method for Green strength for Compacted
Metals Powder Specimens.
ASTM: B331-95 (2002) Standard Test Method for Compressibility of Metal Powders
in Uniaxial Compaction.
ASTM: B925-03 (2003) Standard Practices for Production and Preparation of Powder
Metallurgy (P/M) Test Specimens.
Raf Moors – Peter Adriaensen
-References- 123
ATZOM, M. (1990) Phys Rev Lett., 64, pp. 487-490.
BELLOSI, A. et al. (1997) Mater Sci Forum, 238, pp. 255-260.
BENJAMIN, J.S. (1990) Metal Powder, 45, pp.122-127.
BENJAMIN, J.S. et al. (1983) Metal Powder, 34, pp.279-300.
BENJAMIN, J.S.(1976) Sci Amer 234(5),pp.40-48.
BEYERLAIN, I.J., TOME, C.N. and LEBENSOHN, R.A (2003) Modeling texture and
microstructural evolution in the equal channel angular extrusion process, Mater. Sci.
Eng., pp. 122–138.
BEYERLEIN, I.J. et al. (2004) Heterogeneity in texture development in single pass
equal channel angular extrusion, in: ZHU, Y.T. et al. (Eds.), Ultrafine Grained
Materials III. TMS (The Minerals, Metals & Materials Society), pp. 185–192.
BOWEN, J.R. et al. (2000) Mater. Sci. Eng, pp. 87–99.
BROOKES, K.J.A.(1998) Hard metals and other materials. International Carbide, pp.
76-78.
BUDILOV, I.V. et al. (2004) Three dimensional modeling of plastic deformation flow
during ECAP, in: ZHU, Y.T. et al. (Eds.), Ultrafine Grained Materials III. TMS (The
Minerals, Metals & Materials Society), pp. 193–198.
CALKA, A. et al. (1993) Mechanical alloying for structural applications in CALKA,
A. (eds.) Materials Park. Ohio: ASM International, pp. 189-195.
CHIN, Z.H. and PERNG, T.P. (1997) Mater Sci Forum. pp. 235-238 and pp. 121-126.
Raf Moors – Peter Adriaensen
-References- 124
COTTREL, A. (1995) An introduction to metallurgy. London: Edward Arnold.
DOLGIN, B.P. et al. (1986) Non-Cryst Solids p. 87 and pp. 281-289.
ECKHOFF, R F (2003) Dust Explosions in the Process industries (3rd Edition., Gulf
Professional Publishing, ISBN 0-7506-7602-7.
EGAMI, A. et al. (1993), 13th Intern. Plansee Seminar, Reutte, 3, p. 639
EGAMI,A. et al.(1993),13th Intern. Plansee Seminar, Reutte, 3, p. 641.
EPMA (2004a) EPMA Non-Members Area, http://www.epma.ch.
EPMA (2004b) EPMA Non-Members Area
< http://www.epma.com/non_members_area/non_members.htm > (accessed on 23
September 2004).
EPMA (2004c) EPMA Non-Members
Area<http://www.epma.com/non_members_area/non_members.htm> (accessed on 28
July 2004).
EXNER, H.E.(1970) Methods and Significance of Particle- and Grain Size Control in
Cemented Carbide Technology. Powder Metallurgy, 13(26), pp. 429-448.
FISCHMEISTER, H.F., EXNER, H.E. and LINDELÖF, G. (1966) Particle-size
Analysis in Cemented Carbide Technology. Modern developments in Powder
Metallurgy. New York: Plenum Press, pp. 106-124.
Raf Moors – Peter Adriaensen
-References- 125
FRIEDERICH, K.M. and EXNER, H.E.(1984) Metallographical Investigations on
Tungsten Carbide Powders. Prakt. Met., 21, pp. 334 – 341.
FURUKAWA, M., HORITA, Z. and LANGDON, T.G. (2002) Factors influencing the
shearingpatterns in equal-channel angular pressing, Mater. Sci. Eng., p. 97.
GAVRILOV, D., VINOGRADOV, O., SHAW, W.J.D. (1995) in POUSARTIP, A.and
STREET, K. (eds) Proc. Inter. Conf.on Composite Materials, ICCM-10, vol. 3.London:
Woodhead Publishing, p. 11.
GERMAN, R.M. n.d. Powder Metallurgy Science. p. 7.
GERMAN, R.M. n.d. Powder Metallurgy Science. pp.6-7.
GILLE, G. et al. (1999) Advanced Grades of WC and Binder Powder – their properties
and applications.
GILMAN, P.S.and BENJAMIN, J.S.(1983) Annu Rev, Mater Sci, 13, pp. 279-300.
GLEITER,H and MAQUARDT (1984) Z.Metal p. 263.
GONZALEZ et al. (1995) WC-(Fe-Ni-Co) Hardmetals with improved toughness
through isometral heat treatments. Journal of Materials Science, 30, pp. 3435-3439.
GONZALEZ, R. et al. (1995) Mat. Sci., 30, p. 3435.
HANAYALOGLU, C., AKASAKAL,B. and BOLTOM, J.D. (2001) Production and
identification analysis of WC/Fe-Mn as an alternative binder to Cobalt bonded
hardmetals. Material characterization, 47, pp. 315-322.
HONEYCOMBE, R. (1981) Steels—microstructures and properties. London: Edward
Raf Moors – Peter Adriaensen
-References- 126
Arnold, p25.
HONEYCOMBE, R.W.K.(1981) Steels, microstructures and properties. London:
Edward Arnold, pp35-42.
INSITUTE FOR MATERIALS RESEARCH (2004) X-ray Powder Diffraction
<http://materials.binghamton.edu/labs/xray/xray.html> (accessed on 16 August 2004).
IVANOV, E. (1992) Mater Sci Forum., (88), pp. 475-480.
IVANOV, E., et al.(1999) Mater Res, (14) , pp. 377-383.
IWAHASHI, Y. et al. (1996) Principle of equal-channel angular pressing for the
processing of ultra-fine grained materials. Scr. Mater., 35 (2), pp. 143–146.
JENKINS, I.(1993) Introduction. in THÜMMLER, F. and OBERACKER, R.
(eds.)Introduction to Powder Metallargy. London: The institute of Materials, pp. 1-5.
KALOSHKIN, S.D. et al. (1997) Mater Sci Forum, 235, pp. 565-570.
KIM, H.S. (2002) J. Mater. Res. 17 pp.172–179.
KIM, H.S., HONG, S.I. and SEO, M.H. (2001) J. Mater. Res., pp. 856–864.
KIM, H.S., SEO, H.H. and HONG S.I. (2000) On the die corner gap formation in equal
channel angular pressing. Mater. Sci. Eng., pp. 86–90.
KIS, V & BEKE,D.L.(1996) Mater Sci Forum, pp. 225-227, pp. 465-470.
KOCH, C. C. (1991) Processing of metals and alloys. in CAHN, R.W. (ed.) Materials
Raf Moors – Peter Adriaensen
-References- 127
science and technology - a comprehensive treatment. Weinheim: VCH
Verlagsgesellschaft GmbH, pp. 193-245.
KOCH, C.C.(1993) Nanostructured Mater, 2, pp. 109-129.
KOCKS, U.F., TOME, C.N. and WENK, H.R. Wenk (2000) Texture and Anisotropy.
second ed., Cambridge: Cambridge University Press.
LAI, M.O. and LU, L. (1998) Mechanical alloying. Boston: Kluwer Academic
Publishers.
LAOUI, T., FROYEN, L., KRUTH, J.P. (1999) Alternative Binders to Co for
WCparticles for SLS process, Proceedings of the 8th European Conference on
RapidPrototyping and Manufacturing, Nottingham, U.K., July 6-8, pp. 299-311.
LE ROUX, H. and KING, R.P. (1987) The exponential distribution of linear intercepts
of Cobalt in Tungsten Carbide – Cobalt compacts using Automatic image analysis,
Acta Stereol., 6(3), pp. 243 – 248.
LEE, P.Y., YANG, J.L., Lin, H.M.J.(1998) Mater Sci 33, pp. 235-239.
LI, S. et al. (2000) Finite element analysis of plastic deformation and deformation
zones in equal channel angular extrusion. Mater. Sci. Eng. A.
LI, S. et al. (n.d.) Heterogeneity of deformation texture in equal channel angular
Extrusion of copper. Acta Mater.
LUYCKX, S. & ALLI, M.S.(2000) comparison between V8C7 and Cr3C2 as grain
refiners for WC-Co. Materials & Design , 22, pp. 507-510.
Raf Moors – Peter Adriaensen
-References- 128
LUYCKX, S. et al.(1996) Powder Metall. pp.39-41.
Nanostructured Materials (2004) n.d.
<http://www.rpi.edu/dept/materials/COURSES/NANO/crawford/> (updated on 11-24-
97,accessed on 23 September 2004).
NICOARA, G. et al. (1997) Mater Sci Forum pp. 235-238 and pp. 145-150.
NISHIYAMA, Z. (1977) Martensitic transformations, London: Academic Press.
OFFICE OF TECHNOLOGY TRANSFER website n.d. < http://ott.rice.edu> (last
accessed on 27september 2004).
OKADA, K. et al. (1992) Mater Sci Lett., p. 11 and pp. 862-864.
OKADA,K. et al. n.d. Proceed. advances in powder metal & particulate materials.
PARK, J. W. and J.-Y. Suh (2001) Effect of die shape on the deformation behaviour in
equal-channel angular pressing. Metal. Mater. Trans., pp. 3007–3014.
PRAKASH, L. (1993) Proc. 12th Int. Plansee Seminar. Reutte, 2, pp.80-109.
PRAKASH, L.(1979) Doctoral Thesis, University Karltuhe.
PRITCHARD, D.K. (2004) Literature review – explosion hazards associate with
nanopowders. Health and safety laboratory.
<http://www.nanotec.org.uk/evidence/nanopowdersReviewI.pdf> Buxton (last updated
2004, accessed on 23 September 2004).
Raf Moors – Peter Adriaensen
-References- 129
QINETIQ NANOMATERIALS Ltd website(2003) < http://www.nano.qinetiq.com>
(last accessed on 25 september 2004).
RÖDIGER et al. (1998) Microwave sintering of hardmetals. International Journal of
Refractory Metals & Hard Materials, 16, pp. 409-416.
ROEBUCK, B. et al. (1999) Comparison of new and conventional grain size
Measurement methods for WC powders and hardmetals. Hard material powders, pp. 47
– 54.
SANTHAMAN, A.T., TIERNAY, P. HUNT, J.L. (1990) in Metals Handbook Vol.
2,ASM International, p. 950.
SCHUMANN, H. (1965) The form of hexagonal epsilon– martensite in austenitic
steels. Rostock: Wiss Z Univ pp. 423-429.
SCHUMANN, H. (1967) Metallography of the gamma–epsilon– alpha transformation
in high-alloy steels. Prakt Metallogr, pp.275-283.
SCHUMANN, H.(1967) Metallography of the gamma–epsilon–alpha
transformationing high-alloy steels. Prakt Metallogr, 4, pp. 275-283.
SEGAL, V.M. (1995) Materials processing by simple shear, Mater. Sci. Eng., pp. 57–
164.
SEGAL, V.M. (2003) Slip line solutions, deformation mode and loading history during
equal channel angular extrusion, Mater. Sci. Eng., pp. 36–46.
SEGMULLER, A. NOYAN, I.C. and SPERIOSU, V.S. (1989) X-ray Diffraction
Studies of Thin Films and Multilayer Structures, Prog. Crystal Growth and Charact.
18 (5), pp. 21-66.
Raf Moors – Peter Adriaensen
-References- 130
SEMIATIN, S.L., DELO, D.P. and SHELL, E.B. (2000) Acta Mater. 48 pp. 1841–
1851.
SHENZHEN J. New Materials Development Co Ltd. N.d.
<http://newnanomaterial.com> (last accessed on 23 september 2004).
SHIMIZU, K. and TANAKA, Y. (1978) The γ →ε→ α martensitic transformations in
an Fe–Mn–C alloy. Trans Jpn Inst Met, 19, pp. 685–693.
SHIMIZU, K. and TANAKA, Y. (1978) The γ→ε→α’ martensitic transformationsin
an Fe–Mn–C alloy. Japan:Inst Met, pp. 685-693.
SHINGU, P.H. et al. (1998) Met Alloys. Suppm. Trans, Jpn. Inst Met. (3) p. 29.
STOICA, G.M. and LIAW, P.K. (2003) Evaluating the induced strain during equal
channel angular processing, in: LIAW, P.K. et al. (eds.), Materials Lifetime Science
and Engineering, TMS, pp. 119–133.
SUBRAMANIAN, R. et al.(1996) Scripta. Metall., 35(5), p. 583.
SUBRAMANIAN, R. and SCHNEIBEL, J.H. (1997) Intermetallic bonded WC-based
cermets by melt infiltration. Intermetallics, 5, pp. 401-408.
SURYANARAYANA, C. (2001a) Mechanical alloying and milling. Progress in
Materials Science, 46, pp. 1-184.
SURYANARAYANA, C. (2001b) Progress in Materials Science (46) pp. 21-29.
SURYANARAYANA, C. (2001c) Progress in Materials Science, (46), pp. 32-33.
Raf Moors – Peter Adriaensen
-References- 131
SURYANARAYANA, C. (2001d) Progress in Materials Science, (46), p. 32
SURYANARAYANA, C. (1995a) Intermetallics. 3, pp.153-160.
SURYANARAYNA, C. (1995b) International Materials, 40, pp. 41-64.
TAKACS, L. (1996) in SURYANARAYANA, C. et al. (eds) Processing and properties
of nanocrystalline materials. Warrendale: TMS, pp. 453-464.
TAKACS, L., PARDAVI – HORVATH, M.J. (1994) Appl Phys., 75, pp. 5864-5866.
TAMU (n.d.) The history of Equal Channel Angular Extrusion.
< http://www.mengr.tamu.edu/Research/ecae/history.htm > (accessed on 17 October
2004).
TETRONICS Ltd website n.d. <http://www.tetronics.com> ( updated 2004, last
accessed on 27 september 2004).
TETRONICS n.d. < http://www.tetronics.com/pdffiles/nano2.pdf > (last accessed on
29 september 2004).
VISWANADHAM, R.K., LINDQUIST, P.G. and PECK, J.A. (1983) in ed. R. K.
Viswanadham, D. J. Rowcliffe and J. Gurland (eds.) Science of Hard Materials.
NewYork: Plenum Press, p. 873.
VISWANDHAM, R.K. and LINDQUIST, P.G. (1987) Mett. Trans., 18 A, p. 2175.
VOGEL, S. et al. (2002) Mater. Sci. Forum. pp. 408–412 and pp.673–678.
VOGEL, S.C. et al. (2003) Investigation of texture in ECAE materials using neutron
Raf Moors – Peter Adriaensen
-References- 132
diffraction. Mater. Sci. Forum p. 2661.
WHISTON, C. (1987) PRICHARD, F. E. (eds.) X-ray methods. Chichester: ACOL,
p.92
WITTMANN, B., SCHUBERT, W.D., LUX, B. (2002) WC grain growth inhibition in
nickel an iron binder hardmetals. International Journal of Refractory Metals & Hard
Materials, 20, pp. 51-60
YAMAZAKI, T. et al. (1997) Mater Sci Lett., 16, pp. 1357-1359.
ZHANG, F.L a,d WANG, C.Y. and ZHU,M. (2003) Nanostructured WC/Co composite
powder prepared by high energy ball milling, Scripta Materiala,49, pp. 125-142.
Raf Moors – Peter Adriaensen
-Appendix A-
-
Appendix A (Website for the project)
Raf Moors – Peter Adriaensen
-Appendix A- A-2
Table of Contents
Table of Contents...............................................................................................................A-2
A. Website for the project...............................................................................................A-3
A. i. The intropage.....................................................................................................A-3
A. ii. Personal file...................................................................................................A-4
A. iii. Navigation page.............................................................................................A-5
A. iv. Photo and movie page....................................................................................A-5
A. v. Links..................................................................................................................A-6
A. vi. Thesis.............................................................................................................A-6
Raf Moors – Peter Adriaensen
-Appendix A- A-3
A. Website for the project
The easiest way to keep in touch with Belgium, was communication through a website.
The website can be accessed on the following adres: http://users.pandora.be/erasmus-
wolverhampton.
Because it was our first experience with web design, more information involving html
programming and software for building websites had to be searched on the internet. In the
end of this effort, the program Dreamweaver MX 2004 was chosen to commence with the
web design because of the good help-index and the functionality of the program and
everything would be programmed in html.
Before the website could be built, there had to be made a plan view of the different pages
and a decision had to be made about the topics that would be handled. Afterwards, the
programming was a lot easier and the website could be developed.
It was very important that the opening page of the website was saved as index.html
otherwise the website couldn’t get accessed.
Because we wanted to give our website a personal tune, we searched the internet for some
applets, i.e. the clock following the mouse-cursor in the navigation page. Those applets
could be copied easily into the Dreamweaver MX 2004 program.
A. i. The intropage
Figure 81 shows the intropage of the website. By clicking in the picture, information of
person in question will be given while clicking on one of the logo’s of the schools or firm
connect you to their personal website. On the bottom of the page is a counter programme,
so the number of visitors is reported, and a song is inserted.
Raf Moors – Peter Adriaensen
-Appendix A- A-4
Figure 81: Homepage of the website
A. ii. Personal file
A window with some personal data is shown while clicking on a person in the picture of
the opening page. Figure 82 gives you a view of this page.
Raf Moors – Peter Adriaensen
-Appendix A- A-5
Figure 82: Personal data page
Raf Moors – Peter Adriaensen
-Appendix A- A-6
A. iii. Navigation page
The navigation page will open by clicking on the red word “FURTHER”. Figure 83 shows
an example of this page in which navigation to every page is possible. The mouse cursor
gives you the date and the hour of the day while in the middle of the page a movie is
shown with a representation of the city and the university. There’s also the opportunity to
write something in the guestbook. The brown buttons are all Flash buttons that could be
inserted easily in Dreamweaver MX 2004. By clicking on one of them, a certain movement
will be made with a matching sound.
Figure 83: Navigation page of the website
A. iv. Photo and movie page
In these pages, different kinds of pictures and movies of us can be seen. They are all
classified by items or dates as shown in figure 84.
Raf Moors – Peter Adriaensen
-Appendix A- A-7
Figure 84: Movies page
A. v. Links
On this page, shown in figure 85, different links involving our thesis project and our home
are given. By clicking on one of them the website of the involved firm will open.
Figure 85: The link page
Raf Moors – Peter Adriaensen
-Appendix A- A-8
A. vi. Thesis
By clicking on the “thesis”-button, an up-to-date PDF-file will be shown.
Raf Moors – Peter Adriaensen
-Appendix B- B-1
Appendix B (Videoconferencing facilities at UoW)
Raf Moors – Peter Adriaensen
-Appendix B- B-2
Table of contents
Table of contents................................................................................................................B-2
B. i. Introduction............................................................................................................B-3
B. ii. Access Grid............................................................................................................B-5
Raf Moors – Peter Adriaensen
-Appendix B- B-3
B. i. Introduction
Video conferencing has in the past been relatively expensive, but prices are now coming
down considerably, as it is possible for anyone with a fast enough internet connection to
operate a video conference. Sometimes the conferencing takes place over a private network
or VPN, which guarantees better performance, but there will be a trend towards running
video conferences over the public internet as technology improves.
A Virtual Private Network, or VPN, is a private communications network usually used
within a company, or by several different companies or organisations, communicating over
a public network. VPN message traffic is carried on public networking infrastructure (ie,
the Internet) using standard (possibly unsecure) protocols.
VPNs use cryptographic tunneling protocols (the transmission of one data protocol
encapsulated in another) to provide the necessary confidentiality, sender authentication and
message integrity to achieve the privacy intended. When properly chosen, implemented,
and used, such techniques can indeed provide secure communications over unsecure
networks.
Note that such choice, implementation, and use are not trivial and there are many unsecure
VPN schemes on the market. Users are cautioned to investigate products they propose to
use very carefully. 'VPN' is a label which, by itself, provides little except a marketing tag.
Video conferencing can be used for:
conducting interviews
holding meetings
setting up meetings
giving lectures
and has the advantage that it can reduce the need for travel (figure 86). Because
participants in a video conference may be working in different time zones, care must be
Raf Moors – Peter Adriaensen
-Appendix B- B-4
taken with the organisation. Video conferencing can also be used within organisations to
provide immediate telepresence, using internal LAN’s as the communications
infrastructure.
Figure 86: Comparison in cost between communication possibilities
Telepresence means a human/machine system in which the human uses of (head-mounted)
displays and body-operated remote actuators and sensors to control distant machinery.
Provides a virtual environment for humans to control devices, robots, etc., in a hostile or
remote real environment.
Transparent telepresence is the experience of being fully present at a live real world
location remote from one's own physical location. Someone experiencing transparent
telepresence would therefore be able to behave, and receive stimuli, as though at the
remote site.
The resulting vicarious interactive participation in activities, and the carrying out of
physical work, brings benefits to a wide range of users. Examples include the emergency
and security services, entertainment and education industries, and those of restricted
mobility such as the disabled or elderly.
Raf Moors – Peter Adriaensen
-Appendix B- B-5
For any telepresence system there are three essential sub-systems, i.e. the home site
technology which interfaces to the user and the communication link, the communication
link itself which interfaces to the home site and the remote site, and the remote site
technology which interfaces with the communication link and possibly a remote site
human.
Raf Moors – Peter Adriaensen
-Appendix B- B-6
B. ii. Access Grid
The aim of Access Grid is to provide an effective environment for remote group-to-group
collaboration. Whilst this includes various audio and video conferencing components, they
are used within a peer-to-peer model that is coordinated via a separate peer-to-peer
services layer presented as virtual spaces.
Figure 87: View over the Access Grid video conference at RIATec office Wolverhampton
Access Grid is a research project that attempts to provide a sense of presence that
approaches that experienced in face-to-face meetings. The project is largely developed and
supported by a worldwide research community that consists primarily of academic
institutions.
Raf Moors – Peter Adriaensen
-Appendix B- B-7
In order to build a wide scale test bed for building collaboration tools, the Access Grid
project specifically defines the minimum set of requirements necessary to be considered an
Access Grid Node. These requirements are specified in a manner that avoids specific
hardware and software implementations, but rather encourages the use of standards based
digital media tools. This provides a level playing field for research into alternative node
devices and the integration of other hardware and software components.
1. Costs
It is possible to buy an Access Grid node piecemeal and install software, etc. 'inhouse'.
This may be appropriate if Access Grid development is to take place. Another route is to
use an established commercial organisation that is fully involved in and accepted by the
Access Grid community. The product supplied is fully integrated with the Access Grid
system and the virtual spaces model and is based upon software in use by the community.
2. Display quality
An Access Grid node provides a large-scale high-resolution display either by projecting
onto a wall or using a rear projection system. The minimum specifications require
3072x768 pixels, at a distance between 2 and 8 times the height of the projected image.
This allows 18 QCIF and 6 CIF video streams to be displayed; additionally an entire XGA
screen is available for collaborative applications to be used simultaneously.
3. Visual Quality
Each Access Grid node transmits four video streams. These can be used for a variety of
purposes, including multiple camera angles of few participants, whole room shots, close-up
views or to transmit video for other media such as VCRs or document viewers. One
Raf Moors – Peter Adriaensen
-Appendix B- B-8
frequent usage is to use one outgoing feed for a presenter, two for the local audience and
one feed to show remote participants the local display.
4. Audio Quality
The audio quality in an Access Grid node is of a very high quality. The audio stream is
sent uncompressed and sampled at 16 bits at 16 KHz. A single audio stream provides mono
audio, but the system is capable of sending multiple audio streams if necessary.
Figure 88: The micros used for the video conference
The audio component (figure 88) utilises a high-end echo canceller (Gentner XAP400 or
XAP800), level balancer and good quality microphones. Participants have hands-free, full-
duplex audio (i.e. many people can speak simultaneously). The resulting sound is as good
as that experienced when the participants are co-located.
Raf Moors – Peter Adriaensen
-Appendix B- B-9
Figure 89: Audio component with the high-end echo canceller
5. Networking issues
An Access Grid node requires connectivity to an IP-based, multicast-enabled network. It is
possible to interact with the Access Grid via a multicast-unicast bridge (or reflector, or
tunnel) even where the Local Area Network is not multicast-enabled, but this is usually
used as a temporary, stopgap solution. However, the quality of experience using a bridge is
indistinguishable from the experience when using full multicast.
The Local Network connection requires 100Mbps connectivity. The Wide Area Network
requires 10Mbps.
6. Security
The Access Grid Toolkit v1.0 has implemented a trivial proof of concept for security
where an Access Control List is used to allow access to a virtual venue. If a particular user
is allowed into the venue, the venue provides keys with which to encrypt the audio and
video streams shared among the participants. Another approach is to share audio using the
public telephone system.
Raf Moors – Peter Adriaensen
-Appendix B- B-10
Securing shared data is tackled on an ad hoc basis. Most collaborative tools have their own
security features that are utilised. For example when PowerPoint is used to share a
presentation, then the presentation itself is separately held at each node. Only control
events, such as 'Page Down', are broadcast and not the data within the presentation itself.
These control events, whilst not encrypted, are of little use in themselves.
7. Appropriate usage
The Access Grid is well suited for 3-6 participants at each of 2-12 sites. It can be used in
very formal meetings, web casts, classroom style interactions, or for unstructured
interactions where the desire is simply to provide a continual sense of presence with
remote collaborators.
The Access Grid does not provide any formal floor control mechanisms, since that would
be in conflict with the premise that if participants feel more presence then standard social
norms can and will govern interactions. Similarly, formal voting mechanisms are not
supported.
Raf Moors – Peter Adriaensen
-Appendix C- C-1
Appendix C (Technical drawings)
Raf Moors – Peter Adriaensen
-Appendix C- C-2
List of contents
List of contents...................................................................................................................C-2
C. i. Assembly die..........................................................................................................C-3
C. ii. Outer die.................................................................................................................C-4
C. iii. Inner die..............................................................................................................C-5
C. iv. Assembly lower punch.......................................................................................C-6
C. v. Bottom plate...........................................................................................................C-7
C. vi. Lower punch.......................................................................................................C-8
C. vii. Assembly upper punch.......................................................................................C-9
C. viii. Top plate...........................................................................................................C-10
C. ix. Upper punch.....................................................................................................C-11
C. x. Assembly die Charpy test.....................................................................................C-12
C. xi. Outer die...........................................................................................................C-13
C. xii. Inner die............................................................................................................C-14
Raf Moors – Peter Adriaensen
-Appendix C- C-3
C. i. Assembly die
Raf Moors – Peter Adriaensen
-Appendix C- C-4
Raf Moors – Peter Adriaensen
-Appendix C- C-5
C. ii. Outer die
Raf Moors – Peter Adriaensen
-Appendix C- C-6
Raf Moors – Peter Adriaensen
-Appendix C- C-7
C. iii. Inner die
Raf Moors – Peter Adriaensen
-Appendix C- C-8
Raf Moors – Peter Adriaensen
-Appendix C- C-9
C. iv. Assembly lower punch
Raf Moors – Peter Adriaensen
-Appendix C- C-10
Raf Moors – Peter Adriaensen
-Appendix C- C-11
C. v. Bottom plate
Raf Moors – Peter Adriaensen
-Appendix C- C-12
Raf Moors – Peter Adriaensen
-Appendix C- C-13
C. vi. Lower punch
Raf Moors – Peter Adriaensen
-Appendix C- C-14
Raf Moors – Peter Adriaensen
-Appendix C- C-15
C. vii. Assembly upper punch
Raf Moors – Peter Adriaensen
-Appendix C- C-16
Raf Moors – Peter Adriaensen
-Appendix C- C-17
C. viii. Top plate
Raf Moors – Peter Adriaensen
-Appendix C- C-18
Raf Moors – Peter Adriaensen
-Appendix C- C-19
C. ix. Upper punch
Raf Moors – Peter Adriaensen
-Appendix C- C-20
Raf Moors – Peter Adriaensen
-Appendix C- C-21
C. x. Assembly die Charpy test
Raf Moors – Peter Adriaensen
-Appendix C- C-22
Raf Moors – Peter Adriaensen
-Appendix C- C-23
C. xi. Outer die
Raf Moors – Peter Adriaensen
-Appendix C- C-24
Raf Moors – Peter Adriaensen
-Appendix C- C-25
C. xii. Inner die
Raf Moors – Peter Adriaensen
-Appendix C- C-26
Raf Moors – Peter Adriaensen
-Appendix C- C-27
Raf Moors – Peter Adriaensen
-Appendix D- D-1
Appendix D
(Dutch Summary)
Raf Moors – Peter Adriaensen
-Appendix D- D-2
Voorwoord
Dit document is bedoeld als bijlage van onze thesis “bereiding en karakterisering van
submicron / nanogestructureerde poeders van wolfraamcarbide- kobalt / alternatieve
bindmiddelen hardmetalen” . Hierin wordt een Nederlandse samenvatting gegeven van de
thesis. Voor diepgaande uitleg rond bepaalde onderwerpen is het daarom ook aangeraden
de volledige thesis te raadplegen.
We hebben ervoor gekozen om onze thesis in het Engels te schrijven omdat we dit
onderzoek hebben gedaan aan “The University of Wolverhampton” in Engeland. Het
spreekt voor zich dat communicatie met onze promotor ter plaatse op deze manier vlotter
verliep.
We hebben voor dit project drie maanden doorgebracht in Engeland in het kader van een
Erasmus project.
Onze promotor van de Xios Hogeschool Limburg was Dr. Ir. A. Van Bael
De promotor van de University of Wolverhampton was Dr. Ir. T. Laoui
Adriaensen Peter
Moors Raf
- Samenvatting Thesis -
-Appendix D- D-3
Abstract
Tot op de dag van vandaag is kobalt (Co) een van de meest geschikte en meest gebruikte
bindmiddelen voor hardmetalen die gebaseerd zijn op wolfraam-carbide (WC). Een
belangrijke reden waarom Co zoveel gebruikt wordt in deze groep van hardmetalen is het
uitgesproken goede bevochtiginggedrag van Co voor WC.
Door onder andere de slechte corrosievastheid, de hoge kost en giftige eigenschappen, is er
uitgebreid onderzoek besteed aan het vinden van geschikte alternatieve bindmiddelen voor
WC systemen. Dit onderzoek heeft als doel de hoeveelheid Co te reduceren of indien
mogelijk, kobalt volledig te vervangen. Er zijn twee veelbelovende vervangers opgenomen
en gebruikt in dit project. De eerste mogelijkheid is een samenstelling van ijzer (Fe), nikkel
(Ni) en kobalt (Co), en de tweede een samenstelling van ijzer (Fe) en mangaan (Mn). In
vergelijking met Co zijn dit goedkope, niet giftige materialen.
Eerst werd er een grondige literatuurstudie uitgevoerd over hardmetalen, poeder-
preparatiemethoden, poedermetallurgie en nanomaterialen. De submicro/nano-
gestructureerde poeders werden vervolgens klaar gemaakt met behulp van het mechanisch
legeringproces. Hiervoor werd gebruik gemaakt van zowel de planetary ball mill als de
high-energy ball mill.
Een serie van experimenten werd uitgevoerd met de planetary ball mill. Er werden
verschillende bewerkingstijden (2.5, 5, 10 uren) en verschillende rotatiesnelheden (250,
400 rpm) gebruikt om WC-10wt%Co, WC-10wt%FeNiCo en WC-10wt%FeMn te
bereiden. Opvallend was dat bij langere bewerkingstijden (langer dan 2.5 uren voor 150
tr/min) grotere concentraties elementen (Fe, Cr) werden opgenomen van de binnenwanden
en de roestvaste stalen ballen. De hoeveelheid contaminatie nam meer toe naarmate de
rotatiesnelheid opgedreven werd naar 400 tr/min. Dat geeft aan dat zowel de snelheid als
de tijd zo laag mogelijk gehouden dient te worden om de contaminatie te minimaliseren
ofwel dient een hardmetalen container gebruikt te worden. Om deze reden werden er extra
poeders klaargemaakt met behulp van de high energy ball mill.
- Samenvatting Thesis -
-Appendix D- D-4
De korrelgrootte van de WC fase werd uitgerekend door gebruik te maken van de Scherer
vergelijking. Deze vergelijking werd toegepast op de overeenstemmende pieken van het X-
stralen diffractie patroon (XRD). De grootte van de WC deeltjes werd geëvalueerd met een
elektronenmicroscoop (SEM). Er werden succesvol composiet poeders gemaakt waarin
fijne WC deeltjes (submicron tot ongeveer 200 nm in grootte) verdeeld waren in de
matrix (Co, FeNiCo of FeMn).
De volgende stap is het samenpersen van de poeders alvorens aan het sinterproces te
beginnen. Voor de compactiefase werden de geschikte matrijzen ontworpen. Het ontwerp
gebeurde met het CAD softwarepakket Inventor. Er werden twee matrijzen ontworpen: een
eerste voor het maken van cilindrische stukken, gebruikt voor microstructurele- en
hardheidonderzoeken, en een tweede matrijs voor het maken van stukken voor de 3-punts
buigtest. Bij dit ontwerp werd rekening gehouden met ASTM normen.
- Samenvatting Thesis -
-Appendix D- D-5
Inhoudsopgave
Voorwoord.........................................................................................................................D-2
Abstract..............................................................................................................................D-3
Inhoudsopgave...................................................................................................................D-5
Lijst van figuren.................................................................................................................D-8
Lijst van tabellen................................................................................................................D-9
Lijst van symbolen.............................................................................................................D-9
1 Inleiding en doelstellingen.......................................................................................D-10
1.1 Inleiding...........................................................................................................D-10
1.2 Doelstellingen..................................................................................................D-11
2 Literatuuronderzoek.................................................................................................D-13
2.1 Alternatieve bindmiddelen...............................................................................D-13
2.1.1 Fe-Mn als alternatief bindmiddel.............................................................D-13
2.1.2 Fe/Ni/Co als alternatief bindmiddel.........................................................D-13
2.2 Korrelgroei.......................................................................................................D-14
2.3 Poedermetallurgie............................................................................................D-14
2.4 Nanogestructureerde materialen......................................................................D-15
2.4.1 Wat zijn nanogestructureerde materialen.................................................D-15
2.4.2 Mechanisch legeringproces......................................................................D-15
2.4.2.1 Mechanisme van mechanisch legeren..................................................D-16
2.5 Korrelgrootte bepaling van WC.......................................................................D-20
2.5.1 Laser Diffractie........................................................................................D-20
2.5.2 X- stralen diffractie (XRD)......................................................................D-21
- Samenvatting Thesis -
-Appendix D- D-6
2.5.3 Microscoop beelden, SEM, TEM............................................................D-21
3 Experimentele procedure.........................................................................................D-23
3.1 Geteste poedersamenstellingen........................................................................D-23
3.2 Mechanisch legeringsproces............................................................................D-24
3.2.1 Planetary ball mill....................................................................................D-24
3.2.2 High energy mill......................................................................................D-24
3.3 Analyse............................................................................................................D-25
3.3.1 X- stralen diffractie..................................................................................D-25
3.3.2 Scanning elektronen microscoop (SEM).................................................D-26
4 Resultaten.................................................................................................................D-27
4.1 Planetary ball mill............................................................................................D-27
4.1.1 Referentie materiaal.................................................................................D-27
4.1.1.1 XRD patroon van het niet verwerkte referentie materiaal...................D-27
4.1.1.2 Berekende korrelgrootte van de start WC deeltjes..............................D-28
4.1.1.3 Onderzoek naar de vervuilingsgraad (XRF)........................................D-30
4.1.2 XRD resultaten van planetary ball milling aan 250 tr/min......................D-31
4.1.2.1 2,5 uur verwerkte staal met planetary ball milling aan 250 tr/min......D-31
4.1.2.2 Het 5 uur verwerkte staal met planetary ball milling aan 250 tr/min. .D-32
4.1.2.3 Het 10h verwerkte staal met planetary ball milling aan 250 tr/min.....D-33
4.1.3 De SEM foto’s van 250 tr/min stalen met planetary ball milling............D-34
4.1.3.1 Het 2,5u verwerkte staal met planetary ball milling aan 250 tr/min....D-34
4.1.4 XRD resultaten van planetary ball milling met 400tr/min......................D-35
4.1.5 Het 5 uur verwerkte staal met planetary ball milling aan 400tr/min.......D-35
4.1.6 Het 10u verwerkte staal met planetary ball milling aan 400tr/min..........D-36
4.1.7 SEM foto’s van 400 tr/min verwerkte stalen met planetary ball milling.D-36
4.1.7.1 Het 10u verwerkte staal met planetary ball milling aan 400tr/min......D-37
4.1.8 Vervuilingsgraad bij planetary ball milling.............................................D-38
4.1.9 Korrelgrootte berekening.........................................................................D-39
4.1.9.1 Poeder stalen van planetary ball mill aan 250tr/min............................D-40
4.1.9.2 Poeder stalen van planetary ball mill aan 400 tr/min...........................D-40
- Samenvatting Thesis -
-Appendix D- D-7
4.1.9.3 Besluiten van de Scherer vergelijking.................................................D-41
4.1.9.4 Weergave van spanningen door piekverschuivingen...........................D-43
4.1.9.5 Besluiten van de Warren Averbach methode.......................................D-44
4.2 High energy horizontal mill.............................................................................D-44
4.2.1 Schatting van de vervuilingsgraad van het verwerkte poeder.................D-44
4.2.2 SEM foto’s van de stalen gelegeerd met de high energy mill.................D-45
4.2.3 SEM foto’s met Fe/Mn als alternatieve binder........................................D-46
4.2.4 SEM foto’s met Fe/Ni/Co als alternatieve binder....................................D-47
5 Besluiten..................................................................................................................D-48
5.1 De planetary ball mill.......................................................................................D-48
5.2 De horizontal high energy mill........................................................................D-49
5.3 Alternatieve bindmiddelen...............................................................................D-50
5.4 Verderzetting van het onderzoek.....................................................................D-50
- Samenvatting Thesis -
-Appendix D- D-8
Lijst van figuren
Figuur 1: Voorstelling van poeder gevangen tussen 2 ballen MA, SURYANARAYANA, C., 2001________D-17
Figuur 2: Fritch Pulverisette 5 planetary ball mill____________________________________________D-18
Figuur 3: Voorstelling van de ballen in de planetary ball mill, Zoz_______________________________D-19
Figuur 4: Bewegingsbaan van de ballen in de container, Courtesy of Gilson Company, p.17___________D-19
Figuur 5: Opstelling van de High energy mill, Metal-Powder___________________________________D-20
Figuur 6: Voorstelling van de ballen in high energy mill_______________________________________D-21
Figuur 7: Laser diffraction analyse________________________________________________________D-22
Figuur 8: XRD- patroon van niet verwerkt WC-Co referentiestaal________________________________D-29
Figuur 9: SEM foto van start WC < 20 micron_______________________________________________D-30
Figuur 10: grafiek van de verschillende elementen in het referentiestaal___________________________D-31
Figuur 11: XRD patroon van 2,5 uur verwerkt materiaal aan 250tr/min___________________________D-32
Figuur 12: XRD- patroon van 5uur verwerkt materiaal aan 250 tr/min____________________________D-34
Figuur 13: XRD- patroon van 10uur verwerkt materiaal aan 250 tr/min___________________________D-35
Figuur 14: SEM foto van 2,5uur verwerkte staal met de planetary ball mill________________________D-36
Figuur 15: XRD patroon van het 5uur verwerkte materiaal aan 400 tr/min_________________________D-37
Figuur 16: XRD patroon van 10uur verwerkt materiaal aan 400tr/min____________________________D-38
Figuur 17: SEM foto van 10u verwerkt materiaal aan 400 tr/min________________________________D-39
Figuur 18: Schatting van de hoeveelheid vervuilende elementen_________________________________D-40
Figuur 19: Vervuilingsgraad in functie van verwerkingstijd_____________________________________D-41
Figuur 20: Excel rekenblad om de gemiddelde korrelgrootte te bepalen m.b.v. de Scherer vergelijking (250
tr/min verwerkt staal)___________________________________________________________________D-42
Figuur 21: Excel rekenblad om de gemiddelde korrelgrootte te bepalen m.b.v. de Scherer vergelijking (400
tr/min verwerkt staal)___________________________________________________________________D-43
Figuur 22: Fourier analyse van niet verwerkte poeder < 20micron_______________________________D-44
Figuur 23: Resultaten van Winfit!v1.2 op <20µ deeltjes________________________________________D-44
Figuur 24: Korrelgrootte van <20µ WC deeltjes met Winfit!v1.2_________________________________D-45
Figuur 25: Piekverschuiving van WC bij 250 tr/min___________________________________________D-45
Figuur 26: Schatting van de hoeveelheid vervuilende elementen in het poeder______________________D-47
Figuur 27: vervuiling na de verwerking____________________________________________________D-47
Figuur 28: SEM foto van het staal WC-Fe/Mn dat gedurende 1h gelegeerd is in de Horizontally high energy
mill_________________________________________________________________________________D-49
Figuur 29: SEM foto van het staal WC-Fe/Mn dat gedurende 1h gelegeerd is in de Horizontally high energy
mill_________________________________________________________________________________D-50
- Samenvatting Thesis -
-Appendix D- D-9
Lijst van tabellen
Tabel 1 Samenstellingen poeders............................................................................................................D-24
Tabel 2: Samenstellingen binders............................................................................................................D-25
Lijst van symbolen
Co KobaltFe IJzer
H2O WaterHCl Waterstofchloride
HNO3 SalpeterzuurMA Mechanisch legerenMn MangaanNi NikkelPM Poeder Metallurgie
SEM Scanning Elektron microscopieTEM Transmissie Elektronen MicroscoopUoW Universiteit van WolverhamptonVC Vanadium CarbideWC Wolfraam CarbideXRD X- stralen diffractieXRF X- straal fluorescentie
- Samenvatting Thesis -
-Appendix D- D-10
1 Inleiding en doelstellingen
1.1 Inleiding
Metaalpoeders vinden hun toepassing in tal van producten. Telkens bewaren de
poederdeeltjes hun oorspronkelijke eigenschappen. Poedermetallurgie is een proces
waarbij doorgaands een metaal, een legering of een composiet in vaste deeltjes van
maximum 150µm wordt omgezet in een bepaalde vorm. De bekomen vorm wordt meestal
niet meer verder bewerkt.
De basisstappen in poedermetallurgie zijn:
- poederproductie
- poedercompactie
- sintering
Hoewel dit proces reeds eeuwen gebruikt wordt, wordt er de laatste jaren toch meer en
meer de aandacht op gevestigd. Er is een proces ontwikkeld, zogenaamd mechanisch
legeren, waarbij de poeders fijn gemalen worden. Hierbij worden hardmetalen deeltjes
verfijnd en verdeeld in een bindend materiaal. Oorspronkelijk was dit proces enkel bedoeld
voor materialen die hoge temperaturen moesten weerstaan, maar vindt nu toepassing in
verschillende ingenieurstoepassingen. Zo worden de bekomen hardmetalen vaak gebruikt
in de lucht- en ruimtevaart of als snijmateriaal. [Jenkins 1993, pp. 1-5].
Tot op de dag van vandaag is kobalt (Co) een van de meest geschikte en meest gebruikte
binders voor hardmetalen die gebaseerd zijn op wolfraam-carbide (WC). Een belangrijke
reden waarom Co zoveel gebruikt wordt in deze groep van hardmetalen is het
uitgesproken goede bevochtiginggedrag van Co.
De reden om voor het WC-Co systeem te kiezen is tweevoudig; het is een klassiek systeem
dat zeer goede bevochtigingeigenschappen vertoont tussen de twee fases, en dit systeem
- Samenvatting Thesis -
-Appendix D- D-11
heeft aantrekkelijke eigenschappen voor tal van toepassingen [Honeycombe 1981, pp.35-
42].
Er zijn nochtans ook negatieve eigenschappen aan dit systeem met kobalt als bindend
metaal. Onder andere: de hoge kost, de giftige eigenschappen en de slechte
corrosieweerstand van kobalt (Gonzalez et al.,1995). Om deze redenen wordt er heel wat
onderzoek gedaan naar mogelijke vervangmiddelen voor kobalt in dit systeem.
1.2 Doelstellingen
Bij aanvang van deze studie werd eerst een uitgebreid literatuur onderzoek verricht. Dit
was nodig om een zekere basiskennis rond het onderwerp te verkrijgen. De resultaten van
deze studie worden dan ook samengevat weergegeven in het hoofdstuk
literatuuronderzoek.
De tweede fase van deze thesis was het vertrouwd geraken met de verschillende machines
en de aanmaak van een aantal poederstalen. Hierbij werd voornamelijk de aandacht
gevestigd op de procesparameters. De voornaamste parameters zijn verwerkingstijd en
snelheid. Verschillende combinaties van deze parameters werden onderzocht en zijn
beschreven in dit werk.
Het doel van dit werk is het bereiden van submicron/ nanogestructureerde poeders met
behulp van het mechanisch legeringproces. Bij dit proces wordt de grootte van de WC-
hardmetaaldeeltjes gereduceerd .
Een ander aspect in deze thesis was de zoektocht naar alternatieve bindmiddelen. Tot op
heden gebruikt men voornamelijk kobalt als bindmiddel voor wolfraam carbide (WC)
hardmetalen. De voornaamste reden hiervoor is de uitstekende bevochtigingeigenschap van
dit systeem.
- Samenvatting Thesis -
-Appendix D- D-12
Jammer genoeg heeft kobalt ook een aantal negatieve eigenschappen, zijnde: de hoge prijs,
de giftigheid en de corrosiviteit [Gonzalez et al.,1995]. Daarom worden twee mogelijke
vervangingsmiddelen, ijzer/mangaan en ijzer/nikkel/kobalt in deze thesis onderzocht.
Om de eigenschappen van de poederstalen te kennen is het belangrijk een zicht te krijgen
op de afmetingen van de poederdeeltjes. Dit onderzoek werd in eerste instantie verricht
met behulp van X- stralen. Uiteindelijk is gebleken dat deze techniek niet geschikt was
omdat de bekomen poederdeeltjes te klein waren voor deze analyse (WHISTON C. 1987,
p. 92). Na deze vaststelling werd er gekozen voor onderzoek met elektronen microscoop
SEM.
De geproduceerde stalen worden later gecomprimeerd en gesinterd. Na deze stappen
moeten dan de mechanische eigenschappen worden gemeten. Voorbeelden van deze
metingen zijn: hardheidsmetingen, buigproef, kerfslagproef enz. De matrijzen die nodig
zijn voor de compactie en de metingen moesten ontworpen worden volgens de ASTM
standaards. Het tekenen van deze matrijzen gebeurde met het softwarepakket Inventor.
- Samenvatting Thesis -
-Appendix D- D-13
2 Literatuuronderzoek
2.1 Alternatieve bindmiddelen
Een onderdeel van onze thesis was het zoeken naar alternatieve bindmiddelen voor kobalt.
Er werd getracht om de negatieve eigenschappen van kobalt weg te werken door andere
alternatieve binders te zoeken voor kobalt. Eventueel het verminderen van de benodigde
hoeveelheden kobalt heeft ook positieve gevolgen, zowel economisch, technisch als
ecologisch.
2.1.1 Fe-Mn als alternatief bindmiddel
Ijzer- Mangaan legeringen vertonen gelijkaardige karakteristieken als kobalt. Onder andere
smeltpunt, kristalstructuur en fasetransformaties bij koelen zijn gelijk.
Omwille van de hoge slijtvastheid van Fe-Mn staal, werd er van uit gegaan dat deze
legering ook een hoge slijtvastheid voor WC hardmetalen kon leveren [Hanayaloglu et al.
2001, pp. 315-322].
2.1.2 Fe/Ni/Co als alternatief bindmiddel
Prakash was de eerste die (Fe/Ni/Co)-legeringen onderzocht en toonde aan dat harde
metalen met een ijzerrijk bindmiddel een aantal verbeterde eigenschappen had. Hogere
hardheid, slijtvastheid en sterkte zijn een aantal voorbeelden van deze eigenschappen
[Prakash 1993, pp. 80-109 – Prakash 1979].
Bindmiddelen gebaseerd op Fe-Ni/Ni-Co en Fe-Cu-Co worden beschreven als goede
vervangmiddelen voor kobalt (Gonzalez et al., 1995) (Gonzalez et al., 1998).
- Samenvatting Thesis -
-Appendix D- D-14
2.2 Korrelgroei
Deze zeer belangrijke eigenschap kan beschreven worden als het verschijnsel waarbij de
hardmetaal deeltjes in grootte toenemen na sinteren. Dit heeft een negatief effect omdat
men met mechanisch legeren een zo fijn en homogeen mogelijke structuur tracht te
bekomen. Om dit verschijnsel tegen te gaan worden korrelgroeiremmers toegevoegd aan
de poeders. In het onderzoek wordt Vanadium Carbide (VC) gebruikt als korrelgroei
remmer. De invloed van VC wordt bestudeerd door de korrelgrootte na sintering te bepalen
bij een staal met, en een staal zonder VC.
2.3 Poedermetallurgie
Poedermetallurgie is een proces waarbij doorgaands een metaal, een legering of een
composiet in vaste deeltjes van maximum 150µm wordt omgezet in een bepaalde vorm. De
bekomen vorm wordt meestal niet meer verder bewerkt.
Naast de verschillende metaalbewerkingtechnieken, vergt poedermetallurgie een totaal
andere aanpak. Een groot voordeel bij poedermetallurgie is de mogelijkheid om met hoge
kwaliteit, complexe onderdelen met nauwe toleranties, te fabriceren op een economisch
verantwoorde wijze [German n.d., pp. 6-7]. De techniek bestaat uit een aantal
verschillende stappen:
Stap 1: Het poeder wordt gemengd met een geschikt smeermiddel (zinc stearate). Deze
stap heeft enkel als doel het verminderen van de wrijving tijdens de compactiefase.
Stap 2: Het poeder wordt gecomprimeerd. Wanneer de poeders onder een bepaalde druk
komen te staan, wordt een vaste vorm verkregen. Door de cohesiekrachten tussen
de poederdeeltjes is het mogelijk deze compacte vorm te bewerken. De termen
dichtheid en sterkte beschrijven de eigenschappen van deze compacte vorm.
- Samenvatting Thesis -
-Appendix D- D-15
Stap 3: De vaste vorm wordt gesinterd. Het sinteren, gebeurt in een inerte atmosfeer of
onder vacuüm. De temperatuur zal onder het smeltpunt van de hardmetaaldeeltjes
gehouden worden maar boven de smelttemperatuur van het bindend metaal.
Hierdoor zullen de hardmetaaldeeltjes in het vloeibaar geworden bindmiddel
opgenomen worden. De hoeveelheid vloeistoffase mag niet te groot worden zodat
de oorspronkelijke vorm behouden blijft.
2.4 Nanogestructureerde materialen
Waar poeders, met een doorsnede van bijvoorbeeld 150µm, hun eigenschappen danken aan
de wetten van de gewone fysica, spelen de wetten van de kwantum fysica hun rol bij de
nanogestructureerde materialen. Hierdoor veranderen zowel de chemische als fysische
eigenschappen van materialen met een doorsnede van 100nm. Zo zullen bijvoorbeeld
nanogestructureerde keramieken harder en sterker zijn dan hun grovere soortgenoten
[Pritchard 2004]. In dit opzicht is het dus belangrijk kennis te hebben over de mogelijke
methoden om nanogestructureerde materialen te vervaardigen en te onderzoeken.
2.4.1 Wat zijn nanogestructureerde materialen
Waar conventionele materialen korrelgroottes hebben van micrometers tot millimeters,
hebben nanogestructureerde materialen slechts korrelgroottes kleiner dan 100 nanometer.
Hierdoor zit ook een enorm verschil in het aantal atomen dat elke korrel bevat. Zo heeft
een conventionele korrel verscheidene biljoenen atomen waar een gemiddelde
nanogestructureerde korrel er slechts minder dan 900 bevat [Nanostructured Materials
2004].
2.4.2 Mechanisch legeringproces
Deze techniek werd in het onderzoek gebruikt om de wolfraam-kobalt hardmetaaldeeltjes
te verkleinen tot een nano-structuur.
- Samenvatting Thesis -
-Appendix D- D-16
De grondstof voor deze techniek zijn zuivere poederdeeltjes met groottes tussen 1 en
200µm. De startafmetingen zijn niet erg belangrijk, de enige voorwaarde is dat de
poederdeeltjes kleiner zijn dan de ballen. De reden waarom de begingrootte niet erg
belangrijk is, is dat de deeltjesgrootte in verhouding tot de verwerkingstijd exponentieel
afneemt [Koch 1991, pp. 193-245].
De onderzochte legeringen zijn gebaseerd op WC-10wt.% (Fe/Ni/Co) en WC-10wt.%
(Fe/Mn) met toevoeging van 1wt% korrelgroei remmers (VC). Ter vergelijking gebruiken
we een referentiestaal gebaseerd op WC-10wt% Co.
Er zijn twee verschillende technieken gebruikt voor het mechanisch legeringproces. De
eerst gebruikte techniek is de planetary ball mill. Als tweede techniek is er de horizontal
high energie mill. Deze principes worden kort uitgelegd in het hoofdstuk
“literatuuronderzoek” van deze samenvatting.
Het probleem van deze techniek is het verzekeren dat alle deeltjes verkleind worden. Deze
techniek resulteert dan ook in een typische Gauss verdeling met een lange start, welke de
niet verkleinde deeltjes afbeeldt.
2.4.2.1 Mechanisme van mechanisch legeren
Tijdens het verwerken worden de deeltjes herhaaldelijk platgedrukt, koud gelast, gebroken
en opnieuw gelast. Telkens 2 ballen botsen, bevind zich er een hoeveelheid poeder
tussenin. Dit wordt schematisch voorgesteld in figuur 1.
Figuur 1: Voorstelling van poeder gevangen tussen 2 ballen MA, SURYANARAYANA, C., 2001
- Samenvatting Thesis -
-Appendix D- D-17
De deeltjes worden plastisch vervormd door de kracht van de botsingen. Dit leidt tot
verharden en breken van de deeltjes. De nieuwe oppervlakken worden opnieuw koud aan
elkaar gelast waardoor de deeltjes terug groter worden. In het begin van het proces zijn de
deeltjes nog zacht waardoor ze makkelijk opnieuw aan elkaar gelast kunnen worden (in
geval van een combinatie van brosse en taaie materialen). Door het verdere verwerken
worden de deeltjes harder en zullen breken onder invloed van het vermoeiingsmechanisme.
Aangezien de ballen blijven inwerken op de deeltjes zal de structuur ervan blijven
verfijnen.
Hoewel de grootte van de deeltjes niet meer verandert, zullen de tussenlagen verkleinen en
het aantal lagen in het deeltje toenemen [Suryanarayana 2001c, pp. 32-33].
2.4.2.1.1 Planetary ball milling
Een vaak gebruikte machine voor mechanisch legeren is de planetary ball mill; afgebeeld
in figuur 2.
Figuur 2: Fritch Pulverisette 5 planetary ball mill
De naam is afgeleidt van de bewegingen die de containers maken. De basisplaat voert een
draaibeweging uit, terwijl de containers afzonderlijk rond hun as draaien. De draaizin van
de containers afzonderlijk is tegengesteld aan de draaizin van de basisplaat.
- Samenvatting Thesis -
-Appendix D- D-18
De centrifugale krachten veroorzaakt door de rotaties weken in op de inhoud van de
containers. Hierdoor bewegen de ballen in de containers zich langs de wanden van de
containers waardoor het poeder geplet wordt.
Figuur 3: Voorstelling van de ballen in de planetary ball mill, Zoz
Figuren 3, 4 laten zien hoe de ballen zich in de containers bewegen voor en tijdens de
werking van de planetary ball mill.
De containers en ballen zijn verkrijgbaar in verschillende materialen waaronder
wolfraamcarbide, siliciumnitride, chroom staal, Cr-Ni staal, en plastic polyamide
[Suryanarayama 2001a, pp. 1-184].
Figuur 4: Bewegingsbaan van de ballen in de container, Courtesy of Gilson Company, p.17
- Samenvatting Thesis -
-Appendix D- D-19
2.4.2.1.2 High energy ball milling
Met deze techniek kunnen al nano- gestructureerde deeltjes bekomen worden in enkele
minuten. Over het algemeen zal bij langere verwerkingstijden en hogere energie, de
deeltjesgrootte verminderen.
De horizontale high energy mill wordt gebruikt voor zowel academische als industriële
toepassingen.
De hoge kinetische energie geleverd door deze machines resulteert in lage
verwerkingstijden. De kortere verwerkingstijden zorgen er ook voor dat er minder
vervuiling optreedt.
Met dit principe is het enkel nodig de inwendige rotor te laten versnellen. In tegenstelling
tot de planetary ball mill waarbij alle containers moeten versneld worden, wordt bij deze
techniek de energie beter benut.
Figuur 5: Opstelling van de High energy mill, Metal-Powder
- Samenvatting Thesis -
-Appendix D- D-20
Het verwerken kan moeiteloos onder inerte argon atmosfeer gebeuren aangezien de
machine is uitgerust met een airlock systeem.
Figuur 5 toont een afbeelding van de machine in werking, volledig met airlock systeem.
De rotor wordt, zoals getoond in figuur 6, gebruikt om de kinetische energie over te
brengen op de ballen en het poeder.
Figuur 6: Voorstelling van de ballen in high energy mill
2.5 Korrelgrootte bepaling van WC
Om te weten of we te maken hebben met nanogestuctureerde poeders is het natuurlijk
belangrijk om te weten hoe groot de korrels van onze poeders zijn. Ook om de geschikte
instelparameters te zoeken van de machines is dit een belangrijk onderdeel. Er zijn
verschillende technieken beschikbaar om korrelgroottes te onderzoeken. Het probleem dat
zich bij het onderzoek stelt is dat de belangrijke hardmetalen deeltjes opgesloten zitten in
de binder. We zijn echter enkel geïnteresseerd in de afmetingen van de hardmetaaldeeltjes.
Een aantal technieken houden hiermee geen rekening en zijn dus niet bruikbaar. De
bruikbare technieken zullen hier kort besproken worden.
2.5.1 Laser Diffractie
Deze techniek is enkel bedoeld voor WC-Co poeders die klaar zijn voor compactie. Er
wordt informatie gegeven over de spreiding van de deeltjes tussen 2 en 1000µm. De
nodige tijd voor een meting is ongeveer 10min. [1970, pp. 429-448 – Friederich and Exner
- Samenvatting Thesis -
-Appendix D- D-21
1984, pp. 334-341 – Fischmeister et al. 1966, pp. 106-124 – Le Roux and King 1987,
pp.243-248] Een voorstelling van het toestel wordt gegeven in figuur 7. Dit toestel is
bruikbaar voor ons onderzoek maar was echter niet beschikbaar in het labo.
Figuur 7: Laser diffraction analyse
2.5.2 X- stralen diffractie (XRD)
Aanvankelijk werd er van uitgegaan dat deze techniek geschikt was voor ons onderzoek.
Er werd dan ook veel tijd in dit onderzoek geïnvesteerd. Toen de analyse van de resultaten
onmogelijk uitkomsten opleverden volgde een zoektocht naar de mogelijke foutoorzaken.
Uiteindelijk is er gevonden dat deze methode niet geschikt is in het gebied tussen 0,1 en 10
µm (WHISTON C. 1987, p. 92)..
2.5.3 Microscoop beelden, SEM, TEM
Dit zijn zeer directe methodes, die vaak worden toegepast in onderzoek. Vooral SEM is
interessant om naar heterogeniteit te kijken. Voor korrels kleiner dan 0,5µm is een hoge
resolutie nodig. TEM is een techniek die beter toelaat polykristallinniteit te zien. Het
nadeel met TEM is de lage statistische nauwkeurigheid, maar de techniek is bruikbaar voor
- Samenvatting Thesis -
-Appendix D- D-22
heel kleine korrels minder dan 50nm. [Anders 1992, pp. 195-204]. Deze techniek werd ook
gebruikt om de deeltjesgrootte te bepalen nadat er werd vastgesteld dat X –stralen
diffractie niet geschikt was voor de deeltjes die wij moesten onderzoeken (tussen 0,1 en 10
µm) (WHISTON C. 1987, p. 92).
- Samenvatting Thesis -
-Appendix D- D-23
3 Experimentele procedure
Het onderzoek bestaat uit een aantal verschillende fasen. Eerst is er het aanmaken van een
aantal testsamenstellingen. Deze zijn gebaseerd op 90 gewichtsprocent (wt%)
wolfraamcarbide plus korrelgroeiremmer en 10 gewichtsprocent (wt%) binder. Er worden
verschillende samenstellingen als binder getest.
Na afwegen van de samenstellingen worden de poeders gemalen in een van beide
vermalingmachines. Hierin vindt het mechanisch legeringsproces plaats.
Na dit proces worden de poeders verzameld en wordt de grootte van de hardmetaaldeeltjes
onderzocht.
3.1 Geteste poedersamenstellingen
Volgende poeders werden klaargemaakt voor onderzoek:
Naam Wt% WC Wt% Binder Wt%
korrelgroeiremmer
A WC-Co 90 wt%
WC
10 wt% Co
B WC-Co met korrelgroeiremmer 89 wt%
WC
10 wt% Co 1 wt% VC
C WC- Fe/Mn 90 wt%
WC
10 wt% Fe/Mn
D WC- Fe/Mn met
korrelgroeiremmer
89 wt%
WC
10 wt% Fe/Mn 1 wt% VC
E WC - Fe/Ni/Co 90 wt%
WC
10 wt%
Fe/Ni/Co
Tabel 1 Samenstellingen poeders
- Samenvatting Thesis -
-Appendix D- D-24
Samenstellingen van de bindmiddelen:
Binder Wt% Fe Wt% Ni Wt% Co Wt% Mn
1 Fe/Ni/Co 75 15 10 /
2 Fe/Mn 86.5 / / 13.5
Tabel 2: Samenstellingen binders
3.2 Mechanisch legeringsproces
Het legeren gebeurde met twee verschillende machines. Voor de planetary ball mill wordt
de invloed van snelheid en verwerkingstijd onderzocht. Voor de high energy mill worden
de parameters ingesteld zoals ze door de leverancier van de machine [ZoZ.] zijn
meegedeeld.
3.2.1 Planetary ball mill
De poeders zijn verwerkt met een Fritch pulveristette 5 planetary ball mill. Om de optimale
parameters te vinden werd er een referentiestaal verwerkt gedurende 2,5; 5 en 10 uur aan
snelheden van 250 en 400 tr/min. We hebben roestvaste stalen containers en ballen
gebruikt. De gebruikte massaverhouding van bal tot poeder was 15/1. De container werd
gevuld met 125ml ethanol.
3.2.2 High energy mill
Het proces werd geprogrammeerd met het programma Matoz. De eerste stap was het
mengen van de poeders. Dit gebeurde gedurende 5minuten op een snelheid van 200tr/min.
Hierna werd een 60 minuten durend programma uitgevoerd. Dit programma bestond uit 12
- Samenvatting Thesis -
-Appendix D- D-25
herhalingen van 4 minuten werken aan 1000 tr/min en 1 minuut werken aan 600 tr/min.
Het leegmaken van de machine gebeurde gedurende 20 min aan een snelheid van 1200
tr/min. De gebruikte massaverhouding van bal tot poeder was 1:10. De ballen en de
verwerkingseenheid zijn vervaardigd uit roestvast staal. Het proces vond plaats onder een
inerte argon atmosfeer
3.3 Analyse
Zoals al aangegeven is het onderzoek naar de deeltjesgrootte van de hardmetalen zeer
belangrijk. Het gebruik van X- straal analyse kon niet worden toegepast in ons onderzoek.
Omdat dit niet onmiddellijk duidelijk was, werd er aan deze techniek in ons onderzoek
toch veel tijd gespendeerd. Daarom hebben we de resultaten hiervan toch weergegeven.
3.3.1 X- stralen diffractie
X- straal zijn elektromagnetische stralen met een golflengte van 10-10 m, dit is ongeveer de
grootte van een atoom. Deze stralen verschijnen in het elektromagnetische spectrum in
vormen tussen gamma stralen en ultraviolet.
X- stralen diffractie wordt gebruikt ter herkenning van kristallijne materialen en bepaling
van de structuur.
Elke kristallijne vaste stof heeft een unieke X- straal karakteristiek welke kan gebruikt
worden als een “vingerafdruk” ter identificatie. Eenmaal een materiaal geïdentificeerd is,
kan X- stralen diffractie gebruikt worden om de structuur te bepalen, m.a.w. hoe de atomen
op elkaar gestapeld zijn in de kristalstructuur, de interatomaire afstandshoeken enz. Deze
techniek heeft de beperking niet bruikbaar te zijn in het gebied tussen 0,1 en 10 µm
(WHISTON C. 1987, p. 92).
- Samenvatting Thesis -
-Appendix D- D-26
3.3.2 Scanning elektronen microscoop (SEM)
Na de vaststelling dat X- stralen diffractie niet geschikt was om de deeltjesgrootte van de
hardmetalen te bepalen was de enige beschikbare oplossing SEM.
Om een sterke vergroting van de stalen te krijgen, zodat de WC korrels in Co goed
zichtbaar zijn, hebben we een ZEIS scanning elektronen microscoop (SEM) gebruikt. Deze
microscoop was uitgerust met een backscatterd detector. Dit zorgt ervoor dat de elementen
met de zwaarste atomaire massa het lichtst verschijnen op de foto. Een spanning tussen
20kV en 25kV werd gebruikt om deze machine te bedienen.
- Samenvatting Thesis -
-Appendix D- D-27
4 Resultaten
De bekomen resultaten worden onderverdeeld in 2 groepen. Enerzijds zijn er resultaten van
de planetary ball mill, waarvoor de geschikte procesparameters worden gezocht. En
anderzijds zijn er de resultaten van de High energy horizontal mill.
Voor beide machines worden de resultaten van X - straal diffractie (XRD) en onderzoek
naar vervuiling (XRF) weergegeven.
4.1 Planetary ball mill
Het eerste onderzoek werd gevoerd naar de instelling van de juiste parameters voor het ball
milling proces. Om de gegevens te bekomen werd een referentie staal verwerkt op
verschillende snelheden en gedurende verschillende verwerkingstijden. De resultaten van
dit onderzoek worden hieronder weergegeven.
4.1.1 Referentie materiaal
Als referentie staal werd een samenstelling van 90 wt% WC en 10 wt% Co (geen
korrelgroei remmers) gebruikt.
4.1.1.1 XRD patroon van het niet verwerkte referentie materiaal
Figuur 8 geeft het XRD- patroon van het niet verwerkte referentie materiaal
Met behulp van deze grafiek kunnen alle elementen en de gemiddelde korrelgrootte
bepaald worden. Op die manier is het eenvoudiger het verband te bepalen tussen
verwerkingstijd, snelheid en de korrelgrootte.
-Samenvatting Thesis-
-Appendix D- D-28
Figuur 8: XRD- patroon van niet verwerkt WC-Co referentiestaal
De WC pieken op 2θ = 31.6° en 2θ = 35.9° samen met de Scherer vergelijking (zie 4.1.1.2)
worden gebruikt om de gemiddelde WC korrelgrootte te bepalen
Deze pieken zijn gekozen, en zullen in ons volledige werk gebruikt worden, omdat
eventuele vervuiling afkomstig van de containers en ballen niet interfereert met deze
pieken. Op deze manier zijn deze twee pieken representatief voor WC.
4.1.1.2 Berekende korrelgrootte van de start WC deeltjes
De Scherer vergelijking werd gebruikt om de grootte van de WC korrels te berekenen.
κ = 1
λ = X-ray golflengte, λCu = 1,5418
β = FWHM van diffractie lijn
θ = diffractie hoek
-Samenvatting Thesis-
-Appendix D- D-29
De waardes van β en θ kunnen in de XRD data gevonden worden. β is de breedte, in
radialen, van de WC piek die zich op 2θ bevindt. En θ (in radialen) is de halve waarde van
2θ, de positie van de piek. De formule wordt gebruikt op de twee pieken(2θ = 31.6° en 2θ
= 35,6°) waarna een gemiddelde bepaald wordt. Op die manier wordt een grotere
nauwkeurigheid bekomen.
In de start WC korrels zijn de FWHM van beide pieken 0,00635 radialen. The diffractie
hoeken zijn respectievelijk 0,275 en 0,310 radialen. Wanneer deze resultaten in de
vergelijking worden gevoegd levert dit voor de twee pieken:
Het gemiddelde van deze resultaten wordt berekend:
Dit is een onverwacht resultaat omdat de startpoeders gezeefd zijn in een 20µm zeef. Ter
verduidelijking is een SEM foto gemaakt waarmee op een snelle manier de afmetingen
bepaald kunnen worden. Figuur 9 toont de SEM foto.
Figuur 9: SEM foto van start WC < 20 micron
-Samenvatting Thesis-
-Appendix D- D-30
Op deze foto zien we dat de gemiddelde korrelgrootte van de startdeeltjes ongeveer 10µm
is. Uit de literatuur is gebleken dat deze afwijking veroorzaakt is omdat XRD niet geschikt
is voor deeltjes met een grootte tussen 0.1 en 10µm (WHISTON C. 1987, p. 92).
4.1.1.3 Onderzoek naar de vervuilingsgraad (XRF)
Om alle elementen in het poeder te identificeren en de hoeveelheid in massapercenten te
schatten, werd gebruik gemaakt van XRF analyse. In figuur 10, zijn de meest belangrijke
elementen en de geschatte hoeveelheden weergegeven.
0
10
20
30
40
50
60
wt%
W Ni Co Fe Mn Cr
Elements
Estimation of the quantities of the contributing elements the not-milled powder
Figuur 10: grafiek van de verschillende elementen in het referentiestaal
De hoeveelheid W is slechts 60 wt% omdat XRF slechts elementen zwaarder dan natrium
detecteert. Hierdoor wordt de hoeveelheid koolstof niet gedetecteerd. De hoeveelheid
kobalt wordt geschat op 10 wt% De onzuiverheden zijn in kleine hoeveelheden aanwezig
en afkomstig van de poederproductie.
-Samenvatting Thesis-
-Appendix D- D-31
4.1.2 XRD resultaten van planetary ball milling aan 250 tr/min
Hier worden de XRD resultaten bekeken die behaald zijn met de planetary ball mill aan
een rotatiesnelheid van 250 tr/min.
4.1.2.1 2,5 uur verwerkte staal met planetary ball milling aan 250 tr/min
Na verwerking is een XRD test gedaan. Het patroon hier van is weergegeven in figuur 11.
Figuur 11: XRD patroon van 2,5 uur verwerkt materiaal aan 250tr/min
Het eerste opvallende aan dit patroon is dat beide pieken lichtjes naar rechts verschoven
zijn. De nieuwe 2θ waardes van de pieken zijn 31.8° en 35.9° .Dit betekent dat er
spanningen in het poeder zijn. De deeltjes zijn dus niet gebroken maar in elkaar gedrukt.
Een andere manier om deze spanningen waar te nemen is het uitrekenen van de lattice
parameters. Het resultaat van deze berekeningen zal verschillend zijn van de lattice
parameters van het niet verwerkte referentiepoeder. Voor uitleg rond de lattice parameters
verwijzen we u naar de volledige thesis onder hoofdstuk literature review.
De FWHM (= Full Width Half Maximum) van de WC pieken op de 2θ hoeken zijn
respectievelijk 0.00714 en 0.00714 radialen. De FWHM is groter dan deze bij het
referentie poeder. Dit betekent dat er een korrelverkleining is opgetreden.
-Samenvatting Thesis-
-Appendix D- D-32
4.1.2.2 Het 5 uur verwerkte staal met planetary ball milling aan 250
tr/min
Het opvallendste aan dit patroon is de zeer kleine verbreding van de piek. Hierdoor zouden
de korrels in principe even groot zijn.
Zoals men kan zien in figuur 12, zijn beide pieken lichtjes naar rechts verschoven. De
nieuwe 2θ waardes van de pieken zijn 31.8° en 35.9° .Dit betekent dat er spanningen in het
poeder zijn. De deeltjes zijn dus niet gebroken maar in elkaar gedrukt. Ook hier kunnen
weer de lattice parameters uitgerekend worden om de spanningen vast te stellen. Het
resultaat van deze berekeningen zal verschillend zijn van de lattice parameters van het niet
verwerkte referentie poeder.
De FWHM (= Full Width Half Maximum) van de WC pieken op de 2θ hoeken zijn
respectievelijk 0.00555 en 0.00555 radialen. De FWHM is kleiner dan deze bij het
referentiepoeder. Dit betekent dat er een korrelvergroting is opgetreden.
Figuur 12: XRD- patroon van 5uur verwerkt materiaal aan 250 tr/min
-Samenvatting Thesis-
-Appendix D- D-33
Ook hier zijn de pieken lichtjes verschoven naar respectievelijk 2θ = 31.7° en 2θ = 35.9°.
Dit is opnieuw te wijten aan het ontstaan van spanningen door samendrukken van de
deeltjes.
4.1.2.3 Het 10h verwerkte staal met planetary ball milling aan 250 tr/min
Het XRD patroon van dit poeder wordt weergegeven in figuur 13.
Figuur 13: XRD- patroon van 10uur verwerkt materiaal aan 250 tr/min
De twee hoeken van de pieken zijn nu respectievelijk 31.8° en 35.8°. Opnieuw zijn de
spanningen weer zichtbaar aan de hand van de verschuivingen.
De FWHM (= Full Width Half Maximum) van de WC pieken op de 2θ hoeken zijn
respectievelijk 0.00748 en 0.00834 radialen. De FWHM is kleiner dan deze bij het
referentiepoeder. Dit betekent dat er een korrelvergroting is opgetreden.
-Samenvatting Thesis-
-Appendix D- D-34
4.1.3 De SEM foto’s van 250 tr/min stalen met planetary ball milling
Hoofdzakelijk wordt de SEM gebruikt om vast te stellen of de WC deeltjes goed omgeven
zijn door kobalt. Een bijkomende reden om deze analyse techniek te gebruiken is een
duidelijk zicht op de afmetingen van de WC korrels. Op die manier kunnen we deze
resultaten vergelijken met de resultaten van XRD. De lichtere vlekken in op de foto stellen
WC voor, de grijze delen zijn Co. Aangezien X- stralen diffractie onbruikbaar is voor het
onderzoek, is SEM de enige optie om de deeltjesgrootte vast te stellen. Jammer genoeg zijn
er maar een klein aantal SEM foto’s gemaakt omdat tijdens het onderzoek de apparatuur
beschadigd raakte. De oorzaak van dit technisch defect is niet achterhaald. Mogelijk is dit
te wijten aan een installatiefout. De machine werd in ons onderzoek voor de eerste maal
gebruikt. Dezelfde bemerking geldt voor het onderzoek naar stalen die aan 500tr/min
verwerkt werden.
4.1.3.1 Het 2,5u verwerkte staal met planetary ball milling aan 250
tr/min
Figuur 14 toont de SEM foto van het 2.5 uur verwerkte materiaal.
Figuur 14: SEM foto van 2,5uur verwerkte staal met de planetary ball mill
-Samenvatting Thesis-
-Appendix D- D-35
Het is duidelijk dat Wc zich in de Co bevindt. De grootte van de WC korrel in het kobalt is
800,4 nm.
4.1.4 XRD resultaten van planetary ball milling met 400tr/min
Hier worden de XRD resultaten bekeken die behaald zijn met de planetary ball mill aan
een rotatiesnelheid van 400 tr/min.
4.1.5 Het 5 uur verwerkte staal met planetary ball milling aan 400tr/min
Figuur 15 toont het XRD patroon van het materiaal na verwerking.
Figuur 15: XRD patroon van het 5uur verwerkte materiaal aan 400 tr/min
Het valt onmiddellijk op dat beide pieken lichtjes naar rechts verschoven zijn ten opzichte
van het referentiestaal. De nieuwe 2θ waardes van de pieken zijn 31,7177° en
35,7926° .Dit betekent dat er spanningen in het poeder zijn. De deeltjes zijn dus niet
gebroken maar in elkaar gedrukt. Een andere manier om deze spanningen weer te geven is
het uitrekenen van de lattice parameters. Het resultaat van deze berekeningen zal
verschillend zijn van de lattice parameters van het niet verwerkte referentiepoeder.
-Samenvatting Thesis-
-Appendix D- D-36
De FWHM (= Full Width Half Maximum) van de WC pieken op de 2θ hoeken zijn
respectievelijk 0,00748 en 0,008311 radialen. De FWHM is groter dan deze bij het
referentiepoeder. Dit betekent dat er een korrelverkleining is opgetreden.
4.1.6 Het 10u verwerkte staal met planetary ball milling aan 400tr/min
Het XRD- patroon wordt weergegeven in figuur 16. Beide pieken zijn nu gelegen op de 2θ
waardes 31,818° en 35,773°. Dit betekent dat er spanningen in het poeder zijn.
De FWHM (= Full Width Half Maximum) van de WC pieken op de 2θ hoeken zijn
respectievelijk 0,008074 en 0,008786 radialen. De FWHM is groter dan deze bij het
referentiepoeder. Dit betekent dat er een korrelverkleining is opgetreden.
Figuur 16: XRD patroon van 10uur verwerkt materiaal aan 400tr/min
4.1.7 SEM foto’s van 400 tr/min verwerkte stalen met planetary ball milling
-Samenvatting Thesis-
-Appendix D- D-37
Hier worden de SEM resultaten bekeken die behaald zijn met de planetary ball mill aan
een rotatiesnelheid van 400 tr/min.
4.1.7.1 Het 10u verwerkte staal met planetary ball milling aan 400tr/min
Figuur 17 toont de SEM foto van een 10 uur verwerkt staal in de planetary ball mill aan
400 tr/min. De meetlijnen in de foto geven weer dat de afmetingen van WC korrels in Co
tussen 2,273 μm en 239,2 nm liggen. Dit is nogal een groot gebied. Maar op een SEM foto
is slechts een dwarsdoorsnede zichtbaar, en niet de volledige korrel in het Co. De
gemiddelde korrelgrootte zal dus dicht bij het gemiddelde van deze metingen liggen.
Het is duidelijk zichtbaar op de foto dat de WC korrels in het Co zitten.
Figuur 17: SEM foto van 10u verwerkt materiaal aan 400 tr/min
-Samenvatting Thesis-
-Appendix D- D-38
4.1.8 Vervuilingsgraad bij planetary ball milling
Omdat de containers en de ballen van roestvast staal zijn, en de inhoud gebaseerd is op
hard WC, is het vanzelfsprekend dat er slijtage is in de containers. Dit veroorzaakt het
verschijnen van andere elementen in de poederstalen. De belangrijkste elementen zijn Fe,
Cr Ni en Mn. De graad van vervuiling wordt onderzocht met behulp van XRF analyse.
Omdat slechts een klein gebied onderzocht wordt, is dit niet de meest nauwkeurige manier
om de exacte gewichtspercenten van de elementen te onderzoeken. Ondanks dit gegeven is
het een goede manier om een schatting te maken van de graad van vervuiling in het poeder.
Figuur 18 toont hoe de vreemde elementen toenemen bij stijgende snelheid en
verwerkingstijd. Het is duidelijk dat 10u aan 400tr/min een te hoge vervuilingsgraad
oplevert.
0
10
20
30
40
50
60
wt%
W Ni Co Fe Mn Cr
Elements
Estimation of the contributing elements in milled powder
not milled5h@250rpm10h@250rpm5h@400rpm10h@400rpm
Figuur 18: Schatting van de hoeveelheid vervuilende elementen
-Samenvatting Thesis-
-Appendix D- D-39
Contamination after milling
0,3545
34,737
47,026
0,3545
60,353
75,071
010203040
50607080
0 2 4 6 8 10
milling time (h)
Cont
amin
atio
n (w
t%)
250rpm
400rpm
Figuur 19: Vervuilingsgraad in functie van verwerkingstijd
Figuur 19 geeft de vervuilingsgraad in functie van de tijd weer. Het is duidelijk dat hoe
langer we verwerken, hoe hoger de vervuilingsgraad wordt. Er wordt ook duidelijk
weergegeven dat de vervuilingsgraad bij verwerking aan 400tr/min veel hoger is als bij
250tr/min.
4.1.9 Korrelgrootte berekening
Om de meest optimale parameters in te stellen moeten we de invloed op de korrelgrootte
kennen. Een eerste methode om deze groottes te bepalen is met behulp van de Sherer
vergelijking. Deze methode bleek niet geschikt te zijn voor de materialen met
korrelgroottes tussen 0,1 en 10 µm (WHISTON C. 1987, p. 92) Omdat onderzoek naar de
afwijking tussen de verwachtte waardes en de berekende waardes m.b.v. XRD veel tijd in
beslag heeft genomen, geven we toch de resultaten van deze analyse.
-Samenvatting Thesis-
-Appendix D- D-40
4.1.9.1 Poeder stalen van planetary ball mill aan 250tr/min
De korrelgrootte wordt berekend met behulp van een Excel rekenblad en de Sherer
vergelijking. Rekenblad en vergelijking voor de stalen verwerkt aan 250tr/min worden
weergegeven in figuur 20.
Grain size of WC for the 2,5h milled powder at 250rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00714 0.277 224.509 0.02241.542 0.00714 0.313 226.989 0.0227 average 225.749 Angström 0.0226 μm
Grain size of WC for the 5h milled powder at 250rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00555 0.277 288.611 0.02891.542 0.00555 0.314 291.860 0.0292 average 290.235 Angström 0.0290 μm
Grain size of WC for the 10h milled powder at 250rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00555 0.278 288.677 0.02891.542 0.00552 0.312 293.565 0.293 average 291.121 Angström 0.0291 μm
Figuur 20: Excel rekenblad om de gemiddelde korrelgrootte te bepalen m.b.v. de Scherer vergelijking (250
tr/min verwerkt staal)
4.1.9.2 Poeder stalen van planetary ball mill aan 400 tr/min
De korrelgrootte wordt berekend met behulp van een Excel rekenblad en de Sherer
vergelijking. Rekenblad en vergelijking voor de stalen verwerkt aan 400tr/min worden
weergegeven in figuur 21.
-Samenvatting Thesis-
-Appendix D- D-41
Grain size of WC for the 5h milled powder at 400rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00748 0.277 214.280 0.0214
1.542 0.00831 0.312 194.940 0.0194
average 204.610 Angström 0.0204 μm
Grain size of WC for the 10h milled powder at 400rpm (90 wt% WC and 10 wt%
λ β θ dXRD dXRD
1.542 0.00807 0.277 198.516 0.0198
1.542 0.00879 0.313 184.453 0.0184
average 191.484 Angström 0.0191 μm
Figuur 21: Excel rekenblad om de gemiddelde korrelgrootte te bepalen m.b.v. de Scherer vergelijking (400
tr/min verwerkt staal)
4.1.9.3 Besluiten van de Scherer vergelijking
Omdat deze resultaten niet overeenstemmen met de resultaten van SEM moest er een
andere methode gezocht worden om de korrelgroottes te bepalen. Een mogelijke reden
hiervoor was omdat er spanningen in de korrels zaten [Segmüller et al. 1989, pp. 21-66].
Daarom zijn we op zoek gegaan naar een vergelijking die rekening houdt met de
spanningen in de korrels. Een methode die hiervoor geschikt is, is de Warren Averbach
formule. Om deze te gebruiken is er eerst een Fourier analyse nodig van het XRD patroon.
Het programma da wij gebruikten voor deze analyse was Winfit! V1.2 geschreven door S.
Krumm. Figuur 22 toont de fourier analyse van het XRD patroon van de WC pieken van
de niet verwerkte poederdeeltjes met een grootte kleiner dan 20µm..
-Samenvatting Thesis-
-Appendix D- D-42
Figuur 22: Fourier analyse van niet verwerkte poeder < 20micron
Figuur 23 toont de resultaten van het programma
Figuur 23: Resultaten van Winfit!v1.2 op <20µ deeltjes
Figuur 24 toont de berekeningen van de WC korrels. Opnieuw is er een groot verschil
tussen de resultaten van SEM en de resultaten met de software.
-Samenvatting Thesis-
-Appendix D- D-43
Figuur 24: Korrelgrootte van <20µ WC deeltjes met Winfit!v1.2
4.1.9.4 Weergave van spanningen door piekverschuivingen.
Zoals al besproken zijn er spanningen in de deeltjes afkomstig van het verwerken. Hier
wordt een duidelijk beeld gegeven van de verschuivingen van deze pieken.
Figuur 25 geeft de grafieken van niet verwerkt poeder, aan 10u verwerkt poeder aan 250
tr/min.
Figuur 25: Piekverschuiving van WC bij 250 tr/min
-Samenvatting Thesis-
-Appendix D- D-44
De verschuivingen van de pieken zijn zichtbaar tussen de 2 rode verticale lijnen.
4.1.9.5 Besluiten van de Warren Averbach methode
Ook de resultaten van deze analyse komen niet overeen met de verwachtte resultaten. De
verschuivingen van de pieken geven aan dat er spanningen ontstaan tijdens het mechanisch
legeringsproces. Uiteindelijk is gebleken dat deeltjes tussen 0,1 en 10 µm niet kunnen
worden opgemeten met XRD apparatuur (WHISTON C. 1987, p. 92). Hierdoor moeten we
overschakelen op een andere techniek.
4.2 High energy horizontal mill
In dit deel wordt een analyse van het poeder met alternatieve bindmiddelen gemaakt. Een
eerste onderzoek gaat uit naar de schatting van de graad van vervuiling van het poederstaal
WC-10 Co. Dit poeder is met de high energy mill verwerkt gedurende een half uur, 1 uur
en 3 uren.
In het deel planetary ball mill werd al aangegeven dat het onderzoek met XRD geen
bruikbare resultaten opleverde. Daarom werden deze resultaten hier niet nog eens vermeld.
4.2.1 Schatting van de vervuilingsgraad van het verwerkte poeder.
Ook met dit type machine treedt het probleem van vervuiling op. Ook hier zijn de
voornaamste elementen afkomstig van de verwerkingseenheid en ballen Fe, Cr, Ni en Mn.
Er wordt opnieuw de nadruk gelegd op het feit dat XRF slechts een schatting is van de
hoeveelheid vreemde elementen in de samenstelling. Maar de techniek is geschikt voor de
trend van vervuiling in functie van de verwerkingstijd uit te zetten.Figuur 26 toont de
schatting van de meest belangrijke vreemde elementen. Het is duidelijk dat naarmate de
verwerkingstijd toeneemt, de vervuilingsgraad stijgt.
-Samenvatting Thesis-
-Appendix D- D-45
0
10
20
30
40
50
60
wt%
W Ni Co Fe Mn Cr
elements
Estimation of the contributing elements in milled powder
unmilled30min high energy1hour high energy
Figuur 26: Schatting van de hoeveelheid vervuilende elementen in het poeder
Figuur 27 toont dat de vervuilingsgraad in gewichtspercent ongeveer verdubbelt wanneer
de verwerkingstijd verdubbelt.
contamination after milling
0,3545
1,4105
3,683
0
0,5
1
1,52
2,5
3
3,5
4
0 0,5 1
milling time (h)
cont
amin
atio
n le
vel (
wt%
)
estimation of thecontamination level
Figuur 27: vervuiling na de verwerking
4.2.2 SEM foto’s van de stalen gelegeerd met de high energy mill
Het nut van de SEM foto’s was tweeledig. De eerste reden was om te kijken of de binder
omringt was met WC. De tweede reden was om een zicht te krijgen over de grootte van de
-Samenvatting Thesis-
-Appendix D- D-46
WC deeltjes. Het is ook belangrijk om te vermelden dat voor deze experimenten de
composiet deeltjes gebruikt werden.
In de foto’s stellen de kleine lichte puntjes WC voor en de donkere vlekken de alternatieve
binder. De zwarte vlekken komen van het bakeliet dat gebruikt werd om de samples te
maken.
4.2.3 SEM foto’s met Fe/Mn als alternatieve binder
Figuur 28 toont de SEM foto van het sample dat gedurende 1 yur gelegeerd is. De maatstaf
onderaan de foto laat zien dat de WC deeltjes een grootte hebben tussen 1,273 μm en 39,2
nm. Dit is een redelijk brede variatie maar op de foto’s is het niet zichtbaat hoe diep deze
particulen omringt zijn met de binder dus zal de effectieve grootte ongeveer overeen
komen met het gemiddelde van de kleine en grootste deeltjes.
Wat ook meteen duidelijk was, is het feit dat de WC deeltjes omringd waren met de Fe/Mn
binder. De eerste conclusies zijn dan ook dat deze binder een goede alternatieve binder is
voor Co maar dat verdere onderzoeken, aangaande mechanische eigenschapen,
noodzakelijk zijn om te weten of dit werkelijk zo is.
Figuur 28: SEM foto van het staal WC-Fe/Mn dat gedurende 1h gelegeerd is in de Horizontally high energy
mill
-Samenvatting Thesis-
Fe/MnWC
-Appendix D- D-47
4.2.4 SEM foto’s met Fe/Ni/Co als alternatieve binder
Figuur 28 toont de SEM foto van het sample dat gedurende 1 uur gelegeerd is. De maatstaf
onderaan de foto laat zien dat de WC deeltjes een grootte hebben tussen 1,273 μm en 39,2
nm. Dit is een redelijk brede variatie maar op de foto’s is het niet zichtbaat hoe diep deze
deeltjes omringt zijn met de binder dus zal de effectieve grootte ongeveer overeen komen
met het gemiddelde van de kleine en grootste deeltjes.
Wat ook meteen duidelijk was, is dat de WC deeltjes omringd waren met de Fe/Mn binder.
De eerste conclusies zijn dan ook dat deze binder een goede alternatieve binder is voor Co
maar dat verdere onderzoeken, aangaande mechanische eigenschapen, noodzakelijk zijn
om te weten of dit werkelijk zo is.
Figuur 29: SEM foto van het staal WC-Fe/Mn dat gedurende 1h gelegeerd is in de Horizontally high energy
mill
-Samenvatting Thesis-
Fe/MnWC
-Appendix D- D-48
5 Besluiten
De literatuurstudie in deze thesis levert bruikbaar materiaal voor onze begeleidende
doctoraatstudent aan de universiteit van Wolverhampton. De poederstalen die wij hebben
klaargemaakt zullen zeker worden gecomprimeerd en gesinterd. Daarna zullen de nodige
metingen erop worden verricht. Om dit uit te voeren zullen de matrijzen die wij ontworpen
hebben worden gebruikt. Wat ons praktisch onderzoek betreft, kunnen we de besluiten
verdelen in enerzijds planetary ball mill en horizontal high energy mill.
Het eindwerk kadert in een doctoraatsonderzoek met betrekking tot alternatieven voor
kobalt als bindmiddel in hardmetalen op basis van wolfraam-carbide deeltjes. Het praktisch
onderzoek in het eindwerk heeft voornamelijk betrekking op het mechanisch legeren met
behulp van twee uitvoeringswijzen, met name “planetary ball mill” en “horizontal energy
mill”.
5.1 De planetary ball mill
Een serie van experimenten werd uitgevoerd met de planetary ball mill. Er werden
verschillende bewerkingstijden (2.5, 5, 10 uren) en verschillende rotatiesnelheden (250,
400 tr/min) gebruikt om WC-10wt%Co, WC-10wt%FeNiCo en WC-10wt%FeMn te
bereiden. Opvallend was dat bij langere bewerkingstijden (langer dan 2.5 uren voor 150
tr/min) grotere concentraties elementen (Fe, Cr) werden opgenomen van de binnenwanden
en de roestvaste stalen ballen. De hoeveelheid contaminatie nam meer toe naarmate de
rotatiesnelheid opgedreven werd naar 400 tr/min. Dat geeft aan dat zowel de snelheid als
de tijd zo laag mogelijk gehouden dient te worden om de contaminatie te minimaliseren
ofwel dient een hardmetalen container gebruikt te worden.
De resultaten van het vervuilingonderzoek (XRF) geven aan dat bij een snelheid van
400tr/min, de vervuiling bijna verdubbelt t.o.v. 250 tr/min (bij constante
verwerkingstijden). Beelden van de elektronenmicroscoop (SEM) tonen aan dat er geen
verdere deeltjes verkleining gebeurt bij stijgende verwerkingstijden (bij 400tr/min). Hieruit
-Samenvatting Thesis-
-Appendix D- D-49
kunnen we besluiten dat snelheden hoger dan 200tr/min, bij gebruik van roestvaste
containers en ballen, geen voordeel opleveren.
X- stralen diffractie (XRD) werd toegepast om de korrelgrootte van de WC fase te bepalen
(gemiddelde grootte van 24.1 nm). Deze techniek geeft ook aan waar er spanningen
optreden in de korrels van de WC fase ( zichtbaar door de verschuivingen van de pieken in
het XRD patroon). Deze spanningen zijn afkomstig van het mechanisch legeringproces.
De grootte van de WC deeltjes werd geëvalueerd met een elektronenmicroscoop (SEM)
(gemiddelde grootte van 1.103 μm). Er werden succesvol composiet poeders gemaakt
waarin fijne WC deeltjes (submicron tot ongeveer 200 nm in grootte) verdeeld waren in
de matrix (Co, FeNiCo of FeMn).
5.2 De horizontal high energy mill
Om het probleem van de grote vervuiling te vermijden wordt gebruik gemaakt van een
hardmetalen container. Daarom worden extra poeders klaargemaakt in de horizontal high
energy mill. Omwille van de grotere verwerkingscapaciteit en hogere snelheden laat dit
toestel toe poeder met een zeer fijne structuur te bereiden, in kortere verwerkingstijden.
Na de eerste test werd onmiddellijk duidelijk dat met deze techniek veel sneller
nanogestructureerde poeders konden gemaakt worden. Dit is te wijten aan de hogere
kinetische energie van het proces (snelheden boven 1000tr/min), waardoor de
verwerkingstijden verminderd konden worden. Verder leverde deze techniek in
vergelijking met de planetary ball mill veel minder vervuiling op. Dit geven de resultaten
van vervuilinganalyse (XRF) duidelijk weer.
Een bijkomend voordeel van deze techniek is de grote inhoud van de container (2liter).
Hierdoor kunnen relatief grote hoeveelheden poeder in een cyclus verwerkt worden.
Uit voorgaande punten kunnen we besluiten dat de horizontally high energy mill een
hogere efficiëntie heeft dan de planetary ball mill.
-Samenvatting Thesis-
-Appendix D- D-50
5.3 Alternatieve bindmiddelen
Twee alternatieve bindmiddelen ter vervanging van kobalt werden onderzocht en verwerkt
met WC. Het eerste alternatief was Fe/Mn. De SEM foto’s laten duidelijk zien dat de WC
deeltjes goed in de Fe/Mn binder gedrongen zijn. Ook werd hierop duidelijk dat de grootte
van de WC deeltjes verkleind werden door het mechanisch legeringproces. Dit is een
veelbelovend resultaat. De compactiefase en het sinterproces zullen in een latere fase
worden uitgevoerd.
De zelfde resultaten werden opgetekend voor het tweede alternatieve bindmiddel
Fe/Ni/Co. SEM foto’s geven weer dat de WC deeltjes goed verdeeld werden over het
bindmiddel. Ook hier worden De compactiefase en het sinterproces zullen in een later fase
worden uitgevoerd. Ook hier worden de compactiefase en het sinterproces een latere fase
uitgevoerd.
5.4 Verderzetting van het onderzoek
Verdere stappen in dit onderzoek zijn de compactiefase en sintering van de verschillende
poederstalen. Parameters die in invloed hebben op de compactiefase en sintering moeten
onderzocht en geoptimaliseerd worden voor elke poedersamenstelling (WC-Co, WC-
FeNiCo, WC-FeMn). Door juiste instelling kan vermeden worden dat de WC korrels met
submicron/ nano- structuur, terug te groot gaan worden.
De fysische en mechanische eigenschappen van gesinterde stalen moeten worden
onderzocht. Dit kan gebeuren door; hardheidstesten, 3punts buigtesten en
dichtheidsmetingen. Ook moet na sintering de microstructuur, waaronder de verdeling en
grootte van WC korrels, opnieuw gecontroleerd worden.
-Samenvatting Thesis-