Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
In-Situ Thermal Analysis Probe
by
Sa'ed Awni Musmar
Department of Mining, Metals and Materials Engineering
McGill University, Montreal
A thesis submitted to the faculty of Graduate Studies and Research in partial fulfillment of the requirements of the degree of Doctor ofPhilosophy
© Sa'ed Awni Musmar
1+1 Library and Archives Canada
Bibliothèque et Archives Canada
Published Heritage Branch
Direction du Patrimoine de l'édition
395 Wellington Street Ottawa ON K1A ON4 Canada
395, rue Wellington Ottawa ON K1A ON4 Canada
NOTICE: The author has granted a nonexclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell theses worldwide, for commercial or noncommercial purposes, in microform, paper, electronic and/or any other formats.
The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.
ln compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis.
While these forms may be included in the document page cou nt, their removal does not represent any loss of content from the thesis.
• •• Canada
AVIS:
Your file Votre référence ISBN: 978-0-494-27823-9 Our file Notre référence ISBN: 978-0-494-27823-9
L'auteur a accordé une licence non exclusive permettant à la Bibliothèque et Archives Canada de reproduire, publier, archiver, sauvegarder, conserver, transmettre au public par télécommunication ou par l'Internet, prêter, distribuer et vendre des thèses partout dans le monde, à des fins commerciales ou autres, sur support microforme, papier, électronique et/ou autres formats.
L'auteur conserve la propriété du droit d'auteur et des droits moraux qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.
Conformément à la loi canadienne sur la protection de la vie privée, quelques formulaires secondaires ont été enlevés de cette thèse.
Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant.
Abstract
A new thennal analysis technique was developed and tested. It makes use of the
improvements in heat transfer characteristics associated with recent advances in heat pipe
technology. Heat is extracted from a liquid sample of a melt taken in-situ from within a
vessel or fumace. The rate of heat extraction is such as to cause the sample to solidify.
The technique was tested both in the laboratory and on an industrial scale (Grenville
Castings, Perth, Ontario). Aluminum alloys including 356, 319, AI-xSi, AI-Si-Cu-xMg,
and 6063 were subjected to various melt treatments and were used to carry out the tests.
Classical thennal analysis was also carried out simultaneously under the same melt
conditions using a preheated graphite cup.
The comparison showed that the new technique has great potential over classical thennal
analysis. The major advantages of the new method are that it conducts the analysis inside
the melt (since it is no longer necessary for a physical sample to be removed from the
melt itself), it consumes less time and the cooling rate can be precisely controlled during
the solidification process. Moreover, it produces curves of greater detail and of better
resolution than conventional techniques. In fact, the detail is of such resolution that, in
sorne cases, the cooling curves may be used to infer the chemical composition of certain
components of the melt, a fact which equates to a fonn of rapid chemical analysis. The
peaks in the signal which refer to intennetallic fonnation are of better resolution and
more identifiable when the new technique is used. The size of the peaks obtained using
the new probe is about three times greater than that obtained by the classical method.
With this new technique it becomes possible to correlate the area below the intennetallic
peak to the concentration of iron or copper in the melt. This is a feature which makes the
new thennal analysis probe act as a rapid chemical analyzer for selected constituents.
11
RÉSUMÉ
Une nouvelle technique d'analyse thermique a été développée et testée. Elle utilise le
perfectionnement dans des caractéristiques de transfert thermique liées à un avancement
récent dans la technologie de caloduc. La chaleur est extraite à partir d'un échantillon
liquide d'une fonte prise à l'intérieur dans un cuve ou un four. Le taux d'extraction de la
chaleur provoque la solidification de l'échantillon. La technique a été examinée au
laboratoire et dans un institut industriel (Fonderie Grenville, Perth, Ontario). Des alliages
d'aluminium de type 356, 319, AI-xSi, AI-Silicium-Cu-xMg et 6063 sont tous soumis à
des traitements divers de métal liquide et ils ont été employés pour réaliser les essais.
L'analyse thermique classique a été également exécutée simultanément à l'aide d'un
moule en graphite préchauffé, et les coulées ont été réalisées dans les mêmes conditions
respectées auparavant.
La comparaison a montré une efficacité meilleure de la nouvelle technique par rapport à
celle de l'analyse thermique traditionnelle. Les avantages principaux de la nouvelle
méthode sont qu'elle conduit à l'analyse à l'intérieur du métal liquide (c.-à-d, l'échantillon
n'est pas enlevé de la fonte), ceci consomme moins de temps et le taux de refroidissement
peut être contrôlé avec précision pendant le procédé de solidification. D'autre part, cette
nouvelle application produit des courbes de solidification avec plus de détails et avec une
meilleure résolution en comparaison à des techniques conventionnelles. En fait, le détail
111
peut être d'une telle résolution que, dans certains cas, les courbes de refroidissement
peuvent être employées pour révéler la composition chimique (quelques composants) de
la fonte (c.-à-d, analyse chimique rapide). Les pics dans le signal qui se rapportent à la
formation d'intermétalliques sont plus clairs et plus identifiables (d'une meilleure
résolution) lorsque la nouvelle technique est employée. Les tailles maximales des pics
obtenues par cette nouvelle méthode sont environ trois fois plus grandes que celles
obtenues par la méthode classique. Avec cette nouvelle technique, il est possible de
corréler l'aire sous le pic d'intermétalliques avec la concentration du fer ou du cuivre
dans la fonte. Ce dispositif laisse la nouvelle technique d'analyse thermique se comporter
comme si c'est un analyseur chimique rapide pour des constituants choisis.
IV
Acknowledgments
1 would like to express my gratitude to my supervisors, Prof essor Frank Mucciardi,
Prof essor Fawzy H. Samuel, and Prof essor John Gruzelski whose expertise,
understanding, and patience added considerably to my graduate experience. 1 appreciate
their vast knowledge in many areas such as light metals, their research skills and ethics,
interaction and their assistance in scientific writing.
Very special thanks go to Prof essor Agnes Samuel for her insight, opinion, precious
discussions, assistance and the family environment she provided during my stay at
UQAC. AIso, 1 would like to acknowledge Dr. Florence Paray for her useful discussions,
kindness and encouragement throughout my study.
Financial and in-kind support received from the Natural Sciences and Engineering
Research Council of Canada (NSERC), REGAL-McGill, and REGAL-UQAC is
gratefu11y acknowledged. 1 would like to thank Dr. Herbert W. Doty of GMPT (USA) for
his stimulating discussions and for providing the chemical analysis for a11 the samples
presented in this study, and Mr. Paul Burke for accommodating us at the Grenville
Castings Ltd. plant to carry out the industrial tests on the production line.
1 wish to acknowledge the valuable technical assistance provided Mr. Walter Greenland
ofMcGill, and MM Alain Bérubé and Regis Boucher ofUQAC.
v
1 would like to thank Ms Marion Sinclair for proofreading the thesis and for providing
valuable language tips which 1 enjoyed and appreciated.
1 enjoyed the friendly multi-cultural environment which 1 experienced at McGill and at
UQAC. Many thanks go to Hicham Tahiri from Morocco who kindly translated the
abstract to French, Kaled Elalem from Libya, Hany Ammar, Mohammad Adel,
Mahmood Tash, Ahmed Nabawy, Yasser Zedan, Emad EI-Gallad and Osama EI-Sebaie
from Egypt, Hujun Zahoo from China, Neivi Andrade and Javier Tavitas from Mexico,
Ehab Samuel, Pietro Navara, and Mathiew Baradies from Canada, Najendra Terepaty and
Shamiek from India, Saeed Shabestari, Reza Hafeshari, Sadeq Fyrozy and Javad from
Iran, and Naser Eloqyly from Saudi Arabia. 1 also wish to acknowledge the
encouragement 1 received from my Jordanian friends, including in particular Mohammad
Twalbeh, Mohammad Alakhrass, Anass Alazzam, Feras Abualhassan, Mohammad
Alkhaleel, and Mohammad Aljarah.
1 wish to thank my brothers Dr. Ayman and Osama, my brother-in-Iaw Mohammad and
my sister Suhier, for their continuous encouragement during my studies, and my fiancée
Mayes for her support during the final stages of writing this thesis.
Finally, but most importantly, 1 wish to express my deep gratitude to my parents to
whom 1 am indebted for their constant encouragement and moral and financial support
without which 1 would not have been able to complete my study.
VI
TABLE OF CONTENTS
ABSTRACT . ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 1
RESUMÉ ... ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 111
ACKNOWLEDGMENT .•............................................................... v
TABLE OF CONTENTS ............................................................... Vll
LIST OF FIGURES . •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• Xl
LIST OF TABLES
NOMENCLATURE
CHAPTER ONE Introduction
1.1 Objectives of the Present Stndy ................................•.........••.•..... 1
1.2 Outline of the Present Study ....................................................... 5
CHAPTER TWO Heat Pipe
2.1 Introduction .......................................................................... 8
2.2 Operating Principle of the Heat Pipe .............................••.•.......... 9
2.3 Types of Heat Pipe ................................................................. 11
2.4 Heat Pipe Applications ............................................................ 12
2.5 Limitations of Classical Heat Pipes ............................................. 13
2.5.1 Film Boiling ................................................ 14
2.5.2 Entrainment ............................................... 17
2.5.3 Sonic Chock ............................................•... 18
Vll
2.6 McGi11 Beat Pipe ................................................................... 20
2.7 Advantages of McGiII Peat Pipe over Classical Heat Pipes ............... 25
CHAPTER THREE Melt Treatment and Thermal Analysis
3.1 Introduction ......................................................................... 31
3.2 Grain Refining ...................................................................... 31
3.2 Eutectic Modification ............................................................. 32
3.3 Effects of Adding Elements to the Melt ........................................ 34
3.4 Thermal Analysis .................................................................. 35
3.4.1 Salient Features of the Cooling Curve .....••......• 38
3.4.2 Classical Thermal Analysis Equipment ....•....... 40
CHAPTER FOUR the New Probe Design
4.1 Introduction ........................................................................... 50
4.2 Characteristics of the Beat Pipe .................................................. 51
4.3 Preliminary Design Considerations ................•..•....•......•.•..•.......• 55
4.4 Selection of the Heat Pipe Components ..........................•.•............ 58
4.4.1 Evaporator and Condenser Material Selection .... 58
4.4.2 Working Fluid Selection ................................ 60
4.4.3 Coolant Fluid Selection ......••.•..••..•.•..••••••••.•.• 61
4.5 Cooling System Design ............................................................. 61
CHAPTER FlVE New Thermal Analysis Technique
Part One Experimental Procedures
5.1 Introduction ........................................................................... 64
5.2 First Probe ............................................................................ 65
5.3 Second Probe ......................................................................... 68
Vlll
5.4 Final Design ........................................................................... 69
S.5 Experimental Setup ................................................................. 74
5.5.1 Measurements and Instrumentation ................... 75
5.5.2 Experimental Methodology ..•..•.............••••...•••. 78
5.5.2.1 The Cooling Period ...................•..•...••...••• 82
5.5.2.2 The Heating or Remelt Period .................... 82
Part Two Primary Tests
5.6 Hot Environment Test .............................................................. 84
5.7 Molten Metal Environment Test ................................................. 87
CHAPTER SIX Results and Discussion
6.1 Introduction ........................................................................ 95
6.2 Grain Refinement ................................................................. 96
6.3 Eutectic Modification ........................................................... 104
6.4 Detecting and Quantifying Iron Intermetallies ............................ 112
6.5 Detecting and Quantifying Copper Intermetallies ...................••.•. 123
6.6 Detecting of Magnesium Intermetallies ....................••.•........•.... 129
6.7 Detecting Liquidus Temperatures ............................................ 134
6.8 Detecting Minor Reactions in Wrought Alloy 6063 ....................... 136
6.9 Effects of Boundary Conditions on the Quality
of the Results Produeed by the Probe ........................................ 149
6.10 Industrial Experiments ......................................................... 153
IX
CHAPTER SEVEN Conclusions, Originality, and Suggested
Future Work
7.1 Conclusions .............................................................•••.•.••••.. 163
7.2 Statement of Originality ...............................................•...•..... 166
7.3 Future Work ....................................................................... 167
x
LIST FIGURES
Fig.2.1 Schematic diagram (a) Heat pipe, (b) Thermosyphon •.•.•...••••.••••••••••.. 11
Fig.2.2 Schematic diagram oflimits on heat pipe performance •.••••••••••••••••.•.• 14
Fig. 2.3 Typical boiling curve for water at atomospheric pressure ••••••••••••••••••• 17
Fig. 2.4 Sketch of McGi11 heat pipe .......................................................... 22
Fig. 3.1 Actual cooling curve for 356 aluminum alloy ••••.•••.•••••••••••••••.•••••.••••• 36
Fig.3.2 First derivative curve associated with actual ••••••.••.••••••••..•••••••••••••••• 37
cooling curve for 356 aluminum alloy
Fig. 3.3 Schematic diagram of a cooling curve ............................................ 39
Fig. 3.4 Sampling cup used for thermal analysis ••••••••••.••.•••••••••••••••••••••••••. 41
Fig. 4.1 Schematic diagram for the tirst thermal analysis probe .•••.•••••••••..••••••• 53
Fig. 4.2 Picture ofthe sodium thermal analysis probe (second probe) .•••••••••.••••• 54
Fig. 4.3. Evaporator section of the new probe .............................................. 56
Fig 4.4. Beat pipe evaporator components .................................................. 57
Fig. 4.5 A schematic drawing of the condenser cooling arrangement •.••.•••••••.•••••..••• 62
Fig. 5.1. First design, evaporator section .................................................... 66
Fig. 5.3. Outer surface of tirst probe after experiments
in a Iiquid metal environment ....................................................... 68
Fig. 5.4 Second probe ........................................................................... 69
Fig. 5.5 Third design ........................................................................... 73
Fig. 5.6 Experimental Setup .................................................................. 7S
Fig. 5.6 First probe un der elevated tempe rature conditions .•......••.•.•...•••..•.•••• 84
Fig. 5.7 Temperatures at 5mm from both sides ofthe probe surface •.••••••••••••.•• 85
Fig. 5.8 Energy absorbed by the probe in a hot environment •••••••••••••••••••••••••• 86
Fig. 5.9 Cooling curves obtained by the tirst probe and sampling cup ••..•••.•.•..••• 88
Fig. 5. 10. Second probe design, with four thermocouples .•..•..•....•••.••.•••••.•.•.• 89
Xl
Fig. 5.11. Two concentric heat pipes ......................................................... 90
Fig. 5.12(a) Cooling curve obtained by the third probe ......•.••.••...•••...•...•••..•.. 91
Figure 5.12 (b) Cooling curve obtained by sampling cup technique ..••...•.••••••••.• 91
Figure 5.13 (a) First derivative of the curve in Fig. 5.12 ..••.••.••.•....•..•••.•.•..•••.• 92
Figure 5.13 (b) Traditional thermal analyses, first derivative curve •..•••.••••••••••• 92
Fig. 6.1 Microstructures of 356 Al alloy with addition of Ti •.••.••••••.••••••••••••.•••• 98
Fig.6.2. Variation of 356 grain size with the amount of Ti added to melt .•••.•.•..•• 99
Fig.6.3 Undercooling portion of the cooling curve ..•..•••••.•..•..••.••..•••.•••.•••..•. 100
Fig.6.4 Variation of the undercooling portion parameters
with Ti concentration in the melt ................................................ 102
Fig.6.5 Eutectic regions of the cooling curves with various Sr levels •.•.....•••..•... 105
Figure 6.6 Variation in undercooling with Sr concentration in the .•...••.•.•••.••.• 106
Figure 6.7 Image analysis results ............................................................ 1 08
Figure 6.8 Eutectic temperature as it varies with the strontium ••••••.•••••.••••••.•• 109
Figure 6.9 Variation ofDAS with levels of Sr content in the melt ••••••.•.•.•.••.•..• 110
Figure 6.10 Microstructure at two different levels of modification .................. 111
Figure 6.11 Micrographs for two different levels of modification .................... 111
Fig. 6.12 Cooling curves and associated first derivative curves for 356 Al with 0.93
wt% Cu ............................................................................... 116
Fig. 6.13 Cooling curves for different iron concentrations in 356 Al alloy ••••••••.. 117
Figure 6.14 Formation temperature of AlsFeSi at different iron concentrations •• 118
Fig. 6.15 Portions of the first derivative curves from in-situ probe
corresponding to the iron intermetallic phase •.•••••••••••••••••••••.••.•.••..••••••••••••• 119
Figure 6.16 Cooling curve and associated second derivative curve for 356 Al alloy
with 0.93 wt% Fe ................................................................................ 119
XlI
Fig. 6.17 Average surface fraction and areas below the intermetallic peaks as they
vary with the iron concentrations in the melt •.••.••.••.••••••.•.••••••••••••• 120
Figure 6.18 Backscattered images ........................................................... 122
Figure 6.19 Cooling curves and associated first derivative curves for
AI-Si altoy with 3.7 wt% Cu ................................................... 124
Fig. 6.20 Average surface fraction and area below the intermetallic
peak as they vary with the copper concentration in the melt .•.•.•.•.••••• 125
Figure 6.21 Cooling curve and fint derivative curve for the
AI-Si altoy with 2.81 wt% Cu ................................................ 126
Fig. 6.22 Backscattered electron images ................................................... 128
Fig. 6.23 Cooling CUrve for Al-Si-Cu ternary alloy with 0.2 wt% Mg •...•...••...•. 130
Figure 6.24 Cooling curve for Al-Si-Cu ternary alloy with 0.2 wt% Mg .•••.•.•.••. 131
Fig. 6.25 Cooling curve for Al-Si-Cu ternary alloy with 0.3 wt% Mg ..••...•.•.••.• 132
Figure 6.26 Cooling curve for Al-Si-Cu ternary alloy with 0.6 wt% ••.••..•••...••.• 134
Fig. 6.27 Variation of liquidus temperature ............................................... 136
Fig. 6.28 Cooling cUJ-v'es for 6063 alloy ..................................................... 140
Fig. 6.29 Cooling curves of 6063 alloy with 200 ppm Sr ••••••••••••••••••••••••••••••••• 141
Fig. 6.30 Cooling curves of 6063 alloy with 0.5 wt % added Fe •.••.•.•...•••..•••..•.. 143
Fig. 6.31 Cooling curves for 6063 alloy with 0.7 % Mn •.••.•••••.••..••.••.•.••.•••••••• 145
Fig. 6.32 Backscattered images ............................................................... 149
Fig. 6.33 Thermocouple locations in the solidifying sam pIe ••.••.••.•..•.•••••.••••••• 151
Fig. 6.34 Cooling curves for 319 alloy at different thermocouple locations •••••..•. 152
Fig. 6.35 Cooling curves and associated fint derivatives for 319 alloy ..••••.•••••••• 153
X1l1
Fig. 6.36 Experimental setup at Grenville Castings Perth Plant •••••••••••••••••••••• 154
Fig. 6.37 Cooling curve and associated first derivative of 356
aluminum alloy (as used by Grenville Castings Ltd) •••••••••••••••••••••••• 159
Fig. 6.38 Cooling and heating cu.-ves ....................................................... 159
Fig. 6.39 Solidified physical sampling as taken by probe .••.•••••••.••.••••••••••••••• 160
XIV
LIST OF TABLES
Table 5. 1. Chemical analysis ofthe alloys used in the present study .••••..••••••.•.• 80
Table 6.1 Chemical analysis of phases 1 and 2 which appear
on the backscattered image of the base alloy (6063) •••••.••.•.••••.••.•••••••••••• 146
Table 6.2 Chemical analysis of phases 1, 2 and 3which appear
in the backscattered ............................................................... 147
Table 6.3 Chemical analysis of phases 1 and 2 which appear
on the backscattered image for base alloy (6063) ••••••••••••.•••.•.••••••.• 147
Table 6.4 Chemical composition of the 356 Al alloy used in industrial tests ••••.•• 154
xv
NOMENCLATURE
d
f
g
H
qc,ver
flow cross-sectional area
inside diameter
free space dimension;
inner diameter of the outer cylinder;
outer diameter of the inner cylinder;
flow modifier diameter;
return line diameter.
friction factor
gravitational acceleration
twist pitch
latent heat of evaporation
effective thermal conductivity of the liquid- wick combination;
effective evaporator length;
maximum capillary pressure
Critical heat flux
Critical heat flux from a vertical surface
axial heat flux
inner radius of the heat pipe wall;
XVI
vapour core radius
nucleation site radius
effective capillary radius.
difference between the surface and fluid temperatures
vapor temperature
v vapor velocity,
y . . d 1 H twist ratIO an equa to-d
z dimension related to the wick spacing
(j surface tension
Pv,Pt density of vapor and liquid, respectively
Weber Number
Re Reynolds number
Pr Prandtl number
NUtr:> Nusselt number for a straight tape and insert (y = 00).
aFe AIls (Mn Fe)3 Sh phase
pFe AIs Fe Si phase
thickness of the twisted tape
dynamic viscosity at bulk fluid and tube wall temperatures, respectively
viscosity ratio equal to (;: r, n ~ 0.18 for liquid heating, and n ~ 0.3 for
liquid cooling.
latent heat of vaporization
XVll
Chapter One
Introduction
1.1 Objectives of the Present Stndy
Quality control is now considered to be one of the most important steps in the casting
production process in any foundry due to intense competition on the global markets. in
the past after casting a certain product it was inspected, and if any defect was found the
product would be rejected. Subsequently, it would either be recycled if possible, or
scrapped, thereby increasing the production costs and reducing the productivity of the
foundry. Thus, it is of primary importance to control the melt quality from the outset of
the casting process.
Thermal analysis is deemed one of the least costly ways used in the industry to monitor
the level of grain refiners and eutectic modifiers in the melt prior to casting. The
temperature of a solidifying sample is recorded as it cools down from a completely liquid
state, passing through the solidifying region, until it becomes completely solid. The
temperature-time curve is the basic output of thermal analysis. The shape of this curve is
dependent to a high degree on the metallic phases which form during solidification. As
liquid solidifies, it evolves latent heat. The quantity of this heat depends on the solid
phases formed during solidification. The emergence of the solid phase affects the rate of
decrease in the sample temperature, and consequently, thermal arrests appear on the
temperature-time curve.
1
The automation of industrial processes has been of major interest to researchers over the
years. The specific objective of this thesis is to explore the potential for automated
thermal analysis and to develop a technique which is capable of carrying out pertinent
tests, in-situ, which can be controlled simply by pressing a button in the control room.
Among the several subsidiary objectives of this research may be included a means for
minimizing human intervention, enhancing accuracy, and increasing productivity while at
the same time reducing costs and increasing profitability for aluminum foundries.
This technique should, ideally, have the ability to extract heat from a specific sample
inside the melt until it is frozen completely without affecting the remaining portion of the
melt. In order to accomplish this, a heat exchanger is required. Such a heat exchanger
should be able to absorb the heat from the sample inside the melt, and transfer it to the
outside environment. Because of the extremely harsh conditions prevailing within the
environment of the molten metal, no viable in-situ thermal analysis technique has yet
been made possible from an industrial point view.
One of the early trials carried out by the Mucciardi-Gruzleski group at McGill involved
utilizing classical heat pipe technology to extract the heat from a designated sample (1].
The device itself and the operating procedure, however, were too complicated to be used
as a substitute for the simple classical thermal analysis technique in general use at the
present time.
2
A heat pipe is an evaporator-condenser system in which the liquid phase of the working
substance is driven by gravity, capillary action or a pump. In its simplest form, it is a tube
with a number of layers of wire screening along the wall to act as a wick. The screen is
filled with a wetting liquid such as sodium or lithium for high temperature applications.
For moderate temperature applications, the screen is filled with water, ammonia, or
methanol. Expressed simply, the working principle oftraditional heat pipes involves one
end of the heat pipe being exposed to a heat source while the other end is subjected to
cooling; the liquid then evaporates at the heated end and condenses at the cooled end. As
the liquid diminishes in the evaporator segment, cavities form in the wick in that section
of the pipe, thereby generating a low pressure area causing the liquid to trickle along the
screen. Meanwhile, in the condenser segment, the screen becomes flooded. The surface
tension acting on the concave liquid-vapor interface causes the pressure to be higher in
the vapor than it is in the liquid. This pressure is transmitted by the vapor to the flooded
condenser section, where the vapor and the liquid pressures are approximately equal, so
that the liquid is driven from the condenser section to the evaporator section through the
wick. In a gravit y field, the evaporator may be placed below the condenser to assist the
liquid flow. It may even happen that occasionally, the gravitational force is capable of
causing the liquid to circulate, and consequently, there is no need to use a wick. Strictly
speaking, when there is no wick in the heat pipes it is called a thermosyphon.
Over the last few years, the Mucciardi-Gruzleski Group at McGill have developed a new
design for heat pipes which makes their use feasible in a molten metal environment. The
3
design makes use of the enhanced heat transfer characteristics which are associated with
vortex flow in order to improve heat pipe performance.
A special improved version of the McGill heat pipe was designed to make it suitable for
carrying out thermal analysis tests yielding greater detail. The recently modernÎzed
features of the modified McGill heat pipe make it possible for the new system:-
i) to carry out thermal analysis tests inside the crucible as required;
ii) to extract a sample from any location within the crucible;
iii) to freeze a specific sample inside the crucible without affecting the rest of the
melt, then to remelt the solidified sample after completing the thermal
analysis;
iv) to freeze the sample at an approximate1y constant and predetermined cooling
rate;
v) to automate the thermal analysis process so that no sample needs to be
extracted manually, and so that the cooling rate can be regulated by valves
which control the flow inside the system.
The new in-situ technology for conducting the thermal analysis of aluminum alloys was
thus developed. AIso, a comparison was instituted between the new upgraded technique
and classical thermal analysis by sampling cup. The new technique displays numerous
advantages over the standard procedures for thermal analysis. The most significant of
these advantages is the in-situ feature which makes it possible for the technique to
produce rapid chemical analyses.
4
1.2 Outline of the Present Study
The present study has two main components:-
The first is the design and the construction of the new heat-pipe probe. The second phase
is thermal analysis together with a study of the microstructure of samples obtained for
different aluminum alloys as used for this study.
Chapter One: Discusses the objectives and provides a general oudine of the topic under
investigation.
Chapter Two: Presents a short introduction to heat pipe technology, and briefly touches
on such topics as operating principles, types of heat pipes, and current applications. The
section includes a few observations on the theoretical background of the McGill heat-pipe
and the limitations of classical heat pipes.
Chapter Three: Several of the most widely applied treatments of liquid aluminum
silicon alloys are presented in this chapter. These include eutectic modification, grain
refinement, and the addition of elements. The chapter also provides definitions and
details of the operational setup and methods of procedure as weIl as a description of the
classical approach to carrying out thermal analysis tests.
5
Chapter Four: This chapter gives a detailed description of the experimental
methodology required for carrying out thermal analysis experiments which incorporate
the newly-designed heat-pipe probe.
Chapter Five: This chapter describes the stages in the evolution of the in-situ heat-pipe
probe for the thermal analysis of aluminum alloys; it also lists and discusses the
advantages and shortcomings of the two probe designs.
Chapter Six: The experimental results from both laboratory and industrial tests are
discussed in this chapter. It is divided into several subsections according to the alloy
under investigation and the nature of the material used for treatment. The discussion
covers grain refiners, melt modifiers, the level of impurities in the melt, and alloying
elements.
Chapter Seven: The main conclusions of the present work are presented in this chapter,
together with sorne carefully considered suggestions for possible future research projects.
6
References
1. Mahfoud, M. "Controlled Thermal Analysis Using Heat Pipe Technology"
Ph.D. Thesis, McGill University, Montreal, Canada,1997.
2 Elalem, K. Mucciardi, F. Gruzleski, J. Zhang, Z. Crescent, R. "Industrial
Applications of Heat Pipe Technology to the Permanent Mold Casting of
Magnesium Alloys", Proceedings of the 42nd Annual Conference of
Metallurgists ofCIM, Vancouver, B. C., Canada, pp 243-259, 2003.
7
Chapter Two
HeatPipe
2.1 Introduction
Heat pipes are commercially available in a wide range of sizes and for a number of
different applications. They provide satisfactory solutions to a variety of thennal
problems where the requirements include the need for high heat flux to be dissipated over
a small area.
Eastman (1968) identified several characteristics ofheat pipes which make them useful in
numerous specific applications [Il. These characteristics include, firstly, a closed-Ioop
cycle of operation in which the heat-transfer capacity is several times higher in order of
magnitude than the best-known solid conductors. This causes the thennal resistance along
the heat pipe to be kept to a minimum, and also makes it possible for both of the main
segments of the heat pipe (the evaporator and the condenser) to remain physically
separate. Secondly, increases in the heat flux in the evaporator lead to an increase in the
rate at which the working substance is vaporized, with a relatively small rise in the
operating temperature. Consequently, the heat pipe operates under quasi-isothennal
conditions, that is to say, the evaporation rate is self-adjusting and able to adapt to a
relatively wide range of power inputs, and is also able to maintain an approximately
8
constant source temperature. Thirdly, both segments of the heat pIpe operate
independently with a common two-phase working substance, thus the area from which
the heat is extracted may be of a different size and shape from those of the area from
which the heat is dissipated. Hence, high heat fluxes developed over a small area, such as
an evaporator surface, may be released over a larger area with a much smaller heat flux,
depending on the area ratio. Fourthly, the response time is independent of the distance
between the heat source and the heat sink, and it is also less than that required by solid
conductors [2, and 3].
Heat transfer characteristics, isothermal behavior, the ability to maintain an
approximately constant evaporator temperature over a range of heat flux levels, and the
variability of the evaporator and condenser sizes, are all factors which make heat pipes
and thermosyphons effective devices which may be used for many engineering
applications [3, 4, and 5].
2.2 Operating Principle of the Heat Pipe
Utilizing the latent heat associated with phase transformation (as for example from liquid
to vapor) is the key feature that gives both the heat pipe and the thermosyphon their
superiority over other types of heat exchangers in dissipating heat loads from specific
surfaces[2-5]. Although most of the components of these two devices are the same, their
manner of operation is substantially different.
9
Figure 2.1 shows a schematic diagram of both the heat pipe and the thermosyphon. The
heat applied at the evaporator surface forces the liquid phase of the working substance to
evaporate. During this phase transformation process, the liquid acquires the energy
associated with its heat of vaporization. Because the temperature of the condenser surface
is less than that of the saturated vapor in the evaporator, a pressure gradient in the
direction of the condenser is established, causing the vapor to flow from the evaporator to
the condenser. Inside the condenser, the hot vapor loses its heat of vaporization to the
colder condenser surface where it is converted to the liquid phase [4]. This condensed
liquid forms a fluid layer on the condenser surface, increasing in thickness with the
amount of condensed vapor. In thermosyphon configurations, the condenser section
should always be placed higher than the evaporator section in such a way as to cause the
liquid in the condenser to be driven by the force of gravity towards the bottom of the
evaporator. Although thermosyphons may contain grooves to promote a return of the
liquid to the evaporator, they depend upon gravitational acceleration to feed the
evaporator section with the liquid phase of the working substance which condenses there.
As regards heat pipes, they exploit the capillary forces associated with a capillary
wicking structure to guide the movement of the liquid from the condenser to the
evaporator. In cases where the evaporator is placed above the condenser, or in those
applications where the environment possesses micro-gravitational forces, a capillary
force or an extemal pumping force is essential to pumping the liquid from the condenser
to the evaporator either against gravity or in a microgravity environment. The only
10
difference between a thermosyphon and a heat pipe is the extent of their dependence on
the gravitational field to force the liquid to move from the condenser to the evaporator.
(')
1 ...
--+ Qout --+ --+
Figure 1 (a) Heat Pipe
m <
Liquid Film
1=t ...
Figure 1 (b) Thermosyphon
Fig.2.1. Schematic diagram (a) Heat pipe, (b) Thermosyphon
2.3 Types of Heat Pipe
There are several approaches to selecting and classifying heat pipes. Firstly, they may be
identified by their operating temperature range. Thus, there are low-temperature or
cryogenic heat pipes, and other types which function at moderate and high temperatures.
Cryogenie heat pipes were first introduced by Haskins (1966) [2,3], who used a nitrogen-
based heat pipe to control the temperature of an infrared detector. Since that time, low-
temperature heat pipes have been of interest to researchers for use in space technology to
cool optical surfaces and otherwise regulate them thermally. In moderate- temperature
heat pipes, either water or oil is customarily used as the working substance, while for
high-temperature heat pipes, liquid metals may be used [2, 3, and 6]. The second category
11
includes heat pipes which are characterized by their wicking structure and may use either
arterial or composite wicks. It should be remembered that pipes which have no wick are
called thermosyphons. The third way to group them is by function: variable conductance
or gas-Ioaded heat pipes; flexible heat pipes; rotating heat pipes; micro heat pipes; and
chemical heat pipes.
Choosing the most appropriate type to use in any given case will ultimately depend on
the specific application and the desired objective for using a heat pipe as a heat
exchanger.
2.4 Heat Pipe Applications
Heat pipes have been investigated and validated for a wide variety of applications. The
most important of these are, first, a separation of the heat source from the heat sink. The
high conduction performance of heat pipes makes it possible for heat to be transferred
over relatively long distances [2-4, and 6]. These pipes are, therefore, necessary in many
electrical applications where it is inconvenient to dissipate the heat by installing a
radiator in the close vicinity of the component. Radiators may not, however, be used in
the case where the neighboring parts are temperature sensitive.
A second application involves temperature flattening. Since a heat pipe works under
saturated temperature and pressure conditions, it tends to have a uniform temperature
profile. Because of this characteristic, heat pipes are used to reduce the temperature
12
gradients between unevenly heated areas of a surface. Heat pipes are, thus, also used in
the outer shell of satellites to cool the part exposed to sunlight and to dissipate the heat
towards the portion of the shell which lies in shadow [4] •
A third application relates to heat flux transformation. Heat pipes may also be used in
energy-saving and energy-recovery measures, as in the case where they have been used
successfully to redirect geothermal energy in heating the permafrost layer beneath the
Trans-Alaska Pipeline [2,4].
Lastly, heat pipes may be used for temperature control applications. Gas-Ioaded or
variable-conductance heat pipes may be used to regulate and modify the temperature of
devices mounted on heat pipe evaporator surfaces [2-6].
2.5 Limitations of Classical Heat Pipes
In addition to fundamental limitations on heat transport by a heat pipe such as capillary
wicking, there are sorne other factors which, under high heat flux conditions, limit heat
transport capacity. These factors include film boiling, entrainment, and sonic limitations.
Figure 2.2 shows how these limitations affect the performance of a typical heat pipe.
13
. : ..... . .. ...
.: .. ": ...• : .... : .. - .... . " .... 2 •• tI~·· ..•
.. s.:: .. : Ss . .
.: .. . . : ... . . " " . .. - " .
Heat flow limits (1-2) Sonic (2-3) Entrainment (3-4) Wicking (4-5) Boiling
Operating Temperature, T
Fig. 2.2 Schematic diagram of limits on heat pipe performance
2.5.1 Film Boiling
Film boiling is the main reason for not implementing water heat pipes in molten metal
environment. In order to appreciate fully the limitation imposed by film boiling, one
should be aware of the regimes that constitute the boiling curve. A typical water boiling
curve is shown in Figure 2.3. Below point A, which indicates the onset of nuc1eate
boiling, ONB, there is insufficient vapor to cause boiling at saturation temperatures, since
natural convection within the liquid is sufficient to transport the heat from the wall.
Beyond point A, vapor bubbles start to nuc1eate on the heated surface, and as the
difference between surface and saturated water temperatures increases, more nuc1eation
sites become active and increase bubble formation. Interference between the highly
populated bubbles then slows up the motion of liquid near the surface. In this region, the
values of convection heat transfer coefficients are considerably higher than those
associated with convection in the absence of phase change [5, 8].
14
The nucleate boiling heat flux at its maximum, or in other words the critical heat flux
(CHF), is reached at point C where a further increase in the surface temperature leads to
faster bubble formation. In this densely populated bubble region, an unstable vapor
blanket starts to form on the surface. The condition at any of these locations on the
surface may vary between nucleate and film boiling. The film boiling condition,
however, becomes dominant as the surface temperature increases (i.e. the fraction of the
total surface covered by the vapor film increases with increasing surface temperatures).
At point D the surface is fully covered by the vapor film and the heat flux is at its
minimum value. Conduction through the vapor layer is the orny effective heat transfer
mode between the surface and the liquid. Further increases in the surface temperature
activate the effect of the radiation mode of the heat transfer across the vapor film. Given
that radiative heat transfer is proportional to the fourth power of temperature, the heat
flux from the wall increases rapidly in the film boiling regime [1,4,5 and 6].
Chi[2,3] derived the following equation for the critical heat flux (CHF) by establishing a
pressure balance on any given bubble and using the Clausius-Clapeyron equation to relate
the temperature and the pressure.
2.1
where Le is the effective evaporator length;
Ke is the effective thermal conductivity of the liquid-wick combination;
15
ri is the inner radius of the heat pipe wall;
rv is the vapor core radius, and rn is the nuc1eation site radius which is about
2.54*10-7 for conventional heat pipes [2];
Mc,m is the maximum capillary pressure which is equal to 20", and rc is the rc
effective capillary radius. For the thermosyphonrc = r, while for the wire-screen
. k wire spacing + wire diameter WlC structure rc = ---=---...::....------
2
Another equation for the critical heat flux (CHF) of pool boiling for vertical
surfaces was derived by Chang [6] and is given as:
1 1
qc,ver =C{pJïhfg {og[PI-PvD4 2.2
where C is a constant ( 0.0012 ml/4/ SII2);
hfg is the latent heat of evaporation (J/kg);
0" is the surface tension (N/m);
g is the gravitational acceleration (m1s2); and
pvand PI are the densities ofvapor and liquid respectively (kg/m3).
16
1.E+07
1.E+06
-NE l 1.E+05 :-~
1.E+04
1.E+03
-
V V
~ 1
: Nucleate Boiling : Transition ! ;... ~... ~; ... Film Boiling : : : : :
{
li V '\
v/' \ / \ V
/ ~
~ ./
V / D
1 V A
10 100 1000
Fig. 2.3 Typical boiling curve for water at atmospheric pressure: surface heat flux as a
function of excess temperature Il T e [7].
2.5.2 Entrainment
In heat pipes, both liquid and vapor flow in opposite directions. The shear force occurring
at the liquid vapor interface may inhibit the return of the liquid to the evaporator. When
this occurs, a further increase in the heat input to the evaporator leads to liquid drop lets
being entrained in the vapor flow and carried to the condenser, eventually causing a
dryout of the evaporator. This shear force depends mainly on the inherent properties of
the vapor and its velocities [4]. The tendency of liquid droplets to be entrained is resisted
by the surface tension in the liquid phase. The ratio between the forces of inertia in the
vapor and the forces of liquid surface tension, expressed by the Weber number, may
provide a convenient method for predicting the entrainment limitation, as follows.
17
2.3
where P. is the vapor density, V is the vapor velocity, 0'/ is the liquid surface tension,
and Z is a dimension related to the wick spacing (in the case of the heat pipe this is the
hydraulic diameter of the wick structure). In order to avoid entrainment in heat pipes We
must be less than unity [2,3, and 41.
The axial heat flux is related to the vapor velocity by the following:
2.4
By substituting equation 2.2 in equation 2.3 and assuming that We is equal to one, then
the maximum axial heat flux which can be transported by a heat pipe as a result of the
entrainment limitation may be predicted by the following equation.
2.5
2.5.3 Sonic Limitation
The sonic limitation serves as an upper boundary for the axial heat transport capacity of
the heat pipe [21. The sonic limitation, which is known as the sonic choke, takes place
when the main flow velocity reaches the speed of sound. In a heat pipe, velocity
variations result from a variable mass flow through a constant area. Thus, the greater the
18
heat load on the evaporator or the cooling load on the condenser the greater is the mass
addition to the main stream from evaporation. The sonic choke occurs at the evaporator
exit where the main flow reaches a maximum speed [2, 4]. When the vapor at the
evaporator exit reaches sonic velocity, a further increase in the pressure difference
between the evaporator and the condenser segments has no effect on the velocity or the
flow rate of the main flow, and also the heat transfer rate across the pipe is not subject to
increase as a result of the existence of the choked flow [2, 3and 4]. However, increasing the
cooling load on the condenser beyond the sonic limit lowers the condenser temperature,
induces supersonic vapor flow, and creates a relatively sharp axial temperature gradient
across the pipe, although it does not affect the heat transfer rate across it [4]. As a general
rule, sonic conditions may be reached when the heat pipe operates at low vapor densities
and high vapor velocities [2].
The maximum axial heat flux due to the sonic limitation is as follows:
2.6
It should be noted that the formation of bubbles in the evaporator wick is to be avoided
because any hot spots formed will obstruct the liquid flow. In the evaporator segment,
sonic and entrainment limitations affect the axial heat flux, and the boiling limitation
affects the radial heat flux. The boiling limitation is not significant with regard to Iiquid-
metal heat pipes, but it can become a major problem for water-based heat pipes. One of
the aims of this research was to obtain a water-based heat pipe which would work
19
efficiently in the molten aluminum environment associated with high heat flux, and
which would operate in the nuc1eate boiling regime.
2.6 McGiII Heat Pipe
The new heat pipe technology which was developed at McGill University [9-11] (see figure
2.4) now makes it possible to use this technology with a high degree of efficiency. The
feature which makes this heat pipe unique is the design of the evaporator section, where
the problems associated with c1assical heat pipes have been resolved by adding a return
line and by modifying the heat transfer characteristics of the main flow. The evaporator is
supplied with a separate return line which overcomes the problem of the countercurrent
flow oftwo phases (i.e. vapor and liquid) [9-11], and thus the shear forces between the
reverse flow streams have been eliminated. A further function of the separate return Hne
is that it feeds the bottom part of the evaporator with a continuous flow of Hquid so that
the probe will never run dry. lntroducing a gate valve into the return Hne facilitates the
control of the flow which feeds the evaporator. The use of the gate valve enables the
probe to stop extracting heat from the sample when the valve is c1osed, whereas by
adjusting the valve position, different heat extraction rates may be attained.
AIso, this new technology is able to deal with relatively high heat fluxes, whereas
c1assical water-based heat pipes suffer from a film boiling limitation [2-7, and 9-11] and thus
are severely constrained as to heat extraction capacity. The evaporator was designed to
avoid the formation of a stable gas film on the heat transfer surface during operation,
which allows the new probe to operate efficiently in the molten metal environment in
20
contrast to classical heat pipes where high heat fluxes obstruct their operation. This new
technology enhances heat removal along the heat pipe. This improvement is attained by
initiating a vortex flow inside the evaporator section. The vortex flow in the heat pipe is
the mechanism which makes such a development a valuable contribution in the field of
heat pipe technology. A negative pressure gradient in the direction of the center of
rotation is induced to balance the centrifugai force which arises from the circular motion
of the two phase fluid. Since the density of the liquid is significantly higher than that of
gases, liquid droplets will be driven toward the pipe wall and will consequently form a
liquid film which slides up the wall under the influence of the axial pressure gradient in
the system. Therefore, a pressure field is established with its maximum pressure against
the pipe wall. This phenomenon breaks the gas film that might be generated at the heat
transfer surface. Consequently, this improvement in flow characteristics enhances the
capacity of the heat pipe to transfer energy across its wall and overcomes the boiling
limitation. The McGill heat pipe is thus enabled to operate efficiently over a wide range
of operating temperatures where classicaI water-based heat pipes are severely affected by
film boiling.
21
Condenser
Main Line
Evaporator Flow Modifier
Fig. 2.4. Sketch of McGill Heat Pipe [9]
22
Creating a vortex flow in heat exchangers in order to enhance heat transfer characteristics
is a topic which has been studied, reported upon, and implemented by a number of
researchers in the field [12-18]. A swirl flow may be created by inserting a twisted tape in
the flow regime [12-15,17, and 18]. Shell and tube heat exchangers were the main focus of
these studies, since introducing swirl tape to upgrade the thermal resistance of the tube
sides may be carried out without modifying the design [17]. In this case, smaller heat
exchangers may be designed for a given heat load.
Manglik and Bergles (1993) proposed an experimental correlation for the Nusselt number
associated with turbulent flow in isothermal tubes using a twisted tape insert [18].
Nu = [1 + 0.769] Nu oo y
2.7
where y is the twist ratio and equal to H , His the twist pitch (m), d is the tube inside d
diameter (m), and Nu oo is the Nusselt number for a straight tape insert (y = 00).
[ j
O.8 [ 8jO.2 7l'+2-2-Nu oo = 0.023 ReO.8 PrO.4 7l'48 8 d t/J
7l'-- 7l'-4-d d
2.8
where "is the thickness of the twisted tape and (J ~ (;: J ' n ~ 0,18 for liquid he.ting,
and n = 0.3 for liquid cooling. /-lb and /-lware the dynamic viscosities of the fluid at the
bulk temperature and at the wall temperature.
23
A different correlation for Nu was proposed by Fujita and Lopez, 1995 [11], as follows:
Nu = 0.023 Reo.8 PrO.4
H wherey=-.
d
( )
0.2
1+; 1-~
1-(~)
0.8
1
1-(~) 2.9
There is a tradeoff between using the improvement in thermal characteristics associated
with inserting twisted tapes and the increase in the pressure drop caused by them [17,18].
The influence of the twist ratio, y, on the friction factor of turbulent flow may be obtained
from the following experimental correlation [18].
[ ]
1.75 [ 28]1.25 f= 0.0791 7r 7r+2- d (1+ 2.752J
Re 0.25 48 t5 y 1.29 7r-- 7r-4-
d d
2.10
The key feature of the McGill heat pipe is the flow modifier implemented in the
evaporator segment. In its simplest form, this is a helical spring placed inside the
evaporator of the probe to form a helical path for the vaporized water to follow. The
existence of the flow modifier inside this section initiates a vortex flow which enhances
the heat transfer characteristics across the evaporator walls. It is important to note that the
flow modifier parameters depend strongly on the application environment involving the
amount ofheat flux associated with it.
24
In a molten aluminum environment where high heat flux of up to 2 MW/m2 [11] is
encountered, a helpful general rule for choosing flow modifier parameters based on
experiments with McGill heat pipe, may be: the pitch (H) and the diameter of the flow
modifier are equal to the inner diameter and to one tenth of the inner diameter of the
evaporator, respectively. A further increase in the flow modifier diameter will increase
the friction effect and obstruct the initiation of a vortex flow inside the evaporator. By
substituting the recommended parameters for the flow modifier in equations 2.9 and 2.10,
the flow inside the evaporator segment of Mc Gill heat pipe is characterized by the
following equations:
Nu = 0.0586 Reo.8 Pr°.4 2.11
f = 0.7866 Re -0.25 2.12
2.7 Advantages of the Mc Gill Heat Pipe over a Classical Heat Pipe
The enhancement in heat transfer associated with the new heat pipe technology may be
ascribed to a number of effects:
i) The vortex motion of the fluid, as generated by the helical flow modifier together with
the associated secondary motion of the fluid leads to better contact between the liquid
phase of the working substance and the pipe wall. This vortex motion expands the
25
nucleate boiling regime by preventing the vapor from accumulating on the pipe surface,
and at the same time causing a uniform distribution of the liquid over the hot surface.
ii) Because of the nature of the helical path, a longer longitudinal path is achieved.
iii) The swirl and the mixing associated with the vortex flow enhance the convective heat
transfer coefficient.
iv) Using a separate return Hne to feed the evaporator with the liquid phase of the
working substance eliminates the entrainment forces between the main flow and the
return flow.
v) Introducing a valve in the return line provides precise control over the rate at which
heat is extracted from the evaporator [19].
26
References
1. Yale Eastman, G. "The Heat Pipe", Scientific American 218(5), 1968, pp. 38-46
2. Peterson, G. "An Introduction to Heat Pipes Modeling, Testing, and Applications" ,
John Wiley & sons, Inc., New York, NY, USA,1994.
3. Chi, S. "Heat Pipe Theory and Practice", Hemisphere Publishing Corporation,
Washington, USA, 1976
4. Dunn, P., Reay, D. "Heat Pipes", Third Edition, Pergamon Press, UK, 1982.
5. Faghri,A., "Heat Pipe Science and Technology", Taylor & Francis, Washington D.C.,
1995.
6. Marten, T., Johan, G. "Heat Pipes: Construction and Application; A Study of Patents
and Patent Applications", Elsevier Applied science Publishers Ltd., Essex,
England, 1987.
7. Incropera, D., Witt, D. "Fundamentals ofHeat and Mass Transfer", Third Editition,
John Wiley & Sons Inc., Singapore, 1990.
8. Tong, L., Tang,Y. "Boiling Heat Transfer and Two Phase Flow", 20d Edition,
Taylor&Francies, 1997
9. Elalem, K. Mucciardi, F. Gruzleski, J. Zhang, Z. Crescent, R. "Industrial Applications
of Heat Pipe Technology to the Permanent Mold Casting of Magnesium Alloys".
Proceeding of the 420d Annual Conference of Metallurgists ofCIM, Vancouver, B. C.,
Canada, 2003, pp 243-259.
27
10. Zhang, C. Mucciardi, F. Gruzleski, J. "An Overview of the Controlled Cooling of
Pennanent Mold Castings of Aluminum Alloys", Proceedings of the 42nd Annual
Conference of Metallurgists of CIM, Vancouver, British Columbia, Canada, pp 307-
321,2003.
Il. Zheng, G. "A Novel Flow-Modified Heat Pipe Development and Experimental
Investigation", Ph. D. Thesis, McGill University, Montreal, P.Q.,Canada, 2003
12. France, D., Minkowycz, W., Chang, C. "Analysis of Post-CHF Swirl Flow Heat
Transfer", International Journal of Heat and Mass Transfer, Vol.37, Suppl.1,pp. 31-
40, 1994.
13. Inasaka, F. Nariai, H. "Critical Heat Flux in Subcooled Flow Boiling in Swirl Tubes
Relevant to High Heat Flux Components", Fusion Technology, Vol. 29, pp. 487-498,
1996.
14. Agarwal, S., Rao, M. "Heat Transfer Augmentation for the Flow ofViscous Liquid in
Circular Tubes Using Twisted Tape Inserts", International Journal of Heat and Mass
Transfer, Vol. 39, No. 17, pp 3547-3557,1996.
15. Weisman, J., Yang, J., Usman, S. "A Phenomenological Model For Boiling Heat
Transfer and the Critical Heat Flux in Tubes Containing Twisted Tapes",
International Journal of Heat and Mass Transfer, Vol. 37, No. 1, pp. 69-80,1994
16. Solnordal, C., Gray, N. "An Experimental Study of Fluid Flow and Heat Transfer in
Decaying Swirl through a Heated Annulus", Experiments in Fluid, Vol. 18, pp. 17-
25, 1994.
28
17. Manglik, R., Bergles, A. "Heat Transfer and Pressure Drop Correlations for Twisted
Tape Inserts in Isothennal Tubes: Part 1- Laminar Flows", Journal of Heat Transfer,
Vol. 115, pp. 881-889, Nov. 1993.
18. Manglik, R., Bergles, A. "Heat Transfer and Pressure Drop Correlations for Twisted
Tape Inserts in Isothennal Tubes: Part 11- Transition and Turbulent Flows", Journal of
Heat Transfer, Vol. 115, pp. 890-896, Nov. 1993.
19. Musmar, S., Mucciardi, F., Gruzleski, J., Samuel, F. "Investigation of Iron and
Copper Intennetallics in 356 Aluminum Alloy and in AI- 7% Si Binary alloy by an
In-Situ Thennal Analysis Probe", proceedings of 1l0th Metal Casting Congress
(AFS), April. 18-21, Ohio,USA, 2006.
29
Chapter Three
Melt Treatment and Thermal Analysis
3.1 Introduction
The widespread use of aluminum alloys, especially those containing silicon as the major
alloying element, may be attributed to their high strength-to-weight ratio, high fluidity,
low shrinkage in casting, good corrosion resistance, machinability, and weldability as
weIl as their electrical and thermal conductivity (1].
In view of the fact that aluminum-silicon alloys are widely used in the automotive
industry, improving melt quality has constantly been of major concem to researchers.
Grain refinement, eutectic silicon modification, and degassing techniques have been
extensively investigated and reported for the purpose of enhancing the cast quality [1-20]. It
is a well-established fact in the domain of aluminum technology that the thermal analysis
of aluminum alloys provides an assessment of the level of grain refiners in the melt and
the extent of eutectic modification.
30
3.2 Grain Refining
Small grain size is a desirable feature in the cast. It ensures uniform mechanical
properties, reduces hot tearing, enhances machinability, and improves the distribution of
second phases as weIl as micro-porosity on a fine scale [2].
In order to reduce the grain size, a common practice in foundries is to add inoculants
deliberately to the melt before casting [2-10]. The master alloys Al-Ti and Al-Ti-B are
usually used as the grain refining inoculants for aluminum alloys, because they promote
heterogeneous nucleation. Since each grain is nucleated by a single foreign particle, a
greater number of particles or nuclei will yield a greater number of grains and thus a
smaller grain size will be obtained. It should be mentioned that not all foreign solid
particles in the melt are capable of promoting heterogeneous nucleation. It is believed
that the interfacial energy between the nucleant and the liquid metal has a key role in
successful grain refinement. In the absence of sufficient heterogeneous inoculants in the
melt, a driving force for initiating solidification is required. Thus, a drop in temperature is
the thermal force which drives the nucleation and growth of the grain, a phenomenon
known as homogeneous nucleation. The higher the concentration of heterogeneous
inoculants in the melt, the less the thermal force is needed. Thermal analysis which is
based on measuring this force pro vides an assessment of the level of grain refiners, or
heterogeneous inoculants, in the melt. Extensive and detailed discussion on the theory
and mechanism of grain refiners may be found in the literature [1-20].
31
3.2 Eutectic Modification
Eutectic modification refers to the transformation process occurring in the morphology of
eutectic silicon from acicular to fibrous. The unmodified alloy contains the silicon phase
in the form of large plates with sharp sides and ends, known as acicular silicon.
Modification brings about a significant improvement in the mechanical properties of the
cast. It improves the impact strength, tensile strength, and ductility. The ductile aluminum
matrix which separates the brittle silicon phase is responsible for the impact strength of
the material. Any process which reduces the size of the brittle phase particles or increases
their separation will improve impact properties [l, Il].
Fine eutectic structure may be produced either by rapid solidification or by adding
chemical elements of sorne groups lA, lIA, and rare earth elements known to cause
modification [11-18]. This addition of small amounts of the modifying agent causes the
eutectic silicon to solidify into a fine morphological structure and to form an
interconnected network.
ln practice, sodium and strontium are the most common elements industrially used for
modification purposes. These are effective at low concentration levels, typically in the
order of 0.007 to 0.02 wt% [Il].
Although the use of sodium as a modifying agent produces the fmest modified structures
at the lowest concentrations, it has several drawbacks. Sodium has low solubility in
aluminum, also its high reactivity requires a special packaging technique. Pure strontium
32
is reactive with air and water vapor, and the oxide formed prevents the dissolution of
strontium in the melt which, in turn, prevents the modification of the eutectic structure.
However, master alloys which contain less than about 45% strontium are not reactive in
the air. Alloys with 20% to 60% strontium have a melting point in excess of 900°C
making them of no practical use in aluminum foundries where the temperatures approach
only 750°C, which is one of the reasons why a 10%Sr-90%AI master alloy is commonly
used for modification purposes in aluminum silicon alloys.
3.3 Effects of Adding Elements to the Melt.
Strengthening of aluminum silicon alloys is achieved by adding small amounts of Cu, Mg
or Ni to the composition of hypoeutectic alloys; and silicon provides excellent casting
properties while copper improves tensile strength, machineability, and thermal
conductivity at the expense of a reduction in ductility and corrosion resistance. Moreover,
the strength and machinability may be improved by heat treatment.
It is recommended that the magnesium level in the cast be kept below 0.3% to avoid the
formation of the Mg2Si phase which leads to a decrease in the tensile strength, although
this effect may be minimized by heat treating the cast using the T 4 solution treatment and
quenching. Heat treating results in a uniform distribution of Mg2Si precipitates through
the aluminum dendrites. Such an alloy is endowed with excellent castability, pressure
tightness, and corrosion resistance.
33
Since iron is known to have a detrimental effect on the properties of the cast, lower levels
of iron in the cast is one of the aims of the alwninwn industry. The formation of brittle
and hard plates of the a -iron phase, AlsFeSi, leads to machining difficulties, and
potential sites for machine tool failure. In practice, it is difficult to get rid of iron
contamination prior to casting, although the undesirable effects of iron can be minimized
by adding such chemical agents as Mn, Cr, Co, Be, and Mo to react with Fe, Si, and Al to
form a Chinese script intermetallic AI1s(Mn,FehSh. Manganese is the element most
frequently used to reduce the effect of iron contamination. The existence of manganese in
the melt in sufficient quantity expands the a-phase region, and increases the possibility of
a-phase crystallization even at high levels of iron in the melt. The reported ratio of Mn to
Fe which is sufficient to ensure the formation of the a-phase rather than the p-phase is
1 :2. [28,29]
3.4 Thermal Analysis
Thermal analysis of metallic alloys involves the acquisition and then the analysis of the
temperature-time trace of a control volwne of molten metal as it cools down and passes
through the mushy zone until it becomes completely solid. The shape of the temperature
time curve depends strongly on the solid metallic and intermetallic phases that are formed
during solidification. The rate of change in temperature, in other words the cooling rate,
is affected by the latent heat evolved during the formation of the solid metallic phase. The
quantity of heat released is not only dependent on the metallic phase which forms but
also on the quantity of that phase. The quantity of energy given off affects the rate of
34
decrease in the temperature of the sample either by slowing down the cooling rate, or by
totally arresting the cooling of the metal, or by heating the sample. Each of these features
will be observed in the temperature curve [2, 15, and 16]. The features which appear in the
solidification reflect not only the phase formed, but also the quantity of this phase.
In solid solution alloys the solidification occurs over a range of temperatures, while in
pure metals and eutectic compositions the rate of decrease drops to zero throughout the
freezing process, during which a corresponding plateau will be evident on the cooling
curve. Although eutectic alloys contain more than one solid phase, they behave as pure
metals.
When considering commercial alloys, the chemical complexity of a commercial alloy
plays a considerable role in shaping the solidification curve [6, 7, and 16]. Such alloys contain
multiple components and several solid phases, all of which have an effect on the shape of
the cooling curve [16]. A simple example is the cooling curve for a 356 aluminum alloy
consisting of 7% Si, 0.35% Mg, 0.2% Cu, 0.2% Fe, and 0.1 % Zn, as presented in Figure
3.1. The presence of minor elements such as Mg may lead to a noticeable change in the
cooling curve, as is indicated by the number 3 in Figure 3.1.
35
T (OC)
620 ~
600
seo
560
540
520
500
480
250
solldlf<catlon time {S)
300 350
1: primary phase nucleation; 2: binary Al-Si eutectic plateau; 3: eutectic AI-Mg2Si-Si.
Fig. 3.1. Actual cooling curve for 356 aluminum alloy, (Gruzleski, 1990).
A close inspection of the cooling curve may be carrled out using the first derivative
curve. The inflection points present on this curve may be much more clearly identified by
examining the first derivative, which is the slope of the cooling curve. Every peak which
appears on the first derivative curve corresponds to an inflection point on the cooling
curve which reflects the formation of a new phase (see Fig. 3.2). Peak 1 refers to primary
aluminum nucleation, while the eutectic Al-Si phase is the plateau denoted by the number
2, and the AI-M~Si-Si temary eutectic is indicated by the inflection point 3.
The first derivative curve magnifies the significant slope changes on the cooling curve
making them more identifiable [29,42, and 46]. For instance, the formation of the AI-Mg2Si-
Si temary phase is more evident on the first derivative curve than on the cooling curve.
The second derivative curve, which is the slope of the first derivative curve, is also used
to determine the start and the end of the reactions [27].
36
700 0.2
680 0.15 660
~ 0.1 u
640 QI ..!!!
620 0.05 ~ e
-> .. 600 0 = ·C .. I! QI
"t:I 8. 580 -0.05 il E
560 iL: ~ -0.1
540
520 -0.15
500 -0.2 0 100 200 300 400 500 600 700 800 900
time (sec)
Fig. 3.2. First derivative curve associated with actual cooling curve for 356 aluminum
alloy
Most of the literature on the thennal analysis of aluminum alloys focuses on how to make
use of the thennal analysis technique to control the quality of the cast [1-4, 6-7, 9-10, and 14-16].
Many of the studies have focused on matching the shape and features of the curve with
several parameters which affect the quality of the cast, such as the extent of grain
refining, corresponding to the amount of grain refiners in the melt, and the level of
modification which refers to the morphology of the eutectic phase.
3.4.1 Salient Features of the Cooling Curve
There are several parameters which are used in analyzing and studying the cooling curve.
These are recalescence temperature, minimum temperature, undercooling, apparent time,
and the eutectic temperature. Figure 3.3 shows these parameters.
37
-Recalescence temperature is denoted by T2 in Figure 3.3. This temperature is the
maximum reached as a result of the evolution of latent heat following nucleation.
-The minimum temperature is denoted by Tl in Figure 3.3. The minimum temperature is
simply the point at which heat evolution exceeds heat loss to the surrounding
environment and the system starts to heat up to recalescence temperature. In sorne of the
literature in the field it is defined as the nucleation temperature (T n) at which the
temperature of liquid drops below the freezing temperature to derive nucleation thermally
[1]
-Undercooling is a departure from equilibrium necessary to cause the first solid to form.
In the present study undercooling refers to the difference between recalescence and
minimum temperatures (undercooling = T2-Tt).
-The apparent time is the time consumed by the sample to heat up from the minimum
temperature to recalescence temperature, and is equal to (trtl).
38
CI)
~ T2 ~ t e ~ .......................... .
Tl ~ ............................ -:-....... .
.
.
Teutectic
time
Fig. 3.3. Schematic diagram of a cooling curve.
In aluminum silicon alloys, thermal analysis is used mainly to control the grain size and
the level of eutectic modification [l, 3, 4, and 7-9]. Grain refinement affects the cooling curve
in the early stages near the liquidus temperature. Both the undercooling and the apparent
time decrease with finer grain sizes while the recalescence temperature increases. In a
well-refined alloy, with a sufficient number of nuclei, nucleation will occur in a short
time while almost no undercooling is required to start primary solidification
(approximately O.2°C).
Eutectic modifiers affect the eutectic portion of the cooling curve. Drops in both the
nucleation temperature and the eutectic temperature are observed when there is an
increase in the level of modification. This factor is the one most frequently reported in the
literature when assessing levels of modification. Tenekedjiev and Gruzleski [12] use
39
strontium as a modifying agent for severa! aluminum-silicon alloys at different cooling
rates. They observed an increase in the eutectic undercooling and a considerable
enlargement of the eutectic reaction in the presence of modifiers. They also found that the
primary arrest was not affected by the strontium.
3.4.2 Classical Thermal Analysis Equipment
The simplicity and the relatively low cost of the setup are what make thermal analysis
applicable in most foundries. A sampling cup, thermocouples, a data acquisition system,
and standard computer are the basic components for the thermal analysis setup.
The sampling cup is of a simple design as shown in Figure 3.4. Thermal analysis is
carried out by pouring a sample of molten metal into the sampling cup then the
temperature of the sample is recorded as it cools down and solidifies. Control of the
cooling rate is important in thermal analysis so as to reveal further information on the
cooling curve. The cooling rate at which the sample solidifies is related to the physical
and thermal characteristics of the sampling cup. High cooling rates may be achieved by
using metallic sampling cups which are normally made of thin-walled steel, while slow
cooling rates may be achieved by using cups of shell-molded sand. Preheating the sample
cup enhances the controllability over the cooling rate; however, in terms of
reproducibility, this aspect of controlling the cooling rate is still poor. Cooling rates near
equilibrium condition may be achieved by insulating the sampling cup.
40
A thennocouple placed at the center of the sampling cup is used to detect the temperature
at the center where a minimum temperature gradient exists. Sorne researchers used two
thennocouples, one of which was placed at the center and the other close to the cup wall
to detect any changes in the cooling rates. The thennocouples are connected to a data
acquisition and computer system so as to store and analyze the data, and then to display
the results.
Sampling cup
Thennocouple
Fig 3.4 A typical sampling cup used for thennal analysis
Even though classical thennal analysis involving the pouring of a sample of the melt into
a cup has been, to a certain extent, successfully implemented in industry, it exhibits
several negative aspects. These include poor control of the rate of heat extraction, and
consequently of the cooling rate during solidification, as weIl as the unfortunate tendency
to poor reproducibility [10, Ill. No viable in-situ systems have yet to be developed, due to
the extreme environment involved in using liquid metals.
41
Two new techniques were introduced during the last decade in an attempt to overcome
shortcomings associated with traditional thermal analysis methods. These were based on
making use of traditional heat pipe technology for use as a heat exchanger that is able to
handle the high heat extracted during liquid metal solidification. These probes provided a
number of advantages such as converting the process into a semi-continuous test,
providing a heating curve in addition to a cooling curve, also the same sample may be
used many times for the test, and the techniques offer the possibility of controlling the
cooling rate during solidification [11]. Using traditional heat pipes in thermal analysis,
however, has major drawbacks, the most important ofwhich, is that these techniques are
able to handle heat removal only from relatively small sampling volumes due to the
boiling limitation restricting this type of heat pipe. Further major drawbacks inc1ude the
fact that controlling the cooling rate has proved to be an over-complicated process so far,
and that the traditional heat pipe is obliged to operate non-stop, whether the desired
function is freezing the sample or melting it. Other lesser drawbacks inc1ude the fact that
many variables need to be controlled during the freezing process, which adds greater
complexity to the entire operation.
42
References
1. Gruzleski, J., Closset, B. "The TreatInent of Liquid Aluminum-Silicon Alloys",
American Foundrymen's Society, Inc., Illinois, USA, 1990.
2. Djurdjevic, M., Hasenbusch, R and Sokolowski, J. "Assessment of the
Hydrogen level in 319 Aluminum Alloy Melts Using the Thermal Analysis
Technique", Light Metals, pp. 889-896,2002.
3. Mohanty, P., Gruzleski, J. "Grain Refinement Mechanisms of Hypoeutectic AI
Si Alloys". Acta Mater. Vol. 44, No. 9, pp. 3749-3760, 1996.
4. Johnsson, M. " Influence of Si and Fe on the Grain Refinement of Aluminum".
Zeitschrift fuer Metallkunde, 85(11), pp. 781-785, 1994.
5. Vali, M. Abdel-Azim, A. and Rejf, V. "Effect of Ultrasonic Processing on the
Structure ofSome Al Alloys", Protsessy Lit'ya, 1, pp. 52-58,2001.
6. Pasciak, K., Sigowrth, G., "Role of Alloy Composition in Grain Refining
Aluminum 319 Alloy", AFS Transactions, pp.567-576, 2001.
7. Simensen, C., "Grain Refining of AI-7wt%Si Alloys", Light Metals, pp. 679-
684,1999.
8. Lee, Y., Dahle, K., St John, D., Hutt, J., "The Effect of Grain Refinment and
Silicon Content on Grain Formation in Hypoeutectic Al-Si Alloys", Material
Science and Engineering A259, pp. 43-52, 1999.
43
9. Apelian, D., Sigowrth, G., Whaler, K., "Assessment of Grain Refinement and
Modification of Al-Si Foundry Alloys by Thermal Analysis", AFS
Transactions, pp. 297-307,1984.
10. Gloria, D. and Gruzleski, J. E. "A Study of the Thermal Analysis Parameters
Applied to the Grain Refmement of Al-Si Casting Alloys", Proceeding of the
International Symposium on Light Metals, Quebec City, QC, pp. 315-329,
1999.
Il. Gruzleski, J. and Closset. B. "The Treatment of Liquid Aluminum-Silicon
Alloys", American Foundrymen's Society, Inc, 1990, Des Plaines, Illinois,
U.S.A
12. Tenekedjiev, N., Gruzleski, J. "Thermal Analysis of Strontium-Treated
Hypereutectic and Eutectic Aluminum-Silicon Alloys", AFS Transactions,
Vol. 99, USA, pp. 1-6, 1991.
13. Tenekedjiev, N. Mulazimoglu, H. Closset, B. and Gruzleski, J.,
"Microstructures and Thermal Analysis of Strontium-Treated Aluminum
Silicon Alloys", 1995, American Foundrymen's Society, Inc. U.S.A.
14. Alexopoulos, N., Pantelakis, "Evaluation of Effects of Variations in Chemical
Composition on the Quality of AI-SiMg, Al-Cu, and AI-Zn-Mg Cast
Aluminum Alloys", Joumals of Materials Engineering and Performance,
Vo1.12, No.2, pp. 1996-205,2003
15. Das-Gupta, R., Brown, C., Marek, S., "Analysis of Overmodified 356
Aluminum Alloy", AFS Transactions, pp. 297-296,
44
16. Argyropoulos, S., Gruzleski, J., Oger, H., "The Quantification Control of
Modification in Al-Si Foundry Alloys Using Thermal Analysis Technique",
AFS Transactions, pp. 351-357, 1983.
17. Mondolfo, L., "Aluminum Alloys, Structure and Properties", Butterworths,
London, pp 213-624, 1979
18. Jeng, S. C. and Chen, S. "The Solidification Characteristics of 6061 and A356
Aluminum Alloys and their Ceramic Particle-Reinforced Composites", Acta
Materialia, Dec., 45, pp. 4887-4899. 1997.
19. Mackay, R. 1. and Sokolowski, J. H. "The Development of Thermal Analysis
Partitioned Parameters for the Determination of Cast Aluminum Structures",
Proceedings of the Advances in Aluminum Casting Technology II, Materials
Solutions Conference, Columbus. U.S.A, 2002.
20. Joenoes and Gruzleski, "Magnesium Effects on the Microstructure of
Unmodified and Modified Al-Si Alloys", Cast Metals, 4, pp. 62-71, 1991.
21. Wang, L. Apelian, D. and Makhlouf, M., "Effect of Alloy Chemistry and
Cooling Rate on the Solidification Characteristics of Al-Si Cu Die Casting
Alloys", Proceedings of the 5th International AFS Conference on Molten
Aluminum Processing, Orlando, U.S.A, Nov., 1998.
22. Gowri, S. and Samuel, F. H. "Effect of Magnesium on the Solidification
Behavior of Two AI-Si-Cu-Fe-Mg (380) Die Casting Alloys", Transactions
of the American Foundrymen's Society, Chicago, Illinois, U.S.A. Apr.,
2003.
45
23. Mackay, R. Sokolowski, J. and Evans, W. "Aluminum-Silicon-Zinc
Magnesium Casting Alloys: A Preliminary Investigation", Proceeding of the
40th Annual Conference ofMetallurgists ofCIM, pp. 467-478, 2001.
24. Tan D. Q., "Precipitated Phases and Thermodynamic Analysis during
Solidification of AI-Fe-X System at Low Cooling Rates". Transactions of
the Nonferrous Metals Society of China, Vol. 13, no. 5, pp. 1133-1139, Oct.
2003.
25. Gowri, S. and Samuel, F. H. "Effect of Alloying Elements on the
Solidification Characteristics and Microstructure of AI-Si-Cu-Mg-Fe 380
Alloy", Metallurigical and Materials Transactions, Vol. 25A, Feb., 1994.
26. Han, Y. S. "Studies on Grain Refmement and Intermetallic Phase Formation
in Al-Si-Fe Based Alloys". Ph.D Dissertation, Univ. Nottingham,
Nottingham, UK, 2002.
27. Musmar, A., Mucciardi, F., Samuel, F. and Gruzleski, J. "Investigation of
Iron and Copper Intermetallics in 356 Aluminum Alloy and in AI-7% Si
Binary Alloy by an In-Situ Thermal Analysis Probe", 110 Metal Casting
Congress Proceedings, Columbus, OH., 2006.
28. Shabetari, S., Gruzleski, J., "Gravity Segregation of Complex Intermetallic
Compounds in Liquid Aluminum-Silicon Alloys", Metallurgical and
Materials Transactions A. Vol. 26A, pp 999-2006. 1995.
29. Mackay, R, Gruzleski, J., "Quantification of Iron in Aluminum-Silicon
Foundry Alloys via Thermal Analysis", International Journal of Cast Metals,
10, pp. 131-145, 1997.
46
30. Das-Gupta, R., "Influence of Iron on Microstructures and Mechanical
Properties of Strontium-Modified 356 Aluminum Alloy", Die Casting
Engineering, Vol. 40, N. 3, pp. 65-67, 1996.
31. Mualzimoglu, M., Tenekedjiev, N., Closset, B., Gruzleski, 1., "Studies of the
Minor Reactions and Phases in Strontium-Treated Aluminum-Silicon
Casting Alloys", Cast Metals, Vol. 6, No.ll, 1993.
32. Shabestari, S., Ghodrat, S., "Thermal Analysis and Microstructural
Evaluation of Intermetallic Compounds Formed During Pre- and Post
Eutectic Reactions in 319 Aluminum Alloy", 43rd Annual Conference of
Metallurgists of CIM, Light Metals, Hamilton, Ontario, pp.299-313, 2004.
33. Narayanan, A., Samuel, F., Gruzleski, J., "Crystallization Behavior of Iron
Containing Intermetallic Compounds in 319 Aluminum Alloy",
Metallurgical and Materials Transactions A, Vol. 25A, pp. 1761-1773, 1994.
34. Gonzalez, C. Baez, J. Chavez, R. and Jurez, J."Quantification of the SiCp
Content in Molten AI-SilSiCp Composites by Computer Aided Thermal
Analysis", Journal of Materials Processing Technology, Vol 143, n 1, Dec.
20, pp 860-865, 2003.
35. Li, Z. Samuel, A. Samuel, F. and Valtierra, S., "Effect of Alloying Elements
on the Segregation and Dissolution of CuAh Phase in Al-Si-Cu 319 Alloys",
Journal ofMaterials Science, 38 (6), pp. 1203-1218,2003.
36. Charbonnier, J. "Microprocessor Assisted Thermal Analysis Testing of
Aluminum Alloy Structures", AFS Transactions, 92, pp. 907-922, 1984.
47
37. Gowri, S. "Comparison of Thennal Analysis Parameters of 356 and 359
Alloys". Transactions of the American Foundrymen's Society, Hamilton,
Canada, Vol. 102, pp. 503-508, 1994.
38. Sparkman, and D. Keamey, A. "Breakthrough in Aluminum Alloy Thennal
Analysis Technology for Process Control", Transactions of the American
Foundrymen's Society, Vol. 102, pp. 455-460, 1994
39. Dedavid, B. Costa, E. and Ferreira. C. "A Study of Precipitates Fonnation in
AA 380 Aluminum Alloys Modified by the Addition of Magnesium",
Proceeding of the 20d Brazilian Congress on Thennal Analysis and
Calorimetry, Pocos de Cladas, Brazil, April, 2000.
40. Ocansey, P. Bamberger, M. and Minkoff, 1. "Solidification, Thennal Analysis
and Properties of a-SiC Partic1e Reinforced AISil1.7 Alloy Composites",
Giessereiforschung, 48 (3), 76, pp. 82-88, 1996.
41. Emadi, D., Whiting, L., "Detennination of Solidification Characteristics of
Al-Si Alloys by Thennal Analysis", Transactions of American Foundry
Society, Vol. 110,2002.
42. Barlow, J., Stefanescu, D., "Computer-Aided Cooling Curve Analysis
Revisited", Transactions of the American Foundrymen's Society, Vol. 105,
1997.
43. Backerud, L., Sigowrth, G., "Recent Developments in Thennal Analysis of
Aluminum Casting Alloys", AFS Transactions, Vol. 97, pp. 459-464, 1987.
48
44. Zhang, C., Musmar, S., Mucciardi, F., Gruzleski, J., Samuel, F., "In-Situ
Thermal Analysis Technology for Aluminum Foundry Alloys", 43rd Annual
Conference of Metallurgist ofCIM, Hamilton, ON, 2004.
45. Mahfoud, M. "Controlled Thermal Analysis Using Heat Pipe Technology",
Ph.D Thesis, McGill University, Montreal, Canada,1997.
46. Backerud, L. KroI, E. and Tamminen, J. "Solidification Characteristics of
Aiuminum Alloys", Tangen Trykk, Norway, 1986.
47. Hatch, J.E., "AIuminum Properties and Physicai Metallurgy", American
Society for Metals (ASM), p. 424, 1984.
49
Chapter Four
The New Probe Design
4.1 Introduction
In order to overcome the disadvantages of traditional thermal analysis, major changes
need to be made in the approach of handling a sample. Consequently, a new design is
strongly to be recommended. This design should incorporate such features as reducing
labor, making thermal analysis a controllable process, and increasing the overall
effectiveness ofthe analysis process itse1f.
Reducing labor may be accompli shed by carrying out the thermal analysis inside a
crucible by means of an in-situ probe. Such a procedure would obviate the need for
human intervention when extracting a sample from the crucible located inside the fumace
and when pouring it into a sampling cup outside the fumace.
Thermal analysis is a quality control technique used to verify the quality of the cast prior
to the casting process. It is based on certain features which occur on the solidification
curve of a small sample of the me1t. Thus, regulating the solidification rate makes it
possible to focus on the lesser details of the curve. Regulating also reduces the time
needed for carrying out the test, which should start with an elevated cooling rate until the
required feature is reached, after which the cooling rate is reduced to a specified limit
where this feature emerges.
50
The thermal analysis process may be rendered more efficient by creating a probe which
(a) is capable ofreducing human exposure to molten metal, (h) produces a melting curve
in addition to the cooling curve, (c) controls the amount of heat extracted from the
sample, (d) makes it possible to extract a sample from any location in the crucible,
(e) carries out the tests inside the fumace, (t) reduces the time required for carrying out
the test, and, (g) incorporates a simple operating procedure with the improved design.
These are all criteria which have been taken into consideration seriously in designing and
constructing the new thermal analysis probe.
4.2 Characteristics of the Heat Pipe
The main heat pipe characteristics which relate to the development of the new thermal
analysis technique include the following:
(1) Heat Transfer Capacity
The new probe was designed and built based on innovative heat pipe technology
developed by the Mucciardi and Gruzleski work group at McGill University [1, 2]. This
new technology enhances the capability of heat removal along the heat pipe which may
be obtained by initiating a vortex flow inside the evaporator section. This type of flow, as
created in the heat pipe, is the very mechanism which makes the new probe so distinctive.
The flow increases the operating temperature range of the heat pipe and reduces the
importance of most of the limitations in selecting a working substance. For instance,
water-based heat pipes suffer from film boiling if they operate in a molten aluminum
environment, while the enhanced heat pipe operates efficiently under these same
conditions.
51
(2) Safety Considerations
One of the reasons for discarding the earlier designs developed at McGill is the fact that
dealing with such working substances as cesium, potassium or sodium, involves safety
considerations which become of major concem when using such a device either in the
laboratory or in the foundry. Cesium is not a realistic choice because of its potential for
environmental hazards, while sodium vapor is flammable. Thus, one of the objectives of
the present work is to build a heat pipe probe utilizing an environmentally friendly
working substance (water).
(3) Simplicity of Operation
One of the major shortcomings of the former designs previously developed at McGill was
the complexity of the setup and the operating procedures. The first probe built by
Meritian, M. [3] (see Fig. 4.1), used a thermosyphon for analyzing the solidification of
aluminum alloys. Although the probe was innovative at that time, it had numerous
limitations. One of these was the need to control the operative condenser area by
introducing a pressurized inert gas which tended to act as a plug and effectively shut off
the portion of the condenser which it filled. During operation, the inert gas was released
to the environment and sorne of the working substance seeped out with the inert gas. This
eventuality made it a distinct possibility that the work place would become contaminated
with the working substance. Moreover, its operation tended to require excessively
painstaking work due to the many variables which needed to be modified in the course of
a single test. Such serious considerations necessitated developing a new design which
52
would betler correspond to contemporaneous needs. Later on in time, another probe was
built by Mahfoud [4], (see FigA.2), using a sodium-based gravity-assisted heat pipe with a
wick incorporated into the design of the updated probe. This probe, however, also
suffered from serious drawbacks, one of which was the complexity of the operating
procedures. The condenser required heating up to melt temperature in order to stop the
probe from operating, while the cooling rate was regulated by controlling the heating
elements on the condenser segment. Furthermore, it was not possible to apply cooling
measures during the entire solidification process for certain aluminum alloys. This may
be ascribed to limitations in the operating temperature range. Thus, it was difficult to
detect intermetallic compounds forming at later stages of solidification (500°C - 520°C)
without replacing the working substance in the probe [4].
NqIWSCALE
Il ---,il;;~------, (D1hennocoupIe Inside the pipe
CD Thermocouple inside the liquid AI
CD ~ transducer
T 1
1 CIJnd-.r
1 1
1 1
-1-t~or
-L-
Fig. 4.1 Schematic diagram of the first thermal analysis probe developed at
McGill [3].
53
Fig. 4.2 Picture of the sodium thermal analysis probe (second probe) [4].
The present design, as described in this thesis, remedies or overcomes the shortcomings
of the previous probes. Probe operation may be halted simply by closing the return line
valve, which then prevents the liquid water from feeding the evaporator. AIso, the
cooling rate may be regulated by controlling the return line valve, and hence controlling
the amount of liquid water returning to the evaporator.
(4) Flexibility of the System
The new probe is designed to cool down the sample at predetermined cooling rates which
may be altered fairly simply during the cooling process as required. This new version of
the probe is exhibits greater ease of maneuverability than the preceding systems. The
evaporator segment is connected to the condenser by flexible hoses thus making it
possible to move the evaporator segment without having to shift the condenser.
54
Furthermore, there is the added advantage that the condenser section may be mounted
some distance away from the molten metal.
Bearing the above characteristics in mind, the new heat pipe was also designed to
monitor and control the quality of the melt.
4.3 Preliminary Design Considerations
There are several factors which play an important role in limiting the design options of
the evaporator and the condenser sections:
1. The size of the sam pie. The sample should be significant enough to represent the
whole batch of molten metal, while at the same time, small enough to be frozen and
reheated easily. The weight and the volume of the sample should also be comparable in
weight and volume to the sample obtained using the available c1assical sampling cup.
With regard to the probe, several shapes and sizes of heat pipe were taken under
consideration for the new design and construction. A heat pipe with a donut-shaped
sampling chamber and a concentric heat pipe with a cylinder-shaped sampling chamber
were built. Both designs succeeded in cooling down the sample efficiently. It was
decided, however, to carry out the experiments using the concentric heat pipe so as to
eliminate the influence of geometry and cooling behavior on the cooling curve produced.
This was deemed essential in order to compare the results obtained by both the probe and
the c1assical cup methods. In this regard, the inner diameter of the heat pipe, which is the
sampling chamber diameter, was selected at 9cm, while the height of the sample was
subject to a variation of up to 20 cm, according to user requirements. For the present
55
study, 6 cm of height were sufficient to produce a sample close enough in weight and
volume to the one produced using a sampling cup, (see Fig. 4.3).
Fig. 4.3. Evaporator section of the new probe
2. Dimensions of the working-substance chamber. This is the factor which affects
evaporator size or the outer diameter. The evaporator chamber contains the flow
modifier, the return line, and a free space into which the vapor may flow and which
should be sizable enough to avoid the sonic limit. The dimensions of this free space may
be determined by subtracting the outer diameters of the return line and of the flow
modifier from the difference between the inner diameter of the outer cylinder and the
outer diameter of the inner cylinder, (see Fig 4.4), as expressed in the following equation:
4.1
where d pis the free space dimension;
56
Do is the inner diameter of the outer cylinder;
D;n is the outer diameter of the inner cylinder;
d f is the flow modifier diameter;
dr is the return line diameter.
For the McGill heat pipe to operate efficiently, the diameter of the flow modifier should
be about one tenth of the evaporator chamber gap, while the outer diameter of the return
line should be half that of the evaporator chamber gap [1,2]. The flow modifier may be
constructed, simply, out of a helical coil placed in the evaporator chamber adjacent to the
inner wall of the evaporator or heat transfer surface. Based on past experience with the
McGill heat pipe, the cross-sectional diameter of the flow modifier was selected at 2.5
mm, the outer diameter of the return line was 8 mm, and the evaporator gap was 20 mm.
Fig 4.4. Heat pipe evaporator components
3. Operating temperature. This is a significant factor, in view of the fact that high
operating temperatures lead to high vapor pressures inside the probe which is a situation
57
to be avoided. Evacuating the probe to a negative pressure before immersing it in the
molten metal reduces the evaporation temperature and the associated saturation pressure
of the working substance. In a molten aluminum environment (600DC - 800°C), the
decrease in the evaporation temperature of the working substance is sufficient to avoid
high pressure build-up inside the evaporator section. Also, an increase in the cooling
effect inside the condenser causes the working temperature of the probe to decrease
satisfactorily. The operating temperature range was selected between a minimum of 80DC
and a maximum of 120DC, equivalent to about 0.47 atm and 2 atm of absolute saturation
pressures, respectively [5].
5. Materials and production costs. The materials used to manufacture the evaporator
and the condenser sections should be good thermal conductors and be able to withstand
thermal cycles, which may occur with great severity across the evaporator walls. Initially,
the evaporator temperature is in equilibrium with the bath temperature (approximately
750°C), then it decreases when the probe is put into operation. The inner surface, close to
the working substance, cools down rapidly to a temperature which approaches the
temperature of this working substance, then it increases again to the bath temperature
when remelting the sample is required, in this case the probe is in the off position. It
should be emphasized here that the material used to construct the probe must be of
relatively low cost and easy to handle during assembly.
58
4.4 Selection of Heat Pipe Components
The probe, once developed, is composed of five main items. These are the evaporator
chamber, the condenser chamber, the connecting hoses, the coating material, and the
working fluid. Each one of these items plays an essential role in the successful outcome
of this heat pipe project.
4.4.1 Selection of Evaporator and Condenser Materials
The chemical compatibility, working environment, degree of difficulty in production, and
the cost are all major factors which govem the choice of the material out of which the
pipe is to be constructed. The selection of the heat pipe container, in terms of
compatibility with the working fluid, was based on the literature in the field and McGill's
experience with heat pipes in the pasto Water was reported to be completely non-reactive
with copper in a copper water heat pipe [6]. However, the problem of the high dissolution
of copper, when it cornes into contact with molten aluminum, prevents it from being
considered as an option. Water will also react with chromium and nickel in stainless steel,
while the hastealloys will yield oxides of these metals and free hydrogen [7]. Past
experience with heat pipes at McGill, however, reveals that the se reactions have a limited
effect on the performance of stainless steel heat pipes, especially when medium scale heat
pipes are considered. Elalem and Zhang [1,2] experimented earlier with stainless steel heat
pipes and experienced no signs of failure over the course of their work. Based on
information provided by them, and due to the relatively low cost and immediate
availability of the product in many shapes and sizes, Type 316 stainless steel was chosen
as the most suitable material for producing the heat pipe evaporator and condenser in the
59
case of the second and third probes. Nonetheless, when the compatibility between the
heat pipe container and the liquid metal was considered, the problem of stainless steel
dissolution had to be factored into the selection process. The problem was overcome by
using boron nitride to coat the surface which cornes into contact with the liquid
aluminum. Boron nitride is customarily used to coat the molds because it adheres firmly
to stainless steel and does not tend to become wetted by liquid aluminum. As regards the
case of hastealloy X, there have been no previous attempts, so far, to use this material in
constructing heat pipes at McGill, although its capacity to withstand thermal cycles
disposed us to use this material in constructing the evaporator section of the first probe.
4.4.2 Working Fluid Selection
In conventional heat pipes, temperatures of the application environment and chemical
compatibility are the main constraints which limit the choice of a working substance. In a
molten aluminum environment, at a temperature range of 500°C-800°C, conventional
water-based heat pipes suffer from film boiling when in operation, which tends to reduce
heat transfer across the pipe wall dramatically [6]. Thus, the choices are limited to the
appropriate liquid metals: sodium, potassium, and cesium. The use of liquid metals inside
the heat pipe complicates the operating process and reduces the controllability factor. For
instance, direct control over the quantity of liquid returning to the evaporator is rendered
over-complicated, while certain of the liquid metals, such as cesium, are not a valid
option for this project because of their potential for producing environmental hazards.
Because water is not a liquid metal, it would be the best choice in this case, on condition
that the film boiling problem could be submitted to a functional solution, since it has high
60
latent heat, low liquid and vapor viscosities, and a moderate boiling temperature (about
100°C at 1 atm). Moreover, the flow of water may be controlled simply by introducing a
valve into a return line. The new heat pipe technology developed at Mc Gill overcomes
the film boiling limitation associated with using water as a working substance. Based on
the new heat pipe characteristics, a solution of 4% dish-washing detergent and distilled
water was settled on as the working fluid. The soap was added to the water in order to act
as a lubricant for the flow, thereby reducing the surface tension. Liquid detergent has an
important role to play in relatively small-scale evaporators where the surface tension of
the water is capable of blocking the return line. AIso, this is the first time that water has
been integrated as a working fluid in pipes used for thermal analysis, particularly in view
of the fact that, hitherto, regular water heat pipes have encountered limitations as a result
of film boiling. The addition of a lubricant is, therefore, one of the key features which
helped elevate the present probe over conventional heat pipes and the first two heat probe
models.
4.4.3 Coolant Fluid Selection
Both water and compressed air were used to extract heat from the condenser. Water has
the advantage of having a higher heat capacity than air; thus, a much lower flow rate
would be required to extract the same amount of heat using water as that which would be
required using compressed air. On the other hand, compressed air has the advantage of
being available in every foundry and can be handled safely and easily. The exhaust can
be discharged into the surrounding ambient air, since it is environmentally friendly,
whereas the hot waste water would need to be drained off.
61
4.5 Cooling System Design
Two condensers of the shell-cylinder type were built for the purposes of
dissipating heat to the surrounding environment. Each one of the se two condensers was
built according to different specifications and using different criteria. The first uses water
as the secondary fluid, while the other uses compressed air and a solid block of stainless
steel for cooling. The advantage of using this type of condenser is that a highly efficient
cooling of the primary fluid may be achieved for a relatively short period of time, and
consequently, a smaller size condenser is needed throughout the period in question. The
advantages of using a water-cooled condenser is that the temperature of the working fluid
remains relatively low, that is to say, lower than the working fluid temperature associated
with the use of compressed air to cool the condenser. It should be noted that water may
be used to cool both condensers; also, that using a thick block inside the condenser will,
however, tend to add considerable thennal resistance across the heat transfer path causing
a noticeable difference in temperatures across the block itself, and as a result, the
temperature of the working fluid will rise.
The coolant fluid is directed to flow through a cooling jacket designed to envelope the
condenser segment of the heat pipe in which the walls are made of Type 304 grade
stainless steel. The wall of the first condenser was 1 mm thick, based on a compromise
between strength and thennal resistance, whereas the walls of the second condenser were
30 mm thick. The proposed cooling system is illustrated in Figure 4.5.
62
Vacuum Control Port
Cooling Fluid Outlet Port
Cooling Fluid Inlet Port
Vent Port
Working Fluid Return Port
Working Fluid Main Port
Fig. 4.5 A schematic drawing of the condenser cooling arrangement.
63
References
1. Zheng, G. "A Novel Flow Modified Heat Pipe Development and
Experimental Investigation", Ph.D. Thesis, Department of Mining,
Metals and Materials Engineering, McGill University, 2003.
2. Elalem, K. "Applications of Heat Pipe Technology in Permanent Mold
Casting of Nonferrous Alloys", Ph.D. Thesis, Department of Mining,
Metals and Materials Engineering, McGill University, 2004.
3. Meritian, M. "Thermal Analysis of Aluminum Foundry Alloys by a
Novel Heat Pipe Probe", Ph.D. Thesis, Department of Mining, Metals
and Materials Engineering, McGill University, 1995.
4. Mahfoud, M. "Controlled Thermal Analysis Using Heat Pipe
Technology", Ph.D. Thesis, Department of Mining, Metals and
Materials Engineering, McGill University, 1997.
5. Peterson, G. "An Introduction to Heat Pipes; Modeling, Testing and
Applications", Wiley and Sons Inc., New York, NY, 1994.
6. Faghri, A. "Heat Pipe Science and Technology", Taylor&Francies,
Washington D.C., 1995.
7. Silverstein, C. "Design and Technology of Heat Pipes for Cooling and
Heat Exchange", Hemisphere Publishing Corp., Washington, D.C.,
1992.
64
Chapter Five
Part One
Experimental Procedures
5.1 Introduction
As mentioned earlier in Chapter Four, in order to fulfill the objectives ofthis study, a new
in-situ thermal analysis setup was designed and implemented to replace the classical
thermal analysis technique used so far for aluminum alloys. The new setup uses enhanced
heat pipe technology to freeze the sample inside the container vessel or furnace.
The evaporator segment in this system acts as a mold for the molten metal sample, and
may be cooled down as desired. The probe evaporator is immersed in the liquid metal and
once in the probe is set in running position, by introducing the working fluid is into the
evaporator. It subsequently evaporates, and thereby establishes the temperature of the
working substance, namely the probe operating temperature. Because of the temperature
gradient across the evaporator wall, latent heat from the fusion of molten metal is
extracted and the pre-selected sample of molten metal becomes solidified inside the probe
core. The temperature of the solidifying sample is recorded instantaneously by means of a
thermocouple placed at the center of the sample. During the freezing and remelting
processes, time and temperature are recorded by a data acquisition system. When the
sample is completely frozen, the cooling process is halted by preventing the working
65
substance from feeding the evaporator. Consequently, heat transfer between the
condenser and the evaporator becomes insignificant compared to the amount transferred
to the sample from the molten pool in which the sample remains immersed, and as a
result, the sample remelts. The semi-continuous nature of this system may be attributed
to the remelting process.
Three heat pipes, each different in terms of configuration and shape, were successively
designed and built. The unsatisfactory quality of the cooling curves obtained by the
original heat pipe probes was the main reason for further modifying the design.
5.2 First Probe
The first probe was designed based on the new heat pipe technology developed at McGill
University. The probe parameters were chosen exactly according to McGill heat pipe
specifications [1] [2]. A separate return line with an outer diameter of 10 mm, together with
a helical flow modifier composed of a helical spring 2 mm in diameter and a 20 mm pitch
were used. This helical spring was positioned adjacent to the inner wall of the evaporator
to assist the main two-phase flow, to initiate spinning behavior and to overcome film
boiling drawbacks. The separate return line was used to eliminate the entrainment effect.
In order to minimize the influence of the feeding or return line on the behavior of the
main vortex flow, the return tine was positioned exactly at the center of the pipe with a 1
cm gap from the bottom of the evaporator.
In this probe, two flanges were used with a copper nng In between to facilitate
assembling and disassembling the evaporator section. A coating of graphite was used to
66
avoid direct contact between the melt and the material out of which the heat pipe was
constructed.
The first design shown in Figure 5.1 was made of Hastelloy X, which is composed of
48.3% Ni, 19.19% Fe, 21.69% Cr, 9% Mo, and 1.5% Co. This material was used because
of its capacity to withstand the high thermal cycles resulting from cooling and heating,
and also because of its high thermal conductivity compared to stainless steel.
Air Gap
6 in
A-+-3cm
Fig. 5.1. First design, evaporator section.
Unfortunately, a number of serious problems arose while testing this probe both in a hot
environment and in a molten aluminum environment. The first of these was the cost of
the material which was substantially higher than that of stainless steel. The second was
67
the oxidation of the material which forms a layer of oxides on the heat transfer surface
and reduces the efficiency in extracting heat from the sample volume (see Figure 5.2).
The third difficulty was that sorne of the Hastelloy X material tended to dissolve in the
molten aluminum (see Figure 5.3). The final problem was that invariability of the
sampling volume, as displayed by this probe model, indicated a certain lack of flexibility
in the design.
a) Oxide layer on the return line and outer surfaces offlow modifier.
b) Oxide layer on the inner surface of the evaporator
Fig. 5.2. Oxides accumulated on the evaporator surfaces
a) return line and flow modifier, b) the effective heat transfer area (evaporator wall).
68
Fig. 5.3. Outer surface offirst probe after experiments in a liquid metal
environment.
5.3 Second Probe
Stainless steel was used to manufacture the second probe (shown in Figure 5.4),
which was designed and built to overcome the shortcomings of the first probe. Thus,
since an adjustable sample volume was one of the objectives to be met, a movable outlet
jacket was introduced into the design with the aim of fulfilling two main functions:
limiting the volume of the sample and insulating the rest of the evaporator from direct
contact with the molten metal. An air gap of 3 cm was introduced between the outlet
jacket and the outer wall of the evaporator. This air gap provides thermal resistance in the
direction of the heat transfer. Thus, heat removal would occur mainly through the sample
chamber wall, while a small fraction of heat passes through the remaining portion of the
69
pipe. The botlom of the pipe was also insulated by an air gap trapped between two discs
welded to the botlom end of the pipe. Preliminary tests showed that this design made it
possible to capture cooling curves which were similar to those obtained by traditional
thermal analysis techniques using a sampling cup. The only difference to be observed
was in the ability to detect the undercooling which occurs prior to solidification; it should
also be noted that the tests were carried out using Type 356 aluminum alloy. This
particular model of the probe was capable of measuring temperatures at four different
locations away from the pipe surface, ranging from 1.7 cm to 4.7 cm from the center.
Fig. 5.4. Second probe
5.4 Final Design
The third probe was designed so as to simulate the manner in which heat is
transferred during thermal analysis using the sampling cup method, while at the same
time maintaining all the advantages inherent in the original versions. Figure 5.5 shows
70
two views and the dimensions of the third probe evaporator segment. For the purposes of
this new model, two major modifications in design were made. The first one incorporated
a hollow concentric cylindrical evaporator into the design. The inner wall of the cylinder
is 2 mm thick and 9 cm (3.5 in) in diameter, which is about the same as the diameter of
the cylindrical sample, while the outer wall is also 2 mm thick but has an inner diameter
of 14 cm (5.5 in). The second major modification was to reposition the return line and
place it adjacent to the outer wall of the evaporator. It is believed that this particular
positioning of the return line will play an important role in redirecting the flow towards
the surface of the inner wall. Consequently, any gas layer which might form on the inner
surface of the evaporator would be destroyed and the operating range thus lengthened, in
that the effects of film boiling as a limitation would be much reduced. This enhancement
in design places the new generation of Mc Gill heat pipes in a new individual category. It
is now possible, for the first time, to use annular shapes to extract heat efficiently either
from the inner surface by adding flow redirectors, and/or from the outer surface as a
result of vortex behavior due to the presence of a flow modifier.
The evaporator section, as shown in Figure 5.5, was built of several parts. The metallic
wall isolates the sample of molten aluminum, thereby preventing it from mixing with the
working substance while also conducting the heat from the sample to the working fluid. It
is important to mention here that the poor heat transfer characteristics across this wall,
when operating conventional heat pipes in such an environment, is the main reason for
not using them in the thermal analysis of aluminum alloys.
71
a) Evaporator, third design.
(Two Phase Line)
0.0.1.65em 1.0. 1.25 em
gem
b) Configuration of second and third designs
in which high cooling rates are required.
(Return Line)
0.0.0.8 em 1.0.0.6 em
0.0.0.3em Piteh 2 em
C) Sketch ofthird probe ( shown in Fig. 5.5.a)
Fig. 5.5 Third design
72
In our capacity as research engmeers and developers of industrial designs, we are
interested in extracting the heat from a limited volume, which is represented by the
sample alone. Since the main aim of this project is to create an in-situ probe, it is
necessary to reduce the amount of heat transferred from the remaining melt in the
crucible either to the heat pipe surface or to the sample itself. This reduction may be
achieved simply by adding a thermal resistance factor such as an air gap or an insulating
layer trapped in the outer sleeve, a concept which has already been included in one form
or another in all of the three designs produced to date. Thus, the outer surface of the
evaporator is insulated using a 5 mm layer of insulating wool covered by a stainless steel
sleeve 2 mm thick. The sleeve also acts as a barrier to prevent the molten metal from
wetting the insulation. This insulating of the outer surface of the evaporator is an
essential step because it significantly reduces the heat transferred from the molten bath to
the probe. The outer sleeve is coated with boron nitride to prevent it from dissolving in
the molten aluminum.
Two return lines placed opposite each other inside the evaporator, close to the outer wall,
supply the evaporator with the liquid phase of the working substance. These return lines
are made of two lengths of piping, 8 mm in outer diameter and 1 mm thick. Although the
pressure is higher in the evaporator section than in the condenser section, the flow inside
the return line goes towards the evaporator. This movement is due to the pressure of the
water column and the addition of a vent line so as to prevent the vapor from blocking one
end of the return line.
73
The main line begins with a rectangular section which then converges at the end of the
evaporator, and is connected at the other end to a flexible hose which is attached to the
condenser as shown in Figure 5.5 (b). The convergent section was chosen as suitable for
avoiding any stationary gas in the evaporator, or at least minimizing its quantity. The
presence of stationary gas will tend to act as an insulating layer; thereby reducing the
efficiency of the evaporator in extracting heat from the sample.
The connecting hoses, which are made of flexible metallic material, connect the
evaporator segment to the condenser. Such hoses afford a wide range of maneuverability
when placing or moving the heat pipe inside the crucible, and for keeping the condenser
at a safe distance from the melt.
The evaporator chamber is provided with an adjustable lower plate for two reasons. The
first is to reduce the amount of heat transferred from the melt to the sample during the
freezing process; furthermore, it was found that the presence of the plate has an effect on
the quality of the signal produced by the thermocouple: it reduces thermal noise, or
convection, which occurs naturally within the melt. The second reason is that an
adjustable lower plate would facilitate the extraction of a physical sample of the melt for
further investigation.
A preliminary test was carried out and the results were promising. It is important to
mention here, that both the second and third designs may be used in tandem when high
cooling rates are required (see Fig. 5.5.b).
74
5.5 Experimental Setup
The same setup was used for all three probes (see Fig. 5.6). The evaporator segment is
connected to the condenser by means of flexible hoses. As a security measure, a stand
was mounted to keep the condenser away from the furnace. In view of the fact that a
frequent cause for concem related to heat pipe performance is the existence of non
condensable gases in the heat pipe system, a vacuum pump was used to reduce the
pressure inside the probes and to eliminate this superfluous gas. If it is present, this gas
will accumulate at the condenser end of the pipe and form a gas plug which shuts off a
part of the condenser, reduces the condenser effective area, and ultimately reduces the
condenser cooling load which leads to an increase in the temperature and pressure of the
working substance.
There are two ways of removing the air from the heat pipe system. The tirst is to heat the
system to about 100°C and then to open the vacuum valve at the top of the condenser
until a stream of vapor emerges from it. The second method is to rai se the temperature of
the heat pipe to 80°C and then to apply a vacuum pump while at the same time
monitoring the vacuum pressure gauge until it reaches 40.5 kPa. Once the evacuation
process is completed the vacuum valve should be closed again.
In terms of pipe mobility, a clamping system or holder was used to place the evaporator
in the liquid metal bath, and to raise and lower it according to experimental requirements.
75
Fig.5.6 Experimental Setup
5.5.1 Measurements and Instrumentation
Temperature Sensor
A number of considerations influence the choice of the type of thermocouple used to
detect the temperature for a specifie application. The most important of these is the time
constant which is defined as the time required by a sensor to reach 63.2% of a step
change in temperature under a specifie set of conditions. An exposed junction
thermocouple is recommended where a rapid response time is sought for. The junction
extends beyond the protective ceramic sheath to provide accurate yet rapid responses.
76
The sheath insulation prevents the infiltration of fluid which could cause faulty readings.
The thermocouple time constant and response time are functions of the thermocouple
diameter.
Type K thermocouples have an operating range between -100°C and 1250°C. The time
constant for this type of 0.01" diameter thermocouple is 0.15 seconds and the response
time is 0.8 seconds as measured for exposure between 93.3°C and 37.8 oC in a water
medium. These thermocouples were thus used to detect temperatures at the following
locations:
1) For the aluminum sample, a 0.01" diameter thermocouple was placed in the probe
core where the aluminum sample was located.
2) For the working substance temperature, two thermocouples 0.02" in diameter
were used here. The tirst was inserted into the middle of the condenser through
the condenser end cap, while the second was inserted inside the evaporator
section. This disposition of the thermocouples provided instantaneous monitoring
of isothermal conditions inside the pipe. Hence any changes in the fluid in the
evaporator section could be detected promptly and rectified either by increasing
the return flow rate or reducing the cooling load in the condenser.
3) For the inlet and outlet fluids, four thermocouples were mounted at the inlet and
outlet ports of the condensers and were used to calculate its cooling load.
4) For the liquid metal bath, the thermocouple used to measure the temperature of
the molten metal bath was sheathed with 3.2 mm stainless steel sheath and sealed
inside a ceramic tube for protection.
77
Pressure Sensor
A vacuum pressure gauge coupled with a vacuum pump was used to evacuate the air
from the system prior to the tests. It was also used to monitor the pressure inside the
system during the test proper. The operating range of the pressure was set at between
40.5 kPa and 202.3 kPa and could be modified either by reducing the amount of liquid
feed to the evaporator or increasing the cooling load on the condenser segment; or a
judicious combination of the se steps.
Flow Sensor
Two types of rotary flow meters were used for monitoring the flow. The first is suitable
for compressed air, and is able to read up to 90 standard cubic feet per minute; the second
one is suitable for measuring the water flow and reads up to 0.1 l/sec. The flow meter is
an essential part of the setup of the experimental apparatus considering that the amount of
heat extracted from the condenser is directly proportional to the mass flow rate of the
cooling fluid. The quantity of this coolant as it passes through the condenser is controlled
by introducing a gate valve at the condenser inlet port. The flow rate is used to calculate
the amount of heat rejected at the condenser segment, which is also equal to the amount
ofheat absorbed by the evaporator.
5.5.2 Experimental Methodology
A former attempt, made by the Gruzleski and Mucciardi Research Group to develop an
in-situ thermal analysis probe, was intended to have the capability of regulating the
78
cooling rate of the solidification phase [3, and4]. They endeavored to make use of
conventional heat pipes in conducting the thermal analysis of aluminum alloys. An
elaborate device was assembled for this purpose and a doctoral thesis reported on the
development and outcome of the project. The complexity of its operating procedure,
however, as weIl as the design itself, make it unsuitable for industrial applications, thus
the concept was withdrawn for the then foreseeable future.
After McGill's great success in developing a new heat pipe for the aluminum industry to
cool permanent molds, and also in developing a new type of oxygen lance for the steel
industry, the need to design a functionally updated probe for in-situ thermal analysis re
emerged once again[2].
The survey of the literature which was carried out in this regard revealed that, to
date, in-situ thermal analysis does not appear to be used either in industrial foundrles or at
the laboratory level. To the best of the author's knowledge, the latest technique used in
laboratories is based on melting the sample in a small induction furnace in such a way as
to control the temperature and consequently the cooling rate of the sample.
To date, only three individual designs have been implemented in departmental machine
shops and tested in the MMPC laboratory (McGill Metals Processing Centre) at McGill
University, in the TAMLA group laboratory (Technologie avancée des métaux legers) at
University of Quebec at Chicoutimi, and also at the industrial plant of Grenville Castings,
Ontario.
79
At the laboratory level, a melt of about 33 kg of 356 aluminum alloy was prepared in a
35-kg capacity silicon carbide crucible which was coated with boron nitride. Table 5.1
shows the chemical composition of the alloy used in this study. The crucible was heated
in an electric resistance furnace. AIl thermal analysis tests were carried out in the
following two ways: 1) using the classical sampling cup, and, 2) using the new in-situ
probe. The tests were thus carried out in pairs and generated simultaneous cooling curves.
A preheated graphite sampling cup (650°C) was used to carry out the classical thermal
analysis, whereas the evaporator section of the modified heat pipe was used as an in-situ
mold for the new technique.
Element
Al% Si% Mg% Cu% Mn% Fe% Zn% Ti% StJ/o
Alloy
356 92.6 6.79 0.38 0.006 0.0009 0.062 0.006 0.1 0
390.1 77.47 17.3 0.54 4.33 0.06 0.32 0.06 0.07 0.001
319 91.6 6.3 0.07 >1.5 0.05 0.35 0.05 0.007 0
6063 98.7 0.43 0.52 0.06 0.037 0.18 0 0 0
Al-Si binary alloy 92.8 6.82 0.016 0.11 0.02 0.21 0.002 0.005 0
Table 5. 1. Chemical analysis of the alloys used in the present study
The heat pipe was regulated to cool down the solidifying sample at a steady heat
extraction rate. In order to obtain the results reported in this study, the samples were
80
cooled from 750°C to 500°C at an average cooling rate of 0.8°C/s. This data reflects the
conditions prevailing in the permanent mold casting process. The melt temperature was
measured at the center of the sample by a K-type thermocouple which was linked to a
data acquisition system and recorded at 0.2 second intervals. The cooling curves
obtained at this cooling rate revealed a greater quantity of information than did those
recorded under conditions of more rapid cooling. The results were then compared with
the cooling curves obtained by the conventional sampling cup technique which involved
a 0.8°C initial cooling rate and a O.3°C average cooling rate.
For each cooling curve, the first derivative curve was plotted, and the portion of the
cooling curve pertinent to this study was expanded to allow the retrieval of data from
both the cooling curve and the first derivative curve. This last, obtained by the classical
method, was smoothed by averaging the values over 2-second time intervals. It was
necessary to do this to reduce the thermal noise that the classical technique generated.
Physical samples from both the new and classical techniques were extracted and prepared
for metallographic study by cutting a piece from the center where the thermocouple was
placed. Sorne samples were observed after etching and others were examined after
polishing. A JEOL scanning electron microscope (Japan Electron Optics Laboratory) was
used to determine the surface fraction of the iron and copper phases. This microscope
was coupled with an EDS (Energy Dispersive Spectrometer) to assess phase
identification.
81
Once the EDS analysis was used to identify the phases in the microstructure, an image
analyzer in conjunction with microstructure photography was used to examine the extent
of the modification taking place. Comparisons between the proposed updated technique
and the classical thermal analysis method were carried out based on the considerations
outlined in the following steps:
1. Monitor the response of the cooling curve to determine the effect of several
concentrations of grain refiners in the melt. This was achieved by adding
titanium to the melt in the form of AI-1 0% Ti master alloy and investigating its
impact on the cooling curve. Two concentrations of titanium were used: 0.2
wt% and 0.42 wt%.
2. Monitor the effect of a modifier in the melt on the cooling curve. The level of
the modifier in the melt was changed by adding strontium in the form of AI-
10% Sr master alloy.
3. Examine the sensitivity of each of the two techniques in order to detect minor
reactions which might occur because of the presence of impurities in the melt.
This operation may be carried out: by adding iron to the 356 aluminum melt in
the form of a master alloy (25% Fe-75% Al); by adding magnesium to the
temary alloy melt (Al-Si-Cu) in the form of a master alloy (25% Mg-75%Al);
and by adding copper to the binary alloy melt (Al-Si) in the form ofpure copper
metal.
4. Examine the operating range for the new probe. As mentioned earlier, the
temperature of the environment is the reason why conventional heat pipes are
not used in this context. Thus, one of the critical issues in designing the
experiments for this project was ascertaining that the new probe would operate
82
efficiently in covering the solidification range for the aluminum alloys. The
temperature of the working environment may be increased by raising the
liquidus temperature of the 390 aluminum eutectic alloy. This is achieved,
according to the alloy phase diagram, either by increasing or reducing the
concentration of the alloying element (Si) in the melt. Five different
concentrations of silicon were investigated; the concentrations, in weight
percent, varied between Il % and 20%. The maximum liquidus temperature was
about 695°C associated with a 20 wt% concentration of silicon in the melt.
5. Test the capability of the new technique to pinpoint minor reactions during the
solidification of 6063 wrought aluminum alloy.
5.5.2.1 The Cooling Period
Once the temperature of the sample and that of the evaporator reach the bath temperature,
or molten metal temperature, the test is initiated. The liquid water is fed to the evaporator
by gradually opening the return line valve. This procedure determines the amount of
liquid flowing to the evaporator and consequently the rate at which the heat is extracted
from the sample. Thus, the setting of the return line valve was predetermined by the
required cooling rate of the aluminum sample. High flow rates from a wider opening
result in high cooling rates, and vice versa. Based on this, it is possible to obtain a range
of cooling rates simply by changing the return valve setting. It is important to mention
here that the cooling rate is strongly dependent on the condenser cooling load, or in other
words on the cooling fluid flow rate. For thls reason, the cooling loads during the tests
were decided upon by setting the same flow rate for the cooling fluid in all the tests. In
83
fact, it is the working substance that is cooled by the cooling fluid and not the aluminum
sample itself, therefore, any modification in the temperature or pressure of the condenser
will necessarily affect the temperature of the sample.
5.5.2.2 The Heating or Remelt Period
To ensure that the results are reproducible, the thermal analysis test was repeated several
times for the same sample and under the same conditions. The return Hne valve is closed
after freezing the sample and obtaining the cooling curve, and as a result, the cooling will
cease. The heat transferred to the sample from the molten bath causes an increase in the
sample temperature and will eventually remelt the sample. Once the sample regains its
earlier temperature, as recorded prior to the cooling process, the test is repeated once
again. Most of the experiments were conducted at a superheat of about 150°C except for
the industrial tests in which the superheat was approximately 100°C causing the sample to
remelt within a very reasonable time frame.
84
Chapter Five
Part Two
Preliminary Tests
Hot Environment Tests
The first probe underwent two series of experimental trials, the first of which involved
testing under the hot environment conditions of a natural gas furnace with an operating
temperature of up to 1300°C. First of all, the evaporator section of the heat pipe was
introduced into the furnace. lnitially, the probe was in the off-mode until its temperature
approached that of the environmental temperature inside the furnace, upon which the
color of the probe turned from black to red (see Figure 5.6).
(a) Probe as seen inside gas furnace (b) Probe as seen removed from furnace
(c) Water fed to evaporator
Fig. 5.6 First probe under elevated temperature conditions
85
Two thermocouples, opposite to each other, were used to detect the temperature at a
distance of 5 mm from the probe surface. The working substance, water, was fed to the
evaporator by opening the return line. The respective temperatures of the heat pipe inside
the condenser, the inlet cooling water, and the outlet cooling water were recorded by
means of the thermocouples and a Data Acquisition System. Figure 5.7 shows the
variation in these temperatures, as a function of time, while the heat pipe was set in
operation. The temperature readings from both the thermocouples located near the probe
surface, decrease sharply, although there is a 5 mm gas gap between them. This indicates
that the cooling was highly effective in that specifie region.
1200 1150 1100 1050 1000
950 900 850
i.J 800 I!! 750 ::s 700 i! 650 G) 600 ~ 550 G) 500 1- 450
400 350 300 250 200 150 100
50 o
h.
o
Thermoco'f'le ~
/' '" AfP' \. ~. /~ ~
/.:\ ~..../'-fu ,1' " // "'----------'~ ,~ 1 // \ r
// \ // ThermOCOlflle
Il /1
1 Il
Il
1000 2000 3000 4000 5000 time (sec)
--~ ;/
1/ J
"
6000
Figure 5.7. Temperatures recorded 5mm from both sides of the probe surface as it varied
with the opening of the return valve.
86
Figure 5.8 shows the variations in the dissipated energy in the condenser with the
opening of the return line valve. The figure shows the feasibility of regulating the heat
absorbed by the probe through controlling the amount of liquid water fed to the
evaporator section. Four cases are c1early illustrated in this figure: (i) the rapid response
of the heat pipe which is represented in area A; (ii) the amount of heat extracted and how
this corresponds to a slightly open return valve, approximately 3800 W, in area B; (iii)
area C shows the decline in the amount of heat extracted as the valve opening is reduced;
(iv) areas D and E show the response of the probe to cutting off the return fluid, and then
the reintroducing of a small amount, respectively; (v) areas F and G show the rapid
response of the probe when the return line is opened to the full, and then the maximum
heat which can be extracted by the probe when it is operated under similar conditions.
This indicates that the amount ofheat removed by the probe is proportional to the amount
of water fed to it.
Diaslpatad Energy
6000~---------------------------------------.~----------~
A
5000
4000 l l "00 evaperator la ............. -~ oMokmg_ookw
2000
The_ ".lavA la
1000 fully opened 1 The valve la allghtlyTh. valve-Ia openeci .. --__ The valve la closed completelv
allghtly op.ned and th.n It 1. slightly opened O+-~--~~--~~~--~~--~~~~~~~~~--~~--~~ 160017001800190020002100220023002400250026002700 2800 2900 3000 3100 3200 3300 3400 3500
TlmeC_1
Fig. 5.8 Energy absorbed by the probe in a hot environment.
87
Molten Metal Environment Test
The second set of tests was carried out inside molten 356 aluminum alloy. Obtaining the
resultant cooling curves was the primary objective, while regulating the cooling rate was
the second. Figures 5.9 (a) and (b) show the cooling curves of 356 aluminum alloy at two
different cooling rates; and (c) the cooling curve obtained using the conventional thermal
analysis method. These figures show the ability of the probe to detect the start of
solidification, the eutectic reaction, and the end of the solidification process. However,
the probe is still not able to detect the occurrence of the undercooling phenomenon prior
to the beginning of solidification or its occurrence prior to eutectic solidification. This
lack may be due to a deficiency in the size of the sample whose volume may be
insufficient for the thermocouple to sense the small amount of energy released;
otherwise, it may be due to the high temperature gradient in the direction of
solidification, with the solid-liquid interface propagating against the direction of cooling
in the manner of a wave.
700
680 1
660
U 640 0 ~
620 .. ::1 .... Ga .. 600 ~ CI. E 580 ~
E-
t\
\ i 1 1
\ '-- 1
------- ----560
540
520
~ !
~ 1
500 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900
time (sec)
(a) Cooling curve obtained by the first probe at O.3°C/s
88
700 690 680 670 660 650 640 630
~ 620 GO ~ 610 ~ 600 ~590 ; 580 1- 570
560 550 540 530 520 510 500
\\ ~ ~
~--........ ~
-------~ -
!
=------- --.......
-....... ~
3300 3350 3400 3450 3500 3550 3600 3650 3700 3750 3800 3850 3900
time (sec)
700 690 660 670 660 650 640 630 620 610
U600 0 .. 590 ; 580
ë 570 8,560 E550 ~540
530 520 510 500 490 480
b) Cooling curve obtained by the first probe O.6°C/s.
1\ \ \ \ ~ ~ ~
"" ~
'" "" "'" """ "'" '" 3135 3335 3535 3735 3935 4135 4335 4535 4735 4935 5135 5335
time (sec)
(c )Cooling curve obtained by conventional thermal analysis with sampling cup
Fig. 5.9 Cooling curves obtained by the first probe and sampling cup method
89
The second probe sampling chamber is three times bigger than the first one. Four
thermocouples were used to measure the temperature at four different locations away
from the probe surface. Figure 5.10 shows four cooling curves, each representing the
solidification phenomenon at that particular position of the thermocouple. The curves do
not show the undercooling feature due either to the halting of the process prior to
solidification or to the eutectic reaction. Moreover, the high temperature gradient
displayed at the four thermocouple locations, about 1.45°C/mm, may be observed clearly.
This observation is in keeping with the nature of the propagation of the solid-liquid
interface.
700 690 680 670 660 650 640 630
.~ ~
P 620 I!! 610 .a 600 l! 590 III
~ -------~ 580 ~ 570
560 550 540 530 520 510 500
3429
\~~
"" ~----~ ~-------~ ---=-'-/~
v
-.....---.
3529 3629 3729 3829 time (8)
==== ~ = ~
~,'\ '" ,\ \~ -'l~
3929 4029 4129
Figure 5. 10. Second probe design, with four thermocouples.
90
The third probe was designed to overcome the temperature gradient problem by applying
radial cooling toward the center, for which purpose a further heat pipe was added instead
of an outer jacket. Figure 5.11 shows the cooling curves for this combination of two
concentric heat pipes. The rapid speed of cooling caused the temperature of the sample to
decrease from 750°C to 500°C within a time lapse of 120 seconds, although the outer heat
pipe was put into operation after 70 seconds. In order to reduce the cooling speed, the
inner or central pipe was removed. Figure 5.12 shows the cooling curve obtained both by
the probe with only an outer heat pipe in operation, and by the classical sampling cup
method; the features of both the curves seem to match those obtained by the traditional
method. The first derivative for both curves is plotted against time (see Figure 5.13).
750
700
650
600
550
u 500 o
f450 ::1
f400 cu Ë'350
~3oo 250
200
150
100
50
~
"
o 8380
- -
Middle 1 h .... t nin ..
te~perature
/ 1
8400 8420
out~r HP on
.""
8440 time (sec)
""
outer heat pipe
;~:::~;;'~" both ~
~, --::::
"" ~
8460 8480
Figure 5.11. Two concentric heat pipes
91
~ "
-
8500
650
640
630
620
610
600
~ 590
::: 580 .s f 570 ~ g.560 E ~ 550
~
"" ~ ~
~ ~ ~
~ \ ~
-'1 ~
540
530
520
510
500
l \
860 910 960 1010 1060 1110 1160 1210
time (sec)
Figure 5.12 (a) Cooling curve obtained by the third probe with oruy outer HP in
680 675 670 665 660 655 650 645 640 635 630 625 620 615 610 605 600 595 590 585 580 575 570 565 560 555 550 545 540 535 530 525 520 515 510 505 500
"-
operation.
...... ........ - - -....
" "--"'. ~
"-
" ~ ~
710 790 870 950 103 111 119 127 135 143 151 159 167 175 183 191 199 207 o 0 0 0 0 0 0 0 0 0 0 0 0 0
time (sec)
Figure 5.12 (b) Cooling curve obtained by sampling cup technique.
92
3 2.8 2.6 2.4 2.2
2 1.8 1.6 1.4 1.2
.-. 1 ~ 0.8 U 0.6 ~ 0.4 ~ 0.2 ~ 0 .. -O.~ .~ -0.4 .. -0.6 ~
~ -0.8 .... -1
NI. .. A • A .. 1 • ·V· "W ... 1IM.NI'of'rOI ~
l1li \~\J'" 910 îT .- 110 1160' •• 1210 V lA. A \
. .... J' r .., -1.2 ..
~ -1.4 -1.6 -1.8
-2
/if T7
\1 -2.2 -2.4 -2.6 -2.8
-3
Figure 5.13 (a) First derivative of the curve in Fig. 5.12.
3 2.8 2.6 2.4 2.2
2 1.8 1.6 1.4 1.2
1 0.8
..-. 0.6
O~ 0.4 _ 0.2 ~ 0 .~ -0.27 ~ -0.4 .. -0.6 ] -0.8 .... -1
-1.2 "E -1.4 ~ -1.6
-1.8 -2
-2.2 -2.4 -2.6 -2.8
-3
\
\' '1"
.IL ...... u . Lili
910 1110 1310 1510 17f(f ..,.910'
1
time (sec.)
,
time (sec.)
Figure 5. 13 (b) Traditional thermal analysis, first derivative curve.
Based on the results previously shown in this section, it was decided to use the outer heat
pipe alone, as the thermal analysis probe, to carryout both the laboratory and industrial
tests which are presented in the results and discussion chapter.
93
References
1. Zheng, G. "A Novel Flow Modified Heat Pipe Development and
Experimental Investigation", Ph.D. Thesis, Department of Mining,
Metals and Materials Engineering, McGill University, 2003.
2. Elalem, K. "Applications of Heat Pipe Technology in Permanent Mold
Casting of Nonferrous Alloys", Ph.D. Thesis, Department of Mining,
Metals and Materials Engineering, McGill University, 2004.
3. Meritian, M. "Thermal Analysis of Aluminum Foundry Alloys by a
Novel Heat Pipe Probe", Ph.D. Thesis, Department of Mining, Metals
and Materials Engineering, McGill University, 1995.
4. Mahfoud, M. "Controlled Thermal Analysis Using Heat Pipe
Technology", Ph.D. Thesis, Department of Mining, Metals and
Materials Engineering, McGill University, 1997.
94
Chapter Six
Results and Discussion
6.1 Introduction
After confirming that the new system was worthy of confidence, the next step was to test
the degree of sensitivity of the probe in detecting a number of melt parameters as
imposed in aluminum foundries for enhancing cast quality. Normally, these parameters
are detected by means of the c1assical thermal analysis technique or the application of
sophisticated and costly devices such as emission spectrometers.
Several melt treatments were investigated for the purposes of developing this thesis,
inc1uding assessment of grain refiners, assessment of the level of modification of eutectic
silicon, and detection of the intermetallic reactions which take place during the
solidification of the aluminum alloy.
Since the probe is a heat transfer device and its applicability depends on the working
temperature range, five liquid aluminum alloys were chosen for testing: 356,390, 6063,
and 319 aluminum alloys, and an Al-Si binary alloy. See Table 5.1 for the chemical
95
concentrations of the alloying elements. Among the se, AI-20 wt%Si alloy has the highest
melting point of about 690°C.
6.2 Grain Refinement
One of the comparison parameters between the new probe and the classical thermal
analysis technique is to monitor the response of the cooling curve, as generated by each
of the two techniques, in order to determine the effect of several concentrations of grain
refiner in the melt. This was achieved first by adding titanium to the melt in the form of
AI-10%Ti master alloy and then investigating its impact on the cooling curve.
Two different levels of titanium were used for grain refinement: 0.2 wt% and 0.42 wt%.
An Al-lO%Ti master alloy was used to introduce titanium to the melt, and was added in
two separate stages. The amount of titanium to he added was calculated on the basis of
the melt weight of 32 kg. A graphite plunger was used to add small pieces of the Al-
10% Ti master alloy which was introduced into the bottom of the crucible so as to avoid
oxidation. The melt was then stirred for three minutes to avoid sedimentation of the
titanium and to improve homogeneity. After that, the probe was inserted into the melt,
and when the temperature of the probe reached that of the melt, the test was triggered. A
scoop of the melt was then poured out into a sample cup which was a graphite mold in
the form of a truncated cone, Il cm high, and 7 x 5 cm in diameters; at the same time the
probe evaporator was wetted with the working substance by gradually opening the return
line valve. Ralf a turn of the valve was found suitable to cool the sample at a rate of
0.8°C/sec when the condenser water-flow rate was 15 gis. The temperatures of the
96
solidified samples, as extracted by the c1assical cup and probe methods, were recorded by
means of two thermocouples placed at the center of each of the two samples. A computer
interface card together with a Lab View software program installed on a regular personal
computer were used to record and store the measured data with the corresponding time
values in a data file. The data were then analyzed and plotted against time using either
Excel or Matlab software programs.
The results were as expected. The grain size decreased according to the increase in the
grain refiner concentration in the melt. Moreover, the rate of the reduction of the average
grain size diminished with an increasing concentration of grain refiner. Figures 6.1 and
6.2 show the grain size and how it varies with the various concentrations of titanium in
the melt. In order to verify the results generated using both techniques, a sample for
metallographic investigation was sectioned from the solidified samples obtained. These
extracted samples were then ground using 180, 320, 400, 600, and finally 800 grit
sandpaper. After that, they were polished, etched and photographed to measure the grain
size. The results show that the grain size decreases as the concentration of the grain
refiner (Ti) in the melt is increased.
97
a) 356 Al with 0 wt% Ti
c) 356 Al with 0.4 wt% Ti
b) 356 Al with 0.2 wt% Ti
Fig. 6.1 Microstructures of356 Al alloy with addition of Ti to the melt in concentrations
of(a) 0%, (b) 0.2%, and (c) 0.4 %.
98
1400 ....... In-situ
1300 ___ Classical
1200
E 1100
:::1. CI) 1000 .t! CI)
c 900 1! (!)
800
700
600
~ I~ ~ ~~ ~~ ~ jl'-----~
-.~
k -----500 ...... ~
o 0.2 0.4 0.6
Ti Concentration ( Ti%)
Fig.6.2. Variation of356 Al alloy grain size with the amount of Ti added to melt.
It is possible to monitor such behavior from an analysis of the cooling curve, and more
specifically by studying the primary undercooling portion of the cooling curve as
obtained by both the classical and the proposed techniques shown in Figures 6.3 (a) and
(b). These two figures are in fairly satisfactory agreement with each other. The cooling
curves for the solidifying samples, as obtained by both techniques simultaneously and
under the same melt conditions, exhibit similar behavior and characteristics, including
undercooling, nucleation temperature, recalescence, and apparent time.
99
620~~--------------------'-------------~ - 0% AddedTi
619 +---\\,c--------------i
- 0.2% Added Ti 618 +------\-~~~t____----~
- 0.4% Added Ti 0617+-~~~~~~~~--~~~~~~~~ o
! 616 +---"'\-------~-----"'~---------____l :::s ! 615 +-----'~------~------="'o.::_-------____I CD E 614 +----+-------~.__--_"::-----__l
~ 613 +------\;----------+--------''----____l
612+-----~~--~~~-T----~~-----l
611 +-------~~----~~----~r___l
610+----,---,----,--~~~--~-~~
o 10 20 30 time (sec)
(a) In-situ Probe
40 50 60
620TT.----------------------------,-------------, - 0% AddedTi
619~~-------------~
618 - 0.2% Added Ti
0617 - 0.4% Added Ti o
~616+--~---~~~~------------~ ... ~ ~615+--~-----~--~~---------~ ~
~614+---+------~---~.-----------l ~
~613t_--~---_=~~~~~~~~----__l
612+_---~~~~---~----~~----__l
611 +------------~-----~-----=~~~
610+-----~------~------~~~--~----~~L-----l
o 10 20 30 time (sec)
40
(b) Classical Thermal Analysis
50 60
Figure 6.3. Successive plots for the undercooling portion of the cooling curve refer to the
formation of primary aluminum as it changed with the addition of the titanium
(a) as obtained by the probe; (b) as obtained by the classical technique
100
Figures 6.4 (a), (b), (c), and (d) show how the level oftitanium affects the features of the
primary undercooling which appear as a result of the formation of the a-aluminum phase.
The undercooling decreased with increasing concentrations of grain refiners in the melt,
as a result of an increase in the number of the heterogeneous nuclei. Consequently, the
thermal driving force required is less, and this leads to a lesser amount of undercooling as
measured on the solidification curve. Subsequently, the undercooling will disappear
completely when there is a sufficient number of heterogeneous nuclei in the melt.
The energy released during nucleation strongly affects the apparent time as shown in
Figure 6.4 (b). Under the same cooling conditions, the less the energy that is released, the
faster it is consumed by the cooling effect. In these experiments, the case of 'sufficient'
heterogeneous nuclei in 30 kg of molten metal was obtained by adding 0.4 wt% Ti to the
melt. At this titanium level, the undercooling approached zero. Figure 6.4 (c) shows the
recalescence temperatures as they increase with increasing concentrations of grain
refiner. This is the expected behavior with the highest temperature occurring at the point
of zero undercooling.
101
1.8
1.6
1.4
1.2
0° Cl .5 '0 0.8 0 ~ GI
"CI 0.6 c ::1
0.4
0.2
0
-0.2
18
16
14
Û 12 CD fi) -CI)
10 E ; -c 8 e cu c.. c.. 6 cC
4
2
0
___ In-situ
-.- Classical
I~
~
0
~ ~ ~ T
~ .L
---~ l 0.05 0.1 0.15 0.2 0.25 0.3 0.35 ~
Ti %
a) Primary undercooling vs Ti concentration
0.05 0.1 0.15 0.2 0.25
Ti%
0.3
___ In-situ
-.- Classical
0.35 0.4
b) Apparent time vs Ti concentration
102
O.
0.45
620
(J 619 o l! 618 :::J 1ii 617 ... CIl
~ 616
~ 615 CIl
g 614 CIl u '" CIl li u CIl
0:::
613
612
611
610
620
619
(J 618 o l! 617 :::J
!616 CIl
~615 ~614 c ~613 CIl
.!! 612 u :::J z 611
610
609
1 ~
/X--....
// ....
-In-situ
// -.-Classical
/L -L/ V
o 0.1 0.2 0.3 0.4 0.5 Ti 0/.
c) Recalescence temperature vs Ti concentration
----- -i /~ ~ -In-situ
// // --.- Calssical
// // .y
o 0.1 0.2 0.3 0.4 0.5 Ti %
d) Nucleation temperature vs Ti concentration
Figures 6.4 (a),(b), (c), and (d) Variation of the undercooling, apparent time, recalescence
and nucleation temperatures, as a function of the Ti concentration in the melt.
103
6.3 Eutectic Modification
Two concentrations of strontium in 356 Al melts were used to test the ability of the
proposed technique to assess the level of modification of the melt. These two strontium
concentrations were 70 ppm and 135 ppm, while the master alloy used to add this
element was AI-IO%Sr. The equipment and procedure employed for the se additions were
the same as those used in grain refinement tests. After each addition, samples for
chemical analysis were also taken so as to determine the actual Sr concentration.
In order to carry out the comparison between the proposed technique and the classical
technique, both experiments are set in motion simultaneously under the same melt
conditions. The main difference is that the proposed technique is able to maintain the
same heat extraction level during solidification, while the classical technique displays a
variable cooling effect during the same test period. The reason for this variability is the
formation of an air gap between the solidifying sample and the cup walls during freezing.
Thus, the heat extraction level is at its maximum at the beginning of the process, then it
decreases sharply when the gap is formed. Figures 6.5 (a) and (b) show the eutectic
portions of the cooling curves obtained by both techniques.
104
571
569
567 0 0
! 565 ::::1
~ 563 CD a. E 561 {!!.
559
557
555
571
569
567 0 0
CD 565 ... ::::1 'tG 563 ... CD a.
561 E CD 1-
559
557
555
0 5 10
0 25
-OppmSr -70ppmSr -135ppmSr
15 20 25 30 35 time (sec)
a) In-situ technique.
-0 ppm Sr - 70 ppm Sr -135 ppm Sr
50 75 time (sec)
b) Classical Technique
100 125
40
150
Figure 6.5 Eutectic regions of the cooling curves with various Sr levels (a) as
obtained by the in-situ technique, and (b) as obtained by the classical cup method.
105
When considering the cooling curve obtained using the in-situ technique, the eutectic
temperature decreased from 568°C to about 560°C and the undercooling increased from
1.2°C to 2°C when the strontium leve1 was increased from 0 ppm to 70 ppm. When the
strontium level was raised to 135 ppm, the eutectic temperature increased to 561.7°C and
the undercooling dropped to approximately OAoC (see Figure 6.6).
2.4
2.2
2
1.8
P 1.6 tJ)
.: 1.4 "0 o 1.2 u :;; 1 'tJ :5 0.8
0.6
0.4
0.2
o
.....-/' ~
ff /
o 50
--- --...... • In-situ • Classical
• ~ •
" ~ " '\.
'\. \. \. \.
\. -
\
100 Srppm
\ \" •
150 200
Figure 6.6 Variations in undercooling with varying concentrations of Sr in ppm.
With the addition of 70 ppm of strontium, the eutectic temperature decreased
from 568.5°C to 563°C at 0 level of strontium, when the c1assical cup method was used.
A further increase of strontium to 135 ppm raised the eutectic temperature to 563.9°C.
Satisfactory agreement may be noted when comparing the cooling curves obtained by
means of each of the two techniques; there is a noticeable difference, however, in the
undercooling when the case of 135 ppm strontium is considered. This arises from the
106
difference in the cooling behavior of the two techniques. It should be noted that the new
probe is able to provide an approximately constant cooling effect during solidification.
Moreover, the cooling rate is higher than that provided by the classical technique at the
same eutectic region of the cooling curve, as is clear from the image analysis results for
the silicon particles shown in Figures 6.7 (a) and (b).
The difference in the average area and average length of the silicon decreased with
increasing concentrations of strontium in the melt, considering that the effect of the
modifying agent balances out the difference in the cooling effect. However, because of
the fonnation of an air gap, this effect is not evident at the very beginning of
solidification, since both techniques will then have the 'same' cooling effect. The air gap
becomes the rate controlling factor which controls the heat transfer step and thus
regulates the cooling for the classical technique. Whereas, the fact that the test is
conducted in-situ for the heat pipe probe, the liquid aluminum from the bath flows to fill
any air gap which might have fonned during solidification.
107
20 -'-In-situ
18 -II- Classical
16
4
.. E 2 :::1. ni I!! 0 ni
" CIl I! 8 ~
00( 6
4
2
~
~ ~ ~
~ ~ ~~l ~ ---. -
0 o 50 100 150 200 250
Sr Concentration ppm
a) Average area of the silicon partic1es
10
..... In-situ 9
8 -II- Classical
7
E 6 :::1. oC .. CIl 5 c .!! CD
4 CIl I! ~ 3 00(
2
~
~ ~ ~ ~ ~~ ~~ ~ ~ - -
o o 50 100 150 200 250
Sr Concentration ppm
b) Average length of the silicon partic1es
Figure 6.7 Image analysis results: (a) average area of the silicon partic1es,
(b) average length of the silicon partic1es.
108
In order to verify further the applicability of the probe in detecting the level of
modification for 356 aluminum, the results from the present study were compared with
published results for the same strontium concentration range. The eutectic temperature is
the main feature of the cooling curve which undergoes alteration with the level of
modifier in the melt and it has, consequently, been used as a comparison parameter.
Figure 6.8 shows a comparison between the results obtained by the probe, the classical
cup method, and the results published by Gupta [1]; in the figure, three curves show
similar behavior. The eutectic temperature decreased by about 10°C when the modified
structure was reached, and then it increased slightly when the concentration of Sr
increased in the melt.
572.--------------------------,----------------------• In-situ
570~--------------------------
o o
• Classical
-*"" Ref. (Gupta) ! 568~~------------------------~--------------------~
.a ~ 566+-~~--------------------------------------------~ a. E S 564+-~~~--------~~------------------------------~ u :e s 562+----4~--------~~------------------------------~ ::s w 560+---~~~~~----------------------------~
558+-----~----~----~----~---,-----,----~----~
o 50 100 150 200
Srppm
250 300 350 400
Figure 6.8 Eutectic temperature as it varies with the strontium level in the melt as
recorded by both techniques compared with results published by Gupta [1].
109
Figure 6.9 shows the variation of the dendrite arm spacing (DAS) with the level of
modification. A maximum difference of about 12% in this spacing may be observed
between the two techniques which were applied. Dendrite arm spacing is strongly
affected by the cooling rate, and since the solidification starts off at approximately the
same cooling rate for both techniques, a reasonable difference in DAS was obtained.
Figures 6.10 and 6.11 illustrate the microstructures of the metallographic samples
sectioned from the solidified samples obtained by applying both techniques.
90
80
70
60
fi) 50 cC C 40
30
20
10
o
~-
o
-- -. -. -. -'- - . - .---------1'
-Classical
-.. 'In-situ
20 40 60 80 100 120 140
Sr Concentration ppm
Figure 6.9 Variation ofDAS with levels of Sr content in the melt
110
160
(a) 70 ppm Sr (b) 135 ppm Sr
Figure 6.10 Microstructure at two different levels of modification. Samples obtained by
the proposed technique: (a) contains 70 ppm Sr, and (b) contains 135 ppm Sr.
(a) 70 ppm Sr (modified) (b) 135 ppm Sr (over-modified)
Figure 6.11 Micrographs for two different levels of modification. Samples obtained by
the c1assical technique: (a) contains 70 ppm Sr, and (b) contains 135 ppm Sr.
111
6.4 Detecting and Quantifying Iron Intermetallics
Iron quantification in the molten metal prior to casting is a topic of major interest to
aluminum foundries. Usually, iron exists in the melt in the form of hard brittle platelets
which adhere weakly to the surrounding matrix [2-61. In a metallographic section, the
morphology of the iron in this phase appears acicular or needlelike, and its presence is
detrimental to the mechanical properties of the alloy. Because ofits shape and nature, this
phase is highly inconvenient when it occurs in the cast structure. The platelet shape acts
as a stress raiser, while the brittle nature of these plates leads to machining difficulties,
and provides potential sites for machine tool failure [2,3 and 71.
In practice, it is difficult to get rid of iron-contamination pnor to casting. The
undesirable effects of iron, however, may be minimized by rapid solidification, melt
superheating, and the addition of chemical agents to the melt [3, and 41. In the first of these
cases, rapid solidification, the p-iron phase forms in a greater amount of smaller size
particles which are distributed more homogeneously [21.
The other two techniques endeavor to change the morphology of the iron intermetallic
phase to a Chinese script shape. In the melt superheating technique, the temperature is
increased beyond the norm prior to pouring. Awano and Y oshihiro [71 found that, when
using this technique, the iron phase tended to crystallize in a Chinese script shape rather
than a needlelike shape under high cooling rate conditions.
112
The third technique uses chemical agents such as Mn, Cr, Co, Be, and Mo to react with
iron, silicon, and aluminum to form a Chine se script intermetallic phase, Alts(Mn,Fe)3Sh.
Manganese is frequently used to reduce the effects of iron contamination. The presence
of manganese in the melt in sufficient quantities expands the a-phase region, and also
increases the possibility of a-phase crystallization even at high levels of iron in the melt [
6,7]. The reported ratio of Mn to Fe which is sufficient to ensure the formation of the a
phase rather than the p-phase is 1: 2 [2, 3, and4]. In order to use the addition of manganese
to neutralize the effects of iron in the melt, a measuring technique is required to quantify
the amount of iron present.
A spectrometer is used in large foundries to analyze a sampling of the melt and to obtain
the exact chemical analysis. However, a spectrometer is relatively expensive and is not
available in all foundries, especially medium- and small-size plants. Also, doing this type
of analysis is time-consuming; since the casting process must be held until a physical
sample is extracted and then analyzed by the spectrometer at a laboratory which is
located at sorne distance from the furnace.
The thermal analysis technique has the potential for making an approximate prediction as
to the amount of iron in the melt[2 , 7 , and 8]. Earlier studies carried out by a number of
researchers including Mackay, Narayanan, Tenekedjiev, Samuel, and Gruzleski, show
that thermal analysis can be a reliable method for quantifying the iron content of
al · ·1· all 1· . [29 and 10] ummum-sl lcon oy me ts pnor to castmg " .
113
IRON INTERMETALLICS
Figure 6.12 (a) shows the cooling curve and the associated frrst derivative curve obtained
by perfonning classical thennal analysis, while Figure 6.12 (b) shows the cooling curve
and the associated first derivative curve generated by the new probe. Both test techniques
were carried out simultaneously under the same melt conditions and with 0.93 wt% iron
in the melt. The peak corresponding to the f3-iron intennetallic phase is more clearly
defined and identifiable in the case where the new probe was used. Moreover, the time
needed to generate a cooling curve which provided a useful level of infonnation was
substantially less using the new probe. The time frame required for applying classical
thennal analysis techniques so as to generate approximately the same amount and quality
of infonnation was about three times greater than that required using the new probe. This
may be seen clearly by comparing the time axes in Figures 6.12 (a) and (b). About 550
seconds were required for the sampling cup to generate the cooling curve shown in
Figure 6.12 (a), while about 150 seconds were sufficient for the new technique to
generate the curve shown in Figure 6.12 (b). Although it would have been feasible to
carry out the classical test in less than 550 seconds, this specific time value was used
because it made it possible to start both tests with an initial cooling rate of 0.8°C/s.
This reduction in test time may be ascribed to the differences in thennal behavior
between both the techniques applied. In the classical technique, an air gap fonns on the
heat transfer surface shortly after the sample starts to solidify and causes high thennal
114
resistance in the direction of the heat flow. When using the new probe, however, fresh
material from the surrounding molten environment flows in to fill any shrinkage induced
by solidification. The interface thermal resistance between the sample and the probe wall
is thereby reduced significantly, and there is a minimal decrease in the heat transfer
across the probe wall. Thus, there will be very little modification in the rate of the heat
extracted from the solidified sample during the solidification process.
700
680
660
640
~ 620
e ! 600 !. ~ 580
560
540
520
500
1Ilr------------------------------------r~Ccd~~3 cu",,"
0
Firsl derivative
2
Iron Intennetallic Peak
-2
L-----~------~------~--------~------~--~~_4-3
100 200 300
time (sec)
400
(a) Classical thermal analysis
115
500 600
Fig. 6.12 Cooling curves and associated first derivative curves for 356 Al with 0.93 wt%
Cu as obtained by (a) c1assical sampling cup, and (b) new in-situ probe
Figure 6.13 shows the cooling curves obtained by the new probe for the different iron
concentrations used in this study. The arrows show the location of the inflection points
which appear on the cooling curves as a result of the fonnation of the (3-iron intennetallic
phase. The higher the concentration of iron in the melt, the more evident the inflection
point is. This is due to the quantity of latent heat released when this solid intennetallic
phase is fonned. Higher concentrations of iron in the melt lead to the fonnation of a
greater quantity of intennetallic AlsFeSi, and consequently to greater heat release.
116
620
-0.15%Fe 610
-0.93%Fe 1.56%Fe 1.3%Fe
600 -1.14%Fe 1.36%Fe
-1.3%Fe 590
,",,~ 1.36%Fe
cP 580 -1.56%Fe ! = ~ 570 • CL
E ~560
550
540
530
520 20 30 40 50 60 70 80 90 100
time (sec)
Fig. 6.13 Cooling curves for difIerent iron concentrations in 356 Al alloy.
The position of the inflection point on the cooling curve which corresponds to the iron
intennetallic phase was observed to be dependent on the concentration of iron in the melt
as shown in Figure 6.13. It will be noted that the higher the iron concentration, the higher
the temperature at which the intennetallic phase fonns. Figure 6.14 shows the
temperatures at which the p-iron phase fonns, and how they vary according to the
concentrations. These tests were repeated three times using the new in-situ probe and
reheating the same solidified sample, whereas the classical thennal analysis was
perfonned only once for each iron concentration. The results were reproducible with an
acceptable standard deviation of about 1.2°C on average.
117
604
602
600
cP 598 e .a 596 I! !. E 594 ~ g 592 ;; ~ 590 o u..
~ 588 rt. ...Jl' oC( 586
584
582
580 0.6 0.8
/. .. ' 1
Y = 23.932x + 564.83 t // ! ~=0.9517
/Y /Y
.// //.
// t 1/
1.2
Concentration of Fe wt%
1.4
y = 24.27x + 563.21 R2 = 0.93
• New Probe
• Classical Tech .
1.6
Figure 6.14 Fonnation temperature of A1sFeSi at different iron concentrations
1.8
Figure 6.15 shows the first derivative curves corresponding to that part of the cooling
curve, as produced by the probe, where the fonnation of the iron intennetallic phase
occurs for the different iron concentrations studied. A curve-fitting technique was used to
obtain an equation for each curve which presents the data in a time zone ranging from
two seconds prior to the inflection point to two seconds after it. The beginning and end of
the phase fonnation were detennined from the corresponding second derivative curves.
Figure 6.16 shows the cooling curve for 356 aluminum alloy with O.93wt% Fe together
with the second derivative of this curve. The start and the end of the reaction were
detennined from the boundaries of the inverted peak on the second derivative curve.
118
0.1
1.3%Fe -0.3
-0.7
Û
J -1.1 ~ al -1.5 .~ .. . ~ -1.9 al c , .. \ r! -2.3 i! \ \
-2.7 , \
-3.1
-3.5 -.. , 20 20.8 21.6 22.4 23.2 24 24.8 25.6 26.4 27.2 28 28.8 29.6
time (sec)
Fig. 6.15 Portions of the tirst derivative curves from in-situ probe corresponding to the
iron intermeta1lic phase.
620r-~==========~--------------------------------------~
600+-----------------------~----4---------------------------------_+
Forming Temperature
580+---------------------------r_~~r_----------------------------_+
~ portion corresponding 10 the Iron f Intermetallic ::s !OOO~------------------~--~~*=~----~~====~-=------------+ !. E ~ MO~~~~~~~~
520+-----~L---~--------------~~~r_------------------------~--_+
2
1.5
0.5
0
-0.5
500 ·1 o 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120
tlme (sec)
.. ~ .!! 0 0
• > = • > '1:
~ 'C c 0 ~ (1)
Figure 6.16 Cooling curve and associated second derivative curve for 356 Al alloy with
0.93 wt%Fe
119
The areas below the iron intermetallic peaks on the first derivative curves shown in
Figure 6.15 were calculated and are presented in Figure 6.17. Samples from the solidified
materials, previously frozen and obtained by both the techniques under discussion, were
poli shed and analyzed by electron probe microanalysis (EPMA). The average surface
fractions of the AlsFeSi intermetallic phase were also measured and the results are
presented in Figure 6.17.
8,-----------------------r---------~~----------~--------__.8
+Area below Fe intermetallic peak .Average surface fraction
~ 7+---------------------------------------------~~--------__+7
i y=5.73x-1.94 R2 = 0.99 ~
~ 6+----------------------------------=~~~--~~~--------__+6e i ~ - y=4.69x-1.18 ..J4' E R
2=0.97 5 ~ .! 5 c
~ 0 CI) ~ = • ... 4 4.t;:
~ B § ~ f 3 3 i • & CD I! ë 2 2 ~ l ~
o+---------~--------~--------~--------~--------,_------__+o
0.6 0.8 1.2 1.4 1.6 1.8
Concentration of Fe in the melt (wt. %)
Figure 6.17 . Average surface fraction and areas below the intermetallic peaks as they
vary with the iron concentrations in the melt
120
Backscattered electron images for the samples with different iron concentrations are
shown in Figure 6.18. These images show the intensity of the p- iron intermetallic phase
as this intensity increased with the iron concentration in the melt. Also, the fragmentation
of this phase becomes more frequent when the iron concentration is increased. The
images also visually correlate the surface area of the p- iron intermetallic phase with the
concentration of iron in the melt.
121
(a) 0.123%Fe (b) 0.932%Fe
(c) 1.147%Fe (d) 1.302 %Fe
(e)1.357 %Fe (f)1.566%Fe
Figure 6.18 Backscatiered images of the samples extracted using the new technique at
five concentrations ofiron: (a) 0.93% Fe, (b) 1. 15%Fe, (c) 1.30% Fe, (d) 1.35% Fe, and
(e) 1.56 %Fe.
122
6.5 Detecting and Quantifying Copper Intermetallics
Al-Si binary alloy (see Table 5.1 for composition) was used as a base alloy to test the
sensitivity of the new probe in detecting an intermetallic copper reaction. The
concentration of copper in the melt was increased by adding pure copper pieces to it. The
same procedure as the one detailed in the previous section was followed in order to
analyze the cooling curves for the copper intermetallic phase. Figure 6.19 shows the
cooling curve and the first derivative curve obtained by both the new probe and the
classical technique for Al-Si alloy with 3.7 wt% copper.
The cooling curve for this melt displays the following reactions:
(i) formation ofprimary aluminum dendrites at about 603°C;
(ii) the main silicon eutectic reaction which occurs at 566°C;
(iii) a reaction at 500°C in which AhCu forms through the reaction
L ~ AI+ AhCu+ Si.
Figure 6.20 shows the relationship between copper concentrations in the Al-Si melt and
the area below the corresponding copper intermetallic peak, which appears on the first
derivative of the cooling curve. The trend of this relationship is linear; the area below the
peak increases with increasing copper concentrations in the melt. The area measurements
from the first derivative curves are reported for the in-situ data only. Because of the noise
in the temperature data obtained when using the classical cup technique, the results from
comparable area measurements show only a slight trend which is associated with
123
substantial scatter and are not presented here. The conclusion may thus be drawn that the
proposed analysis which correlates area with concentration is not convenient for use with
the classical cup technique, although it is applicable to the new in-situ heat pipe probe, as
is shown here.
650
600
0 "I! 550 ~ I! . ra. ~ 500 ...
450
400
40 60 80
640
~ 620
600
~ ~
fl580
1580 F
5040
520
SOO
460 0 100
100 120
Ilmelnc)
140
(a) In-situ probe
160 160
Copper Intermetallic peak
L.L... .• J ,d \..
1
200
~T ' ... --..-.---
~ ~
300 tlmela)
400
(b) Classical CUp
~
soo
3.7l\Cu
0.5
ô • -0.5 ~
'-• -1 ~ · >
~ -1.5 ..
~ ... -2
-2.5
-3
200
3.7" Cu
0.5
o
..... .--~ -0.5
~
-1
-1.5
-2
-2.5
-3 600
Figure 6.19 Cooling curves and associated fIfst derivative curves for Al-Si alloy with 3.7
wt«>.10 Cu obtained by (a) new probe, and (b) classical sampling cup
124
9",,------------------------~1.40
8 • Area below the copper peak • Average surface fraction "-+----------------------±-----+J 1.20 ~
~ 7+------------------~~~~-~ ~ !. ~ y=196x+0.46 1.00 ~ c. CG 6 R' =0.96 ....
8 !. 3 ! .~ 5 +----------~-----'&o"7''_:f'1_-------____I. 0.80-i! _ - y=0.35x-0.1 ~
~ .; 4 R'=0.96:
Qi Ë 0.60 ~ .Q CD 3 • CG ë > CD ._ 0.40 cC
~ 2 ~ 1 +----~~-----------------~ 0.20 ;(
o -!---.,-------,-----;-----r-----,---.,.---..,.-----!- 0.00
o 0.5 1 1.5 2 2.5 3 3.5 4
Cu Concentration wt%
Fig. 6.20 Average surface fraction and area below the intennetallic peak as they vary
with the copper concentration in the melt
Figure 6.20 also shows a comparison between the area below the copper intennetallic
peak and the average surface fraction of the copper intennetallic phase, as measured by
an Electron Bearn Probe (or JEOL). The results show satisfactory agreement. Both curves
display a linear trend, although the two lines tend to converge when the concentration of
copper increases. The slope of the line which represents the area below the curve is less
than the slope of the line which represents the average surface area of the intennetallic
copper. It has been postulated that this effect may be due to a limited fonn of chromium
contamination which took place upon the dissolution of a portion of the stirring ladle in
the melt. At the time the tests were conducted, these researchers were not aware of the
125
possibility ofthis contamination by Cr. Nevertheless, the results are revealing and worthy
of consideration. With the presence of dissolved Cr, another phase fonned during
solidification and a peak corresponding to this phase appeared just prior to the copper
intennetallic peak. The area below this peak was excluded from the calculations for the
area below the copper intennetallic peak (AhCu), as seen in Figure 6.21.
700,-----------------------------------------------------,
0.5 650
A12Cu
A118C~Silea(MnNiCr) 0 û • ~600
.!! Co) 0
CI> .. GO .. - -0.5 ~ .. .. ..
CI> > D- .;: E • .!550 'a -..
-1 . ~ u..
500 -1.5
~~----r_----r_----r_----r_----r_----,_----,_----,_--~-2
o 20 40 60 80 100 120 140 160 180
time( sec)
Figure 6.21 Cooling curve and first derivative curve for the Al-Si alloy
with 2.81 wt% Cu.
126
The backscattered images for the samplings of different copper concentrations are
presented in Figure 6.22. The images show larger amounts of the copper intermetallic
phase as the concentration of copper increased in the samples.
0.13 wt% Cu 0.78wt% Cu
1.3 wt% Cu 1.79 wt% Cu
127
2.24 wt% Cu 2.8 wt% Cu
3.53 wt%Cu
Figure 6.22 Backscattered electron images of Al-Si alloy with varying copper
concentrations
128
6.6 Detecting of Magnesium Intermetallics
A base alloy of AI-Si-Cu-Fe was used to test the sensitivity of the new probe to detecting
the AIsMgsCu2Si6 intermetallic reaction. The copper concentration in the melt was fixed
at about 3.5 wt% while the magnesium concentration was increased incrementally by
adding pure magnesium pieces to the melt. The magne sium concentrations were 0.2 wt%,
0.3 wt%, 0.4 wt%, and 0.6 wt%. Samples were taken from the melt and chemical analysis
was performed. Samples were also sectioned from the solidified samples obtained using
both the probe and the classical cup. Figures 6.23 and 6.24 show the cooling curves and
the corresponding first derivatives obtained using both techniques.
The cooling curves are characterized by the following reactions:
1) the formation of primary aluminum at 603°C;
2) a primary AlsFeSi formation at 579°C;
3) the main silicon eutectic reaction 563.5°C;
4) a reaction at approximately 500°C in which AhCu, AlsFeSi, and Si form;
129
?
i •
0.5
o 0
.; p • ~
-0.5 • N-------------------------~~~--------~ j ..
: -1 ...
-1.5
400 +--""T"""""-___r_----,-------,-----,--~-_,__-_r_-___,__:'--___I_ -2 50 100 150 200 300 350 400 450 500
(a) Cooling curve obtained by the probe.
6OO~~~----------------------------------------------~
3
550~~------------~---~=_----------~~------------~
~500~--------------------------------~~------------~
{!.
~~-----,-----""T"""""---_r_---_r_---___r_---~--1
o 100 200 300
time(sec)
500
(b) Cooling curve obtained by the classical cup method.
600
0.5
0 ô
~ -~
-0.5 li >
! i!
-1 iL
-1.5
-2
Fig. 6.23 Cooling curve for Al-Si-Cu temary alloy with 0.2 wt% Mg obtained by the
(a) probe, and (b) classical cup.
130
The heat effect associated with the formation of primary AlsFeSi is evident only at slow
cooling rates. Figure 6.24 shows the cooling curve for AI-Si-Cu-Fe with 0.2 wt% Mg at a
higher cooling rate than that shown in Figure 6.23. It should be noted that the peak
corresponding to the formation of the AlsFeSi intermetallic phase has vanished in Figure
6.24.
cP !
55O~~-L------------------------------------~-4------~
t500~--------------------------------------~~~-----!
o 50 100 150 200 250 300 350
tlme(sec)
(b) Cooling curve obtained using the classical cup method.
0
-1 f ~
-21 i
-3 1! iL
-4
-5
Figure 6.24 Cooling curve for Al-Si-Cu ternary alloy with 0.2 wt% Mg obtained by the
probe.
131
It will be observed that the peak corresponding to the magnesium intermetallic reaction,
i.e. the formation of AlsMgsCu2Si6 phase, has changed its position with the increase in the
Mg concentration in the melt. In the case where the 0.2wt% magne sium concentration
was analyzed, the peak corresponding to the magnesium intermetallic phase did not
appear as a significant peak immediately following the copper intermetallic phase
however thermal anomaly appears at approximately 476°C. When the magnesium
concentration in the melt was increased to 0.3 wt%, two thermal anomalies appeared, one
just before the AhCu intermetallic peak at 505°C and the other just after it at
approximately 585°C. When 0.6 wt% magnesium concentration melt was analyzed, the
peak appeared just before the final reaction peak, i.e. before the formation of AhCu, at
approximately 505°C. Figures 6.25 and 6.26 show the cooling curve and the associated
frrst derivative for the melt with 0.3 wt% and 0.6 wt«'/o magnesium, respectively. Both the
thermal anomaly and the peak corresponding to the formation of AIsMgsCu2Si6 phase are
denoted by number 5 on Figures 6.25 and 6.26.
MO+-~=-----~------------------------------------------
0.5
550 +-----JY\-.J~:-:-------------.i~~'""ç"_----_\_--------_/_----------____:J 0 1 p P I .. .f...----.l...-~~~~~ -0.51
F ! -1 IL
450~--------~~----------------------------~~"~--~
-1.5
400 +L-------~--------~------_.__--------,.___------~------___+ -2
o 50 100 150
tlme(sec )
200
(a) Cooling curve obtained by the probe.
132
250 300
600
3 4 0.5
0 l p
550
P •
~ -0.5 !
!
~ f500
~
-1 1! iL:
450
-1.5
400~------~--------~------~--------~------~--------~--~ -2 o 100 200 300 400 500 600
t1me (sec)
(b) Cooling curve obtained using the classical cup method.
Fig. 6.25 Cooling curve for Al-Si-Cu temary alloy with 0.3 wt% Mg obtained using
(a) probe, and (b) classical cup method.
610~--------------------------------------------~4----------~
0.5 590+--A--------~--------_1--------------------~----------~
o
490+------------------------+---r----~~==----------------~-2
470 ~------------------------+_+_----------------~--------__t -2.5
450+-----~----~----~----~------~----~----~----~---=~-3
o 20 40 60 80 100 120 140 160 180
tlme(.ec)
(a) Cooling curve obtained using the probe method.
133
610
590
570
o De 550
.a (530
i .. 510
490
470
450
\ \
o
.-; 4 -........
ï ~ ) \ /
0.5 f, ~~ ~ " ." - ~ ----...-'-./ '-
~ ~ ~ -1.5
~ -2
100 200 300 400 500 600
tlme (sec)
b) Cooling curve obtained using the c1assical cup method.
Figure 6.26 Cooling curve for Al-Si-Cu temary alloy with 0.6 wt% Mg obtained using
(a) probe, and b) classical cup method.
6.7 Detecting Liquidus Temperatures
It will be recalled that the main reason for not using c1assical water-based heat pipes in a
molten aluminum environment is the temperature factor. Since the probe is basically a
heat transfer device and its applicability depends on the working temperature range and
heat flux regardless of the nature of the medium, it was decided to push the probe to its
limit by raising the temperature of the working environment.
A 32-kg melt of 390 alloy was prepared according to the chemical composition provided
in Table 5.1, and a superheat of 150°C was maintained. Thermal analysis tests were
carried out using both techniques simultaneously. Then the concentration of silicon in the
134
melt was changed to 20%, 15%, 13%, and 11%. The cooling curves produced by both
techniques were similar under conditions where the liquidus temperature reached a
minimum value of about 580°C, at which point the Si concentration was situated
approximately at the eutectic point (12.6 wt %). A maximum value was reached when the
Si concentration increased to 20 wt% which was the maximum concentration investigated
in this study. The melt temperature was maintained at over 850°C throughout.
Figure 6.27 shows how the liquidus temperature varies with the concentration of Si in the
melt, as predicted by both of the thermal analysis techniques. These results are compared
with the liquidus temperature as obtained from the aluminum-silicon phase diagram. It is
important to mention here that 390 alloy contains elements, other than Al and Si, which
might affect the value of the liquidus temperature, since this alloy also contains 4 wt%
copper as weIl as traces of iron and magnesium. This fact may account for the difference
which arises between the liquidus temperatures predicted by both thermal analysis
techniques and the one obtained from the phase diagram.
135
710
690
670
0 . f 650 .a I!
~630 {! • ::1
;g 610 0 II)
590
570
550 10 11 12 13 14 15 16 17 18
SI Concentration
19
- • - Phase diagram
• In-situ " Classical
20 21 22
Fig. 6.27 Variation ofliquidus temperature (390 Al-Si alloy) as Si concentrations
increased in the melt, compared with the solidus temperature from Al-Si phase diagram.
6.8 Detecting Minor Reactions in Wrought Alloy 6063
After the encouragmg performance of the new probe in producing more detailed
information on the cooling curves for sorne of the cast alloys investigated in this study, it
was decided to test the sensitivity of the probe in detecting minor reactions associated
with the solidification of wrought alloys. The fact that such alloys tend to contain
relatively low levels of alloying elements (see Table 5.1) influenced our decision in this
regard.
136
The 6063 wrought alloy, which is an alloy of the 6000 series, was selected as the base
alloy for these tests. Wrought alloys of this series are usually used in the automotive
industry, or the manufacture of furniture and for structural appliances, where corrosion
resistance and strength are essential [12].
Sorne of the minor reactions associated with the solidification of 6063 aluminum alloy
cannot be detected by classical thermal analysis [12], because the heat associated with the
formation of a certain phase during solidification is insufficient for recording purposes,
and hence is not visible on the cooling curve.
The reactions which take place during the solidification of 6063 alloy as stated by
Backerud [12] are the following:
(i) formation of a dendritic network of aluminum at 655°C;
(ii) formation of AIsFe2Si intermetallic phase at 618°C;
(iii) formation of AlsFeSi intermetallic phase at 612°C;
(iv) formation of Mg2Si phase at 576°C as predicted from the phase
diagram.
In this study, the effects of adding Sr, Mn, and Fe to the 6063 melt were investigated.
Thermal analysis tests were carried out in an electric furnace using 32 kg of 6063 melt.
For this purpose, 200 ppm of Sr were added to the melt and the test was performed using
both the new probe and the classical cup method. Then 0.5 wt% of Fe was added to the
137
melt and the thermal analysis was effectuated. Finally, 0.7 wt% of Mn was added to the
melt.
Figure 6.28 shows the cooling curves and their derivatives for 6063 aUoy obtained by the
probe at two different average cooling rates which are 1.7°C/s and 0.6°C/s and the cooling
curve obtained by the classical cup. The two curves produced in-situ and that produced
by the classical technique were able to detect the formation of a dendritic network of
aluminum at 653°C, 652.9°C, and 651.4°C, respectively. The appearance of a plateau on
the first derivative curve of these curves, which is denoted by i, is evidence of this
reaction. Peak iii which appears on the first derivative of the cooling curves generated by
the probe refers to the formation of Al5FeSi intermetallic. When the first derivative curve
is superimposed on the cooling curve, the temperatures at which Al5FeSi intermetallic
formed were found to be 628°C, and 595.4°C when 1. 7°C/s, 0.7°C/s when average
cooling rate were applied, respectively. In the case where classical thermal analysis was
used a tiny anomaly, namely Band pointed by an arrow on Figure 6.28 (c), appeared on
the first derivative curve. The temperature corresponding to this feature as obtained from
the cooling curve was 596.3°C. The final reaction which forms Mg2Si phase was detected
by both techniques; peak iv which appears on all the first derivatives of the cooling
curves produced by both techniques was used to set the formation temperature of M~Si
phase on the cooling curve. The formation temperature of this phase was found from the
classical thermal analysis technique to be 528.6°C, which is similar to the formation
temperature obtained by the probe (529.2°C) when 0.7°C/sec cooling rate was considered.
While in the case when 1.7°C/s cooling rate was used, peak iv was corresponding to a
138
forming temperature of 585.6°C. This difference in the forming temperature, when
comparing the two curves generated by the in-situ probe, is due to the difference in the
instantaneous cooling rate prior to the formation of the phase which is more than the five
times in the first case (lO°C/s vs 2°C/s).
750
700
650
cP f '" ë600 !. E • ....
550
500
450
700
650
S' 600 !
t ! 550
500
450
1 - Cooling Curve - First Derivative 0
-2
i .!!! cP
--4 • +-----~----------------------------~~--------------~ i
.!l -6 i!
~--------------------------------~;-~----~c-------~ ~
~-----------------------------------tlr---------------~ -6
+-------~------~------~------~--~L-~------~----__+-10
0 20 40 60 80 100 120 140
dm. Ieee)
(a) Cooling curve obtained by the probe (1.7°C/s)
- Cooling Curve
- First Derivative 0.5
-2
+-----r-----.,-----.,-----.,-----..----~---__+ -2.5
0 50 100 150 200 250 300 350
tlme(.ec)
(b) Cooling curve obtained bOy the probe (O.7°C/s)
139
700~----------------------------------------~--------~
650
1
~ 600 ! .:1 I! • Go
~ 550
500
- cooIing Curve
- First Derivative. 0.5
o i ~ -~
-0.5 i
i -1 ~
-1.5
450~------~------~--------~------~------~----~--~-2
o 100 200 300
time (sec)
400 500
(c) Cooling curve obtained by the classical cup method.
600
Fig. 6.28 Cooling curves for 6063 alloy obtained by both techniques (a), (b) for the new
probe, and (c) for the classical cup method.
Figure 6.29 shows the cooling curve and their derivatives for 6063 alloy with 200ppm Sr
obtained by the probe and those obtained by the classical cup. The two curves behave
similarly and it looks like the addition of Sr to the melt has no significant effect on the
features appearing on the cooling curves, generated by both techniques, and the
associated first derivative curves. Although it was noticed that the peak corresponding to
the formation of the AlsFeSi intermetallic phase became smaller for the case of the in-
situ probe when compared to the peak corresponding to the same reaction in the absence
of Sr in the melt. Aiso the same remark was noticed in the case of classical thermal
analysis, the tiny anomaly that appeared on the first derivative curve vanished completely
when adding the Sr to the melt.
140
7oo.-------------------------~----------------------------~
o ~0t-----~4W~~~~==~~========~~--------------------_J
-1
-2 •
P600+-~~----------------------------~~-+------------~~~- P ! -~ ~i !550+-----------------------------------+-----~~~P-------~ ~ ~
-5 ~O+-----------------------------------~~---------+------~
-6
450+-------~----~------~------~------~----~------~--~~
o 20 40 60 60
tlma(sec)
100 120
(a) Cooling curve obtained using the new probe
140
700.-----------------------------------------------------~
650~==~--~--------------------------------------------~ 0.5
0
! li:
J -~.5 j -1
~O~--------------------------------------~~----------~ -1.5
450~------~--------~--------~------~--------~----~~ -2
o 100 200 300
tlm.(seo)
400 ~o 600
(b) Cooling curve obtained by the classical cup.
Fig. 6.29 Cooling curves of 6063 alloy with 200 ppm Sr obtained by both
techniques; (a) new probe and (b) classical cup.
! li:
Figure 6.30 shows the cooling curves and their derivatives for 6063 alloy after
adding 0.5 wt% iron to the melt. Two different average cooling rates of 1.3°C/s and
141
O.6°C/s were used with the probe and these are compared to the cooling curve obtained
by the c1assical cup. The cooling curves obtained by the in-situ probe detected three
reactions which are the formation of aluminum dendritic network (i), AlsFeSi
intermetallic phase (iii), and Mg2Si phase which indicates the end of the solidification.
The formation temperature of the AlsFeSi intermetallic phase was found to be altered
from 595°C to 628°C. Another small anomaly appeared on the first derivative of the
cooling curves, however it was difficult to identify these peaks due to the difficulty in
analyzing the tiny partic1es that appeared in the microstructure. The size of these
partic1es, which is in the order of 1 ~m, is less than the diameter of the electron beam used
to analyze the samples which is 2~m.
700~--------------------------------------------------~
650 -l------------==========F-~----------------------__l 0.5
111
cP 600 I! :::1
! & ! 550
500 -1.5
450 .f.-------.------.,------,.------r--------.--------j.-2
o 50 100 150
time(lec)
200 250
a) Cooling curve obtained by the new probe (O.6°C/s)
142
300
700~---·_------------_·_---------~
0.5
-0.5
~ 600 +-+----lt------------------+--I--\---------Jl~ -1.5 ! i i f ~~ ~~ ~
! -3.5 iL
500~--------------------~~~~~--~
-<4.5
450 +---,----r------,----,-------.------,-----,-----,----~'---____+ -5.5
o 20 40 60 80 100 120 140 160 180 200
time (sec:1
b) Cooling curve obtained by the new probe (1.3°C/s)
700~------------------------~
S'600 !
t E {!!.~
500
450
0.8
~----"=::--------_;_::'------------__+ 0.6
~L---~----__,__---__ ---_----,---~-1 0 100 200 300
tlme(sec:)
(c) Cooling curve obtained by the classical cup.
500
Fig. 6.30 Cooling curves of 6063 alloy obtained by both techniques with 0.5 wt % added
Fe; (a), (b) new probe, and (c) classical cup method.
143
Figure 6.31 shows the effect of adding 0.7wflo manganese to the preceding melt (6063
Al + 0.5 wt% Fe) on the cooling curves and their derivatives obtained by the probe at the
two different average cooling rates of 1.3°C/s and O.3°C/s, along with the cooling curve
obtained by the classical cup. The comparison between these curves shows that the in-situ
probe is able to detect more phases than the classical cup. AH four reactions associated
with the solidification of 6063 and reported in the literature are detected by the in-situ
probe [11]. On the other hand, the classical technique was not able to detect the formation
of Piron (AlsFeSi) phase during the solidification of the sample. Microscopic study,
however, provided proof of the presence of Piron phase in the sample. In Figure 6.32 (d),
Piron is signaled by ii in this figure.
700=-~~~~~~~~~~~~~~~~~~~~~~~--,
0.5
650 -1--~t-------;I;-____ ~::::=~=======~~~~~~~~~-+ -0.5
-1.5 li p~ l I!
t ~t ~ 550 +-~~~~~~~~~~~~~~~~~~t--t-~~~~~-r -3.5 -3
! ü:
-4.5
500+-~~~~~~~~~~~~~~~~~~~~~~~~
-5.5
450 +-~~~~~~~~~~~~~~~~~~~~~~~.---"- -6.5
o 20 40 60 80 100 120 140 160 180 200
t1me (sec,)
(a) Cooling curve obtained using the new probe.
144
700 ---... ------.... --.--.-.. -------.---.--------.-.. ----.. ----.. --.-------------.---- 0.5 1
0.3
650 +----=t=F.=;=====1~=-_=:::_____\-------------___l_ 0.1
500 -l+1-4-----------------'--'--f'---~-----__!_ -1.1
-1.3
450 +----,-----,-----,-----r----~~--~---===_I_ -1.5
o 100 200 300 400 500 600 700
tlme (sec.)
(b) Cooling curve obtained by the new probe (slow cooling rate).
750-----------------···------·---------- 0.2
1 0.1
700~-~L------------------------~
o
650 -JI-I--è*=......:I11111:--=,-------,r--------------------~ ~.1 j P • ! ~ t '! 600 #-t----lII.Wt.it--------=:-:,..~------iv-------_t ~.2 i ~ ~ ~ . ~.3 ~ 550~+--------~ .. ~----~~-+--------~ !
ii:
!III ~_~::-:-----~~~ ~.4 500lll-
~.5
450 .J-I-'-----~----_,_----,__---__,_----~---~ ~.6
o 100 200 300 400 500 600
tlme (sec.)
(c) Cooling curve obtained using the classical cup method.
Figure 6.31 Cooling curves for 6063 alloy obtained using both techniques with 0.7 %
Mn; (a) and (b) newprobe; and (c) classical cup method
145
In order to identify the reactions associated with the solidification of this alloy, samples
were sectioned from the ones frozen using both the new probe and the classical cup
method. They were then examined using the electron beam probe technique.
Metallographic analysis was carried out on the samples; then, backscattered images of the
polished samples were obtained and may be seen in Figure 6.32. An energy dispersive
spectroscopic analysis of the samples (EDS) was made so as to identify the phases
present in the microstructure. A quantitative analysis was also performed using an
electron probe microanalyzer (EPMA) , and in the case of the base alloy, two phases
appear as seen in the backscattered image of Figure 6.32 (a). Table 6.1 gives the
composition of both these phases.
Phase
(1)
(2)
Approximate Composition (atomic %)
Stoichiometry Al Fe Si Mg Mn
AhoFe2Sh 66.9 12.8 18.6 0.12 0.5
AIsShMg 57 0.8 26.4 12 0.3
Table 6.1 Chemical analysis of phases 1 and 2 which appear on the backscattered
image of the base alloy (6063).
Figure 6.32 (b) shows the backscattered image for the sample with 200 ppm Sr.
Two phases appear in the image in addition to the strontium (bright white-colored points
in the image). Table 6.2 shows the chemical analysis for both these phases as weIl.
146
Approximate Composition (atomic %)
Phase Stoichiometry Al Fe Si Mg Mn Sr Cu
(1) Ahs(FeMn)JS4 70 lIA 17.2 0.08 0.7 0.01 0.07
Ah IMg2(FeCuMn) (2) 55.9 1.6 26 10.8 0.6 0.3 2.7
Sis
Table 6.2 Chemical analysis of phases 1 and 2 which appear on the backscattered
image for base alloy (6063) containing 200 ppm of strontium
For the case where 0.5 wt% of iron was added to the melt, Figure 6.32 (c) shows the
backscattered image, in which only B-phase iron could be distinguished, whereas all the
other phases were not significant enough to be detected by the electron beam. In view of
the fact that the beam diameter is 3 microns and the size of these particles is less than
that, the beam will therefore cover part of the matrix in addition to the particle under
investigation.
Figure 6.32 (d) shows the effect ofadding 0.7 wt% of Mn to the melt. Three phases were
identified, the most common of which is that of a-iron. Table 6.3 shows the chemical
analysis of these phases.
Approximate Composition ( atomic % ) Phase
Stoichiometry Al Fe Si Mg Mn Sr Cu
(1) AhsFes Mn3 Sh 61 20 8.1 0 10.5 0.02 0.34
(2) AIsShM~(FeCuMn) 56 1.63 26.7 10.5 0.56 0.253 4041
(3) Ah(FeMn)Si 73 lIA 9.22 0 0.29 0.008 0.2
Table 6.3. Chemical analysis ofphases 1,2 and 3 which appear in the backscattered
image ofbase alloy (6063) containing 0.7 wt% Mn.
147
(a) Backscattered image for 6063 alloy
(b) Backscattered image for 6063 alloy with 200 ppm Sr
(c) Backscattered image for 6063 alloy with 0.5 wt% added Fe
148
(d) Backscattered image of 6063 alloy with 0.7 wt% Mn
Figure 6.32 Backscattered images for (a) 6063, (b) 6063 with 200 ppm Sr, (c) 6063 with
0.5 wt% added Fe, (d) 6063 with 0.7 wt % Mn
6.8 Effects of Boundary Conditions on the Quality of the
Results Produced by the Probe
In order to understand the effect of boundary conditions on the quality of the data
provided by the new probe, thermal analysis tests were carried out under two different
boundary conditions. The first test involved operating the probe in-situ, inside the molten
metal which was poured into the evaporator section in a procedure similar to the one
followed when applying c1assical thermal analysis techniques.
In the tirst case, the probe extracts heat from the solidifying sample while heat is
transferred from the molten bath to the sample, affecting the radial temperature gradient
inside it. In the second case, where the thermal analysis test is performed outside the
149
crucible, the heat is transferred to the pipe from the sides of the sample, while a small
amount is transferred through the lower surface of the sample, due to the thermal
resistance created by the insulated plate installed there.
A 319 aluminum alloy was used in this part of the study because of its chemical
composition. The cooling curve for this alloy and the associated first derivative curve
show features corresponding to the formation of intermetallic copper and iron.
Monitoring the changes in these peaks when the probe is exposed to different boundary
conditions may exp Iain the enhancement in the quality of the signals generated by the
probe when operating in-situ.
Six thermocouples were used to measure the temperature at different locations in the
sample during both freezing and cooling processes. Figure 6.33 shows the distances from
the probe surface at which each thermocouple is placed. Thermocouple 1 is placed at the
center of the sample while thermocouple 5 is placed at a 5-mm distance from the probe
wall. Thermocouples 2,3, and 4 were placed between thermocouples 1 and 5 at distances
of lem from each other. Thermocouple 0 was placed on the same plane but on the
opposite side at a distance of 1 cm from thermocouple 1 matching the distance between
thermocouple 2 and thermocouple 1.
150
+._+._.+_.+._+._._._._._._._._._ .. _._._._._._._.-1
321 0
Figure 6.33 Thermocouple locations in the solidifying sample
The cooling curves detected by the six thermocouples and generated by the probe when
operating in-situ and outside the crucible are presented in Figure 6.34 (a) and (b), while
Figure 6.35 (a), and (b) show the cooling curves and the associated tirst derivative curves
detected by thermocouple 1 in both cases when the probe operating in-situ and operating
as a cooled classical cup.
700 -.-------_ .. __ ._-_._-----_._-------_ ...... --............ _ ... _-_ ...... - ...... _ .... --_ .. _ ... _ ... _------_ ...... --_ ........ _--... _--~~~ -Canter(Q)
-Center(1) -Center(2)
650 +----"~-----------------------1 Center(3) -Center(4) -Center(5)
6OO~---~~~-=~--~~---------------~
500+-------------~~~~~---------~
~+--------------------~~=-----~
~+--~--~--~--~-~--~--~--~-~--~
o 20 40 60 80 100
t1me(sec)
120 140
(a) molten metal environment (in-situ)
151
180 180 200
700
650
600
P I!
~ 550 8. E ~
500
450
400 0
--.------.-... -------------.. - ... -.-.-.. --....... ------.... -.-.-.. --... -.. -....... -..... ---.......... -r----,
20 40 60 80 100
tlme (sec)
120
(b) room environment
-CentIll(O)
Center(2)
-Center(l)
Center(3)
-CenIel(4)
-CenIel(5)
140 160 180 200
Figure 6.34 Cooling curves for 319 a1loy at different locations in the fcozen sample. (a) the probe operating in-situ, and (b) the probe operating outside the crucible.
700
650
600
P I! :s f 550 & ~
500
450
400 0 20 40 60 80 100
tlme (sec)
120 140 160
Fig. 6.35 (a) molten metal environment, thermocouple 1
152
0
-1
1 -2 e.
~
1 -Ji
'1:1
11 i:
-4
-5
-6 180 200
700 ,----------.,-------------------------- ----------~---------
0
-1
-21 i
-3 'i 'a
! ii:
-4
-5
4~~~-~---~--~---~--~--~~--~--~ -6 o 20 40 60 80
time (sec)
100 120
Fig. 6.35 (b) room environment, thermocouple 1
140 160
Figure 6.35 Cooling curves and associated first derivatives for 319 alloy generated by the probe operating in (a) molten metal environment, and (b) at room temperature
Comparing the cooling curves presented in Figures 6.34 and 6.35, one can notice that the
final reaction, due to the formation of AhCu phase, is more evident when the probe
operates in-situ since the peak corresponding to this phase is clearer and more
identifiable. Further focusing on the final reaction portion of the cooling curves presented
in Figure 6.34 (a) and (b) one can notice that the temperature gradient in the sample when
the probe is operating in-situ is much less than that when the probe operates outside the
molten metal environment. The temperature difference as detected by the thermocouples
placed close to the probe wall and that placed at the center of the sample is 20°C in the
case where the probe operates in-situ, while a difference of 90°C is obtained when
operating the probe outside the crucible. It is believed that this reduction in the
temperature gradient, associated with the in-situ application of the probe, is the main
153
reason for getting better quality of information. In order to investigate this hypothesis two
thermal analysis tests were carried out using a weIl insolated mold, see Figure 6.a, at a
very slow cooling rate (0.04°C/s) on 356 and 319 aluminum alloys with 1% Fe
concentration and under slow cooling rates conditions. 9000 sec was needed to solidify 7
Kg sample, which is a condition that approximates equilibrium solidification (quazi
equilibrium). Figure 6.36 shows a comparlson between these curves and the cooling
curve generated by the in-situ probe. The comparlson shows that the quality of
information obtained by the in-situ probe is similar to that obtained by the classical cup
under a quasi-equilibrium condition. The reaction temperatures are approximately the
same with a difIerence of about 2°C in the final reaction temperature for the case when
319 Al was used while about 4°C difIerence is observed for the case of 356 Al. Aiso the
significance of the peaks, especially those which refer to the final reaction, is similar
when the first derivative curves for the cooling curves generated by both techniques are
compared.
Fig. 6.a. WeIl insulated mold used to carryout quazi-equilibrium classical thermal
analysis tests and the place of the thermocouples used in these tests.
154
Q.W
498.5 Oc Q.az ~~~------------------------------~--------~
0.01
o M -001 P
-Oaz i -0031!i
j500 ~&Or-~------------+-------------~~~~------~ !
-004 0:
-005
-Qo)
~r-~~----'---~----~--~----~--~----~--~ -Q07
Q 1000 2000 llOO 4000 500l BaX) 7000 moo 9000
time(sec)
(a) Cooling curve for 319 Alloy, Classical cup
700,----------------------------------------------, 602.1 Oc 562.5 Oc
496.5 Oc
-1 ~eoo~--~~==~-=~--~~----~--~~~----------~ i
-2 P
I~ ~f F 5OO~----~--------------~~~------------------~ ~
-4
-5
400~--~----~--~--~----~~~--~----~--~--~~
o 100
tirre(sec)
140 111) 100
(b) Cooling curve for 319 Alloy, In-situ probe
155
:!DO
.P • ...
700 ,.--------------------------,0.03
6oo+-~---------------------~
592.5 Oc
1 600+-~b.-~~_+1---.~~ .. ~---------~
549°C
3 ~ m 500 +----I-----+---""'+~---_\~.____---------t ~.02 i ~ ~
i 5 ~~ ~
ü:
400 ~.07
o 1000 2000 3000 4000 5000 6000 7000 8000 9000
tlme (seC)
(c) Cooling curve for 356 Alloy, Classical cup
700~---~------------------~
594.8 Oc 567°C 0.5
650~~-~---+_----~-----_=~~--~
-1.5 C -1: II.
-2
450+--m--~---~----+__-----~~~--~ -2.5
400+--WL~--~-~-~--~-~-~--~-~-__+~
o 20 40 60 80 100 120 140 160 180 200
time (sec)
(d) Cooling curve for 356 Alloy, In-situ probe
Fig. 6.36 Cooling curves generated by the classical cup and the in-situ probe for 319 and
356 aluminum alloys with 1 % Fe concentration.
156
From the data presented in Figure 6.36, one can postulate a theory as to why the probe is
superior to the cup which is used in the practice of the classical technique. The difference
between the two is rooted in the nature of the contact at the solid/solid interface that is
formed when liquid metal freezes on the wall of the cup or the probe.
ln the case of the cup, solidification occurs on the side walls as weIl as the bottom. As a
solid shell is formed, it pulls off the walls because of the bridging by the solidified shell
on the bottom. Thus, the contact at the solid/solid interface is subjected to transient
variations as the shell shrinks. The overall heat transfer at the interface will exhibit
variability (i.e. 'noise') as the interfacial contact varies.
For the probe, the solidification occurs only on the side walls and not on the bottom.
There is no solidification on the bottom because it is not cooled. Thus, the solid shell on
the side walls is not pulled from the walls because there is no bridging by solidified
material on the bottom. As a result, the contact at the solid/solid interface is stable. The
net result is that the cooling curves (i.e. first derivative curves) have much less noise than
those generated with the cup. The consequence is that thermal events are much more
visible when acquired with the probe than they are when acquired with the classical cup.
6.9 Industrial Experiments
Industrial testing was carried out at the Perth Plant of Grenville Castings Ltd in Ontario.
Figure 6.36 shows the industrial setup where the probe was immersed in a 5000-kg melt
157
of 356 aluminum alloy whose chemical composition may be seen in Table 6.4. A number
of thermal analysis tests were then performed under different cooling conditions.
Elementwt%
Al Si Mg Cu Mn Fe Zn Ti Sr
92.6 6.6 0.31 0.001 0.06 0.18 0.004 0.18 0.013
Table 6.4 Chemical composition of the 356 AI alloy used in industrial tests
Once the probe was introduced into the production line, melt conditions could no longer
be altered. Thus, the addition of grain refiners, modifying-agents or impurities to the melt
was not subject to being changed. Consequently, the only parameter open to modification
in this case was the cooling rate. Such a context thus led to the decision to apply different
cooling conditions when conducting the thermal analysis tests at the plant. The new probe
has the capability of extracting heat from a given sample at a constant rate throughout the
whole solidification process; it is also possible to alter this rate at any time during
solidification simply by adjusting a valve.
Fig. 6.36 Experimental setup at Grenville Castings Perth Plant
158
Two K-type thermocouples were used to capture the temperature of the solidified sample
during the freezing and melting processes. The upper thermocouple was placed at the
center of the sample 4.5 cm from the pipe walls and 2 cm below the top surface of the
sample, while the lower thermocouple was placed on the same plane as the first one at a
distance of 3 cm from the surface.
In the industrial tests, three different cooling conditions were investigated. Cooling
occurred relatively rapidly in the first case, where a period of 110 sec was required to
reduce the temperature from about 650°C to 500°C. In the second case the cooling
occurred relatively slowly (see figure 6.37).
The cooling curve shown in Figure 6.37 (a) reveals that the cooling rate increased during
the solidification process. Once the temperature of the solidifying sample reached the
eutectic temperature, which is 577°C, the return flow was increased and consequently the
cooling effect increased. This action affects the cooling curve in that it both reduces the
eutectic portion or plateau, and causes a sharp decline in the first derivative curve
immediately following the eutectic plateau.
Figure 6.37 (b) shows the cooling curves and the corresponding first derivatives under the
same conditions which prevailed in case (a), although this time the cooling was
maintained at a constant rate throughout the entire solidification process. Monitoring the
portion immediately beyond the eutectic plateau on the first derivative curve, it will be
159
observed that the curve drops more than 2.5 times faster in case (a) than it does in case
(b).
Cooling occurred at a relatively slow rate in the second case, at an average of O.3°C/s,
which is similar to the cooling conditions associated with sand casting [111. Figure 6.37 (c)
shows the cooling curve corresponding to slow cooling conditions. A time lapse of about
600 sec is needed to cool down the solidifying sample from about 680°C to about 500°C.
As in the first case, denoted by (a), the cooling rate was increased by opening the return
valve slightly once the eutectic temperature was reached. This increase in cooling rate
had a significant effect on the emergence of the last peak in the fust derivative curve, as
denoted by the number 4, and which corresponds to the formation of the M~Si-Al-Si
temary eutectic phase. Upon monitoring this peak it will be observed that in case (a),
where the first derivative reads about -7°C/sec, the peak vanishes. On the other hand, for
cases (b) and (c), a significant peak, denoted by the number 4, appears on the first
derivative curve for both cases. The clarity of the appearance of this peak is associated
with the slower cooling rate, as may be seen in Figure 6.3 7 (c).
The basic trend for cooling curves presented in Figure 6.37 may be described as follows.
The cooling curves show their first thermal arrest at about 626°C when the growth of
aluminum nuclei occur. An undercooling of OAOC was detected at that fust arrest. This
undercooling is due to an insufficiency of heterogeneous nuclei to start formation of (l
aluminum dendrites in the melt. Table 6.3 shows that this alloy has approximately 0.18
wt% of titanium. Our experience in the laboratory suggests that more than 0.3 wt"110 of Ti
concentration is needed to ensure the existence of sufficient heterogeneous nuclei in the
160
melt (see Figure 6.4). After this arrest, the temperature of the solidifying sample
continues to decrease. During this time, the liquid is progressively enriched in silicon
until a thermal anomaly occurs at about 608°C, caused by the latent heat of fusion of the
iron intermetallic phase. The cooling rate changes again to nearly zero. This event
continues for sorne time, depending on the cooling rate, as the latent heat of fusion for
eutectic aluminum-silicon evolves. An evident plateau appears at this stage. The final
reaction occurs at 547°C, when the Mg2Si phase is formed. As mentioned earlier,
however, the ability to detect this reaction by thermal analysis depends strongly on the
cooling rate. The peak corresponding to the formation of this phase is not evident on the
cooling curve, nor does it appear on the associated first derivative curve presented in
Figure 6.35 (a), where the cooling rate is relatively high. While significant peaks appear
on the first derivative curves associated with the cooling curves presented in Figures 6.35
(b) and (c), these peaks in fact reflect the different temperatures at which this phase forms
(Mg2Si), namely, 554°C, 552°C and 541°C.
Figure 6.38 shows the cooling and heating curves for 356 aluminum alloy. A span of 200
second is sufficient to remelt and to superheat the sample to melt conditions.
Figure 6.39 shows the actual size of the solidified sample frozen in the probe. The sample
is cylindrical in shape with a height of 4.2 cm and a radius of 4.5 cm. The photograph
also shows both the thermocouples introduced into the center of the sample.
161
.----------------_._~-------'
- l.ower thennocoople - Upper thennocoople
-1
~+-------------------~~~-----~----~ -2 _
540
520
627.' +---Ai"u,....AJ"""-__ ~-__4
l'
t" :: +----"r-----~-----' -3 i -4= ~ 825.' +-----\;.,--.,---+-- _5 ;
.. 825 --el 024.5 +----------1..--\-----1 -7
024+_-_-__ -_-~~L-L-___l-. 10 20 30 40 50 eo
lm. ( •• cr- Lower thermocoupte
1 è)
-3':.. ~
-41 .. 1
-5"-
~
-7
~~========~========~~~~~I~hennoc~~oopIe~==~--~-------~~--~--4 -8 o 20 60 60 100 120
a .... 1 MC)
(a)
700,-----------------,------------~
- upper thennocouple -Iower thermocouple 3
~~~-------~--------J--------------~
3
~~~------------------------++--~
~+_----~-----__ ----_.-----_----_.-----4~
o 20 60
amelMe)
(b)
162
60 100 120
700,--------------------------------,---------------, - Upper thermocouple
0.5
650+---~~_r ... ~----------~--------------~----~ o
~ cP 600 fl1_II"f~---'----'-----------------""".....:'"'~,___------__I__+__I -0.5 a;
i -1 ·i & ~ E ~ t! 550 -1.5.il
~
-2 500+--------------------------------------------+~
-2.5
450 +-------.,...-------,-------,-------,.-------,--------\- -3
o 100 200 300 400 500 600 tlme (sec)
(c)
Figure 6.37 Cooling curve and associated first derivative of356 aluminum alloy (as used
by Grenville Castings Ltd).
700 -------------------------,----------------,-------------------
--Uppor~
650
~600 !
~ !. !550
500
Cooling Curve ; Heating Curve i '~------------------------------~~14 ~i 4
450 0 100 200 300 400 500 600 700 800 900 1000
tlme (sac:)
Figure 6.38 Cooling and heating curves
163
Figure 6.39 Solidified physical sampling as taken by probe
References
1. Das-Gupta, R., Brown, C., Marek, S., "Analysis of Ovennodified 356
Aluminum Alloy", AFS Transactions, pp. 297-296
2. Mackay, R., Gruzleski, J., "Quantification of Iron ln Aluminum-Silicon
Foundry Alloys Via Thennal Analysis", International Journal of Cast Metals,
10, pp. 131-145, 1997
3. Shabetari, S., Gruzleski, J., "Gravity Segregation of Complex Intennetallic
Compounds in Liquid Aluminum-Silicon Alloys", Metallurgical and Materials
Transactions A. Vol. 26A, pp 999-2006. 1995
164
4. Mualzimoglu, M., Tenekedjiev, N., Closset, B., Gruzleski, J., "Studies of the
Minor Reactions and Phases in Strontium-Treated Aluminum-Silicon Casting
Alloys", Cast Metals, Vol. 6, No.11, 1993.
5. Narayanan, A., Samuel, F., Gruzleski, J., "Crystallization Behavior of Iron
Containing Intermetallic Compounds in 319 Aluminum Alloy", Metallurgical
and Materials Transactions A, Vol. 25A, pp. 1761-1773, 1994.
6. Shabestari, S., Ghodrat, S., "Thermal Analysis and Microstructural Evaluation
of Intermetallic Compounds Formed During Pre- and Post-Eutectic Reactions
in 319 Aluminum Alloy", 43rd Annual Conference of Metallurgists of CIM,
Light Metals, Hamilton, Ontario, pp.299-313, 2004.
7. Das-Gupta, R. "Influence of Iron on Microstructures and Mechanical
Properties of Strontium-Modified 356 Aluminum Alloy", Die Casting
Engineer, Vol. 40, No. 3, pp. 65-67, 1996.
8. Tenekedjiev, N. Mulazimoglu, H. Closset, B. and Gruzleski, J.
"Microstructures and Thermal Analysis of Strontium-Treated Aluminum
Silicon Alloys". 1995, American Foundrymen's Society, Inc. U.S.A.
165
9. Narayanan, A., Samuel, F., Gruzleski, J., "Crystallization Behavior of Iron
Containing Intermetallic Compounds in 319 Aluminum Alloy", Metallurgical
and Materials Transactions A, Vol. 25A, pp. 1761-1773, 1994
10. Tenekedjiev, N., Gruzleski, J. "Thermal Analysis of Strontium-Treated
Hypereutectic and Eutectic Aluminum-Silicon Alloys". AFS Transactions,
Vol. 99, USA, pp. 1-6, 1991.
11. Backerud, L., Krol, E., Tamminen, J., "Solidification Characteristics of
Aluminium Alloys; Volume 1: Wrought Alloys", Tangen Trykk, Norway,
1986.
166
Chapter Seven
Conclusions, Originality, and Suggested Future Work
7.1 Conclusions
An innovative state-of-the-art water-based heat pipe was successfully developed
and implemented in the field of molten aluminum technology with the aim of
automating thermal analysis and providing improved control over heat extraction
rates.
A new technique using an in-situ probe based on updated heat pipe technology
was therefore developed so as to carry out the required thermal analysis tests. The
key novelty in this respect is the in-situ feature of the device in view of the fact
that the probe can remain positioned in the melt, in a furnace or in a vessel, for an
indefinite period of time without harm coming to it. A cooling curve of the melt
may be initiated at the push of a button. Likewise, the heating curve for the
remelting of the same sample may be obtainedjust as easily.
Comparisons drawn between the in-situ technique using the heat pipe probe
versus the classical technique are all most encouraging from a number of aspects.
166
The quality of the infonnation obtained from the proposed new technique is better
than that obtained using the sampling cup. In fact, when using the new probe only
about one third of the time needed by the classical technique is required to
generate a similar quality of infonnation.
The proposed technique shows a predictable response with respect to the effect of
the grain refiner in the melt. The new probe has confinned that undercooling
decreases and recalescenece temperatures increase with increasing concentrations
of titanium in the melt.
The in-situ technique and probe may be used to replace the classical sampling cup
technique to carry out thennal analysis tests of aluminum alloys at industrial
plants as weIl as on the research laboratory scale. This new device will allow the
thennal analysis process to be automated, semi-continuous, and fully controllable.
The sensitivity of the probe in detecting minor reactions has been tested on:
-356 aluminum alloy with different concentrations ofiron;
-AI-7 wt% Si binary alloy with different concentrations of copper; and
-6063 wrought alloy.
Classical thennal analysis tests were also carried out simultaneously using a
preheated graphite cup. The comparisons show the greater potential inherent in
the new technique over the classical thennal analysis method. The peaks in the
signal that refer to intennetallic fonnation are clearer and more identifiable and of
167
better resolution when the new technique is applied. The amplitude of the peaks
obtained by the new probe is about three times higher than that obtained when
using the classical method. With this new technique, it is possible to correlate the
area below the intermetallic peak to the concentration of iron or copper in the
melt. This feature causes the new thermal analysis probe to behave as if it were a
rapid chemical analyzer for selected constituents.
This new thermal analysis probe which has the capacity of performing
thermal analysis in-situ, is also endowed with enough sensitivity to detect minor
reactions, such as iron or copper intermetallic formation, and has already
undergone the relevant tests for this. In the case of iron, the results were compared
with those obtained using classical thermal analysis techniques. The main
conclusions which may be drawn from this are the following:
1. The new probe is able to freeze a controlled sample of molten aluminum in a
molten metal environment. This makes it a viable option for performing the
thermal analysis test in-situ.
2. The quality of the information obtained using the new probe is better by far
than that obtained using the classical sampling cup technique. The peaks
corresponding to minor reactions, such as iron and copper intermetallic
formation, and obtained by the new technique, are on a better scale and of
greater resolution, as compared to those obtained by earlier techniques. The
168
new probe magnifies the minor peaks on the cooling curve as a result of the
heat flow from the bottom of the sample which tends to decrease the radial
temperature gradients in it. Consequently, more material solidifies at the same
time and more heat is detected during the formation of minor phases.
3. The heat removal rate from the sample obtained with the in-situ probe is
approximately constant during the freezing process. This amount of heat
removal may easily be controlled by regulating the evaporator return line and /
or modulating the heat removal on the condenser side of the probe.
4. The area below the first derivative curve, when the new probe is used to
generate the cooling curve, may be used to quantify the concentration of iron
or copper in the melt.
7.2 Statement of Originality
Contribution to Original Knowledge
1. A novel, first of its kind, in-situ probe which provides semi-continuous online
thermal analysis of aluminum melts was designed, built and tested.
2. The probe is based on a patent pending heat pipe technology that uses a swirl flow
to enhance heat transfer. However, a further contribution to the new heat pipe
technology was made by adding another flow modifier that is almost
169
perpendicular to the existing swirler. This made possible, for the first time, a
swirled flow heat pipe with heat extraction from the inner wall of the annulus of
the heat pipe.
3. Also novel was the method of controlling the rate of heat extraction. The device
was titted with a valve on a separate return line that fed the evaporator. In this
way the thermal analysis unit was transformed into one which featured
controllable heat extraction rates.
4. The successful use of water as the working substance in a heat pipe that was
immersed in molten aluminum was a novel contribution. There are no reports in
the literature of the use of water in a high heat flux system such as molten
aluminum.
5. The thermal analysis results obtained with the heat pipe probe are substantially
more detailed than those obtained with the c1assical technique. As a general rule,
it was shown that the heat pipe probe can magnify certain peaks and it can obtain
such data in a fraction of the time. This is the tirst time that such a finding has
been reported. Moreover, this is a major innovation that may lead to the
commercialization of such heat pipe probes in the future.
6. A theory as to why the probe is superior to the c1assical technique has been
formulated. It focuses on the nature of the solid/liquid interface that is formed
when a sample is soliditied. The probe maintains a relatively constant contact
resistance while the conventional unit has a contact resistance that increases with
time.
170
7. The solid/solid interface in the probe maintains a relatively stable contact
resistance during the solidification process. This is the first time that a thermal
analysis unit has been able to achieve this. The net result has been that the thermal
noise, which is generated when interface is subjected to a combination of high
heat fluxes and solidification, has been greatly reduced. This contribution to the
science should help researchers in the future to design thermal analysis devices
that are more sensitive because of the reduction in the thermal noise.
7.3 Future Work
The applicability ofthis new thermal analysis technique both in the laboratory and on an
industrial scale, has been born out by results from the present thesis. There are, however,
still a few points remaining which require further investigation. These points inc1ude the
following:
1. Modifying the probe configuration in order to make it more compact. One
suggestion is to merge both the main Hnes into one main line, and both the
retum lines into one. Another suggestion would be to modify the condenser
design to make it more compact.
2. Making use of the temperature of the main flow to create an automated signal
for exerting control over the retum line valve.
3. Further study is required to understand fully the thermal behavior ofboth the
heat pipe and the solidified sample.
171
4. Extend the thennal analysis study using different aluminum alloys such as
alwninwn copper-based alloys.
The author of this thesis believes that an open-ended range of research may be defined,
from now on, based on sorne of the findings described in this study which has been of
constant interest throughout. 1 am confident that this work will lead others to apply the
new technological approaches in original and successful metallurgical applications.
172