Eutectic Modification of Al-Si Casting Alloys

Linköping Studies in Science and Technology
Dissertation No. 2014
Eutectic modification of
Al-Si casting alloys
Jenifer Barrirero
Linköping Studies in Science and Technology
Dissertation No. 2014
Eutectic Modification of Al-Si casting alloys
Jenifer Barrirero
Nanostructured Materials
Department of Physics, Chemistry and Biology (IFM)
Linköpings University, Sweden
2019
© Jenifer Barrirero, 2019
ISBN 978-91-7519-007-5
ISSN 0345-7524
Printed by LiU-Tryck, Linköping, Sweden, 2019
ii
ahora
en esta hora inocente
yo y la que fui nos sentamos
en el umbral de mi mirada
Alejandra Pizarnik
iii
iv
Abstract
Aluminum alloys with silicon as the major alloying element are the most widely used
aluminum casting alloys. The eutectic phase in these alloys is formed by hard and brittle
silicon plates in an aluminum matrix. Such silicon plates can act as crack propagation
paths deteriorating the toughness of the material. To enhance ductility, silicon can be
modified to a coral-like microstructure by addition of a modifying agent. Amongst the
elements proposed as modifiers, only strontium, sodium and europium induce a plate-to-
coral transition, while others such as ytterbium, only refine the silicon plates. The exact
mechanism for the remarkable plate-to-coral change, and the reason why certain
elements only refine the structure, is still not completely understood.
In this investigation, atom probe tomography and transmission electron microscopy

were used to analyze and compare the crystal structure and the distribution of solute
atoms in silicon at the atomic level. An unmodified alloy and alloys modified by
strontium, sodium, europium and ytterbium were studied. Elements inducing silicon
plate-to-coral transition were found to contain nanometer sized clusters at the defects in
silicon with stoichiometries corresponding to compounds formed at the ternary eutectic
reaction of each system. In contrast, the addition of ytterbium, that only refines the
silicon plates, is unable to form clusters in silicon. We propose that the formation of
ternary compound clusters AlSiNa, Al2Si2Sr and Al2Si2Eu at the silicon / liquid interface
during solidification restrict silicon growth. The formation of clusters on silicon facets
create growth steps and increase growth direction diversity. The incorporation of clusters
in silicon explains the high density of crystallographic defects and the structural
modification from plates to corals.
The parallel lattice plane-normals 011Si // 0001Al2Si2Eu, 011Si // 6 10Al2Si2Eu and 111Si //
6 10Al2Si2Eu were found between Al2Si2Eu and silicon, and absent between Al2Si2Yb and
silicon. We propose a favorable heterogeneous formation of Al 2Si2Eu on silicon. The
misfit between 011Si and 0002Al2Si2X interplanar spacings shows a consistent trend with
the potency of modification for several elements such as strontium, sodium, europium,
calcium, barium, ytterbium and yttrium.
v
vi
Popular Summary
Air pollution is one of the top environmental concerns. Cars are responsible for around
12% of the total CO2 emissions in Europe. One way to reduce CO2 emissions and fuel
consumption is to reduce the mass of vehicles, also called light-weighting. Light-
weighting also plays an important role in the developing electro-mobility branch
compensating for the high weight of batteries and improving energy efficiency.
To significantly reduce the weight of a vehicle, we can focus on the materials’ selection,
for example, by replacing some steel or cast iron parts by light-weight aluminum parts.
Aluminum has one third the density of iron allowing a reduction of up to 50% weight
without compromising safety. This replacement is, however, not trivial. Strong and tough
alloys based on aluminum need to be designed and optimized for this purpose.
The mechanical properties and, consequently, the performance of an alloy can be

controlled and improved by designing their microscopic structure. In the aluminum-
silicon (Al-Si) alloys studied in this investigation, for instance, silicon grows in aluminum
in the form of hard and brittle plates that can act as crack propagation paths
deteriorating the resistance to fracture of the material. To enhance ductility, the
morphology of silicon can be modified to a coral-like structure by adding a modifying
agent. Amongst the several elements that have been proposed as potential modifiers,
only strontium, sodium and europium induce the plate-to-coral transition, while other
elements such as ytterbium, only refine the silicon plates. Although this modification has
been used at the industrial practice in the last decades, the exact underlying mechanism
for the remarkable plate-to-coral change, and the reason why certain elements only
refine the structure, is still not completely understood. This lack of knowledge hinders
the control of the microstructure homogeneity in more complex alloys such as Al-Si-Mg
and Al-Si-Mg-Cu.
The reason why this structural modification has not been completely understood, in spite
of almost 100 years of heavy investigation, is that the key of the effect lies at an extremely
small length-scale. Only now with the possibility of combining two characterization
methods with spatial, structural and chemical resolutions down to the atomic scale, we
are capable of gaining further understanding. Atom probe tomography and transmission
electron microscopy among other methods, were used in this investigation for a detailed
study of the distribution of atoms in the silicon crystal of Al-Si alloys. In alloys containing
vii
elements that transform the silicon structure from plates to corals (strontium, sodium
and europium), groups of atoms with fixed compositional relationships, were found
inside the silicon crystal. In contrast, with the addition of ytterbium, which only refines
the silicon plates, no clusters of atoms in silicon were found. We propose that the
formation of clusters during the solidification of the casting parts restrict the silicon
growth. These clusters lead to increased growth direction diversity, explaining the
formation of a coral-like structure.
This new understanding contributes to the future control of the microstructure evolution
of complex alloys at the industrial practice and the further enhancement and
optimization of aluminum casting parts.
viii
Populärvetenskaplig sammanfattning
Luftföroreningar är en av de viktigaste miljöfrågorna. I Europa står bilar för ungefär 12%
av det totala utsläppet av CO2. Ett sätt att minska bränsleförbrukning och CO2-utsläpp är
att minska fordonets massa, så kallade lättviktskonstruktioner. Lättviktskonstruktioner
spelar också en viktig roll för elektrifieringen av transportsektorn genom att kompensera
för batteriernas vikt och öka energieffektiviteten.
Vikten på ett fordon kan signifikant minskas genom materialvalen, t.ex. kan vissa stål-
och gjutjärnskomponenter bytas mot lättviktskomponenter i aluminium. Aluminiums
densitet är ca en tredjedel av järns vilket möjliggör en viktsreduktion på 50% utan att
tumma på säkerheten. Detta byte av material är dock inte trivialt. Höghållfasta och sega
legering av aluminium måste designas och optimeras för detta ändamål.
De mekaniska egenskaperna och följaktligen prestandan hos en legering kan kontrolleras

och förbättras genom design av dess mikrostruktur. Till exempel, i de legeringar mellan
aluminium och kisel (Al-Si) som studerats i denna avhandling så växer kislet in i
aluminium och bildar hårda och spröda plattor som kan verka som spricktillväxtvägar
som lättare ger upphov till materialbrott. För att öka segheten så kan kislets morfologi
ändras till en koralliknade morfologi genom att tillsätta en modifierare. Av de många
ämnen som föreslagits som möjliga modifierare är det endast strontium, natrium och
europium som inducerar den eftertraktade övergången från plattor till koraller medans
andra ämnen så som ytterbium endast förfinar kiselplattorna.Trots att denna
modifiering har använts industriellt de senaste årtiondena så är fortfarande de
underliggande mekanismerna för denna remarkabla förändring av plattorna inte
förstådd. Avsaknaden av denna kunskap hindrar att oss från att kunna kontrollera
mikrostrukturen hos mer komplexa legeringar så som Al-Si-Mg och Al-Si-Cu.
Orsaken till denna strukturförändring är ännu inte förstådd, trots nästan 100 år av
studier. Det beror på att nyckeleffekterna återfinns på en extremt liten längdskala.
Endast nu är det möjligt att kunna studera detta genom att kombinera två
karakteriseringsmetoder som tillsammans ger nödvändig spatial, strukturell och kemisk
upplösning, dvs på atomär nivå. Atomsond och transmissionselektronmikroskopi, jämte
andra metoder, användes i denna avhandling för detaljerade studier av atomernas
placering i kiselkristallerna i Al-Si legeringar. I legeringar innehållande ämnen som
modifierar Si-strukturen (strontium, natrium och europium) hittades endast dessa
ix
ämnen grupperade till små områden i kiselkristallen med en och samma
sammansättning, s.k. kluster. Det står i kontrast till de ämnen som enast förfinanade
kiselplattorna där sådana kluster inte kunde hittas. Vi föreslår att bildandet av kluster
under framställningsprocessen av de gjutna detaljerna begränsar tillväxten av
kiselkristaller. Dessa kluster leder till en större variation i tillväxtriktning som också
förklarar bildandet av en koralliknande struktur.
Denna nya förståelse bidrar till en framtida kontroll av mikrostrukturutvecklingen av

gjutna komplexa aluminiumlegeringar så att förbättrade och optimerade lättviktdetaljer
kan framställas industriellt.
x
Vereinfachte Zusammenfassung
Die Luftverschmutzung ist eines der größten Umweltprobleme. Pkw sind für rund 12%
der gesamten CO2-Emissionen in Europa verantwortlich. Eine Möglichkeit, den
Kraftstoffverbrauch und damit den CO2-Ausstoß zu reduzieren, besteht darin, die Masse
der Fahrzeuge zu reduzieren, was auch als Leichtbau bezeichnet wird. In der zunehmend
an Bedeutung gewinnenden Elektromobilitätsbranche spielt der Leichtbau ebenfalls eine
wichtige Rolle, da er das hohe Gewicht der Batterien kompensieren und damit die
Energieeffizienz verbessern kann.
Eine deutliche Reduzierung des Fahrzeuggewichts ist durch eine geeignete

Materialauswahl möglich, indem beispielsweise einige Stahl- oder Eisengussteile durch
leichte Aluminiumteile ersetzt werden. Aluminium hat ein Drittel der Dichte von Eisen,
was eine Gewichtsreduzierung von bis zu 50% ermöglicht, ohne die Sicherheit zu
beeinträchtigen. Dieser Ersatz ist jedoch nicht trivial, da feste und zähe Legierungen auf
Aluminiumbasis für diesen Zweck entwickelt und optimiert werden müssen.
Die mechanischen Eigenschaften und damit die Leistungsfähigkeit einer Legierung

können durch die Gestaltung ihrer Mikrostruktur, dem Gefüge, kontrolliert und
verbessert werden. In den in dieser Untersuchung untersuchten Aluminium-Silizium (Al-
Si) -Legierungen liegt Silizium beispielsweise in Form von harten und spröden Platten im
Aluminium vor, die als Rissausbreitungspfade fungieren können und die Bruchfestigkeit
des Materials beeinträchtigen. Um die Duktilität zu erhöhen, kann die Morphologie des
Siliziums durch Zugabe eines sogenannten Veredelungsmittels in eine korallenartige
Struktur überführt werden. Unter den zahlreichen Elementen, die als potenzielle
Veredelungselemente vorgeschlagen wurden, erzeugen nur Strontium, Natrium und
Europium den Übergang von einer plattenförmigen zu einer korallenförmigen
Morphologie, während andere Elemente wie Ytterbium nur die Siliziumplatten
verfeinern. Obwohl diese Veredelung in den letzten Jahrzehnten in der industriellen
Praxis angewendet wurde, ist der genaue zugrunde liegende Mechanismus für diese
bemerkenswerten Morphologieänderung und der Grund, warum bestimmte Elemente
die Struktur nur verfeinern, noch nicht vollständig verstanden. Dieses fehlende Wissen
behindert die Kontrolle der Mikrostrukturhomogenität bei komplexeren Legierungen wie
beispielsweise Al-Si-Mg und Al-Si-Mg-Cu.
xi
Der Grund, warum diese Morphologieänderung trotz fast 100-jähriger intensiver
Forschung nicht vollständig verstanden wurde, ist, dass sich die Ursache des Effekts auf
einer extrem kleinen Längenskala abspielt. Erst heute sind wir durch die Kombination
zweier Charakterisierungsmethoden mit räumlichen, strukturellen und chemischen
Auflösungen bis in den atomaren Bereich in der Lage, dieses Verständnis zu erweitern.
Die Atomsondentomographie und die Transmissionselektronenmikroskopie wurden in
dieser Untersuchung unter anderem für eine detaillierte Untersuchung der
Atomverteilung in der Siliziumphase in Al-Si-Legierungen eingesetzt. In Legierungen mit
Veredelungselementen die eine korallenförmige Morphologie des Siliziums erzeugen
(Strontium, Natrium und Europium), wurden in der Siliziumphase Atomcluster mit
festem Verhältnis der Zusammensetzung gefunden. Im Gegensatz dazu wurden bei der
Zugabe von Ytterbium, das die Siliziumplatten nur verfeinert, keine Atomcluster im
Silizium gefunden. Wir schlagen vor, dass die Bildung von Clustern während der
Erstarrung der Gussteile das Siliziumwachstum behindert. Diese Cluster führen zu einer
erhöhten Vielfalt der Wachstumsrichtungen und erklären damit die Bildung einer
korallenartigen Struktur.
Dieses neue Verständnis trägt zur zukünftigen Steuerung der Mikrostruktur komplexer
Legierungen in der industriellen Praxis und zur weiteren Verbesserung und Optimierung
von Aluminiumgussteilen bei.
xii
Preface
This thesis is the result of the Joint European Doctoral Program in Advanced Materials
Science and Engineering - DocMASE. The work was accomplished in the Chair of
Functional Materials at the University of Saarland, Saarbrücken, Germany; and in the
group of Nanostructured Materials at the Department of Physics, Chemistry and Biology
(IFM) of Linköping University, Linköping, Sweden.
The work was supported by the Erasmus Mundus program of the European Commission
(DocMASE), the EU funding in the framework of the project AME-Lab (European
Regional Development Fund C/4-EFRE-13/2009/Br). The atom probe was financed by
the German Research Society (DFG) and the Federal State Government of Saarland
(INST 256/298-1 FUGG).
xiii
xiv
Included papers
Paper I.
Comparison of segregations formed in unmodified and Sr-modified Al-Si alloys studied

by atom probe tomography and transmission electron microscopy.
Barrirero J., Engstler M., Ghafoor N., de Jonge N., Odén M., & Mücklich F. Journal of
Alloys and Compounds, 611, 410–421 (2014).
https://doi.org/10.1016/j.jallcom.2014.05.121
Paper II.
Cluster formation at the Si/liquid interface in Sr and Na modified Al–Si alloys.
Barrirero J., Li J., Engstler M., Ghafoor N., Schumacher P., Odén M., & Mücklich F.
Scripta Materialia, 117, 16–19 (2016). https://doi.org/10.1016/j.scriptamat.2016.02.018
Paper III.
Eutectic modification by ternary compound cluster formation in Al-Si alloys.
Barrirero J., Pauly C., Engstler M., Ghanbaja J., Ghafoor N., Li J., Schumacher P., Odén
M., & Mücklich F. Scientific Reports 9, 1-10 (2019). https://doi.org/10.1038/s41598-
019-41919-2
Paper IV.
Nucleation and Growth of Eutectic Si in Al-Si Alloys with Na Addition.
Li J. H. H., Barrirero J., Engstler M., Aboulfadl H., Mücklich F., & Schumacher P.
Metallurgical and Materials Transactions A, 46(3), 1300–1311 (2014).
https://doi.org/10.1007/s11661-014-2702-6
Paper V.
Phase selective sample preparation of Al-Si alloys for Atom Probe Tomography.
Barrirero J., Engstler M., Odén M., Mücklich F. Practical Metallography, 56(2), 76 - 90
(2019) https://doi.org/10.3139/147.110557
xv
xvi
My contribution to the papers
Paper I.
I was responsible for the planning of the project, prepared the samples for SEM / FIB,
EBSD, APT and TEM. I performed the APT and EBSD measurements and analyses and
participated in the TEM analysis. I wrote the first draft of the paper and was in charge of
the submission and revision processes.
Paper II.
APT and TEM. I performed the APT measurements and analyses and participated in the
TEM analysis. I wrote the first draft of the paper and was in charge of the submission and
revision processes.
Paper III.
EBSD, APT and TEM. I performed the APT and EBSD measurements and analyses and
participated in the TEM. I wrote the first draft of the paper and was in charge of the
submission and revision processes.
Paper IV.
I prepared the samples for APT. I performed the APT measurements and analyses. I
contributed to the writing of the paper.
Paper V.
I developed the method, prepared the samples, and applied it successfully in my work. I
wrote the first draft of the paper and was in charge of the submission and revision
processes.
xvii
xviii
Other papers not included in the thesis
Barrirero J, Engstler M, Mücklich F. Atom Probe analysis of Sr distribution in AlSi
foundry alloys. In: Light Metals 2013 - TMS. ; 2013:291-296.
doi:10.1002/9781118663189.ch50.
Liang S-M, Engstler M, Groten V, Barrirero J, Mücklich F, Bührig-Polaczek A, Schmid-

Fetzer R. Key experiments and thermodynamic revision of the binary Al–Sr system. J
Alloys Compd. 2014;610:443-450. doi:10.1016/j.jallcom.2014.05.018.
Mücklich F, Engstler M, Britz D, Barrirero J, Rossi P. Why We Need All Dimensions to

Solve Both Very Old and Very New Questions in Materials at the Micro-, Nano- and
Atomic Scales. Pract Metallogr. 2015;52(9):507-524. doi:10.3139/147.110360.
Li JH, Barrirero J, Sha G, Aboulfadl H, Mücklich F, Schumacher P. Precipitation

hardening of an Mg–5Zn–2Gd–0.4Zr (wt. %) alloy. Acta Mater. 2016;108:207-218.
doi:10.1016/j.actamat.2016.01.053.
Shulumba N, Hellman O, Raza Z, Alling B, Barrirero J, Mücklich F, Abrikosov I A, Odén

M. Lattice Vibrations Change the Solid Solubility of an Alloy at High Temperatures. Phys
Rev Lett. 2016;117(20):205502. doi:10.1103/PhysRevLett.117.205502.
Yalamanchili K, Wang F, Aboulfadl H, Barrirero J, Rogström L, Jiménez-Pique E,

Mücklich F, Tasnadi F, Odén M, Ghafoor N. Growth and thermal stability of TiN/ZrAlN:
Effect of internal interfaces. Acta Mater. 2016;121:396-406.
doi:10.1016/j.actamat.2016.07.006.
Menezes CM, Bogoni N, Barrirero J, Aboulfadl H, Mücklich F, Figueroa CA. Influence of

the surface chemistry-structure relationship on the nanoscale friction of nitrided and
post-oxidized iron. Surf Coatings Technol. 2016;308:220-225.
doi:10.1016/j.surfcoat.2016.07.095.
Roa JJ, Aboulfadl H, Barrirero J, Turon-Vinas M, Mücklich F, Anglada M. Chemical

segregation in a 12Ce-ZrO 2 /3Y-ZrO 2 ceramic composite. Mater Charact. 2017;132:83-
91. doi:10.1016/j.matchar.2017.07.045.
Calamba KM, Pierson JF, Bruyère S, Febvrier AL, Eklund P, Barrirero J, Mücklich F,
Boyd R, Johansson Jõesaar MP, Odén M. Dislocation structure and microstrain
evolution during spinodal decomposition of reactive magnetron sputtered heteroepixatial
c-(Ti 0.37 ,Al 0.63 )N/c-TiN films grown on MgO(001) and (111) substrates. J Appl Phys.
2019;125(10):105301. doi:10.1063/1.5051609.
El Azhari I, Barrirero J, García J, Soldera F, Llanes L, Mücklich F. Atom Probe

Tomography investigations on grain boundary segregation in polycrystalline Ti(C,N) and
Zr(C,N) CVD coatings. Scr Mater. 2019;162:335-340. doi:10.1016/j.scriptama
t.2018.11.041.
xix
Calamba KM, Barrirero J, Johansson Jõesaar MP, Bruyère S, Boyd R, Pierson JF, Le
Febvrier A, Mücklich F, Odén M. Growth and high temperature decomposition of
epitaxial metastable wurtzite (Ti1-x,Alx)N(0001) thin films. Thin Solid Films. 2019;
137414. doi.org/10.1016/j.tsf.2019.137414
xx
Acknowledgements
First and foremost, I would like to express my gratitude to my two supervisors, Prof.
Frank Mücklich and Prof. Magnus Odén.
Prof. Frank Mücklich is a leadership role model for me. I have learned from his ever
positive approach and communication skills. I would like to thank him for his trust and
unrestricted access to all I needed to fulfill this research.
Prof. Magnus Odén is an inspiring scientist, I learned from his experience and
confidence. I would like to thank him for the fruitful discussions, academic guidance,
personal advice and endless support.
I am especially grateful to Michael Engstler for believing in me, always finding a solution
to whatever came up and accompanying me during all these years.
I thank all the co-authors of the publications included in this thesis. This work would not
have been possible without their contributions. Thank you to Christoph Pauly for giving
me insight on materials characterization and for the fruitful discussions, and to Michael
Engstler, Flavio Soldera, Andrés Olguín and Pranav Nayak for proof reading the
manuscript of the thesis.
Thank you to the colleagues at the Chair of Functional Materials at Saarland University
and the Nanostructured Materials group at Linköping University that made all these
years such an enjoyable time.
I thank the Erasmus Mundus Joint European Programme for giving me the possibility of
performing my doctoral studies in two countries and working in multi-cultural
international environments.
I am very grateful to Andrés Olguin for holding me in his embraces and encouraging me
to go forward in difficult times.
I am particularly grateful to Roberto Tannchen who stood wholehearted by my side

helping me to sail through the storm.
Thank you to Corinna Markmann for her inspiring light.
I thank my parents Carmen Caffaratti and Domingo Barrirero for their love and
unconditional support. Thanks to my brother Exe Barrirero for teaching me with his
strength and determination. Thanks to Caro Rey for her openness, sweetness and peace.
xxi
I am extremely grateful to all my dearest friends who are the biggest treasure that I have
in this life: Pauli Kuschnir, María Laura Cenci, Anne Villamil, Moni Echeverry, Marian
De Giovanni, Prisi Zanetti, Pauli Sierra, Anita Barbotti, Cele Karchesky, Juli Trivelli,
Gaby Fertonani, Celi Bratovich, Hisham Aboulfadl, Fede Benedetto, Marco Francesconi,
Oscar Deccó, Feli Giussani, Almila Özügürler, Micha Agthe, Tiny Walther, Gonza
Schierloh, Eugenia Dalibon, Naureen Ghafoor, Nina Shulumba, Ana Chaar, Fei Wang,
Gustavo Barrirero, Lucía Campo, Nico Souza, Flora Kiss, Fede Lasserre, Ire Morales,
Agus Guitar, Gyöngyi Andras, Flavio Soldera, Fede Miguel, Seba Suárez.
xxii
Acronyms and symbols
A Solidification interface area

APT Atom Probe Tomography
AR Atomic Resolution
BSE Back-Scattered Electrons
∆HF Enthalpy of fusion
∆SF Entropy of fusion
EBSD Electron Backscattered Diffraction
EDX Electron Dispersive X-ray
ETD Everhart Thornley Detector
FIB Focused Ion Beam
G Temperature gradient
HR High Resolution
HV High Voltage
IIT Impurity Induced Twinning
IPF Inverse Pole Figure
L Latent heat of fusion per unit volume
LEAP Local Electrode Atom Probe
µXRF Micro X-Ray Fluorescence
PF Pole Figure
q Total rate of heat extraction
R Ideal gas contant
SE Secondary Electrons
SEM Scanning Electron Microscopy
SIMS Secondary Ions Mass Spectroscopy
TEM Transmission Electron Microscopy
ToF Time-of-Flight
TTL Through The Lens
UHV Ultra-High Voltage
V Average interface velocity
V Growth rate
vCD Low voltage - high Contrast Detector
VF Volume Fraction
Z Atomic number
α Jackson's factor
η Nearest neighbors at a growing solid/liquid interface
ν Coordination number
ξ Orientation factor
xxiii
xxiv
Content
Abstract ................................................................................................................................ v
Popular Summary .............................................................................................................. vii
Populärvetenskaplig sammanfattning ................................................................................ ix
Vereinfachte Zusammenfassung......................................................................................... xi
Preface............................................................................................................................... xiii
Included papers ................................................................................................................. xv
My contribution to the papers ......................................................................................... xvii
Other papers not included in the thesis............................................................................ xix
Acknowledgements ........................................................................................................... xxi
Acronyms and symbols ................................................................................................... xxiii
1. Introduction ................................................................................................................. 1
1.1 Scope ........................................................................................................................... 2
1.2 Outline of the thesis .................................................................................................... 2
2. Aluminum casting alloys .............................................................................................. 5
2.1 Al-Si alloys ..................................................................................................................8
3. Eutectic solidification .................................................................................................. 11
4. Eutectic solidification of Al-Si alloys .......................................................................... 17
4.1 Plate-like structure and TPRE mechanism ..............................................................20
5. Eutectic modification of Al-Si alloys .......................................................................... 25
5.1 Enhancement of mechanical Properties ................................................................... 25
5.2 Elements modifying eutectic silicon and their effect on the microstructure ...........28
5.3 Porosity and shrinkage ............................................................................................. 31
5.4 Theories explaining eutectic modification of Al-Si alloys ........................................ 32
5.4.1 Early theories for eutectic modification............................................................. 32
5.4.2 Modification effect on eutectic growth .............................................................. 35
5.4.3 Modification effect on eutectic nucleation ........................................................ 39
5.4.4. Recent developments........................................................................................ 46
6. Sample preparation and Characterization ................................................................. 47
xxv
6.1 Al-Si alloys ................................................................................................................ 47
6.1.1 Al – 7 wt% Si ....................................................................................................... 47
6.1.2 Al – 5 wt% Si ...................................................................................................... 47
6.2 Scanning electron microscopy – SEM .....................................................................48
6.2.1 Secondary electron imaging – SE ......................................................................48
6.2.2 Backscattered electron imaging – BSE ............................................................. 49
6.2.3 Energy dispersive X-ray spectroscopy – EDX ...................................................50
6.2.4 Electron backscattered diffraction – EBSD ...................................................... 51
6.3 Focused ion beam – FIB .......................................................................................... 53
6.3.1 Sample preparation for TEM ............................................................................. 54
6.3.2 Sample preparation for APT .............................................................................. 55
6.4 Transmission electron microscopy – TEM .............................................................. 56
6.5 Atom probe tomography – APT ............................................................................... 57
6.5.1 Working principle and experimental setup ....................................................... 57
6.5.2 Data reconstruction, visualization and analysis ............................................... 60
6.5.3 Trajectory aberrations and local magnification effects ..................................... 61
7. Summary of included papers and contribution to the field ....................................... 63
Paper I. Comparison of segregations formed in unmodified and Sr-modified Al-Si

alloys studied by atom probe tomography and transmission electron microscopy ... 64
Paper II. Cluster formation at the Si / liquid interface in Sr and Na modified Al-Si
alloys ........................................................................................................................... 65
Paper III. Eutectic modification by ternary compound cluster formation in Al-Si

alloys ........................................................................................................................... 65
Paper IV. Nucleation and growth of eutectic Si in Al-Si alloys with Na addition ...... 66
Paper V. Phase selective sample preparation of Al-Si alloys for atom probe
tomography ................................................................................................................ 66
7.1 Contribution to the field ........................................................................................... 67
7.2 Outlook and future work .......................................................................................... 67
7.2.1 APT analysis of modified Al-Si eutectic phase ...................................................68
7.2.2 Calculation of phase diagrams and simulations ................................................68
xxvi
7.2.3 Correlative characterization of the solidification front ..................................... 69
7.2.4 Interaction of modifying elements in Al-Si alloys with magnesium and iron... 69
8. References .................................................................................................................. 71
Paper I ......................................................................................................................83
Paper II .................................................................................................................... 97
Paper III ................................................................................................................. 103
Paper IV .................................................................................................................. 115
Paper V ................................................................................................................... 139
xxvii
xxviii
Eutectic modification of Al-Si casting alloys
1. Introduction
Improving the design of structural alloys is a never-ending task as industrial applications
continuously demand increased performance. Aluminum alloys are used in a wide range
of applications due to their attractive low density in comparison to steels, which offers
the possibility to design light-weight components.
Aluminum alloys with silicon as the major alloying element are the most widely used for
casting applications. Their excellent castability and high strength-to-weight ratio makes
them an appealing material for the automotive industry [1]. Al-Si alloys have an irregular
eutectic phase formed by faceted silicon in a non-faceted aluminum matrix [2–4]. During
the solidification of this phase, several microstructures and growth modes can be
obtained depending on the temperature gradient and cooling rate [5,6].
In the solidification processes generally used in industry, such as sand casting or die
casting, eutectic silicon grows in the form of plates, also known as flakes. Since silicon
plates are hard and brittle and, often act as easy paths for cracks, a microstructural
modification towards rounded silicon branches is desirable [7]. Almost 100 years ago, a
peculiar phenomenon was patented by Pacz [8], who found that the addition of low
concentrations of alkaline fluorides, particularly sodium fluoride changed the silicon
structure in these alloys from plate-like to fibrous or coral-like. This modification
significantly improves ductility and therefore, it is industrially used nowadays [7].
However, challenges in its application are found when the composition of the alloy gets
more complex.
In general, industrially relevant Al-Si alloys contain ternary and quaternary elements
such as magnesium and/or copper. The addition of these elements can make the alloy
heat treatable and improve its strength and machinability [1]. All of these alloys benefit
from the modification of the eutectic microstructure, but in the case of ternary and
quaternary alloys, the change is often inhomogeneous showing well-modified regions
and non-modified regions with coarse silicon parts. Inhomogeneities in the
microstructure are detrimental for the properties of the alloys and hinder its further
development. The lack of understanding of the fundamental mechanism underlying
eutectic modification impedes further improvement of these alloys.
1
During the last decades, several alkaline earth and rare earth metals were investigated
and two groups of elements were differentiated based on their effects: (1) elements which
change the silicon morphology into corals such as sodium [9,10], strontium [11,12] and
europium [13]; and (2) others which only refine the plate-like silicon or can only partially
change the silicon structure such as calcium [14,15], yttrium [16], ytterbium [17,18],
barium [19] and most rare-earths [20]. The reason why these elements behave differently
is still unknown.
1.1 Scope
The aim of this thesis is to widen our understanding of the eutectic modification in Al-Si
alloys. Several investigations have been performed regarding this topic, however, no
systematic research revealed and compared the three-dimensional distribution of the
atoms in the silicon phase. I seek to analyze, for the first time ever, atomically resolved
chemical information in these alloys by atom probe tomography (APT) and propose a
model for the multiplication of crystal defects in silicon and its morphological change.
The study will focus on three cases:
- Al-Si alloy with no addition of any modifier as a reference for the identification of
changes and characteristics related to addition of a modifier agent.
- Alloys with addition of the three most powerful modifiers known to date: sodium,
strontium and europium. By studying and comparing the solute distribution in
eutectic silicon in these alloys, it is possible to pinpoint the common
characteristics to further understand the plate-to-coral transition.
- An alloy with ytterbium addition, where the silicon plate structure is refined
without corals formation, to understand the differences with coral-forming
elements.
1.2 Outline of the thesis

This thesis contains a background about aluminum and Al-Si alloys highlighting their
relevance, classification and general applications in chapter 2. An introduction to eutectic
solidification of irregular systems in general and, of Al-Si alloys in particular is given in
chapter 3 and 4 respectively. Chapter 5 presents a literature review about eutectic
modification. Chapter 6 goes through the sample preparation and characterization
methods used during this investigation. Chapter 7 gives a summary of the papers
2
included in the thesis and the contribution to the field, together with an outlook and
suggestions for further work. Finally, the outcome of the thesis is presented in five
scientific papers.
3
4
2. Aluminum casting alloys

Several useful properties make aluminum the base metal of choice for engineering
solutions. It is light weight, with a density of approximately one-third that of steel; it has
high electrical and heat conductivities; good corrosion resistance and can be used in
wrought and casting applications. Its high recyclability also plays an important role with
a remelting process out of scrap requiring only about 5 % of the energy needed to extract
the same amount of primary metal from the bauxite ore [1,21].
Aluminum is the most abundant metallic element, the third most abundant element in
earth’s crust and the second most industrially used after iron. Surprisingly, its existence
was not acknowledged until the beginning of the 19th century and it took almost until the
end of the century to develop an economically viable production route [21]. Although its
short history of about 200 years in comparison to thousands of years of iron use,
aluminum metallurgy evolved extremely fast to cover a wide range of applications. The
constant growth in the importance of aluminum is tightly related to the increasing
demands for mass reduction in vehicles to improve fuel consumption and lower CO2
emissions [1,21].
Pure aluminum has low strength and has to be alloyed for use in structural applications.
The major alloying elements are copper, manganese, magnesium, silicon and zinc. The
mechanical, physical and chemical properties of these alloys are determined by their
composition and microstructure. Depending on the composition, manufacturing
processes and mechanism of properties development, alloys can be classified into
wrought or casting alloys and further into heat treatable or not heat treatable alloys. The
most important difference between wrought and casting composition is the alloy’s
castability. Wrought parts are casted in round or rectangular cross sections with a
uniform solidification front. On the contrary, casting compositions are able to achieve
dimensionally accurate near-net-shape parts with complex geometries and designed
properties, in order to fulfill specified requirements [1]. The good castability of aluminum
castings is related to their relative high fluidity, low melting point, short casting cycles
and relatively low tendency for hot cracking [22]. Since the present work focuses on the
eutectic modification of casting alloys, only this type will be further described. Aluminum
castings present a wide range of compositions with great versatility in the achievable
properties (Table 2.1).
5
The most common classification of aluminum castings is based on the major alloying
elements. Although no international standard nomenclature is available, the designation
given by the Aluminum Association of the United States is well known [23]. This
designation has four numeric digits separated by a period between the third and fourth.
Table 2.2 shows the main categories defined by the first digit based on the major alloying
constituent. The second and third digits do not have any significance but are unique to
each alloy. The fourth digit describes whether it is casting (0) or ingot (1, 2). For fine
variations in the composition limits, a letter preceding the numbers is added (A, B, C).
Table 2.1: Range of mechanical properties for aluminum casting alloys [1]
Tensile strength 70 – 505 MPa
Yield strength 20 – 455 MPa
Elongation < 1 – 30 %
Hardness 30- 150 HB
Electrical conductivity 18 – 60 %IACS
Thermal conductivity 85 – 175 W/m*K at 25 °C
Fatigue limit 55 – 145 MPa
Coefficient of linear thermal expansion at 20 – 100 °C (17.6 – 24.7) x 10-6/°C
Shear strength 42 – 325 MPa
Modulus of elasticity 65 – 80 GPa
Specific gravity 2.57 – 2.95
Table 2.2: Casting alloys designation of the Aluminum Association (ANSI H35.1) [23]
1xx.x pure aluminum (99% or greater)
2xx.x aluminum-copper alloys
3xx.x aluminum-silicon + copper and/or magnesium
4xx.x aluminum-silicon
5xx.x aluminum-magnesium
7xx.x aluminum-zinc
8xx.x aluminum-tin
9xx.x aluminum + other elements
6xx.x unused series
Table 2.2 summarizes the most relevant composition families for aluminum castings.
Some of their remarkable characteristics and applications are [1]:
6
- Aluminum-copper alloys primarily fulfill requirements of strength and toughness.

They exhibit high strength and hardness at room and elevated temperatures.
Mostly used in the aerospace industry.
- Aluminum-silicon-copper alloys exist in a wide range of compositions. Copper is
added to improve strength and machinability, while silicon contributes to the
castability. They dominate the market for powertrain components such as engine
blocks, cylinder heads or pistons.
- Aluminum-silicon binary alloys have excellent fluidity, castability and corrosion
resistance, but limited strength and poor machinability. They show low specific
gravity and coefficients of thermal expansion. In hypoeutectic alloys, the strength,
ductility and castability can be improved by modification of the eutectic phase.
- Aluminum-silicon-magnesium alloys combine remarkable casting characteristics,
outstanding properties after heat treatment, good corrosion resistance and low
level of thermal expansion. Eutectic modification is also used for these alloys to
increase elongation.
- Aluminum-magnesium are binary alloys with moderate to high strength and
toughness with excellent corrosion resistance. They have good weldability,
machinability and an attractive appearance. They are used to produce high-
pressure die cast automotive steering wheels and structural components.
- Aluminum-zinc-magnesium have the particularity of naturally aging, showing full
strength after approximately 30 days at room temperature. Their machinability
and corrosion resistance is good in general, but they often show poor castability.
- Aluminum-tin are alloys used for bearing applications. The light weight and good
heat dissipation are beneficial characteristics. Alloys with 5.0 to 7.0 wt% Sn are
often used when low friction, low compressive and fatigue strengths and good
resistance to corrosion are needed.
The steady increase in the production of aluminum responds in a great part to its light
weight. The reduction in weight by the use of aluminum alloys in automotive designs,
improves the efficiency of energy consumption without compromising performance and
safety, with a minimal impact on costs [21]. Castings in the transport sector are used in
applications such as engine blocks, cylinder heads, pistons, wheels or suspension
components, just to mention some examples. In addition to the automotive sector,
aluminum alloys are also used in aerospace applications, construction, machinery,
packaging, cooking utensils, and housing for electronics or pressure vessels.
All casting manufacturing methods can be applied to cast aluminum alloys. The choice of
the most convenient method depends on the size, complexity of the design and number
7
of parts to be manufactured per year [1,21]. The three most important methods are die
casting, permanent mold and sand casting. Figure 2.1 shows application sectors and their
relative percentages for these three casting processes.
High pressure die casting accounts for more than 50 % of aluminum castings produced
[21]. It is a fast near-net shape manufacturing process well suited for large production
volumes. High cooling rates are desirable to get short manufacturing cycles. This method
enables the fabrication of dimensionally accurate parts with excellent surface finish [22].
The tooling and automation costs are rather high, but they are compensated by the
production volume.
Permanent mold and sand castings are used for thicker wall products or for parts
requiring internal hollow sections that strictly need a sand core to be fabricated (e.g.
cylinder heads) [21]. These two methods show slower cooling rates than the high
pressure die casting and therefore, for Al-Si parts manufactured by these methods, the
modification of the eutectic phase treated in this thesis is used to refine the structure and
improve strength, ductility and machinability [22]
Figure 2.1: Applications for aluminum cast products separated by casting process, data
collected from [24].
2.1 Al-Si alloys

Al-Si alloys are the most widely used aluminum castings, especially in automotive
applications. Silicon provides good castability with improved fluidity, elevated-
temperature resistance to cracking and good feeding characteristics [1]. The amount of
silicon depends on the desired properties but also on the casting process used to
8
manufacture the part. Processes that need higher heat flux use higher silicon content to
improve fluidity, which in turn assists the filling of narrow cavities and intricate designs.
Good castability is associated with alloys of reduced solidification range. Addition of
silicon also reduces the specific gravity and the coefficient of thermal expansion.
Commercial alloys may contain silicon from hypoeutectic concentrations to hypereutectic
with up to about 30 wt% Si [1].
Binary Al-Si alloys have low density, are weldable and resistant to corrosion although
sometimes difficult to machine. Binary alloys often range between 5 to 12 wt%. Some
applications include architectural panels, marine components, cooking utensils, tire
molds and, medical and dental equipment. Besides the binary Al-Si, alloys with
additional elements such as Al-Si-Cu, Al-Si-Mg and Al-Si-Cu-Mg are extremely relevant
for industrial applications (Figure 2.2). The addition of copper to Al-Si results in good
castability, higher strength and hardness, and improved machinability; but it reduces
ductility and resistance to corrosion. Typical applications are in transmission cases,
engine blocks, gear blocks and cases, fuel pumps and cylinder heads. Al-Si-Mg, which
include the very well-known 356.0 and A356.0, have outstanding casting properties and
good corrosion resistance. The remarkable combination of tensile and physical
properties that can be obtained by heat treatments, makes them appealing for aerospace,
machinery, automotive and military applications. Some examples of parts produced by
these alloys are automotive space frames and wheels, pump and compressor bodies,
cylinder heads, impellers or missile bodies.
Finally, in some cases, the addition of both, copper and magnesium, to Al-Si is
advantageous. These alloys have excellent strength and hardness, with some sacrifice in
ductility and corrosion resistance. Optimal properties are achieved after heat treatment.
The alloy most commonly used for pistons in passenger cars and light trucks belongs to
this category (332.0-T5 / Al-9.5 wt% Si-3 wt% Cu- 1 wt% Mg), showing a good
combination of mechanical and physical properties at elevated temperatures including
low thermal expansion. Other applications of Al-Si-Cu-Mg are found in crankcases,
structural aerospace components, air compressor pistons or compressor cases [1].
9
Figure 2.2: Schematic of the Al-Si phase diagram with some of the most common casting
alloys. Adapted from [25].
All these general-purpose alloys may be subjected to modification of the eutectic silicon
phase morphology. Ultimately, the commercial use of these materials depends on the
control of the microstructure of the silicon phase. The detailed study of this modification
is the focus of this thesis and will be further explained in the next sections.
10
3. Eutectic solidification
The microstructure of an alloy is influenced by the solidification conditions and the
composition. Generally speaking, two basic growth morphologies exist during alloy
solidification: dendritic and eutectic [26]. Depending on the composition on a phase
diagram, one can find (a) a pure substance that can solidify in a planar or dendritic
manner; (b) solid-solution dendrites; (c) dendrites with interdendritic eutectic; and (d)
eutectic (Figure 3.1).
The design of casting alloys requires good castability to get small hot tearing, low
shrinkage and good mold filling [1,22]. The best castability is given by pure metals, or by
alloys with near eutectic compositions [26]. Multicomponent systems with an invariant
eutectic point have a long history in the casting of components. These systems offer
relatively low temperature melting when mixing pure elements enabling the fabrication
of near-net shape parts of high performance [27]. Near eutectic compositions have a
short freezing range offering better fluidity than long freezing range.
Figure 3.1: Representation of part of the Al-Cu system between aluminum and θ (Al2Cu).
Letters in the diagram highlight different microstructures depending on the composition:
(a) planar or dendritic solidification of a pure component; (b) solid-solution dendrites;
(c) dendrites plus interdendritic eutectic; (d) eutectic. (Adaptation from [26])
11
During solidification of a binary eutectic, two solid phases form cooperatively from the
liquid. There are numerous morphologies in which eutectic phases may evolve depending
on the growth characteristics of the individual phases [28,29]. Eutectic structures can be
classified depending on the entropy of fusion of the components, which is the difference
of entropy between the liquid and the solid phases at the melting point. Jackson et al
[30] proposed a parameter, α, to evaluate this:
where ξ is the orientation factor defined as the ratio between the number of nearest
neighbors for a growth unit at the solid / liquid interface of the crystal (η) and the
coordination number (ν); ΔHF the enthalpy of fusion (or latent heat of fusion); TM the
melting temperature; R the ideal gas constant and (∆SF/R) the dimensionless entropy of
fusion. According to this factor, related to the roughening transition of a crystal surface,
phases with α > 2 (high entropy of fusion) grow in a faceted manner with an atomically
smooth interface, while phases with α < 2 (low entropy of fusion) grow isotropically
showing no facets and atomically rough interface. Based on this, eutectic structures can
be broadly classified into regular (or normal) and irregular (or anomalous). Regular
eutectics are formed by two non-faceted phases (low entropies of fusion), while irregular
eutectics have one faceted phase with a high entropy of fusion. Aluminum-silicon, for
example, presents an irregular faceted / non-faceted eutectic, with a metallic aluminum-
rich phase with α < 2 (∆SF/R = 1.35) and a faceted silicon α > 2 (∆SF/R = 7.15) [11].
Another difference that distinguishes regular and irregular eutectics is that the former
generally occurs for symmetric phase diagrams with a symmetric eutectic coupled zone
(Figure 3.2 (a)). The coupled zone represents the solidification conditions under which
the two eutectic phases can grow together with similar velocities [31]. For two non-
faceted phases, both phases have a similar undercooling and therefore, the coupled zone
is symmetric. Differently, in an irregular system where one phase is faceted, its growth
and consequently, that of the eutectic phase will need a higher undercooling. Dendrites
of the non-faceted phase can grow faster and they can grow even at eutectic composition.
Because of this reason, pure eutectic microstructures can be obtained only at
hypereutectic compositions forming an asymmetrical coupled zone (Figure 3.2 (b)) [32].
A further finer classification can be done if the volume fraction (VF) of the solute phase is
considered. When the minor phase in a regular eutectic has a VF smaller than 30%, the
structure will be rod-like; and when the VF is higher than that, a lamellar structure will
form. For irregular eutectics, when VF < 30%, the faceted phase grows with a rod-like or
12
fibrous morphology, while for VF > 30%, branched flakes or acicular structures are
generally present. Figure 3.3 shows a schematic of the classification considering
Jackson’s factor (α) and VF [26].
Figure 3.2: Representation of the coupled zones for: (a) eutectic system with regular
structure and (b) eutectic system with irregular structure. Adapted from [32].
Figure 3.3: Schematic of the four broad categories of eutectic structures based on
Jackson’s factor, α; and volume fraction (VFβ). The top part shows the regular structures:
rod (left) and lamellar (right); and at the bottom, the irregular structures: fibrous (left)
and lamellar (right) . Reproduced from [26].
13
In general, rod-like or fibrous structures are formed for small VF of one phase because
the interfacial area decreases with decreasing VF of the fibers, while it is constant for
lamellae. The interfacial area of fibers is smaller than that of lamellae at VF smaller than
30% [26]. These criteria are approximated, and lamellae can also form for lower VF if the
specific interfacial energy is strongly anisotropic. Such is the case of the irregular Al-Si
system, where the silicon phase represents only about 11 % of the eutectic structure, but
still forms a plate-like silicon structure. Irregular eutectics can present a wide range of
morphologies, depending on the solidification conditions [29].
Figure 3.4: Classification of eutectic morphologies as a function of the entropy of fusion

(ΔSα) and volume fraction (VF) for a growth rate of 5 x 10-4 cm/s. Six regions are shown
corresponding to: (1) regular lamellar; (2) regular rod; (3) broken lamellar; (4) irregular;
(5) complex regular; and (6) quasi regular. Reproduced from [28].
The α factor has considerable success in predicting whether a eutectic structure would
grow in a normal or anomalous manner. However, since the eutectic structure grows
from solution at a considerably lower temperature than the melting point of its pure
constituents, the tendency to facet may be higher. Then, the Jackson’s factor can be
recalculated for the growth from solution by replacing the latent heat of fusion of the
separate constituent by the latent heat of fusion of the solid solution and the melting
temperature by the eutectic temperature [33]. Based on this improvement, Croker et al.
14
[28] developed a more detailed classification of eutectic structures. In this approach, an

entropy of solution of 23 J/(mol.K) was found as the transition value between non-
faceting and faceting behavior. Due to the difficulty in calculating the orientation factor
(η/ν) for complex structures, it is advantageous to use this entropy of solution instead of
the α factor. Systems with an entropy of solution ∆S < 23 J/(mol.K) present a normal
growth, while ∆S > 23 J/(mol.K) are anomalous. The structures in each group,
particularly the anomalous, depend on the VF and the growth velocity. Figure 3.4 shows a
schematic of the structures in the different regions for a growth rate of approximately 5 x
10-4 cm/s as presented by Croker et al. [28].
Regions (1) and (2) show regular lamellar structure and rod structure, respectively.
Because of the increased surface energy anisotropy towards higher faceting tendency, the
boundary VF between regions 1 and 2 becomes smaller when ∆S rises.
Anomalous structures show a wider variety of morphologies. They can be broadly divided
into four types:
(3) broken lamellar or sometimes fibrous (VF < 10%);
(4) irregular phases with a number of morphological types which may coexist. This is the
region for the Al-Si eutectic structure.
(5) complex regular, array of regular plates or fibers over small areas and generally
surrounded by a spine. This structure grows with macrofaceted cellular projections at the
solid/liquid interface (VF higher than approximately 20%)
(6) quasi-regular structure. Sheets or fibers of a non-faceted minor phase in a matrix of

the faceted phase
The transitions between regions in figure 3.4 are not sharp and, in the case of anomalous
eutectic structures, they also depend on growth rate. The Al-Si eutectic is an example of
this dependence on solidification conditions.
15
16
4. Eutectic solidification of Al-Si alloys

The Al-Si phase diagram has a simple eutectic point with two solid solution phases:
aluminum (fcc) and silicon (diamond cubic). The eutectic reaction occurs at 12.2 ± 0.1
at% Si and 577 ± 1°C [34]. The maximum solubility of silicon in aluminum at the eutectic
temperature is 1.5 at%, and decreases to 0.05 at% at 300 °C. The solubility of aluminum
in silicon is extremely low at about 0.04 ± 0.02 at% [35]. Figure 4.1 shows the phase
diagram of the Al-Si system as presented by [34] with the metastable extensions of the
liquidus and solidus lines.
Figure 4.1: Equilibrium phase diagram of the Al-Si system with the extensions of the
metastable liquidus and solidus lines. Reproduced from [34].
The formation temperature of the eutectic phase in this system is cooling rate dependent.
At high cooling rates, the eutectic temperature is depressed and the eutectic point is
shifted towards higher silicon concentrations [36]. This behavior is explained by the
presence of an asymmetric coupled zone (introduced in section 3). Since silicon is a
nonmetal that grows anisotropically in a faceted fashion forming directed covalent
17
bonds, it needs a higher undercooling than the non-faceted aluminum phase and a
growth asymmetry arises for changing solidification rates.
The irregular or anomalous Al-Si eutectic system formed by faceted silicon and non-
faceted aluminum can grow with a variety of morphologies depending on the
solidification conditions. Silicon is capable of several crystal growth mechanisms in metal
solutions. These mechanisms were rationalized and classified by Day and Hellawell [5,6].
Figure 4.2 depicts four distinctive regions as a function of temperature gradient (G) and
growth rate (V).
Figure 4.2: Classification of eutectic microstructures in Al-Si alloys as rationalized and

presented by [5]
- Region A: the two eutectic phases grow independently showing a long-range-diffusion

front at high G/V. The solidification front is formed by a planar metal liquid interface
with uncoupled massive silicon crystals projecting forward into the liquid. The silicon
crystals were showed to be interconnected and twinned in {111} planes. Several silicon
particles in this region show elongation in the <110> or <211> orientations.
- Region B: eutectic growth with a lower G/V than region A presenting short-range
diffusion. Silicon shows a variety of morphologies with highly preferred <100> texture.
Two of the most common forms found in this region are thin plates with {100} faces
closed at the growing end by <110> edges; and corrugated crystals with {111} faces with
the axis of corrugation being the <110>. Silicon plates are inclined with a variety of
18
angles that account for twin configurations. For these structures to be stable, the
aluminum must wet the {100} external surfaces up to the aluminum-silicon-melt
junction as shown in figure 4.3 (a).
- Region C: irregular plate-like silicon structure. This structure occurs when the growth
rate is increased above a critical value where a very large undercooling is produced.
Silicon flakes grow faster in the <112> orientation and project ahead of the solidification
interface forming a non-isothermal front. They have {111} growth habit and contain flat
twins across the plates that allow a variety of orientations. This type of growth presents a
wide range of inter-particle spacing as a result of the rigid growth anisotropy.
- Region B+C: shows a gradual transition between <100> texture and the {111} growth
habit. This transition in the morphology of the microstructure is the result of the change
of growth mechanism of the faceting phase. Figure 4.3 shows the <100> texture of region
B with facets on the {111} crystal faces. This type of morphology can occur only if the
metal phase wets the {100} external faces of the silicon crystal up to the solid / liquid
interface, that is, if the growth takes place with an iso-thermal solidification front. For
decreasing G/V, the silicon phase starts to project ahead of the solidification front in a
non-isothermal front and the rapidly growing {100} grow laterally giving rise to a
transition structure. If the G/V decreases even further because of an increase in V, the
irregular structure of region C is formed by close packed {111} faces at the external walls
of the plates and at the growing end.
Figure 4.3: Simplified schematic view of the crystallography of eutectic silicon growing in
regions B and C. Reproduced from [5].
- Region D: for higher freezing velocities in the range of 0.2 to 1 mm /sec., there is a
further transition in the microstructure called sometimes chill- or quench-modification
19
in literature [5,37]. Silicon appears as continuous irregular fibers with rounded cross-
section similar to the modified structure obtained by the addition of impurity elements.
In the case of such a high cooling velocity, it is proposed that the kinetic undercooling
increases sharply and eventually, an inversion at the solidification front occurs such that
the silicon grows behind the metal phase [37]. This change in the front will also affect the
liquid diffusion.
Some differences in the twinning of silicon in this structure has been reported. There is
evidence showing some twinning mostly parallel to silicon growth axes [37,38], other
showing non-faceted silicon with no twins [39,40], and considerable higher twinning
density than in region C [41].
4.1 Plate-like structure and TPRE mechanism

The irregular plate-like structure from region C (Figure 4.4) corresponds to the
solidification conditions most generally found in typical unmodified industrial casting
processes and therefore, this type of growth deserves closer attention.
Figure 4.4: Alloy microstructure of Al-7 wt% Si. (a, b) Optical microscopy images
showing α-aluminum dendrites and plate-like eutectic. (c) Dark field TEM image
showing repeated parallel twinning on {111}Si planes in agreement with the TPRE growth
mechanism.
As shown in figure 4.3 (b) and 4.4 (c), the growth rate of silicon in region C will be
limited by the nucleation rate on the slow growing {111} faces. One way the crystal
increases its nucleation rate is by the presence of coherent twin planes across the silicon
plates. In 1960, Wagner [42], and Hamilton and Siedensticker [43] thoroughly explained
the faceted growth mechanism of germanium crystals in contact with a supercooled melt,
20
which was later also well-accepted for silicon growth. This is called the twin plane re-
entrant edge (TPRE) mechanism and it is based on the formation of twin planes through
the germanium or silicon lamella, generating self-perpetuating grooves that function as
nucleation and growth sites. The growth on grooves is assumed to be rapid because steps
are generated at the center of the grooves. The authors showed that germanium growth
occurs readily if the crystal contains at least two parallel twin planes [42]. Figure 4.5 (a)
shows a crystal with two twin planes bounded by {111} faces at the growth interfaces.
These twinned crystals form re-entrant corners with an angle of 141°, and ridges with an
angle of 219° enabling rapid growth in <211> orientations as shown in figure 4.5 (a).
Considering the three <211> preferred growth orientations, six re-entrant corners are
formed. Figure 4.5 (b) shows examples of nucleation events on two of these corners (sites
I) and the further generation of new re-entrant corners with an angle of 109.5° between
{111} planes (sites II). The simultaneous existence of two self-perpetuating re-entrant
corners ensures the permanent presence of steps for a continuous growth in preferred
sites.
Figure 4.5: Schematic diagram of the twin plane re-entrant edge (TPRE) mechanism. (a)
Re-entrant corners (141°) and ridges (219°) formed by the presence of twin planes. (b)
View of the growth after nucleation in two “type I” sites, which perpetuate nucleation by
the formation of “type II” corners. Reproduced from [43].
Shamsuzzoha et al [44] confirmed the TPRE mechanism and showed that all active {111}
twin planes are cozonal or coplanar, i.e. parallel to a single <110> zone axis (Figure
4.4(c)). The growth of the silicon under these conditions is purely two-dimensional. To
maintain an approximately constant average inter-plate spacing, repeated branching and
direction changes occur by multiple twinning. Kobayashi and Hogan [45] explained
branching by a 70.5° change in direction due to twinning at the bounding {111} plane.
Figure 4.6 (a) shows a representation of a horizontal plate formed by the twinned crystals
A and B, and a branch containing the twinned crystals B and C. In this example, crystal B
commenced the branching nucleating on A, and the C crystals formed subsequently.
21
Repeated branching such as in figure 4.6 (b) allow for changes in the silicon growth in
almost any angle while retaining the <211> direction and the {111} surface planes.
Figure4.6: Schematic representation of the high-angle branching of the eutectic silicon

plate structure. (a) 70.5° branching by the formation of a twinned crystal on a {111}Si
surface plane. (b) Repeated branching by twinning allowing for almost any growth
direction [45].
Shamsuzzoha and Hogan [46] showed a further mechanism for the adjustment of the
inter-plate spacing called “displacement twinning”. Figure 4.7 gives a schematic
representation showing two mutually twinned crystals (C and D) that stop their original
growth while a new twin is formed laterally on the external {111}D face. A crystal on the C
orientation protrudes and growth continues in the <121> direction parallel to the original
by the formation of two further twin events with self-perpetuating grooves. Repeated
displacements result in branching in any arbitrary angle.
Figure 4.7: Schematic representation of displacement twinning [46]
Recently, a thorough analysis of the silicon plate-like growth done by EBSD supports the
TPRE model [47]. However, in this study the authors show that, microscopically, the
22
silicon plates elongate in a <110> direction rather than the <112> assumed in the model.
They argue that a zigzag paired <112> growth from parallel twinning planes result in
<110> growth habit. They explain this by an alternate disappearance and creation of 141°
re-entrants, schematically shown in Figure 4.8.
Figure 4.8: Silicon plate with two twin planes TP1 and TP2. (a) Two 141° re-entrants are
shown between faces 2 and 3, and between 4 and 5, while two 219° ridges complete the
structure between 1 and 4, and between 3 and 6. (b) Further growth causes the
disappearance and subsequent creation of 141° re-entrants that result in a <110> growth
extension. [47]
TPRE mechanism was shown experimentally for germanium crystals [42,48,49]. Figure
4.4 (c) shows several parallel twin boundaries crossing through the silicon plate and
parallel to the outer {111}Si plate bounding planes. This evidence is in favor of the TPRE
mechanism for silicon plate growth. It is however not possible to ensure that this
mechanism is the main responsible for the plate-like growth. Lu and Hellawell [39,40]
for example, showed that the spacing between twins might be too wide to be determining
for the kinetics of molecular attachment and stated that the TPRE mechanism might be
just incidental. Kitamura et al [50] argue that the TPRE mechanism was developed for
perfect crystals without considering the influence of dislocations and therefore, its
relevance is questionable. Although it is not completely elucidated, whether just twins or
the presence of twins plus dislocation is responsible for unmodified silicon growth, in
any case, twins show to readily form and aid the growth of silicon.
23
24
5. Eutectic modification of Al-Si alloys

Al-Si alloys are commercially relevant not only because of their good castability,
strength-to-weight ratio or corrosion resistance, but also because the silicon
microstructure can be modified by the addition of low concentrations of certain elements
[7]. This modification is used to enhance properties such as ductility and toughness in
these alloys and because of this, it was the subject of hundreds of studies since its
discovery in 1921 [8].
5.1 Enhancement of mechanical Properties

Unmodified silicon adopts a plate-like structure for industrial solidifying conditions
(Figure 5.1 (a,c)) [51,52]. The coarse, hard and brittle silicon plates diminish the alloy’s
ductility. The facets of the silicon plates are often on the cleavage plane {111} and
therefore, cracks propagate easily across them [37]. The unmodified alloys’ elongation is
often no more than a few percent and toughness is deteriorated [53,54]. In contrast, the
modified silicon structure shows a fine interconnected coral-like structure formed by
fibers with rounded cross-section (Figure 5.1 (b,d)) [51,52,55]. This transformation of the
silicon structure improves the elongation significantly [7,56,57]. Table 5.1 shows the
modification effect on the tensile yield strength, ultimate tensile strength and elongation
of some representative alloys.
The mechanical properties of Al-Si castings strongly depend on the form, size and
distribution of silicon in its eutectic phase. When comparing modified and unmodified
alloys, the modified structure showed a reduced amount of fractured silicon particles
under the same testing conditions. This is related to the lower aspect ratio and particle
size of modified silicon [54]. Larger and longer particles are more prone to cracking
rapidly at low strains in coarse structures, in contrast to finer structures where the
progression of particles’ cracking is more gradual [58–60]. High sphericity of silicon
particles is favorable to the resistance of interface debonding and plastic deformation of
the aluminum matrix [61]. Fractography shows a direct relationship between the size and
aspect ratio of the silicon particles and the dimples implying that the elongation of silicon
particles is transmitted to the dimples and the cross-section area of the silicon particles
directly influences the fracture surface geometry [62]. The mechanical properties of
hypoeutectic Al-Si castings are also influenced by the grain size and the secondary
25
dendrite arm spacing (SDAS). The combination of eutectic modification together with
grain refiners and small SDAS were shown to exert a significant improvement of tensile
[63–65] and impact properties [53,60,66].
Figure 5.1: Comparison between unmodified and modified microstructure in an Al – 7

wt% Si alloy. Both images show primary α-Al dendrites surrounded by eutectic phase.
(a,c) 2D and 3D images of unmodified eutectic silicon plates in an aluminum matrix.
(b,d) Modified eutectic phase by the addition of 150 wt ppm Sr. 2D and 3D images of the
silicon coral-like structure.
In industrial practice, ductility can also be improved by heat treatment producing a

“thermal modification”. During the heat treatments of Al-Si alloys, the spheroidization of
silicon crystals is a time consuming part of the process. Silicon plates first break up into
smaller parts, then coarsen and become spherical. If the chemical modification studied in
this thesis is combined with a heat treatment, the time needed for the spheroidization
can be reduced in half. The reduction in cost by shortening the heat treatment can be
26
about 10 times the cost of strontium addition and the throughput of the furnaces doubled
[67,68].
Table 5.1: Comparison of mechanical properties for unmodified and modified cast alloys
as presented in reference [69]
Tensile
Ultimate
Alloy and Modification yield Elongatio
Product tensile
temper treatment strength n (%)
strength (MPa)
(MPa)
Sand cast test None … 124 2.0
bars Na-modified … 193 13.00
13 % Si Permanent None … 193 3.6
mold test
Na-modified … 221 8.0
bars
Permanent None … 180 5.5
359.0 mold test
0.07% Sr … 210 12.0
bars
Sand cast test None 208 289 2.0
bars 0.07% Sr 238 293 3.0
356.0-T6 Bars cut from None 213 284 4.4
chilled sand
0.07% Sr 218 291 7.2
casting
A356.0- Sand cast test None 179 226 4.8
T6 bars 0.01% Sr 207 297 8.0
Permanent None … 151 24.0
A444.0-
mold test
T4 0.07% Sr … 149 30.0
bars
None 112 137 1.8
Sand cast test
0.005-0.05%
bars 108 159 8.4
Sr
Permanent None 125 168 6.0
A413.2 mold test 0.005-0.08%
125 191 12.0
bars Sr
Test bar cut 0.05% Sr 121 193 10.6
from auto
0.06% Sr 126 193 12.8
wheel
Apart from the modification of the eutectic microstructure, the addition of strontium was
also reported to have an effect on iron-rich intermetallics. Iron is a common impurity in
commercial alloys which often cannot be avoided. Several iron-rich intermetallics with
different compositions and morphologies such as the needle/plate-like β (Al5FeSi),
Chinese script α (Al8Fe2Si or Al15(Fe,Mn)3Si2), π (Al8Mg3FeSi6) or δ (Al4FeSi2) can be
formed in Al-Si alloys and they are generally detrimental to the mechanical behavior of
the alloy [70–75]. The β-phase (Al5FeSi) is brittle and cracks propagate along the
boundary of its needle morphology [76]. Manganese is often added to the alloy to inhibit
the formation of coarse β phase and replace it by the less harmful α phase which is more
compact, but strontium addition was also reported to lower the volume fraction of Fe-
27
intermetallics and refine the needle-shaped β (Al5FeSi) making it shorter and thinner
[72,76,77]. Samuel et al [78] noted the dissolution of β in the aluminum matrix and an
enhanced effect when magnesium and strontium are added together by the additional
transformation of β to π phase. In wrought aluminum alloys of the 6xxx series, addition
of strontium was reported to promote the formation of the α phase instead of β, leaving
more excess silicon available for the precipitation of Mg2Si during aging and, therefore,
enhancing its mechanical properties [79].
5.2 Elements modifying eutectic silicon and their effect on the

microstructure
Sodium was the first element extensively studied as a modifier [5,9,10,40,80–85]. It has
a powerful effect and only around 50 ppm are enough to show full modification. The
main challenge of adding sodium is its high vapor pressure and consequent tendency to
fade [7,11,40]. This makes the amount of effective added sodium difficult to control and
therefore not so convenient for industrial practice. For this reason, since the 1970s
strontium replaced sodium and its effect on Al-Si was increasingly studied [11,12,38,86–
88]. Strontium is generally added in the form of Al-Sr master alloys [7,89,90] with
almost no loss during solidification. Detailed information on the effect of sodium and
strontium is reviewed in section 5.4.
On 1987, in the search for a systematic that could explain the modification of silicon, Lu
and Hellawell [40] introduced the impurity induced twinning mechanism. This model
considered that elements with an appropriate size would force the silicon growth to miss
one close packed position and create a twin. Based on this assumption, several elements
such as barium, calcium, yttrium and most rare-earths were proposed as potential
modifiers. Since then, several investigations were carried out on the effect of these
elements in Al-Si alloys. Calcium was found to modify the silicon structure but with a
weaker effect than sodium or strontium [14–16,19,91–94]. Ludwig et al [14,15,94]
concluded that a fully coral silicon structure cannot be achieved with calcium additions of
up to 300 ppm, but a strong refinement of the silicon plates is observed. Besides silicon
refinement, calcium was found to refine detrimental iron intermetallics in Al-Si alloys
[95–97] and to improve elongation and impact strength, but often deteriorate ultimate
tensile strength due to porosity and formation of Al-Si-Ca intermetallics. Abdollahi and
Gruzleski [98], and Ludwig et al [94] observed that silicon modification with calcium
only refines the silicon plate structure in alloys solidified with low cooling rates and
achieves fibrous silicon for higher cooling rates.
28
Barium addition can produce a full modification of the silicon plates [16,19,40,91].
However, a relatively high amount of modifier is needed in comparison to sodium or
strontium. Knuutinen et al [16] showed that full modification needs additions of at least
890 ppm Ba to occur, while lower additions produce only partial modification or only
refinement of the silicon plates.
Rare earths were also tested as possible modifiers [20]. From all of them, only europium
showed the ability to transform the silicon plates into corals, while all others showed only
a refinement of the silicon plates (Figure 5.2). Besides, much lower concentrations of
europium (~ 600 ppm in an Al-10 wt% Si) are enough for silicon modification compared
to the amount of the other rare-earths elements needed just to refine the structure (1.3 to
3.0 wt%) [20]. Alloys modified by europium show similar characteristics as strontium
modified alloys. Some of the characteristics of this modification are: high density of
crystallographic defects in silicon, formation of pre-eutectic and eutectic Al2Si2Eu
intermetallic phases, and improved ultimate tensile strength and elongation [13,56,99].
The question that remains unanswered is: why do the other rare earths only refine
eutectic silicon and cannot form silicon corals. Similarly to europium, alloys with
ytterbium also form Al2Si2Yb phases [18]. However, Nogita et al [17] showed by micro X-
ray fluorescence (µ-XRF) that, while europium is present in the silicon phase like in the
case of strontium [100], ytterbium is not. Ytterbium was tested in a wide range of
concentrations from 0 to 6400 ppm in an A356.0 alloy and every composition showed
only an increasingly refined silicon plate structure [16]. The absence of ytterbium in the
silicon phase was pointed as one possible explanation for the lower density of twins in
silicon [18].
29
Figure 5.2: Optical micrographs of Al-Si alloys with addition corresponding to the
maximum modification for each element: (a) unmodified, (b) 17300 ppm La, (c) 19700
ppm Ce, (d) 17200 ppm Pr, (e) 14800 ppm Nd, (f) 19700 ppm Sm, (g) 600 ppm Eu, (h)
30300 ppm Gd, (i) 25300 ppm Tb, (j) 14300 ppm Dy, (k) 18700 ppm Ho, (l) 17900 ppm
Er, (m) 12900 ppm Tm, (n) 13900 ppm Yb, (o) 12700 ppm Lu [20].
30
5.3 Porosity and shrinkage

A downside of the addition of a modifying agent to Al-Si castings is the alteration of the
amount, size and distribution of porosity. A considerable percentage of scrap is caused by
undesirable casting defects such as porosity and shrinkage. Low elongation, loss of
tensile strength or poor fatigue strength are some of the problems related to porosity and
/ or shrinkage [101]. An increased level of porosity is often reported for modified alloys
and several approaches to find an explanation to this behavior can be found in literature
[102–106]. A decrease of hydrogen solubility in the liquid was reported in strontium
modified alloys [107]; degassing the melt after strontium addition was suggested to
decrease the content of hydrogen [108]. Emadi et al [107,109] showed lower surface
tension and higher viscosity of the liquid, as well as a higher volume shrinkage due to
increased solid density. Inclusions such as strontium oxides (Al2OSr3) [105] and
strontium-containing intermetallics (Al2Si2Sr) [110] were suggested to act as sites for
heterogeneous nucleation of pores.
In alloys with a significant volume fraction of eutectic phase, porosity was showed to be
redistributed from centrally located irregular and interconnected pores in unmodified
alloys, to well dispersed and rounded pores in strontium and sodium modified alloys
[103,104,111]. The redistribution of the porosity was explained by an alteration of the
solidification mode. As it will be presented in more detail in section 5.4.3, eutectic grains
in hypoeutectic unmodified alloys, which nucleate near the tip of primary dendrites, are
very numerous and small; in contrast, few and very large eutectic grains nucleate
independently from the dendrites in modified alloys. An “isolation” model was proposed
by McDonald et al [103] to account for the effect of the solidification mode on porosity.
Pores are showed to form after a significant amount of solidification has occurred. The
authors argue that during solidification, liquid pools will become trapped and isolated
from the riser when eutectic grains impinge each other. After isolation, no feeding will
occur and solidification shrinkage leads to pore formation. In the unmodified alloy, the
massive number and small size of the eutectic grains will allow a good permeability of the
liquid until the last stage of solidification. Very small pools may form, but porosity will
mainly concentrate at the center of the casting. In contrast, the dramatic decrease in
number and increase in size of the eutectic grains in the modified alloys, cause the
isolation of well distributed relatively large pools of liquid where rounded pores will form
due to shrinkage. The impact of solidification mode on interdendritic flow was also
evaluated for additions of barium, calcium, yttrium and ytterbium to Al-Si alloys [102].
Significant differences were found in the distribution of porosity in the castings that
support the effect of solidification modes on the permeability of the mushy zone. On the
31
one hand, barium and ytterbium showed nucleation of independent eutectic grains that
reduced the flow paths causing small, rounded and well distributed pores similarly to
strontium. While on the other hand, the eutectic phase with calcium and yttrium evolved
from the surface of the casting towards the center of the hot spot and the freezing of the
central path caused the formation of porosity in the center such as in unmodified alloys.
The amount of silicon plays a role on porosity in hypoeutectic and eutectic alloys. Higher
silicon contents decrease the volume fraction of primary aluminum dendrites, leaving
more space for feeding during a longer time. In this sense, near-eutectic Al-12.5 wt% Si
alloy shows less porosity than Al- 7 wt% Si [106]. The combination of casting design
together with the feeding efficiency, given by the solidification modes and the silicon
content, influences the amount of porosity in these alloys [111]. Based on this knowledge,
casting design can be used to promote directional solidification towards the feeder in
order to get sounder castings of increased integrity [102,106].
5.4 Theories explaining eutectic modification of Al-Si alloys

Several conflicting theories have been proposed through the years showing the
complexity and controversy of the subject [112]. In spite of 100 years of research, the
underlying phenomena for modification are still not fully understood.
5.4.1 Early theories for eutectic modification
Shortly after the discovery of the eutectic modification in 1921 [8], one of the first
explanations considered the ternary system as a key factor. Otani [80] suggested a
ternary monotectic reaction in alloys modified by sodium. He proposed that, during
solidification, aluminum and silicon solid phases are formed simultaneously with a small
amount of a β-liquid rich in sodium. If this very small amount of β-liquid is just ignored,
the invariant reaction may be taken as equivalent to the binary eutectic reaction and, as
the reaction proceeds, the β-liquid gradually accumulates contiguous to the crystal
hindering the free growth. Also during the 1920s, Gwyer and Phillips [113] suggested a
condition of transient stability, where formation of aggregates of colloidal dimensions in
the nanometer range would play an important role. Ransley and Neufeld [81] corrected
the Al-Si-Na phase diagram to show that the system had a ternary eutectic instead of a
monotectic reaction and found the formation of a ternary compound [AlSiNa]. At that
time, they pointed out that the idea of a colloidal dispersion would provide an adequate
explanation if it were assumed that this [AlSiNa] compound is the hindering agent. Later
development of the ternary phase diagram of the Al-Si-Na [114] and Al-Si-Sr [115]
32
systems confirmed ternary eutectic reactions forming ternary compounds instead of a β-

liquid.
In 1949, Thall and Chalmers [82] proposed a model to explain eutectic modification
based on the balance of interfacial tensions during solidification. They assumed that
aluminum is the leading phase during eutectic solidification. This assumption was based
on the statement that the rate of advance of the solidification interface depends on the
balance between the heat flow from the liquid to the solid and the latent heat evolved
during solidification. Based on this, the higher thermal conductivity and the lower the
latent heat of fusion, the faster the interface will advance. Because aluminum has
considerably higher thermal conductivity and lower latent heat than silicon, it was
concluded that the aluminum / liquid interface would freeze in advance more rapidly
than the silicon / liquid interface. The second assumption made by the authors is that a
small amount of added sodium reduces the surface energy between solid aluminum and
solid silicon, which in turn increases the interfacial angle θ (Figure 5.3). They suggested
that this reduction of surface energy causes the aluminum phase growing ahead of the
solidification front to enclose the silicon crystal and, therefore, the silicon particle will be
“sealed off” from the liquid and will need to re-nucleate repeatedly. Despite the logical
train of thought of the authors, it was later showed that the silicon phase is the leading
phase during eutectic solidification and silicon grows in a continuous manner instead of
constantly re-nucleating [5,116,117].
Figure5.3: Eutectic growth proposed by Thall and Chalmers: (a) Alloy with no
modification. Aluminum grows in advanced of silicon. Interfacial angle θ between the
growing aluminum and silicon phases. (b) Modified alloy with a reduced surface energy
between the phases that translates into a larger interfacial angle θ. (c) Because of the
reduced surface energy, aluminum encloses silicon which will need to re-nucleate [82].
Following the idea of Thall and Chalmers [82], Davies and West [9,84] studied interfacial
energies in deeper detail. The interfacial angle between solid silicon and the melt was
shown to be equal to the grain boundary energy in silicon in an unmodified alloy. For the
sodium modified melt, no typical angles were found and the angle depended on the
crystallographic orientation of the silicon substrate. This showed that, even though a
33
lower surface tension was found for sodium modified alloy in comparison to unmodified
[84], the change in interfacial energy cannot determine the interface profile during
solidification. The authors claimed that sodium modifies the crystal habit of silicon by
being incorporated into certain crystallographic facets. They showed the restriction of
fast-growing crystal faces in the presence of sodium (Figure 5.4), which reduces the
growth rate of the eutectic phase to that of the slowest facets in silicon. To the best of our
knowledge, Davies and West [9] were the first to propose the progressive poisoning of
silicon facets by sodium atoms, explaining the supercooling of the melt as the increased
driving force needed by the crystal to further grow. They also concluded that the
modification is found for elements with sufficient affinity for silicon to form ternary
compounds confirming the early idea [80,81] that the ternary system plays a crucial role
for modification.
Figure 5.4: Dihedral angles at the interface between polycrystalline silicon and eutectic
(Dash etch). (a) Grain boundary without sodium. (b) Twin boundaries without sodium.
(c) Grain boundary and twins in presence of sodium. (d) Silicon growth layer with
sodium. Apparent restriction to certain crystal faces [9].1
Around 1970, Day and Hellawell [5,6] agreed with the proposal of the poisoning of silicon
by sodium reporting that sodium poisons re-entrant {111} twin grooves characteristic for
1Magnifications reported in the original publication are: (a) x450, (b) x450, (c) x450, (d)
x240. Note possible small deviations from these values due to print size.
34
the growth of plate-like silicon in region C (Figure 4.2). They explained that the selective
adsorption of sodium retards silicon growth and increases undercooling.
From the early theories mentioned above, two hypotheses stand out:
(i) the ternary system Al-Si-X (X = modifier) and the formation of ternary
phase/s play an important role to understand modification; and
(ii) the modifier restricts silicon growth by poisoning certain crystallographic
facets.
In the following decades, hypothesis (i) was abandoned because of lack of proof and great
attention was put on (ii) to understand silicon growth and its restriction. Next section
presents a brief summary on this topic.
5.4.2 Modification effect on eutectic growth
Similar to Davies and West [9], and Day and Hellawell [5,6], several other studies based
their models and explanations on the silicon growth restriction due to the presence of the
modifier at the growth front [11,118–120]. During the 1970s and 1980s, even though
some studies still used sodium to modify Al-Si alloys, more and more investigations
focused on strontium modification because of strontium’s low tendency to volatilize.
Strontium was proven to segregate preferentially into silicon in the eutectic phase [87].
Steen and Hellawell [118] analyzed eutectic growth and reported that strontium
interfered with facet formation on {111} silicon planes and increased the frequency of
twinning events. They suggested that the effect of sodium and strontium is to inhibit the
growth on planes other than {111} by selective adsorption of the impurity element
poisoning other faces. This situation leads to faceting of other planes implying that the
original plane is no longer the most slowly growing and crystal growth becomes more
isotropic. Hanna et al [119] suggested that the inhibition of molecular attachment on the
most rapidly growing sites would increase the undercooling until growth could occur in
less advantageous planes.
Jenkinson and Hogan [11] as well as Steen and Hellawell [118] concluded that
modification occurs due to a limited diffusion process at the solidification front, caused
by accumulation of the impurity element at the solid-liquid interphase up to a
concentration where the undercooling and growth profile are affected. If the solubility of
silicon in the melt is reduced, the attachment of silicon to the crystal surface is also
reduced until silicon and aluminum growth velocities are equalized and both phases form
by an isothermal continuous growth.
35
The isotropic growth front in modified alloys was shown in several studies [10,85,86,119]
by comparing the shape of the solidification front in unmodified and sodium-modified
eutectics. While unmodified eutectic has an irregular solidification front with silicon
growing ahead, the sodium-modified alloy shows a very smooth one. In 1981, Flood and
Hunt [85] emphasized that not only the growth of silicon was affected by the addition of
sodium, but also the nucleation. From the imaging of quenched samples, these authors
arrived at the conclusion that in an unmodified alloy, nucleation can occur on dendrites
ahead of the main eutectic front, but when sodium is added these nucleation events are
inhibited and growth occurs in a smooth front (Figure 5.5). Some years later, Hanna et al
[119] polished successive layers on similarly quenched samples and showed that, in the
unmodified alloy, eutectic colonies do not necessarily originate as nucleation events in
the liquid, but are connected to the main growth front by outlying extensions of flake
silicon. Although the conclusion of a restricted nucleation given by Flood and Hunt [85]
was rebutted, their analysis still holds. In their study, the growth front of the sodium
modified alloy was 18 times smaller than the growth front of the unmodified alloy.
Because the rate of heat extraction from the solidification front should be similar for both
alloys, the average interface velocity of the modified alloy must be much greater. The
authors suggested that the higher interface velocity due to the smaller interface area,
accounted for the refinement of the microstructure and consequent smaller spacing
between silicon branches.
Figure 5.5: Solid / liquid interfaces after quenching molds during solidification: (a)
Unmodified Al-Si eutectic alloy; (b) Sodium modified Al-Si eutectic alloy; (c) solid line
showing the solid / liquid interface in (a) and dashed line showing the expected smooth
interface [85].
The reason for the change in morphology of the solidification front remained unclear and
the next step forward to understand eutectic modification came with the development
36
and increasing popularity of transmission electron microscopy (TEM). Modified silicon

revealed a remarkably imperfect crystallography with a massive increase in the twin
density [12]. TEM imaging and diffraction were extensively used to study the presence of
twins and to understand whether they are important or incidental for the growth of
silicon and the modification process [38,39,88,121]. Frequent twinning in the silicon
crystal was proposed to enable an isotropic growth by allowing repeated changes of
growth orientation [12,38–40,88].
Shamsuzzoha and Hogan [12,38,88], and Jenkinson and Hogan [11] performed TEM
studies on silicon fibers modified by strontium and showed that modified silicon has
fiber axes in the <100> and <110> orientations and exhibit multiple {111} twinning from
at least two {111} planes simultaneously. The high twin population allows frequent
branching and flexible growth. This permits the axes of the corals to change and
maintain the necessary inter-coral distance. Similarly, Lu and Hellawell [39,40] showed
that silicon in a sodium modified alloy is heavily twinned with the silicon fibers having a
principal axis along a <100> direction, with the four twin systems involved and side
branches extending in <211> directions at an angle of 35.3° to the principal axis. They
observed that modified fibers have rough or microfaceted external surfaces, and that
multiple steps and grooves are expected at the solid / liquid front as twins emerge.
Shamsuzzoha and Hogan [12] postulated that the increased twin density is the result of
sodium or strontium adsorption on {111} Si surfaces lowering the {111} twin energy. Based
on TEM images, they proposed a zigzag growth model. Figure 5.6 (a) shows how the local
growth directions [ 2]A and [112]B are repeatedly alternated with [112]C and [ 2]B. If
twins are nucleated such that each growth segment is of equal length, the resultant fiber
axis will be [00 ] in terms of the common B orientation. Figure 5.6 (b) shows a similar
case but in this case the fiber axis is [110] in terms of A. The difference between figure 5.6
(a) and (b) is given by the twin alternative configurations forming angles of 70.5° and
109.5° between the crystals, respectively. The <100> and <110> fiber orientations are
extreme cases. The fiber axis can vary anywhere between <100> to <112> to <110> if the
lengths of the crystal segments are unequal. Based on this model, Shamsuzzoha and
Hogan [12] also explained the branching of the fibers as an extension of the same
mechanism with the twin configuration governed by the local diffusion conditions.
37
Figure 5.6: Schematic representation of the “zigzag” growth in a strontium modified
alloy. Orientation relationships in a fiber with effective [00 ] (a)and [110] (b) growth
axes [12]
Following the idea of the adsorption of the modifying element on {111}Si facets, in 1987,
Lu and Hellawell [40] proposed the impurity induced twinning (IIT) model. In this
model, the authors looked for a common trait that could explain which elements modify
the eutectic silicon. They considered in detail the possible relevance of the atomic radius
of the modifying agent by calculating what would be a suitable size of the impurity atom
to change the {111} stacking sequence and promote the formation of a twin when
adsorbed on the solid-liquid interface. The authors assumed silicon growth by extension
of step sources across the {111} planes. The modifier “poisons” the growth by the
accumulation of adatoms at those steps. Figure 5.7 shows a 011 projection where an
impurity atom of appropriate size forces the next silicon atom to miss one regular close
packed position and so, continue in the alternative stacking sequence creating a twin.
The geometrical calculation of the necessary hard sphere size to fulfil this condition
resulted in a radius ratio of r(modifier)/r(silicon) ≈ .646. Table 5.2 shows a line between
ytterbium (Yb) and lanthanum (La) for this “ideal” ratio and a cross (x) for elements that
were reported to have some kind of influence on the eutectic phase.
At the time when the impurity induced twinning (IIT) model was proposed, there were
no characterization methods with enough chemical resolution to prove the hypothesis by
direct observation. This model was considered a valid explanation for the formation of
twin in the silicon phase although there were incongruencies with experimental evidence.
Sodium, for example, is much more effective than calcium [14,15] although it has a larger
difference from the “ideal”; and ytterbium and lanthanum only refine the silicon plates
instead of forming the beneficial coral structure seen with sodium or strontium additions
[16–18,20]. Factors other than the atomic radius -such as melting points, vapor
38
pressures or oxide formation-, were also qualitatively considered by Lu and Hellawell

[40], but they were rather inconclusive.
Table 5.2: Possible modifying additions in order of atomic radii as proposed by Lu and
Hellawell [40]
Figure 5.7: Schematic showing the impurity induced twinning model on a 011 plane
projection. Impurity atom (modifying agent) with an appropriate size would promote the
formation of a twin by forcing silicon to grow on the alternative {111} stacking sequence
[40].
5.4.3 Modification effect on eutectic nucleation
The addition of a modifying element also influences the nucleation of the eutectic phase
in Al-Si alloys. The study of the cooling curves during solidification shows a depression of
the nucleation temperature (TN), minimum temperature before recalescence (TMin) and
growth temperature (TG) in modified alloys (Figure 5.8) [119,122–124]. Thermal analysis
39
is routinely used to study the effect of modification on nucleation and growth enabling
the identification of phase formation. This analysis is made by the combination of the
cooling curve (temperature over time) and its first derivative showing the solidification
rate of the alloy. The nucleation temperature (TN) of the eutectic phase is defined as the
point where the first silicon crystals are formed and start to grow releasing latent heat.
This temperature is identified by a change in the slope of the cooling curve that can be
clearly identified by the first derivative of the cooling curve [16]. The minimum
temperature (TMin) shows the point where the latent heat due to eutectic growth equals
the heat flow out of the system. This depends on the cooling rate and the heat capacity of
the system [18]. After this point, the release of latent heat is higher than the heat
extraction from the system and the temperature rises until the growth temperature is
reached (TG). The growth temperature corresponds to the nearly horizontal part of the
eutectic arrest and is defined as the maximum reaction temperature after recalescence
[123]. The difference ∆T = TG - TMin is known as the recalescence of the eutectic arrest.
Thall and Chalmers [82], and Hanna et al [119] showed depressions of ≈6 K in the
nucleation and growth temperatures of sodium modified alloys. They also showed a
depression of primary silicon nucleation in modified hyper-eutectic alloys, while primary
aluminum remains unchanged. They concluded that the depression in the nucleation of
the eutectic phase is predominantly related to a change in the nucleation of the silicon
phase. Another distinctive characteristic of the modified alloy is that, while the eutectic
arrest is depressed, the melting temperature remains almost unaffected [82,83,119]. The
difference between melting and freezing temperatures was shown to be dependent on the
amount of modifier in the sample. Higher concentrations of the modifier in an alloy
result in higher differences between melting and freezing temperatures [83]. Thall and
Chalmers [82] pointed out that when the modifier content was very low, the eutectic
solidification took place in two stages: the first at the normal temperature and the second
at a lower temperature characteristic of a modified alloy. Several studies have confirmed
the depression of the eutectic reaction temperatures for additions of strontium [100,122–
124], calcium [14–16], barium [16], europium [13,17,56], yttrium [16] and rare earth
metals [16–18,20]
40
Figure 5.8: Comparison of cooling curves from unmodified and Sr-modified hypoeutectic
Al-10 wt% Si alloys with commercial purity. The primary aluminum reaction presents
almost no change, while the eutectic reaction is depressed to lower TN, TMin and TG and
higher recalescence [123]
Thermal analyses of Al-Si alloys clearly depict a retarding effect of the eutectic nucleation
by the addition of modifying elements. However, for a complete analysis of the eutectic
nucleation, it is essential to consider the role of phosphorus. Phosphorus is deliberately
added in hypereutectic Al-Si alloys to refine primary silicon by an increased silicon
nucleation on aluminum phosphide (AlP) [7,125,126]. Contrarily, in the eutectic phase,
silicon is coarsened in the presence of phosphorous [123,127,128]. Phosphorus content
also affects the modification of the eutectic silicon and larger modifier concentrations are
needed in alloys with higher phosphorus levels [7,125]. Thus, the control of the
phosphorus content is a major key to ensure the reproducibility of the modification
treatment.
The Al-Si-P ternary system and particularly the L → (Al) + Si + AlP ternary reaction was
calculated showing a phosphorus concentration threshold value between zero and 7.4
ppm for all hypoeutectic compositions [129,130]. For Al-7 wt% Si-P with phosphorus
contents higher than 3.7 ppm, AlP forms prior to eutectic silicon acting as nucleation
sites. For phosphorus contents lower than 3.7 ppm this is not possible, with silicon
starting its precipitation before AlP [131]. Ludwig et al [128] showed that increasing the
phosphorus concentration from 2 to 3 ppm resulted in a transition from nucleation on
the mold walls with planar front and a strong dependence of the thermal gradient, to
nucleation in the vicinity of the primary aluminum dendrites with a dramatic increase in
the number of grains.
41
Because the threshold level of phosphorus that exert a significant difference in eutectic
nucleation is extremely low and the presence of further impurities also play a role on
nucleation, the entrained droplet experiment is often used to study nucleation kinetics in
these alloys [13,127,132–134]. In this setup, impurities can be isolated and clean
heterogeneous nucleation can be studied. The technique consists of two stages: first, the
alloy is rapidly solidified by melt spinning to produce a microstructure consisting of
finely dispersed eutectic droplets embedded in a higher melting point aluminum matrix;
and subsequently, the low melting point eutectic droplets are re-melted and slowly
cooled in a thermal analyzer to record the solidification characteristics of the eutectic
phase. The advantage of this setup is that impurities segregate into a small number of
droplets, leaving a large number of droplets where clean nucleation conditions can be
studied. The nucleation undercooling can then be quantified by differential scanning
calorimetry (DSC). A DSC is a thermal analysis technique that shows the change of heat
capacity of a material as a function of temperature. In the case of Al-Si eutectic, the DSC
curve shows normally two exotherms: one first sharp peak that corresponds to the
solidification of eutectic droplets at the grain boundaries, and a second broader peak at
lower temperature that corresponds to the entrained droplets. The undercooling of this
second exotherm is related to the ability of heterogeneous nucleation in clean conditions.
Ho and Cantor [127] used this method in 1995 to show a strongly increased undercooling
when the content of phosphorus was decreased to less than 2 ppm in Al-3 wt% Si.
Commercial purity alloys (up to 10 ppm P) showed centrally located particles containing
aluminum and phosphorus consistent with AlP and increasingly higher nucleation
undercoolings with decreasing amount of phosphorus [122,124,128]. Insufficient number
of AlP particles to effectively nucleate silicon heterogeneously causes a strong reduction
in the number of eutectic grains and considerable refinement of the microstructure
(Figure 5.9 (a,b)) [123,127]. The refinement is explained by the higher solidification front
interface velocity derived from the low number of nucleation events. Given a constant
rate of heat extraction, the inverse relationship between the solid-liquid interface velocity
and surface area of the eutectic grains results in a finer eutectic structure [122,123,128].
The addition of modifying elements to commercially pure alloys was shown to have a
comparable effect than decreasing the phosphorus content (comparison of Figure 5.9 (a,
b and c)). In alloys with sodium addition, DSC analyses showed increased undercooling
for eutectic nucleation. A possible explanation is the formation of the competing phase
Na3P that consumes phosphorus, decreasing AlP nucleation sites. Similarly, other
competing phases were proposed in alloys with other modifiers such as Sr3P2 [133,134] or
YbP [18]. However, no direct evidence of these phases was ever shown.
42
The effect of strontium was extensively studied in literature depicting a density of

eutectic grains two orders of magnitude lower than in the equivalent unmodified alloy
[122,124] and the size of the eutectic grains at least one order of magnitude larger
[123,135]. The local structure of strontium was shown to be that of Al2Si2Sr by means of
synchrotron based extended X-ray absorption fine structure (EXAFS) spectroscopy [136].
Several studies proposed the poisoning or removal of AlP by the formation of Al2Si2Sr
[122,124]. Pre-eutectic Al2Si2Sr surrounding a P-rich particle was shown by EDX analysis
on a TEM sample [137]. The formation of pre-eutectic Al2Si2Sr was corroborated by DSC
and Scheil simulation with Thermo-Calc [133,134]. Eiken et al [131] showed a critical
threshold of strontium of 80 ppm in hypoeutectic alloys with 7 wt% Si and 5 ppm P. If
the strontium content is lower than 80 ppm, AlP solidifies before eutectic silicon.
However, for higher strontium contents, the Al2Si2Sr phase precipitates before eutectic
silicon and after AlP, leaving a solidification sequence L → (Al) → AlP → Al2Si2Sr → Si.
This calculation agrees with the postulate that Al2Si2Sr can nucleate on AlP particles
reducing silicon’s nucleation potential. DSC analysis of Al-5 wt% Si modified with 500
ppm Eu also showed the formation of pre-eutectic Al2Si2Eu phases suggesting the
possibility of poisoning of AlP also in this system [13].
As an alternative for AlP nucleation sites of the eutectic phase, Shankar et al [138]
proposed that eutectic silicon nucleates on β-(Al,Si,Fe) phases forming ahead of primary
aluminum dendrites. The addition of strontium increases the viscosity of the eutectic
liquid ahead of the α-aluminum dendrites [139]. Since the viscosity is proportional to the
surface tension, the authors argue that the presence of strontium changes the wetting
angle between the eutectic liquid and the β-(Al,Si,Fe) particles. If the viscosity is
sufficiently high, the eutectic phases cannot nucleate at the eutectic temperature and
significant undercooling of the melt occurs. However, Lu and Dahle [72]and Li et al [134]
could not find evidences suggesting that β needles nucleate eutectic silicon. On the
contrary, it was proven by Cho et al [137] that both, pre-eutectic β-Al5FeSi and eutectic
silicon, nucleate on phosphorus-rich nuclei. Higher phosphorus contents cause an
increase in the nucleation temperature of β-Al5FeSi suggesting that this iron-rich
intermetallic nucleates at lower undercooling because of the larger number of
phosphorus-based nuclei. The Al-Si eutectic nucleation frequency decreases further
when the iron concentration in the alloy increases. This means that the AlP particles are
consumed by both, Al2Si2Sr and β-Al5FeSi, before eutectic solidification. Shankar et al
[139] showed a depression of the eutectic nucleation temperature for higher iron content
that supports these results. The inter-relationship between strontium and iron on
eutectic nucleation events shows that eutectic growth undercooling is not an exclusive
characteristic of strontium modification, but is strongly related to the presence of
43
additional impurities in commercial alloys as it was stated before by McDonald et al
[123] based on the analysis of cooling curves.
Figure 5.9: Eutectic microstructures: unmodified alloys with (a) commercial purity, and
(b) high purity; strontium modified alloys with (c) commercial purity, and (d) high purity
[123]
The decrease in eutectic nucleation frequency is accompanied by changes in the

nucleation mode. Dahle et al [140] showed three solidification modes (figure 5.10): (a)
nucleation at or adjacent to the wall, and growth opposite to thermal gradient; (b)
nucleation of the eutectic phase on primary aluminum dendrites; and (c) heterogeneous
eutectic nucleation on nucleant-particles in the interdendritic liquid. Quenched samples
showed growth type (b) in hypoeutectic unmodified alloys with the eutectic phase formed
close to the tips of the dendrites (figure 5.10(b)). In this case, primary and eutectic
aluminum had the same crystallographic orientation [122,124,135,140,141]. In contrast,
alloys with 70 and 110 ppm Sr showed eutectic grains starting solidification at the center
of interdendritic channels in an “independent” nucleation as depicted in figure 5.10 (c)
44
[122,124,135,140]. Small eutectic aluminum grains in these modified samples showed no

orientation relationship to the surrounding aluminum dendrites [141]. Interestingly, a
sample modified with 500 ppm Sr exhibited growth associated to the dendrites, like
unmodified alloys (figure 5.10 (b)). The authors argue that strontium containing particles
such as Al2Si2Sr in samples with a low strontium level, might promote independent
nucleation, while these particles become too large at higher strontium levels causing the
reversion in nucleation mode [140].
Figure 5.10: Possible eutectic growth modes in Al-Si alloys (eutectic phase illustrated in
black): (a) front growth opposite to the thermal gradient; (b) nucleation on primary
aluminum dendrites; and (c) independent heterogeneous nucleation in interdendritic
spaces. [140]
In summary, it has been shown by several studies, that the nucleation of the eutectic
phase is inhibited by the addition of a modifying element. Considering the following
equation:
where q is the total rate of heat extraction, the average interface velocity, A the
solidification interface area and L the latent heat of fusion per unit volume. If fewer
eutectic grains are nucleated in a modified alloy, the solid / liquid interface area (A) will
be smaller than in the unmodified case. In order to maintain a constant rate of heat
extraction (q), the interface velocity ( ) of the modified alloy will increase to compensate
for the decrease in area (A). The higher interface velocity has been accounted in literature
for the refinement of the structure inside the eutectic grains. Several elements, for
example all rare earths (except for europium) only refine the eutectic microstructure but
are not able to produce the coral-like growth of silicon [20]. The plate-to-coral transition
45
is believed to be independent of the change in the nucleation mode [123,131,140] and it is
rather a consequence of the additional effect of the modifier during growth.
5.4.4. Recent developments
Efforts to reveal the distribution of the modifier elements in the eutectic phase show sub-
micrometer strontium-rich intermetallics (most likely Al2Si2Sr) adjacent to eutectic
silicon crystals [100,142]. A relatively homogeneous distribution of strontium and
sodium is shown to be exclusively incorporated into the silicon phase [100,142,143].
Although these studies were able to confirm the relevance of the modifier in the silicon
phase, the reported lateral resolutions of about 50 nm are not enough to disclose how
crystallographic defects in this phase are formed.
The first APT results showing the local distribution of the modifying element in the Al-Si
eutectic showed segregations of strontium together with aluminum and silicon [144–
147]. These studies showed segregations with varying compositions of 2.6-2.8 at% Sr,
6.2-11 at% Al and 86.4-91 at% Si, and proposed the formation of SrAl4Si33 or SrAl2Si88
intermetallic phases to cause the multiplication of twins and restrict silicon growth.
These phases have never been shown for the Al-Si-Sr system, although several ternary
phases were calculated [148]. Nevertheless, these early investigations proved the
potential of APT to study eutectic modification.
46
6. Sample preparation and Characterization

A combination of characterization techniques has been used in this thesis for the study of
the eutectic modification in Al-Si alloys. This chapter gives a basic description of the
sample preparation and the methods used including: mechanical polishing, SEM, EDX,
EBSD, FIB, (S)TEM and APT.
6.1 Al-Si alloys

The compositions of the analyzed alloys are summarized in Table 6.1.
Table 6.1: List of the alloys analyzed in this thesis.

Modification Si (wt%) X = Sr, Na, Eu, Al
Yb (wt. ppm)
Unmodified 7 -- Balance Directional
solidification
Unmodified 5 -- Balance Conventional casting
Strontium 7 150 Balance Directional
solidification
Sodium 5 19 Balance Conventional casting
Sodium 5 160 Balance Conventional casting
Europium 5 500 Balance Conventional casting
Ytterbium 5 6100 Balance Conventional casting
6.1.1 Al – 7 wt% Si
Al-7 wt.% Si with and without 150 wt. ppm Sr were produced at the Foundry Institute at
RWTH Aachen by directional solidification in a Bridgman furnace at a cooling rate of
0.25 K/s and a temperature gradient of 15 K/mm. High purity silicon and aluminum
(Al5N5) were melted at the corresponding proportions. An aluminum master alloy
containing 15 wt% Sr was added to modify the samples with 150 wt. ppm Sr.
6.1.2 Al – 5 wt% Si
Al – 5 wt% Si with no additions and with 19 wt. ppm Na, 160 wt. ppm Na, 500 wt. ppm
Eu and 6100 wt. ppm Yb were manufactured with high purity silicon chips (5N) and Al
(5N). Sodium was added using elemental sodium (2N8, 99.8) supplied in vacuum-
47
packed aluminum foils. Europium and Ytterbium were added by the aluminum master
alloys Al - 2 wt% Eu and Al – 5 wt% Yb.
Conventional casting samples were produced by electric resistance melting of the charge
material in a boron-nitride coated clay-graphite crucible at 720 °C - 750 °C and then
casted using gravity die casting. Prior to casting, no degassing treatment was performed.
The average cooling rate was about 20 K / min (20 C / min).
6.2 Scanning electron microscopy – SEM

Scanning electron microscopy (SEM) is a standard technique to characterize alloys from
nanometer to micrometer length scales. The sample is irradiated with a high energy (1 –
30kV) focused electron beam that interacts with the electrons of the atoms in the sample
analyzed. Secondary electron, backscattered electrons and characteristic x-rays are
produced from the interaction of the primary electron beam with the sample [149]. For
this investigation, a FIB / SEM Dual-station Helios 600 (FEI company) was used.
6.2.1 Secondary electron imaging – SE
Secondary electrons (SE) are generated from a narrow volume near the primary beam
impact area and, therefore, the imaging resolution can approximate the size of the
focused electron beam. Secondary electrons are loosely bound outer shell electrons from
the specimen atoms that are excited by the inelastic scattering of beam electrons. They
have low energy of less than 50 eV, with 90 % of them emitted with less than 10 eV. The
important consequence of their low kinetic energy is the shallow escape depth in the
order of few nanometers [149]. SE imaging was performed at a voltage of 5 kV and
currents in the range of 86 pA to 1.4 nA. Two secondary electrons detectors were used
during this investigation: an Everhart-Thornley detector (ETD) and a through-the-lens
detector (TTL).
ETD uses a scintillator to convert electrons to light and transmit them through reflection
in a light guide to a photomultiplier. Photons are then transformed back into electrons
multiplied in a cascade until they reach the final collector. The face of the scintillator is
metal coated with a bias positive potential to accelerate the SE to a sufficient energy to
generate light in the scintillator. The scintillator is surrounded with a Faraday cage with a
separate bias potential to increase the collection of SE with higher geometric efficiency.
The TTL detector is a scintillator with a high bias positioned above the objective lens.
Only the SE that pass through the lens bore are detected. The disposition of the detector
48
completely eliminates the SE coming from remotely produced electrons from BSE
colliding with the bottom of the lens and chamber walls. The TTL detector is especially
advantageous when used together with the “snorkel”-type lens. The strong magnetic field
of this lens extends from the polepiece directly to the specimen below the physical lens.
Lens aberrations are small due to the reduced working distance (reduced focal length)
and, therefore, high-resolution images are possible [149].
6.2.2 Backscattered electron imaging – BSE
Backscattered electrons (BSE) are beam electrons that interacted with the electrical field
of the specimen atoms and were elastically scattered along new trajectories with small or
no loss of kinetic energy. The beam electron can suffer several deflections before leaving
the specimen backscattered. Because of these repeated deflections, the interaction
volume from which BSE emerge is larger than the size of the incident beam and,
consequently, the resolution lower than SE. However, the important characteristic of
BSE is that the probability of elastic scattering increases with the atomic number (Z) by a
relationship of approximately Z2. Primary electrons interact strongly with the positive
charge of the nucleus of heavy atoms. This characteristic allows for compositional
contrast or also called Z contrast. Due to their penetration depth, BSE can carry
information from subsurface features of the specimen and therefore, this signal is
generally not considered surface-sensitive [149]. BSE imaging was performed at voltages
of 5 and 10kV and currents in the range of 0.34 to 1.4nA.
A high sensitivity solid state backscattered detector was used for this investigation (low-
voltage / high-contrast detector - vCD). Silicon diode detectors operate on the principle
of electron-holes production induced in a semiconductor by energetic electrons. They
have the form of a thin wafer placed parallel to the surface of the specimen below the
electron column.
Figure 6.1 compares SE and BSE contrasts for the eutectic phase in unmodified and
strontium modified Al- 7 wt% Si-0.3 wt% Mg alloys. The compositional and channeling
contrasts of BSE electron allows for the clear identification of eutectic Al2Si2Sr ternary
phases and twin boundaries in silicon.
49
Figure 6.1: Scanning electron microscopy images of unmodified and strontium modified
Al-7 wt% Si-0.3 wt% Mg alloys. Comparison of secondary electron (SE) and back-
scattered electrons (BSE) contrasts.
6.2.3 Energy dispersive X-ray spectroscopy – EDX
Beam electrons can interact with the specimen atoms and eject inner shell electrons. The
atom is left in an excited state and to go back to the ground state, an electron of an outer
shell can transition to fill the inner-shell vacancy. This energy difference between
electron shells is a characteristic value for each element and it can be released as a
photon of electromagnetic radiation called characteristic x-ray [149]. Characteristic x-
rays emitted from the sample were used in energy-dispersive x-ray spectroscopy (EDX)
as a routine analytical technique for qualitative elemental detection and phase
assignment. For this investigation, an Apolo XV detector (EDAX Company) was used for
EDX measurements at voltages of 15 and 20 kV and currents of 5.5 and 11 nA.
The compositional information in EDX comes from a region ~1 µm in both diameter and
depth with accuracy of approximately 1-2% in composition. At the typical beam currents
in SEM, EDX spectrometers are capable of detecting x-rays of all elements above an
atomic number of 4. Figure 6.2 shows an example of Al2Si2Eu phase recognition in an Al-
5 wt% Si with 500 ppm Eu.
50
Figure 6.2: Al2Si2Eu phase identification by energy dispersive X-ray (EDX) spectroscopy
in an Al-5 wt% Si hypoeutectic alloy with 500 ppm Eu. Acquisition with 11 nA and 15 kV
electron beam.
6.2.4 Electron backscattered diffraction – EBSD
In EBSD, the last elastic scattering event of BSE at the last plane before escaping the
specimen is analyzed. To get a maximal intensity in the diffraction patterns, the surface
of the specimen is tilted 70° from the horizontal. Primary electrons do not penetrate as
deeply in the strongly tilted specimen and electrons with small scattering angles can
escape the surface and be backscattered, increasing the backscattering coefficient. For
every plane in the specimen, BSE fulfill the Bragg diffraction condition in all directions
and Kossel-cones are formed. The so-called Kikuchi lines, are a section of the Kossel-
cone wall on the detector’s screen. For every plane in the specimen there are two
Kikuchi-lines and the distance between them is determined by the crystallographic
structure. A large part of the stereographic triangle is imaged at the same time which
facilitates the determination of the crystal orientation. Kikuchi lines are transformed into
51
points by a Hough transformation. Hough peaks are detected and back-transformed to
index the Kikuchi lines. The relative orientation of the grains and phases in the specimen
is determined with respect to a reference orientation. A charge-coupled-device (CCD)
camera is used for the acquisition and a specialized software makes an automated
indexing of the backscattered Kikuchi patterns received from the sample [149].
Transmission Kikuchi diffraction (TKD) was also used in this study. This is a similar
technique to EBSD but uses scattered electron that are forward transmitted from a TEM
sample for the diffraction analysis. The advantage of this alternative is the improved
resolution due to the small interaction volume in an approximately 50 nm thick TEM
lamella [150,151]. Because transmitted electrons are used, the tilt of the specimen is of
45° in the opposite direction of the 70° used for EBSD. The working distance is of 4-5
mm in comparison to the 9-11 mm used for EBSD [151].
For the analysis of results in this work, the data was presented in two ways: pole figures
(PFs) and inverse pole figures (IPFs). PFs represent a stereographic projection of the
crystallographic directions in the analyzed sample. A stereographic projection is a two-
dimensional projection of the three-dimensional unit cell such that the angular
relationship between planes and directions in the crystal can be read from the projection.
Each pole shows the direction of the crystal plane normal with respect to a sample
reference frame. An (hkl) pole figure shows the distribution of the {hkl} poles in the
sample. IPFs are orientation maps that show the distribution of crystallographic
orientations in relationship to a reference sample direction [152].
In this investigation, an EDAX Hikari detector within the FIB/SEM workstation was used
to record EBSD and TKD data using an electron beam of 20 kV / 22 nA and 30 kV / 5.5
nA, respectively; and post-processing was performed using the OIM Analysis software
(EDAX). IPFs were used to identify twin boundaries and Σ9 boundaries in eutectic silicon
for different samples (figure 6.3(b)); and PFs from silicon and neighboring Al2Si2X
(X=Eu, Yb) phases were used to find parallel orientations during the growth of these
phases as evidence of heterogeneous epitaxial growth. Al2Si2X crystal structures from
literature [153–155] were added to OIM Data Collection 7 database by means of the built-
in structure creation wizard. Figure 6.3 shows an example in the Al- 5 wt% Si with 500
ppm Eu. For the calculation of pole figures, data points from the intermetallic compound
and neighboring silicon were extracted using the grain highlighting feature of OIM Data
Analysis 7 software. This feature allows a point-and-click selection of data points which
belong to continuous grains based on a predefined tolerance angle (here 5°) (figure 6.3
(c, d)). Selected data points were then used for pole figure calculation. In contrast to
52
classic cleanup techniques which involve a more or less severe data modification, the
highlighting technique is no more than an orientation based selection of data points.
Figure 6.3: EBSD analysis in an Al-5 wt% Si with 500 ppm Eu. Crystal orientation
relationship between silicon and Al2Si2Eu: (a) SEM image of Al2Si2Eu in contact with
silicon and aluminum. (b) Overlaid IPF (color) and image quality (IQ) (brightness) maps
of region in (a). Uncleaned raw data, step size 30 nm. Black line depicts a ∑3 twin
boundary between the silicon orientations. (c,d) Orientation-based extracted IPF maps of
neighboring silicon and Al2Si2Eu. (e) PFs of Al2Si2Eu and both silicon orientations in (c)
and (d). Circles and squares highlight parallel plane normals in the crystals.
6.3 Focused ion beam – FIB

A focused ion beam (FIB) with a liquid metal ion source (LMIS) allows the controlled
sputtering of small volumes of material in a sample and for this reason, is especially
useful for site-specific sample preparation for TEM and APT. In this work, a SEM / FIB
53
dual-beam station Helios Nanolab 600 (FEI Company) with a gallium (Ga+) source was
used.
6.3.1 Sample preparation for TEM
An in-situ lift-out procedure in the SEM/FIB dual station was used for site-specific TEM
sample preparation. The region of interest (ROI) can be chosen by surface inspection
with SE or BSE and the orientation of the lamella can be determined by surface EBSD.
Figure 6.4 shows an example of preparation in the Al-Si eutectic phase. Electron and ion
induced platinum were deposited on the ROI to protect the region from gallium
implantation and avoid later curtain effect during thinning. A script was used to
automatically cut-out an approximately 1 µm thick lamella and the specimen was lifted-
out and transferred to a 3 mm TEM grid with an Omniprobe micromanipulator. Cleaning
cross-section patterns with progressively decreasing currents are used for thinning on
both sides. A last low voltage cleaning step with 2 kV is done to minimize artifacts like
amorphization , ion implantation and creation of atomic defects in the lamella after
thinning with 30 kV.
Figure 6.4: FIB sample preparation for TEM lamella. (a) Electron induced platinum
deposition. (b) Lamellas after TEM-Wizard. (c) In-situ lift-out of a lamella with a
micromanipulator. (d) Transfer to a TEM Grid for thinning. (e) Lamella perpendicular to
the ion beam for thinning. (f) Thinned lamella ready for analysis.
54
6.3.2 Sample preparation for APT
APT requires a high electric field on the specimen to field evaporate atoms from its
surface. For this reason, the specimen needs to be needle-shaped with an end radius
smaller than 100 nm. If it is assumed that the specimen ends in a hemispherical cap of
radius R and a high voltage (V) is applied, the electric field (F) at the apex of the
specimen can be estimated by the idealized equation of a charged sphere:
where kf is a field factor between 3 and 5 accounting for the specimen shape and the
electrostatic environment [156]. The end radius of the specimen is practically limited by
the required evaporation fields of around 10 to 50 V/nm, and a maximal applied voltages
of up to around 10 kV.
Figure 6.5: FIB sample preparation for APT. (a) Cut-out of a triangular wedge in the
eutectic phase. (b) Array of micro-posts to mount the lift-out. (c) Triangular wedge
attached to a micromanipulator being transferred to one of the posts shown in (b). (d)
Cut piece attached by platinum deposition to the holder. (e) First thinning steps by
annular milling. (f) APT specimen ready to run.
Needle-shaped specimens for APT can be prepared in two ways: electro-chemical

polishing and FIB-assisted methods. The main advantages of electro-chemical polishing
are avoiding the implantation of gallium ions and its low-cost; however, no site-specific
preparation is possible. For Al-Si alloys, a further problem of electro-polishing is the
55
selective etching that leads to inhomogeneities in the form of the specimen. For these
reasons, FIB methods were used in this work. The lift-out method described by
Thompson et al [157] was used to prepare specimens of very fine eutectic structure and to
study eutectic aluminum. Figure 6.5 shows the main steps of this method: a 20 µm long
triangular wedge is cut-out from the sample and transferred to an APT sample holder
with an array of 36 posts. The wedge is then divided into 6 to 8 pieces attached to single
posts and an annular milling process is then used to gradually thin the specimens up to
the final required size. A last low-energy milling step with 2 kV is used to minimize
gallium induced damage.
Since this work specially focuses on the analysis of eutectic silicon, a phase-selective
sample preparation method was developed to prepare silicon specimens. In this method,
aluminum was deep etched and single silicon branches or plates were lifted out by FIB.
Details concerning this method can be found in Paper V of this thesis.
6.4 Transmission electron microscopy – TEM

In transmission electron microscopy (TEM) and scanning transmission electron
microscopy (STEM), a high-energy electron beam interacts with a thin specimen. If the
specimen is thin enough to be electron transparent (~20-100 nm), beam electrons are
transmitted through the specimen and collected after interacting with its atoms. Due to
the small interaction volume, (S)TEM resolution is significantly increased with respect to
SEM. TEM can be used to observe the crystal structure, crystal defects or grain
boundaries down to the atomic scale. TEM’s high-resolution makes it an adequate
technique for correlation with APT. In this investigation, TEM and STEM were used to
characterize the crystal defects in modified silicon. Since stacking faults (SF) / nanotwins
and ∑3 twin boundaries, all of them on (111)Si planes, are the most frequent defects,
selected area electron diffraction (SAED) patterns were used to align the specimens in
<110>Si zone axes to see the defects on edge (figure 6.6).
TEM and STEM imaging and analysis in this work were performed in a Tecnai G2 TF 20
UT FEG (FEI Company) in micro and nanoprobe modes, an image-side aberration-
corrected JEOL 2100F microscope and a JEM - ARM 200 F Cold FEG TEM/STEM
equipped with a spherical aberration (Cs) probe and image correctors. All of them
operated at 200 kV.
56
Figure 6.6: Eutectic silicon in an Al-7Si-0.3Mg with 250 ppm Sr. (a) Twinned silicon
imaged with bright field diffraction contrast. Signal through a 20 µm objective aperture.
(b) Selected area electron diffraction (SAED) pattern aligned in <110>Si zone axis. Red
and blue circles highlight {111}Si reflections for the two mutually twinned silicon crystals
in the specimen. (c) High resolution image showing twins, nanotwins and a grain
boundary in detail.
6.5 Atom probe tomography – APT

APT is an analytical technique that provides three-dimensional mapping of materials
with atomic resolution. Surface atoms from a needle-shaped specimen are ionized,
removed from the material and accelerated towards a detector. The removal of atoms
from their own lattice, is field-induced by the combined effect of a standing electrostatic
field and either high-voltage or laser pulses. In this investigation a LEAP 3000X HR
(CAMECA Company) was used. Measurements were performed at repetition rates of 200
– 250 kHz, pressure lower than .33 × 0−8 Pa, evaporation rate of 5 atoms per 000
pulses and temperatures between 40 and 60 K. Laser-pulsed APT was applied to
measure silicon specimens using a laser with a wavelength of 532 nm, pulse length of 10
ps, and a pulse energy of 0.3–0.4 nJ. Voltage measurements were use to measure
aluminum specimens at 20% pulse fraction. Datasets were reconstructed and analyzed
with IVAS™ 3.6.8 sofware (CAMECA)
6.5.1 Working principle and experimental setup
Müller [158] explained field evaporation by the image hump model. In this model, a
sufficiently strong electric field fully ionizes a surface atom and lowers the energy barrier
of the ion to escape the surface. Later, Gomer [159,160] proposed the alternative charge-
exchange model that assumes that ionization and escape take place simultaneously while
the atom is desorbing. In any case, the energy barrier that bonds the atom to the surface
57
is overcome by thermal activation energy, although it is possible that in low-temperature
regime, ion tunneling through the energy barrier also plays a role. One further model
worth mentioning is the post-ionization theory [161,162]. The early models, assumed the
direct transition of the surface atom into an n-charged ion. However, in the 1980s, it was
shown that the ion leaves the surface with a single charge and is then subjected to one or
more field ionization processes as it moves away from the surface. The number of post-
ionization events increase with field. This theory was used to calculate the relative
frequency of charge-states as a function of field for several elements in the periodic
system. The resultant curves are called Kingham curves and are used nowadays as a
rough estimation of the evaporation field during a measurement. For a deeper review
about field evaporation theories refer to [163].
Figure 6.7: Schematic of the required electric field to evaporate as a function of

temperature (thermal activation energy). Pulsing modes are depicted by arrows starting
from a standing point.
In a practical sense, the removal of atoms from the specimen is a function of the applied
field and temperature. Each element exhibits a characteristic behavior and a slope as
shown schematically in figure 6.7. To achieve a controlled evaporation atom-by-atom,
two mechanisms can be used: field-pulsing and thermal-pulsing. Starting from a
standing point at a DC field (base field) and base temperature, the electric field can be
increased by high-voltage pulses at a constant temperature or, on the contrary,
temperature can be increased by laser pulses at a constant field. In both cases, enough
energy is transferred to the specimen to evaporate an atom from the surface (figure 6.7).
Figure 6.8 shows the basic experimental setup of an atom probe. The needle-shaped
specimen is mounted on a three-axis translation stage and cooled to cryogenic
58
temperatures between 20 to 80K. The base electrostatic field is generated by a high-

voltage (HV) supply connected to the specimen and the voltage pulses are applied to a
local electrode positioned in front of the specimen. The system operates under ultra-high
vacuum (UHV) below 10-8 Pa.
Figure 6.8: Basic experimental setup of an atom probe. [UHV: ultra-high vacuum; HV:
high voltage; ToF: time of flight].
The detection system is able to record the position of impact of each ion. Ions are
converted into electrons and multiplied in a stack of microchannel plates (MCP). The
electrons then come to a delay-line position-sensitive detector, where they are focused on
two or three polarized conductive lines. The electrons induce electric signals on all lines
and are propagated along them till the end of the line. Using the time and distance
covered, the X and Y coordinates of the impact on the detector are calculated.
The chemical nature of the ions is determined by a time-of-flight (ToF) spectrometer.

The time of flight of an ion (tflight) is defined as the time between its departure from the
specimen -given by the pulse that triggers evaporation- and its arrival, when the ion
impacts the detector. The ion will be accelerated in the electrostatic field by the potential
energy available , where V is the total voltage and ne, the charge of the ion. The
total kinetic energy after departure will be equal to the potential energy, and
the velocity can be defined as , where L is the flight length between the
specimen and the detector [164]. The mass-to-charge ratio ( ) of the ion can be then
directly determined by:
59
By knowing the mass-to-charge ratio of each ion, it is possible to identify their chemical
nature. All mass-to-charge values in a measurement are showed in a single spectrum,
where the natural isotopic abundance ratios assist in the assignment of the peaks. The
peaks were ranged at the full width tenth maximum (FWTM) in this investigation. Figure
6.9 shows a mass-to-charge spectrum from a eutectic silicon specimen of an Al- 7 wt% Si
with 150 ppm Sr.
Figure 6.9: Mass-to-charge spectrum of eutectic silicon in Al- 7 wt% Si with 150 ppm Sr.
6.5.2 Data reconstruction, visualization and analysis
After acquisition, the raw data is reconstructed and analyzed in the dedicated software
IVAS 3.6.8 (CAMECA). This software uses an algorithm that calculates the position of the
ions by assuming that they fly in a straight line from the surface of the specimen, with
these lines originating from a single projection-point along the axis of the specimen
[165]. The position of the projection-point depends on the compression of the field lines
during evaporation, which is related to the distribution of the field at the surface of the
specimen and its shape. It is also assumed that the atoms are evaporated layer-by-layer
from the outermost surface inwards. This means that the sequence of detection reveals
the depth position of the atoms in the specimen, the z-coordinate. This information
together with the lateral position (x,y) deduced from the reverse projection of the impact
coordinates given by the position-sensitive detector determine the (x,y,z) coordinates for
each atom in the volume [166].
Four parameters are set by the operator during reconstruction: the evaporation field F e,
image compression factor ICF, field factor kf (section 6.3.2) and detection efficiency. In
this investigation, the detection efficiency of 37% given for the LEAP 3000 XR system
60
was used. Silicon specimens were reconstructed with Fe of 30-33 V/nm, ICF of 1.65 and
kf of 3.3. For aluminum specimens Fe of 19 V/nm, ICF of 1.5 and kf of 3.3 were adopted.
The tools used for visualization and analysis of the data in this investigation are iso-
concentration surfaces, one-dimensional (1D) concentration profiles, proximity
histograms and background corrected compositions in regions of interest (ROI).
An iso-concentration surface is a 3D representation of a threshold value. Concentrations

are interpolated from average voxel values and connected triangles build an interface for
a set value [167]. This is a powerful means of visualization and also the first step for the
construction of proximity histograms. Proximity histograms represent concentration
profiles with respect to distance from an interface allowing the analysis of sophisticated
features in a reconstruction. Compositions are computed normal to the iso-surface for
each point and, therefore, independent from the interface geometry [168].
1D concentration profiles are an alternative method to show the evolution of

composition. Concentrations are plotted along the length of a rectangular or cylindrical
ROI across a feature in the reconstruction. A fixed number of counts was used in this
work to divide the ROI into slices called blocks. Concentrations in each block are
calculated and displayed as a function of length. The block size was optimized to
maximize the signal-to-noise ratio. In principle, proximity histograms might seem as a
more advanced representation of the data, but in cases where iso-concentration surfaces
are difficult to obtain, for example due to low solute concentrations, 1D profiles are a
better choice.
To count the amount of solute atoms in small clusters from one to a couple of lattice
distances in size, tight ROI around each cluster were constructed and exported. The
solute atoms were counted after a time-independent background subtraction included in
IVAS software.
6.5.3 Trajectory aberrations and local magnification effects
Differences between the original structure and the reconstructed data may arise from a
number of reasons in APT [164]. In the samples analyzed in this work, the main concern
was trajectory aberrations. The deflection in the flight of the ions is related to the
distribution of the electric field at the surface of the specimen. Localized variations of the
electric field can induce undesirable lateral displacement of the ion just after
evaporation, which induces zones of high or low density of ions in the reconstruction
[169,170].
61
When a precipitate requires a considerably different field to evaporate than the matrix,
trajectory aberrations may occur. In this work, clusters with a high content of aluminum
need a significantly lower field than the silicon matrix around them. This produces a
flattening of the surface and an inward deflection of the ion trajectories that causes an
increased density of hits in the detector. Low-evaporation field solutes may also be
evaporated prematurely, i.e. before more protruding atoms, which would result in an
artefact in the depth coordinate. When considering 2 - 4 nanometer clusters, a
convolution of the matrix with the precipitate may occur resulting in an overestimation
of the matrix element [171–173]. Because of this uncertainty, relative solute ratios Al:M
(M = Sr, Na or Eu) rather than absolute concentrations were considered in this work. The
standard deviations in the Al:M ratios are attributed to the combined effect of the limited
number of solute atoms in the small clusters and the detection efficiency of APT (~ 37 %).
62
7. Summary of included papers and contribution

to the field
This investigation contributes to understand the growth change of eutectic silicon when
small concentrations of modifying elements are added to hypoeutectic Al-Si alloys.
Figure 7.1 shows a graphical summary of the three cases of alloys studied in this research
revealing the structure and atomic distribution in the silicon phase. A model explaining
the multiplication of crystal defects in silicon and its microstructural change is proposed
as the outcome of this thesis.
Figure 7.1: Overview of the cases studied: (a) Al-Si alloy with no addition of modifier as a
reference structure; (b) Alloy with ytterbium addition. Refined plates, but no coralline
63
structure; (c) Alloys with addition of the three most powerful modifiers known to date:
sodium, strontium and europium. Coral-like structure.
Paper I. Comparison of segregations formed in unmodified and Sr-modified Al-

Si alloys studied by atom probe tomography and transmission electron
microscopy
The atomic distribution of solutes in eutectic silicon of an Al – 7 wt% Si alloy with and
without strontium addition was studied by APT and (S)TEM.
The silicon phase in the unmodified alloy showed straight twins across the plates and
spherical coherent particles. APT depicted small particles (5 to 10 nm) with 5-15 at% Al,
larger particles (13-18 nm) with a core-shell structure with 40-50 at% Al in the shell, and
a ring-segregation decorated with 6 at% Al. These aluminum enrichments are small
pockets of liquid entrapped during solidification and impurity clouds on silicon defects
such as dislocation loops.
The modified alloy with 150 ppm Sr showed irregular twin boundaries and a high density
of crystal lattice defects correlated to Al-Sr-Si segregations. Depending on their
morphology, the segregations were divided into three categories: rod-like, planar and
particle-like (rounded). Rod-like segregations have one dimension significantly larger
than the other two and often show a v-shape arrangement with precipitates at their
intersection. They are responsible for changes in the silicon lattice stacking sequence and
contribute to the irregular growth and smoothening of the boundary between eutectic
aluminum and silicon. Planar segregations are the result of the progressive accumulation
of aluminum and strontium at the silicon solid-liquid interface and favor the formation
of grain boundaries and irregular twin boundaries. Rounded segregations are also found
in the modified alloy, with the distinction that all contained strontium atoms together
with aluminum and regions with high density of small segregations responsible for the
silicon growth obstruction.
The amount of aluminum and strontium in solid solution in silicon was of 430 ± 160 at-
ppm and 40 ± 10 at-ppm, respectively. The total content of aluminum trapped in silicon
was nearly 4 times higher in the modified alloy and most of the strontium in the alloy
segregated to the silicon phase. Based on the Al:Sr content ratio in the rod-like and
planar segregations a local ternary reaction is proposed to assist in the formation and
pinning of crystal defects.
64
Paper II. Cluster formation at the Si / liquid interface in Sr and Na modified Al-
Si alloys
Atom probe tomography was used to compare the effect of strontium and sodium
additions in the eutectic silicon phase of hypoeutectic Al-Si alloys. Spherical precipitates,
stacking faults (SFs) and twin lamellas (TLs) depicted with TEM in both alloys show one-
to-one correlation with APT results showing spherical, rod-like and planar clusters in
silicon. The silicon growth is obstructed in both cases by the formation of crystal defects
with solute-enriched clusters. However, the solute ratios Al:M (M= Sr or Na) in the
clusters were 2.92 ± 0.46 in the case of strontium and 1.07 ± 0.23 for sodium, showing a
distinctly different interaction of aluminum with each of the modifiers. The heterogenous
formation of clusters of the ternary compounds AlSiNa and Al2Si2Sr at the silicon / liquid
interface is proposed to alter the silicon microstructure by imparting growth flexibility.
These ternary compounds are reported in literature for ternary invariant reactions in Al-
Si-Sr and Al-Si-Na systems at nearly the same temperature and composition as the Al-Si
binary eutectic reaction. The efficiency of a modifier is suggested to be closely related to
its ability to form ternary compound clusters at the faceted silicon solidification front
during growth.
Paper III. Eutectic modification by ternary compound cluster formation in Al-

Si alloys
The effects of europium and ytterbium additions in hypoeutectic Al-Si alloys were
studied. The addition of europium modifies the eutectic silicon to a coral-like structure,
while ytterbium yields a refined plate-like silicon structure. Backscattered electron
images showed that both alloys form pre-eutectic and eutectic Al2Si2Eu and Al2Si2Yb
phases. A substantial difference was found when comparing the silicon phases with APT.
The alloy with europium addition presented numerous clusters containing europium and
aluminum with the three morphologies consistent with Paper I and II: spherical, rod-like
and planar. The average of the Al:Eu ratio for 35 rod-like and 4 planar clusters was of
2.25 ± 0.42, which agrees with the formation of Al2Si2Eu clusters at the crystallographic
defects. On the contrary, no ytterbium was found in the silicon phase and only aluminum
segregations similar to the unmodified alloy were present. It is proposed that the
difference between these alloys is their ability to form early stage Al 2Si2X clusters on the
facets of silicon. The formation of clusters obstructs the silicon growth and increases
growth flexibility causing a coral-like silicon structure. The crystallographic orientation
relationship between silicon and Al2Si2Eu or Al2Si2Yb was studied by EBSD / TKD. The
parallel plane normals: 111Si // 6 10Al2Si2Eu, 011Si // 6 10Al2Si2Eu and 011Si // 0001Al2Si2Eu
65
with misfits of the d-spacings smaller than 6 % were found between silicon and Al2Si2Eu,
while no consistent orientation relationship was found between silicon and Al2Si2Yb. The
interplanar spacing misfits between the d(0002)Al2Si2X and d(110)Si are 5.79% for Al2Si2Eu
and 10.50% for Al2Si2Yb. The higher mismatch of the ytterbium phase adds strain energy
to its nucleation, what makes Al2Si2Yb clusters less stable than Al2Si2Eu clusters on
silicon facets. Based on these results, it is postulated that Al2Si2Yb needs a higher energy
to heterogeneously form sub-critical clusters on silicon than Al2Si2Eu.
Paper IV. Nucleation and growth of eutectic Si in Al-Si alloys with Na addition
The nucleation and growth of sodium modified Al-Si alloys was studied in controlled
sand casting and melt spun samples. TEM and APT analyses of controlled sand casting
alloys showed a high density of twins and rod-like and spherical clusters containing
sodium and aluminum atoms with Al:Na ratios of about 1-1.5. It is proposed that the
spherical clusters might agree with the impurity induced twinning idea changing the
silicon stacking sequence and promoting twin formation. The rod-like clusters can be
related to the poisoning of twin plane re-entrant edges as proposed by previous models.
The rejection of sodium ahead of the solidification front was confirmed by a sodium
enrichment of about 0.25 at% at the interface between eutectic aluminum and silicon.
Melt spun samples were used to study eutectic nucleation for sodium addition. DSC was
used to measure the undercooling for nucleation of entrained droplets. Al-5 wt% Si
without sodium showed an undercooling of 24 °C, while the addition of 19 ppm Na lead
to a significant depression of the nucleation temperature (∆T= 49 °C). This depression
can be explained by the removal of AlP nucleation sites by the formation of the
competing phase Na3P. However, the entrained droplet undercooling for the alloy with
160 ppm Na decreased to 29 °C, indicating that the formation of Na3P alone is not
enough to explain the effect of sodium on nucleation.
Paper V. Phase selective sample preparation of Al-Si alloys for atom probe
tomography
A method to selectively prepare eutectic silicon specimens for APT is presented. It is

often said that in APT, the specimen itself is the “lens” of the microscope. The trajectory
of the ions removed from the surface determine the magnification and are responsible for
the later 3D-image formation. To obtain long and stable measurements of the silicon
phase with a high quality mass-spectrum, aluminum eutectic surrounding the silicon
branches was deep etched and silicon pieces were lifted-out attached to a
66
micromanipulator in the SEM / FIB dual station. This specimens were transferred to an
APT sample holder for final thinning. Two slightly adapted methods are given for
rounded modified silicon branches and coarser unmodified silicon plates. Specimens
prepared by this method had an extremely high success rate and yield during APT
measurements. Density fluctuations and preferential retention due to the large difference
between the evaporation field of aluminum and silicon were minimized.
7.1 Contribution to the field

The present investigation shows that all three elements that induce silicon plate-to-coral
morphology transition - strontium, sodium and europium – contain nanometer sized
clusters of solute atoms at the defects in silicon. The clusters have stoichiometries
corresponding to compounds formed in the ternary eutectic reaction of each system. In
contrast, the addition of ytterbium, which only refines the silicon plates but do not form
corals, is unable to form ternary compound clusters in silicon.
The evidence of clusters containing aluminum apart from the modifying element,
falsified the impurity induced twinning model which assumed the adsorption of single
modifying atoms at the silicon growth interface, but confirmed the idea that the modifier
is responsible for the multiplication of crystallographic defects in silicon.
We propose that the formation of ternary compounds clusters AlSiNa, Al 2Si2Sr and
Al2Si2Eu at the silicon / liquid interface during solidification restrict the silicon growth.
The formation of ternary compound clusters on silicon facets create growth steps and
increase growth direction diversity. The incorporation of clusters onto the silicon surface
explains the high density of crystallographic defects in silicon and the modification from
plates to corals.
The parallel lattice plane-normals 011Si // 0001Al2Si2Eu, 011Si // 6 10Al2Si2Eu and 111Si //
6 10Al2Si2Eu were found between Al2Si2Eu and silicon, and absent between Al2Si2Yb and
silicon. We propose a favorable heterogeneous formation of Al2Si2Eu on Si. The misfit
between 011Si and 0002Al2Si2X interplanar spacings shows a consistent trend with the
potency of modification for several elements such as strontium, sodium, europium,
calcium, barium, ytterbium and yttrium.
7.2 Outlook and future work

Based on the results obtained in this thesis some ideas for further work are presented.
67
7.2.1 APT analysis of modified Al-Si eutectic phase
APT proved to be a powerful analytical tool in the study of eutectic modification in Al-Si
alloys. Continuing with the results in this thesis, the analysis of the silicon phase in high
purity alloys with additions of calcium, yttrium and barium will give a complete outlook
of the interaction of the modifying elements with aluminum and silicon at the
solidification front and the formation of clusters and defects.
Further APT analysis of the Al / Si boundary in the eutectic phase for the different
modification elements is also of interest. Paper IV shows an enrichment of sodium at the
Al / Si interface. The comparison with additions of other modifying elements would
enable a deeper understanding of the role that this enrichment plays during growth.
7.2.2 Calculation of phase diagrams and simulations
The results in this work show the relevance of phase formation in the ternary systems Al-
Si-X, with X = Sr, Na, Eu or Yb. To the best of our knowledge, the Al-Si-Sr [115] and Al-
Si-Na [114] phase diagrams were reported in literature, but Al-Si-Eu and Al-Si-Yb are
still missing. These phase diagrams, particularly the liquidus surfaces, will assist in the
interpretation of phase formation and solidification pathways during eutectic reaction.
Atomistic simulations of the crystallization at the eutectic solidification front are also
needed. These simulations will unravel the structure of the melt just ahead of the
crystallization interface and the cluster formation process. Li et al [99] performed
calculations based on density functional theory with europium atoms placed in four
different positions of a silicon twin boundary. However, in these calculations no
aluminum atoms where considered, which differs with the experimental results shown in
Paper III. Using another approach, Suzuki-Yamamoto et al [174] studied the stability of
the crystal structure of the ternary intermetallics Al2Si2X (X = Ca, Sr, Ba, Eu, Y, Yb) over
a wide range of temperatures showing their stability and proposing their possible
relevance for modification, but no calculation in the silicon lattice was performed.
Moreover, no simulations were performed to visualize the undercooled melt in addition
to the solid, with atomic-scale resolution. Force-field based molecular-dynamics (MD)
simulations can be employed as an option to facilitate this [175]. Force field based
simulations, rather than density-functional theory (DFT) would give access to relevant
number of atoms over sufficiently large times to “equilibrate” the slightly undercooled
melt close to the crystalline surface. An example of the potential of force field based
simulations can be found in the work of Gao et al [176], who studied the moving
crystallization front of CuZr from a high viscous melt.
68
7.2.3 Correlative characterization of the solidification front
The heterogeneous formation of clusters of ternary compounds on the eutectic

solidification front is proposed to be one of the keys of modification. This hypothesis is
based on “post-mortem” analysis, i.e. on completely solidified alloys. As a further study,
we propose the comparison of the topology, crystal structure and concentration profile at
the solidification front of unmodified and modified alloys in samples quenched during
eutectic solidification. The study of the three-dimensional (3D) “frozen” solidification
front at various length scales will allow the comparison of the eutectic growth modes and
the clarification of the role of aluminum together with the modifier agent in the silicon
crystal. Nano-CT (X-Ray computer tomography) or plasma FIB are techniques capable of
giving 3D images to unveil the microstructural evolution and show the change of
morphology of the solidification front. This analysis enables the study of the growth
restriction due to the presence of the modifier at the growth front. EDX maps can give a
rough idea of the concentration profile at both sides of the solidification front and help to
recognize the exact position of the front. APT specimens at the growth front and at
increasing distances from it can be used to determine the accumulation of the modifying
element at the front during solidification. TEM can provide the structure of the
crystallographic defects in the silicon structure and at the solidification front. The
correlation of the atomic evolution at the growth front with the evolution at the
micrometer scale leads to a global understanding of the solidification process and the
formation of the coralline structure.
7.2.4 Interaction of modifying elements in Al-Si alloys with magnesium and

iron
In this thesis, high purity binary Al-Si alloys were studied to understand the interaction
between the modifying agent and silicon avoiding the effects of further impurities or
alloying elements. The next step is to analyze more complex alloys such as Al-Si-Mg and
Al-Si-Cu-Mg. These alloys are industrially relevant and the interaction between the
modifier agent and these ternary and quaternary elements has not been studied at the
atomic scale. APT analysis of the Mg2Si and Al2Cu phases and its boundary with eutectic
aluminum and silicon would reveal whether a part of the modifying element is consumed
by these phases. Additionally, the interaction of iron and its phases with the modifying
elements has also not been studied at the atomic level.
69
70
8. References
[1] J.G. Kaufman, E.L. Rooy, Aluminum Alloy Castings: Properties , Processes , and
Applications., 2004.
[2] M.M. Makhlouf, Solidification of Eutectic Alloys - Aluminum-Silicon, in: ASM

Handbook, Vol. 15 Cast., 2008: pp. 312–316. doi:10.1361/asmhba0005212.
[3] S. Khan, R. Elliott, Solidification kinetics of the unmodified aluminium-silicon

flake structure, Acta Metall. Mater. 41 (1993) 2433–2439. doi:10.1016/0956-
7151(93)90323-K.
[4] W. Kurz, R. Trivedi, Overview No. 87 Solidification microstructures: Recent

developments and future directions, Acta Metall. Mater. 38 (1990) 1–17.
doi:10.1016/0956-7151(90)90129-5.
[5] M.G. Day, A. Hellawell, The Microstructure and Crystallography of Aluminium-

Silicon Eutectic Alloys, Proc. R. Soc. A Math. Phys. Eng. Sci. 305 (1968) 473–491.
doi:10.1098/rspa.1968.0128.
[6] A. Hellawell, The growth and structure of eutectics with silicon and germanium,
Prog. Mater. Sci. 15 (1970) 3–78. doi:10.1016/0079-6425(70)90001-0.
[7] J.E. Gruzleski, B. Closset, The treatment of liquid Aluminum-Silicon alloys,

American Foundrymens Society, 1990.
[8] A. Pacz, Alloy, 1,387,900, 1921.
[9] V.L. Davies, J.M. West, Factors affecting the modification of the aluminium-
silicon eutectic, J. Inst. Met. 92 (1963) 175–180.
[10] H. Fredriksson, M. Hillert, N. Lange, The modification of aluminium-silicon alloys

by sodium, J. Inst. Met. 101 (1973) 285–299.
[11] D.C. Jenkinson, L.M. Hogan, The modification of aluminium-silicon alloys with
strontium, J. Cryst. Growth. 28 (1975) 171–187. doi:10.1016/0022-
0248(75)90233-X.
[12] M. Shamsuzzoha, L.M. Hogan, The crystal morphology of fibrous silicon in

strontium-modified Al-Si eutectic, Philos. Mag. A. 54 (1986) 459–477.
doi:10.1080/01418618608243605.
[13] J.H. Li, X.D. Wang, T.H. Ludwig, Y. Tsunekawa, L. Arnberg, J.Z. Jiang, P.
Schumacher, Modification of eutectic Si in Al–Si alloys with Eu addition, Acta
Mater. 84 (2015) 153–163. doi:10.1016/j.actamat.2014.10.064.
[14] T.H. Ludwig, E. Schonhovd Dæhlen, P.L. Schaffer, L. Arnberg, The effect of Ca and
P interaction on the Al-Si eutectic in a hypoeutectic Al-Si alloy, J. Alloys Compd.
586 (2014) 180–190. doi:10.1016/j.jallcom.2013.09.127.
71
[15] T.H. Ludwig, P.L. Schaffer, L. Arnberg, Influence of Some Trace Elements on
Solidification Path and Microstructure of Al-Si Foundry Alloys, Metall. Mater.
Trans. A. 44 (2013) 3783–3796. doi:10.1007/s11661-013-1694-y.
[16] A. Knuutinen, K. Nogita, S.D. Mcdonald, A.K. Dahle, Modification of Al – Si alloys

with Ba , Ca , Y and Yb, J. Light Met. 1 (2002) 229–240. doi:10.1016/S1471-
5317(02)00004-4.
[17] K. Nogita, H. Yasuda, M. Yoshiya, S.D. McDonald, K. Uesugi, a. Takeuchi, Y.

Suzuki, The role of trace element segregation in the eutectic modification of
hypoeutectic Al–Si alloys, J. Alloys Compd. 489 (2010) 415–420.
doi:10.1016/j.jallcom.2009.09.138.
[18] J.H. Li, S. Suetsugu, Y. Tsunekawa, P. Schumacher, Refinement of Eutectic Si

Phase in Al-5Si Alloys with Yb Additions, Metall. Mater. Trans. A. 44 (2012) 669–
681. doi:10.1007/s11661-012-1410-3.
[19] K. Nogita, J. Drennan, A.K. Dahle, Evaluation of Silicon Twinning in Hypo-

Eutectic Al-Si Alloys, Mater. Trans. 44 (2003) 625–628.
doi:10.2320/matertrans.44.625.
[20] K. Nogita, S.D. McDonald, A.K. Dahle, Eutectic Modification of Al-Si Alloys with
Rare Earth Metals, Mater. Trans. 45 (2004) 323–326.
doi:10.2320/matertrans.45.323.
[21] R.N. Lumley, Introduction to aluminum metallurgy, in: R. Lumley (Ed.), Fundam.
Alum. Metall. Prod. Process. Appl., First Edit, Woodhead Publishing, Cambridge,
UK, 2011: pp. 1–20.
[22] S. Otarawanna, A.K. Dahle, Casting of aluminium alloys, in: L. R.N. (Ed.),
Fundam. Alum. Metall. Prod. Process. Appl., Woodhead Publishing, Cambridge,
UK, 2011: pp. 141–154.
[23] R.B.C. Cayless, Alloy and Temper Designation System for Aluminum and
Aluminum Alloys, in: ASM Handbook, Vol. 2 Prop. Sel. Nonferrous Alloy. Spec.
Mater., ASM International, 1990: pp. 15–28. doi:10.1361/asmhba0001058.
[24] K. Kirgin, Feeling the domino effect, Mod. Cast. 99 (2009) 34–36.
[25] European Aluminium Association. The Aluminium automotive manual - Cast

alloys and products, (2002). https://www.european-
aluminium.eu/media/1544/aam-products-6-cast-alloys-and-products.pdf
(accessed November 15, 2018).
[26] W. Kurz, D.J. Fisher, Fundamentals of Solidification, Third Edit, Trans Tech
Publications, Switzerland, 1992.
[27] S. Shankar, Eutectic Solidification - An Introduction, in: ASM Handbook, Vol. 15

Cast., ASM International, 2008: pp. 307–311. doi:10.1361/asmhba0005211.
[28] M.N. Croker, R.S. Fidler, R.W. Smith, The Characterization of Eutectic Structures,
Proc. R. Soc. A Math. Phys. Eng. Sci. 335 (1973) 15–37.
doi:10.1098/rspa.1973.0111.
[29] M.N. Croker, D. Baragar, R.W. Smith, Anomalous eutectic growth, J. Cryst.
Growth. 30 (1975) 198–212. doi:10.1016/0022-0248(75)90090-1.
[30] K.A. Jackson, Liquid metals and solidification, American Society for metals,
72
Cleveland, Ohio, 1958.
[31] R. Elliott, Eutectic solidification, Int. Mater. Rev. 22 (1977) 161–186.

doi:10.1179/095066077791100363.
[32] D.M. Stefanescu, Science and Engineering of Casting Solidification, Second Ed.,
Springer, New York, 2009.
[33] M.R. Taylor, R.S. Fidler, R.W. Smith, Broken lamellar eutectic growth; Structure
of the silver-bismuth eutectic, J. Cryst. Growth. 3–4 (1968) 666–673.
doi:10.1016/0022-0248(68)90242-X.
[34] J.L. Murray, A.J. McAlister, The Al-Si (Aluminum-Silicon) system, Bull. Alloy
Phase Diagrams. 5 (1984) 74–84. doi:10.1007/BF02868729.
[35] J. Barrirero, M. Engstler, N. Ghafoor, N. de Jonge, M. Odén, F. Mücklich,

Comparison of segregations formed in unmodified and Sr-modified Al-Si alloys
studied by atom probe tomography and transmission electron microscopy, J.
Alloys Compd. 611 (2014) 410–421. doi:10.1016/j.jallcom.2014.05.121.
[36] M.M. Makhluf, Solidification of Eutectic Alloys - Aluminum-Silicon, in: ASM

Handbook, Vol. 15 Cast., ASM International, 2008: pp. 312–316.
doi:10.1361/asmhba0005212.
[37] H.A.. Steent, A. Hellawell, Structure and properties of aluminium-silicon eutectic

alloys, Acta Metall. 20 (1972) 363–370. doi:10.1016/0001-6160(72)90030-2.
[38] M. Shamsuzzoha, L.M. Hogan, Twinning in fibrous eutectic silicon in modified Al-
Si Alloys, J. Cryst. Growth. 72 (1985) 735–737. doi:10.1016/0022-0248(85)90229-
5.
[39] S. Lu, A. Hellawell, Growth mechanisms of silicon in Al-Si alloys, J. Cryst. Growth.
73 (1985) 316–328. doi:10.1016/0022-0248(85)90308-2.
[40] S. Lu, A. Hellawell, The mechanism of silicon modification in aluminum-silicon

alloys: Impurity induced twinning, Metall. Trans. A. 18 (1987) 1721–1733.
doi:10.1007/BF02646204.
[41] M. Shamsuzzoha, L.M. Hogan, The twinned growth of silicon in chill-modified

Al−Si eutectic, J. Cryst. Growth. 82 ( 98 ) 598–610. doi:10.1016/S0022-
0248(87)80004-0.
[42] R.S. Wagner, On the growth of germanium dendrites, Acta Metall. 8 (1960) 57–60.
doi:10.1016/0001-6160(60)90145-0.
[43] D.R. Hamilton, R.G. Seidensticker, Propagation Mechanism of Germanium

Dendrites, J. Appl. Phys. 31 (1960) 1165. doi:10.1063/1.1735796.
[44] M. Shamsuzzoha, L.M. Hogan, D.J. Smith, P.A. Deymier, A transmission and
high-resolution electron microscope study of cozonally twinned growth of eutectic
silicon in unmodified Al-Si alloys, J. Cryst. Growth. 112 (1991) 635–643.
doi:10.1016/0022-0248(91)90119-P.
[45] K.F. Kobayashi, L.M. Hogan, The crystal growth of silicon in Al-Si alloys, J. Mater.
Sci. 20 (1985) 1961–1975. doi:10.1007/BF01112278.
[46] M. Shamsuzzoha, L.M. Hogan, Crystal morphology of unmodified aluminium-

silicon eutectic microstructures, J. Cryst. Growth. 76 (1986) 429–439.
doi:10.1016/0022-0248(86)90390-8.
73
[47] X. Liu, Y. Zhang, B. Beausir, F. Liu, C. Esling, F. Yu, X. Zhao, L. Zuo, Twin-
controlled growth of eutectic Si in unmodified and Sr-modified Al-12.7%Si alloys
investigated by SEM/EBSD, Acta Mater. 97 (2015) 338–347.
doi:10.1016/j.actamat.2015.06.041.
[48] E. and P.J.H. Billig, Some observations on growth and etching of crystals with the
diamond or zincblende structure, Acta Crystallogr. 8 (1955) 353–354.
doi:10.1107/S0365110X55001084.
[49] E. Billig, P.J. Holmes, New Observations on the Structure Germanium Dendrites,
Acta Metall. 5 (1957) 53–54.
[50] M. Kitamura, S. Hosoya, I. Sunagawa, Re-investigation of the re-entrant corner

effect in twinned crystals, J. Cryst. Growth. 47 (1979) 93–99. doi:10.1016/0022-
0248(79)90162-3.
[51] F. Lasagni, a. Lasagni, E. Marks, C. Holzapfel, F. Mücklich, H.P. Degischer, Three-

dimensional characterization of ‘as-cast’ and solution-treated AlSi12(Sr) alloys by
high-resolution FIB tomography, Acta Mater. 55 (2007) 3875–3882.
doi:10.1016/j.actamat.2007.03.004.
[52] F. Lasagni, A. Lasagni, C. Holzapfel, F. Mücklich, H.P. Degischer, Three

Dimensional Characterization of Unmodified and Sr-Modified Al-Si Eutectics by
FIB and FIB EDX Tomography, Adv. Eng. Mater. 8 (2006) 719–723.
doi:10.1002/adem.200500276.
[53] A.M. Samuel, E.M. Elgallad, H.W. Doty, S. Valtierra, F.H. Samuel, Effect of
metallurgical parameters on the microstructure, hardness impact properties, and
fractography of Al-(6.5-11.5) wt% Si based alloys, Mater. Des. 107 (2016) 426–439.
doi:10.1016/j.matdes.2016.06.051.
[54] S. Joseph, S. Kumar, A systematic investigation of fracture mechanisms in Al-Si

based eutectic alloy-Effect of Si modification, Mater. Sci. Eng. A. 588 (2013) 111–
124. doi:10.1016/j.msea.2013.09.019.
[55] F. Lasagni, a. Lasagni, M. Engstler, H.P. Degischer, F. Mücklich, Nano-

characterization of Cast Structures by FIB-Tomography, Adv. Eng. Mater. 10
(2008) 62–66. doi:10.1002/adem.200700249.
[56] F. Mao, G. Yan, Z. Xuan, Z. Cao, T. Wang, Effect of trace Eu addition on the
microstructures and mechanical properties of A356, J. Alloys Compd. 650 (2015)
896–906. doi:10.1016/j.jallcom.2015.06.266.
[57] A.M. Samuel, G.H. Garza-Elizondo, H.W. Doty, F.H. Samuel, Role of modification
and melt thermal treatment processes on the microstructure and tensile properties
of Al-Si alloys, Mater. Des. 80 (2015) 99–108. doi:10.1016/j.matdes.2015.05.013.
[58] C.H. Caceres, J.R. Griffiths, Damage by the cracking of silicon particles in an Al-
7Si-0.4Mg casting alloy, Acta Mater. 44 (1996) 25–33. doi:10.1016/1359-
6454(95)00172-8.
[59] Q.G. Wang, C.H. Caceres, J.R. Griffiths, Damage by eutectic particle cracking in
aluminum casting alloys A356/357, Metall. Mater. Trans. A Phys. Metall. Mater.
Sci. 34 (2003) 2901–2912. doi:10.1007/s11661-003-0190-1.
[60] A.M.A. Mohamed, F.H. Samuel, A.M. Samuel, H.W. Doty, Influence of additives
on the impact toughness of Al-10.8% Si near-eutectic cast alloys, Mater. Des. 30
(2009) 4218–4229. doi:10.1016/j.matdes.2009.04.041.
74
[61] J.F. Su, X. Nie, V. Stoilov, Characterization of fracture and debonding of Si

particles in AlSi alloys, Mater. Sci. Eng. A. 527 (2010) 7168–7175.
doi:10.1016/j.msea.2010.07.106.
[62] A. Kruglova, M. Roland, S. Diebels, T. Dahmen, P. Slusallek, F. Mücklich,

Modelling and characterization of ductile fracture surface in Al-Si alloys by means
of Voronoi tessellation, Mater. Charact. 131 (2017) 1–11.
doi:10.1016/j.matchar.2017.06.013.
[63] M. Riestra, E. Ghassemali, T. Bogdanoff, S. Seifeddine, Interactive effects of grain

refinement, eutectic modification and solidification rate on tensile properties of
Al-10Si alloy, Mater. Sci. Eng. A. 703 (2017) 270–279.
doi:10.1016/j.msea.2017.07.074.
[64] S.M. Jigajinni, K. Venkateswarlu, S.A. Kori, Effect of a grain refiner cum modifier
on mechanical properties of Al-7Si and Al-11Si alloys, Met. Mater. Int. 19 (2013)
171–181. doi:10.1007/s12540-013-2006-1.
[65] S. Tao, Y. Pan, T. Lu, Y. Chen, J. Wu, Microstructures and mechanical properties
of La added Al–Si casting alloys and mechanism of grain refinement, Int. J. Cast
Met. Res. 28 (2015) 375–381. doi:10.1179/1743133615Y.0000000011.
[66] A.M. Samuel, H.W. Doty, S. Valtierra, F.H. Samuel, Effect of grain refining and Sr-
modification interactions on the impact toughness of Al-Si-Mg cast alloys, Mater.
Des. 56 (2014) 264–273. doi:10.1016/j.matdes.2013.10.029.
[67] S. Shivkumar, S. Ricci, J.C. Keller, D. Apelian, Influence of Solution Treatment on

Tensile Properties of Cast Aluminum Alloys, J. Heat Treat. 8 (1990) 63–70.
[68] D.L. Zhang, L.H. Zheng, D.H. StJohn, Effect of a short solution treatment time on
microstructure and mechanical properties of modified Al-7wt.%Si-0.3wt.%Mg
alloy, J. Light Met. 2 (2002) 27–36. doi:10.1016/S1471-5317(02)00010-X.
[69] D.L. Zalensas, ed., Aluminum Casting Technology, 2nd ed., American
Foundrymens Society, 1993.
[70] W. Khalifa, F.H. Samuel, J.E. Gruzleski, Iron intermetallic phases in the Al corner
of the Al-Si-Fe system, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 34 (2003)
807–825. doi:10.1007/s11661-003-1009-9.
[71] T.O. Mbuya, B.O. Odera, S.P. Ng’ang’a, Influence of iron on castability and
properties of aluminium silicon alloys: Literature review, Int. J. Cast Met. Res. 16
(2003) 451–465. doi:10.1080/13640461.2003.11819622.
[72] L. Lu, A.K. Dahle, Iron-rich intermetallic phases and their role in casting defect
formation in hypoeutectic Al−Si alloys, Metall. Mater. Trans. A. 36 (2005) 8 9–
835. doi:10.1007/s11661-005-1012-4.
[73] M. Timpel, N. Wanderka, B.S. Murty, J. Banhart, Three-dimensional visualization

of the microstructure development of Sr-modified Al-15Si casting alloy using FIB-
EsB tomography, Acta Mater. 58 (2010) 6600–6608.
doi:10.1016/j.actamat.2010.08.021.
[74] C.M. Dinnis, J.A. Taylor, A.K. Dahle, As-cast morphology of iron-intermetallics in
Al-Si foundry alloys, Scr. Mater. 53 (2005) 955–958.
doi:10.1016/j.scriptamat.2005.06.028.
[75] J.A. Taylor, Iron-Containing Intermetallic Phases in Al-Si Based Casting Alloys,
75
Procedia Mater. Sci. 1 (2012) 19–33. doi:10.1016/j.mspro.2012.06.004.
[76] S.-N. Yie, S.-L. Lee, Y.-H. Lin, L. Jing-Chie, Mechanical Properties of Al-11%Si
Casting Alloys Containing Trace Be and Sr, Mater. Trans. JIM. 40 (1999) 294–
300.
[77] M. Tash, F.H. Samuel, F. Mucciardi, H.W. Doty, Effect of metallurgical parameters
on the hardness and microstructural characterization of as-cast and heat-treated
356 and 319 aluminum alloys, Mater. Sci. Eng. A. 443 (2007) 185–201.
doi:10.1016/j.msea.2006.08.054.
[78] F.H. Samuel, P. Ouellet, A.M. Samuel, H.W. Doty, Effect of Mg and Sr additions on
the formation of intermetallics in AI-6 Wt Pet Si-3.5 Wt Pet Cu-(0.45) to (0.8) Wt
Pet Fe 319-Type Alloys, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 29 (1998)
2871–2884. doi:10.1007/s11661-998-0194-y.
[79] M.H. Mulazimoglu, A. Zaluska, F. Paray, J.E. Gruzleski, The effect of strontium on
the Mg2Si precipitation process in 6201 aluminum alloy, Metall. Mater. Trans. A.
28 (1997) 1289–1295. doi:10.1007/s11661-997-0265-5.
[80] B. Otani, Silumin and its structure, J. Inst. Met. (1926) 243–267.
[81] C.E. Ransley, H. Neufeld, The solubility relationship in the aluminium-sodium and
aluminium-silicon-sodium systems, J. Inst. Met. (1950) 25–46.
[82] B.M. Thall, B. Chalmers, Modification in Aluminum-Silicon Alloys, J. Inst. Met. 77

(1949) 79–97.
[83] R.C. Plumb, J.E. Lewis, The modification of Aluminium-Silicon alloys by Sodium,
J. Inst. Met. 86 (1958) 393–400.
[84] V.L. Davies, J.M. West, Influence of small additions of sodium on the Surface
tension of aluminium and aluminium-silicon alloys, J. Inst. Met. 92 (1963) 208–
210.
[85] S.C. Flood, J.D. Hunt, Modification of Al-Si eutectic alloys with Na, Met. Sci. 15
(1981) 287–294. doi:10.1179/030634581790426813.
[86] S.M.D. Glenister, R. Elliott, Strontium modification of AI-12.7wt-%Si alloys, Met.

Sci. 15 (1981) 181–184. doi:10.1179/030634581790426732.
[87] L. Clapham, R.W. Smith, Segregation behaviour of strontium in modified and

unmodified Al-Si alloys, J. Cryst. Growth. 92 (1988) 263–270. doi:10.1016/0022-
0248(88)90458-7.
[88] M. Shamsuzzoha, L.M. Hogan, The role of non-cozonal twinning in the growth of
fibrous silicon in strontium-modified Al−Si eutectic, J. Mater. Sci. 24 ( 989)
2849–2859. doi:10.1007/BF02385637.
[89] B. Closset, H. Dugas, M. Pekguleryuz, J.E. Gruzleski, The Aluminum-Strontium

Phase Diagram, Metall. Trans. A. 17A (1986) 1250–1253.
[90] S.-M. Liang, M. Engstler, V. Groten, J. Barrirero, F. Mücklich, A. Bührig-Polaczek,

R. Schmid-Fetzer, Key experiments and thermodynamic revision of the binary Al–
Sr system, J. Alloys Compd. 610 (2014) 443–450.
doi:10.1016/j.jallcom.2014.05.018.
[91] K. Nogita, A. Knuutinen, S.D. McDonald, A.K. Dahle, Mechanisms of eutectic

solidification in Al–Si alloys modified with Ba, Ca, Y and Yb, J. Light Met. 1 (2001)
76
219–228. doi:10.1016/S1471-5317(02)00005-6.
[92] S.S.S. Kumari, R.M. Pillai, B.C. Pai, Role of calcium in aluminium based alloys and
composites, Int. Mater. Rev. 50 (2005) 216–238. doi:10.1179/174328005X14366.
[93] S.S.S. Kumari, R.M. Pillai, B.C. Pai, Structure and properties of calcium and
strontium treated Al-7Si-0.3Mg alloy: A comparison, J. Alloys Compd. 460 (2008)
472–477. doi:10.1016/j.jallcom.2007.05.085.
[94] T.H. Ludwig, J. Li, P.L. Schaffer, P. Schumacher, L. Arnberg, Refinement of

Eutectic Si in High Purity Al-5Si Alloys with Combined Ca and P Additions, Metall.
Mater. Trans. A Phys. Metall. Mater. Sci. 46 (2015) 362–376. doi:10.1007/s11661-
014-2585-6.
[95] T. Kobayashi, J.H. Kim, M. Niinomi, Effect of calcium on mechanical properties of

recycled aluminium casting alloys, Mater. Sci. Technol. 13 (1997) 497–502.
doi:10.1179/mst.1997.13.6.497.
[96] S.S. Sreeja Kumari, R.M. Pillai, B.C. Pai, A study on the structural, age hardening
and mechanical characteristics of Mn and Ca added Al-7Si-0.3Mg-0.6Fe alloy, J.
Alloys Compd. 453 (2008) 167–173. doi:10.1016/j.jallcom.2006.11.145.
[97] S.S. Sreeja Kumari, R.M. Pillai, K. Nogita, A.K. Dahle, B.C. Pai, Influence of
calcium on the microstructure and properties of an Al-7Si-O.3Mg-xFe Alloy,
Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 37 (2006) 2581–2587.
doi:10.1007/BF02586230.
[98] A. Abdollahi, J.E. Gruzleski, An evaluation of calcium as a eutectic modifier in

A357 alloy, Int. J. Cast Met. Res. 11 (1998) 145–155.
doi:10.1080/13640461.1998.11819269.
[99] J. Li, F. Hage, M. Wiessner, L. Romaner, D. Scheiber, B. Sartory, Q. Ramasse, P.

Schumacher, The roles of Eu during the growth of eutectic Si in Al-Si alloys, Sci.
Rep. 5 (2015) 13802. doi:10.1038/srep13802.
[100] K. Nogita, H. Yasuda, K. Yoshida, K. Uesugi, a. Takeuchi, Y. Suzuki, a. K. Dahle,

Determination of strontium segregation in modified hypoeutectic Al–Si alloy by
micro X-ray fluorescence analysis, Scr. Mater. 55 (2006) 787–790.
doi:10.1016/j.scriptamat.2006.07.029.
[101] G.K. Sigworth, Modification of Aluminum-Silicon Alloys, in: ASM Handbook, Vol.
15 Cast., 2008: pp. 240–254. doi:10.1361/asmhba0005301.
[102] A. Knuutinen, K. Nogita, S.D. McDonald, A.K. Dahle, Porosity formation in

aluminium alloy A356 modified with Ba, Ca, Y and Yb, J. Light Met. 1 (2001) 241–
249. doi:10.1016/S1471-5317(02)00006-8.
[103] S.D. Mcdonald, A.K. Dahle, J.A. Taylor, D.H. Stjohn, A.H. Effects, Modification-
Related Porosity Formation in Hypoeutectic Aluminum-Silicon Alloys, 35 (2004)
1097–1106.
[104] L. Lu, K. Nogita, S.D. McDonald, A.K. Dahle, Eutectic solidification and its role in
casting porosity formation, Jom. 56 (2004) 52–58. doi:10.1007/s11837-004-0254-
8.
[105] L. Liu, A.M. Samuel, F.H. Samuel, H.W. Doty, S. Valtierra, Influence of oxides on
porosity formation in Sr-treated alloys, Int. J. Met. 11 (2017) 729–742.
doi:10.1007/s40962-016-0118-3.
77
[106] N.S. Tiedje, J.A. Taylor, M.A. Easton, Feeding and distribution of porosity in cast
Al-Si alloys as function of alloy composition and modification, Metall. Mater.
Trans. A Phys. Metall. Mater. Sci. 43 (2012) 4846–4858. doi:10.1007/s11661-012-
1308-0.
[107] D. Emadi, Porosity Formation in Sr-Modified Al-Si Alloys, 1995.
[108] Z. Zhang, X. Bian, X. Liu, Effect of strontium addition on hydrogen content and
porosity shape of Al-Si alloys, Int. J. Cast Met. Res. 14 (2001) 31–35.
doi:10.1080/13640461.2001.11819422.
[109] D. Emadi, J.E. Gruzleski, J.M. Toguri, The effect of na and Sr modification on
surface tension and volumetric shrinkage of A356 alloy and their influence on
porosity formation, Metall. Trans. B. 24 (1993) 1055–1063.
doi:10.1007/BF02660997.
[110] M. De Giovanni, J.M. Warnett, M.A. Williams, P. Srirangam, 3D imaging and

quantification of porosity and intermetallic particles in strontium modified Al-Si
alloys, J. Alloys Compd. 727 (2017) 353–361. doi:10.1016/j.jallcom.2017.08.146.
[111] C.M. Dinnis, M.O. Otte, A.K. Dahle, J.A. Taylor, The influence of strontium on
porosity formation in Al-Si alloys, Metall. Mater. Trans. A Phys. Metall. Mater. Sci.
35 A (2004) 3531–3541. doi:10.1007/s11661-004-0190-9.
[112] M.M. Makhlouf, H.V. Guthy, The aluminum–silicon eutectic reaction:

mechanisms and crystallography, J. Light Met. 1 (2001) 199–218.
doi:10.1016/S1471-5317(02)00003-2.
[113] A.G.C. Gwyer, H.W.L. Phillips, The constitution and structure of the commercial
aluminium-silicon alloys, J. Inst. Met. 36 (1923) 283–324.
[114] N. Bochvar, P. Budberg, F. Hayes, Y. Liberov, R. Schmid-Fetzer, Al-Na-Si Ternary

Phase Diagram, MSI, Materials Science International Services GmbH, Stuttgart,
1993.
[115] R. Ferro, O. Kubaschewski, H. Hubert, G. Ibe, Aluminium-Silicon-Strontium,

VCH. 8 (1993) 270–278.
[116] J.A.E. Bell, W.C. Winegard, Structure of Pure Aluminium-Silicon Eutectics, J. Inst.
Met. 93 (1964) 318–319.
[117] M.G. Day, A. Hellawell, The Structure of Modified Aluminium-Silicon Eutectic

Alloy, J. Inst. Met. 95 (1967) 377.
[118] H.A.H. Steen, A. Hellawell, The growth of eutectic silicon-contributions to

undercooling, Acta Metall. 23 (1975) 529–535. doi:10.1016/0001-6160(75)90093-
0.
[119] M.D. Hanna, S.-Z. Lu, A. Hellawell, Modification in the aluminum silicon system,
Metall. Trans. A. 15 (1984) 459–469. doi:10.1007/BF02644969.
[120] F. Yilmaz, O.A. Atasoy, R. Elliott, Growth structures in aluminium-silicon alloys II.
The influence of strontium, J. Cryst. Growth. 118 (1992) 377–384.
doi:10.1016/0022-0248(92)90086-X.
[121] J.H. Li, M.Z. Zarif, G. Dehm, P. Schumacher, Influence of impurity elements on
the nucleation and growth of Si in high purity melt-spun Al–Si-based alloys,
Philos. Mag. 92 (2012) 3789–3805. doi:10.1080/14786435.2012.687840.
78
[122] S.D. McDonald, A.K. Dahle, J.A. Taylor, D.H. StJohn, Eutectic grains in
unmodified and strontium-modified hypoeutectic aluminum-silicon alloys, Metall.
Mater. Trans. A. 35 (2004) 1829–1837. doi:10.1007/s11661-004-0091-y.
[123] S.D. McDonald, K. Nogita, A.K. Dahle, Eutectic nucleation in Al–Si alloys, Acta
[124] A.K. Dahle, K. Nogita, S.D. McDonald, C. Dinnis, L. Lu, Eutectic modification and
microstructure development in Al–Si Alloys, Mater. Sci. Eng. A. 413–414 (2005)
243–248. doi:10.1016/j.msea.2005.09.055.
[125] D.M. Stefanescu, ed., Modification of Aluminum-Silicon Alloys, in: ASM Met.
Handb. - Vol. 15, Fourth pri, ASM INternational, 1998: pp. 1052–1064.
[126] S.-Z. Lu, A. Hellawell, Modification of Al-Si Alloys: Microstructure, Thermal

Analysis, and Mechanisms, JOM. (1995) 38–40.
[127] C.R. Ho, B. Cantor, Modification of hypoeutectic Al-Si alloys, J. Mater. Sci. 30
(1995) 1912–1920. doi:10.1007/BF00353013.
[128] T.H. Ludwig, P.L. Schaffer, L. Arnberg, Influence of phosphorus on the nucleation
of eutectic silicon in Al-Si alloys, Metall. Mater. Trans. A Phys. Metall. Mater. Sci.
44 (2013) 5796–5805. doi:10.1007/s11661-013-1945-y.
[129] S.-M. Liang, R. Schmid-Fetzer, Phosphorus in Al–Si cast alloys: Thermodynamic

prediction of the AlP and eutectic (Si) solidification sequence validated by
microstructure and nucleation undercooling data, Acta Mater. 72 (2014) 41–56.
doi:10.1016/j.actamat.2014.02.042.
[130] S.M. Liang, R. Schmid-Fetzer, Validated thermodynamic prediction of AlP and

eutectic (Si) solidification sequence in Al-Si cast alloys, IOP Conf. Ser. Mater. Sci.
Eng. 117 (2016). doi:10.1088/1757-899X/117/1/012003.
[131] J. Eiken, M. Apel, S.-M. Liang, R. Schmid-Fetzer, Impact of P and Sr on

solidification sequence and morphology of hypoeutectic Al–Si alloys: Combined
thermodynamic computation and phase-field simulation, Acta Mater. 98 (2015)
152–163. doi:10.1016/j.actamat.2015.06.056.
[132] D.L. Zhang, B. Cantor, Heterogeneous nucleation of solidification of Si by solid AI

in hypoeutectic Al-Si alloy, Metall. Trans. A. 24 (1993) 1195–1204.
doi:10.1007/BF02657251.
[133] M. Zarif, B. Mckay, P. Schumacher, Study of Heterogeneous Nucleation of Eutectic

Si in High-Purity Al-Si Alloys with Sr Addition, Metall. Mater. Trans. A. 42 (2010)
1684–1691. doi:10.1007/s11661-010-0553-3.
[134] J.H. Li, M.Z. Zarif, M. Albu, B.J. McKay, F. Hofer, P. Schumacher, Nucleation
kinetics of entrained eutectic Si in Al-5Si alloys, Acta Mater. 72 (2014) 80–98.
doi:10.1016/j.actamat.2014.03.030.
[135] C.M. Dinnis, A.K. Dahle, J.A. Taylor, Three-dimensional analysis of eutectic grains
in hypoeutectic Al-Si alloys, Mater. Sci. Eng. A. 392 (2005) 440–448.
doi:10.1016/j.msea.2004.10.037.
[136] P. Srirangam, S. Chattopadhyay, a. Bhattacharya, S. Nag, J. Kaduk, S. Shankar, R.

Banerjee, T. Shibata, Probing the local atomic structure of Sr-modified Al–Si
alloys, Acta Mater. 65 (2014) 185–193. doi:10.1016/j.actamat.2013.10.060.
79
[137] Y.H. Cho, H.-C. Lee, K.H. Oh, a. K. Dahle, Effect of Strontium and Phosphorus on
Eutectic Al-Si Nucleation and Formation of β-Al5FeSi in Hypoeutectic Al-Si
Foundry Alloys, Metall. Mater. Trans. A. 39 (2008) 2435–2448.
doi:10.1007/s11661-008-9580-8.
[138] S. Shankar, Y.W. Riddle, M.M. Makhlouf, Nucleation mechanism of the eutectic
phases in aluminum–silicon hypoeutectic alloys, Acta Mater. 52 (2004) 4447–
4460. doi:10.1016/j.actamat.2004.05.045.
[139] S. Shankar, Y.W. Riddle, M.M. Makhlouf, Eutectic solidification of aluminum-

silicon alloys, Metall. Mater. Trans. A. 35 (2004) 3038–3043. doi:10.1007/s11661-
004-0048-1.
[140] A.K. Dahle, K. Nogita, S.D. McDonald, J.W. Zindel, L.M. Hogan, Eutectic
nucleation and growth in hypoeutectic Al-Si alloys at different strontium levels,
Metall. Mater. Trans. A. 32 (2001) 949–960. doi:10.1007/s11661-001-0352-y.
[141] K. Nogita, A.K. Dahle, Eutectic solidification in hypoeutectic Al-Si alloys: Electron
backscatter diffraction analysis, Mater. Charact. 46 (2001) 305–310.
doi:10.1016/S1044-5803(00)00109-1.
[142] J. Manickaraj, A. Gorny, Z. Cai, S. Shankar, X-ray nano-diffraction study of Sr

intermetallic phase during solidification of Al-Si hypoeutectic alloy, Appl. Phys.
Lett. 104 (2014). doi:10.1063/1.4865496.
[143] C.J. Simensen, Ø. Nielsen, F. Hillion, J. Voje, NanoSIMS Analysis of Trace

Element Segregation during the Al-Si Eutectic Reaction, Metall. Mater. Trans. A.
38 (2007) 1448–1451. doi:10.1007/s11661-007-9165-y.
[144] M. Timpel, N. Wanderka, G.S. Vinod Kumar, J. Banhart, Microstructural

investigation of Sr-modified Al-15 wt%Si alloys in the range from micrometer to
atomic scale., Ultramicroscopy. 111 (2011) 695–700.
doi:10.1016/j.ultramic.2010.12.023.
[145] M. Timpel, N. Wanderka, R. Schlesiger, T. Yamamoto, N. Lazarev, D. Isheim, G.

Schmitz, S. Matsumura, J. Banhart, The role of strontium in modifying
aluminium–silicon alloys, Acta Mater. 60 (2012) 3920–3928.
doi:10.1016/j.actamat.2012.03.031.
[146] M. Timpel, N. Wanderka, R. Schlesiger, T. Yamamoto, D. Isheim, G. Schmitz, S.

Matsumura, J. Banhart, Sr-Al-Si co-segregated regions in eutectic Si phase of Sr-
modified Al-10Si alloy., Ultramicroscopy. 132 (2013) 216–21.
doi:10.1016/j.ultramic.2012.10.006.
[147] J. Barrirero, M. Engstler, F. Mücklich, Atom Probe analysis of Sr distribution in

AlSi foundry alloys, in: Light Met. 2013 - TMS, 2013: pp. 291–296.
doi:10.1002/9781118663189.ch50.
[148] A.M. Garay-Tapia, A.H. Romero, G. Trapaga, R. Arróyave, First-principles

investigation of the Al–Si–Sr ternary system: Ground state determination and
mechanical properties, Intermetallics. 21 (2012) 31–44.
doi:10.1016/j.intermet.2011.09.001.
[149] J.I. Goldstein, C.E. Lyman, D.E. Newbury, E. Lifshin, P. Echlin, L. Sawyer, D.C.
Joy, J.R. Michael, Scanning Electron Microscopy and X-Ray Microanalysis, Third
Edit, Kluwer Academic / Plenum Publishers, New York, NY, 2003.
[150] R.R. Keller, R.H. Geiss, Transmission EBSD from 10 nm domains in a scanning
80
electron microscope, J. Microsc. 245 (2012) 245–251. doi:10.1111/j.1365-

2818.2011.03566.x.
[151] S. Suzuki, Features of transmission EBSD and its application, Jom. 65 (2013)
1254–1263. doi:10.1007/s11837-013-0700-6.
[152] S. Suwas, R.K. Ray, Crystallographic Texture of Materials, Eng. Mater. Process. 88
(2014) 151–177. doi:10.1007/978-1-4471-6314-5.
[153] C. Kranenberg, D. Johrendt, A. Mewis, R. Pöttgen, G. Kotzyba, C. Rosenhahn, B.D.

Mosel, Structure and properties of the compounds LnAl2X2 (Ln=Eu, Yb; X= Si,
Ge ), Solid State Sci. 2 (2000) 215–222. doi:10.1016/S1293-2558(00)00135-7.
[154] S. Bobev, P.H. Tobash, V. Fritsch, J.D. Thompson, M.F. Hundley, J.L. Sarrao, Z.
Fisk, Ternary rare-earth alumo-silicides - Single-crystal growth from Al flux,
structural and physical properties, J. Solid State Chem. 178 (2005) 2091–2103.
doi:10.1016/j.jssc.2005.04.021.
[155] P. Villars, K. Cenzual, eds., YbAl2Si2 Crystal Structure: Datasheet from “PAULING
FILE Multinaries Edition – 20 2” in SpringerMaterials
(https://materials.springer.com/isp/crystallographic/docs/sd_1210515), (2016).
[156] R. Gomer, Field emission, field ionization, and field desorption, Surf. Sci. 299–
300 (1994) 129–152. doi:10.1016/0039-6028(94)90651-3.
[157] K. Thompson, D. Lawrence, D.J. Larson, J.D. Olson, T.F. Kelly, B. Gorman, In situ
site-specific specimen preparation for atom probe tomography., Ultramicroscopy.
107 (2007) 131–9. doi:10.1016/j.ultramic.2006.06.008.
[158] E. Müller, Field Desorption, Phys. Rev. 102 (1956) 618–624.

doi:10.1103/PhysRev.102.618.
[159] R. Gomer, Field desorption, J. Chem. Phys. 31 (1959) 341–345.

doi:10.1063/1.1730354.
[160] R. Gomer, L.W. Swanson, Theory of field desorption, J. Chem. Phys. 38 (1963)
1613–1629. doi:10.1063/1.1776932.
[161] R. Haydock, D.R. Kingham, Post-Ionization of Field-Evaporated Ions, Phys. Rev.

Lett. 44 (1980) 1520–1523. doi:10.1103/PhysRevLett.44.1520.
[162] D.R. Kingham, The post-ionization of field evaporated ions: A theoretical

explanation of multiple charge states, Surf. Sci. 116 (1982) 273–301.
doi:10.1016/0039-6028(82)90434-4.
[163] R. Forbes, Field evaporation theory: a review of basic ideas, Appl. Surf. Sci. 87–88
(1995) 1–11. doi:10.1016/0169-4332(94)00526-5.
[164] B. Gault, M.P. Moody, J.M. Cairney, S.P. Ringer, Atom Probe Microscopy,
Springer New York, New York, NY, 2012. doi:10.1007/978-1-4614-3436-8.
[165] P. Bas, A. Bostel, B. Deconihout, D. Blavette, A general protocol for the

reconstruction of 3D atom probe data, Appl. Surf. Sci. 87–88 (1995) 298–304.
doi:10.1016/0169-4332(94)00561-3.
[166] D. Blavette, A. Bostel, J.M. Sarrau, B. Deconihout, A. Menand, An atom probe for
three-dimensional tomography, Nature. 363 (1993) 432–435.
doi:10.1038/363432a0.
81
[167] D.J. Larson, T.J. Prosa, R.M. Ulfig, B.P. Geiser, T.F. Kelly, Local Electrode Atom
Probe Tomography, Springer New York, New York, NY, 2013. doi:10.1007/978-1-
4614-8721-0.
[168] O.C. Hellman, J.A. Vandenbroucke, J. Rüsing, D. Isheim, D.N. Seidman, Analysis
of Three-dimensional Atom-probe Data by the Proximity Histogram, Microsc.
Microanal. 6 (2000) 437–444. doi:10.1007/s100050010051.
[169] W. Lefebvre, F. Danoix, G. Da Costa, F. De Geuser, H. Hallem, A. Deschamps, M.

Dumont, 3DAP measurements of Al content in different types of precipitates in
aluminium alloys, Surf. Interface Anal. 39 (2007) 206–212. doi:10.1002/sia.2516.
[170] F. Vurpillot, A. Bostel, D. Blavette, Trajectory overlaps and local magnification in

three-dimensional atom probe, Appl. Phys. Lett. 76 (2000) 3127.
doi:10.1063/1.126545.
[171] B. Gault, F. de Geuser, L. Bourgeois, B.M. Gabble, S.P. Ringer, B.C. Muddle, Atom
probe tomography and transmission electron microscopy characterisation of
precipitation in an Al-Cu-Li-Mg-Ag alloy., Ultramicroscopy. 111 (2011) 683–9.
doi:10.1016/j.ultramic.2010.12.004.
[172] G. Sha, H. Möller, W.E. Stumpf, J.H. Xia, G. Govender, S.P. Ringer, Solute
nanostructures and their strengthening effects in Al–7Si–0.6Mg alloy F357, Acta
[173] J. Rüsing, J.T. Sebastian, O.C. Hellman, D.N. Seidman, Three-dimensional

Investigation of Ceramic / Metal Heterophase Interfaces by Atom-probe
Microscopy, Microsc. Microanal. 6 (2000) 445–451. doi:10.1007/s100050010050.
[174] Y. Suzuki-Yamamoto, R. Ozaki, M. Yoshiya, T. Nagira, H. Yasuda, Yet Another

Marked Difference among Impurities as Modifier Elements for Refinement of
Eutectic Si in Al-Si Alloys, 56 (2015) 1475–1483.
doi:10.2320/matertrans.MA201576.
[175] M.A. González, Force fields and molecular dynamics simulations, École
Thématique La Société Française La Neutron. 12 (2011) 169–200.
doi:10.1051/sfn/201112009.
[176] X.Z. Gao, M.H. Müser, L.T. Kong, J.F. Li, Atomic structure and energetics of
amorphous-crystalline CuZr interfaces: A molecular dynamics study, Model.
Simul. Mater. Sci. Eng. 22 (2014). doi:10.1088/0965-0393/22/6/065007.
82
Papers
The papers associated with this thesis have been removed for
copyright reasons. For more details about these see:
http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-160235
FACULTY OF SCIENCE AND ENGINEERING
Linköping Studies in Science and Technology, Dissertation No. 2014, 2019
Department of Physics, Chemistry and Biology (IFM)
Linköping University
SE-581 83 Linköping, Sweden
www.liu.se

Eutectic Modification of Al-Si Casting Alloys

Uploaded by

Copyright:

Available Formats

Eutectic Modification of Al-Si Casting Alloys

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eutectic Modification of Al-Si Casting Alloys

Uploaded by

Copyright:

Available Formats

Linköping Studies in Science and Technology

Dissertation No. 2014

Eutectic Modification of Al-Si casting alloys

In this investigation, atom probe tomography and transmission electron microscopy

The mechanical properties and, consequently, the performance of an alloy can be

De mekaniska egenskaperna och följaktligen prestandan hos en legering kan kontrolleras

Denna nya förståelse bidrar till en framtida kontroll av mikrostrukturutvecklingen av

Eine deutliche Reduzierung des Fahrzeuggewichts ist durch eine geeignete

Die mechanischen Eigenschaften und damit die Leistungsfähigkeit einer Legierung

Comparison of segregations formed in unmodified and Sr-modified Al-Si alloys studied

Cluster formation at the Si/liquid interface in Sr and Na modified Al–Si alloys.

Eutectic modification by ternary compound cluster formation in Al-Si alloys.

Nucleation and Growth of Eutectic Si in Al-Si Alloys with Na Addition.

Liang S-M, Engstler M, Groten V, Barrirero J, Mücklich F, Bührig-Polaczek A, Schmid-

Mücklich F, Engstler M, Britz D, Barrirero J, Rossi P. Why We Need All Dimensions to

Li JH, Barrirero J, Sha G, Aboulfadl H, Mücklich F, Schumacher P. Precipitation

Shulumba N, Hellman O, Raza Z, Alling B, Barrirero J, Mücklich F, Abrikosov I A, Odén

Yalamanchili K, Wang F, Aboulfadl H, Barrirero J, Rogström L, Jiménez-Pique E,

Menezes CM, Bogoni N, Barrirero J, Aboulfadl H, Mücklich F, Figueroa CA. Influence of

Roa JJ, Aboulfadl H, Barrirero J, Turon-Vinas M, Mücklich F, Anglada M. Chemical

El Azhari I, Barrirero J, García J, Soldera F, Llanes L, Mücklich F. Atom Probe

I am particularly grateful to Roberto Tannchen who stood wholehearted by my side

Thank you to Corinna Markmann for her inspiring light.

A Solidification interface area

Popular Summary .............................................................................................................. vii

Populärvetenskaplig sammanfattning ................................................................................ ix

Included papers ................................................................................................................. xv

My contribution to the papers ......................................................................................... xvii

Other papers not included in the thesis............................................................................ xix

Acknowledgements ........................................................................................................... xxi

Acronyms and symbols ................................................................................................... xxiii

1.1 Scope ........................................................................................................................... 2

1.2 Outline of the thesis .................................................................................................... 2

2. Aluminum casting alloys .............................................................................................. 5

2.1 Al-Si alloys ..................................................................................................................8

3. Eutectic solidification .................................................................................................. 11

4. Eutectic solidification of Al-Si alloys .......................................................................... 17

4.1 Plate-like structure and TPRE mechanism ..............................................................20

5. Eutectic modification of Al-Si alloys .......................................................................... 25

5.1 Enhancement of mechanical Properties ................................................................... 25

5.3 Porosity and shrinkage ............................................................................................. 31

5.4 Theories explaining eutectic modification of Al-Si alloys ........................................ 32

5.4.1 Early theories for eutectic modification............................................................. 32

5.4.2 Modification effect on eutectic growth .............................................................. 35

5.4.3 Modification effect on eutectic nucleation ........................................................ 39

5.4.4. Recent developments........................................................................................ 46

6. Sample preparation and Characterization ................................................................. 47

6.1.1 Al – 7 wt% Si ....................................................................................................... 47

6.1.2 Al – 5 wt% Si ...................................................................................................... 47

6.2 Scanning electron microscopy – SEM .....................................................................48

6.2.1 Secondary electron imaging – SE ......................................................................48

6.2.2 Backscattered electron imaging – BSE ............................................................. 49

6.2.3 Energy dispersive X-ray spectroscopy – EDX ...................................................50

6.2.4 Electron backscattered diffraction – EBSD ...................................................... 51

6.3 Focused ion beam – FIB .......................................................................................... 53

6.3.1 Sample preparation for TEM ............................................................................. 54

6.3.2 Sample preparation for APT .............................................................................. 55

6.4 Transmission electron microscopy – TEM .............................................................. 56