6RemovalofIronFromAluminum Review

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/232878267
Removal of Iron From Aluminum: A Review
Article · March 2012

DOI: 10.1080/08827508.2010.542211
CITATIONS READS
129 10,576
4 authors, including:
Lucas Damoah
38 PUBLICATIONS 454 CITATIONS
SEE PROFILE
All content following this page was uploaded by Lucas Damoah on 19 July 2014.
The user has requested enhancement of the downloaded file.

This article was downloaded by: [Missouri S & T], [Lucas Damoah]
On: 28 December 2011, At: 09:01
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Mineral Processing and Extractive

Metallurgy Review
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/gmpr20
Removal of Iron From Aluminum: A

Review
a b b c b
Lifeng Zhang , Jianwei Gao , Lucas Nana Wiredu Damoah &
b
David G. Robertson
a
School of Metallurgical and Ecological Engineering, University of
Science and Technology Beijing, Beijing, China
b
Department of Materials Science and Engineering, Missouri
University of Science and Technology, Rolla, Missouri, USA
c
Shanghai Institute of Quality Inspection and Technical Research,
Shanghai, China
Available online: 01 Jun 2011
To cite this article: Lifeng Zhang, Jianwei Gao, Lucas Nana Wiredu Damoah & David G. Robertson
(2012): Removal of Iron From Aluminum: A Review, Mineral Processing and Extractive Metallurgy
Review, 33:2, 99-157
To link to this article: http://dx.doi.org/10.1080/08827508.2010.542211
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.
Mineral Processing & Extractive Metall. Rev., 33: 99–157, 2012
Copyright # Taylor & Francis Group, LLC
ISSN: 0882-7508 print=1547-7401 online
DOI: 10.1080/08827508.2010.542211
REMOVAL OF IRON FROM ALUMINUM: A REVIEW
Lifeng Zhang1,2, Jianwei Gao2,3, Lucas Nana

Wiredu Damoah2, and David G. Robertson2
1
School of Metallurgical and Ecological Engineering, University of Science
Downloaded by [Missouri S & T], [Lucas Damoah] at 09:01 28 December 2011
and Technology Beijing, Beijing, China

2
Department of Materials Science and Engineering, Missouri University of
Science and Technology, Rolla, Missouri, USA
3
Shanghai Institute of Quality Inspection and Technical Research, Shanghai,
China
In this paper, the Fe-rich phases in and their detrimental effect on aluminum alloys are
summarized. The existence of brittle platelet b-Fe-rich phases lowers the mechanical
properties of aluminum alloys. The methods to neutralize the detrimental effect of iron
are discussed. The use of high cooling rate, solution heat treatment, and addition of ele-
ments such as Mn, Cr, Be, Co, Mo, Ni, V, W, Cu, Sr, or the rare earth elements Y, Nd,
La, and Ce are reported to modify the platelet Fe-rich phases in aluminum alloys. The
mechanism of the modification is briefly described. Technologies to remove iron from alumi-
num are reviewed extensively. The precipitation and removal of Fe-rich phases (sludge) are
discussed. The dense phases can be removed by methods such as gravitational separation,
electromagnetic (EM) separation, and centrifuge. Other methods include electrolysis,
electro-slag refining, fractional solidification, and fluxing refining. The expensive three-
layer cell electrolysis process is the most successful technique to remove iron from aluminum
so far.
Keywords: alloys, aluminum, iron removal, modification
INTRODUCTION
One of the greatest challenges to aluminum recycling is the widely varying
composition of scraps. Most scraps are mixture of alloys with a wide range of com-
positions. Table 1 (Gesing et al. 2002) shows the representative composition of pre-
sorted wrought and cast alloy scraps. Particularly, iron is a serious challenge because
of its gradual accumulation during repeated scrap recycling (Green 2007). Also, the
difficulty (and therefore, cost) to remove it from the aluminum alloy increases with
decreasing Fe content (Crepeau 1995). Most aluminum alloy production requires
tight composition controls on iron. For example, iron content level above 0.15 wt%
is unacceptable in premium aerospace alloys such as 7050, 7055, and 7475, and high
performance automotive alloys, like 5474 and 6111, also restrict both iron and
Address correspondence to Lifeng Zhang, Professor, School of Metallurgical and Ecological

Engineering, University of Science and Technology Beijing, Beijing 100083, China. E-mail:
zhanglifeng@ustb.edu.cn
99
100 L. ZHANG ET AL.
Table 1. Representative composition of presorted wrought and cast scraps (Gesing et al. 2002)
Al Cu Fe Mg Mn Si Zn Others
Wrought 1 97.1 0.11 0.59 0.82 0.21 0.51 0.45 0.19

Wrought 2 96.7 0.30 0.60 0.60 0.20 0.90 0.50 0.10
Wrought 3 93.1 0.95 1.01 0.89 0.12 2.41 1.25 0.27
Wrought 4 93.1 1.20 0.70 0.70 0.30 2.60 1.20 0.20
Cast 1 83.5 4.40 1.10 0.40 0.30 8.0 1.90 0.40
Cast 2 86.0 3.90 1.00 0.10 0.20 6.30 2.30 0.30
Cast 3 88.4 2.50 0.75 0.58 0.26 5.18 1.27 1.09
Mixed wrought and cast 90.1 2.30 0.80 0.50 0.20 4.50 1.20 0.30
silicon to 0.40 wt% as maximum. Thus, secondary aluminum alloys are not sources
for producing premium quality aerospace alloys (Starke and Staley 1996). The recy-
cling and refining of scraps must reduce the impurities, especially iron, to an accept-
able level. Iron is the most pervasive impurity element in aluminum alloys, which
stems from the bauxite and steel tools used during both primary and secondary pro-
duction. Iron usually forms second phases in the aluminum alloys owing to its low
equilibrium solid solubility in the aluminum (max. 0.05%), such as Al3Fe, a-AlFeSi,
and b-AlFeSi (Khalifa, Samuel, and Gruzleski 2003). These Fe-rich phases have an
appreciable deleterious effect on the mechanical properties of the alloy. The ductility
and tensile strength progressively decreases with the increasing Fe content. Parti-
cularly, the ductility decreases dramatically after a critical Fe content is surpassed.
The techniques applied to lessen the deleterious effect of iron have been the subjects
of extensive researches for decades. The largest efforts have been made on the
suppression of the detrimental impact of the iron phases, for example, structural
modification by adding some chemical reagents, such as Mn, Co, and Cr (Flores
et al. 1998; Yang and Yin 2000; Yin et al. 2000; Shabestari et al. 2002; Mahta
et al. 2005). However, these metals have only a partial effect because they increase
the total amount of the Fe-rich phases, which is also detrimental to Al alloys. Other
elements like Be, Sr, and Ni have an influence on Fe-rich phases too but they are far
less effective than Mn. There are no effective practical methods to directly remove
iron from aluminum alloys by conventional refining. The techniques or theories
on direct iron removal from aluminum have made no satisfactory progress so far.
Thus, the study on effective removal of iron from aluminum is urgent. Nevertheless,
iron is present in most traditional die casting alloys as an impurity, yet a very useful
impurity (Mbuya, Odera, and Ng’ang’a 2003; Apelian 2009). A minimum of 0.8% Fe
is beneficial to the high pressure die casting alloys, such as (e.g., 380, A380, C380,
A383, A384, 360, A360, etc.) to prevent die sticking (i.e., soldering of the alloy to
the die), improve die casting surface finish and improved resistance to hot tearing
during solidification (Wang, Makhlouf, and Apelian 1995). However, concerning
Fe improving the resistance to hot tearing of the alloy there are conflicting reports
(Apelian 2009). The Al-Fe-Si ternary eutectic composition occurs at about 0.8%
Fe. Apalian reported that, theoretically, when iron is alloyed to a little above that
amount, the supersaturated molten alloy has little tendency to dissolve the relatively
unprotected tool steel while the molten alloy and die are in intimate contact (Apelian
2009).
REMOVAL OF IRON FROM ALUMINUM 101
However, in high-integrity die casting variations such as high-vacuum, squeeze,

and semi solid casting iron is not tolerated at high concentrations due to the forma-
tion of b-Al5FeSi that destroys ductility (a major goal in those alloys) by providing
numerous stress raisers and points of crack initiation (Apelian 2009).
In this paper, the following topics are reviewed
. Iron-containing intermetallic phases in aluminum alloys

. The detrimental effect of iron on the mechanical properties of aluminum alloys
. Methods to modify Fe-rich phases
. Methods to remove iron from aluminum alloys
Fe-RICH PHASES IN ALUMINUM AND ALUMINUM-SILICON ALLOYS

Since the solid solubility of iron in aluminum is less than 0.05% at equilibrium,
almost all iron forms second phases in aluminum (Mondolfo 1976). The binary
Al-Fe and ternary Al-Fe-Si phases are the main Fe-rich phases in aluminum alloys
(Khalifa, Samuel, and Gruzleski 2003). Hence, the scope of this discussion is mainly
about the Al-Fe-Si system.
A number of Fe-rich phases in Al-Fe-Si ternary system have been identified as
shown in Table 2. Al3Fe (also reported as h-Al3Fe or h-Al13Fe4 (Mondolfo 1976;
Stefaniay, Griger, and Turmezey 1987)), a common equilibrium phase, forms an
eutectic with aluminum at about 655 C (Khalifa, Samuel, and Gruzleski 2003).
The metastable Al6Fe also forms an eutectic with aluminum (Young and Clyne
1981), and its eutectic temperature is a few degree lower than that of the Al-Al3Fe
eutectic (Jones 1969; Backerud, Chai, and Tamminen 1990). The transition between
Al-Al3Fe and Al-Al6Fe eutectic reaction is common. Moreover, the typical micro-
structure, precipitated morphologies and Fe content of Al6Fe and Al3Fe are very
similar (Young and Clyne 1981). Thus, it is difficult to identify between the Al6Fe
and Al3Fe. Metastable phases AlmFe [m 4.0 4.4 (Allen et al. 1998)], Al6Fe, and
AlxFe [x 4.5 or 5 (Allen et al. 1998)] usually form at a high cooling rate (Kosuge
and Mizukami 1975; Miki, Kosuge, and Nagahama 1975; Mondolfo 1976; Young
and Clyne 1981). AlmFe is reported to be stable at the cooling rate above 10, 18,
or 20 C=s; Al6Fe is stable at the cooling rate in the ranges 1–10 C=s, 3–18 C=s,
and 2–20 C=s; AlxFe is stable at 0.5—6 C=s (Khalifa, Samuel, and Gruzleski
2003). However, AlmFe, Al6Fe, and AlxFe were also observed at a slow cooling rate
0.16 C=s in low Fe=Si Al alloy (0.35 wt% Si, 0.23 wt% Fe). While Al3Fe, supposed to
be stable at the cooling rate below 1 C=s (Liu, Thorvaldsson, and Dunlop 1986),
was not found; Khalifa stated that higher content of Si (Fe=Si < 1) stabilized AlmFe,
Al6Fe, and AlxFe and destabilized the Al3Fe at slow cooling rates as shown in
Figure 1 (Khalifa, Samuel, and Gruzleski 2003). AlpFe [p 4.5 (Allen et al. 1998)]
was observed by Liu in the Al alloy containing 0.511 wt% Fe and 0.127 wt% Si at
a high cooling rate 10 C=s (Liu, Thorvaldsson, and Dunlop 1986).
q1-AlFeSi was observed in alloys containing 0.23 wt% Fe þ 0.49 wt% Si or
1.03 wt% Fe þ 0.62 wt% Si alloy at a high cooling rate of 14 C=s as shown in
Figure 2 (Khalifa, Samuel, and Gruzleski 2003). Liu found q1-AlFeSi in the alloy
with 0.25 wt% Fe, 0.125 wt% Si at the cooling rate 10 C=s and q2-AlFeSi was
102 L. ZHANG ET AL.
Table 2. Identified Fe-rich phases in aluminum alloys
Fe-rich intermetallics Crystal structure Reference
AlmFe Bct (Liu, Thorvaldsson, and Dunlop 1986; Khalifa,

Samuel, and Gruzleski 2003)
Al6Fe Orthorhombic (Mondolfo 1976; Khalifa, Samuel, and Gruzleski
C-centered 2003)
orthorhombic (Liu, Thorvaldsson, and Dunlop 1986)
AlxFe (Young and Clyne 1981; Khalifa, Samuel, and
Gruzleski 2003)
AlqFe Bcc (Liu, Thorvaldsson, and Dunlop 1986)
h-Al3Fe or h-Al13Fe4 Monoclinic (Mondolfo 1976; Khalifa, Samuel, and Gruzleski
C-centered 2003; Shabestari 2004)

monoclinic (Liu et al. 1986)
a-Al8Fe2Si or Hexagonal (Mondolfo 1976; Stefaniay, Griger, and Turmezey
a-Al12Fe3Si2 Bcc 1987; Crepeau 1995; Khalifa, Samuel, and
a-Al15Fe3Si2 Gruzleski 2003; Shabestari 2004)
(Crepeau 1995) (Cooper 1967; Skjerpe 1987; Kral, Mcintyre, and
Smillie 2004)
b-Al5FeSi Monoclinic (Mondolfo 1976; Stefaniay, Griger, and Turmezey
Al9Fe2Si2 (Philips 1976) B-face centered 1987; Crepeau 1995; Khalifa, Samuel, and
orthorhombic Gruzleski 2003; Shabestari 2004; Lu and Dahle
Orthorhombic 2005)
Tetragonal (Carpenter and LePage 1993)
(Zheng et al. 2000)
(Kral, Mcintyre, and Smillie 2004)
d-Al4FeSi2 Tetragonal (Mondolfo 1976; Crepeau 1995; Khalifa, Samuel,
d-Al3FeSi2 (Panday and and Gruzleski 2003; Kral, Nakashima, and
Schubert 1969; Kral, Mitchell 2006)
Nakashima, and
Mitchell 2006)
q1-AlFeSi C-centered (Liu, Thorvaldsson, and Dunlop 1986; Khalifa,
orthorhombic Samuel, and Gruzleski 2003)
q2-AlFeSi monoclinic (Liu et al. 1986)
c-Al8FeSi C-centered (Mondolfo 1976; Stefaniay, Griger, and Turmezey
monoclinic 1987)
p-Al8Mg3FeSi6 (Crepeau 1995)
p-Al8Mg3FeSi6 (Barresi et al. 1993)
observed after further heat treatment at 600 C for 24 h, as shown in Figure 3 (Liu,
Thorvaldsson, and Dunlop 1986). c-Al8FeSi was reported to usually exist in the high
silicon and high iron Al alloy (Mondolfo 1976). Plate-shaped d-Al4FeSi2 was
reported in earlier references (Phragmen 1950; Mondolfo 1976). Mondolfo reported
that d-Al4FeSi2 formed in high Si content Al alloys (Mondolfo 1976). However,
Khalifa observed the d-Al4FeSi2 phases in low Si alloys at a cooling rate >10 C=s
in alloys with Fe content ranging from 0.23 to 0.56 wt% and Si content ranging
from 0.35 to 0.9 wt%, as shown in Figure 4 (Khalifa, Samuel, and Gruzleski
2003). d-Al4FeSi2 was also described as d-Al3FeSi2 (Kral, Nakashima, and Mitchell
2006).
The most important Fe-rich phases in aluminum alloys containing silicon
are b-phase and a-phase. The a-phase is identified most commonly as a-Al8Fe2Si,
Figure 1 Morphology of AlmFe, Al6Fe, and AlxFe phases in Al alloys (0.35Si, 0.23Fe) at cooling rate
0.16 C=s (Khalifa, Samuel, and Gruzleski 2003).
a-Al12Fe3Si2 (Mondolfo 1976), a-Al15Fe3Si2 (Crepeau 1995) or generally a-AlFeSi

(Liu, Thorvaldsson, and Dunlop 1986; Stefaniay, Griger, and Turmezey 1987; Kha-
lifa, Samuel, and Gruzleski 2003). There is a contradiction about the structure of the
a-phase. Skjerpe (1987) proved the a-phase was body-centered cubic and Cooper
(1967) described it as Al19Fe4MnSi2, with space group Im3, and a ¼ 1.256 nm. Kral
substantially proved it as Al19 (FeMn)5Si2, space group Im3, a ¼ 1.256 nm (Kral,
Mcintyre, and Smillie 2004; Kral 2005). While in other references, the structure of
the a-phase was reported as hexagonal (Mondolfo 1976; Stefaniay, Griger, and
Turmezey 1987; Crepeau 1995). The a-phase has a compact morphology such as Chi-
nese script, star like and polygon [shown in Figure 5 (Lu and Dahle 2005) and
Figure 6 (Dinnis, Taylor, and Dahle 2005)], which was thought to be much less
harmful than the platelet b-Al5FeSi to the mechanical properties of Al alloys. Dinnis
found the a-phase exhibited two distinct morphologies: coupled Al-a-AlFeSi eutectic
with convoluted arm structure and polyhedral crystal as seen from Figure 6 (Dinnis,
Figure 2 Morphology of q1-AlFeSi phase in the aluminum alloy with 0.23 wt% Fe þ 0.49 wt% Si or
1.03 wt% Fe þ 0.62 wt% Si at 14 C=s (Khalifa, Samuel, and Gruzleski 2003).
104 L. ZHANG ET AL.
Figure 3 TEM morphology of q2-AlFeSi phase in an aluminum alloy (Liu, Thorvaldsson, and Dunlop
1986).
Taylor, and Dahle 2005) when it precipitates as primary phase (Dinnis, Taylor, and
Dahle 2005). The platelet b-phase is usually identified as b-Al5FeSi (Crepeau 1995;
Mondolfo 1976; Cao and Campbell 2006), Al9Fe2Si2 (Philips 1976) or generally
b-AlFeSi (Stefaniay, Griger, and Turmezey 1987; Tang and Sritharan 1998; Khalifa,
Samuel, and Gruzleski 2003). There is also conflicting conclusion regarding the
structure of b-phase. Monoclinic was accepted by many researchers (Mondolfo
1976; Murali et al. 1994; Crepeau 1995). Murali stated that b-Al5FeSi is monoclinic
with a ¼ 0.5792 nm, b ¼ 1.2273 nm, c ¼ 4.313 nm, and b ¼ 98.93 deg (Murali et al.
1994).
However, Carpenter claimed that b-phase was B-face centered orthorhombic
with a ¼ 0.6184 nm, b ¼ 0.6250 nm, and c ¼ 2.069 nm (Carpenter and LePage 1993);
Zheng observed that the b-phase was orthorhombic with a ¼ 0.618 nm, b ¼ 0.620 nm,
Figure 4 Morphology of d-Al4FeSi2 phase in an aluminum alloy (Khalifa, Samuel, and Gruzleski 2003).
Figure 5 Typical morphology of a-phase and b-phase in aluminum (Lu and Dahle 2005).
and c ¼ 2.08 nm (Zheng, Vincent, and Steeds 2000). Kral claimed the b-phase was
consistent with tetragonal Al3(Fe, Mn)Si2 with space group I4=mcm, a ¼ 0.607 nm
and c ¼ 0.950 nm (Kral, Mcintyre, and Smillie 2004). However, Kral (2005) claimed
that the structure of b-phase was consistent with that of the d-phase in his further
study.
Figure 6 Three-dimensional reconstruction of a-phase: (a) original two-dimensional photo; (b) three-
dimensional a-phase with high convoluted arms observed (Dinnis, Taylor, and Dahle 2005).
106 L. ZHANG ET AL.
Figure 7 Three-dimensional reconstruction of b-phase, (a) original two-dimensional phases; (b)

three-dimensional b-phase (Dinnis, Taylor, and Dahle 2005).
Among all the Fe-rich phases, b-AlFeSi is thought to be the most deleterious,
and most efforts have been devoted on how to avoid the formation of b-AlFeSi.
b-AlFeSi has an undesirable platelet morphology as shown in Figure 5 (Lu and
Dahle 2005) and Figure 7 (Dinnis, Taylor, and Dahle 2005), which is brittle and gen-
erally assumed to act as stress raisers and points of weak coherence. Usually, higher
iron content and slow cooling rate result in the increasing size of b platelets (Tang
and Sritharan 1998). The domination of platelet b-phases results in serious loss of
strength and ductility in Al-Si cast alloys. It is notable that a- and b-phase may
not clearly exhibit the dendrite or platelet shape respectively and are difficult to be
identified by their morphology, especially when the alloys are eutectic modified by
Na or Sr (Kral, Mcintyre, and Smillie 2004).
EFFECT OF IRON ON THE MECHANICAL PROPERTIES OF ALUMINUM

ALLOYS
In some alloys, iron is added intentionally. For example, iron is commonly
added to Al-Cu-Ni alloys to improve the high-temperature strength. For Al – low
Mg (<2 wt%) alloys iron is used to reduce germination. For Al-Fe-Ni alloys, iron
is added to reduce the corrosion in steam at elevated temperature. In the aluminum
conductor materials, iron is added to improve strength without substantial conduc-
tivity loss (Mondolfo 1976). Specifications normally permit considerably more Fe to
be present in permanent mould and pressure die-cast alloys compared to sand-cast
alloys because the cooling rate is higher in these cast conditions, thus, the size of
the constituents is finer. Particularly, in the commercial process of pressure die cast-
ing, the Fe content is above 0.8 wt%, and the precipitated eutectic Al-Si-Fe compo-
sition prevents the molten alloys from sticking (called ‘‘soldering’’ to the die that
causes a deterioration in the mould) and hence improves the surface quality of the
component (Suarez-Pena and Asensio-Lozano 2006). In these cases, Fe is an alloying
element.
The mechanical properties of cast aluminum alloy are generally imperiled by
the presence of iron (Couture 1981). Three-dimensional morphological results of
Fe-rich intermetallic compounds observed by co-authors of this paper in their paper
on inclusions in aluminum, Figure 8 (Damoah and Zhang 2010), suggests that
Fe-rich intermetallic phases have much more complex morphologies, with fragile
and brittle appearance than what is shown in two-dimensional observation. These
morphologies could explain why they are detrimental to the mechanical properties
of aluminum.
Al-Cu
The presence of Fe is generally reported to have a detrimental effect on the
ductility, strength, and fatigue properties of aluminum alloys. But a few investiga-
tions reported a slight improvement in strength and ductility with the increasing
of Fe content as shown Table 3. In Al-Cu alloys, needle-like Cu2FeAl7 phases pre-
cipitate, (Couture 1981) which are brittle and act as stress raisers like b-Fe phases,
reduce the strength and ductility of the alloy. The formation of Cu containing
Fe-phases also reduces the effective Cu content of the alloy. Since it is impossible
to dissolve this compound even with the subsequent solution heat treatment, the
amount of Cu available for hardening is considerably reduced.
Al-Mg
The influence of Fe on the Al-Mg alloy is shown in Table 4. Usually, increasing
of Fe content reduces the elongation (ductility) and tensile strength. However, it was
(Couture 1981) observed that the elongation and tensile strength improved with iron
Figure 8 Three-dimensional morphologies of Fe-rich intermetallic phases (a) (Chinese script) morpholo-
gies of b-Al (Fe, Mn)3Si and (b) p-Al8FeMg3Si6 intermetallics (Damoah and Zhang 2010). These are SEM
micrographs from 30% HCl solution-etched aluminum alloy (Mg-4.5%, Si-0.1%, Fe-0.1%) for 35 s.
Table 3. Effect of Fe on Al-Cu alloys
Tensile
UTS (Sampath YS (Sampath Elongation Hardness Fatigue

Al-Cu (wt%) Fe (wt%) and Dwarakadasa 1992) and Dwarakadasa 1992) (%) (HB) resistance Reference
Cu, 5 0.25 370–395 NA 12–13 NA NA Grand

0.6 260–285 3–4
Cu, 4.5 0.03 300 375 180 220 8 17 NA NA Passmore et al.
0.24 230 365 180 220 3 13
Cu, 3.5–4.2; 0.5–2.6 NA NA NA NA – Roth
108
Mg, 0.62–1.61;
Mn, 0.65–1.1;
Si, 0.45–0.55
Cu, 4.5; Mn, 0.55 0.73 250 NA 5 NA NA Bonsack
1.06 275 NA 5.5
Cu, 3 0.76–2 þ þ NA NA Hyman
2–2.91 – –
NA NA NA – þ NA Guillet et al.
Cu 4; Mg, 0.6; 0.25–1 – – – NA Cook, M et al. (Cook and
Ni 0.01–1, Wrought Chadwick 1944)

Data from Couture (Couture 1981). þ: increase with increasing Fe content; –: decrease with increasing Fe content; NA: unavailable data.
Table 4. Effect of Fe on Al-Mg alloys
Tensile (Sampath Elongation Corrosion Hardness Fatigue

Al-Mg (wt%) Fe (wt%) and Dwarakadasa 1992) (%) Creep resistance (HB) resistance Ref.
Mg, 5 or 10 NA – – NA – NA NA Kato et al.

Mg, <2 NA Slight NA NA NA NA NA Mondolfo (Mondolfo 1976)
>2 NA – – –
Mg, 10; Si < 0.1 initial 0.75 Slight – NA NA NA NA Parker, Cox and Tuner
Mg, 10; Si, 0.5 initial 0.6 þ þ
109
>0.6 – –
Mg, 3 or 6 0.2–0.8 Slight þ No NA NA NA NA Grand
Mg, Zn initial-2 þ – NA NA þ NA Bonsack
Mg, 0.7; 0.03–0.35 þ – NA – NA NA Chadwick et al. (Chadwick,
Si, 1; Muir, and Grainger 1953)
Mn, 0.2–0.6;
Cr, 0.2–0.6.
Wrought

Data from Couture (Couture 1981). þ: increase with increasing Fe content; –: decrease with increasing Fe content; No: No influence; NA: unavailable data.
110 L. ZHANG ET AL.
content increased to 0.6 wt% in Al-10Mg-0.5Si alloy. In Al-3 wt% Mg or 6 wt% Mg

alloys, an increase in iron content from 0.2 to 0.8 wt% resulted in a slight improve-
ment in tensile strength (Couture 1981). An increase in the tensile strength with iron
content, result was also obtained by Bonsack (Couture 1981) in Al-Mg-Zn alloy and
Chadwick and others (Chadwick, Muir, and Grainger 1953) in Al-0.7 Mg-1Si
wrought alloy.
Al-Zn
The ductility of Al-Zn alloys decreases with increasing iron content in
aluminum alloys due to the formation of needle like FeAl3 phase (Couture 1981).
Iron was also added as anti-recrystallizer to Al-Zn-Mg-Cu alloys (Fridlyander,

Tkachenko, and Matveets 1981). Kholnova (1968) found 0.2–0.4 wt% addition of
iron ensured satisfactory ductility and fracture toughness. With Fe content
0.1 wt%, the alloy had poor casting properties and a coarse grained structure,
low impact toughness and corrosion resistance; when Fe content 0.4 wt%, the frac-
ture toughness is again impaired. Some data are shown in Table 5.
Al-Si
Since the Al-Si alloys are the most common and important, there are many stu-
dies on the effect of Fe. The literature has been thoroughly reviewed by Couture
(Couture 1981). A basic consistent conclusion was that Fe has a small effect on
the properties up to a certain level, and when the iron content exceeds that critical
level the properties drop rapidly. When the Fe level exceeds about 0.7 wt%
(Mondolfo 1976), the b-Al5FeSi usually forms at roughly the location of the eutectic
trough on the Al-Fe-Si phase diagram (Figure 9 (Mondolfo 1976; Khalifa, Samuel,
and Gruzleski 2003)). The apparent critical Fe level is related to the formation of pri-
mary b-Al5FeSi, which is brittle and detrimental. Cooling rate and certain alloying
elements can increase the critical Fe content.
Although the negative effects of iron are clear, the mechanisms involved are
not fully understood. Mascre (1955) originally proposed the ‘‘restricted feeding’’
Table 5. Effect of Fe on Al-Zn alloys
Elongation Tensile (Sampath and Hardness

Al-Zn (wt%) Fe (wt%) (%) Dwarakadasa 1992) (HB) Ref.
Zn, Mg Chill cast 0.36 14 þ þ Bonsack

1 5
1.5 6
2.02 3 Slight Slight
(Zn þ Mg), 1.5 NA þ NA Krupkowski
6.5–10 Chill cast
Zn, 5; Mg Initial-1 No No NA Grand
>1 NA –

Data from Couture (Couture 1981). þ: increase with increasing Fe content; –: decrease with increasing
Fe content; No: No influence; NA: unavailable data.
Figure 9 Al corner of ternary Al-Fe-Si system diagram.
theory, suggesting that b-Al5FeSi increases porosity by blocking feeding channels

since the platelet b-Al5FeSi forms in the interdendritic channels during solidification.
Roy, Samuel, and Samuel (1996) claimed that the b platelets were active sites for
Figure 10 Growth of b-Fe phase (the white rod) on the entrained oxide film (the black line in the white
rod) (Cao and Campbell 2004b).
112 L. ZHANG ET AL.
pore nucleation and growth, also influenced the ultimate shape of pores—called
‘‘pore nucleation theory’’. While Taylor, Schaffer, and St. John (1999a, 1999b, 1999c)
reported their results conflicted with the former two theories for Al-5Si-1Cu-0.5Mg
alloys, and they substantially proved that the porosity was due to the effect that
the platelets have on the nucleation and growth of eutectic silicon. Cao and Campbell
(2004b, 2006) thought that the porosity was due to the presence of double-over oxide
bi-films, which were common in the aluminum melt. Its wetted side is the favored
substrates for the nucleation and growth of b-Fe platelets. After solidification, the
interface between the two sides of an entrained oxide film becomes the crack, as
shown in Figure 10 (Cao and Campbell 2004b). As an aside–Miresmaeili et al.
(2005) also found the nucleation and growth of Sr-rich compounds on the wetted
side of double-over oxide bi-films, which may give another explanation for the
increasing tendency of microporosity formation by Sr modification in Al-Si alloy.
MODIFICATION OF PLATELET IRON PHASES WITH ADDITION OF

OTHER METALS
Mbuya et al. discussed that the addition of a suitable neutralizer like Mn, Cr,
Be, Co, Mo, Ni, V, W, Cu, Sr, or the rare earth elements Y, Nd, La, and Ce could
sever to control the deleterious effect of Fe in aluminum alloys (Mbuya, Odera, and
Ng’ang’a 2003). Platelet Fe-rich phase is mainly the brittle b-Al5FeSi phase in Al-Si
cast alloy. In Al-Cu alloys, Cu2FeAl7 also has a kind of long needles shaped mor-
phology, which tends to embrittle the alloy (Mondolfo 1976), as mentioned earlier.
Modification=neutralization of the Fe-rich phases can be obtained by modifying the
platelet morphology and promoting the compact morphology such as Chinese script,
polyhedral or star like shape. Some earlier studies (Couture 1981) and recent studies,
which mainly focused on Al-Si-Mg cast alloys (Murali, Raman, and Murthy 1994,
1995a, 1995b, 1996; Murali et al. 1996, 1997a, 1997b; Yie et al. 1999; Wang and
Xiong 2000; Yin et al. 2000; Zhang et al. 2006d), showed the modification of
needle-like phases. However, the iron content usually maintains the same level after
the treatment. The total proportion of Fe-rich phases increases with Fe content (plus
Mn and Cr) and gradually imperils the properties of alloys even after the mod-
ification=neutralization treatment (Couture 1981).
Manganese (Mn)
Although Mn by itself is also harmful to the mechanical properties of alumi-
num alloys, it is widely used to neutralize Fe in Al-Si cast alloys. Manganese has
low cost and is readily available. Mn has an atomic radius and crystal structure close
to Fe as seen in Table 6 (Mahta et al. 2005). Literature on the neutralization of plate-
let Fe-rich phases by Mn addition is abundant. Couture has written a detailed review
on this topic (Couture 1981).
Mondolfo claimed that with the addition of sufficient manganese the platelet
Fe-rich phases transformed from the brittle platelet to the compact morphology such
as Chinese script, globule or polyhedral with a resulting restoration of the tensile
strength and elongation (Mondolfo 1976). b-Al5FeSi tends to be suppressed and
a-Al15 (Fe, Mn)3Si2 is formed after the Mn addition (Mondolfo 1976). In the ternary,
Table 6. Atomic number, Bravais lattice, and atomic radius of some transition metals in aluminum
Transition metal Atomic number Bravais lattice Atomic radius, Å
Ti 22 HCP 1.47
V 23 BCC 1.34
Cr 24 BCC 1.27
Mn 25 SC 1.26
Fe 26 BCC 1.26
Co 27 HCP 1.25
Ni 28 FCC 1.24
Al-Si-Mn alloy, a-Al15Mn3Si2 forms. In quaternary Al-Si-Fe-Mn alloys, iron may

replace up to 90 pct of manganese in this phase to form a-Al15 (Fe, Mn)3Si2 (Dinnis,
Taylor, and Dahle 2006). The morphology of a-Al15 (Fe, Mn)3Si2 changes with the
cooling rate and the concentration of Si and Mn (Granger, Sawtell, and Kersker
1984; Gustafsson, Thorvaldsson, and Dunlop 1986; Narayanan, Samuel, and
Gruzleski 1994; Crepeau 1995; Kim et al. 2006). The crystal structure of a-Al15
(Fe, Mn)3Si2 in wrought alloy is reported to be a-AlMnSi, simple cubic (sc) with a
Pm 3 space group (Cooper and Robinson 1966) or a-AlFeSi, body-centered cubic
(bcc) with a disordered Im 3 space group (Cooper 1967). Yoo reported that the crys-
tal structure of precipitated a-Al15 (FeMn)3Si2 depended on the Mn=Fe ratio (Yoo,
Shan, and Moon 1999). Kim found the rod-type a-Al15 (FeMn)3Si2 had a bcc
(a-AlFeSi)=sc (a-AlMnSi) dual structure or just Pm3 structure in skeleton-type as
shown in Figure 11 (Kim et al. 2006). It was reported that the increase of Mn in
a-Al15 (FeMn)3Si2 changed its crystal structure from disordered Im3 to ordered
Pm 3 (Donnadieu, Lapasset, and Sanders 1994; Cooper 1967). Kim claimed the criti-
cal Mn=Fe value for disorder–order transition of a-Al15 (FeMn)3Si2 lay within the
range of 0.11–0.66 (Kim et al. 2006). Murali observed that b-platelets transformed
into very coarse polyhedral with the presence of 1.3 wt% Fe and 0.53 wt% Mn in
the sand cast in Al-7Si-0.3Mg alloy, while more Chinese scripts were found at lower
Fe content 0.87 wt% and 0.35 wt% Mn at a high cooling rate (chill castings) (Murali,
Raman, and Murthy 1994). Ashtari studied the effect of Mn and Sr additions on
b-Al5FeSi modification in Al-6.5Si-3.5Cu-1.0Fe alloy. With the 0.015 wt% addition
of Sr in the Mn free alloy, only the refinement of eutectic Si and b-Al5FeSi was
observed. With the combined addition of 0.015 wt% Sr and 0.3 wt% Mn, both
a-AlFeSi and b-Al5FeSi was observed. No evidence was stated that Sr improved
the transformation of b-Al5FeSi (Ashtari, Tezuka, and Sato 2003).
According to Shabestari (2004), high Fe content in aluminum encourages the
formation of star like a-AlFeSi in Al-12.7Si alloys as shown in Figure 12. A linear
relationship between segregation factor [‘‘SF’’(Gobrecht 1975; Czikel et al. 1985;
Awano and Shimizu 1990)] and the amount of intermetallic compounds at short
holding time was obtained. Figure 13 (Shabestari 2004) shows the relationship and
the function can be described as Volume (%) ¼ 1.33SF 1.53 (Shabestari 2004),
and the cooling rate was stated as 10 C=s. In Al-7.0 Si-0.38 Mg (A356) alloy, with
the addition of Mn from 0.07 to 0.20 wt%, the platelet b-Fe were replaced by globu-
lar or Chinese scripts a-AlFeSi as shown in Figure 11 (Kim et al. 2006). Zhang
observed that the platelet b-Al5FeSi was converted to a-AlFeSi with the condition
114 L. ZHANG ET AL.
Figure 11 Morphology of a-Al15 (Fe, Mn)3Si2 particles (a) Al-7.0 Si-0.38 Mg–0.20 Fe, (b) Al-7.0 Si-0.38
Mg-0.20 Fe-0.07 Mn, (c) Al-7.0 Si-0.38 Mg-0.20 Fe-0.13 Mn, and (b) Al-7.0 Si-0.38 Mg-0.20 Fe-0.20 Mn
(Kim et al. 2006).
Figure 12 Effect of Mn and Fe on the formation of intermetallic compounds in Al-12.7 Si-0.1 Cr alloys
(Shabestari 2004).
Figure 13 Volume fraction of intermetallics versus the segregation factor (SF) (Shabestari 2004).
of Mn=Fe ¼ 1.28, Fe ¼ 1.09 wt% in Al-12 Si-0.3 Mg alloy, the cooling rate was not
stated in the report (Zhang 2007).
The amount of manganese needed to neutralize iron is still not well established
(Couture 1981; Crepeau 1995) because the neutralization process is deeply affected
by many factors, such as cooling rate, heat treatment, initial Si and Fe content,
and other elements. Generally, it is desirable to have a weight percentage Mn=Fe
ratio of 1:2 if the Fe content exceeds 0.45 wt% (Colwell and Kissling 1961). Mascre
arrived at a different neutralization formula for sand and permanent mold castings:
Mn% ¼ 2 (Fe% 0.5) (Mascre 1955). However, Cao found the Chinese script
a-AlFeSi even with the Mn=Fe ¼ 0.17 in an Al-11.6 Si-0.37 Mg-1.16 Fe alloy as
shown in Figure 14 (Cao and Campbell 2004b). Mn modified Fe-rich compounds
have a greater tendency to segregate and to form coarse primary a-Fe, which imper-
ils the tensile strength and ductility of Al alloys (Couture 1981; Crepeau 1995).
In Al-Cu alloys, Mn addition encourages the formation of Chinese script
phases and suppresses the platelet Cu2FeAl7 (Mondolfo 1976). Tseng reported that
Mn addition recovered the ductility but caused a further loss in the yield strength of
Al-Cu alloys. The primary constituents were needle shaped Cu2FeAl7 at low content
of Mn (0.29 wt%). Upon the further addition of Mn (0.66 wt%), almost all the
Cu2FeAl7 transformed into the Chinese script Mn-bearing particles either in low
Fe alloy or in high Fe alloy as shown in Figure 15 (Tseng et al. 2002). However,
the structure or stoichiometry of the Mn-bearing particles was not mentioned. In
Al-Mg alloys, Mn transformed Al3Fe into (Fe, Mn)Al6 and improved the corrosion
resistance of the alloy (Mondolfo 1976).
Cobalt
Fe, Cr, and Co have similar atomic radii and Cr also has the same Bravais lat-
tice with Fe as shown in Table 6. Thus, Cr and Co can also be used to precipitate
116 L. ZHANG ET AL.
Figure 14 SEM image of Al-11.6 Si-0.37 Mg-1.16 Fe-0.29 Mn showing typical Chinese script of a-Fe (Cao
and Campbell 2004b).
Figure 15 CuAl2, Cu2FeAl7, and Mn-bearing compounds in aluminum alloys, (a) 4.64Cu-0.05Fe-0.29 Mn;
(b) 4.67Cu-0.05Fe-0.66 Mn; (c) 4.58Cu-0.31Fe-0.29 Mn; (d) 4.52Cu-0.30Fe-0.67 Mn; (arrow1: CuAl2;
arrow2: Cu2FeAl7; arrow3: Mn-bearing) (Tseng et al. 2002).
Fe-rich phases. There are few investigations on the effect of Co and Cr on the modi-
fication of Fe-rich phases. Co was thought less effective than Mn and a higher
amount of addition was needed. Co=Fe 0.5–1.0 was suggested as an appropriate
addition to modify the platelet Fe phases to globular shape (Couture 1981). Mahta
et al. (2005) reported that the optimum ratio of Co=Fe was 1.0 for A413 Al-11Si
alloys and the Co-Fe phases were identified approximately as Al15 (Fe, Co)4Si2, while
Murali, Raman, and Murthy (1994) stated the Co-Fe phase was Al14Co2 (Fe,Si). It
was reported that almost all the transformed Chinese script a-AlFeSi phases precipi-
tated within a-Al dendrites (Figure 16) (Couture 1981; Murali, Raman, and Murthy
1994; Crepeau 1995; Mahta et al. 2005), which reduced the segregate tendency. Thus,
Co is much less harmful than Mn. However, the presence of Co (<1 wt%) at low con-
centration was reported to encourage the formation of primary Si as well as the a-

AlFeSi phases in Al-11Si alloys as shown in Figure 16(b) (Mahta et al. 2005). This
work has not been referred by earlier references (Couture 1981; Murali, Raman, and
Murthy 1994; Crepeau 1995).
Figure 16 Microstructure of Co transformed platelet Fe phases, (a) 11Si, 2.4Fe, 2.4Co; (b) 0.6Fe, 0.3Co
(Mahta et al. 2005).
118 L. ZHANG ET AL.
Chromium
Mondolfo (1976) claimed that Cr was a possible additive to neutralize Fe
phases in aluminum. Cr was believed more effective than Co, and a ratio of Cr=
Fe ¼ 0.33 can prevent the formation of b-AlFeSi (Mahta et al. 2005). The function
of Cr is similar to that of Co. Mahta reported that the identified transformed Chi-
nese script a-Fe phase was close to Al15(Fe, Cr)3Si2, agreeing with Crepeau’s report
(Crepeau 1995) while different from Murali’s report [Al25(Fe, Cr)3Si4] (Murali,
Raman, and Murthy 1994). As with Co, a low amount of Cr also resulted in the for-
mation of primary Si in Al-11Si alloys, as show in Figure 17 (Mahta et al. 2005). Dif-
ferent from Co, the transformed a-phases by Cr were observed both in
a-Al dendrites and the interdendritic region (Murali, Raman, and Murthy 1994;
Mahta et al. 2005), which was the same as during Mn modification [Figure 17(a)
(Mahta et al. 2005)].
Beryllium
Murali considered beryllium was a more effective neutralization element than
Mn, Co, and Cr (Murali, Raman, and Murthy 1995b). Crepeau reported that the
Figure 17 Microstructure of Cr transformed platelet Fe phases, (a) 11Si, 0.56Fe, 0.23Cr; (b) 11Si, 2.68Fe,
0.83Cr (Mahta et al. 2005).
addition of Be > 0.4 wt% was required (Crepeau 1995), while other references
showed that the trace addition of Be 0.06–0.27 wt% was enough (Murali, Raman,
and Murthy 1995b; Murali et al. 1997b; Yie et al. 1999; Wang and Xiong 2000).
Beryllium along with Fe, Al, and Si can form Chinese script or polygon Be-Fe phases
identified as Al8Fe2BeSi while other literatures described as Al4Fe2Be5 (Crepeau
1995). It is notable that Be-Fe phases precipitated inside of the primary a-Al den-
drites (like Co) (Murali, Raman, and Murthy 1994, 1995a, 1995b; Murali et al.
1996, 1997b; Wang and Xiong 2000) as shown in Figure 18 (Wang and Xiong
2000), which favors fracture toughness, whereas the b-AlFeSi phase is in the inter-
dendritic region. Wang and Xiong (2000) observed that the addition of beryllium
to Al-Si-Mg alloys hindered the crack propagation on Be-Fe phases. Particularly,
small amount of Be addition results in the formation of stable Chinese script

a-Al8 (Fe, Mn)2BeSi in 5XXX series alloys. The modification decreases the platelet
iron phases formation during annealing and improves the mechanical properties as
shown in Figure 19 (Samuel et al. 2003). However, Yie et al. (1999) did not report the
Be-Fe phases with an addition of 0.11 wt% Be in Al-11Si alloys, but the platelet
phases were transformed to a-AlFeSi.
Figure 18 Be-Fe phases in Al-7Si-0.3 Mg alloys, (a) Be 0.23 wt% Fe 0.21 wt%; (b) Be 0.26 wt% Fe 0.82 wt%
(Wang and Xiong 2000).
120 L. ZHANG ET AL.
Figure 19 Microstructure of 5xxx alloys (0.42Si-0.46Fe-0.044Cu-0.291 Mn-3.47 Mg-0.003Cr) containing

0.009 wt% Be (Samuel et al. 2003).
The addition of beryllium changes the solidification sequence of Fe-rich

phases. The ternary eutectic reaction to generate b-AlFeSi is substituted by a peritec-
tic reaction that forms Be-Fe phases, which replaces the b-AlFeSi phases in Al-Si-Mg
cast alloys (Murali, Raman, and Murthy 1995b; Wang and Xiong 2000). In the stir-
red cast and extruded Al-7Si-0.3 Mg alloys (Murali et al. 1997b), the beryllium
addition lowered the detrimental effect of Fe-rich phases on the ductility and impact
energy in sand cast alloys by a small amount as shown in Table 7, and it was
explained that the stirring eliminated the peritectic reaction, and Be-Fe phases was
found in the interdendritic region as well as in the primary a-Al.
Table 8 shows the effect of Be addition on mechanical properties of the sand
cast Al-Si-Mg alloys. The deleterious effect of Fe has been neutralized such that
the strength and elongation remained more or less unaffected as Fe content increased
from 0.1 to 0.94 wt%. Figure 20 (Wang and Xiong 2000) shows the mechanical pro-
perty changes in Al-7Si-0.3Mg-0.2Ti alloys with the Fe content (Wang and Xiong
2000), indicating that the UTS (Ultimate Tensile Strength) of beryllium-added alloys
decreases little with the increasing iron content. Moreover, the addition of beryllium
led to grain refinement in the Al-Si-Mg alloys (Depierre and Bernstein 1951). Thus,
the refinement effect of beryllium also contributed to the mechanical properties
improvement (Murali, Raman, and Murthy 1995b; Wang and Xiong 2000).
Table 7. Mechanical properties of stirred cast and extruded

Al-7Si-0.3 Mg alloys (Murali et al. 1997b)
Alloy (wt%) Elongation, % ET, J
0.4Fe 15.0 0.5 14.5 1

0.96Fe 11.3 0.5 9.2 1
0.96 Fe þ 0.26 Be 11.3 0.5 8.0 1
Table 8. Mechanical properties of sand-cast Al-Si-Mg alloys with trace Be addition

(Murali, Raman, and Murthy 1995b)
Without Be addition With Be addition
0.2 wt% Fe 1.0 wt% Fe 0.1 wt% Fe 0.94 wt% Fe
Ultimate tensile strength (Fukuda 270 3 210 4 260 3 260 3

et al. 1998) (Sampath and
Dwarakadasa 1992)
Yield strength (YS) (Sampath and 220 3 185 3 205 3 205 3
Dwarakadasa 1992)
Elongation (E) (%) 4 0.5 2 0.5 4 0.5 3.5 0.5
Strontium
Strontium has been used to modify the eutectic silicon phases from acicular to
lamellar in Al-Si cast alloys (Closset and Gruzleski 1982a, 1982b; Hanna, Lu, and
Hellawell 1984; Shabestari and Shahri 2004). In aluminum wrought alloys, Sr was
also applied to transform the platelet Fe-rich phases to a-AlFeSi (Al8Fe2Si) (Morris
and Miners 1975; Mulazimoglu, Gruzleski, and Closset 1992; Mulazimoglu et al.
1994; Closset et al. 1996). In 6063 alloys (0.2–0.6 Si, 0.45–0.9 Mg), the 0.01–
0.5 wt% addition of Sr promoted the formation of a-AlFeSi and improved the
extrusion characteristics significantly (Morris and Miners 1975). It was reported that
the 0.01–0.10 wt% addition of Sr to Al-Cu-Mg-Zn wrought alloys refined the inter-
metallic phases (William and Bernard 1987). Tensile strength and conducting
property were improved by approximately 0.1 wt% Sr addition into 6201 alloys
(0.45–0.65 Si, 0.50–0.65 Mg) (Closset et al. 1996). Figure 21 shows the comparison
of UTS and YS with or without Sr addition in 6201 alloys (Closset et al. 1996).
The Sr-treated alloys exhibit 5% higher UTS and YS than that of the Sr-free alloys
at high magnesium content (0.6 wt%). Mulazimoglu observed that 0.015 wt% Sr
addition was enough to modify the platelet phases in 6063 and 6201 alloys
Figure 20 Mechanical property of aluminum alloys changes with the content of Fe and Be (Wang and
Xiong 2000).
122 L. ZHANG ET AL.
Figure 21 Mechanical properties of synthetic 6201 wires (aging at 160 C for 5 h), (a) Tensile strength; (b)
Yield strength (Closset et al. 1996).
(Mulazimoglu, Gruzleski, and Closset 1992). The modification of b-AlFeSi was

achieved at 0.013 wt% Sr addition in synthetic 6061 alloys (0.7 Si, 0.5 Fe, 0.88
Mg, 0.11 Mn) with the cooling rate 2 C=s, while in commercial 6061 alloys (0.7Si,
0.3Fe, 0.9 Mg, 0.02 Mn) the Sr addition must be increased to 0.031 wt%
(Mulazimoglu et al. 1994). A slower cooling rate favored the modification of
b-AlFeSi with Sr addition in 6061 alloys (Mulazimoglu et al. 1994). In 1XXX series
alloys (0.3–1.0Si, 0.3–1.0Fe), where Fe and Si are the major impurities, Sr addition
effectively promoted the formation of a-AlFeSi, as shown in Figure 22
(Mulazimoglu et al. 1994). A higher cooling rate (>2 C=s), 0.06 wt% addition of
Sr, and Si=Fe ratio <1 resulted in almost 100% of a-AlFeSi formation
(Mulazimoglu et al. 1994). The mechanism of the modification may be associated
with the adsorption of Sr as a surface active element. It is presumed that Sr can
be absorbed on the surface of primary a-Al8Fe2Si phase to act as a barrier for the
dissolution. Thus, the peritectic reaction L þ Al8Fe2Si ! Al þ Al5FeSi will be
hindered.
Sr was also reported to modify b-AlFeSi phases in Al-Si cast alloys (Couture
1981; Shabestari and Gruzleski 1996). Lu and Tang claimed that the addition of
Sr suppressed the formation of b platelets (Tang and Sritharan 1998; Lu and Dahle
2005). Shabestari reported that (0.04–0.06 wt%) Sr addition not only transformed
b-AlFeSi to Chinese script a-phase but also reduced the size of platelet b-AlFeSi
by fragmentation in Al-12Si (413) alloys (Shabestari and Gruzleski 1996). Yie and
Pucella proved the fragmentation of platelet b-AlFeSi and marginal reduction of
platelet by Sr addition in Al-11Si alloys (Pucella et al. 1998; Yie et al. 1999). Ashtari
indicated that the size and amount of platelet b-AlFeSi were reduced after Sr
addition but no transformation of platelet b-AlFeSi in Al-Si-Cu-Fe alloy (Ashtari,
Tezuka, and Sato 2003). Samuel observed that only primary b-AlFeSi dissolved in
the a-Al matrix by peritectic reactions at 0.03–0.06 wt% Sr addition in Al-6 Si-3.5
Cu (A319) alloys (Samuel et al. 1998). However, many studies showed that there
was no transformation or reduction of platelet b-AlFeSi phase after Sr addition
(Samuel et al. 1998; Yie et al. 1999; Ashtari, Tezuka, and Sato 2003; Suarez-Pena
and Asensio-Lozano 2006). Suarez-pena’ and Lu also showed that no obvious
Figure 22 Effect of Sr on 1XXX series alloys. (a) No Sr; (b) 0.06 wt% Sr (Mulazimoglu et al. 1994).
transformation occurred except that the b-AlFeSi phase became thinner after the
0.04 wt% Sr addition in Al-12Si alloy (Lu and Dahle 2005; Suarez-Pena and
Asensio-Lozano 2006). Cho reported that the combined addition of phosphorus
and strontium could suppress the formation of b-AlFeSi phases by the pre-existing
of Al2Si2Sr phases, which was caused by deactivation of AlP nuclei, in Al-10Si alloys
(Cho et al. 2008). The fragmentation mechanism of b-AlFeSi is related to nucleation
and growth: The addition of Sr causes a high supersaturation of the element Fe and
Si, thus, the primary b-AlFeSi phase precipitates before the primary a-Al in the
liquid. Then, the primary a-Al dendrites nucleate onto the primary b-AlFeSi phase
and reject Fe into the residual liquid. The secondary needle-like b-AlFeSi phase
nucleates due to a generalized undercooling induced by Sr (Closset et al. 1996;
Cho et al. 2008). The growth rate of the secondary phase is much lower than that
of the primary phase, thus, the smaller b-AlFeSi phases precipitate (Suarez-Pena
and Asensio-Lozano 2006). It was also reported that the added Sr was absorbed
on the nucleation sites of the b-AlFeSi phase (called poison), thus, the amount of
the b-AlFeSi phase was reduced (Shabestari and Gruzleski 1996; Pennors et al.
1998; Cho et al. 2008).
124 L. ZHANG ET AL.
Molybdenum (Mo), Nickel (Ni), and Sulfur (S)

Molybdenum, Nickel, and Sulfur are much less used to modify iron phases in
aluminum alloys. However, Mo was claimed to convert platelet iron phases to com-
pact, star-shaped compounds (Dichtl 1970), and to be much more effective than Mn
(Couture 1981). A combined addition of Mo (0.05–0.3 wt%) and S (0.01–0.2 wt%)
was reported to improve the mechanical properties of Al-6 to 12 Si alloys (Couture
1981). Ni was mentioned as a ‘‘neutralizer’’ in some references (Mondolfo 1976;
Couture 1981; Crepeau 1995).
MODIFICATION OF PLATELET IRON PHASES WITH THERMAL

INTERACTIONS
Molten Aluminum Superheating
Mondolfo reported that the size of Fe intermetallic compounds decreased as
the superheating temperature increased up to 500 C above the melting point
(Mondolfo 1976). Crepeau claimed that Chinese script a-phases formed instead of
platelet b phases when the melt temperature was >800 C, the subsequent thermal
cycling of the melt slightly affected the Chinese script morphology (Crepeau
1995). The conclusion was confirmed by other references (Awano and Shimizu
1990). Narayanan claimed that a-phases formed at the temperature around
200–300 C above the liquidus temperature for Al-6Si-3Cu alloys due to the domi-
nation of a alumina inclusions at the superheating temperature in the melt, which
was a poor nucleus for b phase. While c alumina inclusions, acting as a nucleus
for crystallization of b phase and forming at low melt temperature 750 C, trans-
formed to a alumina inclusions at the superheating temperature (Narayanan,
Samuel, and Gruzleski 1994). However, the high superheating temperature in alumi-
num alloys is not recommended owing to the increasing of gas pick-up, oxidation
loss and the formation of oxide inclusions as well as higher energy costs and furnace
wearing.
Cooling Rate
According to the Al-Fe-Si phase diagram [Figure 9 (Mondolfo 1976; Khalifa,
Samuel, and Gruzleski 2003)], the b-Fe rich phase is a stable phase under the equi-
librium cooling rate. When casting is conducted under a very high cooling rate and=
or the melt is superheated to a high temperature, however, the iron rich phase crys-
tallizes into the a-phase in metastable forms (Iglessis, Frantz, and Gantois 1977;
Narayanan, Samuel, and Gruzleski 1994). The cooling rate increases as the critical
content level of Fe to form b-Fe increases. For example, 0.75 wt% Fe-1 C=s,
0.9 wt% Fe 5 C=s, and 1.0 wt% Fe 10 C=s (Mascre 1955). Passmore et al.
reported that the Cu2FeAl7 phase was fine and evenly distributed in the chilled
Al-Cu alloy while in the slowly cooled metal the Cu2FeAl7 phase was coarse and
concentrated at the interdendritic region, which was detrimental to the mechanical
properties (Passmore, Flemings, and Taylor 1958). Narayanan reported that with
increasing cooling rate and melt superheating temperature, the precipitation tem-
perature of the b-Fe phase decreased until the b-Fe rich phase merged with the
silicon eutectic temperature. The chemical composition of the remaining liquid

favored the crystallization of the a-Fe rich phase in Al-6Si-3Cu alloys (Narayanan,
Samuel, and Gruzleski 1994). The refinement and fragmentation of the b-Fe phase
also can be obtained by a high cooling rate. Sigworth reported that the rapid solidi-
fication not only reduced the dendrite arm spacing but also refined the b-AlFeSi
phase (Sigworth, Shivkumar, and Apelian 1989).
Heat Treatment
Fe-rich phases can be dissolved, fragmented and globularized during the
solution heat treatment at the solidus temperature (Crepeau 1995). High treatment
temperature is required because of the low diffusivity of Fe in the solid aluminum.

Barresi et al. (1993) reported that the p-Al8FeMg3Si6 phase was transformed to the
b-Al5FeSi phase by the solution heat treatment. After the treatment, the b-Al5FeSi
phase became the predominant phase. The reduction of the p-Al8FeMg3Si6 phase
implied that Mg went into solid solution, which was favorable for the age hardening
to improve the strength of the cast alloy (Barresi et al. 1993). But this method is
limited by the possible damage caused by incipient melting (Crepeau 1995).
IRON REMOVAL TECHNOLOGIES

Precipitation of High Fe-Rich Phases
The method is mainly applied to purify Al-Si cast alloys. Iron is removed by
the formation of primary Fe-rich intermetallics, generally, primary a-Al15 (Fe,
Mn)3Si2 or a-Al15 (Fe, Mn, Cr)3Si2 (called ‘‘sludge’’). The crystallization of the
sludge under an appropriate holding temperature and an initial Fe and Mn content
has proved to be an effective method to remove Fe from Al-Si alloys, as shown in
Figure 23 (Flores et al. 1998). An empirical equation is presented to simulate the
kinetics of the nucleation and growth rate of the sludge. This method is a
Figure 23 Effect of holding time and initial Fe, Mn content on the formation of sludge; alloy I: 9.5 Si-0.6
Mn-0.8 Fe; alloy II: 9.5 Si-0.8 Mn-1.3 Fe; alloy III: 9.5 Si-1.5 Mn-1.5 Fe; alloy IV: 9.5 Si-2.2 Mn-1.6 Fe.
(a) mass of sludge segregate isothermally at 605 C; (b) mass of sludge segregated from alloy IV at different
holding time (Flores et al. 1998).
126 L. ZHANG ET AL.
well-established technique for Al-Si cast alloys. The process consists of two steps:
formation of primary Fe-rich intermetallic particles, followed by the removal of
the particles. The alloy is melted and held at a high temperature (750 800 C) for
some time in order to be well homogenized. Then, the melt is cooled to a holding
temperature in the range of 600–650 C for the formation and growth of the sludge
(Flores et al. 1998; Pucella et al. 1998; de Moraes et al. 2006). After the formation of
the sludge, further steps including gravity separation, filtration, electromagnetic
(EM) separation or centrifuge separation are conducted to remove the sludge phase
from the molten aluminum (Shabestari and Gruzleski 1995; Flores et al. 1998; Li
et al. 2003; Cao, Saunders, and Campbell 2004; de Moraes et al. 2006). The iron con-
centration generally decreases from 1 to 2 wt% to at most 0.4 wt% after the treatment
(Cao and Campbell 2004b; de Moraes et al. 2006).

Mn or the combination of Mn and Cr are used to form sludge (Shabestari and
Gruzleski 1995; Van Der Donk, Nijhof, and Castelijns 1995; Nijhof et al. 1996; de
Moraes et al. 2006). The formation of sludge can be predicted by an iron equivalent
value (IEV, referred to elsewhere as the ‘‘sludge factor’’ [Czikel et al. 1985)] in the
initial Al-Si alloy (Gobrecht 1975). The IEV represents an attempt to integrate the
combined effects of some important impurities: Fe, Mn, and Cr. The generally
accepted formula was reported as (Dunn 1965; Gobrecht 1975; Jorstad 1986; Awano
and Shimizu 1990):
IEV ¼ ðwt%FeÞ þ 2ðwt%MnÞ þ 3ðwt%CrÞ ð1Þ
A high IEV implies the easy formation of sludge phase. Czikel et al. and
Jorstad predicted the sludge formation by temperature and IEV as shown in
Figure 24 (Czikel et al. 1985; Jorstad 1986). Jorstad discussed that the Fe:Mn:Cr
ratio of 1:2:3 gives a realistic view of the sludge factor for determining the tolerable
limits for a given melt. For instance, a sludge factor of 1.8 is indicated as suitable for
a furnace operating temperature of 660–663 C (1220–1225o F) whereas a sludge
factor of 2.2 is considered too high for a furnace temperature of 660–663 C
Figure 24 Sludge factor (SF) diagram showing the temperature and composition at which the sludge can
form (Czikel et al. 1985).
(1220–1225o F) (Jorstad 1986). In Al-12.7Si alloys with a constant 0.30 wt% Mn and
0.1Cr, the temperature variation of for the formation of sludge versus iron content is
shown in Figure 25 (Shabestari 2004). However, some researchers treated the ratio of
Mn=Fe as the most critical process parameter, and believed that it was necessary to
increase Mn addition to obtain a low Fe-level (Nijhof et al. 1996). Flores (Flores et al.
1992a) claimed that with a holding temperature of 620–640 C the addition of Mn
(2.2 wt%) decreased the Fe content from 1.5 to <0.3 wt% in Al-9.2 Si alloys, while
with the addition of 0.6 wt% Mn decreased the Fe content from 1.2 to about
0.7 wt%. The presence of 0.2 wt% Cr accelerated the gravity segregation process
without improving Fe reduction ratio. Further studies on the nucleation and growth
of sludge in Al-9.5Si alloys were also conducted by Flores (Flores et al. 1998).
The formation of primary AlFeMnSi phases can be predicted more correctly

by Al-Fe-Mn-Si phase diagrams [Figure 26 (Balitchev et al. 2003) and Figure 27
(Balitchev et al. 2003)]. Balitchev et al. (2003) obtained good agreement between
the calculated phase diagram (using ThermoCalc) and experimental data as shown
in Figure 26 and Figure 27. As in Figure 26(a), the formation of primary a-Al
(Fe, Mn)Si can occur at high temperature with high Mn content. With a low
addition of Mn (0.3 wt%) and 600 C, the primary b-Al (Fe, Mn)Si phase can
precipitate from the liquid. When the iron content increases, the composition range
of the primary a-Al (Fe, Mn)Si phase changes, and the high temperature favors this
process (Figure 26b). According to Figure 27, the content of Si also changes the for-
mation range of sludge.
There are some new theories about the formation and segregation of Fe-rich
phases. It was reported (Cao and Campbell 2000) that Fe content decreased from
0.57 to 0.47 wt% after holding for 10 min at 600 C with the presence of 0.59 wt%
Mn in Al-11Si-0.4Mg eutectic alloy, and the nucleation and growth of Fe phases
on the wetted side of the oxide films led to the reduction in the amount of both oxide
Figure 25 Temperature for sludge formation in Al-12.7Si-0.3Mn-0.1Cr-xFe alloys [the curve can be
expressed by T ( C) ¼ 645.7 þ 34.2 (wt% Fe)2] (Shabestari 2004).
128 L. ZHANG ET AL.
Figure 26 Calculated isopleths at 10.0 wt% Si, (a) 1.0 wt% Fe; (b) 3.0 wt% Fe (D: experimental data)
(Balitchev et al. 2003).
films and Fe-rich phases remaining, which resulted in a moderate improvement in

tensile properties of the alloy.
In addition to the Mn and Cr, Sr also was used to form sludge combined with
Mn (Pucella et al. 1998; Cao and Campbell 2004a). Pucella reported that the
addition of Sr at 620 C pouring temperature precipitated the high Fe sludge of
Figure 27 Calculated isothermal sections at 660 C, (a) 10.0 wt% Si; (b) 14 wt% Si (D: experimental data)
(Balitchev et al. 2003).
a-Al12 (Fe, Mn, Cr)7Si2 as well as transformed polyhedral sludge particles (Pucella
et al. 1998). Cao and Campbell reported that the mass of sludge particles with Sr
addition was significantly increased at a temperature of 900 C compared with Sr-free
samples, and under the same condition, the number of sludge particles was dramati-
cally increased, which resulted in an effective Fe reduction (Cao and Campbell
2004a).
Gravity Separation
Gravity separation is a method employed to remove the sludge. The melt is
held at the sludge formation temperature for a relative long time. Longer settling
time and lower holding temperature encourages higher iron removal efficiency.
Van Der Donk, Nijhof, and Castelijns (1995) observed that sludge formed and
segregated immediately during the cooling from a high temperature (840 C) to a
holding temperature of 600 C. Reported removal efficiency data were summarized
in Table 9. Flores and Cao reported an Fe removal fraction over >70% with Mn=
Fe > 1 and relative high Fe content (Flores et al. 1992a; Flores et al. 1998; Cao
and Campbell 2004b). When the Mn=Fe ¼ 1, the removal fraction of Fe decreased
to 50% (Cao, Saunders, and Campbell 2004; Cao and Campbell 2004b).
Figure 28a (Li 2004) shows a microstructure of an aluminum alloy after gravity sep-
aration for a long time. Most of the sludge settled to the bottom and thus the upper
part of the alloy was purified. Figure 28b (Cao, Saunders, and Campbell 2004) illus-
trates the precipitated primary a-Fe particles. The white polygons and Chinese
scripts are the sediment sludge phases.
Figure 29 (Shabestari and Gruzleski 1995) shows the variation of sludge vol-
ume percentage, the average size and number of particles per area at different
location of the mold of Al-12Si-1.2Fe-0.29Mn-0.1Cr alloys at different holding
times. Obviously, the volume percentage, the number, and average size of sludge
increase with distance from the melt surface (Flores et al. 1992a; Shabestari and
Gruzleski 1995; Flores et al. 1998). The removal fraction by settling varies with
the position in the melt due to the depletion of the sludge near the surface and the
increase near the bottom. The settling speed of sludge particles can be roughly cal-
culated by Stokes’ law. As shown in Figure 30 (Shabestari and Gruzleski 1995), the
final terminal velocity can be reached at very short time, but the velocity is low.
Filtration
Primary, Fe-rich intermetallic particles can also be removed by porous filters
similar to the removal of nonmetallic inclusions by filtration. Figure 31 (de Moraes
et al. 2006) shows the schematic steps of the filtration operation. After a short time
(10–20 min) holding at sludge formation temperature, the melt is decanted through a
preheated filter. By the way, small amount of sludge also precipitates at the bottom
of the melt during the holding time. The holding time used for filtration is much
shorter than holding time used for gravity separation and so filtration is suitable
for the continuous treatment.
Van Der Donk, Nijhof, and Castelijns (1995) concluded that finer-pore foam
filters can remove small size sludge particles but actually only slightly increased
Table 9. Reported removal efficiency of iron from aluminum alloys
Holding After Purification (wt%)
Composition (wt%, Temp Time Fe removal Filter Sludge

Authors and reference Al balance) Mn=Fe ( C) (min) Mn Fe fraction (%) (ppi) separation method
Kim (Kim and Yoon 2000) 6Si-1.64Fe-1.66 Mn 1 690 20 NA 0.45 72.6 NA EM separation
Li (Li et al. 2001; Li et al. 2003; Li 12Si-1.13Fe-1.22Mn 1.1 640 30 0.26 0.41 63.7
2004)
Jiao (Jiao, Zhang, and Yao 2006) 10Si-1Fe-1.1Mn 1.1 NA NA 0.39 0.30 70
Xu (Xu et al. 2003) 11.7Si-1.2Fe-1.8Mn 1.5 NA NA 0.26 0.41 65.8
Cao (Cao, Saunders, and Campbell 12Si-1.23Fe-1.11Mn 0.9 600 240 0.36 0.52 57.7 NA Gravitational
2004; Cao and Campbell 2004b) separation
Cao (Cao and Campbell 2004b) 12Si-1.11Fe-1.03Mn 0.9 600 240 0.29 0.57 48.6
Cao (Cao and Campbell 2004b) 12Si-1.22Fe-2.15Mn 1.8 600 240 0.30 0.35 71.3
Flores (Flores et al. 1998) 9.5Si-1.6Fe-2.2Mn 1.4 605 380 0.5 0.37 77
Flores (Flores et al. 1992a) 9.2Si-1.5Fe-2.2Mn 1.5 640 180 NA 0.4 73
130
Nijhof (Nijhof et al. 1996) 12Si-1.1Fe-1Mn 0.9 605 20 0.42 0.64 42 NA Filtration
8Si-1.22Fe-1.12Mn 0.9 630 20 0.67 0.82 32.8 NA
12Si-1.1Fe-1.8Mn 1.6 605 20 0.57 0.47 57.2 NA
Van Der Donk (Van Der Donk 8Si-1.22Fe-1.12Mn 0.9 640 30 0.63 0.79 35.2 30
1998) 11.5Si-1.12Fe-0.99Mn 0.9 605 15 0.42 0.64 42.9 10
11.4Si-1.09Fe-1.86Mn 1.7 605 15 0.58 0.47 56.9 10
11.2Si-1.58Fe-1.9Mn 1.2 605 15 0.50 0.61 61.4 10
12Si-2.07Fe-1.91Mn 0.9 605 15 0.53 0.85 58.9 10
De Moraes (de Moraes et al. 2006) 12Si-1.32Fe-1Mn 0.75 605 30 NA 0.4 69.7 20
12Si-1.30Fe-1Mn 0.77 625 30 NA 0.48 63.1 20
9.5Si-1.34Fe-1.5Mn 1.1 605 30 NA 0.25 81.3 30
9.5Si-1.14Fe-1.5Mn 1.4 610 30 NA 0.41 64.0 20
9.5Si-0.98Fe-1.5Mn 1.5 605 30 NA 0.18 82 20
Matsubara (Matsubara, Izawa, 11Si-2.07Fe-2.03Mn 1 NA NA 0.30 0.36 82.6 NA Centrifuge
and Nakanishi 1998) 11Si-2.07Fe-2.53Mn 1.22 0.20 0.23 88.9
11Si-2.07Fe-3.00Mn 1.45 0.12 0.15 92.8
11Si-2.07Fe-4.15Mn 2 0.1 0.13 93.7
Figure 28 Microstructure of an aluminum alloy after Fe-rich phase precipitation and subsequent gravity
separation. (a) overall view (Li 2004) (b) SEM view (Cao, Saunders, and Campbell 2004).
Figure 29 Sludge volume, average size, and particle number distribution in the mold during the
gravitational segregation (Shabestari and Gruzleski 1995).
132 L. ZHANG ET AL.
Figure 30 Settling speed of the sludge as a function of time (Shabestari and Gruzleski 1995).
the removal efficiency because the captured small particles easily blocked the filter
pores. Data in Table 9 proved this conclusion. The efficiency of the filtration process
is also affected by the initial aluminum composition. The efficiency increases with
increasing Mn=Fe (>1) or initial Fe and Mn content (Table 9) (Nijhof et al. 1996;
Van Der Donk 1998). The entire efficiency of the filtration process for the same
experimental condition is relatively low compared with other separation methods
Figure 31 Schematic of the filtration process (T1: melting temperature, T2: holding temperature)
(de Moraes et al. 2006).
according to the studies by Nijhof et al. (1996) and Van Der Donk (1998) (Table 9).
The filtration efficiency can be substantially improved by a preliminary gravity sep-
aration step. Flores et al. (1992b) firstly, used gravity separation to purify the alumi-
num melt (initial composition wt%: 1.5 Fe, 2.25 Mn, 9.2 Si, 0.2 Cr, 0.76 Zn, and 3.53
Cu) for 180 min at 640 C and the iron content decreased from 1.5 to 0.73 wt%; then
the upper part of the purified melt was filtered and Fe content was decreased to
0.27 wt%. Approximately, 82% Fe was removed overall. High efficiency was also
obtained in the same way by de Moraes et al. (2006) (Table 9). A slow cooling rate
from the melting temperature to the holding temperature and a longer holding time
(30 min) encouraged the formation and growth of the primary iron intermetallic par-
ticles, which resulted in a high removal efficiency (de Moraes et al. 2006).
Centrifugal Separation
The centrifugal separation technique was applied to directly remove iron-rich
phases from the partially solidified aluminum alloy melts without any other elements
addition, which will be discussed later. Matsubara et al. (Matsubara, Izawa, and
Nakanishi 1998) studied the sludge removal from Al-11Si alloys by the centrifugal
separation technique. The centrifugal separation apparatus that was used is shown
schematically in Figure 32 (Matsubara, Izawa, and Nakanishi 1998). The iron-rich
phases moved to the edge side of the melt and the central part was purified as shown
in Figure 33 (Matsubara, Izawa, and Nakanishi 1998). Figure 34 (Matsubara, Izawa,
and Nakanishi 1998) shows the microstructure of the inner and outer part of the
melt. The rotational speed has a great influence on the purification efficiency as
shown in Figure 35 (Matsubara, Izawa, and Nakanishi 1998). When the rotational
speed is 8.3 s1, small sludge particles are unable to be expelled from the central
Figure 32 Schematic of vertical centrifugal separation apparatus (Matsubara, Izawa, and Nakanishi
1998).
134 L. ZHANG ET AL.
Figure 33 Microstructure of the transverse cross section of the centrifugal separated melt (Matsubara,
Izawa, and Nakanishi 1998).
region and the Fe and Mn content in the central region remained at a high level
>1 wt%. As the rotational speed increases over 16.6 s1, the Fe and Mn content in
the center are slightly affected by the rotation speed (Figure 35). The ratio of Mn=Fe
affects the formation and growth of primary Fe-rich phases and thus the iron
removal fraction varied with Mn=Fe as seen in Figure 36 (Matsubara, Izawa, and
Nakanishi 1998). With increasing the ratio of Mn=Fe, the iron removal fraction
increases and Mn=Fe ¼ 1.5 is the optimum value removing 86% Fe. The removal
fraction of Fe slightly decreases when Mn=Fe exceeded 1.5.
EM Separation
The principle of EM separation of particles from liquids was first proposed by
Kolin in 1953, and by Leenov and Kolin in 1954. All of the other studies are based
on their theory. They considered a single spherical particle or a cylinder having infi-
nite length, placed in conducting liquid with imposition of DC EM field. In this field,
direction of the magnetic field is perpendicular to that of the electric current, thus the
direction of EM force is perpendicular to both directions. They calculated the
migration force by analyzing the EM-flow around the particle. Their analysis has
Figure 34 Typical microstructure of inner and outer parts by centrifugal separation (Al-11Si-2Fe-3Mn)
(Matsubara, Izawa, and Nakanishi 1998).
Figure 35 Effects of rotational speed on the concentration distribution of Fe and Mn in Al-11Si-2Fe-3Mn

(Matsubara, Izawa, and Nakanishi 1998).
been further modified by some researchers. EM separation of inclusions from molten

metal has been investigated since 1961 (Verte 1961; Makarov, Ludwig, and Apelian
2000). In the 1980’s, the principle of applying EM forces to inclusion separation was
Figure 36 Removal fraction of Fe versus Mn=Fe ratio (Matsubara, Izawa, and Nakanishi 1998).
136 L. ZHANG ET AL.
further investigated by Moreau (1980), Marty and Alemany (1982), and Korovin
(1985). Patel and El-Kaddah (1997) analyzed migration force to a single particle
in magnetic field having various angles with respect to the current direction. EM sep-
aration of inclusions from molten aluminum is an emerging and promising method
for the production of ultraclean aluminum alloys to meet the growing demand for
clean aluminum alloys (Park et al. 1994; El-Kaddah, Patel, and Natarajan 1995;
Patel and El-Kaddah 1997; Makarov, Ludwig, and Apelian 1999; Takahashi and
Taniguchi 2003; Ogasawara, Yoshikawa, and Taniguchi 2004). Using numerical
technique, Shu et al. (2000a, 2000b) recently analyzed the single particle with wall
proximity effect, and they also studied the problem of the two particles with their
direction in a row parallel to EM force.
Figure 37 shows the principle of EM separation. When a uniform EM force is

applied to a liquid metal, the metal is compressed by the EM force (Lorentz force)
and a pressure gradient is generated in the metal. The non- or less-conductive par-
ticle suspended in the liquid metal receives only the pressure force because it does
not experience the EM body force. As a result, the particle is forced to move in
the opposite direction of EM force. Based on the Leenov–Kolin’s theory (Kolin
1953; Leenov and Kolin 1954), a non- or less-conductive spherical particle receives
a force in a conductive fluid with the imposition of a uniform EM force. This force
is expressed by
3 pdp3
Fp ¼ F ð2Þ
4 6
Figure 37 Principle of electromagnetic separation of inclusions from the molten metal.

where FP is the force acting on a particle, dP particle diameter, F imposed EM force,

which can be expressed by
Fp ¼ J B ð3Þ
where J is the imposed electric current density (A=m2), and B is the imposed
magnetic flux density (T).
As this force acts inversely against the EM force, it can be applied for the
separation of non- or less-conductive particles from the liquid metal.
Several researchers have proposed the application of EM force to inclusion
separation. The methods of the imposition of the EM force are classified into the
following types:
a. simultaneous imposition of direct current and stationary magnetic field, (Marty

and Alemany 1982; Park et al. 1994; Yoon et al. 2002a, 2002b)
b. imposition of Alternating Current (AC), (Taniguchi and Brimacombe 1994)
c. imposition of alternating magnetic field, (El-Kaddah, Patel, and Natarajan 1995;
Li et al. 2002; Shu et al. 2002; Yamao et al. 1997)
d. imposition of traveling magnetic field, (Tanaka et al. 1995)
e. simultaneous imposition of alternating current and alternating magnetic field,
(Alemany et al. 1980; Taniguchi and Kikuchi 2000)
f. imposition of stationary magnetic field, (Pillin and Gillon 1994) and
g. more recently, strong magnetic field created by a superconducting DC coil
(Makarov, Ludwig, and Apelian 2001) were also presented.
Compared with other methods, AC magnetic field (type c) shows great advan-
tages because no loop circle is needed for the imposed current and no metal electrode
is needed, which usually contaminates the melt. (Shu et al. 1999b)
Korovin (1985, 1986) deduced the theory of separating nonmetallic inclusions
in the melt by using alternating magnetic field induced in an infinitely long solenoid.
El-Kaddah (1988) designed an AC magnetic field for the separation of inclusions
from melt, by which the adjustments of the direction of imposing magnetic field
become easier. He calculated the particle motion in the molten pool of the EM sep-
aration process and found that a vigorously stirred melt may cause problems with
the separation of small inclusions due to the entrapment of the inclusion particles
within flow eddy. Also, due to the weak magnetic field in most areas opposite to
the coil, the removal efficiency is badly influenced (El-Kaddah, Patel, and Natarajan
1995). The flexibility of applying the AC magnetic field generated in an induction coil
to the particle separation was theoretically demonstrated by Yamao et al. (1997).
They also found that the higher removal efficiency was obtained for a circular cru-
cible in a strong magnetic field, where complete mixing is assumed to account for the
effects of melt flow on the inclusion separation efficiency. Shu et al. (1999b) studied
the EM force exerted on the particles in a cylindrical melt and the relation between
various processing parameters and the removal efficiency theoretically.
The advantages of EM separation include:
. Efficient, even for <50 mm small inclusions: The main advantage of EM separ-
ation is that the EM expulsive force exerted on inclusions is only dependent on
138 L. ZHANG ET AL.
the difference in electrical conductivity between the inclusions and melt, and
independent of the density, chemical composition, or different phases (solid,
liquid, or gas) of the inclusions. (Leenov and Kolin 1954; Marty and Alemany
1982). Thus, for the purification of molten aluminum, EM separation can, in prin-
ciple, remove even micrometer-sized inclusions at an almost constant rate by using
high-intensity EM force fields (El-Kaddah, Patel, and Natarajan 1995; Patel and
El-Kaddah 1997; Shu et al. 1999b).
. Environmentally clean (no salt input): Another advantage is that it offers very
clean processes in view of environmental protection, because the use of fluxes,
generally chlorides or fluorides, is avoided, that is, no salt input. With growing
concerns about the role of materials in sustainable development (Schwartz
1999; Garnier 2000; Yagi and Halada 2001), materials processing industries are
paying more attention to the environmental friendliness of the processing methods
while pursuing material performance. EM separation really satisfies this point.
. No contamination to the metal: The melt is also protected against contamination
if induction technology is utilized, since there is no contact between the melt and
the power source due to the intrinsic properties of the magnetic fields, (Garnier
2000) and there is no extra lining refractory introduced.
. Easy adjustment: Unlike conventional separation techniques, the magnitude and
direction of the separation force on the particle can be easily modified by control-
ling the direction and intensity of the electric and magnetic fields in the liquid.
Since the conductivity of Fe-rich phases is less than that of molten aluminum,
they can be separated by EM forces. The mechanism is shown in Figure 37 (Li 2004).
Some Chinese and South Korean researchers (Kim and Yoon 2000; Li et al. 2001;
Xu et al. 2001; Yao, Sun, and Qiu 2001; Li et al. 2003; Xu, Li, and Zhou 2003; Li
2004; Zhang et al. 2005a, 2006a; Jiao and Zhang 2006; Jiao, Zhang, and Yao
2006) have applied EM separation method to remove iron (in the form of sludge)
from the melt. Particularly, the compact morphology of sludge is favorable for the
uniform direction of repulsive force. It is necessary for the successful removal of
sludge (Li 2004). Li et al. (2003) observed that the iron content decreased form
1.13 wt% to 0.41 wt% after the treatment in Al-12Si-1.1Fe-1.2Mn alloys, the micro-
structure before and after the process is shown in Figure 38 (Li 2004). However, the
platelet b-Fe was unable to be separated because its needle-like morphology resulted
in the disorder repulsive force and the platelets moved in random direction [show in
Figure 38(b)] (Li et al. 2003). Li et al. (2003) theoretically proved that the removal
efficiency increased with increasing EM forces and the diameter of sludge particles,
and decreased with increasing the velocity of flowing melt and the height of separ-
ator. Equation (4) gives the definition of removal efficiency (Li et al. 2003). Here,
it is simply assumed that uf ¼ 0.5uM and uM is the maximum velocity of the melt.
The gravity influence was neglected because the EM force is much larger than that
of the gravitational force. It was reported that the removal efficiency of 20 mm
sludge particles was 100% with 8.7 105 N=m3 EM force, uf ¼ 121.3 mm=s and
5 mm separator height (Li et al. 2003; Li 2004).
3zBJ dp2
g¼ ð4Þ
8huf 24l
Figure 38 Electromagnetic separation of Al-12Si-1.1Fe-1.2Mn alloys, (a) Before EM separation; (b) After
EM separation (Li 2004).
where g is the removal efficiency; z is the melt flow direction (Figure 37); B is the
magnetic flux density vector; J is the current density; h is the height of the separator;
uf is the average velocity of the melt; dp is the diameter of sludge particles; and l is
the viscosity of the melt.
Xu, Li, and Zhou (2003) also proposed the removal efficiency of EM
separation as follows:
3xvt
g¼ ð5Þ
4huM

dp2 3
vt ¼ ðqp qÞg þ BJ ð6Þ
18l 4
where g is the removal efficiency; x is the melt flow direction; vt is the terminal velo-
city of particles; B is the magnetic flux density vector; J is the current density; h is the
140 L. ZHANG ET AL.
height of the separator; uM is the maximum velocity of the melt; dp is the diameter of
sludge particles; l is the viscosity of the melt; qp is the density of sludge particles; q is
the density of melt; g is the acceleration due to gravity. Xu’s model included the
effect of gravity. If the gravity influence is neglected, Eq. (6) is simplified into
Eq. (7). Substituting vt in Eq. (5) by Eq. (7) and with uf ¼ 0.5uM, Eq. (4) will be
obtained.
3dp2 BJ
vt ¼ ð7Þ
72 l
Actually, Xu also observed that the gravity influence was rather weak.
Figure 39 (Xu, Li, and Zhou 2003) shows the removal efficiency under different
technological parameters. Figure 39a illustrates that the removal efficiency drops
Figure 39 Relationship between removal !efficiency!

and technological parameter: (a) uM, maximum
velocity of the melt; (b) f, EM force (f ¼ j B J j); (c) x, action length of f; (d) 2h, height of separation
chamber (Xu et al. 2003).
quickly with increasing uM. Figure 39b shows that the removal efficiency of 30 mm
particles is over 90% when the f was 1.5 105 N=m3. Figure 39c shows that >30 mm
particles can all be removed if x ¼ 120 mm. The best iron removal fraction was approxi-
mately 65.8% in Xu’s study (Xu, Li, and Zhou 2003). Jiao reported an iron removal
fraction of 70%. (Jiao, Zhang, and Yao 2006). Kim and Yoon (2000) successfully pur-
ified Fe from 1.64 to 0.45 wt% in Al-7Si scraps. The iron decreases with increasing of
electric current as shown in Figure 40 (Kim and Yoon 2000), which confirms the
removal efficiency theory of Li. Figure 41 (Kim and Yoon 2000) shows the Al-7Si scrap
microstructure processed by the EM separation for 60 s with a magnetic field of 0.3
Tesla. With 10–30 A induced currents, the angular sludge settled on the side and bot-
tom of the tube, but accumulated at the bottom of the tube at 40 A. Therefore, a higher
current favors iron removal. However, the main factor controlling the removal
efficiency of iron was initial Fe and Mn content as well as the size of sludge (Li 2004).
Table 9 shows reported iron removal efficiency by different sludge separation
methods. Comparing with gravity and filtration, EM separation appears more
efficient under similar experimental conditions. Furthermore, EM separation can
also remove nonmetallic inclusions, which has been proven by many researchers
(Shu et al. 1999a; Li et al. 2000b; Shu et al. 2000a, 2000b). EM separation is suitable
for large industrial scale of continuous treatment, and has already been used for
aluminum refining by several Chinese aluminum companies. Gravitational separ-
ation can achieve a high iron removal but needs long settling time. Gravitational set-
tling is nonfeasible for large scale treatment because the sediment sludge causes
serious built-up and the low holding temperature is inappropriate for industrial
operation, and it is a batch process. The combination of gravitational separation
and filtration gives impressive iron removal results but the disadvantages during
gravitational separation step are still there, in addition to the problem of clogging
of filter pores. Although the centrifuge has relatively high removal efficiency, it is
also unpractical for the large scale processing due to challenges with temperature
Figure 40 Removal efficiency of Li from aluminum by EM separated with different currents (Kim and
Yoon 2000).
142 L. ZHANG ET AL.
Figure 41 Distribution of sludge with induction (a) 10, (b) 20, (c) 30, and (d) 40 A (Kim and Yoon 2000).
control, rapid increasing of contracted solid phases, the need for centrifuge cleaning,
and the need for further processing of the compact solid residues (Singleton and
Robinson 1971).
Electrolysis
The expensive process of three-layer-cell electrolysis is the most successful tech-
nique for the removal of iron and silicon from the molten aluminum so far (Van Der
Donk, Nijhof, and Castelijns 1995). Figure 42 shows a sketch of a three-layer
refining cell. The impure molten aluminum anode (9), the barium-sodium-aluminum
halide electrolyte (7) and the pure molten aluminum (5) form three layers in the cell.
Only aluminum in the anode can be dissolved through the reaction Al 3e ¼ Al3þ
during the electrolysis process and Al3þ deposits on the cathode—the carbon elec-
trode at the top of the cell (6). Because the pure molten aluminum is the lightest,
it will stay on the top of the three layers. Thus, the purified aluminum is obtained.
However, the energy consumption for this process is relatively high, 13–14 kWh
kg1 so far (Lu et al. 2004; Benkahla et al. 2008; Martin et al. 2008; Zhao and Lu
2008). This technique is also applied to purify commercial aluminum with low level
of initial impurities into high purity (Vire and Gauckler 1985).
Fractional Solidification
The technique is based on the distributions coefficient (k) of impurities. For k
less than unity, a much lower contents of impurities will result in the solidified than
Figure 42 Sketch of three layers electrolysis process.
that in the residual liquid (Lux and Flemings 1979a, 1979b). The distributions coef-
ficient of iron is 0.03 therefore; it can be removed from aluminum by this method.
However, the technique is not suitable for normal aluminum foundries owing to
their low productivity, thus economically unfavorable. Large scale fractional solidi-
fication is used to produce extreme purity aluminum (99.999 wt% Al) (Dawless et al.
1988; Zhang et al. 2005b; Zhang et al. 2006b; Zhang et al. 2006c) and the aluminum
used for fractional solidification is also required to have a relatively high purity
(99.993 wt%). It is reported that iron content decreased from 747 to 24 ppm after
2 h processing in a fractional solidification apparatus (Toshiaki Iuchi et al. 1988).
Zhang et al. (2006c) developed a new fractional solidification apparatus as shown
in Figure 43. High purity grade of Al (99.993 wt%) and extremely low Fe content
0.6–0.9 ppm was achieved (Zhang et al. 2006c).
Electroslag Refining
Electroslag refining process (ESR) is a secondary refining process in which the
slag or flux is used both as a heating source and as a refining medium, as shown in
Figure 44 (Benz and Hills 1992). The process is already well established for ferrous
metals but has not been used for aluminum refining. Early attempts (Bhutani,
Abraham, and Nair 1975; Shrikhande et al. 1986; Sampath and Dwarakadasa
1992) indicated that no iron removal was observed, except that Mohanty’s investi-
gation (Mohanty, Subramanian, and Hajra 1986) reported 26% iron removal (from
0.22 to 0.16 wt%) from commercial aluminum by ESR with a flux containing alumi-
num phosphide. Further investigations on the use of ESR in aluminum refining have
not been reported.
Fluxing Refining
Few references have reported that significant iron removal can be achieved by
using flux. The studies of the current author showed that the addition of Na2B4O7
144 L. ZHANG ET AL.
Figure 43 Sketch of the fractional solidification apparatus developed by Zhang (Zhang et al. 2006b).
Figure 44 Sketch of the electroslag refining apparatus (Benz and Hills 1992).
flux with NaCl and KCl significantly lowered the iron content from 0.33 to 0.18 wt%
(Gao et al. 2007) in laboratory experiments. The possible mechanism was the forma-
tion of stable Fe2B phases by the reaction between iron and the Na2B4O7 and the
simultaneous capturing process by the flux. However, industrial scale experiments
showed little removal of iron. Nijhof et al. (1996) reported iron removal with a mixed
flux of NaCl, KCl, and NaF as shown Figure 45, an iron removal from 0.9 to
0.7 wt% was obtained. However, the mechanism and detailed processing were
not mentioned in his paper and no further studies were reported.
Direct Centrifugal Separation of Fe-Rich Phases

During the method of centrifugal separation, the melt is cooled to a tempera-

ture at which solid phases form, that is, the melt is partially solidified. The first sol-
idified phases include mainly silicon, iron, or titanium. Then, the centrifuge is started
and cooling is continued. Liquid is expelled from the solid framework and collected.
Temperature is an important factor controlling the removal efficiency of impurities
because the precipitation of solid crystals contributes to the success of extraction.
Singleton and Robinson (1971) reported that near-eutectic phase was obtained after
the successful centrifugal separation of raw aluminum alloys (30 wt% Si, 4 wt% Fe,
2 wt% Ti) in a centrifuge bowl apparatus (Figure 46). The separated solidified
phases, which were primary silicon and d-AlFeSi, are shown in Figure 47 (Singleton
and Robinson 1971). The formation of network structure was due to the different
density of precipitated phases. This rigid and porous structure was desirable for
the flow of liquid without blocking, which was an important requirement to use cen-
trifugal separation technology in real practice. The composition of filtrate from the
Figure 45 Removal of impurities from aluminum using a molten salt mixture after 16 h at 750 C (Nijhof
et al. 1996).
146 L. ZHANG ET AL.
Figure 46 Centrifugal apparatus, (a) Components of the removable-wall centrifuge bowl; (b) The centri-
fuge bowl assembled (Singleton and Robinson 1971).
centrifuge varied with the separation temperature, as shown in Figure 48 (Singleton

and Robinson 1971). Iron content decreased from 4 to 1 wt% after the centrifuge
was at 600 C. A higher separation temperature imperiled the separation efficiency
of impurities. Guldin, Arnold, and Matsenko 1979) carried out a centrifugal separ-
ation experiment for an aluminum alloy of composition (12.52 wt% Si, 1.87 wt% Fe,
0.94 wt% Ti) and reported a removal of over 50% iron and 90% Ti at 580 C separ-
ation temperature, while no iron removal at > 650 C as shown in Table 10. The
extracted impurity containing solidified phases was claimed to contain TiAl3 and a
mixture of TiAl3 and b-AlFeSi at 650 C and 580 C respectively. Li et al. (2000a)
developed a mathematical model to predict the migration velocity of the solid phase
at every moment during the solidification process of Al-Fe alloys under centrifugal
casting. Brun et al. (2007) studied the relationship between the collected mass of par-
ticles and the centrifugal time for Al-2.5Fe-1.5Mn alloys by water model, laboratory
scale and pilot scale experiments, however, no iron removal efficiency was stated in
his paper.
The energy used in the process of direct centrifugal separation is much lower
than the electrolysis process. In addition to the separation of primary Fe-rich phases,
Figure 47 Extracted solidified phases by centrifugal process: left, predominantly silicon; right, silicon plus
d-AlFeSi (Singleton and Robinson 1971).
Figure 48 Relationship between filtrate composition and separation temperature (Singleton and Robinson
1971).
Table 10. Composition of Fe, Si, and Ti in aluminum after centrifugal

separation (Guldin, Arnold, and Matsenko 1979)
Impurities Initial 650 C 580 C
Si (wt%) 12.52 12.4 11.66

Fe (wt%) 1.87 1.9 0.81
Ti (wt%) 0.94 0.38 0.09
the centrifugal separation method is also applied to remove sludge in Al-Si alloys
from the melt as described in section A.3 of the current paper.
SUMMARY
It is well known that iron has a detrimental effect on aluminum alloys. The for-
mation of brittle needle-like b-Fe-rich phases, as stress raiser and points of fracture,
is the main reason for the decline in the properties of the alloys. The iron removal
techniques are mainly focused on the formation and separation of primary a-Al
(FeMn)Si. More efforts are devoted on the modification of platelet iron rich phase
in aluminum.
The detrimental effect of iron can be partially reduced by increasing cooling
rate, solution heat treatment and the addition of neutralization elements. The modi-
fication process is complicated and unpredictable because it is deeply affected by
casting conditions and the initial content of impurities. Mn is the most common
element used for neutralization. Manganese in the alloy increases the total amount
148 L. ZHANG ET AL.
of Fe-rich phases, which is also harmful to aluminum alloys. Cr and Co have equal
or more modification ability than Mn, but increase the total amount of Fe-rich
phases. Be seems a good choice but it is toxic. Available data shows the neutraliza-
tion effect of Sr as marginal and still needs further investigation before being applied.
So far, three-layer-cell electrolysis is the most successful technique to remove
iron from the aluminum but it is expensive and only suitable for high purity alumi-
num production, because all the impurity elements are removed from the feed
material. The precipitation of Fe-rich phases (sludge) is a well-established method
for laboratory scale but difficult for continuous treatment process. The technology
of EM separation can efficiently remove Fe-rich phases and could be a continuous
process. EM separation is faster than gravitational separation, and avoids the prob-
lem of pore blockage during the filtration process. Due to the disadvantages of Mn
addition, new elements that can efficiently form ‘‘sludge’’ should be investigated. The
elements close to Fe in the Periodic Table may be good choices. The feasibility for
the formation of Fe-rich compounds in Al-Fe-X system can be thermodynamically
evaluated. Since it was reported that the electroslag technique removed 26% Fe from
aluminum with the presence of aluminum phosphides, aluminum phosphide seems a
clue for the removal of iron too. The flux refining is also a good way to remove iron,
which is easy to be applied by industry. Na2B4O7 flux processing is a possibility and
more effort should be devoted on it.
In summary, the following further studies will be carried out:
. Exploring new elements that can efficiently neutralize iron phases in aluminum
alloys or form primary Fe-rich phase at high temperature (700–750 C).
. Exploring possible flux to directly remove iron from aluminum alloys.
. Theoretical and experimental study on the removal of sludge [primary a-Al15 (Fe,
Mn)3Si2] from Al-Si alloy melt through EM separation.
ACKNOWLEDGMENTS
This research is supported by the Research Board Grant, Laboratory of Green
Process Metallurgy and Modeling (GPMM), Material Research Center (MRC),
Intelligent Systems Center (ISC) at Missouri University of Science and Technology
(Missouri S&T).
REFERENCES
Alemany, A., Argous, J. P., Barbet, J., Ivanes, M., Moreau, R., and Poinsot, S., 1980, ‘‘Elec-
tromagnetic device for the separation of inclusions contained in an electrically conducitve
fluid.’’ French Patent No. 80400 4430.
Allen, C. M., O’reilly, K. A. Q., Cantor, B., and Evans, P. V., 1998, ‘‘Intermetallic phase
selection in 1xxx Al alloys.’’ Progress in Materials Science, 43, pp. 89–170.
Apelian, D., 2009, ‘‘Aluminum cast alloys: Enabling tools for improved performance.’’ In
Worldwide Report, Wheeling, IL: NADCA, 2009.
Ashtari, P., Tezuka, H., and Sato, T., 2003, ‘‘Influence of Sr and Mn additions on intermetallic
compound morphologies in Al-Si-Cu-Fe cast alloys.’’ Materials Transactions, 44,
pp. 2611–2616.
Awano, Y. and Shimizu, Y., 1990, ‘‘Non-equilibrium crystallization of alfesi compound in

melt-superheated Al-Si alloy castings.’’ In Transactions of the American Foundrymen’s
Society, Detroit, Michigan, MI: American Foundrymen’s Society, pp. 889–895.
Backerud, L., Chai, G., and Tamminen, J., 1990, Solidification Characteristics of Aluminum
Alloys. Vol. 2. Foundry Alloys, Detroit: American Foundrymen’s Society, p. 266.
Balitchev, E., Jantzen, T., Hurtado, I., and Neuschutz, D., 2003, ‘‘Thermodynamic assess-
ment of the quaternary system Al-Fe-Mn-Si in the Al-rich corner.’’ Calphad, 27, pp.
275–278.
Barresi, J. G., Couper, M. J., St. John, D. H., Edwards, G. A., and Wang, H., 2002, ‘‘Alumi-
numalloy for producing high performance shaped castings.’’ US Patent No. 2002=
0155023.
Benkahla, B., Caratini, Y., Mezin, H., Renaudier, S., and Fardeau, S., 2008, ‘‘Last develop-
ment in Ap50 cell.’’ Light Metals, pp. 451–455.

Benz, M. G. and Hills, B., 1992, ‘‘Direct Processing of Electroslag Refined Metal.’’ US Patent
No. 5325906.
Bhutani, A. K., Abraham, K. P., and Nair, K. D., 1975, ‘‘Electroslag Refining of Al Alloy.’’
Transactions of the Indian Institute of Metals, 28, pp. 316–318.
Cao, X. and Campbell, J., 2000, ‘‘Effect of precipitation of primary intermetallic compounds
on tensile properties of cast Al-11.5Si-0.4Mg alloy.’’ Transactions of the American
Foundry Society and the 104th Annual Castings Congress, Pittsburgh, PA, April 8–9,
pp. 391–400.
Cao, X. and Campbell, J., 2004a, ‘‘Effect of Sr on primary a-Fe phase in liquid Al-11.5Si-
0.4Mg cast alloy.’’ Materials Science and Technology, 20, pp. 514–520.
Cao, X. and Campbell, J., 2004b, ‘‘The solidification characteristics of Fe-rich intermetallics in
Al-11.5Si-0.4Mg cast alloys.’’ Metallurgical and Materials Transactions a-Physical Metal-
lurgy and Materials Science, 35A, pp. 1425–1435.
Cao, X. and Campbell, J., 2006, ‘‘Morphology of beta-Al5FeSi phase in Al-Si cast alloys.’’
Materials Transactions, 47, pp. 1303–1312.
Cao, X., Saunders, N., and Campbell, J., 2004, ‘‘Effect of iron and manganese contents on
convection-free precipitation and sedimentation of primary alpha-Al(Fe Mn)Si phase
in liquid Al-11.5Si-0.4Mg alloy.’’ Journal of Materials Science, 39, pp. 2303–2314.
Carpenter, G. J. C. and Lepage, Y., 1993, ‘‘Revised cell data for the beta-FeSiAl phase in
aluminum alloys.’’ Scripta Metallurgica et Materialia, 28, pp. 733–736.
Chadwick, R., Muir, N. B., and Grainger, H. B., 1953, ‘‘Effect of iron, manganese, and chro-
mium on properties in sheet form of aluminium alloys containing 0.7% magnesium and
1.0% silicon.’’ Institute of Metals—Journal, 82, pp. 75–80.
Cho, Y. H., Lee, H. C., Oh, K. H., and Dahle, A. K., 2008, ‘‘Effect of Strontium and Phos-
phorus on Eutectic Al-Si Nucleation and Formation of beta-Al5FeSi in Hypoeutectic
Al-Si Foundry Alloys.’’ Metallurgical and Materials Transactions A., 39, pp. 2435–2448.
Closset, B. and Gruzleski, J. E., 1982a, ‘‘Mechanical properties of A356.0 alloys modified with
pure strontium. (Retroactive coverage).’’ Transactions of the American Foundrymen’s
Society, 90, pp. 453–464.
Closset, B. and Gruzleski, J. E., 1982b, ‘‘Structure and properties of hypoeutectic Al–Si–Mg
alloys modified with pure strontium.’’ Metallurgical Transactions A, 13, pp. 945–951.
Closset, B., Paray, F., Gruzleski, J., and Mulazimoglu, H., 1996, ‘‘Microstructures and
properties of strontium treated aluminium electrical conductor alloys.’’ Light Metals,
pp. 737–744.
Colwell, D. L. and Kissling, R. J., 1961, ‘‘Die and permanent mold casting aluminum alloy
minor elements.’’ AFS Transactions, 69, pp. 610–615.
Cook, M. and Chadwick, R., 1944, ‘‘Effect of nickel and iron on properties of wrought
aluminium-copper-magnesium alloy.’’ Institute of Metals—Papers, pp. 423–433.
150 L. ZHANG ET AL.
Cooper, M., 1967, ‘‘The crystal structure of the ternary alloy alpha.’’ ACTA Crystallogr, 23,
pp. 1106–1107.
Cooper, M. and Robinson, K., 1966, ‘‘The crystal structure of the ternary alloy alpha-
AlMnSi.’’ ACTA Crystallogr, 20, pp. 614–617.
Couture, A., 1981, ‘‘Iron in aluminum casting alloys–a literature survey.’’ International Cast
Metals Journal, 6, pp. 9–17.
Crepeau, P. N., 1995, ‘‘Effect of iron in Al-Si casting alloys: A critical review.’’ Transactions of
the American Foundrymen’s Society, 103, pp. 361–366.
Czikel, J., Pfeiffer, W. D., Sabath, G., and Steinhaufl, B., 1985, ‘‘Effects of zinc, iron and
manganese on the properties of AlSi8Cu3 type alloys.’’ Aluminium, 61, pp. 917–922.
Damoah, L. N. W. and Zhang, L., 2010, ‘‘Removal of inclusions from aluminum through
filtration.’’ Metallurgical and Materials Transactions B, 41, 886–907.
Dawless, R. K., Troup, R. L., Meier, D. L., and Rohatgi, A., 1988, ‘‘Production of
extreme-purity aluminum and silicon by fractional crystallization processing.’’ Journal
of Crystal Growth, 89, pp. 68–74.
De Moraes, H. L., De Oliveira, J. R., Espinosa, D. C. R., and Tenorio, J. A. S., 2006,
‘‘Removal of iron from molten recycled aluminum through intermediate phase filtration.’’
Materials Transactions, 47, pp. 1731–1736.
Depierre, V. and Bernstein, H., 1951, ‘‘Grain refinement of aluminum-silicon (5% Si) and
aluminum-silicon-magnesium (7% Si, 0.3% Mg) casting alloys.’’ Transaction of the
American Society of Metals, 43, pp. 635–643.
Dichtl, H. J., 1970, ‘‘Use of molybedenum to compensate for the adverse effect of iron in
aluminium-silicon alloys.’’ Aluminium, 46, pp. 362–364.
Dinnis, C. M., Taylor, J. A., and Dahle, A. K., 2005, ‘‘As-cast morphology of iron-
intermetallics in Al-Si foundry alloys.’’ Scripta Materialia, 53, pp. 955–958.
Dinnis, C. M., Taylor, J. A., and Dahle, A. K., 2006, ‘‘Interactions between iron, manganese,
and the Al-Si eutectic in hypoeutectic Al-Si alloys.’’ Metallurgical and Materials
Transactions A, 37, pp. 3283–3291.
Donnadieu, P., Lapasset, G., and Sanders, T. H., 1994, ‘‘Manganese-induced ordering in
the alpha-(Al-Mn-Fe-Si) approximant phase.’’ Philosophical Magazine Letters, 70,
pp. 319–326.
Dunn, R., 1965, ‘‘Aluminum melting problems and their influence on furnace selection.’’ Die
Casting Engineer, B9, pp. 8, 10–12, 14, 16, 20–26, 28, 30.
El-Kaddah, N., 1988, ‘‘Comprehensive mathematical model of electromagnetic separation of
inclusion in molten metals.’’ Conference Record – IAS Annual Meeting (IEEE Industry
Applications Society), 35, pp. 1162–1167.
El-Kaddah, N., Patel, A. D., and Natarajan, T. T., 1995, ‘‘Electromagnetic filtration of molten
aluminum using an induced-current separator.’’ JOM Journal of the Minerals Metals and
Materials Society, 47, pp. 46–49.
Flores, A., Escobedo, J., Mendez, J., and Mendez, M., 1992a, ‘‘Kinetic mechanisms of iron
segregation from Al-Si-Cu-Fe-Mn melts.’’ Light Metals, pp. 845–850.
Flores, A., Escobedo, J. C., Mendex, J., Mendex, M., and Mercado, H., 1992b, ‘‘Application
of the segregation process for the iron removal from aluminum-silicon scrap.’’ 3rd Inter-
national Conference on Molten Aluminum Processing, Orlando, FL, November 9–10,
pp. 293–300.
Flores, A., Sukiennik, M., Castillejos, A. H., Acosta, F. A., and Escobedo, J. C., 1998, ‘‘A
kinetic study on the nucleation and growth of the Al8FeMnSi2 intermetallic compound
for aluminum scrap purification.’’ Intermetallics, 6, pp. 217–227.
Fridlyander, I. N., Tkachenko, E. A., and Matveets, E. N., 1981, ‘‘The influence of iron, nickel
and zirconium on the structure and properties of type V93pch aluminum alloy.’’ Metal-
lovedenie i Termicheskaya Obrabotka Metallov, 2, pp. 47–49.
Fukuda, J., Ohtani, Y., Kiyose, A., Kawase, T., and Tsutsumi, K., 1998, ‘‘Improvement of
slab quality with in-mold electromagnetic stirrer.’’ 3rd European Conference on Continu-
ous Casting, Madrid, Spain, pp. 437–445.
Gao, J. W., Shu, D., Wang, J., and Sun, B. D., 2007, ‘‘Effects of Na2B4O7 on the elimination
of iron from aluminum melt.’’ Scripta Materialia, 57, pp. 197–200.
Garnier, M., 2000, ‘‘Present and future prospects in electromagnetic processing of materials.’’
Proceedings of the 3rd International Symposiumon Electromagnetic Processing of Materials
(ISIJ), Tokyo, Japan, pp. 3.
Gesing, A., Berry, L., Dalton, R., and Wolanski, R., 2002, ‘‘Assuring continued recyclability
of automotive aluminum alloys: grouping of wrought alloys by color, X-ray absorption
and chemical composition-based sorting.’’ TMS 2002 Annual Meeting: Automotive Alloys
and Aluminum Sheet and Plate Rolling and Finishing Technology Symposia. Seattle, WA:
Minerals, Metals and Materials Society (TMS), pp. 3–17.

Gobrecht, J., 1975, ‘‘Settling-out of Fe, Mn and Cr in Al-Si casting alloys. Pt. 1.’’ Giesserei, 62,
pp. 263–266.
Granger, D. A., Sawtell, R. R., and Kersker, M. M., 1984, ‘‘Effect of beryllium on the
properties of A357.0 castings. (Retroactive coverage).’’ Transactions of the American
Foundrymen’s Society, 92, pp. 579–586.
Green, J. A. S., 2007, Aluminum Recycling and Processing for Energy Conservation and Sustain-
ability, Materials Park, OH: ASM International, pp. 53.
Guldin, I. T., Arnold, A. A., and Matsenko, Y. A., 1979, ‘‘Refining aluminum-silicon alloys
free of iron and titanium.’’ Tsvetnye metally (The Soviet Journal of Non-Ferrous Metals),
pp. 48–50.
Gustafsson, G., Thorvaldsson, T., and Dunlop, G. L., 1986, ‘‘The influence of iron and
chromium on the microstructure of cast Al–Si–Mg alloys.’’ Metallurgical Transactions
A, 17, pp. 45–52.
Hanna, M. D., Lu, S. Z., and Hellawell, A., 1984, ‘‘Modification in the aluminum–silicon
system.’’ Metallurgical Transactions A, 15, pp. 459–469.
Iglessis, J., Frantz, C., and Gantois, M., 1977, ‘‘Conditions for the formation of the Fe phases
in commercial purity Al-Si alloys.’’ Mem. Sci. Rev. Metal, 74, pp. 237–242.
Iuchi, T., Sugita, K., Sagisaka, E., Tuchihashi, N., Kanamori, T., and Sugiyama, M. 1988,
‘‘Process and apparatus for refining aluminum.’’ US Patent No. 4734127.
Jiao, W.-L. and Zhang, L., 2006, ‘‘Purification of iron rich impurity from aluminum-
silicon alloy by electromagnetic seperation.’’ Foundry Technology (Chinese), 27,
pp. 269–272.
Jiao, W.-L., Zhang, L., and Yao, G.-C., 2006, ‘‘Seperated effection iron-rich impurity from
aluminum-silicon alloy by parameter of electromagnetic seperated equipment.’’ Foundry
(Chinese), 55, pp. 179–181.
Jones, H., 1969, ‘‘Observations on a structural transition in aluminum alloys hardened by
rapid solidification.’’ Materials Science and Engineering, 5, pp. 1–18.
Jorstad, J. L., 1986, ‘‘Understanding sludge.’’ Die Casting Engineer, 30, pp. 30–36.
Khalifa, W., Samuel, F. H., and Gruzleski, J. E., 2003, ‘‘Iron intermetallic phases in the Al
corner of the Al-Si-Fe system.’’ Metallurgical and Materials Transactions A, 34,
pp. 807–825.
Kholnova, V. I., 1968, ‘‘The high-strength forging aluminium alloy, V93.’’ Paper from
Alyumin Splavy Metallurgiya, Moscow, 5, pp. 169–176.
Kim, H. Y., Park, T. Y., Han, S. W., and Lee, H. M., 2006, ‘‘Effects of Mn on the crystal
structure of alpha-Al (Mn,Fe)Si particles in A356 alloys.’’ Journal of Crystal Growth,
291, pp. 207–211.
Kim, J. H. and Yoon, E. P., 2000, ‘‘Elimination of Fe element in A380 aluminum alloy scrap
by electromagnetic force.’’ Journal of Materials Science Letters, 19, pp. 253–255.
152 L. ZHANG ET AL.
Kolin, A., 1953, ‘‘An electromagnetokinetic phenomenon involving migration of neutral

particles.’’ Science, 117, pp. 134–137.
Korovin, V. M., 1985, ‘‘Separation of particles, suspended in a conducting liquid with the help
pf an alternating electromagnetic field.’’ Magnetohydrodynamics, 21, pp. 321–326.
Korovin, V. M., 1986, ‘‘Motion of a drop in a conducting liquid under the action of a variable
electromagnetic field.’’ Magnetohydrodynamics, 22, pp. 17–19.
Kosuge, H. and Mizukami, I., 1975, ‘‘Formation of ‘Fir-Tree’ structure in D.C. Cast ingots of
Al-0.6per cent Fe alloys.’’ Journal of Japan Institute of Light Metals, 25, pp. 48–58.
Kral, M. V., 2005, ‘‘A crystallographic identification of intermetallic phases in Al-Si alloys.’’
Materials Letters, 59, pp. 2271–2276.
Kral, M. V., Mcintyre, H. R., and Smillie, M. J., 2004, ‘‘Identification of intermetallic phases
in a eutectic Al-Si casting alloy using electron backscatter diffraction pattern analysis.’’
Scripta Materialia, 51, pp. 215–219.

Kral, M. V., Nakashima, P. N. H., and Mitchell, D. R. G., 2006, ‘‘Electron microscope studies
of Al-Fe-Si intermetallics in an Al-11 Pct Si alloy.’’ Metallurgical and Materials
Le Brun, P., Krautlein, C., Rombach, G., Pouly, P., De Vries, P., and Luyten, J., 2007,
‘‘Removal of intermetallic particles for the purification of aluminum alloys.’’ Light
Metals, pp. 657–664.
Leenov, D. and Kolin, A., 1954, ‘‘Theory of electromagnetophoresis. I. Magnetohydrody-
namic forces experienced by spherical and symmetrically oriented cylindrical particles.’’
Journal of Chemical Physics, 22, pp. 683–688.
Li, K., Wang, J., Shu, D., Sun, T. X., Sun, B. D., and Zhou, Y. H., 2002, ‘‘Theoretical and
experimental investigation of aluminum melt cleaning using alternating electromagnetic
field.’’ Materials Letters, 56, pp. 215–220.
Li, R. D., Xu, Y. Q., Ma, B., and Bai, Y. H., 2000a, ‘‘Migration of solid phase during solidi-
fication process of Al-Fe alloy under centrifugal casting.’’ Chinese Journal of Nonferrous
Metals, 10, pp. 353–357.
Li, T. X., Xu, Z. M., Sun, B. D., Shu, D., and Zhou, Y. H., 2000b, ‘‘Remove inclusions
from aluminum melt in electromagnetic field.’’ Acta Metallurgica Sinica, 13, pp. 1068–
1074.
Li, T. X., 2004, ‘‘Application of electromagnetic seperation in aluminum melt.’’ PhD thesis,
Shanghai Jiao Tong University, Shanghai, China.
Li, T. X., Xu, Z. M., Sun, B. D., Shu, D., and Zhou, Y. H., 2003, ‘‘Electromagnetic separation
of primary iron-rich phases from aluminum-silicon melt.’’ Transactions of the Nonferrous
Metals Society of China, 13, pp. 121–125.
Li, T. X., Xu, Z. M., Zhang, X. P., Sun, B. D., and Zhou, Y. H., 2001, ‘‘Reducing iron content
In Al-Si alloy by electromagnetic seperation.’’ Journal of Shanghai Jiao Tong University,
35, pp. 664–667.
Liu, P., Thorvaldsson, T., and Dunlop, G. L., 1986, ‘‘Formation of intermetallic compounds
during solidification of dilute Al–Fe–Si alloys.’’ Materials Science and Technology, 2,
pp. 1009–1018.
Lu, H., Yuan, W., Xu, C., Liu, Z., Wu, B., and Zhang, Y., 2004, ‘‘The development of 65KA
three-layer electrolysis cell for refining of aluminium.’’ Light Metals, pp. 303–305.
Lu, L. and Dahle, A. K., 2005, ‘‘Iron-rich intermetallic phases and their role in casting defect
formation in hypoeutectic Al-Si alloys.’’ Metallurgical and Materials Transactions A, 36,
pp. 819–835.
Lux, A. L. and Flemings, M. C., 1979a, ‘‘Refining by fractional melting.’’ Metallurgical trans-
actions. B, 10, pp. 79–84.
Lux, A. L. and Flemings, M. C., 1979b, ‘‘Refining by fractional solidification.’’ Metallurgical
Transactions. B, 10, pp. 71–78.
Mahta, M., Emamy, M., Daman, A., Keyvani, A., and Campbell, J., 2005, ‘‘Precipitation of
Fe rich intermetallics in Cr- and Co-modified A413 alloy.’’ International Journal of Cast
Metals Research, 18, pp. 73–79.
Makarov, S., Ludwig, R., and Apelian, D., 1999, ‘‘Inclusion removal in molten aluminum:
mechanical, electromagnetic, and acoustic techniques.’’ Transactions of the American
Foundrymen’s Society, 107, pp. 727–735.
Makarov, S., Ludwig, R., and Apelian, D., 2000, ‘‘Electromagnetic separation techniques in metal
casting. I. Conventional methods.’’ IEEE Transactions on Magnetics, 26, pp. 2015–2021.
Makarov, S., Ludwig, R., and Apelian, D., 2001, ‘‘Electromagnetic separation techniques in
metal casting. II. Separation with superconduting coils.’’ IEEE Transactions on
Magnetics, 37, pp. 1024–1031.
Martin, O., Despinasse, S., Ritter, C., Santerre, R., and Tomasino, T., 2008, ‘‘The fecri
approach and the latest developments in the Ap3x technology.’’ Light Metals,
pp. 255–260.
Marty, P. and Alemany, A., 1982, ‘‘Theoretical and experimental aspects of electromagnetic
separation.’’ Proceedings of the Symposium of the IUTAM, Cambridge, UK, September
6–10, pp. 245–259.
Mascre, C., 1955, ‘‘Influence of iron and manganese on type A-S13 (Alpax) alloys.’’ Fonderie,
108, pp. 4330–4336.
Matsubara, H., Izawa, N., and Nakanishi, M., 1998, ‘‘Macroscopic segregation in Al-11
mass% Si alloy containing 2 mass% Fe solidified under centrifugal force.’’ Journal of
Japan Institute of Light Metals, 48, pp. 93–97.
Mbuya, T. O., Odera, B. O., and Ng’ang’a, P. S., 2003, ‘‘Influence of iron on castability and
properties of aluminum silicon alloys: literature review.’’ International Journal of Cast
Metals Research, 16, pp. 451–465.
Miki, I., Kosuge, H., and Nagahama, K., 1975, ‘‘Supersaturation and Decomposition of Al-Fe
alloys during solidification.’’ Journal of the Japan Institute of Light Metals, 25, pp. 1–9.
Miresmaeili, S. M., Campbell, J., Shabestari, S. G., and Boutorabi, S. M. A., 2005, ‘‘Precipi-
tation of Sr-rich intermetallic particles and their influence on pore formation in
Sr-modified A356 alloy.’’ Metallurgical and Materials Transactions A, 36, pp. 2341–2349.
Mohanty, B. P., Subramanian, S., and Hajra, J. P., 1986, ‘‘Electroslag refining of commercial
aluminium. (Retroactive coverage).’’ Trans. Indian Inst. Met, 39, pp. 646–651.
Mondolfo, L. F., 1976, Aluminum Alloys: Structure and Properties, London: Butterworths.
Moreau, R., 1980, ‘‘Metallurgical Applications of Magnetohydrodynamics.’’ Proceedings 15th
Congress Theoretical Applied Mechanics, Toronto, ON, August 17–23, pp. 107–118.
Morris, L. R. and Miners, F. B., 1975, Aluminum Alloys. Montreal, QC: Alcan Research and
Development Limited.
Mulazimoglu, M. H., Gruzleski, J. E., and Closset, B., 1992, ‘‘Evaluation of the metallurgical
effects of strontium on cast 6000 series aluminum alloys.’’ Aluminium, 68, pp. 489–493.
Mulazimoglu, M. H., Paray, F., Gruzleski, J. E., and Kulunk, B., 1994, ‘‘Modification of
intermetallic phases by strontium in aluminum wrought alloys.’’ Light Metals, 27,
pp. 1047–1056.
Murali, S., Arvind, T. S., Raman, K. S., and Murthy, K. S. S., 1997a, ‘‘Fatigue properties of
sand cast, stircast and extruded Al-7Si-0.3 Mg alloy with trace additions of Be and Mn.’’
Materials Transactions, JIM, 38, pp. 28–36.
Murali, S., Guru Row, T. N., Sastry, D. H., Raman, K. S., and Murthy, K. S. S., 1994,
‘‘Crystal structure of beta-FeSiAl sub 5 and (BeFe)-BeSiFe sub 2 Al sub 8 phases.’’
Scr. Metall. Mater, 31, pp. 267–272.
Murali, S., Muthukkaruppan, S., Raman, K. S., and Murthy, K. S. S., 1997b, ‘‘Stir cast and
extruded Al-7Si-0.3Mg alloy containing iron and beryllium.’’ Materials Science and
Technology, 13, pp. 337–342.
154 L. ZHANG ET AL.
Murali, S., Raman, K. S., and Murthy, K. S. S., 1994, ‘‘Morphological studies on beta-
FeSiAl5 phase in Al- 7Si-0.3Mg alloy with trace additions of Be, Mn, Cr, and Co.’’
Materials Characterization, 33, pp. 99–112.
Murali, S., Raman, K. S., and Murthy, K. S. S., 1995a, ‘‘Effect of trace additions (beryllium,
chromium, manganese and cobalt) on the mechanical properties and fracture toughness
of iron-containing Al-7Si-0.3Mg alloy.’’ Ljevarstvo, 37, pp. 3–14.
Murali, S., Raman, K. S., and Murthy, K. S. S., 1995b, ‘‘The formation of beta-FeSiAl 5
and Be-Fe phases in Al-7Si-0.3 Mg alloy containing Be.’’ Materials Science &
Engineering A (Structural Materials: properties, Microstructure and Processing), A190,
pp. 165–172.
Murali, S., Raman, K. S., and Murthy, K. S. S., 1996, ‘‘Al-7Si-0.3Mg cast alloy: Formation
and crystal structure of beta-FeSiAl sub 5 and (Be-Fe)-BeSiFe sub 2 Al sub 8 phases.’’
Materials Science Forum, 217–222, pp. 207–212.

Murali, S., Trivedi, A., Shamanna, K. S., and Murthy, K. S. S., 1996, ‘‘Effect of iron and com-
bined iron and beryllium additions on the fracture toughness and microstructures of
squeeze-cast Al-7Si-0.3Mg alloy.’’ Journal of Materials Engineering and Performance, 5,
pp. 462–468.
Narayanan, L. A., Samuel, F. H., and Gruzleski, J. E., 1994, ‘‘Crystallization behavior
of iron-containing intermetallic compounds in 319 aluminum-alloy.’’ Metallurgical
and Materials Transactions a-Physical Metallurgy and Materials Science, 25, pp. 1761–
1773.
Nijhof, G. H., Van Der Donk, H. M., Castelijns, C. A. M., Bouvet, S., and Bertaud, Y., 1996,
‘‘Purification of remelted aluminum packagings.’’ Light Metals, pp. 1065–1069.
Ogasawara, T., Yoshikawa, N., and Taniguchi, S., 2004, ‘‘Flow of conducting liquid around
two nonconducting particles in DC electromagnetic field and the electromagnetic
migration force.’’ Metallurgical and Materials Transactions, 35B, pp. 847–855.
Panday, P. K. and Schubert, K., 1969, ‘‘Structure investigation in a mixed compound T-B3-
B4 T equals Mn Fe Co Ir Ni Pd B3 equals A1 Ga T1 B4 equals Si Ge.’’ Journal of the Less
Common Metals, 18, pp. 175–202.
Park, J.-P., Morihira, A., Sassa, K., and Asai, S., 1994, ‘‘Elimination of non-metallic inclu-
sions using electromagnetic force.’’ Tetsu-to-Hagane, 80, pp. 31–36.
Passmore, E. M., Flemings, M. C., and Taylor, H. F., 1958, ‘‘Fundmental studies on effects of
sloution treatment, iron content, and chilling of sand cast aluminum-copper alloy.’’ AFS
Transaction, 66, pp. 96–103.
Patel, A. D. and El-Kaddah, N., 1997, ‘‘Kinetics of inclusion removal from molten aluminum
under an applied alternating magnetic field.’’ TMS Annual Meeting, Proceedings of Ses-
sions—Light Metals, Orlando, FL, February 9–13, pp. 1013–1018.
Pennors, A., Samuel, A. M., Samuel, F. H., and Doty, H. W., 1998, ‘‘Precipitation of beta-Al
sub 5 FeSi iron intermetallic in Al-6% Si-3.5% Cu (319) type alloys: role of Sr and P.’’ One
Hundred Second Annual Meeting of the American Foundrymen’s Society, Atlanta, GA;
USA, American Foundrymen’s Society, pp. 251–264.
Philips, H. W., 1976, Annotated Equilibrium Diagrams of Some Aluminum Alloy Systems,
London: The Institute of Metal.
Phragmen, G., 1950, ‘‘On phases occurring in alloys of aluminium with copper, magnesium,
manganese, iron, and silicon.’’ Institute of Metals—Journal, 77, pp. 489–552.
Pillin, B. and Gillon, P., 1994, ‘‘Electromagnetic separtion fo particles from a conducting
liquid using a strong DC magnetic field.’’ Proceeding of 1st International Symposiumon
Electromagnetic Processing of Materials, Nagoya, Japan, pp. 97–102.
Pucella, G., Samuel, A. M., Samuel, F. H., Doty, H. W., and Valtierra, S., 1998, ‘‘Sludge for-
mation in Sr-modified Al-11.5 Wt% Si diecasting alloys.’’ One Hundred Third Annual
Meeting of the American Foundrymen’s Society, pp. 117–125.
Roy, N., Samuel, A. M., and Samuel, F. H., 1996, ‘‘Porosity formation in Al-9 wt% Si-3 wt%
Cu alloy systems: metallographic observations.’’ Metallurgical and Materials Transactions
A, 27, pp. 415–429.
Sampath, V. and Dwarakadasa, E. S., 1992, ‘‘On the feasibility of the application of electro-
slag refining to an aluminium alloy.’’ Materials Forum, 16, pp. 81–85.
Samuel, F. H., Ouellet, P., Samuel, A. M., and Doty, H. W., 1998, ‘‘Effect of Mg and
Sr additions on the formation of intermetallics in Al-6 wt% Si-3.5 wt% Cu-(0.45) to
(0.8) wt% Fe 319-type alloys.’’ Metallurgical and Materials Transactions A, 29, pp.
2871–2884.
Samuel, F. H., Samuel, A. M., Doty, H. W., and Valtierra, S., 2003, ‘‘Influence of compo-
sition, Sr modification, and annealing treatment on the structure and properties of cast
Al-4 pct Mg alloys.’’ Metallurgical and Materials Transactions A, 34, pp. 115–129.
Schwartz, L. H., 1999, ‘‘Sustainability: The materials role.’’ Metallurgical and Materials
Transactions B, 30B, pp. 157–170.
Shabestari, S. G., 2004, ‘‘The effect of iron and manganese on the formation of intermetallic
compounds in aluminum-silicon alloys.’’ Materials Science and Engineering A, 383,
pp. 289–298.
Shabestari, S. G. and Gruzleski, J. E., 1995, ‘‘Gravity segregation of complex intermetallic
compounds in liquid aluminum-silicon alloys.’’ Metallurgical and Materials Transactions
A, 26, pp. 999–1006.
Shabestari, S. G. and Gruzleski, J. E., 1996, ‘‘Modification of iron intermetallics by strontium in
413 aluminum alloys,’’ Transactions of the American Foundrymen’s Society, Volume 103,
Kansas City, MO: American Foundrymen’s Society, Inc, pp. 285–293.
Shabestari, S. G., Mahmudi, T., Emamy, M., and Campbell, J., 2002, ‘‘Effect of Mn and Sr on
intermetallics in Fe-rich eutectic Al-Si alloy.’’ International Journal of Cast Metals
Research, 15, pp. 17–24.
Shabestari, S. G. and Shahri, F., 2004, ‘‘Influence of modification, solidification conditions
and heat treatment on the microstructure and mechanical properties of A356 aluminum
alloy.’’ Journal of Materials Science, 39, pp. 2023–2032.
Shrikhande, V. J., Balasubramanian, M. S. N., Gupt, K. M., and Rao, P. K., 1986, ‘‘Electro-
slag refining of aluminium alloys.’’ Trans. Indian Inst. Met, 39, pp. 57–59, 61.
Shu, D., Li, T. X., Sun, B. D., Wang, J., and Zhou, Y. H., 1999a, ‘‘Study of electromagnetic
separation of nonmetallic inclusions from aluminum melt.’’ Metallurgical and Materials
Shu, D., Li, T. X., Sun, B. D., Zhou, Y. H., Wang, J., and Xu, Z. M., 2000a, ‘‘Numerical
calculation of the electromagnetic expulsive force upon nonmetallic inclusions in an
aluminum melt. I. Spherical particles.’’ Metallurgical and Materials Transactions B, 31,
pp. 1527–1533.
Shu, D., Li, T. X., Sun, B. D., Zhou, Y. H., Wang, J., and Xu, Z. M., 2000b, ‘‘Numerical
calculation of the electromagnetic expulsive force upon nonmetallic inclusions in an
aluminum melt. II. Cylindrical particles.’’ Metallurgical and Materials Transactions B,
31, pp. 1535–1540.
Shu, D., Sun, B. D., Li, K., Xu, Z., and Zhou, Y. H., 2002, ‘‘Continuous separation of
nonmetallic inclusions from aluminum melt using alternating magnetic field.’’ Materials
Letters, 55, pp. 322–326.
Shu, D., Sun, B. D., Wang, J., Li, T. X., and Zhou, Y. H., 1999b, ‘‘Study of electromagnetic
separation of nonmetallic inclusions from aluminum melt.’’ Metallurgical. Materials
Transactions A, 30A, pp. 2979–2988.
Sigworth, G. K., Shivkumar, S., and Apelian, D., 1989, ‘‘The influence of molten metal
processing on mechanical properties of cast Al–Si–Mg alloys.’’ In 93rd AFS Casting
Congress, San Antonio, TX: American Foundrymen’s Society, pp. 811–824.
156 L. ZHANG ET AL.
Singleton, O. R. and Robinson, G. C. J., 1971, ‘‘Refining aluminum-silicon alloys by centrifu-

ging.’’ Journal of the Institute of Metals, 99, pp. 155–159.
Skjerpe, P., 1987, ‘‘Intermetallic phases formed during DC-casting of an Al–0.25Fe–0.13Si
alloy.’’ Metallurgical Transactions A, 18, pp. 189–200.
Starke, E. A. and Staley, J. T., 1996, ‘‘Application of modern aluminum alloys to aircraft.’’
Progress in Aerospace Sciences, 32, pp. 131–172.
Stefaniay, V., Griger, A., and Turmezey, T., 1987, ‘‘Intermetallic phases in the aluminium-side
corner of the AlFeSi-alloy system.’’ Journal of Materials Science, 22, pp. 539–546.
Suarez-Pena, B. and Asensio-Lozano, J., 2006, ‘‘Influence of Sr modification and Ti grain
refinement on the morphology of Fe-rich precipitates in eutectic Al-Si die cast alloys.’’
Scripta Materialia, 54, pp. 1543–1548.
Takahashi, K. and Taniguchi, S., 2003, ‘‘Electromagnetic separation of nonmetallic inclusion
from liquid metal by imposition of high frequency magnetic field.’’ ISIJ International, 43,
pp. 820–827.
Tanaka, Y., Sassa, K., Iwai, K., and Asai, S., 1995, ‘‘Separation of non-metallic inclusions
from molten metal using traveling magnetic field.’’ Tetsu-to-Hagane, 81, p. 1120.
Tang, S. K. and Sritharan, T., 1998, ‘‘Morphology of beta-AlFeSi intermetallic in Al-7Si alloy
castings.’’ Materials Science and Technology, 14, pp. 738–742.
Taniguchi, S. and Brimacombe, J. K., 1994, ‘‘Application of pinch force to the separation of
inclusion particles from liquid steel.’’ ISIJ International, 34, pp. 722–731.
Taniguchi, S. and Kikuchi, A., 2000, ‘‘Removal of nonmetallic inclusion from liquid metal by
AC-electromagnetic force.’’ Proceedings of the 3rd International Symposiumon Electro-
magnetic Processing of Materials, Nagoya, Japan, p. 315–320.
Taylor, J. A., Schaffer, G. B., and St. John, D. H., 1999a, ‘‘The role of iron in the formation of
porosity in Al-Si-Cu-based casting alloys. I. Initial experimental observations.’’ Metallur-
gical and Materials Transactions A, 30, pp. 1643–1650.
Taylor, J. A., Schaffer, G. B., and St. John, D. H., 1999b, ‘‘The role of iron in the formation of
porosity in Al-Si-Cu-based casting alloys. II. A phase-diagram approach.’’ Metallurgical
and Materials Transactions A, 30A, pp. 1651–1655.
Taylor, J. A., Schaffer, G. B., and St. John, D. H., 1999c, ‘‘The role of iron in the formation of
porosity in Al-Si-Cu-based casting alloys. III. A microstructural model.’’ Metallurgical
and Materials Transactions A, 30, pp. 1657–1662.
Tseng, C. J., Lee, S. L., Tsai, S. C., and Cheng, C. J., 2002, ‘‘Effects of manganese on micro-
structure and mechanical properties of A206 alloys containing iron.’’ Journal of Materials
Research, 17, pp. 2243–2250.
Van Der Donk, H. M., 1998, ‘‘Method for refining an aluminum scrap smelt.’’ US Patent No.
741348.
Van Der Donk, H. M., Nijhof, G. H., and Castelijns, C. A. M., 1995, ‘‘The removal of iron
from molten aluminum.’’ Third International Symposium. Recycling of Metals, and Engi-
neered Materials, Point Clear, AL: Minerals, Metals, and Materials Society, pp. 651–661.
Verte, L. A., 1961, ‘‘Author’s certificate, Byull. Izobret. (in Russian), no. 19.’’ USSR Patent
141592-13c.
Vire, S. and Gauckler, L., 1985, ‘‘Cell for the refining of aluminum.’’ US Patent No. 4552637.
Wang, L., Makhlouf, M., and Apelian, D., 1995, ‘‘Aluminium die casting alloys: Alloy
composition, microstructure, and properties-performance relationships.’’ International
Materials Reviews, 40, pp. 221–238.
Wang, Y. and Xiong, Y., 2000, ‘‘Effects of beryllium in Al-Si-Mg-Ti cast alloy.’’ Materials
Science and Engineering a-Structural Materials Properties Microstructure and Processing,
280, pp. 124–127.
William, D. V. and Bernard, W., 1987, ‘‘Aluminum base alloys of the Al-Cu-Mg-Zn type.’’ US
Patent No. 4711762.
Xu, Z., Li, T., Shu, D., and Zhou, Y., 2001, ‘‘Electromagnetic filtration of primary Fe-rich
phases from Al-Si alloy melt.’’ Cailiao Kexue Yu Jishu, 17, pp. 306–310.
Xu, Z., Li, T., and Zhou, Y., 2003, ‘‘Elimination of Fe in Al-Si cast alloy scrap by electromag-
netic filtration.’’ Journal of Materials Science, 38, pp. 4557–4565.
Yagi, K. and Halada, K., 2001, ‘‘Materials development for a sustainable society.’’ Materials
and Design, 22, pp. 143–146.
Yamao, F., Sassa, K., Iwai, K., and Asai, S., 1997, ‘‘Separation of inclusions in liquid metal
using fixed alternating magnetic field.’’ Tetsu-to-Hagane, 83, pp. 30–35.
Yang, J. B. and Yin, F., 2000, ‘‘Improvement of the morphology of iron phase in Al-Si
alloys.’’ Foundry, 49, pp. 454–459.
Yao, G. C., Sun, Z. Q., and Qiu, Z. X., 2001, ‘‘Electromagnetic separation of iron-rich
phases from molten aluminum.’’ Journal of Northeastern University, Natural Science,
22, pp. 127–129.

Yie, S. N., Lee, S. L., Lin, Y. H., and Lin, J. C., 1999, ‘‘Mechanical properties of Al-11% Si
casting alloys containing trace Be and Sr.’’ Materials Transactions, JIM, 40, pp. 294–300.
Yin, F., Yang, J. B., Wang, Y. X., and Wang, G. X., 2000, ‘‘Effect of Mn and Be on the
iron-phase in Al-Si alloys.’’ Foundry, 49, pp. 829–831.
Yoo, J. E., Shan, A., and Moon, I. G., 1999, ‘‘A study on composition and crystal structure of
dispersoids in AlMgSi alloys.’’ Journal of Materials Science, 34, pp. 2679–2683.
Yoon, E.-P., Choi, J.-P., Kim, J.-H., Nam, T.-W., and Kitaoka, S., 2002a, ‘‘Continuous elim-
ination of Al2O3 particles in molten aluminium using electromagnetic force.’’ Materials
Science and Technology, 18, pp. 1027–1035.
Yoon, E.-P., Kim, J.-H., Choi, J.-P., and Kwon, H.-R., 2002b, ‘‘Effects of electromagnetic
force on the removal of alumina particles in molten A356 aluminum alloy.’’ Journal of
Materials Science Letters, 21, pp. 739–742.
Young, R. M. K. and Clyne, T. W., 1981, ‘‘An Al–Fe intermetallic phase formed during
controlled solidification.’’ Scr. Metall, 15, pp. 1211–1216.
Zhang, H., 2007, ‘‘Effect of iron on the microstructure and mechanical properties of Al-Si
alloys.’’ Light Metals, pp. 775–780.
Zhang, L., Jiao, W.-L., Wei, H., and Yao, G.-C., 2005a, ‘‘Effects of the shape of Fe enrich-
ment phase on purification of Liquid Al in Electromagnetic Field.’’ Special Casting &
Nonferrous Alloys (Chinese), 25, pp. 131–132.
Zhang, L., Jiao, W.-L., Wei, H., and Yao, G.-C., 2006a, ‘‘Electromagnetic seperation of
magnetic purity particles in aluminum melt.’’ Materials Science & Technology (China),
14, pp. 524–526.
Zhang, J., Shu, D., Rao, Q.-L., Sun, B.-D., and Chen, G., 2006b, ‘‘Nucleation and growth of
high purity aluminum grains in directional solidification bulk sample without electromag-
netic stirring.’’ Transactions of the Nonferrous Metals Society of China, 16, pp. 1–7.
Zhang, J., Sun, B., He, B., Mao, H., Chen, G., and Ge, A., 2006c, ‘‘Principle and control of
new-style purification equipment of 5 N high purity aluminum.’’ Chinese Journal of Mech-
anical Engineering, 42, pp. 64–68.
Zhang, J., Sun, B.-D., Rao, Q.-L., Chen, G., and Ge, A.-J., 2005b, ‘‘Growth of high purity
aluminum grains during directional solidification.’’ Shanghai Jiaotong Daxue Xuebao,
39, pp. 1787–1791.
Zhang, L.-L., Xiao, Y.-D., Zhou, J., and Li, W.-X., 2006d, ‘‘Effects of be on microstructures
and mechanical properties of as-cast Al-Fe-V-Si alloys.’’ Zhuzao, 55, pp. 65–68.
Zhao, H. and Lu, H., 2008, ‘‘The development of 85ka three-layer electrolysis cell for refining
of aluminium.’’ Light Metals, pp. 533–535.
Zheng, J. G., Vincent, R., and Steeds, J. W., 2000, ‘‘Crystal structure of an orthorhombic
phase in beta-(Al-Fe-Si) precipitates determined by convergent-beam electron diffrac-
tion.’’ Philosophical Magazine A, 80, pp. 493–500.
View publication stats

6RemovalofIronFromAluminum Review

Uploaded by

Copyright:

Available Formats

6RemovalofIronFromAluminum Review

Uploaded by

Document Information

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

6RemovalofIronFromAluminum Review

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Removal of Iron From Aluminum: A Review

Article · March 2012

The user has requested enhancement of the downloaded file.

Mineral Processing and Extractive

Removal of Iron From Aluminum: A

Available online: 01 Jun 2011

To link to this article: http://dx.doi.org/10.1080/08827508.2010.542211

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

REMOVAL OF IRON FROM ALUMINUM: A REVIEW

Lifeng Zhang1,2, Jianwei Gao2,3, Lucas Nana

and Technology Beijing, Beijing, China

Keywords: alloys, aluminum, iron removal, modiﬁcation

Address correspondence to Lifeng Zhang, Professor, School of Metallurgical and Ecological

Wrought 1 97.1 0.11 0.59 0.82 0.21 0.51 0.45 0.19

However, in high-integrity die casting variations such as high-vacuum, squeeze,

. Iron-containing intermetallic phases in aluminum alloys

Fe-RICH PHASES IN ALUMINUM AND ALUMINUM-SILICON ALLOYS

Table 2. Identiﬁed Fe-rich phases in aluminum alloys

Fe-rich intermetallics Crystal structure Reference

AlmFe Bct (Liu, Thorvaldsson, and Dunlop 1986; Khalifa,

C-centered 2003; Shabestari 2004)

a-Al12Fe3Si2 (Mondolfo 1976), a-Al15Fe3Si2 (Crepeau 1995) or generally a-AlFeSi

Figure 7 Three-dimensional reconstruction of b-phase, (a) original two-dimensional phases; (b)

EFFECT OF IRON ON THE MECHANICAL PROPERTIES OF ALUMINUM

Table 3. Effect of Fe on Al-Cu alloys

UTS (Sampath YS (Sampath Elongation Hardness Fatigue

Cu, 5 0.25 370–395 NA 12–13 NA NA Grand

Table 4. Effect of Fe on Al-Mg alloys

Tensile (Sampath Elongation Corrosion Hardness Fatigue

Mg, 5 or 10 NA – – NA – NA NA Kato et al.

content increased to 0.6 wt% in Al-10Mg-0.5Si alloy. In Al-3 wt% Mg or 6 wt% Mg

Iron was also added as anti-recrystallizer to Al-Zn-Mg-Cu alloys (Fridlyander,

Table 5. Effect of Fe on Al-Zn alloys

Elongation Tensile (Sampath and Hardness

Zn, Mg Chill cast 0.36 14 þ þ Bonsack

Figure 9 Al corner of ternary Al-Fe-Si system diagram.

theory, suggesting that b-Al5FeSi increases porosity by blocking feeding channels

MODIFICATION OF PLATELET IRON PHASES WITH ADDITION OF

Transition metal Atomic number Bravais lattice Atomic radius, Å

Al-Si-Mn alloy, a-Al15Mn3Si2 forms. In quaternary Al-Si-Fe-Mn alloys, iron may

centration was reported to encourage the formation of primary Si as well as the a-

small amount of Be addition results in the formation of stable Chinese script

Figure 19 Microstructure of 5xxx alloys (0.42Si-0.46Fe-0.044Cu-0.291 Mn-3.47 Mg-0.003Cr) containing

The addition of beryllium changes the solidiﬁcation sequence of Fe-rich

Table 7. Mechanical properties of stirred cast and extruded

Alloy (wt%) Elongation, % ET, J

0.4Fe 15.0  0.5 14.5  1

Table 8. Mechanical properties of sand-cast Al-Si-Mg alloys with trace Be addition

Without Be addition With Be addition

0.2 wt% Fe 1.0 wt% Fe 0.1 wt% Fe 0.94 wt% Fe

Ultimate tensile strength (Fukuda 270  3 210  4 260  3 260  3

(Mulazimoglu, Gruzleski, and Closset 1992). The modiﬁcation of b-AlFeSi was

Molybdenum (Mo), Nickel (Ni), and Sulfur (S)

MODIFICATION OF PLATELET IRON PHASES WITH THERMAL

silicon eutectic temperature. The chemical composition of the remaining liquid

temperature is required because of the low diffusivity of Fe in the solid aluminum.

IRON REMOVAL TECHNOLOGIES

(Cao and Campbell 2004b; de Moraes et al. 2006).

0.4Fe 15.0 0.5 14.5 1

Ultimate tensile strength (Fukuda 270 3 210 4 260 3 260 3