Академический Документы
Профессиональный Документы
Культура Документы
review
S. Lun Sin, A. Elsayed and C. Ravindran*
High volume application of lightweight materials is the key to improving fuel efficiency and vehicle
performance, and decreasing exhaust emissions to address environmental concerns. Currently,
magnesium alloys, which are the lightest structural materials, represent only y0?3% of an
automobile weight. One of the most important parameters in controlling the properties of
magnesium and its alloys is melt cleanliness, manifested mainly in inclusions. The presence of
such inclusions will strongly influence the mechanical properties and corrosion resistance of
structural components. This paper gives an overview of the current state of knowledge pertaining
to magnesium melt cleanliness. It describes the nature and origin of the inclusions, the methods
for assessing the cleanliness of magnesium, the methods used to control inclusions in the melt
and the relationship between melt cleanliness and the properties of magnesium castings.
Keywords: Magnesium alloys, Inclusions, Melt cleanliness assessment, Melt cleanliness control, Mechanical properties, Fluidity, Corrosion resistance
manganese to molten Mg. The melt is then cooled, melting point are 0?65 cm3/100 g and 0?034 cm3/100 g
resulting in the precipitation of Fe–Mn–Al intermetallic. respectively. For Mg, they are 27 cm3/100 g and 19 cm3/
Control of melt temperature is thus crucial. 100 g respectively.17 The solubility of hydrogen in
molten Mg is significantly higher than in Al. However,
Hydrogen molten Mg saturated with hydrogen rejects only 32% of
Hydrogen solubility is greater in the liquid than in the hydrogen during solidification as compared to alumi-
solid state for Al and Mg alloys. For Al, the maximum nium that rejects 95%.14 If the concentration of
solubility of hydrogen in the liquid and the solid near the hydrogen is above the maximum solid solubility,
A380-F 2.76 324 160 4.0 227 145 10.0 71.0 0.08
319.0-T6 2.79 248 165 2.0 N/A N/A N/A 74.0 N/A
332.0-T5 2.76 248 193 1.0 214 165 2.0 77.0 N/A
Al 356.0-T6 2.68 262 186 5.0 145 117 10.0 72.4 N/A
A356.0-T61 2.69 283 207 10.0 145 117 20.0 72.4 N/A
383.0-F 2.74 310 150 3.5 N/A N/A N/A 71.0 N/A
390.0-T5 2.73 296 265 1.0 N/A N/A N/A 81.2 N/A
AZ91D 1.81 250 160 7.0 169 105 16.0 45.0 2.7
AM60B 1.80 240 130 13 N/A N/A N/A 45.0 N/A
AM50A 1.77 230 125 15 N/A N/A N/A 45.0 N/A
Mg AM20 1.75 210 90 20 N/A N/A N/A 45.0 N/A
AS41 1.77 240 140 15 153 94 17 45.0 0.05
AS21 1.76 220 120 13 130 87 20 45.0 0.19
AE42 1.79 230 145 11 140 88 23 45.0 0.06
*All die cast, except 319?0-T6, 332?0-T5, 356?0-T6 and A356?0-T61 (permanent mould).
N/A: none available.
formation of pores can occur upon solidification.2 This Mg melts and covers their assessment, control and
represents a serious problem as porosity affects the effects on properties of the alloys.
mechanical properties and the surface quality of
castings. The acceptable hydrogen levels for Al and Inclusions in magnesium alloys
Mg alloys are 0?1 cm3/100 g or lower18 and 15–20 cm3/
100 g,4,19 respectively. The level of hydrogen can be Characterisation of inclusions in magnesium
reduced by bubbling chlorine or an inert gas through the alloys
melt.12,20 Inclusions present in molten Mg can be categorised into
There is no commercial instrument available to two major groups:31
measure hydrogen content in Mg melts. Researchers (i) non-metallic inclusions: these include oxides and
have utilised different techniques (e.g. helium bubbling, nitrides; sodium, magnesium and potassium
vacuum extraction, complete combustion of a Mg based chlorides; aluminium and calcium based
sample)21 to measure hydrogen concentrations, some- carbides, magnesium based sulphides (MgS),
times making the results difficult to compare and fluorides (MgF2) and sulphates (MgSO4).
unreliable.22 Oxides are the most predominant non-metallic
inclusions, followed by nitrides
Inclusions (ii) intermetallic inclusions: these include iron rich
Unlike Al alloys, Mg alloys form a loose, permeable film intermetallic phases, which precipitate during
which does not protect the molten metal from further iron removal. Almost all intermetallic inclusions
oxidation. Consequently, molten magnesium will oxidise contain iron.
and burn rapidly in contact with air if special precau- Both inclusion types are potentially harmful to the
tions are not taken (e.g. use of flux, cover gas or alloying properties of Mg alloys, in particular tensile strength,
additions such as beryllium).1 The high oxidation elongation and corrosion resistance. Further, descrip-
potential of Mg usually results in inclusion contents tion of the influence of inclusions on the properties of
10–20 times higher than that of Al (0?1–10 ppm for pure Mg alloys is discussed in the section on ‘Effect of
Al, 10–200 ppm for pure Mg).23 Inclusions in Mg alloys inclusions on properties of magnesium alloys’. The
reduce mechanical properties, are detrimental to surface characteristics of the various inclusions found in Mg
finish, increase porosity and exhibit a tendency to alloys are summarised in Table 4, and the typical
increase corrosion.24,25 inclusions are shown in Fig. 2.
The maximum allowable concentrations of the major
impurities in Mg alloys and the methods to remove them Origin of inclusions in magnesium alloys
are summarised in Table 3. Reaction with air
There is a need to understand and control all aspects During the melting process, molten Mg reacts with
of metal handling and processing to improve the oxygen to form magnesium oxide (MgO) when the melt
consistency of quality and properties of the cast product. is exposed to the atmosphere.31,35 Molten Mg can also
This also applies to Mg recycling, for which the large react with the moisture in air to form MgO and
amount of scrap runners, overflows, flash and rejected hydrogen, which can result in a fire or explosion.
castings from Mg die casting processes requires careful Therefore, all tools to be used must be dry/preheated
control of melt practices to ensure metal quality.29,30 before immersion in the melt. Water should not be used
Many efforts have been devoted to improve and assess to fight magnesium fires.1 The casting process, depend-
the melt cleanliness of Al alloys. In comparison, limited ing on the pouring technique and gating design, almost
information is available for Mg alloys. The techniques always leads to some melt turbulence, which entraps
used for Al were generally applied to Mg. However, due inclusions and exposes a fresh melt surface to the
to the reactive nature of Mg, the adaptability and atmosphere, further increasing the number of inclusions
applicability of these techniques to Mg melts can be in the melt. Stirring, charging, ladling, poor venting,
challenging. This review paper is written with a focus on removal of dross and sludge and excessive movement of
the characterisation and formation of inclusions within the mould before solidification is complete12 also
Table 3 Impurity levels in Mg alloys
Impurity Typical level Acceptable levels Methods for achieving acceptable limits
Dissolved elements
Fe 200–400 ppm14 0.0035–0.004 ppm26 Addition of Mn and Al additions to form
Al–Fe–Mn phases14
Ni Solubility in Mg at 650uC: 32 wt-%1 0.001–0.01 ppm26 Dilution, addition of Al and Zr to reduce
Cu solubility in Mg and promote formation
of AlNi precipitates14
Cu Solubility in Mg at 650uC: 70 wt-%1 0.008–0.25 ppm26 Dilution, removed alongside Fe in Al–Fe–Mn
phase formation14
Hydrogen In pure Mg: up to 24 cm3/100 g 15–20 cm3/100 g4,19 Gas bubbling (chlorine or inert gas) through
of metal21 the melt12,18
In AZ91 alloy (between 650 and
750uC): 6–14 cm3/100 g27
Inclusions Variable (depends on processing Less than 500 ppm28 Addition of flux or fluxless refining (protective
conditions) atmospheres, filtration, inert gas bubbling,
melt sedimentation, ultrasonic purification,
electromagnetic separation)
2 Optical micrographs of a an oxide cluster in AZ91 alloy, b a ‘snaky’ oxide in AM50 alloy and c Mn–Al particles in
AM60 alloy34
promote the entrapment of oxides. Oxide defects exhibit nitrogen. Magnesium nitride will not form unless the
various morphologies, from particles to films. The oxide partial pressure of oxygen becomes extremely low.35
films can measure up to 50 mm in length with a thickness There is no indication that nitride inclusions appear
of 0?5 mm. The oxide particles are commonly seen as independently of oxide inclusions.
clusters less than 50 mm in diameter. Both oxide films
and clusters have high surface to volume ratios and Reactions with fluxes
cannot be removed by melt settling.35 Fluxes are used to protect magnesium melts from
Magnesium nitrides (Mg3N2), which form as a result oxidation and for refining (removal of inclusions).
of the reaction of Mg with nitrogen, may appear Flux inclusions (e.g. MgCl2, CaCl2) occur if unabsorbed
together with oxide clusters. As can be seen in Table 5, protective flux remains on the metal surface, the flux is
the Gibbs free energy associated with the oxidation of not viscous enough, there is brittle or powdery flux due
magnesium is more negative than the Gibbs free energy to long handling, pouring is too fast or if there is
associated with the formation of magnesium nitride, incomplete removal of the flux adhering to the lip of the
meaning that Mg preferentially reacts with oxygen over pot before pouring.38 Magnesium chloride in the flux
Oxides
MgO Particles 10–300 3.58
Films 0.5–1 (thickness)
50–400 (length)
MgO.Al2O3 (spinel) Particles 3.58
Nitrides
Mg3N2 Particles 10–300 2.71
Films 0.5–1 (thickness)
50–400 (length)
Carbides
Al4C3 Particles 0.1–10 2.36
CaC2 Particles 2–20 2.22
Chlorides and salts
MgCl2 2.32
NaCl Particles 10–50 2.17
CaCl2 2.15
KCl 1.98
Other non-metallic inclusions
Fluorides (MgF2) 3.15
Sulphides (MgS) Particles 2.68
Sulphates (MgSO4) 2.66
Borides (FeB) 7.15
Iron rich intermetallic compound
Al8(Mn,Fe)5, a-AlMnFe, (Mn,Fe)5Si3, Particles, needles ,20 4–7
Al8(Mn,Fe)4RE, a-Fe, Fe2(Si,B),
Fe3(Al,Si), (Fe,Mn)3Si
Table 5 Standard Gibbs free energy DGreaction as function of temperature T of selected reactions36,37
can react with oxygen and water in the air to form MgO, Table 6, but none of these methods is considered as a
which can become entrapped in the casting.39 Similarly, standard for determining Mg melt cleanliness.
the use of boride containing fluxes could lead to the
formation of FeB inclusions.40 Considerations related to Metallographic techniques
flux entrapment and melt loss41 led to a shift to the use Classic metallographic techniques
of protective atmospheres.42 Classic metallographic techniques involve the examina-
Reactions with protective gases tion of ingot slices to determine the presence of
inclusions. It allows the determination of inclusion
Protective atmospheres43–52 are used to prevent oxida-
amount, size, shape and distribution. Metallographic
tion or burning of molten Mg and consequently to
analysis can be combined with image analysis to
reduce the formation of inclusions in the melt. Some
determine the particle size, number of particles, percen-
cover gases, including sulphur dioxide (SO2), sulphur
tage of oxide in the sample, or with scanning electron
hexafluoride (SF6), hydrofluorocarbons (HFC) and
microscopy (SEM) and energy dispersive X-ray spectro-
fluorinated ketones (FK), modify the natural oxide film
scopy (EDS) to determine the nature of the inclusions
such that Mg vaporisation is suppressed and reactive
and the possible source of melt contamination.56
gases are excluded. For example, sulphur dioxide reacts
However, optical metallography requires multiple steps:
with Mg to form MgSO4, MgO or MgS.46 With SF6,
sectioning, grinding, polishing and examination. The
HFC and FK, the reaction products are MgO and
technique is therefore time consuming and expensive.
MgF2.43–45,47–52 All these compounds can enter the melt
Also, the sample size is relatively small, possibly leading
and lead to the formation of inclusions. Although
to results which are non-representative of the sample.
protective gases may lead to the formation of inclusions,
their absence would be more detrimental to the proper- Humidity test
ties of Mg alloys. Proper melting, handling and pouring The humidity test is used in the industry to detect and
procedures should minimise their formation. identify flux inclusions in Mg alloys. The surface to be
Reactions during melt treatment and alloying examined is carefully ground and polished. The pre-
Inclusions can be introduced during the process of pared specimens are then placed in a chamber where
iron removal using manganese. With the addition of humidity conditions are controlled (e.g. 95% relative
manganese, intermetallic particles containing iron, humidity at 38uC1). Corrosion occurs at the site of
manganese and aluminium form.31 Inclusions can also certain inclusions, others being relatively unaffected.
result from degassing and grain refinement by the addi- The corrosion products are then examined by micro-
tion of hexachloroethane (C2Cl6), calcium cyanamide chemical techniques to identify the inclusions.22
(CaNCN) or by the addition of alloying elements (e.g. However, the results are available only after sufficient
Al4C3 in Mg alloyed with Al).31 corrosion on the sample has occurred which can
typically take 24–48 h. Fracture ingot surfaces are
Reactions during casting sometimes exposed to high humidity in a humidity
During casting, poor running system design is generally chamber. Salt inclusions pick up moisture and bloom,
associated with high surface turbulence and greater indicating the presence of fluorides and chlorides.75
entrainment of oxide films. The entrained oxides are
typically double oxide film defects (bifilms) comprised of Techniques based on filtration of liquid metal
folded MgO films.53,54 The high reactivity of Mg results Hydro magnesium inclusion assessment method (HMIAM)
in rapid formation of MgO films during pouring that In the hydro magnesium inclusion assessment method
quickly thickens to produce a tube of oxide that (HMIAM) or vacuum filtration technique, a known
surrounds the falling metal stream. This tube of oxide volume of melt is drawn through a stainless steel filter by
is not in danger of becoming entrapped within the liquid vacuum. The inclusions are retained on top or within the
metal if the oxide film is unbroken. A similar phenom- filter (Fig. 3). Upon solidification and cooling, the filter
enon occurs in Al alloys and is due to the same and its contents are sectioned parallel to the flow
mechanism of rapid thickening of oxide films.18 The direction and examined using an optical microscope.
oxide film forming at the surface of the Al melt is The amount of inclusions is determined in terms of
relatively thin and elastic as compared to the thick and volume of particles per unit weight of metal drawn
cracked oxide film forming at the surface of the Mg through the filter. The major advantage of HMIAM is
melt. As a result, Mg alloys may be more prone to bifilm that sampling is performed directly in the melt.31,35
formation than Al alloys. However, this method can be considered too time
Inclusions can also be generated during die casting consuming.57 The quantity and morphology of oxides
from oxidation in the shot sleeve, entrapment of air in ingots or die cast components may also be different
during casting and absorption of lubricants spread on from those observed on the filter due to different
the die surface.1,11,14,31 In sand casting, turbulent flow solidification conditions (e.g. melt temperature).58
may cause sand grains to detach from the mould wall
and be entrained within the Mg melt, with which they Pressure filtration technique: PoDFA and Prefil
react. The inclusions formed are called ‘reacted sand Both PoDFA (porous disc filtration analysis) and Prefil
inclusions’.14,31,55 techniques involve forcing a known amount of molten
metal under pressure through a filter. In the PoDFA
Assessment of cleanliness of method (Fig. 4), the inclusions concentrated inside or on
top of the filter are examined metallographically in terms
magnesium melts of size, colour and morphology.31,35 In the Prefil test, the
A summary of the methods used or considered to assess flow rate of molten metal through the filter is measured
the metal cleanliness in the Mg industry is shown in and used as an indication of the melt quality: clean
metals will flow quickly, whereas inclusions in molten can be subsequently analysed by optical microscopy.60
melts build up on the filter surface and reduce the The technique is still under development to enable the
flowrate. These two methods have been applied to monitoring of the filtration rate.
measure the melt quality of Al melts, but it was only
recently that they were adapted to Mg melts.59 Techniques based on melt centrifugation and
sedimentation
Self-gravitation filtering technique This technique is based on the separation of phases with
In the self-gravitation filtering technique, specifically different densities. The samples are heated to 720–
designed for Mg alloys, a filter tube containing two 800uC, held isothermally for 5 min and then centrifuged
sintered metal fibre filter discs is immersed in the melt, at high speed (200–2500 rev min21) until the melt
allowed to be preheated for 3–4 min and then lifted up solidifies. The temperature is measured continuously
out of the melt at a constant speed. As the tube rises, the with a thermocouple located below the crucible.61
metal flows through the filters by self-gravitation. The During this period, inclusions are sedimented at the
filter cake thickness is correlated to the inclusion bottom of the crucible. The inclusions are subsequently
content, thereby obtaining an estimation of the melt analysed using optical microscopy. Experimental results
cleanliness. The inclusions collected on the filter discs showed that (Fe,Mn)3Si particles present in Mg melt
Table 6 Methods for metal cleanliness assessment used or considered for use in magnesium industry*
Method Response time Information Detection limits Sample size Cost References
7 a LIBS mapping of oxide inclusions in magnesium samples and b corresponding optical micrograph65
current. Non-conducting inclusions in the melt are non-reactive with Mg alloys; and (6) the cost is very
detected by measurement of the change in the electrical high.31
conductivity as they pass through the aperture.31,35,67,68
The LiMCA technique can be made available online LECO method
very quickly at time intervals on the order of 1 min and The LECO method (oxygen and nitrogen determinator)
is able to detect inclusions as small as 20 mm. However, may be applicable for detecting total oxygen and nitrogen
this method presents several disadvantages: (1) it cannot in solid Mg samples. The method is based on fusion of a
detect conductive inclusions such as intermetallic solid metal specimen in a graphite crucible. At high
particles, which are present in molten Mg; (2) the orifice temperatures (up to 2300uC), magnesium oxide will be
can be blocked by large size inclusions; (3) the amount reduced to magnesium and carbon monoxide (CO) in the
of molten metal that the LiMCA can assess at a time is presence of carbon. The amount of CO developed can be
very limited; (4) it fails to provide information on the quantified and related to the original MgO content.69 The
chemistry, shape or the physical state of the inclusions; detection limit is claimed to be 0?1 ppm for both oxygen
(5) there are difficulties in finding appropriate non- and nitrogen.35 Large inclusions may affect the results, as
conductive materials for sampling tubes that are also the sample size is only about 1 g.
Fluxless refining processes are based on sedimentation Environmental concerns have thus prompted the Mg
(settling), floatation, interception or electromagnetic industry to seek alternatives to SF6. Sulphur dioxide
forces. SO2 (approximately 1?5% in air) was demonstrated
to adequately protect molten magnesium. However,
Protective atmospheres SO2 is a toxic gas causing corrosion of foundry
Protective atmospheres43–52 are used to prevent oxida- equipment.92 Experiments also proved that pure CO2
tion or burning of molten Mg and consequently to effectively protects molten Mg, as long as the CO2
reduce the formation of inclusions in the melt. There are atmosphere is not contaminated with substantial
two types of protective gas atmospheres:88 non-reactive quantities of air.92
gases and oxide film modifiers. Other potential alternatives include
Non-reactive gases, such as nitrogen and argon, HFCs,45,48,50,52,95,96 FKs,49 boron trifluoride BF397 and
prevent burning. However, use of such gases is solid CO2.98,99 Hydrofluorocarbons, which include
impractical outside of the laboratory as they do not HFC-134a (1,1,1,2-tetrafluoroethane C2H2F4),48,50
52
suppress Mg vaporisation which can result in melt loss. HFC-152a (1,1-difluoroethane C2H4F2) and HFC-
In addition, as Mg vaporises, it condenses into 125 (pentafluoroethane C2HF5),96 have global warming
pyrophoric size particles on the cooler surfaces of the potentials 8?5–170 times lower than SF6. The protec-
furnace.39 tion mechanism is similar to that of SF6 with the
The other method of melt protection involves the use of formation of MgF2 and MgO.45,95 Boron trifluoride
a gas which modifies the magnesium oxide film forming (BF3) is not a greenhouse gas and is well known for its
naturally on the surface of the melt, thereby suppressing excellent protective behaviour. However, BF3 is highly
vaporisation. Since the 1970s, the Mg industry has largely toxic and expensive. Boron trifluoride must also be
used sulphur hexafluoride (SF6) to protect Mg stored as a concentrated, highly compressed gas and,
melts.39,43,44,51,88–93 Sulphur hexafluoride is an effective as a result, special storage conditions are required to
cover gas due to its ability to form a dense film containing minimise explosion hazard. The MagShield system was
Mg oxide (MgO) and Mg fluoride (MgF2) on the molten developed to enable safe distribution of BF3.
Mg surface. This film prevents further oxidation and Specifically, the gas was produced in situ by the
evaporation of Mg.43,44,93 Sulphur hexafluoride is also controlled decomposition of KBF4.97 Fluorinated
attractive as a cover gas since it is odourless, colourless, ketones49 have global warming impact similar to that
non-toxic, non-flammable and non-corrosive. It has been of CO2. They thermally decompose on the molten Mg
found that a combination of SF6 and CO2 is better than surface, producing MgF2 and CO2. They have pro-
SF6 alone for pure Mg and most Mg alloys. The optimum vided protection at significantly lower concentrations
SF6 concentration is 0?1–0?2 vol.-% as an excess con- than SF6. The use of CO2 ice pellets to protect
centration of SF6 will cause severe crucible corrosion.39 magnesium was accomplished by injecting CO2 into
However, SF6 is becoming increasingly expensive and is the furnace chamber at high pressure by a specially
also an extremely powerful greenhouse gas, with a 100- designed nozzle. The CO2 pellets precipitate at the
year global warming potential, estimated at 23 900 times molten metal surface and reduce the tendency of
that of carbon dioxide (CO2). In addition, SF6 is magnesium to evaporate. Sublimated CO2 snow also
chemically stable in the atmosphere and remains present causes the CO2 gas to expand and displace all oxygen
for over 3200 years.94 from the bath surface area.98,99
gas bubbling to remove inclusions depends on bubbling preliminary studies show much promise. Shao et al.115
time, gas flowrate, bubble size and melt temperature.111 also examined the effect of ultrasonic power andparticle
Typical usage of Ar gas sparging was a flowrate of size on the efficiency of ultrasonic separation of
1?8 L min21 at a melt temperature of 740uC for a inclusions. Their results revealed that there is anopti-
bubbling time of 30 min.111 With increasing bubbling mum ultrasonic power above which no further improve-
time and/or flowrate, the rate of inclusion removal and ment was observed. Also, at a given ultrasonic power,
hence, mechanical properties improved. Excess gas smaller particles take more time to agglomerate.
bubbling resulted in additional porosity in the melt that If the ultrasonic treatment is based on the formation
reduced strength and elongation. Argon bubbling was of cavitation bubbles, high energy ultrasonic waves
effective in removing inclusions less than 80 mm, but was disintegrate large inclusion particles into finer, less
unable to remove inclusions larger than 800 mm in size.41 detrimental particles more homogeneously dispersed in
The limited availability of quantitative results on the the Mg melts.116
efficiency of gas bubbling has sparked the need to
determine the optimal processing variables (e.g., argon Electromagnetic separation
flowrate, residence time, bubble diameter, melt tempera- Electronic separation implies flotation of non-conduct-
ture) to enable efficient melt refining. ing inclusions in molten metal by electromagnetic force
A combination of Ar gas sparging at 708 L h21 and a and their subsequent removal. This technique has been
stainless steel filter with openings of 1?14 mm (smaller considered as a new technique to produce Mg melts free
openings were more prone to plugging and choking of from inclusions.117,118 Numerical simulations of steady
metal) on recycled AZ91D Mg alloy filter has been state laminar flow of Mg in a vertical rectangular channel
shown to reduce the number of non-metallic inclusions/ have shown that inclusion removal efficiency increases as
cm2 from 124 to 65 for the unfiltered and the filtered and current density increases and channel size decreases.
bubbled melts respectively.41 A similar result was However, the separation ability of the field diminishes as
observed with the simultaneous use of a 20 pores per the velocity of the melt increases and further research is
inch (ppi) Al2O3 based ceramic foam filter and Ar required to optimise the process.117,118
bubbling by Wu et al.112. Argon was bubbled into the
molten Mg using a rotary impeller and the optimum
Alloying additions
bubbling conditions were determined to be 2 L h21 When Mg is melted in air, the oxide film that forms by
flowrate for 30 min at a melt temperature of 740uC. reaction with the air is porous and unable to prevent
catastrophic oxidation. Calcium and beryllium improve
Degassing the protective quality of the reaction film,39,119 or
Melt degassing utilises typically reactive chlorine (Cl2) increase the ignition temperature of the alloy Mg.
gas or hexachloroethane (C2Cl6) to reduce hydrogen gas Additions of CaO have resulted in a significant
content within melts. The addition of C2Cl6 has the improvement in the oxidation and ignition resistance
added advantage of grain refining melts, but causes of Mg melts to the point where no protective gas was
environmental problems due to the emission of chlori- required during casting.120,121 Although no melt cleanli-
nated hydrocarbons.113 Both methods benefit from the ness analysis was conducted, it is expected that a
formation of MgCl2 (common flux component), which reduction of the oxidation potential could result in
can remove inclusions as discussed previously, but may lower inclusion content in the melt.
become an inclusion itself if it becomes entrapped within
the melt during pouring. Typical usage of Cl2 gas Effect of inclusions on properties of
bubbling involves heating the melt to 735–750uC and
bubbling for 5–15 min.12
magnesium alloys
There has been a lack of data on the effects of non-
Melt sedimentation (settling) metallic inclusions on die cast Mg.122 In this section, the
Various high density (4–7 g cm23) Fe, Si and P based data available are reviewed.
intermetallics can be removed from Mg by allowing
them sufficient time to settle in the melt.35 The required Tensile and fatigue properties
settling time depends on the size of the crucible, It is known that inclusions lead to premature failure of
intermetallic particles and the type of holding furnace the material by reducing the effective cross-section of
being utilised. metal under load and by the concentration of stresses at
the inclusion.
Ultrasonic purification The direct impact of inclusions on the mechanical
Ultrasound can be used to treat Mg alloy melt with the properties is, however, difficult to quantify, since casting
aim to enhance and accelerate the separation of defects such as cold shuts, hot tears and porosity will
inclusions from the melt. During ultrasonic treatment normally be present alongside inclusions and dominate
based on acoustic radiation forces, a force is exerted on over inclusions in determining the properties of a cast
the inclusions to promote their agglomeration and part.28
settling. Shao et al.114 found that when ultrasound was This is illustrated in the work of Chen et al.116 and
applied, the inclusion fraction at the top layer decreased Yim et al.111 Chen et al.116 studied the effect of
sharply from 80?6% to 3?3%, while it increases from ultrasonic treatment on the microstructural and
16?5% to 86?6% at the bottom layer. The result indicated mechanical properties of AZ91 and AM60 alloys. It
that most of the inclusions suspended in the melt would was found that the application of ultrasound to the melt
ultimately sink into the bottom layer with the help of led to the fragmentation of oxide films. Inclusion
ultrasonic agglomeration.114 The use of ultrasound to particles became smaller, more numerous and better
reduce melt inclusions is still in its infancy, but distributed in the melt. Smaller inclusion particles were
in turn led to a significant increase in the ultimate tensile observed that the corrosion rate increased with increas-
strength and elongation. Wang et al.82 also studied the ing average volume fraction of inclusions, as shown in
effect of MgO ceramic foam filters on the properties of Table 9. This was attributed to the difference of
GW103K. As compared to the untreated samples, the electronegativity between the inclusions and Mg matrix.
mechanical properties were higher but remained inferior The inclusions act as cathodes and form galvanic
to those obtained by using a flux, as the filter only coupling with the matrix, leading to pitting corrosion.
removed large inclusions. In another study on Mg–Gd– With increasing inclusion content, the cathodic area
Y–Zr alloy, the YS, ultimate tensile strength and increases, thereby adversely affecting the corrosion
elongation were improved from 156 MPa, 200 MPa resistance of the alloy.128
and 3?4% to 167 MPa, 232 MPa and 7?0% respectively
with the use of filters.108 A similar study was performed Machinability
on AZ91 alloy.30 Again, it was found that the use of flux Haerle et al.122 studied the effect of non-metallic
improved the mechanical properties of the alloy. The inclusions on the machinability of AZ91D alloy. They
mechanical properties were further enhanced by com- observed that a larger volume of inclusions contributed
bining the use of flux with ceramic foam filtration. to a greater degree of wear of the cutting tools. As a
It should be noted that small oxide inclusions play a result, the force required to cut AZ91D increased. The
minor role in influencing tensile properties of castings. machinability of AZ91D was thus reduced.
In a study by Wang et al.,123 the results indicate that
oxide contents up to 1000 ppm levels have apparently
no effect on tensile properties when most oxides are Conclusions and perspectives
less than 15 mm and oxide films are less than 50 mm in Types of inclusions and their control
size. Limited information on refining of Mg alloys is
It is also accepted that fatigue life decreases sig- available, as compared to Al alloys. Inclusions present
nificantly with inclusions size. In AZ91E alloys, in molten Mg can be categorised into two major groups:
Horstemeyer et al.125 have shown that if the inclusion non-metallic inclusions (mainly oxides and chlorides)
size is reduced from several millimetres to several and intermetallic inclusions (iron rich phases). These
hundred microns, the fatigue life in the high cycle fatigue inclusions arise from various sources, which include the
regime can be increased by two orders of magnitude. charge (quality of metals and alloys), melting process,
Fluidity reaction with fluxes and protective gases (during
melting, transfer and pouring processes) and reactions
Inclusions can increase the viscosity of the melt, thereby in the mould. Several methods are available to control
impairing its fluidity. Also, if oxides are present, they inclusions in Mg alloys.
have a tendency to form as films and these can block
(i) fluxes, consisting of MgCl2 and other chlorides
feeding channels and result in incomplete filling.126
(e.g. KCl, NaCl), protect Mg melts from oxida-
Bakke et al.28 found that the flow length (measured
tion. However, typical fluxes used for the
using a meander die) of AM50 alloy was decreased by
refining of Mg alloys are not compatible with
10–20% in a contaminated melt. The reduction of
all alloys, such as those containing calcium,
inclusion content in the melt leads to an increase in
strontium and rare earth elements.
the fluidity, as illustrated in Table 8.58 Similar results
(ii) fluxless methods include use of protective gases,
were obtained by Wang et al.82 and Wu et al.127
filtration, gas bubbling and electromagnetic
Corrosion resistance separation
Inclusions have a deleterious effect on the corrosion (iii) addition of some alloying elements (e.g. calcium
resistance of Mg alloys.82–84,108,122,128 As an example, or beryllium) to Mg melts reduces the oxidation
Wang et al.82 studied the effect of inclusions on the potential of Mg and consequently the inclusion
corrosion rate of Mg–10Gd–3Y–0?5Zr alloy. They content in the melt.
4.07 2.0
2.84 1.8
0.87 1.3
Inclusions and properties 2. J. R. Davis (ed.): ‘Aluminum and aluminum alloys’; 1993,
Materials Park, OH, ASM International.
The use of filters, inert gas bubbling, fluxes and degasser 3. H. Westengen and T. K. Aune: in ‘Magnesium technology:
addition are all effective methods to improve melt metallurgy, design data, automotive applications’, (ed. H.
cleanliness and mechanical properties of Mg alloys. Friedrich et al.), 145–204; 2006, Berlin/Heidelberg, Springer-
Improvements of 15% in tensile strength and 100% in Verlag.
4. B. R. Powell, A. A. Luo and P. E. Krajewski: in ‘Advanced
elongation are possible. Combination of filtration and materials in automotive engineering’, (ed. J. Rowe), 150–209;
other techniques could further improve melt cleanliness 2012, Cambridge, Woodhead Publishing Limited.
and mechanical properties. 5. G. Cole: ‘Magnesium vision 2020: a North American automotive
The presence of inclusions can also reduce melt strategic vision for magnesium’; 2006, Southfield, MI, United
States Council for Automotive Research LLC.
fluidity and corrosion resistance of castings. The
6. J. E. Hatch (ed.): ‘Aluminum: properties and physical metallurgy’;
presence of inclusions will increase viscosity of melts 1984, Metals Park, OH, ASM International.
making them more difficult to cast and form galvanic 7. M. O. Pekguleryuz and A. A. Kaya: Proc. Magnesium
coupling with the matrix, leading to pitting corrosion. Technology 2004 Symp., Charlotte, NC, USA, March 2004,
The presence of non-metallic inclusions will also TMS, 281–287.
prematurely wear cutting tools and as a result, reduce 8. NADCA: ‘Product specification standards for die castings’,
NADCA, Wheeling, IL, USA, 2006.
the machinability of Mg alloys. 9. Hydro Magnesium: ‘Magnesium die casting alloys, data sheet’;
There is a need to understand and control all aspects 2005, Livonia, MI, Hydro Magnesium.
of metal handling and processing to improve quality and 10. P. Bakke, P. K. Andersen, J. M. Svalestuen and S. I. Stromhaug:
consistency of the cast product. Mater. Sci. Forum, 2005, 488–489, 65–72.
11. P. Bakke, J. Svalestuen and D. Albright: Proc. SAE 2002 World
Perspectives Cong. & Exhib., Detroit, MI, USA, March 2002, SAE
International, 299–305.
Assessment of melt cleanliness 12. E. F. Emley: ‘Principles of magnesium technology’; 1966, Oxford,
None of the techniques developed to assess the melt Pergamon Press.
cleanliness of Mg alloys has been universally accepted. 13. R. E. Brown: Proc. 4th Int. Symp. on ‘Recycling of metals and
Metallographic techniques, humidity tests and K-mould engineered materials’, Pittsburgh, PA, USA, October 2000, TMS,
1317–1329.
fracture samples have been invaluable tools used in the 14. H. S. Tathgar, P. Bakke and T. A. Engh: Proc. 5th Int. Conf. on
foundry shop, but there is still a need for an efficient and ‘Magnesium alloys and their applications’, Munich, Germany,
economical assessment technique. A combination of September 2000, Wiley-VCH, 767–779.
methods may be needed. Future areas of research should 15. J. D. Hanawalt and C. E. Nelson: ‘Magnesium base alloy’, US
aim at utilising filtration methods such as self-gravita- Patent 2,264,309, 2 December 1941.
16. F.-S. Pan, J.-J. Mao, X.-H. Chen, J. Peng and J.-F. Wang: Trans.
tion and HMIAM as these techniques are relatively Nonferrous Met. Soc. China, 2010, 20, 1299–1304.
easier to implement on a foundry scale and are of 17. A. Ditze and C. Scharf: ‘Recycling of magnesium’; 2008,
relatively low cost. Clausthal-Zellerfeld, Papierflieger Verlag.
18. J. Campbell (ed.): ‘Castings’, 2nd edn; 2003, Oxford/Boston, MA,
Control of inclusions Elsevier Butterworth-Heinemann.
The methods used to control the inclusions in the melt 19. ASM International: ‘ASM handbook’, Vol. 15, ‘Castings’; 1998,
Materials Park, OH, ASM International.
need to be optimised, since some results are contra-
20. P. Bakke, J. L. Lauritzen, T. A. Engh and D. Oymo: Proc. Light
dictory and quantitative results are missing. Most Metals 1991 Conf., New Orleans, LA, TMS, 1015–1022.
importantly, the process costs to achieve acceptable 21. R. S. Busk and E. G. Bobalek: Trans. Am. Inst. Min. Metall. Pet.
quality are sometimes prohibitive. Eng., 1947, 171, 261–276.
22. W. F. Gale and T. C. Totemeier: ‘Smithells metals reference
Understanding of inclusion formation and influence on Mg book’, 8th edn; 2004, Oxford/Burlington, Ont., Elsevier
alloy properties Butterworth-Heinemann.
23. T. A. Engh: ‘Principles of metal refining’; 1992, Oxford/New
The majority of inclusion related issues with Mg alloys
York, Oxford University Press.
appear to originate with the molten Mg being exposed to 24. Y. Kitahara, H. Shimazaki, T. Yabu, H. Noguchi, M. Sakamoto
moisture or oxygen. This is an issue during melting, melt and H. Ueno: Mater. Sci. Forum, 2005, 482, 359–362.
transfer before pouring and mould filling. There is a 25. D. G. L. Prakash, D. Regener and W. J. J. Vorster: Mater. Sci.
need to determine the Mg melt quality at all stages of the Eng. A, 2008, A488, 303–310.
26. ‘Standard specification for magnesium alloys in ingot form for
casting process. Better understanding of the behaviour
sand castings, permanent mold castings, and die castings’, ASTM
of oxide films (in particular bifilms) within the molten B93/B 93M-07, ASTM International, West Conshohocken, PA,
metal would help determine the best foundry practices to USA, 2007.
reduce their formation and entrapment in Mg melts. 27. S. X. Xu, S. S. Wu, Y. W. Mao, P. An and P. Q. Gao: Foundry
World, 2006, 3, 275–278.
28. P. Bakke, K. Pettersen, S. Guldberg and S. Sannes: Proc. 7th Int.
Acknowledgements Conf. on ‘Magnesium alloys and their applications’, Dresden,
Germany, November 2006, Wiley-VCH, 739–745.
The authors are indebted to Professor A. McLean for 29. U. Galovsky and M. Kühlein: Proc. 7th Int. Conf. on ‘Magnesium
his review and valuable comments. They thank the alloys and their applications’, Weinheim, Germany, November
members of the Centre for Near-Net-Shape Processing 2006, Wiley-VCH, 746–751.
of Materials, in particular A. Machin, J. Hill, S. Ahmad 30. G. H. Wu, M. Xie, C. Q. Zhai, X. Q. Zeng, Y. P. Zhu and W. J.
and E. Vandersluis for their suggestions. They sincerely Ding: Trans. Nonferr. Met. Soc. China, 2003, 13, 1260–1264.
31. H. Hu and A. Luo: JOM, 1996, 48, 47–51.
acknowledge the Natural Sciences and Engineering
32. D. Oymo, D. O. Karlsen, P. M. D. Pinfold, T. Mellerud and
Research Council of Canada for financial support. O. Lie: Proc. Light Metals 1994 Conf., San Francisco, CA, USA,
February–March 2004, TMS, 1017–1024.
References 33. C. J. Simensen and B. Oberlander: Prakt. Metallogr., 1980, 17,
125–135.
1. M. Avedesian and H. Baker (eds.): ‘Magnesium and magnesium 34. J. M. Tartaglia and J. C. Grebetz: Proc. Magnesium Technology
alloys’; 1999, Materials Park, OH, ASM International. 2000 Symp., Nashville, TN, USA, March 2000, TMS, 113–121.
35. P. Bakke and D. O. Karlsen: Proc. 1997 SAE Int. Cong. & Expo., 68. C. Carozza, P. Lenard, R. Sankaranarayanan and R. I. L.
Detroit, MI, USA, February 1997, SAE International, 61–73. Guthrie: Proc. 36th Annual Conf. of Metallurgists of CIM – Light
36. O. Kubaschewski and C. B. Alcock: ‘Metallurgical thermochem- Metals Symp., Sudbury, Ont., Canada, August 1997, Canadian
istry, international series on materials science and technology’, Institute of Mining, Metallurgy and Petroleum, 185–196.
Vol. 24; 1979, Oxford/New York, Pergamon Press. 69. H. Westengen: in ‘Magnesium technology: metallurgy, design
37. O. Kubaschewski, C. B. Alcock and P. J. Spencer: in: ‘Materials data, automotive applications’, (ed. H. Friedrich et al.), 633–664;
thermochemistry’, 257–323; 1993, Oxford/New York, Pergamon 2006, Berlin/Heidelberg, Springer-Verlag.
Press. 70. Y. Ono, J. F. Moisan, Y. Zhang, C. K. Jen and C. Y. Su: JOM,
38. M. Pekguleryuz: in ‘Magnesium technology: metallurgy, design 2004, 56, 59–64.
data, automotive applications’, (ed. H. Friedrich et al.), 109–127; 71. Y.-J. Chen and P.-S. Wei: Mater. Trans., 2007, 48, 3181–3189.
2006, Berlin/Heidelberg, Springer-Verlag. 72. C. D. Fuerst and C. J. Dasch: Proc. Magnesium Technology 2000
39. R. S. Busk and R. B. Jackson: Proc. 37th Annual World Conf. on Symp., Nashville, TN, USA, March 2000, TMS, 107–111.
‘Magnesium’, Salt Lake City, UT, USA, June 1980, International 73. D. Mackie, J. Robson, P. Withers and M. Turski: Proc. 9th Int.
Magnesium Association, 1–4. Conf. on ‘Magnesium alloys and their applications’, Vancouver,
40. H. T. Gao, G. H. Wu, Y. Fan, W. J. Ding and Y. P. Zhu: Mater. BC, Canada, July 2012, Wiley-VCH, 881–887.
Sci. Forum, 2005, 488–489, 25–30. 74. K. D. Liss, X. Thibault, H. Li and P. Bendeich: J. Mater. Sci.
41. S. E. Housh and V. Petrovich: Proc. 1992 SAE Int. Cong. & Eng., 2011, 5, 195–199.
Expo., Detroit, MI, USA, February 1992, SAE International, 1–7. 75. D. Argo, P. Forakis and M. Lefebvre: ‘Chemical composition and
42. W. Schubert and H. Gjestland: Proc. 7th Int. Conf. on cleanliness during recycling of the AJ52 magnesium strontium
‘Magnesium alloys and their applications’, Dresden, Germany, alloy’, Proc. Magnesium Technology 2003 Symp., San Diego,
November 2006, Wiley-VCH, 761–766. USA, TMS, February 2003, 33–37.
43. S. P. Cashion, N. J. Ricketts and P. C. Hayes: J. Light Met., 2002, 76. ABB: PoDFA brochure, 2004.
2, 43–47. 77. S. Poynton, M. Brandt and J. Grandfield: ‘A review of inclusion
44. S. P. Cashion, N. J. Ricketts and P. C. Hayes: J. Light Met., 2002, detection methods in molten aluminium’, Light Metals 2009 –
2, 37–42. TMS 2009 Annual Meet. and Exhib., February 2009, San
45. H. Chen, J. Liu and W. Huang: J. Mater. Sci., 2006, 41, 8017– Francisco, CA, USA, 2009, Minerals, Metals and Materials
8024. Society, 681–687.
46. H. Dorsam: Proc. Magnesium Technology 2000 Symp., Nashville, 78. S. Poynton, M. Brandt and J. Grandfield: Mater. Sci. Forum,
TN, USA, March 2000, TMS, 99–105. 2010, 630, 155–164.
47. W. Ha and Y. J. Kim: J. Alloys Compd, 2006, 422, 208–213. 79. J. C. Grebetz and A. G. Haerle: Light Met. Age, 1997, 55, 60–69.
48. P. Lyon, P. D. Rogers, J. F. King, S. P. Cashion and N. J. 80. W. M. Haynes (ed.): ‘CRC handbook of chemistry and physics’,
Ricketts: Proc. Magnesium Technology 2003 Symp., San Diego, 91st edn; 2010, Boca Raton, FL, CRC Press/Taylor & Francis.
CA, USA, March 2003, TMS, 11–14. 81. M. J. Liang, G. H. Wu, W. J. Ding and W. Wang: Trans.
49. D. S. Milbrath: Proc. 2nd Int. Conf. on ‘SF6 and the environ- Nonferrous Met. Soc. China, 2011, 21, 717–724.
ment’, San Diego, CA, USA, November 2002, United States 82. W. Wang, Y. Huang, G. Wu, Q. Wang, M. Sun and W. Ding:
Environmental Protection Agency. J. Alloys Compd, 2009, 480, 386–391.
50. N. J. Ricketts and S. P. Cashion: Proc. Magnesium Technology 83. W. Wang, G. Wu, M. Sun, Y. Huang, Q. Wang and W. Ding:
2001 Symp., New Orleans, LA, USA, February 2001, TMS, 31– Mater. Sci. Eng. A, 2010, A527, 1510–1515.
36. 84. W. Wang, G. Wu, Q. Wang, Y. Huang and W. Ding: Mater. Sci.
51. S. M. Xiong and X. L. Liu: Metall. Mater. Trans. A, 2007, 38A, Eng. A, 2009, A507, 207–214.
428–434. 85. W. Wang, G. H. Wu, Q. D. Wang, Y. G. Huang, M. Sun and
52. Y. W. Zeng, L. M. Peng, X. M. Mao, X. Q. Zeng and W. J. Ding: W. J. Ding: Trans. Nonferrous Met. Soc. China, 2008, 18, s292–
Mater. Sci. Forum, 2005, 488–489, 73–76. s298.
53. A. R. Mirak, M. Divandari, S. M. A. Boutorabi and J. A. Taylor: 86. H. T. Gao, G. H. Wu, W. J. Ding and Y. P. Zhu: Trans.
Int. J. Cast Met. Res., 2012, 25, 188–194. Nonferrous Met. Soc. China, 2004, 14, 530–536.
54. W. D. Griffiths and N. W. Lai: Metall. Mater. Trans. A, 2007, 87. P. Forakis, E. Richard and D. Argo: Proc. Magnesium
38A, 190–196. Technology 2002 Symp., Seattle, WA, USA, February 2002,
55. B. Lagowski: AFS Trans., 1979, 87, 387–390. TMS, 49–54.
56. J. M. Tartaglia, J. H. Howard, R. E. Swartz and R. L. J. Bentz: 88. C. F. Baker: Proc. 1990 SAE Int. Cong. & Expo., Detroit, MI,
JOM, 2001, 53, 16–19. USA, February–March 1990, SAE International, 1–5.
57. C. Brassard and L. Riopelle: Proc. 1997 TMS Annual Meet. on 89. S. Couling, F. Bennett and T. Leontis: Light Metals, 1977, 1, 545–
‘Light metals’, Orlando, FL, USA, February 1997, TMS, 1111– 560.
1114. 90. S. L. Couling and T. E. Leontis: Light Met., 1980, 4, 997–1009.
58. B. Bronfin, N. Polyak, E. Aghion, C. Fuerst and D. Barris: Proc. 91. S. L. Couling: Proc. 36th Annual World Conf. on ‘Magnesium’,
Magnesium Technology 2002 Symp., Seattle, WA, USA, Oslo, Norway, June 1979, International Magnesium Association,
February 2002, TMS, 55–60. 54–57.
59. V. Ohm, S. Griese and P. Johnen: Proc. 9th Int. Conf. on 92. J. W. Fruehling and J. D. Hanawalt: AFS Trans., 1969, 77, 159–
‘Magnesium alloys and their applications’, Vancouver, BC, 164.
Canada, July 2012, Wiley-VCH, 961–966. 93. M. J. Walzak, R. D. Davidson, N. S. McIntyre, D. Argo and
60. S. Shang, A. F. Yu, T. Zhu and X. Mei: Proc. Magnesium B. R. Davis: Proc. Magnesium Technology 2001 Symp., New
Technology 2001 Symp., New Orleans, LA, USA, February 2001, Orleans, LA, USA, February 2001, TMS, 37–41.
TMS, 67–72. 94. M. Maiss and C. A. M. Benninkmeijer: Environ. Sci. Technol.,
61. C. J. Simensen: Metall. Trans. B, 1981, 12B, 733–743. 1998, 32, 3077–3086.
62. A. G. Haerle, B. A. Mikucki and W. E. Mercer II: Light Metal 95. H. Chen, J. Liu and W. Huang: Mater. Charact., 2007, 58, 51–58.
Age, 1996, 54, 22–29. 96. G. You, S. Long and R. Li: Mater. Sci. Forum, 2007, 546–549,
63. C. D. Fuerst and W. D. James: Proc. 15th Int. Conf. on ‘the 119–122.
Application of accelerators in research and industry’, Denton, 97. V. Revankar, P. Baker, A. Schultz and H. Brandt: Proc. 57th
TX, USA, November 1998, The American Institute of Physics, Annual World Magnesium Conf., Vancouver, BC, Canada, May
731–734. 2000, International Magnesium Association, 51–55.
64. W. D. James and C. D. Fuerst: J. Radioanal. Nucl. Chem., 2000, 98. F. W. Bach, A. Karger, C. Pelz and M. Schaper: Proc. Magnesium
244, 429–434. Technology 2005 Symp., Warrendale, PA, USA, February 2005,
65. M. Sabsabi, L. St-Onge, V. Detalle and J. M. Lucas: Proc. 16th TMS, 3–6.
World Conf. on ‘Nondestructive testing’, Montreal, Que., 99. A. Karger, F. W. Bach and C. Pelz: Mater. Sci. Forum, 2005, 488–
Canada, August–September 2004, Canadian Institute for Non- 489, 85–88.
Destructive Evaluation, Paper 679. 100. L. Martins and S. Kannan: Mod. Cast., 2003, 93, 39–41.
66. C. J. Simensen and A. I. Spjelkavik: Fresenius J. Anal. Chem., 101. R. M. Smith: Proc. 42nd Annual Technical Meet. of the
1980, 300, 177–182. Investment Casting Institute, Atlanta, GA, USA, September
67. S. Kuyucak and R. I. L. Guthrie: Proc. 26th Annual Conf. of 1994, Investment Casting Institute, Paper 13.
Metallurgists of CIM – Light Metals Symp., Winnipeg, Man., 102. E. Nazari, S. H. Razavi and S. M. A. Boutorabi: J. Mater.
Canada, August 1987, Pergamon Press, 229–238. Process. Technol., 2010, 210, 461–465.
103. P. Bakke, T. A. Engh, E. Bathen, D. Oymo and A. Nordmark: 116. Y. J. Chen, W. N. Hsu and J. R. Shin: Mater. Trans., 2009, 50,
Mater. Manuf. Processes, 1994, 9, 111–138. 401–408.
104. S. Makarov, R. Ludwig and D. Apelian: IEEE Trans. Magn., 117. M. R. Afshar, M. R. Aboutalebi, M. Isac and R. I. L. Guthrie:
2000, 36, 2015–2021. Mater. Lett., 2007, 61, 2045–2049.
105. G. Wu, M. Sun, J. C. Dai and W. Ding: China Foundry, 2010, 7, 118. A. I. Kulinsky and V. V. Agalakov: Proc. 1st Int. Non-ferrous
400–407. Processing and Technology Conf., St Louis, MO, USA, March
106. Q. Le, Z. Zhang, J. Cui and S. Chang: Mater. Sci. Forum, 2009, 1997, ASM International, 535–539.
610–613, 754–757. 119. B. H. Choi, I. M. Park, B. S. You and W. W. Park: Mater. Sci.
107. S. Tardif, R. Tremblay, D. Dube and Z. Zhang: Proc. 40th Forum, 2003, 419–422, 639–644.
Annual Conf. of Metallurgists of CIM – Light Metals Symp., 120. D. I. Jang and S. K. Kim: Mater. Sci. Forum, 2009, 620–622, 149–
Toronto, Ont., Canada, August 2001, Canadian Institute of 152.
Mining, Metallurgy and Petroleum, 225–235. 121. J.-K. Lee and S. K. Kim: Adv. Mat. Res., 2008, 47–50, 940–943.
108. J. Wang, J. X. Zhou, W. H. Tong and Y. S. Yang: Trans. 122. A. G. Haerle, R. W. Murray, W. E. Mercer II, B. A. Mikucki and
Nonferrous Met. Soc. China, 2010, 20, 1235–1239. M. H. Miller: Proc. 1997 SAE Int. Cong. & Expo., Detroit, MI,
109. J. Wang, X. G. Dong and Y. S. Yang: Int. J. Cast Met. Res., 2012, USA, February 1997, SAE International, 75–84.
25, 165–169. 123. G. G. Wang, B. Froese and P. Bakke: Proc. Magnesium
110. D. Emadi, J. P. Thomson, K. Sadayappan and M. Sahoo: Proc. Technology 2003 Symp., San Diego, CA, USA, March 2003,
Magnesium Technology 2005 Symp., San Francisco, CA, USA, TMS, 65–69.
February 2005, TMS, 335–339. 124. A. Elsayed, S. Lun Sin, E. Vandersluis, J. Hill, S. Ahmad and C.
111. C. D. Yim, G. Wu and B. S. You: Mater. Trans., 2007, 48, 2778–2781. Ravindran: AFS Trans., 2012, 120, 423–429.
112. G. H. Wu, J. C. Dai, M. Sun and W. J. Ding: Trans. Nonferrous 125. M. F. Horstemeyer, N. Yang, K. Gall, D. L. McDowell, J Fan
Met. Soc. China, 2010, 20, 2037–2045. and P. M. Gullett: 2004, 52, 1327–1336.
113. D. O. Karlsen, D. Oymo, H. Westengen, P. M. D. Pinfold and 126. A. L. Bowles, Q. Han and J. A. Horton: Proc. Magnesium
S. I. Stromhaug: Proc. Int. Symp. on ‘Light metals processing and Technology 2005 Symp., San Francisco, CA, USA, February
applications’, Quebec City, Que., Canada, August–September 2005, TMS, 99–104.
1993, Canadian Institute of Mining, 397–408. 127. G. Wu, W. Liu and W. Ding: Proc. 9th Int. Conf. on ‘Magnesium
114. Z. W. Shao, Q. C. Le, J. Z. Cui and Z. Q. Zhang: Trans. Nonferr. alloys and their applications’, Vancouver, BC, Canada, July 2012,
Met. Soc. China, 2010, 20, 382–387. Wiley-VCH, 1235–1241.
115. Z. W. Shao, Q. C. Le, Z. Q. Zhang and J. Z. Cui: Materialwiss. 128. J. Wang, Y. S. Yang and W. H. Tong: Trans. Nonferrous Met.
Werkstofftech., 2012, 43, 220–225. Soc. China, 2011, 21, 949–954.