Chapter 1

Basics of Solidification Processing

of Metallic Alloys

Dmitry G. Eskin

1.1 Structure Formation Upon Casting

The micro- and macrostructure of cast metal is very important from the point of view
of casting and downstream processing performance, as it determines the quality of
the casting and mechanical properties of as-cast and deformed products. Fine grain
structure means uniform distribution of grain size in the billet (ingot) cross-section,
elimination of columnar and feathery grains, lesser macrosegregation, uniform and
improved mechanical properties in the semisolid and solid states, decreased propen-
sity to hot and cold cracks, etc. Intermetallics and other excess inclusions (oxides,
carbides, nitrides, borides, etc.) should also be fine and evenly distributed in the cast
matrix. These inclusions are usually intrinsic to the metallic material and result from
its composition or contamination. Sometimes, however, the foreign inclusions may
be intentionally added to form a composite material or for the purpose of grain
refining (acting as substrates).
In this section, we will consider the main mechanisms of structure formation and
outline the main means to affect or control this structure.
Structure formation on the microscopic level depends on two basic phenomena:
nucleation and growth.
The fundamentals of nucleation are discussed in detail elsewhere [1, 2]. Here, we
only consider some most important parameters of heterogeneous nucleation, as this
type of nucleation is most relevant to real casting practice.
The solidification sites (or substrates, or heterogeneous nuclei) that ease the
nucleation are specially introduced or naturally form in the melt. Heterogeneous

D. G. Eskin (*)
Brunel Centre for Advanced Solidification Technology, Brunel University London,
Uxbridge, Middlesex, UK
e-mail: [email protected]

© Springer Nature Switzerland AG 2018 1

D. G. Eskin, J. Mi (eds.), Solidification Processing of Metallic Alloys
Under External Fields, Springer Series in Materials Science 273,
https://doi.org/10.1007/978-3-319-94842-3_1
2 D. G. Eskin

nucleation is energetically advantageous over homogeneous nucleation as the

undercooling (energy stimulus) required for the formation of the new solid phase
from the melt is lower by the level of energy required to form the new interface.
Hence, heterogeneous nucleation is stimulated by the decrease of the interfacial
energy that is required to form the new solid phase. The substrate that can efficiently
do the job should therefore: (1) be wettable by the liquid phase (active); (2) provide
crystallographic planes that resemble the crystallographic planes of the solid phase;
(3) provide chemistry required for the formation of the new phase; and (4) be of the
size that matches the actual undercooling achieved upon solidification. For example,
the actual undercooling required for activation of all potent heterogeneous solidifi-
cation sites in liquid aluminum is less than 2  C [2]. The type, amount, and size of the
substrates determine the efficiency of heterogeneous nucleation. First, particles have
to be sufficiently large to become active substrates at a given undercooling. In the
case of good wetting and crystallographic match, the required undercooling is very
small. The nucleation rate depends solely on the temperature (not on the time since
nucleation occurs instantaneously), and this type of nucleation is sometimes called
“athermal,” though this term is misleading. Then, growth occurs until the new phase
reaches the size of the substrate and subsequent growth will require further
undercooling. The nature of the substrate (interfacial energy) determines how
efficient the nucleation is, while the size of the substrate determines the ease of
subsequent growth [3].
The undercooling ΔT for the so-called free growth is related to the size of the
substrate d as

4σ sl
ΔT ¼ , ð1:1Þ
ΔSV d
where σ sl is the solid–liquid interfacial energy and ΔSV is the volumetric entropy of
fusion. In reality, there is a size distribution of potential solidification sites that
become active in the melt at a specific degree of undercooling. For the substrate size
of 1 μm, the required undercooling for free growth in liquid aluminum is about
0.5  C, whereas the real measured undercooling in aluminum alloys is about 0.2  C
which corresponds to the size range 3–5 μm [3].
The growth of the nucleated grain is determined to a large extent by the compo-
sition of the melt ahead of the solid–liquid interface. Despite the common notion that
the diffusion in liquid occurs almost instantaneously, there is always an accumula-
tion of solute elements at the solid–liquid interface as compared to the bulk compo-
sition. This accumulation causes on the one hand the phenomenon called
constitutional undercooling, and on the other hand prevents the growth of the solid
phase, as the growth can only occur when the compositions of the solid and liquid
phases at the solid–liquid interface are in equilibrium, according to the phase
diagram. The degree of solute accumulation depends on the alloy composition
(partitioning coefficients of alloying elements) and on the solidification conditions
(diffusion and convection), which explains why the final structure formed in
a casting depends on both the alloy composition and the casting conditions
1 Basics of Solidification Processing of Metallic Alloys 3

Fig. 1.1 Dependence of the

grain size in Al–Cu alloys
on the cooling rate and the
composition

Fig. 1.2 Characteristic points on the binary phase diagrams (a) and the variation of composition-
ally determined solidification temperature (TC) and real melt temperature (Tmelt) forming the regions
of constitutional undercooling (or superheating) (b)

(e.g., temperature gradient). Figure 1.1 illustrates this fact for binary Al–Cu alloys
and Fig. 1.2 explains the principle of constitutional undercooling. Early systematic
work to examine the factors controlling the formation of the equiaxed structure was
reported by Northcott who showed for copper alloys that the amount of alloying
elements influenced the length of columnar crystals and the extent of the equiaxed
structure [4]. He demonstrated that alloys with a wider freezing range more readily
4 D. G. Eskin

produced the equiaxed structure. Cibula [5] and Eborall [6] reported the effect of
alloy composition on grain refinement. Nowadays, the growth-restriction theory is
generally accepted as one of the main mechanisms to control the grain size in
castings [3, 7–11]. Let us consider it in brief.
Two parameters have been suggested for quantifying the effect of alloy compo-
sition on grain size, both parameters being proportional to the solidification range.
These are the undercooling parameter:
P ¼ mC 0 ðK 1Þ=K ¼ ΔT 0 ð1:2Þ

and the growth-restriction factor:

Q ¼ mC 0 ðK 1Þ ¼ KΔT 0 , ð1:3Þ

where m is the liquidus slope at the alloy composition C0, K is the partition
coefficient, and ΔT0 is the equilibrium solidification range, i.e., difference between
equilibrium liquidus and solidus temperatures.
The undercooling parameter represents the maximum possible undercooling and
is equal to the solidification range of an alloy with composition C0, while the growth-
restriction factor reflects the solute rejection at the solid/liquid interface and potential
degree of solute accumulation at the interface (see Fig. 1.2b).
The constitutional undercooling at the solid/liquid interface is proportional to the
growth-restriction factor:
ΔT c ¼ ΞQ, ð1:4Þ

where Ξ ¼ (CssL C0)/(CssL (1 K )) is the solutal supersaturation. CssL represents

the liquid constitution at the interface, which is different from the equilibrium
nominal composition C0, when undercooling is present (see Fig. 1.2).
Generally, the constitutional undercooling at the solid/liquid interface will control
the growth according to the general formulation for the maximum growth velocity
[12, 13]:

V growth ¼ A Ds ΔT 2=QΓ, ð1:5Þ

where Ds is the diffusion coefficient of the solute in the melt, ΔT is the undercooling
required for the growth (can be constitutional undercooling or the total
undercooling), Q is the growth-restriction factor, Γ as the Gibbs–Thompson coeffi-
cient reflecting the effect of the dendrite tip curvature on the tip growth velocity.
(Γ ¼ σ ls/ΔSf, where σ ls is the solid/liquid interfacial energy and ΔSf is the volumetric
entropy of fusion, which corresponds to the latent heat of solidification), and A is a
coefficient. We can see, at least qualitatively, that slow solute diffusion (Ds), small
undercooling (ΔTc), large solidification range and partitioning (Q), and large inter-
facial energy (σ ls) would result in limited growth of the solid phase into the liquid.
The growth of the solid phase requires that the solute supersaturation at the
interface is dissipated either by diffusion (convection) or by remelting of the solid
phase, which has been experimentally demonstrated [14]. There may be various
1 Basics of Solidification Processing of Metallic Alloys 5

reasons for the solute accumulation or the solute influx to the solid/liquid interface.
Most frequently, the growth restriction is considered in relation to grain refinement
by inoculants and to the columnar-to-equiaxed transition (CET) [7–9]. In the course
of grain refinement, the solute rejected by the growing grain slows down its growth
and, at the same time, may create the constitutional undercooling sufficient for the
nucleation of a new grain at the substrate that preexists in the melt [9].
In the case of CET, the “added” solute is the solute rejected by the columnar grain
tip, but it can also come from the neighboring columnar grain or the solid phase
nucleated ahead of the columnar front. The constitutional undercooling due to the
solute rejection at the solid/liquid interface creates favorable conditions for hetero-
geneous nucleation at some distance ahead of the solidification front (see Fig. 1.2b).
The nucleated grains start to grow and reject solute as well as generate the latent heat.
As a result, the growth of the “parent” interface stops. The greater the solidification
rate (solid fraction evolution per unit time), the more effective is this growth
restriction. Martorano et al. showed that the rejection of solute by newly formed
grains with a high grain density would, in fact, effectively decrease the constitutional
undercooling initially created at the “parent” interface and stop its advance [15].
One can notice that the area of maximum constitutional undercooling caused by
solute accumulation (or depletion in the case of K > 1, e.g., Ti in Al) is at the some
distance from the solid/liquid interface, where the condition for the nucleation of a
new solid crystal is the most favorable (Fig. 1.2b).
The solute can be also brought to the interface by flows during solidification that
can be natural, e.g., thermo-solutal and shrinkage-driven, or forced by external
fields, e.g., stirring, streaming, mixing, as well as can result from nonequilibrium
solidification conditions. In all cases, the volume around the interface cannot be
considered as a closed system. The extent of this solute accumulation depends on the
effective diffusivity of the solute in the melt (see Eq. (1.5)). In order to maintain the
local equilibrium at the interface and get the momentum for further growth (solid-
ification), the interface should first remelt and absorb the excess of the solute. This
situation is shown in Fig. 1.2b and is often a case upon fragmentation of dendrite
branches [16, 17]. It is worth mentioning here that fragmentation or “grain multi-
plication” is considered as one of the mechanisms of grain refinement in castings,
also under the action of external fields (see Chaps. 4, 5, 6, and 7). Intense convection
during solidification may assist in flattening solute diffusion profiles ahead of the
solid/liquid interface, which may result in a faster grain growth and, according to the
solute-suppressed nucleation concept, in finer grain sizes in commercial alloys [18].
It is also suggested that the solute accumulated at the interface hinders the growth
only until some values of Q, when the solid–liquid interface is relatively smooth
(which is the case of globular and little-branched grains) [9, 19]. In the case of high
Q, and hence high constitutional undercooling, grains start to grow faster and in a
much more branched manner. As a consequence, the solute is rejected from the sharp
and elongated dendrite-branch tips not only to the front but also sideways. This
solute is accumulated in the liquid between the dendrite branches and is, on
decreasing the temperature, absorbed by the solid phase. In such a way, the effective
growth restriction is decreased and the dendrite grains get an opportunity to grow
faster, even if the average concentration of solute in the melt is high.
6 D. G. Eskin

We can summarize that the accumulation of solute (solvent in peritectic systems)

that assists the nucleation at some distance ahead of the solid/liquid interface
simultaneously prevents the growth of the parent particles. The latent heat released
during solidification has the same effect. The larger the nucleation density of the
primary grains, the sooner their thermo-solutal fields start to interact and the lesser is
their dendritic development. As a result, a fine and not well-developed dendritic
equiaxed structure is formed. Sometimes, the branching is so negligible that the
grains are called globular or nondendritic.
One of the main means of controlling the structure is by grain refinement through
the sharp increase of solidification sites for heterogeneous nucleation of the primary
phase. This can be done essentially in four ways:
• By alloying with elements that will form nucleating substrates (intermetallics)
during solidification prior to the primary solidification of the solid solution (e.g.,
Ti and Zr in aluminum and magnesium alloys, respectively);
• By adding special insoluble particles to an alloy that act as substrates either for
nucleating intermetallic particles or for the solid solution (e.g., TiB2 and Al4C3 in
aluminum and magnesium alloys, respectively);
• By activation of insoluble particles (e.g., Al2O3 by wetting and Al4C3 by melt
superheating in aluminum and magnesium alloys, respectively);
• By multiplication of solidification sites by fragmentation of existing solid grains
or fragmentation of nucleating intermetallic particles, or de-agglomeration of
nucleating insoluble particles (e.g., by cavitation melt treatment).
The first two means of grain refinement are by far the most widely used methods
for producing castings with small, equiaxed, and uniformly distributed grains. The
application of external fields may be useful in accelerating the dissolution of alloying
additions or dispersion and de-agglomeration of added particles.
In the two other approaches, the use of external fields maybe decisive for success.
The insoluble inclusions (e.g., oxides or carbides) are usually not very wettable
by the liquid metal, either due to gaseous phase being present, other contamination
absorbed at the surface, or due to too high surface energy. They, therefore, are
excluded from the process of solidification. If, however, they could be wetted with
liquid phase and form substrates either for grain refining phases (e.g., intermetallics
like Al3Ti) or for the solid matrix phase (aluminum or magnesium), then they would
trigger efficient refinement. This process is called activation.
In aluminum, the most likely indigenous inclusions are represented by alumina or
aluminum oxide. Alumina may exist in various crystal modifications, depending on
the temperature range, origin, and presence of hydrogen and other alloying elements.
The transition between the lower-temperature modifications of alumina upon
heating is in reality continuous and related to the dehydration (removal of hydrogen
from the crystal lattice). On increasing the melt temperature, δ-alumina (or
γ-alumina which is sometimes considered to be the same phase) transforms to
χ-alumina (700–850  C) and then the stable α-alumina is formed at above 1000  C
[20]. During cooling from 1000  C, the stable modification α-alumina is either
retained (at high cooling rates) or is transformed to χ-alumina (at low cooling
rates) without δ-alumina forming [20].
1 Basics of Solidification Processing of Metallic Alloys 7

One of the approaches to grain refinement is multiplication of available or

potential solidification sites (substrates for heterogeneous nucleation). Technologi-
cal approaches towards realizing this mechanism usually involve physical action
upon the melt with fragmentation of growing crystals or agglomerates and dispersion
of the formed solidification sites in the melt volume by forced convection.
These approaches include structure refinement through cavitation, shearing,
electromagnetic stirring, and vibration [26–28] and will be considered in detail in
Chaps. 3, 4, 5, and 6. For example, it was shown that the intense shearing (not
covered in this book) of liquid aluminum alloys results in dispersion of oxide films
and facilitates activation of fine oxide particles with respect to (Al) and Al3Ti
[21]. The activation of oxides can be also done by thermal (see Chap. 8) or other
physical action upon the melt (see, e.g., Chap. 5).
The presence of magnesium in aluminum alloys changes the nature of indigenous
oxides as magnesium has greater affinity for oxygen than aluminum [20]. Even small
concentrations of magnesium, 0.01–1 wt%, result in the formation of mixed oxides
MgO
Al2O3 þ MgO. At the magnesium content above 1%, the magnesium oxide
MgO almost completely substitutes alumina.
The role of magnesium oxide in grain refining of aluminum is not reported, but
there is enough evidence to suggest that it may play a significant part in grain
refinement of magnesium alloys. The activation of magnesia particles can be
achieved by ultrasonic cavitation treatment, which is evidenced by grain refinement
of cavitation-treated magnesium melts [22, 23]. Melt shearing is also shown to be
very efficient in activating magnesium oxides. The intense shearing of melt flow
performed by twin screws well above the liquidus of a magnesium alloy results in
efficient fragmentation of oxide films and increases the density of MgO particles by
three orders of magnitude and the amount of particles that become active at the
reached undercooling by a factor of 20 [24].
The formation of new solidification sites can be achieved, in some special cases,
by superheating of the melt as considered in Chap. 8. For example, in Mg–Al–Mn
alloys melt superheating followed by relatively rapid cooling results in the prefer-
ential formation of the refining metastable τ(Al4Mn) phase, thus producing grain
refinement [25]. Similar effect can be achieved in Al–Zr alloys when rapid cooling of
the melt from high temperature produces metastable cubic L12 Al3Zr phase with the
crystal structure similar to that of aluminum. In both cases, the melt superheating
deactivates all available substrates for heterogeneous nucleation of the solid solution
and equilibrium intermetallics. As a result, a deeper undercooling can be achieved
for the primary solidifying phase, deep enough to trigger the formation of a meta-
stable phase.
Castings also have macroscopic structure features, such as zones of fine and
coarse equiaxed grains, columnar and feathery grains, segregates, and porosity. All
these features depend on the kinetics of solidification and can be controlled by the
macroscopic and microscopic melt flow (convection and shrinkage, forced convec-
tion), heat flow (thermal gradient and positions of liquidus and solidus), and
precipitation of solid and gaseous phases. The effect of flow is illustrated in the
next section.
8 D. G. Eskin

1.2 Effects of Flow on Structure Formation During

Solidification

Over the past 50 years, many attempts have been made to design special techniques
with the purpose of improvement and control of the final solidified structure by
forced flow, which is the subject of Chaps. 3, 4, 5, 6, and 7.
During casting, flow generally occurs in the bulk liquid and in the semisolid
regions. Some of the techniques that have been successfully used in academic
research with the aim to study the fundamentals of solidification under forced-flow
conditions are: gravity flow-through systems [29], mechanical stirring [30], centrif-
ugal casting [31], controlled flow forced by an electromagnetic pump [32], applica-
tion of magnetic or electromagnetic field creating the Lorenz force (see Chaps. 3, 4,
and 5), etc. While the experimental research has been developing over the last
50 years, computational modeling and simulation have been widely used in the
last three decades as a cost-saving tool for prediction and interpretation of the
experimental results (see Chaps. 3, 4, and 7). The combination of these two
approaches helps in gaining deeper understanding of the effects of melt flow on
solidification phenomena in metallic alloys, that is, (1) morphology of grains and
their deflection towards incoming flow; (2) columnar to equiaxed transition (CET)
and grain morphology; and (3) change of segregation pattern.
Flemings’ pioneering work [33] and several later studies [29, 34] where the
gravity-driven system was used have reported that the columnar grain structure is
formed in the casting systems where the melt flows perpendicular to the heat
extraction. The grains are deflected towards the incoming flow. The tendency of
growing columnar dendrites to incline into the flow under the influence of some
shear forces at the solidification front is also a common phenomenon in centrifugal
casting [31], electromagnetic stirring [35, 36], and directional solidification [37]. Ini-
tially, it was proposed that the driving forces leading to the preferred growth
directions were due to interface kinetics and anisotropic surface energy. They were
attributed to flow-induced asymmetric thermal or concentration fields around the
growing crystal [33]. The solute-rich liquid is swept away from the upstream side of
the dendrite tip to the downstream side, leading to a local enrichment of the solute
and lowering of the liquidus temperature on the downstream side. This gives
preferential growth conditions on the upstream side, thus causing the crystal tips
to incline upstream [30]. Another theory suggested that the deflection of the growth
direction of columnar grains could be controlled by the dendrite branching mecha-
nism, namely by the competitive growth of primary and secondary arms [29]. The
dendritic evolution of an initially small nucleus affected by an external flow was also
treated using phase-field modeling [38]. It was shown that the local heat transfer at
the solid/liquid interface controlled the dendrite growth and when the forced flow
was increased, the main vertical stem tilted gradually towards the upstream direction.
The CET, various grain morphologies, and complex grain distribution due to melt
flow are the subjects of many studies. Several theories were developed to explain the
CET, including: heterogeneous nucleation in the undercooled melt [39, 40]; floating
1 Basics of Solidification Processing of Metallic Alloys 9

grains [41]; and crystal multiplication due to thermomechanical fragmentation

promoted by forced convection [31, 34, 42–44].
Feathery grains represent a specific growth morphology that is found in alumi-
num alloys [45]. In addition to high thermal gradient about 100 K cm 1 and growth
rate about 1 mm s 1, feathery grains are known to be favored by the presence of
strong fluid flow.
The melt flow is found to have an influence on the microstructure of metallic
alloys. The analytical solution based on asymptotic expansion method showed that
in Pb–Cu and Al–Cu alloys a weak convective flow increased the eutectic lamellar
spacing since the convection produced a slight shift of solute concentration profile at
the interface and decreased the eutectic growth undercooling [46]. Similar results
were obtained in another work [47], where the Jackson–Hunt model of eutectic
growth was extended to allow for different densities of the phases. Additionally, it
was experimentally found that the dendritic shape obtained under conditions of
natural convection changed to a globular form when the melt flow by electromag-
netic stirring was applied [42]. This was further confirmed by simulations
[48]. Forced flow, applied to the bulk of the molten metal, interacts with the growing
solid producing the distortion of the solid–liquid interface [29], altering the shape of
the mushy zone [49] and affecting solidification parameters [50]. Depending on the
nature of the flow and the initial velocity, oscillation of various magnitudes may
occur at the solidification front.
Let us illustrate the effects of melt flow on the structure formed during solidifi-
cation using experimental data obtained using an Al–4.5% Cu alloy and a special
setup where forced controlled flow was directed along the chilled cavity [32, 51].
Figure 1.3 shows two typical macrostructures of the longitudinal section of an
Al–4.5% Cu alloy obtained during solidification in the cavity under conditions of
constant forced flow along the solidification front. Depending on the flow condi-
tions, the macrostructure of samples either comprises a zone of equiaxed grains with
a columnar zone on top in the central part of the sample, expanding in the direction
of the downstream edge with respect to the bulk flow direction (Fig. 1.3a), or
consists entirely of columnar grains deflected towards incoming flow (Fig. 1.3b).
Equiaxed to columnar transition (ECT) is clearly seen at the bottom of the sample
(Fig. 1.3a).
The longitudinal macrostructure of the samples obtained under forced-flow
conditions can be conditionally divided into three zones. Zone 1 is the region
close to upstream edge with regard to the initial flow while zone 3 is the region of
downstream edge of the cavity. Therefore, zone 2 is the central part of the cavity.
During experiment, zone 1 is constantly affected by the hottest melt flow and the
highest thermal gradient. However, due to heat extraction from the bottom and
the side wall, the solidification onset occurs after the first seconds of experiment.
Similar heat extraction conditions are typical for zone 3. According to computational
results, zone 2 is affected by traveling either counter or/and clockwise vortices.
The following phenomena may contribute to formation of the zone of equiaxed
grains in Fig. 1.3a: (1) transport and continuous deposition of the detached dendrites
along the chill wall by the forced flow and (2) detachment of the slender dendrite
10 D. G. Eskin

Fig. 1.3 Macrostructures of an Al–4.5% Cu alloy solidified under forced flow (from left to right) at
(a) 0.15 m/s and (b) 0.05 m/s over the 100 mm water-cooled chill. A diagram in (c) illustrates the
formation of the structure shown in (a). (A.N. Turchin and D.G Eskin, 2008)

arms, and (3) their transport and deposition at the downstream edge of the cavity.
The macrostructural features, such as tiny columnar dendrites, duplex (coarse and
fine) internal structure, etc., observed in this zone [32] support the suggested
mechanisms.
The particular phenomenon of equiaxed-to-columnar transition (ECT) (Fig. 1.3a)
can be attributed to the termination of the fragmentation and settling processes with
the growth of the columnar dendrites under conditions of thermal gradient. After
onset of solidification, the fragments may form as results of detachment of the
equiaxed and columnar dendrites at the chill surface or fragmentation of their slender
dendrite arms. These fragments are either immediately removed from the chill zone
by the constantly coming forced flow or continuously settled in the downstream edge
of the chill. In the meantime, the columnar dendrites keep growing at the upstream
edge due to the highest thermal gradient and at the same time strong heat extraction
in this region (Fig. 1.3c). While the progressing solidification on the one hand
facilitates fragmentation due to branching of dendrites, on the other hand it increases
1 Basics of Solidification Processing of Metallic Alloys 11

Fig. 1.4 Typical microstructures of an Al–4.5% Cu alloy solidified under different forced-flow
conditions onto a water-cooled chill. The thermal gradient to solidification rate ratio G/V is given
(A.N. Turchin and D.G. Eskin, 2008)

the probability of fragments removal from the chill zone by the melt outflow due to
the decreasing geometrical ratio of the chill cavity with the progressed solidification.
At the same time, due to the gap between the family of columnar dendrites at the
upstream corner and the settled fragments at the downstream corner, the preferable
thermal conditions promote the free growth of columnar dendrites that are clearly
distinguished in Fig. 1.3a.
It is also possible to separate some zones in the vertical section (Fig. 1.4). It is
found that the zone close to the chill surface (zone A) may exhibit different
structures, such as: (1) coarse dendritic equiaxed structure, (2) globular equiaxed
structure, and (3) undeveloped columnar structure. The zone affected by forced flow
(zone B) consists of columnar dendrites and in some cases of “feathery crystals”
deflected towards incoming flow. The next zone (zone C) marks the solidification
front. Finally, at the top of the sample there is a zone (zone D) developed after the
end of experiment consisting of columnar dendrites, but with completely different,
always finer, internal structure as compared with the zone B.
12 D. G. Eskin

Fig. 1.5 Feathery grain structure of an Al–4.5% Cu alloy obtained under forced-flow conditions at
solidification rate of 0.6 mm/s and thermal gradient of 9 K/mm (a) and 0.7 mm/s, 15 K/mm (b); flow
direction: left to right (A.N. Turchin and D.G. Eskin, 2008)

The dimensions of each zone may vary depending on the flow conditions. The
grains growing in the flowing melt are generally deflected towards the incoming
flow. Also, the forced flow generally promotes the columnar growth. At the same
time, the increased flow velocity triggers columnar–equiaxed–columnar transitions,
mostly due to the fragmentation of dendrites (Fig 1.3).
Peculiar grain morphologies form upon solidification under forced-flow condi-
tions with a highly superheated melt (>55 K) and upon slow bulk flow (<0.05 m/s).
Figure 1.5a shows typical feathery crystals consisting of parallel lamellae growing
into the incoming flow. These lamellae are separated either by wavy or by straight
boundaries. The latter is a twinned boundary. Another morphology, typical of the
high thermal gradient and the solidification rate less than 1 mm/s, is presented in
Fig. 1.5b. Here, the dendrite arms, growing on the upstream side at 40 from the
primary trunk, exhibit a curved, woven-type shape and are directed to the incoming
flow. This type of morphology formed under forced-flow conditions is discussed in
detail elsewhere [45, 52].
The primary dendrite arm spacing (λ1) is plotted in Fig. 1.6 as a function of
solidification rate for different flow velocities at a thermal gradient of 4–4.5 K/mm,
superheat 55  C. The accelerated flow results in decreasing of primary dendrite arm
spacing and, at the same time, increasing of the solidification rate decreases the
distance between primary trunks. This tendency is valid for different superheats and
thermal gradients.
Dependence of secondary dendrite arm spacing (λ2) on solidification time, flow
velocity, and superheat is shown in Fig. 1.7. Although it has long been known that
the increasing of solidification time coarsens the structure, this plot shows the
pronounced effect of forced flow itself: microstructure coarsening is affected by
both the increasing solidification time and the flow rate [53]. The significant change
of the coarsening exponent with the increasing flow velocity has strong implications
for the formation of structure and should be taken into account.
1 Basics of Solidification Processing of Metallic Alloys 13

Fig. 1.6 Primary dendrite arm spacing vs solidification rate for an Al–4.5% Cu alloy cast under
different flow conditions (see the legend) and at a superheat of 55  C (A.N. Turchin and D.G. Eskin,
2008)

1.3 Main Factors Affecting the Structure Formation;

and Melt Processing and Casting Technologies

There are basic phenomena that can affect the structure formation and therefore the
external fields are used to facilitate or suppress these phenomena. The formation of
structure starts with nucleation. As it will be shown in Chaps. 5, 6, 7, and 8,
improved wetting of potential substrates for heterogeneous nucleation, and their
de-agglomeration and dispersion in the liquid volume (facilitated by electromagnetic
stirring, high-frequency vibration, electric pulses, or high melt temperature) play an
essential role in the observed structure modification. This is considered in Chaps. 3,
4, 5, 6, and 7. In addition, the change of the crystal structure of the substrates with
melt superheat may make the particles either more potent or otherwise, and affect the
structure formation accordingly as is shown in Chap. 8. Next comes the growth that
can be affected, as is discussed in Chaps. 3, 4, 5, and 6, by the melt flow (magne-
tohydrodynamics, electromagnetic stirring, magnetic, acoustic, and thermal convec-
tion), solutal redistribution (stirring, magnetic forces, forced convection, and melt
temperature), and thermal gradients (melt temperature, and forced convection).
Forced convection and mechanical forces (e.g., pressure pulses) in combination
with thermal and concentration gradients may also fragment the growing solid
14 D. G. Eskin

Fig. 1.7 Secondary dendrite arm spacing vs solidification rate for an Al–4.5% Cu alloys cast under
different flow conditions (see the legend) and at superheat (a) 55  C (G ¼ 4–5  C/mm) and (b)
75  C (G ¼ 13  C/mm). Coarsening exponents are given on the plot (A.N. Turchin and D.G. Eskin,
2008)
1 Basics of Solidification Processing of Metallic Alloys 15

phase and therefore change its morphology and multiply the amount of individual
solid crystals, effectively refining the structure. This is particularly important when
applying ultrasonic cavitation, electromagnetic stirring, magnetic force, or electric
pulses to the semisolid material, see Chaps. 4, 5, 6, and 7. The distribution of
alloying elements and reacting substances, substrates, reinforcing particles, and
crystal fragments in the bulk of the liquid as well as the control of the thermal and
solutal homogeneity can be done by convection (mass and heat flow) enforced by
electromagnetic, magnetic, and acoustic forces (Chaps. 3, 4, 5, and 6).
Alternating magnetic and electromagnetic fields have been used in solidification
processes for quite some time, especially in the iron and steel industry and in the field
of crystal growth. However, their application is still limited in the processing of other
metals, e.g., aluminum, magnesium, or even copper alloys. Recent years saw an
increased research and development in this area, which is due to the development of
both new hardware and new applications such as control of as-cast structure,
segregation, and production of gradient or composite materials.
The use of very strong continuous magnetic fields is now limited to small-scale
processes. But, the rapid development of large bore magnets at a lower cost opens up
prospects for their use in industrial applications.
There are special casting and melt processing methods explicitly designed to
incorporate external fields such as casting under pressure (high, low, vacuum,
isostatic, and centrifugal), melt processing in the presence of electromagnetic forces
(electrolysis cells, electromagnetic casting, casting with electromagnetic stirring, and
electromagnetic pumps), and melt processing with ultrasonic cavitation, high mag-
netic fields, or electro-pulse treatment. At the same time, some of these external
fields can be incorporated in the conventional gravity, die, or direct-chill casting
technologies, as it is demonstrated in Chaps. 4 and 5. The use of very high magnetic
fields (Chap. 6), electric and magnetic pulses (Chap. 7), and very high melt temper-
atures (Chap. 8) is currently reserved to specialized technologies applied to relatively
small melt volumes, e.g., rapid solidification or atomization.
It is worth to note that most of conventional melt processing and casting tech-
nologies use the effects of external fields as discussed in this monograph, through
intrinsic features such as thermal and solutal gradients (continuous casting, direct-
chill casting, suspension casting, and semisolid processing), natural thermal and
solutal convection (all casting techniques during mold filling and solidification,
electrolysis cells), and forced convection (low- and high-pressure die casting).

References

1. K.F. Kelton, A.L. Greer, Nucleation in Condensed Matter: Applications in Materials in Biology
(Elsevier, Amsterdam, 2010)
2. J.A. Dantzig, M. Rappaz, Solidification (EPFL Press/CRC Press, Lausanne, 2009)
3. A.L. Greer, Solidification and Casting (Institute of Physics Publishing, Bristol, 2003),
pp. 199–247
4. L. Northcott, J. Inst. Met. 62, 101–136 (1938); 65, 173–204 (1939)
16 D. G. Eskin

5. A. Cibula, J. Inst. Met. 76, 321–360 (1950)

6. M.D. Eborall, J. Inst. Met. 76, 295–320 (1950)
7. T.E. Quested, A.T. Dinsdale, A.L. Greer, Acta Mater. 53, 1323–1334 (2005)
8. M.A. Easton, D.H. St. John, Acta Mater. 49, 1867–1878 (2001)
9. M.A. Easton, D.H. StJohn, Metall. Mater. Trans. A 30A, 1613–1623 (1999)
10. B.S. Murty, S.A. Kori, M. Chakraborty, Int. Mater. Rev. 47, 3–29 (2002)
11. D.H. St. John, M. Qian, A. Easton, P. Cao, Z. Hildebrand, Metall. Mater. Trans. A 36A,
1669–1679 (2005)
12. J.D. Hunt, Mater. Sci. Eng. 65, 75–83 (1984)
13. M. Rappaz, P. Thevos, Acta Metall. 35, 1487–1497 (1987)
14. R.M. Mathiesen, L. Arnberg, Acta Mater. 53, 947–956 (2005)
15. M.A. Martorano, C. Beckermann, C.-A. Gandin, Metall. Mater. Trans. A 34A, 1657–1674
(2005)
16. K. Jackson, J. Hunt, D. Uhlmann, T. Seward, Trans. Metall. Soc. AIME 236, 149–158 (1966)
17. D. Ruvalcaba, R. Mathiesen, D.G. Eskin, L. Arnberg, L. Katgerman, Acta Mater. 55,
4287–4292 (2007)
18. D. Shu, B. Sun, J. Mi, P.S. Grant, Acta Mater. 59, 2135–2144 (2011)
19. M. Johnsson, Z. Metallkde. 85, 781–785 (1994). 786–789
20. V.I. Dobatkin, R.M. Gabidullin, B.A. Kolachev, G.S. Makarov, Gases and Oxides in Aluminum
Wrought Alloys (Metallurgiya, Moscow, 1976)
21. Z. Fan, Y. Wang, Z.F. Zhang, M. Xia, H.T. Li, J. Xu, L. Granasy, G.M. Scamans, Int. J. Cast
Metals Res. 22(1–4), 318–322 (2009)
22. G.I. Eskin, Ultrasonic Treatment of Light Alloy Melts (Gordon & Breach OPA, Amsterdam,
1998)
23. D.H. StJohn, M.A. Easton, M. Qian, Solid State Phenom. 141–143, 355–360 (2008)
24. H. Men, B. Jiang, Z. Fan. Acta Mater. 58, 6526–6534 (2010)
25. D. Qiu, M.-X. Zhang, J.A. Taylor, P.M. Kelly, Solidification Processing, ed by H. Jones.
Proceedings of the 5th Decennial International Conference on Solidification Processing (Uni-
versity of Sheffield, Sheffield, 2007), pp. 153–158
26. C. Vivès, Metall. Mater. Trans. B 20B, 623–629 (1989). 631–643
27. X. Wang, H. Zhang, Y. Zuo, Z. Zhao, Q. Zhu, J. Cui, Mater. Sci. Eng. A 497, 416–420 (2008)
28. Y. Zuo, H. Naganumi, J. Cui, J. Mater. Process. Technol. 197, 109–115 (2008)
29. K. Murakami, T. Fujiyama, A. Koike, T. Okamoto, Acta Mater. 31, 1425–1432 (1983)
30. A. Buchholz, S. Engler, Comput. Mater. Sci. 7, 221–227 (1996)
31. M. Guo, Y. Yang, F. Hua, Z. Hu, Z. Metallkde. 95, 835–839 (2004)
32. A.N. Turchin, Effects of Melt Flow in Casting Processes, PhD Thesis, Delft University of
Technology, Delft, 2008
33. M.C. Flemings, C.M. Adams, E.E. Hucke, H.F. Taylor, AFS 64, 636–639 (1956)
34. L.L. Rishel, A Study of the Effect of a Unidirectional Bulk Flow on the Structure of Selected
Aluminum-Copper Alloys. PhD thesis, University of Pittsburgh, USA, 1993
35. X. Li, Z. Ren, Y. Fautrelle, J. Mater. Process. Technol. 195, 125–134 (2008)
36. S. Steinbach, L. Ratke, Mater. Sci. Eng. A 413–414, 200–204 (2005)
37. X. Li, Y. Fautrelle, Z. Ren, Acta Mater. 55, 5333–5347 (2007)
38. R. Tönhard, G. Amberg, J. Cryst. Growth 194, 406–425 (1998)
39. M.C. Flemings, Solidification Processing, 1st edn. (McGraw-Hill, New York, 1974)
40. W. Kurz, D.J. Fisher, Fundamentals of Solidification, 3rd edn. (Trans Tech Publications Ltd.,
Aedermannsdorf, 1992)
41. C.J. Vreeman, F.P. Incropera, Int. J. Heat Mass Transf. 43, 687–704 (2000)
42. W.D. Griffiths, D.G. McCartney, Mater. Sci. Eng. A 216, 47–60 (1996)
43. W.D. Griffiths, D.G. McCartney, Mater. Sci. Eng. A 222, 140–148 (1997)
44. S. Eckert, B. Willers, P.A. Nikrityuk, K. Eckert, U. Michel, G. Zouhar, Mater. Sci. Eng. A
413–414, 211–216 (2005)
1 Basics of Solidification Processing of Metallic Alloys 17

45. A.N. Turchin, M. Zuijderwijk, J. Pool, D.G. Eskin, L. Katgerman, Acta Mater. 55, 3795–3801
(2007)
46. D. Ma, W.Q. Jie, Y. Li, S.C. Ng, Acta Mater. 46, 3203–3210 (1998)
47. S.R. Coriell, G.B. McFadden, W.F. Mitchell, B.T. Murray, J.B. Andrews, Y. Arikawa, J. Cryst.
Growth 224, 145–154 (2001)
48. A. Das, S. Ji, Z. Fan, Acta Mater. 50, 4571–4585 (2002)
49. P.A. Nikrityuk, K. Eckert, R. Grundmann, Int. J. Heat Mass Transf. 49, 150–1515 (2006)
50. P.A. Nikrityuk, K. Eckert, B. Willers, S. Eckert, Modeling of Casting, Welding and Advanced
Solidification Processes XI (The Minerals, Metals & Materials Society, Warrendale, 2006),
pp. 333–340
51. A.N. Turchin, D.G. Eskin, L. Katgerman, Int. J. Cast Metals Res. 20, 312–318 (2007)
52. S. Henry, T. Minghetti, M. Rappaz, Acta Mater. 46, 6431–6443 (1998)
53. A.N. Turchin, D.G. Eskin, L. Katgerman, Metall. Mater. Trans. A 38A, 1317–1329 (2007)