Fundamentals of Solidification

EXERCISES

(For graduate students)

Xuefeng Guo
2014.3.25

Henan Polytechnic University

1
EXERCISES

1.1 - Discuss the shape of the upper surface of the ingot in figure 1.6. What would happen if the solidifying material
was one of the following substances: water, Ge, Si, Bi?
1.2- From a consideration of the volume element in the mushy zone of figure 1.5 (upper part), define the local
solidification time, tf, in terms of the dendrite growth rate, V, and the length, a, of the mushy zone.
1.3 - Sketch -T diagrams for the crystallisation of a pure metal and for an alloy, and comment on their significance in
each case.
1.4 - Equiaxed dendrites develop freely in an undercooied melt, Discuss the direction of movement of the equiaxed
dendrites in a quiescent melt. Where would most of them be found in solidifying melts of a) steel, b) Bi?
1.5 - Sketch two different phase diagrams having a positive and a negative value of m. Show that the product, m(k-1),
is always positive.
1.6 - Using data from "Constitution of Binary Alloys" (M.Hansen and K.Anderko, McGraw-Hill, New York, 1958) or a
similar compilation, estimate the distribution coefficient, k, of S in Fe at temperatures between 1530 and 900C, and of
Cu in Ni at temperatures between 1400 and 1300C. Discuss the validity of the assumption that k is constant in these
systems.
1.7 - A molten alloy, like any liquid which exhibits local density variations, will be subject to movements known as
natural convection. What is the origin of this convection in a) a pure metal, b) an alloy? Discuss your conclusions with
regard to various alternative solidification processes such as upward (as opposed to downward) directional
solidification (Bridgman - figure 1.4), and casting (figure 1.5).
1.8 - Give possible reasons for the good mould-filling characteristics which are exhibited by pure metals and eutectic
alloys during the casting of small sections. Discuss them with regard to the interface morphology shown in figure 1.7.
1.9- Figure 1.4 illustrates two directional solidification processes which differ with respect to their heat transfer
characteristics. In one case, a steady-state behaviour is established after some transient changes. In the other ease,
changes continue to occur with the passage of time. One process is not limited with regard to the length of the
product, but is limited by its diameter. The other process is not affected by the diameter, but rather by the specimen
length. Sketch heat flux lines, and T, G, and V values as a function of z for various times and discuss the above facts
in the light of the sketches. Note that, in directional solidification, the temperature gradient at the solid/liquid interface
in the liquid must always be positive, as shown in figure 1.7a,c.
1.10 - Illustrate the changes in the temperature distribution of a casting as a function of time between the moment of
pouring of a pure, superheated melt and the establishment of the situation shown in figure 1.7a and b. Discuss the
fundamental differences between a and b.

2
EXERCISES

2-1 Show that setting the first derivative of equation 2.4 equal to zero leads to equation 2.1.
2-2 Compare the dA/dv ratio of spheres and cylinders with equation 1.5, and the curvature of the critical nucleus.
2-3 What is the meaning of the expression, d(G)/dr = 0, in terms of the forces acting upon the critical nucleus?
2-4 Develop an equation for G -n which is analogous to equation 2.4, and calculate Gn and n0.
2-5 Derive equation 2.5. What approximations are made? Why must caution be exercised when applying it to cases
involving high undercooling?
2-6 For small undercoolings, the nucleation rate given by equation 2.11 can be written in the form, I = K 1exp[-K2 /T2 ].
Develop expressions for K1 and K2 and explain the origin of the term, T2 . Determine how much K 1 and K2 must
change in order to double the value of T required to produce one nucleus m -3s-1 when K1 is originally equal to 10 37 m-
3 -1
s .
2-7 In his classic experiments, Turnbull divided a melt up into droplets which were only a few microns in diameter in
order to measure the temperature at which homogeneous nucleation occurred. Why did he use this method?.
2-8 By measuring the temperature required for homogeneous nucleation, it is possible to determine the solid/liquid
interface energy, which is difficult to measure using other methods. Develop the equations needed for this with the
aid of equation 2.11. In order to learn more about this technique, the reader should consult the papers of D.Turnbull:
Journal of Applied Physics 21 (1950) 1022 and J.H.Perepezko et al in: Solidification and Casting of Metals, The Metals
Society, London, 1979, p169.
2-9 - Show that n and G are functions of but not r. Why is this so? (See appendix 3, figure 3.7).
2-10 Determine the curvature and the melting point of a pure iron crystal in a wetted (=30) conical pore in the
crucible surface, as illustrated below.

2-11 What will happen if one stirs 0.1wt% of Ti powder into a pure Al melt at 700C 0 just before casting? (Hint: consult
the phase diagrami). For more details, see T.W.Clyne, M.H.Robert: Metals Technology 7 A980) 177.
2-12 Water and Bi expand during solidification. From this fact alone, one can make predictions concerning the entropy
of melting, and therefore the crystallisation behaviour, in terms of the faceted/non-faceted classification. What
reasoning would you employ?
2-13 Show graphically in two dimensions why rapidly growing planes disappear during crystal growth and leave the
slowest growing ones.
2-14 What is the difference in the shape of an equilibrium crystal, when V = 0 and is a minimum for {111}, and the
shape arising from growth when V>0 and is a minimum for <l 00> ?
2-15 Explain why an atomically smooth interface is usually microscopically rough when growing in a temperature

3
gradient (figure 2.8a - top).

EXERCISES

3.1 - A solid/liquid interface becomes unstable when relation 3.1 (3.9) is obeyed in the case of a pure metal (alloy).
Show that relation 3.1 is implied by relation 3.9. Discuss the differences.
3.2 - Indicate why the constitutional undercooling criterion cannot yield the wavelength of the perturbed interface
resulting from instability.
3.3 - Discuss the advantage of the Bridgman method (figure 1.4a), over conventional casting processes (figure 1.4b),
with respect to the control of interface stability.
3.4 - Determine the phase diagram from the information that the solute distribution ahead of the planar solid/liquid
interface of an Al-Cu alloy has been found to be described by:
Cl [wt%] = 2(1 + [0.86/0.14]exp[-Vz/D])
under steady-state conditions. It has also been determined that the interface temperature is 624, and it is known that
the melting point of Al is 660C. Give the values of k, m, T0 , Tl , and C0
3.5- What is the limit of stability, G/V, of the above alloy if it is given that D = 3x 10 -5cm2/s?
3.6 - Calculate the heat-flux required to produce a value of G which is sufficient to stabilise the planar front in
question 3.5 at a growth rate, V, of 10cm/h.
3.7- By analogy with figure 3.4, draw a diagram for k>1 and discuss the constitutional undercooling criterion for this
situation. Point out differences between this case and the case where k<l.
3.8 - Calculate Cl*, T*, and the minimum value of G required to stabilise the planar interface of a Cu-10wt%Ni and an
Fe-0.01wt%S (see exercise 1.6) alloy during directional growth at a rate of 0.01mm/s. Plot the solute concentration in
the boundary layer existing in each case. (See appendix 12 for the data required.)
3.9- Determine the difference between the critical G/V ratio estimated according to the constitutional undercooling
criterion and that estimated according to the Mullins-Sekerka criterion, for Al-2wt%Cu alloy (see appendix 6).
3.10- Calculate the ratio, m / i, for conditions of high constitutional undercooling and constant V and G values. Here,
m is the wavelength which corresponds to the maximum in figure 3.7. Note that b in equation 3.23 is a function of
and, in appendix 6, is defined as: b = (V/2D) + -{[(V/2D) 2 + 2]}1/2. Under certain conditions, this can be simplified by
comparing the relative magnitudes of V/2D and . In order to estimate the value of , use equation 3.25. Docs the
value of m / i. depend upon the composition of the alloy?
3.11 - Imagine that an experiment is carried out, on the alloy of question 3.4, in which the specimen is solidified using
the Bridgman method with a temperature gradient of G=10K/mm and is maintained at its limit of stability until steady-
state conditions are established. The growth rate is then doubled without changing the value of G. Calculate the
minimum, and the most probable, wavelength which the resultant perturbation is likely to have.
3.12- For the case of pure Al at an undercooling of 1K, determine the order of magnitude of the radius of a growing
sphere (R i.) at which the spherical crystal will become unstable when growing in its undercooled melt. Assume for
the purpose of calculating G that the steady-state solution for tlie temperature field around a sphere is applicable
(appendix 2).

4
EXERCISES

4.1 What will happen when the angle of the dendrite trunk axis in figure 4.4c is at exactly 45 to the heat flow
direction? Sketch a portion of such a dendrite growing under these conditions.
4.2 A microstructure such as that in figure 4.3 is being formed by directional solidification. The material is cubic and
the cube axis is 20 away from the axis of the cells. What will happen when the growth rate is markedly increased?
(Consider the transition from cells to dendrites depicted in figure 4.4)
4.3 What does monocrystalline mean in the case of figure 4.5 (no concentration variation in the solid, absence of low-
angle boundaries, absence of high-angle boundaries)? Compare with the columnar zone of a casting (figure 4.2).
4.4 Design a mould which is suitable for the production of a dendritic monocrystalline casting, e.g. a gas-turbine
blade.
4.5 Sketch a sequence of transverse sections of a dendrite which illustrates the region between the tip and the root of
the dendrite in figure 4.6.
4.6 In appendix 7, the solution (equation A7.10) for solute diffusion around a hemispherical needle crystal is
developed. Derive the equivalent solution for the case of heat diffusion (equation A7.12).
4.7 Imagine a dendrite tip which changes its radius during growth. What will happen to the concentration field of this
dendrite (figure 4.8 where T = Tc ) as the tip becomes sharper and sharper at a given growth rate. Note that T =
Tc + Tr in directional solidification. Sketch the concentration profile for the case where (a) T > Tr, (b) T = Tr,
(c) T < Tr. Which situation corresponds to the critical nucleation radius? Which dendrite grows and which melts?
4.8 Calculate the undercooling for the case of figure 4.9 which represents an Al-2wt%Cu under directional
solidification conditions. From the value of R0 determine the Gibbs-Thomson parameter, , used.
4.9 The unoptimised dendrite growth behaviour can be expressed as, V = f(R) (as in figure 4.10) or as T =g(R). Derive
an equation for T(R) on the basis of equation A8.3. Indicate the range of R values within which diffusion or capillarity
is governing growth.
4.10 In figure 4.11, the results of Trivedi's more exact model, based upon the Ivantsov solution, for directionally
solidifying dendrites are compared with those of the simple model derived in the present text. The discrepancy
between them is large at low and high growth rates. At both extremes, the Pclet number becomes large (P greater
than unity). Discover which of the simplifications made is responsible for the unrealistic predictions of the simple
model. (Consider the low-V part of the curve, where the dendrite tip changes to a plane front, and make use of figure
A7.4).
4.11 Calculate the tip temperature (table 4.2) of a dendritic growth front in Al-2%Cu alloy when V = 0.1mm/s and G =
10K/mm. Determine 1, 2, and the length of the mushy zone, assuming that G is constant in that region and the
melting point of the last liquid is Te. What is the value of the ratio, T'/T0?
4.12 By using the lower limit, C0/k, instead of Clm in equation 4.19, show how 2, varies with C0 in a given alloy system.
For many systems, this simplification is realistic.
4.13 In an Al-5wt%Si alloy casting, measurements of temperature and microstructure gave the results below. Compare
these values to the theoretical ones and estimate the cooling rate.

5
4.14 In experiments involving strong, uniform flow of the melt across the solid/liquid interface (for example the
electromagnetic stirring of steel during continuous casting), it is observed that columnar dendrites are inclined in a
direction which is opposite to the flow direction. How would you explain this observation? Consider the way in which
the boundary layers around the dendrite tip are changed. Would the same effect occur in pure metals?
4.15 Stirring of the melt during solidification is an efficient method of promoting the columnar-to-equiaxed transition
in a casting. The reason for this transition in stirred castings is that the melt becomes more rapidly cooled and, at the
same time, many dendrite branches are detached from the interface. With the aid of figure 4.1, indicate the two
limiting temperatures at which the melt must be in order to make this transition possible.

6
EXERCISES
5.1 Calculate the equilibrium volume fraction of carbon at Te in an Fe-C alloy of eutectic composition.
5.2 Calculate, on purely geometrical grounds, the volume fraction at which fibrous or lamellar structures have the
lower total / interface energy. Assume that the / interface energy is isotropic and that the phase separation, , is
equal and constant in both cases.
5.3 In figure 5.5b, the solute concentration in the liquid at the eutectic solid/liquid interface (z = 0) is given. Draw an
approximate C-y diagram for z= .
5.4 In order to demonstrate the potential value of producing 'in-situ' composites via directional eutectic solidification,
calculate the total length of fibres contained in a composite cube of side 1cm when = 1m.
5.5 An eutectic stores part of the transformation energy in the form of / interfaces (figure 5.4). By what amount, T,
will the melting point of a lamellar eutectic with = 1m and = 5x10-8 J/mm2 be lowered? As shown in appendix 12,
the magnitude of sf, is typically of the order of 106J/m3K.
5.6 Find expressions for Kc and Kr (equation 5.8) for the simple case shown in the main text, and derive the solutions
for extremum growth.
5.7 Using the phase diagram, explain the apparent discontinuity of the liquidus temperature, T l* at the /-junction
shown in figure 5.5c.
5.8 Draw analogous diagrams to those in figure 5.5 for the case where both phases have a positive curvature, and for
the case where both phases have depressions at their centers.
5.9 It can be easily shown that, at least over one half-spacing of the eutectic (figure 5.2), the solid/liquid interface of
the eutectic must be very close to isothermal: an enormous heat flux in the y-direction would be required in order to
change, even slightly, the interface temperature of the two phases. Using the properties of Al (appendix 12), calculate
this lateral heat flux, assuming that Ty = 0.1K and that = 1m.
5.10 Experiments performed on eutectic Fe-C alloys reveal the following relationship for the mean lamellar spacing,
: 2v = 4x10-7 mm3/s. Does this support the extremum criterion? Use equation A9.30.
5.11 Is an effect of G upon the spacing of regular eutectic microstructures at a given growth rate to be expected?
Explain your answer.
5.12 Draw a T-V diagram, similar to that in figure 5.13, for a symmetrical coupled zone under conditions where G>0.
5.13 Repeat exercise 5.12 for the case when G0. Compare your results with figure 5.12. (First study figure 4.13 in
detail).
5.14 Under constrained growth conditions (G>0) at low growth rates, coupled eutectic growth without the appearance
of dendrites is possible at off-eutectic compositions (figure 5.12). Calculate the limit of stability of a hypereutectic Al-
Al2Cu() alloy containing 36wt%Cu at which -dendrites plus eutectic will appear as in figure 5.11a. (Hint: use the
simple constitutional undercooling criterion and replace T0 by Tl Te ). To what line in figure 5.12 does this situation
correspond? What will happen if G is doubled?
5.15 From the growth equations for dendrites (when G0) and eutectic: Td = Kd(V)1/2 and Te = Ke(V)1/2, determine the
limiting growth rate and temperature of the coupled zone as a function of the composition. (Note that both growth
morphologies must grow at the same rate in order to exist side-by-side under steady-state conditions. Assume that

7
the constants, Kd and Ke, are independent of the composition and that Td. and Te are defined as shown in the figure
below.

5.16 Sketch the solid/liquid interface of a unidirectionally and dendritically solidifying steel containing 0.2wt%C. Note
that, due to the rapid solid-state diffusion of carbon in -Fe and -Fe, the liquid/solid and solid/solid transformations
closely follow the behaviour to be expected on the basis of the equilibrium diagram.

8
EXERCISES

6.1 Write an equation for the Tl*-fs relationship for the Lever rule and Scheil equation cases.
6.2 Determine ' values for Al-2wt%Cu and -Fe-0.09wt%C alloys when tf=10s. Which system exhibits the greater
tendency to segregate? (To obtain an estimate for 2 use figure 4.17 for both alloys)
6.3 It is desired to purify part of a cylindrical metal ingot by directional solidification. What conditions are most
favourable: a short initial transient or Scheil-type solidification? What interface morphology is required in order to
accomplish this? Give the maximum growth rate which can be used.
6.4 Devise a method for estimating part of the phase diagram (m,k values) of a transparent organic alloy by solidifying
it under planar interface conditions at the limit of stability and observing it with a microscope. Indicate how one might
perform the experiment (see figure 6.2). It is assumed that the values of C0 and D are known.
6.5 What will happen if the rod in figure 6.1 is solidified under conditions of strong convection? Sketch the solute
profile and indicate which equation applies to this situation.
6.6 In what respect is equation 6.3 an approximation to the initial transient? Examine the assumptions made
concerning the boundary layer for the transient (see appendix 1 0 and figure A2.4) .
6.7 Describe a method for the production of a control sample of a given composition for use in microprobe
measurements. Such a standard should present a composition to the electron beam which is homogeneous at a scale
of the order of 1 m.
6.8 Write the equation, given in this chapter, which describes approximately the concentration variation along the
curved interface of figure 6.6 under growth conditions where the tip concentration in the liquid, C l*, is approximately
equal to C0. In what region would one expect the concentration gradient, perpendicular to the solid/liquid interface, in
the liquid to be (a) close to zero and (b) not zero?
6.9 Indicate, with the aid of figure 4.12, the growth rates for which, in an Al-2wt%Cu alloy, no intercellular or
interdendritic enrichment (segregation) will occur for (a) G 0 and for (b) G = lK/mm.
6.10 Why is equation 6.9 incorrect (under the original assumption that u = 1 - 2k) If >l? What happens in this case?
Sketch the solute profiles present in the solid and liquid under these conditions. Compare with figure A11.2.
6.11 Compare the fractions of eutectic predicted by the Seheil equation, by the Lever rule, and by equation 6.9, in the
case of Fe-0.6wt%C and Al-2wt%Cu. Discuss the differences. For ', use the same conditions as those assumed in
exercise 6.2.
6.12 Zone melting is a process in which the tendency to segregation is exploited in order to produce a solid having a
high purity. In practice, this is done by causing a molten zone with planar solid/liquid interfaces to pass many times
through the material (figure). The equation for the impurity distribution after the first pass with strong mixing
occurring in the liquid zone (Cl= constant) is: Cl/C0 = (l/k)[l (1-k)exp(-kz/L)], where L is the zone length. Discuss this
equation with regard to the equation for the initial transient (equation 6.3)

9
10