Solidification and Metallurgical Microstructures

Introduction

Solidification is the transformation of a liquid metal into a solid crystalline material. It is central to casting, welding, additive manufacturing, brazing, soldering, and the production of ingots and semi-finished products. Although solidification may appear to be simply the removal of heat, it is in fact a coupled thermal, chemical, and structural process. The rate and direction of heat flow determine crystal growth; alloy composition determines phase formation and segregation; fluid flow influences defects and chemical uniformity.

The resulting microstructure—grain size, grain shape, phase distribution, segregation pattern, porosity, and inclusions—strongly affects strength, ductility, toughness, fatigue life, machinability, corrosion resistance, and dimensional stability. Consequently, a component cannot be designed reliably from chemical composition alone. Its processing history must also be considered.

Learning Objectives

By the end of this chapter, the student should be able to:

1. Explain thermodynamic and kinetic requirements for nucleation.
2. Distinguish homogeneous and heterogeneous nucleation.
3. Describe crystal growth and dendritic solidification.
4. Relate cooling rate, thermal gradient, and solidification time to grain structure.
5. Explain grain refinement, columnar and equiaxed structures, and constitutional supercooling.
6. Describe microsegregation, macrosegregation, and phase formation in alloys.
7. Identify common casting defects, their causes, and preventive measures.
8. Connect solidification variables with mechanical and functional properties.
9. Apply solidification principles to the analysis of a practical casting problem.

1. Thermodynamics of Solidification

A pure metal normally solidifies at a definite equilibrium melting temperature, Tₘ. At temperatures above Tₘ, the liquid has the lower Gibbs free energy; below Tₘ, the solid has the lower Gibbs free energy. At Tₘ, the two phases have equal free energy.

For solidification to begin in practice, a small degree of undercooling is usually required. Undercooling, or supercooling, is the temperature difference between the equilibrium melting temperature and the actual liquid temperature:

Here, ∆T is undercooling in kelvin or degrees Celsius, Tₘ is the equilibrium melting temperature, and T is the actual liquid temperature. This expression applies when the equilibrium melting point is known and the temperature scale difference is used; temperature differences in kelvin and Celsius are numerically equal.

Undercooling supplies a thermodynamic driving force for forming solid. However, the formation of a solid particle also creates an interface between solid and liquid. Interface formation requires positive surface energy. Thus, solidification involves competition between:

• a negative bulk free-energy change, which favors solid formation; and
• a positive interfacial-energy change, which opposes formation of very small particles.

A small solid particle may therefore remelt, whereas a particle larger than a critical size can grow spontaneously. For a spherical nucleus, the change in Gibbs free energy may be represented as:

where r is nucleus radius, γ is solid–liquid interfacial energy per unit area, and ∆Gᵥ is the volumetric free-energy change. During undercooled solidification, ∆Gᵥ is negative. The first term is positive and the second negative.

Differentiating with respect to r and setting the result to zero gives the critical radius:

A nucleus with r < r is unstable and tends to dissolve; a nucleus with r > r is stable and can grow. The equation assumes a spherical nucleus, uniform interfacial energy, and a homogeneous liquid. Real castings often nucleate heterogeneously on mould walls, inclusions, or grain-refining particles, so the ideal equation is mainly a conceptual and comparative model.

For modest undercooling, the driving force is often approximated by:

where ∆H_f is molar latent heat of fusion, Vₘ is molar volume, and absolute temperatures are used. This approximation assumes small undercooling and approximately constant thermodynamic properties.

2. Nucleation

Nucleation is the formation of stable, microscopic solid particles in a liquid. It determines the number of crystals available for subsequent growth and is therefore a major control on grain size.

Homogeneous nucleation

Homogeneous nucleation occurs throughout the liquid without preferential sites. It requires relatively large undercooling because the new solid must create a complete solid–liquid interface. It is uncommon in ordinary industrial castings because mould walls, oxide films, inclusions, and impurities provide easier nucleation sites.

Heterogeneous nucleation

Heterogeneous nucleation occurs on a foreign surface. The substrate reduces the interfacial-energy penalty, lowering the required critical undercooling. Effective substrates generally have suitable crystallographic matching, chemical stability, and good wetting by the liquid.

Examples include:

• mould walls, which initiate surface grains;
• inoculant particles in cast irons;
• grain refiners such as titanium- or boron-containing additions in aluminium alloys; and
• oxide or nitride particles, which may either refine grains or become harmful inclusions.

The usefulness of an inoculant depends not only on its presence but also on particle size, distribution, survival during holding, and chemical compatibility. Excessive additions can increase cost, sludge formation, or inclusions.

3. Crystal Growth and Interface Stability

Once a stable nucleus forms, atoms transfer from liquid to the solid–liquid interface and become incorporated into the crystal lattice. Growth requires both heat removal and atomic rearrangement. The interface velocity depends on thermal undercooling, solute concentration, diffusion, convection, and crystallographic orientation.

A planar interface can remain stable when the liquid ahead of it is thermally and constitutionally stable. In alloys, however, the growing solid often rejects solute into the liquid. The liquid immediately ahead of the interface may then have a lower local melting temperature than the bulk alloy. If the actual temperature gradient is insufficient, the liquid can become constitutionally supercooled. Small disturbances at the interface grow preferentially, producing cells or dendrites rather than a flat front.

Two variables are especially important:

• Thermal gradient, G: temperature change per unit distance ahead of the interface, commonly expressed in K/m.
• Interface growth velocity, R: speed at which the solid–liquid boundary advances, commonly expressed in m/s.

The ratio G/R is associated with interface stability and morphology. A high G/R generally favours a more stable, directional interface; a low G/R favours cellular or dendritic growth. The product G×R is related to the rate of cooling experienced by the advancing interface. These are qualitative process relationships rather than universal grain-size equations, because alloy system, convection, nucleation density, and crystallography also matter.

4. Dendritic Solidification

Dendrites are tree-like crystals with a primary stem and secondary and tertiary branches. They are the dominant solidification form in many engineering alloys. The primary arms grow approximately along crystallographic directions that provide favourable atomic attachment. Solute rejection and heat flow control the development of side branches.

Dendritic growth produces several important regions:

1. Dendrite cores: relatively early-formed solid, often compositionally different from the interdendritic liquid.
2. Interdendritic liquid: liquid trapped between growing arms and enriched in elements rejected by the solid.
3. Last-to-freeze regions: sites where shrinkage porosity, low-melting constituents, and eutectic phases may concentrate.

The average spacing between primary or secondary dendrite arms is called dendrite arm spacing. Finer spacing generally results from faster cooling. A commonly used empirical relationship is:

where λ₂ is secondary dendrite arm spacing, t_f is local solidification time, K is an alloy- and process-dependent constant, and n is an empirically determined exponent. Units of K must be selected consistently with the units of λ₂ and t_f. This is not a universal law; it must be calibrated for the alloy and thermal range of interest.

Fine dendrite spacing usually improves chemical homogeneity after homogenization, reduces the dimensions of interdendritic defects, and can improve strength and fatigue performance. It may also increase resistance to hot tearing in some systems, although the effect depends on feeding, alloy composition, and grain morphology.

5. Grain Formation and Grain Structure

A grain is a region in which the crystal lattice has a continuous orientation. Grain boundaries separate crystals of different orientation. Grain structure develops through the competition between nucleation and growth.

Chill zone

At the mould wall, rapid heat extraction produces many nuclei. A thin region of fine, randomly oriented grains, called the chill zone, may form. It is especially prominent when the mould wall is cold and well wetted by the liquid.

Columnar grains

Some grains whose growth directions are favourably aligned with the heat-flow direction outgrow their neighbours. They develop into elongated columnar grains. Columnar growth is common in ingots and castings with strong directional cooling.

Equiaxed grains

Equiaxed grains have approximately similar dimensions in different directions. They form when numerous nuclei survive in the liquid or when fragments of dendrites are transported into the bulk liquid. Equiaxed structures generally reduce directional anisotropy and may reduce the risk of certain hot-tearing patterns.

The transition between columnar and equiaxed growth depends on nucleation density, temperature gradient, growth velocity, convection, alloy composition, and mould geometry. Grain refinement increases the number of nuclei and tends to promote a finer equiaxed structure.

Grain size influences properties through grain-boundary strengthening. For many metals, the Hall–Petch relationship is written as:

where σᵧ is yield stress, σ₀ is the friction stress of the lattice, kᵧ is a material constant, and d is average grain diameter. The units of kᵧ must be consistent with stress and the square root of length. This relationship is an empirical approximation, most reliable over a range where ordinary dislocation slip and reasonably uniform grains control yielding. Very fine, ultrafine, or highly textured materials may not follow it closely.

Fine grains often improve yield strength, toughness, and isotropy. However, grain refinement does not automatically eliminate porosity, inclusions, segregation, or residual stress. Microstructural features must therefore be considered together.

6. Solidification of Alloys and Phase Formation

Pure metals solidify at nearly constant temperature. Most engineering alloys solidify over a temperature interval between the liquidus and solidus temperatures. The liquidus is the temperature above which the alloy is completely liquid; the solidus is the temperature below which it is completely solid under equilibrium conditions.

During non-equilibrium solidification, the first solid may have a composition different from the original liquid. As solid grows, solute is partitioned between solid and liquid. The partition coefficient is:

where C_s is solute concentration in the newly formed solid at the interface and C_l is solute concentration in the adjacent liquid. The definition applies locally at the interface and assumes concentrations are expressed on the same basis, such as mass fraction or mole fraction.

If k < 1, the solute preferentially remains in the liquid. The residual liquid becomes progressively enriched, and the final solidification regions contain higher solute concentration. If k > 1, the early solid is enriched and the remaining liquid may become depleted.

Microsegregation

Microsegregation occurs over dendrite-arm or grain dimensions. It results from solute partitioning and incomplete diffusion during solidification. Homogenization heat treatment can reduce it by solid-state diffusion, but complete removal may be impossible if low-melting phases form or if diffusion distances are large.

Macrosegregation

Macrosegregation occurs over centimetre- or component-scale distances. It is caused by movement of solute-enriched liquid, sedimentation or flotation of crystals, thermal convection, and contraction-driven flow. Unlike microsegregation, it is difficult to remove by heat treatment because the composition varies over large distances.

Segregation can produce local differences in hardness, corrosion resistance, weldability, transformation behaviour, and fracture resistance. It is particularly important in large steel ingots, aluminium billets, and continuously cast products.

7. Heat Flow and Solidification Time

Heat flows from the liquid metal through the solidifying shell, mould, and surroundings. The local cooling rate may be expressed as:

where Ṫ is cooling rate in K/s and T is temperature as a function of time. The sign depends on the chosen convention; cooling rate is often reported as a positive magnitude, |dT/dt|.

A simplified one-dimensional heat-conduction relation is Fourier’s law:

where q is heat-transfer rate in watts, k_t is thermal conductivity in W/(m·K), A is cross-sectional area in m², and dT/dx is the temperature gradient. The negative sign indicates that heat flows toward decreasing temperature. The relation assumes continuum conduction and is local; mould–metal contact resistance, radiation, convection, and latent heat may also be important.

Increasing cooling intensity generally causes:

• faster solidification;
• finer dendrite arm spacing;
• finer grains if nucleation is adequate;
• lower segregation length scales; and
• greater thermal gradients and possible residual stress.

Cooling too rapidly can increase distortion, cracking, or incomplete feeding. Cooling too slowly may produce coarse grains, large dendrite spacing, severe segregation, and prolonged exposure to hot-tearing conditions. The objective is controlled rather than simply maximum cooling.

8. Feeding, Shrinkage, and Casting Defects

Most metals contract when cooling from liquid to solid and then continue to contract in the solid state. If liquid metal cannot flow into the regions undergoing solidification shrinkage, internal cavities or dispersed microporosity develop. A riser, also called a feeder, is designed to remain liquid longer than the casting and supply metal during solidification.

Common defects

Gas porosity and shrinkage porosity may appear similar in radiographs, but their distributions and causes differ. Gas pores are often rounded, whereas shrinkage is commonly irregular and concentrated in interdendritic or hot-spot regions. Metallographic examination, density measurements, computed tomography, and process records help distinguish them.

Hot tearing

Hot tearing occurs near the end of solidification when a partially coherent solid network is unable to accommodate tensile strain caused by contraction or mould restraint. Risk is increased by a wide freezing range, low-melting films at grain boundaries, poor liquid feeding, coarse columnar grains, and unfavourable geometry. Prevention may involve alloy modification, grain refinement, improved feeding, rounded transitions, and reduced restraint.

Mould filling and turbulence

Turbulent flow can entrain oxide films and air, producing defects that act as fatigue-crack initiators. A well-designed gating system promotes controlled filling, avoids sharp changes in flow direction, and limits free-fall and vortex formation. Pouring temperature must be high enough for complete filling but not so high that oxidation, gas absorption, mould reaction, and grain coarsening become excessive.

9. Process–Microstructure–Property Relationships

The central engineering chain is:

Process variables → thermal and chemical history → microstructure → properties → performance.

Examples include:

• Higher cooling rate → finer grains and dendrite spacing → generally higher strength and improved uniformity.
• Strong directional heat flow → columnar grains → anisotropic mechanical and thermal behaviour.
• Poor feeding → shrinkage porosity → reduced tensile, fatigue, and pressure-tight performance.
• Solute enrichment in interdendritic regions → low-melting phases or local hardness variation → impaired ductility or corrosion resistance.
• Grain-refiner addition → more nuclei and finer equiaxed grains → improved isotropy, though possible inclusions or sludge must be controlled.

The relationship is not deterministic from one variable alone. For example, a high cooling rate produces fine structure only if enough nucleation sites exist. Likewise, a fine grain structure cannot compensate for severe porosity. Quality assessment should therefore combine thermal records, chemical analysis, metallography, non-destructive testing, and mechanical testing.

10. Worked Example: Estimating Dendrite Spacing

An aluminium alloy has an experimentally determined relationship λ₂ = 35 t_f^0.33, where λ₂ is in micrometres and t_f is in seconds. Estimate the secondary dendrite arm spacing when local solidification time is 64 s.

Using the calibrated relationship:

Since 64^0.33 is approximately 3.98:

The estimated spacing is approximately 139 μm. This result is valid only for the alloy, range, thermal conditions, and units used to establish the empirical constants. It should not be transferred uncritically to another alloy or casting process.

11. Applied Case: Porosity in an Aluminium Housing

Background

Consider an illustrative industrial case involving a gravity-cast aluminium housing intended to contain hydraulic fluid. Radiographic inspection reveals irregular internal pores near a thick boss. Leak testing fails intermittently, and tensile specimens taken from the boss show lower ductility than specimens from thin sections.

Problem analysis

The boss is a thermal hot spot: it has a larger local modulus and loses heat more slowly than surrounding regions. Solidification therefore progresses from the thin walls toward the boss. As the boss freezes, liquid feeding becomes increasingly difficult. Solute enrichment in the interdendritic liquid further lowers its local freezing temperature, prolonging the weak mushy stage. The irregular pores and location near the hot spot indicate shrinkage microporosity rather than uniformly distributed gas porosity.

Approach and solution

The foundry team first maps solidification using thermocouples and simulation. The gating and riser arrangement is then modified so that the riser remains liquid longer and feeds the boss. A local chill is added to increase heat extraction at the hot spot, while the boss-to-wall transition is redesigned with a smoother geometry. Melt cleanliness is improved through filtration and degassing, ensuring that gas porosity is not an additional source of leakage. Pouring temperature is reduced within the range that still provides complete mould filling.

Outcome

Radiographic inspection shows a substantial reduction in concentrated porosity, leak-test failures decline, and ductility becomes more uniform. The case demonstrates that defect control requires coordinated changes in geometry, feeding, cooling, and melt quality rather than reliance on a single parameter.

Learning points

1. Thick sections govern local solidification time and often determine casting quality.
2. Riser design and directional solidification are central to shrinkage control.
3. Chills can be useful, but excessive thermal gradients may increase cracking or distortion.
4. Simulation should be validated with temperature measurements and inspection data.
5. Microstructure and defect location provide evidence about the underlying process.

12. Industrial Measurement and Contemporary Practice

Modern foundries combine thermal analysis, computational fluid dynamics, solidification simulation, optical emission spectroscopy, metallography, radiography, ultrasonic testing, and computed tomography. Process data can be used to establish links between pouring temperature, mould temperature, filling time, cooling curves, and defect probability.

Machine-learning tools are increasingly used to classify radiographic or tomographic defects and to optimize process windows. Their predictions are useful only when training data are representative, labels are reliable, and the model is validated on unseen production conditions. Explainability, traceability, and human engineering review remain important because a statistical correlation does not establish a physical cause.

Sustainable solidification practice includes reducing remelt, increasing yield, controlling alloying additions, recycling clean scrap, and minimizing energy used for overheating and subsequent heat treatment. Environmental and occupational controls are also essential because melting operations may involve fumes, intense radiation, noise, and reactive additives.

13. Best Practices for Solidification Control

1. Specify composition limits and control tramp elements.
2. Maintain clean, dry charge materials and tools.
3. Measure rather than assume melt and mould temperatures.
4. Use grain refinement and inoculation only at controlled additions.
5. Design gating for calm, complete filling with minimal entrainment.
6. Promote directional solidification and provide adequate feeding.
7. Use chills and insulation strategically to control hot spots.
8. Validate simulation using thermocouples and inspection results.
9. Characterize grain size, dendrite spacing, phases, porosity, and segregation.
10. Link acceptance criteria to service requirements such as fatigue, pressure tightness, and fracture toughness.

Challenges and Limitations

Solidification models often simplify convection, interfacial kinetics, nucleation, mould contact resistance, and three-dimensional geometry. Alloy databases may be incomplete, and empirical constants can vary with cleanliness and processing history. A predicted microstructure is therefore not a substitute for production validation. In addition, improving one characteristic may worsen another: rapid cooling may refine grains but increase stress; high pouring temperature may improve filling but increase gas absorption; grain-refiner additions may reduce grain size but introduce unwanted particles.

Chapter Summary

Solidification is governed by thermodynamic driving force, interfacial energy, heat flow, solute redistribution, and interface kinetics. Nucleation establishes the number of crystals, while growth determines their shape and orientation. Heterogeneous nucleation dominates most industrial processes because mould walls and particles reduce the energy barrier.

Alloy solidification commonly produces cellular and dendritic structures. Solute partitioning generates microsegregation and, when fluid movement occurs over large distances, macrosegregation. Thermal gradient and interface velocity influence columnar or equiaxed growth, while cooling rate controls grain size and dendrite spacing. Grain refinement can improve strength and isotropy, but it cannot by itself remove porosity or severe segregation.

Casting defects—including gas porosity, shrinkage, hot tearing, cold shuts, misruns, inclusions, and segregation—are consequences of interactions among alloy composition, mould design, filling, heat extraction, feeding, and restraint. Effective control therefore requires a process–microstructure–property approach supported by measurement and validation.

Key Terms

Nucleation: Formation of stable solid particles in a liquid.

Undercooling: Difference between equilibrium melting temperature and actual liquid temperature.

Critical radius: Minimum radius required for a nucleus to grow spontaneously.

Homogeneous nucleation: Nucleation occurring without preferential foreign surfaces.

Heterogeneous nucleation: Nucleation assisted by a mould wall, inclusion, or inoculant.

Interfacial energy: Energy associated with the boundary between two phases.

Dendrite: Tree-like crystal formed during unstable alloy solidification.

Dendrite arm spacing: Average distance between dendrite branches.

Thermal gradient: Temperature change per unit distance.

Growth velocity: Speed of an advancing solid–liquid interface.

Columnar grain: Elongated grain growing preferentially along the heat-flow direction.

Equiaxed grain: Grain with approximately equal dimensions in different directions.

Constitutional supercooling: Instability caused by solute-induced reduction of liquidus temperature ahead of the interface.

Microsegregation: Composition variation over grain or dendrite dimensions.

Macrosegregation: Composition variation over component-scale distances.

Riser: Reservoir of liquid metal used to feed solidification shrinkage.

Hot tearing: Cracking of a weak, partially solidified network under tensile strain.

Gas porosity: Voids formed by gas evolution or entrapment.

Grain refinement: Process of increasing nucleation density to produce smaller grains.

Solidification time: Time required for a local region to pass through its solidification range.

Review Questions

Short Answer Questions

1. Why is undercooling generally required for practical solidification?
2. Distinguish homogeneous from heterogeneous nucleation.
3. What is the role of solute partitioning in dendritic solidification?
4. Define microsegregation and macrosegregation.
5. Why are risers used in castings?

Long Answer Questions

1. Explain the formation and growth of a stable nucleus using the concepts of bulk free energy, interfacial energy, and critical radius.
2. Describe how thermal gradient and interface velocity influence planar, cellular, columnar, and equiaxed growth.
3. Discuss the relationship between cooling rate, grain size, dendrite arm spacing, and mechanical properties.
4. Compare gas porosity, shrinkage porosity, hot tearing, and oxide inclusions in terms of mechanism and prevention.
5. Explain how alloy composition and non-equilibrium solidification produce segregation and how processing can reduce its effects.

Analytical / Application Questions

1. A casting contains irregular porosity concentrated beneath a thick section, while thin sections are sound. Identify the most probable defect mechanism and propose a coordinated corrective plan.
2. A process change doubles local cooling rate but produces more distortion and cracking. Analyze the competing microstructural and thermal effects and recommend additional controls.
3. A component requires high fatigue strength and pressure tightness. Develop a solidification-control strategy covering alloy cleanliness, grain structure, filling, feeding, inspection, and validation.

Multiple Choice Questions with Answers

Question 1

What is the primary thermodynamic effect of undercooling during solidification?

A. It increases the free-energy advantage of the solid phase

B. It eliminates solid–liquid interfacial energy

C. It prevents heterogeneous nucleation

D. It raises the equilibrium melting temperature

Correct Answer: A. It increases the free-energy advantage of the solid phase

Question 2

A nucleus smaller than the critical radius will generally:

A. Grow rapidly into a grain

B. Dissolve or remelt

C. Become a riser

D. Cause macrosegregation

Correct Answer: B. Dissolve or remelt

Question 3

Which condition generally favours dendritic growth in an alloy?

A. Low constitutional undercooling

B. A low G/R ratio

C. No solute rejection

D. An infinitely high thermal gradient

Correct Answer: B. A low G/R ratio

Question 4

If the partition coefficient k is less than one, the rejected solute tends to:

A. Enrich the remaining liquid

B. Disappear from both phases

C. Enrich only the mould wall

D. Prevent all grain growth

Correct Answer: A. Enrich the remaining liquid

Question 5

Which feature is most characteristic of a columnar grain structure?

A. Randomly oriented, equiaxed grains throughout the casting

B. Elongated grains aligned with the preferred heat-flow direction

C. Gas pores aligned with the riser

D. A completely amorphous structure

Correct Answer: B. Elongated grains aligned with the preferred heat-flow direction

Question 6

The main purpose of a riser is to:

A. Increase oxide formation

B. Supply liquid metal during solidification shrinkage

C. Reduce nucleation density

D. Create a cold shut

Correct Answer: B. Supply liquid metal during solidification shrinkage

Question 7

Which defect is most strongly associated with turbulent metal flow and oxide-film entrainment?

A. Oxide inclusion

B. Equiaxed grain

C. Homogeneous nucleation

D. Grain refinement

Correct Answer: A. Oxide inclusion

Question 8

Using λ₂ = 35t_f^0.33 with t_f = 64 s, the estimated dendrite arm spacing is closest to:

A. 35 μm

B. 64 μm

C. 139 μm

D. 2,240 μm

Correct Answer: C. 139 μm

Question 9

According to the Hall–Petch relationship, decreasing average grain diameter generally causes yield strength to:

A. Decrease because boundaries vanish

B. Increase, within the relationship’s applicable range

C. Remain exactly unchanged

D. Become independent of composition

Correct Answer: B. Increase, within the relationship’s applicable range

Question 10

Which intervention most directly promotes directional feeding of a thick casting section?

A. Removing all risers

B. Using a properly positioned feeder and thermal control

C. Increasing turbulence during filling

D. Adding moisture to the mould

Correct Answer: B. Using a properly positioned feeder and thermal control