Volume 09 - Metallography and Microstructures Part 14 pot

Major structural features, listed generally in increasing size, are: • Atomic structure: nuclei, atoms • Electronic structure • Crystal structure: perfect crystals, crystal imperfectio

Fig 26 Graphite-epoxy composite (Thornel 300 graphite fibers in 5208 epoxy matrix), unidirectional Standard

autoclave cure of lay-up made from prepreg tape Scanning electron micrograph of a failed tensile surface, which is tiered, as is typical of this particular graphite epoxy Considerable resin adheres to the fibers, indicating good interfacial bond strength 500× (P.R Lee)

Fig 27 Graphite-PPS composite (AS4 graphite fibers in thermoplastic polyphenylene-sulfide matrix), film

stacked and hot pressed Scanning electron micrograph of a failed tensile surface The high ductility of the thermoplastic matrix results in the "tails" of drawn matrix materials shown 500× (A.C Lou)

Fig 28 Graphite-Al composite (Thornel 50 fibers in 6061 Al matrix), unidirectional Liquid-metal infiltration and

consolidation in a liquid-phase press SEM of a failed surface The older version (~1973) of this graphite fiber, with its crenelated cross section, has debonded and pulled out of the Al matrix Pull-out holes are also seen 1000× (L.W Davis)

Fig 29 B-Al composite (25% B fibers in 6061 Al matrix), unidirectional Fiber on foil and diffusion bonded SEM

of a fairly flat, failed tensile surface, characteristic of this material Little matrix adheres to the fiber surface, indicating fairly low interfacial bond strength The Al matrix shows good ductility The tungsten cores of the vapor-deposited fibers are evident 105× (R Moss)

Fig 30 Graphite-silver copper composite (Thornel 300 fiber in 70Ag-30Cu eutectic matrix), unidirectional

Liquid-metal infiltration of fiber bundles followed by diffusion bonding Scanning electron micrograph of a failed tensile surface The matrix shows good ductility, but the lack of matrix adhering to the fibers (~1977 version) indicates low interfacial bond strength 3000× (W.C Harrigan)

Fig 31 Silicon carbide-glass ceramic composite (unidirectional silicon carbide fibers in glass ceramic matrix)

Ply lay-up and hot press densification Scanning electron micrograph of a failed tensile surface Although the ceramic matrix has failed in a brittle manner, the long pull-out length of the fibers indicates high composite toughness 20× (K.M Prewo)

Introduction to Structures in Metals

Michael B Bever, Professor of Materials Science and Engineering, Emeritus, Massachusetts Institute of Technology

Introduction

FOR MORE THAN A CENTURY, dating back to the pioneering contributions of Henry Clifton Sorby, metallurgists have not been satisfied merely to describe their metallographic observations, but have striven to explain them and to understand their implications (Ref 1, 2, 3, 4) In addition, new techniques of structural investigation have yielded new observations and posed new problems The quest for meaningful and precise explanations of metallurgical structures has been the primary driving force in the development of the science of physical metallurgy (Ref 5, 6, 7, 8) Physical metallurgy now comprises a very broad spectrum That portion of the spectrum dealing with the structure of metals is the subject of the articles in this Section

This article will develop the sequence in which the articles in this Section are presented and establish some connections among them, provide background for the subject matter explored more fully in the specialized articles, and furnish general references It will also treat important topics, such as grain structure and substructure, that are not covered systematically and comprehensively in the other articles Finally, it will describe the scale of structural features and introduce the concept of hierarchical relations among them

The term structure, as used here, refers primarily to the study of those microstructural features that can be investigated using optical (light) and electron microscopy (Ref 9, 10, 11, 12, 13, 14, 15, 16) The results of investigations using other techniques, such as x-ray diffraction, are included when pertinent (Ref 17, 18) Macrostructural features, which can be observed with little or no magnification, will also be considered

The purpose of the articles in this Section is to assist in the interpretation of microstructure Such interpretation requires

an understanding of the processes by which various structures are formed; therefore, the articles are organized according

to the major processes that produce characteristic structures A special article describes textures that can result from several of these processes

The principles applicable to various types of structures are illustrated by micrographs in the respective articles; references are also made to micrographs that appear in the Sections "Metallographic Techniques and Microstructures: Specific Metals and Alloys" in this Volume Several works that treat the interpretation of microstructures systematically are cited

1 R.F Mehl, A Brief History of the Science of Metals, American Institute of Mining and Metallurgical

Engineers, Warrendale, PA, 1948

2 C.S Smith, A History of Metallography, University of Chicago Press, 1960

3 C.S Smith, Ed., Sorby Centennial Symposium on the History of Metallurgy, Gordon and Breach, 1965

4 R.F Mehl and R.W Cahn, The Historical Development of Physical Metallurgy, in Physical Metallurgy,

Part I, 3rd ed., R.W Cahn and R Haasen, Ed., North-Holland, 1983, p 1-35

5 R.W Cahn and P Haasen, Ed., Physical Metallurgy, Parts I and II, 3rd ed., North-Holland, 1983

6 A.G Guy and J.J Hren, Elements of Physical Metallurgy, 3rd ed., Addison-Wesley, 1974

7 W.F Smith, Structures and Properties of Engineering Alloys, McGraw-Hill, 1981

8 R.E Smallman, Modern Physical Metallurgy, 4th ed., Butterworths, 1985

9 R.H Greaves and H Wrighton, Practical Microscopical Metallography, 4th ed., Chapman & Hall, 1957

10 H Gleiter, Microstructure, in Physical Metallurgy, Part I, 3rd ed., R.W Cahn and P Haasen, Ed,,

North-Holland, 1983, P 650-712

11 G.F Vander Voort, Metallography: Principles and Practice, McGraw-Hill Book Co., 1984

12 W Rostoker and J.R Dvorak, Interpretation of Metallographic Structures, 2nd ed., Academic Press, 1977

13 J.W Edington, Practical Electron Microscopy in Materials Science, Van Nostrand Reinhold, 1976

14 P.J Goodhew, Electron Microscopy and Analysis, Wykeham Publications, 1975

15 M.H Loretto and R.E Smallman, Defect Analysis in Electron Microscopy, Chapman &

Hall Halsted/Wiley, 1975

16 G Thomas and M.J Goringe, Transmission Electron Microscopy of Materials, John Wiley & Sons, 1979

17 C.S Barrett and T.B Massalski, Structure of Metals, 3rd ed., Pergamon Press, 1980

18 B.D Cullity, Elements of X-ray Diffraction, 2nd ed., Addison-Wesley, 1978

General Features of Structure

The structure of metals comprises features of various magnitudes Major structural features, listed generally in increasing size, are:

• Atomic structure: nuclei, atoms

• Electronic structure

• Crystal structure: perfect crystals, crystal imperfections

• Substructure: subgrains, other cellular structures

• Microstructure: grains of single-phase metals and alloys, shapes and sizes of micro-constituents and

their configurational arrangements in multiphase systems

• Textures

• Structural features related to composition

Crystal imperfections include point defects, such as impurity atoms, vacancies and vacancy aggregates, and interstitial atoms; line defects (dislocations); and area defects, for example, stacking faults, twin interfaces, subboundaries, and grain boundaries They are described in specialized texts on the theory of dislocations and other crystal imperfections (Ref 24, 25, 26)

Examples of various crystal defects are presented throughout this Volume In the article "Transmission Electron Microscopy," dislocations (Fig 21 to 24 and 26 to 29), dislocation dipoles (Fig 30), dislocation net-works (Fig 31 to 33), and dislocation loops (Fig 34 to 36) are shown Dislocations are also shown in Fig 3 to 6, 11, and 12 in the article

"Solidification Structures of Pure Metals" in this Section Stacking faults are shown in Fig 44 to 47 in the article

"Transmission Electron Microscopy."

Subgrains and cellular structures are formed by subboundaries (low-angle boundaries) The simplest of these boundaries consists of periodically spaced dislocations In more complex instances, particularly in structures resulting from deformation, dislocation tangles can form cellular structures Crystal imperfections of all kinds, including subboundaries, may occur in single crystals and within the grains of polycrystalline metals

Grain structure of single-phase polycrystalline metals, which is the most characteristic feature of their microstructure, will be discussed below

Twins, which occur within grains, are special imperfections that may originate during growth processes, for example, the annealing of cold-worked metal, or during deformation

Antiphase domain boundaries occur in solid solutions with long-range order, reducing the perfection of the order

Ferromagnetic domains are characteristic of ferromagnetic materials, as described in the article "Magnetic and Electrical Materials" in this Volume Unlike typical metallurgical processes, a change in ferromagnetic domain structure requires a variation in magnetic field Antiferromagnets also have domain structures

Multiphase Structures. As discussed below, the shapes, sizes, and configurational arrangements of two or more

microconstituents in a multiphase system produce a variety of typical microstructures

Textures combine the crystallographic feature of lattice orientation with the microstructural feature of grain structure In

a metal having a texture, or preferred orientation, the crystal lattices of the grain are arranged in a correlated and organized manner

Chemical composition affects structure through its influence on phase relations Composition is also involved in such structural features as microsegregation in solidified metals and solute-enriched regions at grain boundaries and other crystal imperfections (Ref 27)

Structural gradients reflect changes of structural features with position For example, a plate can have a grain-size gradient from the surface toward the interior Composition gradients can cause structural gradients, as in case-hardened metals Composites present special opportunities for establishing structural gradients by controlling the spatial arrangements of the reinforcing phase for example, fibers (Ref 28) Additional information is provided in the article

"Fiber Composite Materials" in this Volume

Porosity and voids are structural features that are characterized by a large range of sizes

Macrostructure is discussed below and is also considered in the articles "Solidification Structures of Steel,"

"Solidification Structures of Aluminum Alloy Ingots," "Solidification Structures of Copper Alloy Ingots," and "Plastic Deformation Structures" in this Section

References cited in this section

19 H.W King, Structure of the Pure Metals, in Physical Metallurgy, Part I, 3rd ed., R.W Cahn and P Haasen,

Ed., North-Holland, 1983, P 37-79

20 D.G Pettifor, Electron Theory of Metals, in Physical Metallurgy, Part I, 3rd ed., R.W Cahn and P Haasen,

Ed., North-Holland, 1983, p 73-152

21 W.A Harrison, Electronic Structure and the Properties of Solids, Freeman, 1980

22 A Kelly and G.W Groves, Crystallography and Crystal Defects, Addison-Wesley, 1970

23 E Prince Mathematical Techniques in Crystallography and Materials Science, Springer-Verlag, 1982

24 H.G van Bueren, Imperfections in Crystals, North-Holland, 1960

25 J.P Hirth and J Lothe, Theory of Dislocations, 2nd ed., John Wiley & Sons, 1982

26 D Hull, Introduction to Dislocations, 3rd ed., Pergamon Press, 1975

27 R.W Balluffi, Grain Boundary Structure and Segregation, in Interfacial Segregation, W.C Johnson and

J.M Blakely, Ed., American Society for Metals, 1979, p 193-236

28 M.B Bever and P.E Duwez, Gradients in Composite Materials, Mater Sci Eng., Vol 10, 1972, p 1-8

Origins of Structures

The characteristic structures of metals and alloys are produced by (1) transformations in which one or more parent phases are converted into one or more new phases, (2) deformation processes, (3) thermal processes, (4) thermomechanical processes, or (5) diffusion processes that do not result in a transformation, such as sintering A typical deformation process is cold working Examples of thermal processes are the annealing of a coldworked metal and the homogenization

of an alloy with microsegregation, The principles underlying and governing these processes are the province of physical metallurgy (see Ref 5, 6, 7, 8, 29, 30, 31)

The transformations and processes that result in the production of typical structures involve characteristic basic mechanisms The transformations that produce solidification structures and solid-state transformation structures involve several such mechanisms The most important of these are diffusion, nucleation, and growth; more complex mechanisms operate in martensitic and bainitic transformations

Basic deformation mechanisms include slip, twinning, and grain-boundary sliding Annealing processes leading to recovery, recrystallization, and grain growth proceed by the mechanisms of polygonization, nucleation and growth, and grain-boundary migration, respectively

Processes developed in recent years, such as rapid solidification, mechanical alloying, ion implantation, deformation of superplastic alloys, and laser annealing, have introduced new structural morphologies For example, a structure without dendritic or cellular microsegregation was produced in an Ag-5Cu alloy that was electron beam melted and rapidly resolidified at 600 mm/s ( see Fig 21 and 22 in the article "Solidification Structures of Solid Solutions" in this Section)

In addition, rapid solidification techniques, such as melt spinning and splat cooling, can produce metallic glasses, that is,

amorphous (noncrystalline) metals, as described in the article "Amorphous Powder Metals" in Volume 7 of ASM

Handbook, formerly 9th Edition of Metals Handbook

References cited in this section

5 R.W Cahn and P Haasen, Ed., Physical Metallurgy, Parts I and II, 3rd ed., North-Holland, 1983

6 A.G Guy and J.J Hren, Elements of Physical Metallurgy, 3rd ed., Addison-Wesley, 1974

7 W.F Smith, Structures and Properties of Engineering Alloys, McGraw-Hill, 1981

8 R.E Smallman, Modern Physical Metallurgy, 4th ed., Butterworths, 1985

29 J.W Christian, The Theory of Transformations in Metals and Alloys, Pergamon Press, 1965; 2nd ed., Part I,

Fig 1 An outline of solidification structures

Fig 2 An outline of solid-state transformation structures

Fig 3 An outline of deformation and annealing structures

Grain Structure. Grains are small crystals (crystallites) that form a three-dimensional aggregate; they are normally viewed in sections, which by their nature are limited to two dimensions The main characteristics of a grain structure are grain size, grain shape, and grain-shape anisotropy

Types of Grain Structure. Typical grain structures include impingement structure, columnar structure, equiaxed grain structure, mature grain structure, deformed grain structure, inhibited recrystallization structure, and duplex grain structure

Impingement structure forms when grains grow until they meet or impinge, producing characteristic ragged interfaces This type of structure is rarely observed, because the interfaces usually are smoothed while the specimen remains at elevated temperature Impingement grains have been observed after secondary recrystallization (Ref 34)

Columnar structure forms by unidirectional growth processes, especially during solidification, and by a growth process involving diffusion accompanied by a solid-state transformation A columnar structure is shown in Fig 2 in the article "Solidification Structures of Steel" in this Section

Equiaxed grain structure may form by several processes, such as solidification (see Fig 2 in the article

"Solidification Structures of Steel") and recrystallization (see Fig 80 to 87 in the article "Copper and Copper Alloys" and Fig 1, 9, and 17 in the article "Zirconium and Hafnium and Their Alloys" in this Volume as well as Fig 1 in the article

"Textured Structures" in this Section)

Mature grain structure forms when the interfaces for example, those resulting from impingement-adjust themselves under capillarity driving forces

Deformed grain structure is the product of cold working In such a structure, the grain shapes are anisotropic (see Fig 60 to 68 in the article "Carbon and Alloy Steels" in this Volume and Fig 1 in "Textured Structures" in this Section)

Inhibited recrystallization structure forms when second-phase particles arranged in a nonrandom pattern inhibit the motion of grain boundaries and impose their nonrandom pattern on the resulting recrystallized structure (see Fig 29 and 30 in the article "Refractory Metals and Alloys" in this Volume)

Duplex grain structure (see Fig 47 in the article "Carbon and Alloy Steels" in this Volume) consists of discrete regions of larger and smaller grain sizes, that is, a bimodal distribution of grain sizes This structure is not related to microduplex alloys, which have characteristic duplex structures involving composition of two coexisting microconstituents rather than grain size (see the section "Multiphase Microstructures" below)

Three-Dimensional Grain Structure. Grain structures exist in three dimensions In a typical structure, two grains are separated by an interface; three interfaces join along a line or edge, and four edges join at a point or junction Six interfaces and four grains join at a junction in addition to the four edges Junctions of four grain edges are the basic units

of a mature grain structure; these junctions can be connected in innumerable ways without structural symmetry or exact repetition of detail (Ref 34, 35)

The major factors controlling grain structure are the requirement of space filling and the tendency toward minimum interfacial energy Space filling implies that adjoining grains interact to determine each other's shapes The problem of filling space with regular geometrical bodies has been studied over the past 100 years, beginning with Lord Kelvin in

1887 (Ref 34, 35) These studies have contributed to the understanding of grain structure, although actual grains may have irregular shapes

The tendency toward minimum interfacial energy operates by reducing the grain-boundary area as much as possible or, when applicable, by rotating the grain boundary into low-energy orientations The reduced grain-boundary area is an essential characteristic of mature grain structures

Topological relations for three-dimensional grain structures, such as the average number of sides of a grain face, have been analyzed The relations applicable to metal grains resemble those for certain nonmetallic materials, such as biological cell structures and foam structures (Ref 34, 35, 36)

Crystallography of Grain Boundaries. Various models have been proposed for the grain-boundary region, ranging from simple models for low-angle tilt boundaries to complicated transition regions in high-angle boundaries (Ref 37) Coincidence and twin boundaries are discussed in the article "Solidification Structures of Pure Metals" in this Section

Two-Dimensional Grain Structure. Sectioning of a three-dimensional grain structure presents the grain structure in only two dimensions for observation In a typical grain structure, the following simple relations between the three-dimensional and the two-dimensional structures can be established:

• A volume three-dimensional cell or spatial grain becomes an area, that is, a two-dimensional cell or planar grain

• An interface in a three-dimensional structure becomes a line or a grain boundary in a two dimensional structure

• An edge becomes a point

• A corner or junction (zero-dimensional cell) has an infinitesimal probability of being intersected by the plane of observation

• The true dihedral angle becomes an apparent dihedral angle, as discussed below

In the transition from a three- to a two-dimensional grain structure, another basic relation is that a structure consisting of uniformly sized three-dimensional, or spatial, grains becomes a two-dimensional structure in which the planar grains are not of uniform size This is because a random plane cuts grains at random positions, ranging from a corner to the largest cross section However, the resulting two-dimensional distribution of a grain structure of uniform three-dimensional grain size has definite statistical regularity In general, the true three-dimensional grain size is more nearly uniform than the apparent two-dimensional grain size The problems of grain-size measurement and grain-size statistics are covered in the article "Quantitative Metallography" in this Volume and in Ref 38, 39, 40

The topological relations of grains in two dimensions (planar grains) have been observed, demonstrating that the average planar grain in a mature structure is a hexagon Consequently, a seven-sided grain in a microsection must be balanced by

a five-sided grain, a nine-sided grain by a three-sided grain, or by three five-sided grains, and so on In addition, correct sampling for polygon distribution ensures better sampling for size (see Ref 34, 35, 36)

Grain Shape. In three dimensions, the average shape of equiaxed grains may, for some purposes, be approximated by a sphere Similarly, nonequiaxed grains may be represented by ellipsoids When viewed in two dimensions, nonequiaxed grains have extended shapes, as shown in Fig 1 in the article "Textured Structures" in this Section The quantitative determination of grain shape has been discussed (Ref 41)

Dihedral Angles. In three dimensions, the true dihedral angle is the angle between two faces of a grain measured in a plane normal to the edge at which the faces intersect In any actual section, the faces are intersected by planes oriented randomly at all angles Therefore, the apparent angle in two dimensions generally differs from the true angle in three dimensions Stated differently, the apparent or observed angle is the angle between the traces of grain faces in the plane of

a random section The angles in a two-dimensional section are statistically random in the absence of any orientation effect

or preselection

Quantitative relations exist between the true angle in three dimensions and the apparent angle observed in two dimensions If the true angle is 120°, as in a mature grain structure, the probability of finding an angle within 5° of the true angle is greater than the probability of finding an angle in any other 10° range (Ref 42) In fact, four angles out of five are expected to be within 25° of the true angle However, in actual grain structures, the true angles and, to a greater extent, the observed angles will have a distribution range

In two-phase structures, the true dihedral angles may differ from 120° even if the structure is equilibrated The extent to which the true angles differ depends on the relative interfacial tensions between grains of the two phases present It has been suggested that the true angle can be found by matching calculated and observed frequency plots The most probable angle is in every instance the true dihedral angle (Ref 43)

A simpler procedure for finding the true angle uses a cumulative distribution curve The median angle differs only slightly, and correctably, from the true angle In addition, fewer measurements perhaps 25 instead of several hundred are sufficient (Ref 44) Errors in measurement have been systematically analyzed, and dihedral angles with nonunique values have been considered (Ref 45)

References cited in this section

32 R.D Doherty, Stability of Grain Structure in Metals, J Mater Educ., Vol 6, 1984, p 845

33 A.P Sutton, Grain Boundary Structure, Int Met Rev., Vol 29, 1984, p 377

34 C.S Smith, Some Elementary Principles of Polycrystalline Microstructure, Met Rev., Vol 9, 1964, p 1-62

35 C.S Smith, Grain Shapes and Other Metallurgical Applications of Topology, in Metal Interfaces, American

Society for Metals, 1952, p 65-133

36 C.S Smith, Microstructure, Trans ASM, Vol 45, 1953, p 533-575

37 R.W Balluffi, Ed., Grain Boundary Structure and Kinetics, American Society for Metals, 1979

38 F Schückher, Grain Size, in Quantitative Microscopy, R.T DeHoff and F.N Rhines, Ed., McGraw-Hill,

1968

39 E.E Underwood, in Quantitative Stereology, Addison-Wesley, 1970, Chapters 4 and 5

40 H.E Exner, Analysis of Grain- and Particle-Size Distributions in Metallic Materials, Int Met Rev., Vol 17,

March 1972, p 24-52

41 E.E Underwood, in Quantitative Stereology, Addison-Wesley, 1970, p 228

42 D Harker and E.R Parker, Grain Shape and Grain Growth, Trans ASM, Vol 34, 1945, p 156-195

43 C.S Smith, Grains, Phases and Interfaces: An Interpretation of Microstructure, Trans AIME, Vol 175,

795-801

Substructure

In the broadest sense, substructure comprises all imperfections within the grains of a polycrystalline metal or a single crystal In the conventional sense, substructure refers to the subgrains formed by subboundaries (low-angle boundaries) This structure is revealed at intermediate magnifications; crystal imperfections, such as dislocations and stacking faults, can be revealed individually only at much higher magnifications

Examples of special kinds of substructure are:

• Lineage structure, mosaics originating by solidification

• Veining originating by transformation of face-centered cubic (fcc) iron to body-centered cubic (bcc) iron

• The cellular structure resulting from cold work

• Impurity substructure involving solute atmospheres associated with dislocations

• Dislocation networks originating by solidification, cold work, or fatigue (cyclic loading)

• Polygonized structure resulting from cold work followed by annealing

• Imperfections resulting from quenching or radiation damage

The subgrains that constitute substructure in the conventional sense have a large range of possible sizes The angular misorientations resulting from subboundaries range from a fraction of 1° to well over 1°

Multiphase Microstructures

Although many industrial alloys are single-phase materials for example, cartridge brass, silicon steel, and austenitic stainless steels multiphase alloys are more often encountered Most ferrous metals as well as many nonferrous alloys, especially the age-hardening and precipitation-hardening alloys, consist of more than one phase

The characteristic multiphase structures can be related to their modes of origin (see Fig 1 and 2) The major types of multiphase structures are discussed below

Structures in which both phases form entirely distinct grains have been called aggregated two-phase structures or random duplex aggregates They develop most clearly in alloys in which both phases are present in approximately equal volume fractions (Ref 46) In microduplex alloys, the two phases are distributed uniformly such that the boundaries are predominantly interphase interfaces This structure is usually fine scale and resistant to microstructural coarsening

Structures in which each phase is closely interconnected can result from spinodal decomposition (see the article "Spinodal Structures" in this Volume) The scale of these spinodal structures is very small They are characterized principally by their high degree of connectivity and often by crystallographic alignment of the phases (Ref 47)

Structures consisting of one continuous phase and isolated particles of a second phase (the plus-dispersed-phase structure) are the most varied of the multi-phase structures Among their characteristic variables are the relative volumes of the two phases, the size of the particles of the dispersed phase, the interparticle distance, the shape of the dispersed particles, and any special orientation of the dispersed particles with respect to each other and the matrix Some of these variables are interdependent; all of them can be measured Examples of the matrix-plus-dispersed-phase structure are rod-shaped particles embedded in a matrix and cellular precipitates The development

matrix-of high-strength steels has introduced the dual-phase microstructure in which a ferrite matrix contains small islands (approximately 20 vol%) of dispersed martensite A dual-phase steel microstructure is shown in Fig 1 and 2 in the article

"Carbon and Alloy Steels" in this Volume

Structures in which the two phases are arranged in alternate layers or lamellae form as eutectics, as pearlites in steels, and as pearlites in nonferrous eutectoid alloys Their characteristic variable is the interlamellar spacing

or thickness of the lamellae

A second phase can be distributed along the grain boundaries of a matrix phase, as in copper that is contaminated by bismuth Particles of a dispersed phase can also be located at other preferential sites, such as at slip planes after cold work followed by a precipitation process

Crystallography of Interphase Interfaces. The two phases that meet at an interface may differ in lattice constants, lattice type, and orientation These differences result in a mismatch or disregistry at the interface

This mismatch can be accommodated in one of the following three ways (Ref 37, 48): (1) A coherent interface exists when, in two adjoining structures, corresponding rows and planes of lattice points are continuous across the interface However, the rows and planes may change direction, resembling a coherent twin boundary Fully coherent interfaces between crystals of appreciable size are rare However, in limited areas, elastic straining can make it possible for coherency to exist The particles of transformation products with such coherency generally are too small to be observed using optical microscopy (2) At a semicoherent interface, the two lattices are elastically strained into coherence over limited areas; they accumulate misfit that is corrected periodically by discontinuities (dislocations) In other words, regions of forced elastic coherence alternate with regions of misfit (3) At an incoherent interface, the two lattices are discontinuous It was thought that such an interface could be explained in terms of dislocations compensating for the mismatch; however, such explanations have no physical significance, and the dislocation model of incoherent interfaces retains little interest

References cited in this section

37 R.W Balluffi, Ed., Grain Boundary Structure and Kinetics, American Society for Metals, 1979

46 R.W Cahn, Metal Systems, in Composite Materials, L Holliday, Ed., Elsevier, 1966, p 65-90

47 J.W Cahn, A Model for Connectivity in Multiphase Structures, Acta Metall., Vol 14, 1966, p 477-480

48 G.B Olson and M Cohen, Interphase Boundaries and the Concept of Coherency, Acta Metall., Vol 27,

1979, p 1907-1918

Macrostructure

The macrostructure of metals and alloys consists of inhomogeneities on a fairly large scale For example, gradients in a macrostructure exist on a much larger scale than that of the constituents of the microstructure A macrostructure may also comprise other inhomogeneities, such as blowholes or porosity in cast or weld metal (see Fig 54 in the article

"Magnesium Alloys" in this Volume), which originate during solidification, and flow lines in forgings (see Fig 82 in the article "Aluminum Alloys" in this Volume and Fig 11 in the article "Plastic Deformation Structures" in this Section), which originate during deformation Flow lines in forgings may be caused by elongated inclusions or by inhomogeneities

in grain-shape alignment Other examples of macrostructures are presented in the articles in this Volume dealing with metallographic procedures and representative microstructures of specific metals and alloys

Size Scales and Hierarchical Structures. The size scales of structural features of metals extend from the atomic level, ~0.1 nm (1 Ao ) to the size of entire metallic objects, ~1 m (3 1

4 ft) This range spans 10 orders of magnitude The techniques for observing structural features at different levels within this range must have adequate resolving powers Figure 4 shows the sizes of some common structural features of metals and various techniques for their observation with limits of resolution

Fig 4 Size scale relating structural features of metals to techniques of observation (after Ref 49)

Frequently, several structural features on different levels in a given metallic system are of interest For example, a polycrystalline single-phase metal has a grain structure, and within each grain a substructure may be present; or, in a polycrystalline long-range ordered binary alloy, a substructure of anti-phase boundaries may exist within each grain In a forging, the macroscopic flow lines may coexist with a structure of matrix grains in which precipitates are dispersed These examples of structural features that coexist at different levels are typical hierarchical structures

Reference cited in this section

49 S.M Allen and M.B Bever, Structure of Materials, in Encyclopedia of Materials Science and Engineering,

to be published

Introduction

PURE METALS normally solidify into polycrystalline masses, but it is relatively easy to produce single crystals by directional solidification from the melt The three common ways of growing single crystals are the Bridgman method, in which a mold is lowered out of a vertical tubular furnace; the Chalmers method, in which a boat is passed through a horizontal tubular furnace; and the Czochralski method, in which a crystal is pulled from a crucible containing the melt Much effort has been directed toward obtaining high-purity starting materials (often by zone refining) and toward maintaining purity during crystal growth Metal single crystals have been prepared with very low dislocation densities, but because pure metal crystals are very soft, this requires great care to reduce thermal and mechanical stresses during growth and subsequent handling Most metal single crystals have dislocation densities of about 106 to 107 per square centimeter These dislocations result from stresses induced during growth by thermal, mechanical, and composition gradients, as well as from entrapped particles In addition, vacancies can condense to form small dislocation loops subsequent to growth

Dislocations present in a metal crystal often polygonize into subboundaries during growth These subboundaries, which frequently intersect the growth front, are propagated by the growth process and result in subgrains that are elongated in the direction of the growth Subboundaries originating in this way are irregular (Fig 1) if the material is pure, but are regular and straight (Fig 2) in a very dilute alloy in which cellular growth has occurred Subboundaries also are formed where the liquid between two slightly misoriented dendrite arms freezes

Fig 1 Irregular subboundaries in high-purity tin grown without cells Subboundaries similar to these form in

many high-purity metals during solidification Compare with the structure shown in Fig 2 10% FeCl3 + 2% HCl, in H2O 40×

Fig 2 Regular subboundaries in tin of lower purity than that in Fig 1, grown with cells Cellular growth,

resulting from the presence of minute amounts of impurity, makes the subboundaries straight during solidification 10% FeCl 3 + 2% HCl, in H 2 O 35×

Dislocations in subboundaries can be resolved by careful metallography; an example is shown in Fig 3 A specimen etched to reveal the subboundaries and dislocations is depicted in Fig 4 The number of dislocations in the subboundaries often approximately equals the number of isolated dislocations in the subgrains

Fig 3 Individual dislocations (revealed by careful etching) that comprise a subboundary in germanium HNO3 acetic-HF-bromine 1500× (W.G Pfann)

-Fig 4 Dislocations and subboundaries produced by polygonization in germanium annealed after deformation

HNO 3 -acetic-HF-bromine 250× (J.R Patel)

Dislocations produced by thermal or mechanical stresses at low temperature often line up on the traces of slip planes, as shown in Fig 5 Dislocations produced by precipitation and condensation of vacancies during cooling usually are in the form of small loops (Fig 6) These vacancies can also form other clusters, such as stacking-fault tetrahedra

Fig 5 Dislocations aligned on traces of slip planes in germanium deformed at low temperature HNO3 HF-bromine 200× (J.R Patel)

-acetic-Fig 6 Dislocation loops produced by vacancy precipitation in germanium Thin-foil electron micrograph

60,000× (D.M Maher)

Polycrystalline Metals

The shape and size of the grains in a polycrystalline specimen of a pure metal are determined initially by nucleation and growth during solidification and ultimately by grain growth after solidification Castings are usually made by pouring hot liquid into a cold mold The solidification process depends on the degree of superheat that is, the degree to which the pouring temperature exceeds the melting point and on the shape and properties of the mold For a small superheat, crystals will nucleate on the cold mold wall, and solidification will proceed inward from the mold wall For a large superheat, the surface of the mold may be heated above the melting point during pouring so that nucleation occurs in the bulk of the liquid

Nucleating agents added to the melt will promote nucleation at many sites to produce a fine grain structure If the mold shape is intricate, it may be necessary to pour the liquid metal at a high temperature to prevent blockage of some of the channels by freezing before the mold has been filled Proper filling of the mold also depends on its design and material Mold design affords control of the size and shape of grains in various parts of a casting

The initial grain structure of a casting is determined by the distribution of nucleation sites and by the subsequent growth that proceeds from these sites If nucleation occurs in the bulk of the melt, then it must have been supercooled (that is, at a temperature below its melting point); therefore, the initial growth is likely to be dendritic The dendrites from one nucleus grow out until they impinge upon dendrites growing from adjacent nuclei (Fig 7) This process defines the initial shape of the grains, as illustrated in Fig 8 The dendrites continue to grow and thicken until the temperature is raised to the melting point by the heat of fusion, which is released by the freezing process The interdendritic liquid, not frozen at this point, can freeze only as heat is extracted from the sample

Fig 7 Dendrites in cyclohexanol, an organic compound that crystallizes like a metal 45×

Fig 8 Coarse, equiaxed grains produced by dendritic growth in an undercooled melt of pure metal

Nucleation can occur on the cold walls of the mold without most of the liquid being undercooled In this case, there may

be a short zone of dendritic growth, but only in a layer near the cold mold wall where the melt is undercooled Depending

on the thermal conditions, the melting point isotherm will be established somewhere between the hot melt and the cold wall The solidification of the casting will then proceed inward, with a more or less uniform isothermal front The grains that nucleated on the mold wall will elongate in the direction of heat flow, as illustrated in Fig 9, resulting in a columnar grain structure As growth proceeds, some crystal orientations will tend to persist at the expense of others, resulting in a preferred orientation texture

Fig 9 Elongated grains and preferred orientation produced by directional solidification

Dendritic growth occurs in a pure material only if the melt is undercooled This is because dendritic growth results from instability of the solid-liquid interface due to a diffusion process, which for a pure material can only be thermal diffusion

In an alloy, interface instabilities from chemical diffusion often result in dendritic growth

Although the grain boundaries in high-purity metals are very mobile, they may be slowed down or pinned by small concentrations of impurities If the initial grain size was large, grain growth is less likely, but there will usually be some adjustment of the grain boundaries to lower energy configurations

Grain Boundaries

The energy of a single grain boundary as a function of the angle of misorientation is shown in Fig 10 Small-angle boundaries, consisting of dislocation arrays, have much lower energies than large-angle boundaries The dislocation arrays in tilt and twist boundaries are illustrated in Fig 11 and 12, respectively

Fig 10 Calculated energy of the boundary between two grains as a function of the angle of misorientation

between the crystal lattices of the grains The energy becomes maximum (E m) at angle θm

Fig 11 Dislocations in a small-angle tilt boundary in gold Thin-foil transmission electron micrograph See also

Fig 10 24,000 × (R.W Balluffi)

Fig 12 Dislocations in a small-angle twist boundary in gold Thin-foil transmission electron micrograph See

also Fig 10 24,000× (R.W Balluffi)

Large-angle boundaries contain regions of good fit and of bad fit Low-energy, small-angle boundaries usually represent less than 10° to 15° of misorientation This range is a small fraction of the total range of possible misorientations Therefore, most random boundaries, formed by the growing together of two grains will be large-angle boundaries that is, will have more than 10° misorientation All large-angle boundaries except twin or coincidence boundaries have roughly the same energy and so form 120° angles with each other when equilibrated at the junction of three grains, as illustrated in Fig 13 (coincidence boundaries have some lattice sites common to both crystals) In Fig 14, the junction of a small-angle

boundary with two large-angle boundaries is shown The ratios of the boundary energies can be calculated from the angles

Fig 13 Grain boundaries in polycrystalline iron Most of the triple junctions of the grain boundaries form 120°

Solidification Structures of Solid Solutions

William J Boettinger, Metallurgist, Metallurgy Division, National Bureau of Standards

Introduction

SOLIDIFICATION is one of the most common steps in the materials processing cycle of metals and alloys Whether the material is used in service as a casting, a heat-treated casting, or a wrought product, the cast microstructure is important in determining properties and service life The microstructure of an as-solidified alloy can be described on three different size scales of magnification At the largest scale are grain size and gross casting defects such as macroporosity or long-range segregation of alloying additions (macrosegregation) At an intermediate scale, usually found within individual grains, are microscopic nonuniformities of composition termed microsegregation (coring) and associated second-phase segregates, inclusions, and microporosity At the finest scale are modifications of the cast structure by solid-state transformations, such as precipitation, which occur during solid-state cooling

This article discusses only microsegregation For castings solidified slowly, the length scale of microsegregation may be

as large as 1 mm (0.04 in.); in rapidly solidified alloys, the scale of microsegregation may be as small as 0.1 μm

Phase Diagrams and Solute Segregation

Alloys consist of a base metal, such as iron, aluminum, copper, or nickel, to which other elements (solutes) are added to yield desired properties These alloying elements are normally soluble in the liquid metal and the solid metal, but usually

in different concentrations This difference is shown in Fig 1, which is part of a binary phase diagram The top curve is called the liquidus At temperatures above this curve, the base metal and the alloying element are completely soluble as a liquid phase The bottom curve is called the solidus At temperatures below this curve, the two components are

completely soluble as a single solid crystalline phase called a solid solution The ratio, k, between the solidus composition

and liquidus composition at a fixed temperature is called the equilibrium partition (or distribution) coefficient

Fig 1 Terminal portion of a phase diagram, showing the solute concentration of liquid (CL) and solid (CS) that

are in equilibrium at temperature T Solute segregation occurs when, during freezing, solute is rejected into the

remaining liquid

In Fig 1, at the temperature T shown, k = CS/CL Figure 1 shows a case where the liquidus and solidus temperatures fall

with increasing alloying addition In this case, k is less than 1 For the type of partial phase diagram shown, k is near unity when the gap between the liquidus and solidus is narrow When the gap is large, k is small If the liquidus and solidus are

straight lines, the equilibrium partition coefficient is a constant independent of temperature

The existence of a gap between the liquidus and solidus leads to segregation of the alloying elements during the

solidification process Consider the case shown in Fig 1, where k is less than 1 If a liquid alloy of composition CL is

cooled to the liquidus temperature, the first solid to form has a composition, CS, that is equal to kCL Further cooling, especially when diffusion is slow, continues the formation of solid with a composition less than that of the liquid and causes a solute-rich layer of liquid to form near the liquid-solid interface The presence of this solute-rich layer promotes further segregation

Many alloys have phase diagrams containing regions of composition where the liquidus and solidus are separated Figure

2 shows three schematic binary phase diagrams that exhibit complete solid solubility (Fig 2a), a eutectic reaction (Fig 2b), and a peritectic reaction (Fig 2c) The primary (or first) product of solidification will be a solid solution, except for the eutectic composition in Fig 2(b) For the cases shown in Fig 2(b) and (c), solidification may be completed by some other process, such as a eutectic or peritectic reaction However, the details of the solidification of the solid solution remain important in determining the final microstructure

Fig 2 Shematic binary phase diagrams that exhibit (a) complete solid solubility, (b) partial solid solubility with

a eutectic reaction, and (c) partial solid solubility with a peritectic reaction: L, liquid; α and β, solid solutions

enrichment, the liquidus temperature at some positions ahead of the interface may actually lie below the real temperature,

as shown in Fig 3 In this case, the liquid ahead of the interface is said to be constitutionally supercooled Consequently, the liquid-solid interface usually cannot remain flat, and the interface will develop bumps or cells, resulting in microsegregation

Fig 3 Constitutional supercooling Formation of solute-rich layer in the liquid adjacent to the solid-liquid

interface (a) lowers the local liquidus temperature (b) In the region of liquid near interface, the actual temperature may be below the local liquidus, causing the interface to become nonplanar

The conditions necessary to obtain a flat liquid-solid interface can be estimated from the following inequality:

∆

>

where G is the liquid temperature gradient in K/cm, V is the interface speed in cm/s, ∆T is the equilibrium freezing range

of the alloy in K (temperature difference between liquidus and solidus for the original alloy composition), and D is the liquid diffusion coefficient In most cases,D ≈10-5 cm2/S for liquid metals Planar growth of alloys can usually be achieved only in crystal growth furnaces with high temperature gradients and low solidification speeds For example, for

planar solidification of an alloy with ∆T = 5 K and G = 200 K/cm, the solidification speed must be less than 4 μm/s (14

mm/h, or 0.6 in./h) Figure 4 shows a dilute tin-cadmium alloy solidified at high temperature gradient and low solidification speed to achieve planar growth (Ref 1) Note the absence of any microsegregation due to solidification

Planar growth of doped semiconductors, where ∆T is very small due to the low concentration of solute, constitutes the

basis of an entire industry However, most metallurgical alloys solidify with nonplanar interfaces

Fig 4 Directionally solidified Sn-0.6Cd alloy Section parallel to the growth direction shows a quenched planar

liquid-solid interface, indicating the absence of constitutional supercooling G = 320 K/cm, V = 0.85 μm/s, ∆T =

5.7 K 5 mL HNO3, 95 mL lactic acid 80× (C Brady)

Reference cited in this section

1 W.J Boettinger, The Structure of Directionally Solidified Two-Phase Sn-Cd Peritectic Alloys, Met Trans.,

Vol 5, 1974, p 2026

Cellular and Dendritic Structures

When constitutional supercooling is present, the interface between the liquid and solid takes on a cellular or dendritic morphology Figures 5 and 6 show the solidification of a transparent organic "alloy" that freezes like a metallic alloy (Ref

2) For conditions of growth where the ratio G/V is only slightly smaller than the ratio ∆T/D, the interface is cellular, as shown in Fig 5 For conditions of growth where the ratio G/V is much smaller than the ratio ∆T/D, the interface becomes

treelike or dendritic, as shown in Fig 6 The regions between the cells and dendrites, which are still liquid in the micrographs, are greatly enriched in solute and produce microsegregation segregation when freezing is complete It should be noted that the cells or dendrites in each micrograph share a common crystallographic orientation and thus belong to the same grain

Fig 5 Directionally solidified transparent organic "alloy" succinonitrile-5.5 mole% acetone in situ observation

of a growing cellular liquid-solid interface Growth direction is shown horizontal G = 67 K/cm, V = 0.58 μm/s,

∆T = 103 K, D ≈10 -5 cm 2 /s 32× (Ref 2)

Fig 6 Same organic compound as in Fig 5 V was increased to 1.17 μm/s to produce a dendritic liquid-solid

interface 36× (Ref 2)

Cells. Cellular structures are most often observed in dilute alloys where ∆T is small Cellular structures can take on three

characteristic morphologies when viewed in a metallographic section transverse to the growth direction: nodes (Fig 7), elongated cells (Fig 8), and hexagonal cells (Fig 9) In each case, a physical depression in the liquid-solid interface leads

to an increased concentration of solute in the material solidified near the depression The hexagonal cellular structure is observed much more frequently than the node or elongated cell structures

Sn-0.05Pb alloy; liquid decanted to reveal structures of liquid-solid interfaces Fig 7: photographed under oblique illumination, showing nodes in the interface Fig 8: elongated cells in the interface under bright-field illumination Fig 9: under bright-field illumination, revealing fully developed hexagonal cells

in the interface Not polished, not etched 150×

When cellular structures are observed by optical microscopy, contrast is usually developed by the preferential attack of the etchant on the regions of the structure enriched in solute, that is, the cell walls An example of an optical micrograph

of an etched transverse section through a cellular structure is shown in Fig 10 When hexagonal cellular structures are observed in a plane of section that is not transverse to the growth direction, the hexagonal structure may appear elongated

in one direction In the limiting case of a plane of section parallel to the growth direction, the hexagonal appearance is completely lost and the structure will appear more as the elongated cells in Fig 8

Fig 10 Pb-0.26Sb alloy casting Section shows a cellular solidification structure Etched at 1 V in a mixture of

10 mL 70% HClO4 and 50 mL methanol 50× (L.R Morris)

Dendrites. The most commonly observed solidification structure is dendritic These treelike shapes of solid solution are observed when the degree of constitutional supercooling is large, or when the alloy solidifies from a liquid that is cooled

to a temperature below the liquidus prior to the start of solidification Dendritic structures exist within single grains, and their main trunks and branches usually follow specific crystallographic directions within the grains

Depending on the type of phase diagram, dendritic structures may differ If the phase diagram shows complete solid solubility (Fig 2a), the structure will be single phase, containing only dendritic composition variations (Fig 11) If the phase diagram contains a eutectic (Fig 2b), the interdendritic regions will be composed of the two-phase eutectic In some cases, especially when the volume fraction of the interdendritic regions is small, the interdendritic region may be composed of a layer or discrete particles of a single second phase This situation is referred to as a divorced eutectic structure Figure 12 shows a dendritic structure with a large volume fraction of interdendritic eutectic For alloys where primary solidification is followed by a peritectic reaction, the microstructure depends strongly on solid diffusion rates When this diffusion is slow, the dendrites are coated by the peritectic phase

Fig 11 Ni-25Cu (at.%) alloy Section shows dendritic solidification structure 70 mL HNO3 and 30 mL H2O 10×

Fig 12 Ni-25Cu (at.%) alloy Section shows dendritic solidification structure 70 mL HNO3 and 30 mL H 2 O 175× (C Brady)

Because dendrites are complex, three-dimensional structures, plane-section micrographs must be interpreted carefully Figure 13 shows the structure of dendrites when the material surrounding the dendrites has been removed by selective deep etching Figure 14 shows the same structure observed in a plane section Parts of a single dendrite often appear disconnected when viewed in a plane section The high degree of segregation present in this structure is revealed in Fig

15 and 16

Fig 13 Scanning electron micrograph of Cu-10Co (at.%) alloy casting Matrix has been selectively etched to

reveal structure of individual cobalt solid solution dendrites Etchant not reported 150×

Fig 14 Scanning electron micrograph of Cu-10Co (at.%) alloy casting Section through cast specimen shows a

cobalt dendrite (smooth, gray) Image was formed using secondary electrons emitted by the specimen

As-polished 400×

Scanning electron micrographs of Cu-10Co (at.%) casting Fig 15: Image was formed using x-rays of Co-Kα wavelength emitted from the specimen under electron bombardment; the cobalt-rich dendrite appears light Fig 16: Image was formed using x-rays of Cu-Kα wavelength emitted from the specimen under electron bombardment The copper-rich matrix appears light As-polished 400×

Solute Redistribution in Dendritic Solidification. There is a simple method of estimating the compositional nonuniformity that can exist in a dendritic structure A small volume element of the dendritic or "mushy" zone of a casting that contains several dendrite arms can be characterized at any instant by a volume fraction of solid f s The volume fraction ranges from 0 to 1 as solidification proceeds from start to finish Assuming (1) no solid diffusion, (2) complete composition uniformity of the liquid remaining at any instant, (3) fluid flow adequate only to feed shrinkage,

and (4) a constant k, the composition of the solid that forms as a function of the fraction solid is given by the Scheil, or

normal freeze, equation:

CS = kCo(1 - f s)k-1

where CS is the solid composition formed, k is the equilibrium partition coefficient as defined earlier, Co is the initial alloy composition and f s is the volume fraction solidified

Figure 17 shows an example of the use of this equation for an Al-4.5Cu alloy for which Co is 4.5% Cu and k is 0.17

Figure 17(a) shows part of the aluminum-copper phase diagram The equation is used only for values of f s from zero up

to a value at which CS reaches the maximum solubility of copper in solid aluminum (5.65% Cu) From Fig 17, the Scheil equation predicts that the first solid to form (the centers of dendrites) will have a composition of 0.77% Cu and that the composition will increase to 5.65% Cu when f s = 0.91 (near the edge of the dendrites) The remaining fraction of the alloy (0.91 to 1.00) freezes as an interdendritic eutectic with an average composition of 33% Cu Another simple equation follows from the Scheil equation to predict the volume fraction of eutectic, f E, in a cast structure:

1 1

E E o

C f C

Fig 17 (a) The aluminum-rich end of the aluminum-copper phase diagram (b) Solid composition (CS) versus fraction solid (f s ) for Al-4.5Cu L, liquid; α, aluminum solid solution (Ref 3)

The predictions of the Scheil equation should be used carefully Generally, the equation tends to underestimate slightly

the composition at the center of a dendrite and to overestimate the volume fraction of eutectic (for k < 1) These errors can

be simply traced to deviations of real alloys from the assumptions stated above The presence of significant rates of solid diffusion (important for interstitial solutes, such as carbon in iron, or for very slow cooling) or of significant composition gradients in the liquid phase (important in chill casting and rapid solidification) is an effect that generally lessens microsegregation Also, if extensive fluid flow exists through the dendritic or "mushy" zone, the average composition of the solidified castings may be significantly altered in regions that are large compared to the dendrite scale This is called macrosegregation Details regarding this subject and the field of solidification can be found in Ref 4 and 5

References cited in this section

2 R Trivedi, Interdendritic Spacing: Part II A Comparison of Theory and Experiment, Met Trans A, Vol 15,

1984, p 977

3 M.C Flemings and R Mehrabian, Segregation in Castings and Ingots, in Solidification, American Society

for Metals, 1971

4 M.C Flemings, Solidification Processing, McGraw-Hill, 1974

5 W Kurz and D.J Fisher, Fundamentals of Solidification, Trans Tech., 1984

Rapid Solidification

The microstructural scale of solidified alloys generally decreases as the rate of heat extraction (cooling rate) increases The term rapid solidification is normally applied to casting processes in which the liquid cooling rate exceeds 103 K/s This definition is rather vague, because different alloys respond very differently to high rates of cooling Also, some microstructures observed in rapidly solidified alloys can be achieved by slow cooling when large liquid supercoolings are achieved prior to nucleation (Ref 6)

Techniques usually used to produce rapidly solidified alloys are melt spinning, planar flow casting, or melt extraction, which produce thin (~25- to 100-μm) ribbon, tape, sheet, or fiber; atomization, which produces powder (~10 to 200 μm); and surface melting and resolidification, which produce thin surface layers These methods may be considered casting techniques where at least one physical dimension of the final product is small Consolidation is used to yield large products from rapidly solidified alloys This consolidation often alters the solidification microstructure in final products; however, as with ordinary castings, many features of the solidification structure can remain in the final product For more

information on rapid solidification processes and properties and applications of rapidly solidified metals, refer to Powder

Metal Technologies and Applications, Volume 7 of the ASM Handbook

Many rapidly solidified structures differ little from those solidified at slow rates except for scale However, the details of the microsegregation (composition) profile within cells or dendrites, the volume fraction of intercellular or interdendritic material, and/or the actual identity of phases found in intercellular or interdendritic regions may differ from those found in more slowly solidified alloys Figure 18 shows a transverse section of a fine cellular structure of the silver-rich phase in Ag-15Cu alloy (Ref 7) In this figure, most of the intercellular regions are filled with the copper-rich phase, not the eutectic of silver and copper

Fig 18 Cellular microsegregation pattern observed in Ag-15Cu alloy, electron beam melted and resolidified at

approximately 25 mm/s (1 in./s) Thin foil transmission electron micrograph prepared by ion milling 32,000× (Ref 7)

A common occurrence in some rapidly solidified alloys is a change in the identity of the primary solidification phase from that observed for slow solidification Many excellent examples are found in hypereutectic aluminum alloys containing transition elements such as iron, manganese, or chromium If the alloy is hypereutectic, slowly cooled castings will contain intermetallics such as Al3Fe or Al6Mn as the primary (or first) phase to solidify However, under rapid solidification conditions the primary phase in these alloys is the aluminum solid solution usually found in a cellular structure with an intermetallic in the intercellular regions Figures 19 and 20 show cellular structures of the α-aluminum solid solution in hypereutectic aluminum-manganese alloys (Ref 8) This transition from an intermetallic to an aluminum solid solution as the primary phase can be understood by a careful examination of the kinetics of the competitive nucleation and growth of the intermetallic and α-aluminum solid solution (Ref 9)

Fig 19 Thin foil transmission electron micrograph of cellular structure of α-aluminum seen in a melt-spun 12Mn alloy Small particles of another phase decorate the cell walls Electropolished at -30 °C (-20 °F) in 950

Al-mL methanol, 50 Al-mL HClO 4 , and 15 mL HNO 3 16,000× (Ref 8)

Fig 20 Thin foil transmission electron micrograph of elongated cellular structure in a melt-spun Al-15Mn alloy

The contrast between some cells indicates crystallographic misorientation (subgrains) Electropolished at -30 °C (-20 °F) in 950 mL methanol, 50 mL HClO 4 , and 15 mL HNO 3 13,000× (Ref 8)

In some cases, an intermetallic that is not on the equilibrium phase diagram may compete with α-aluminum In iron alloys, a metastable phase, Al6Fe, rather than the stable phase, Al3Fe, can form under some rapid solidification conditions This situation is analogous to the appearance of cementite rather than graphite in some cast irons The use of metastable phase diagrams to assist in the interpretation of rapidly solidified microstructures is described in Ref 10

aluminum-Microsegregation-Free Structures. A particularly dramatic microstructural change that occurs in some rapidly solidified crystalline alloys is the complete absence of dendritic or cellular microsegregation Figures 21 and 22 show Ag-5Cu alloys solidified using electron beam surface melting and resolidification at speeds of 300 and 600 mm/s (12 and 24 in./s), respectively Figure 21 shows a longitudinal view of a cellular soldification structure In Fig 22, the alloy has solidified with a planar interface to produce a microsegregation-free alloy The fine particles are the result of a solid-state precipitation This type of microstructure can be understood by a theory that incorporates the effects of liquid-solid interfacial energy in the constitutional supercooling analysis described earlier (Ref 11)

Fig 21 Cellular microsegregation pattern in Ag-5Cu alloy revealed by dislocation net-works along cell walls

Specimen was electron beam melted and resolidified at approximately 300 mm/s (12 in./s) Thin foil transmission electron micrograph prepared by ion milling 18,000× (Ref 7)

Fig 22 Same as Fig 21, except resolidified at approximately 600 mm/s (24 in./s) The cellular structure is

absent, and the solid produced is uniform in composition except for fine copper precipitates formed during solid-state cooling 87,000× (Ref 7)

Other rapidly solidified alloys have microsegregation-free structures formed by a liquid-solid transformation similar to a massive solid-solid transformation Solidification by this mechanism is called partitionless (diffusionless) solidification, whereby the liquid transforms to solid without a change in composition In other words, the ratio of the solid composition

at the interface to the liquid composition is 1, rather than the equilibrium partition coefficient Velocities required to produce partitionless solidification range from 1 to 10 m/s (40 to 400 in./s) The liquidus and solidus of the phase diagrams obviously do not apply in this situation For more information on massive transformation, see the article

"Massive Transformation Structures" in this Volume

Figures 23 and 24 show optical and transmission electron micrographs, respectively, of a single-phase free solid solution of a silver-copper alloy of eutectic composition (28% Cu) that formed by partitionless solidification The alloy is not only free of microsegregation, but also has a solid solubility of copper in silver far in excess of the equilibrium solubility limit (~9% Cu) Solubility extension is commonly seen in many aluminum alloys (Ref 13)

microsegregation-Fig 23 Columnar grains of single-phase solid solution of melt-spun Ag-28Cu alloy (eutectic composition)

Section of full ribbon cross section Chill (wheel) side is at bottom 20 mL NH 4 OH, 10 mL 3% H 2 O 2 , 10 mL H 2 O (used fresh) 2000× (Ref 12)

Fig 24 Same alloy as Fig 23 Transmission electron micrograph of thin foil parallel to the chill surface

prepared by ion milling Three grains are shown The fine mottled structure is the result of solid-state decomposition 100,000× (D Shechtman)

Structure of Atomized Powders. Rapidly solidified alloy powders exhibit a broad spectrum of solidification structures, depending on alloy composition and solidification conditions Figure 25 shows a single powder particle of Al-4.5Cu in which dendritic structure radiates from a point on the surface where nucleation has occurred The scale of the structure is relatively uniform across the powder particle

Fig 25 Al-4.5Cu alloy atomized powder Optical microscopy shows dendritic structure in rapidly solidified

powder particle Keller's reagent 390× (S Wright)

Other rapidly solidified powders often show significant microstructural variations across individual powder particles Figure 26 shows an entire submicron Al-6Si powder particle with three microstructural zones On the lower left is a zone containing only fine precipitates formed by solid-state transformation of an initially uniform supersaturated solid solution

On the upper right is a zone with a cellular solidification structure An intermediate region in the center of the particle is a transition zone All of the aluminum solid solution in the particle has the same crystallographic orientation; therefore, the particle is a single grain This particle is estimated to have been supercooled by approximately 200 K while in liquid form before nucleation occurred on the surface at the lower left Initial growth of the solid occurred very rapidly in a partitionless manner The interface speed is reduced as the liquid-solid interface crosses the particle due to the release of latent heat of fusion and warming of the powder particle Because of this reduction in interface speed, the solidification front becomes cellular

Fig 26 Electrohydrodynamic (EHD) atomized Al-6Si powder Transmission electron micrograph of unthinned

particle mounted on TEM replication tape, carbon coated and tape dissolved Rapid solidification has produced (A) a supersaturated zone, (B) a transition zone where solute has built up in front of the interface, and (C) cells formed after the droplet has recalesced Thin foil preparation of the specimen was not necessary because of the small diameter of the particle and the low atomic number of aluminum 80,000× (Ref 14)

Figure 27 shows a two-zone microstructure observed in a larger diameter (~ 10 μm) hypereutectic Al-8Fe powder particle (Ref 15) A thin foil was prepared by electropolishing a 3-mm (0.12-in.) diam green powder compact (Ref 16) In this alloy, however, the zone to the left where nucleation occurs contains a very fine cellular structure, and the zone to the right contains a coarse cellular structure of α-aluminum with Al6Fe between the cells Larger powder particles of the alloy frequently do not supercool significantly before the start of solidification and contain AlFe as the primary phase

Fig 27 Thin foil transmission electron micrograph of vacuum-atomized Al-8Fe powder A single nucleation site

of the supercooled particle (left) initiated solidification at a high interface rate to produce a fine cellular structure Recalescence slowed the interface rate to produce the coarser cellular structure at the right The green powder compact was electropolished at -30 °C (-20 °F) in 950 mL methanol, 50 mL HClO4, and 15 mL HNO 3 6300× (Ref 15)

References cited in this section

6 J.H Perepezko and J.J Paike, Under-cooling Behavior of Liquid Metals, in Rapidly Solidified Amorphous and Crystalline Alloys, B.H Kear, B.C Giessen, and M Cohen, Ed., North Holland, 1982, p 49

7 W.J Boettinger, D Shechtman, R.J Schaefer, and F.S Biancaniello, The Effect of Rapid Solidification

Velocity on the Microstructure of Ag-Cu Alloys, Met Trans A, Vol 15, 1984, p 55

8 D Shechtman, R.J Schaefer, and F.S Biancaniello, Precipitation in Rapidly Solidified Al-Mn Alloys, Met Trans A, Vol 15, 1984, p 1987

9 I.R Hughes and H Jones, Coupled Eutectic Growth in Al-Fe Alloys: Part I Effects of High Growth

Velocity, J Mater Sci., Vol 11, 1976, p 1781

10 J.H Perepezko and W.J Boettinger, Use of Metastable Phase Diagrams in Rapid Solidification, Mat Res Soc Symp Proc., Vol 19, 1983, p 223

11 W.W Mullins and R.F Sekerka, Stability of a Planar Interface During Solidification of a Dilute Binary

Alloy, J Appl Phys., Vol 35, 1964, p 444

12 W.J Boettinger, Growth Kinetic Limitations in Rapid Solidification, in Rapidly Solidified Amorphous and Crystalline Alloys, B.H Kear, B.C Giessen, and M Cohen, Ed., North Holland, 1982, p 15

13 H Jones, Rapid Solidification of Metals and Alloys, Institute of Metallurgists, London, 1984

14 C Levi and R Mehrabian, Microstructure of Rapidly Solidified Aluminum Alloy Submicron Powders, Met Trans A, Vol 13, 1982, p 13

15 W.J Boettinger, L Bendersky, and J.G Early, An Analysis of the Microstructure of Rapidly Solidified

Al-8 wt% Iron Powder, Met Trans A, submitted for publication

16 D Shechtman and E Gutmanas, Transmission Electron Microscopy of Metallic Powder, Prakt Metallogr.,

Vol 18, 1981, p 587

Solidification Structures of Eutectic Alloys

Franklin D Lemkey, Senior Consulting Scientist, United Technologies Research Center, Adjunct Professor of Engineering, Dartmouth College; R Wayne Kraft, Professor of Metallurgy and Materials Science, Lehigh University

Introduction

EUTECTIC STRUCTURES can form beautiful and regular arrays of lamellae or rods, depending on the nature and amounts of the phases and the conditions of solidification As many as four phases have been observed to grow simultaneously from the melt; however, most technologically useful eutectic alloys consist of two phases In general, eutectic microstructures have certain similar characteristics that allow arbitrary classification by structure When interpreting eutectic structures, it is necessary to consider (1) the types of phases, dictated by their amounts and growth kinetics, (2) the scale, or size range, of the phases, which is determined by solid/liquid interface shape and solidification velocity, (3) the operating compositional range, and (4) the sectioning techniques

Solidification and Size of Eutectic Structures

The schematic phase diagram in Fig 1 shows a binary eutectic invariant point at temperature, Te, and composition, Ce, At this point, solid phases α and β simultaneously solidify from the liquid, L For most eutectic systems, if the solidification occurs directionally, an aligned two-phase solid can be produced At least three conditions must be simultaneously satisfied, however, to produce parallel duplex microstructure: (1) the heat must be removed from the melt unidirectionally, (2) a sufficiently positive temperature gradient must be maintained ahead of the solidifying interface to prevent unwanted nucleation and preserve solid/liquid interface planarity, and (3) cooperative nucleation and growth processes must occur between the phases

Fig 1 Phase diagram for a eutectic system showing the eutectic invariant point at temperature (Te ) and

composition (Ce)

Examining this process reveals that the major solidification parameters for example, the thermal gradient, G, at the liquid/solid interface and the growth rate, R, or the velocity at which the liquid/solid interface advances are keys to the

control and size of the structure If both phases occur with approximately equal volume fractions, a situation encouraged

by a symmetrical phase diagram, a preference for the formation of lamellar structures is found (for example, Al-CuAl,

lead-tin, and Ni3Al-Ni3Nb) On the other hand, if one phase is present in a small volume fraction, there is a tendency to the formation of fibers of that phase (for example, chromium in copper-chromium, molybdenum in Ni3Al-Mo) When the volume fraction of one phase is less than 0.3, the eutectic will probably be fibrous If the volume fraction of one phase is between 0.3 and 0.5, the eutectic will tend to be lamellar Because eutectic structures are often more complex than this simplified model implies, it is useful to consider three of their features individually: particle structure, colony structure, and grain structure Their nominal size ranges are shown in Fig 2; the extremes of the ranges overlap somewhat

Fig 2 Size ranges of eutectic structures

Particle Structure. Individual phase particles typically have micron dimensions Their size and shape strongly affect the mechanical and physical properties of the eutectic aggregate Size and shape are affected by solidification rate, thermal gradients, atomic bonding, relative amounts, crystallographic factors, interfacial energies, impurity content, and alloy composition

One condition that affects the size of phase particles, which has been explained theoretically and confirmed

experimentally, is illustrated in Fig 3 The interparticle spacing, λ, and the solidification rate, R, for an alloy system are

related by the equation λ2R = constant Because the volume percentage of the phases in a particular alloy is fixed, the

equation permits determination of effect of solidification rate on particle size from size measurements made on only one specimen solidified at a known rate

Fig 3 Eutectic spacings as a function of growth rate (Ref 1)

The shape of a particle can be defined by the three principal dimensions: x, y, and z Describing the phase particles as fibrous (x ?y ≈z), lamellar (x ≈y ?z), or equiaxed (x ≈y ≈z) is useful Illustrated in Fig 4, 5, 6, and 7 are the lamellar

and fibrous forms Each casting was unidirectionally solidified to produce specimens in which the nominal size and shape could be established by examining the microstructures of sections parallel to and normal to the direction of solidification

Fig 4

Lamellar eutectic structures Fig 4: Casting of Ni3Al-Ni3Nb solidified unidirectionally (left to right) Fig 5: Same casting as Fig 4, but a section taken normal to the direction of solidification, showing alternate lamellae of Ni3Al and Ni3Nb Both etched in 5 parts HCl, 1 part HNO3, 6 parts glycerol 1000×

Fig 6 Fibrous eutectic in section taken normal to growth Faceted dark phase is molybdenum in a solid-solution

(γ) matrix of nickel containing cuboidal precipitates of Ni 3 Al (γ') Electrolytic etch: 10 parts H 3 PO 4 , 50 parts

H 2 SO 4 , 40 parts HNO 3 15,000×

Fig 7 Scanning electron micrograph of fibrous eutectic unidirectionally solidified vertically Section shows

exposed tantalum carbide particles in a superalloy nickel matrix Etchant not identified 40,000× (M Henry)

Colony Structure. Eutectic colonies are aggregates of phase particles with a characteristic arrangement Colonies are not present in all eutectic structures They are frequently confused with eutectic grains, particularly in specimens cut from material that was not solidified unidirectionally Eutectic colonies are formed when the alloy solidifies with a cellular, rather than an essentially planar, freezing interface (Fig 8) The cellular freezing face is caused by constitutional undercooling brought about by rejection of an impurity from the solidifying eutectic phases into the liquid (see the article

"Solidification Structures of Solid Solutions" in this Volume)

Fig 8 Colony structure of rod eutectic that was solidified unidirectionally (right to left) Section parallel to

direction of growth shows fanlike arrangement of niobium carbide rods (dark) in nickel matrix (light) resulting from curved liquid-solid interface Murakami's reagent 30×

The arrangement of phase particles in a eutectic colony is typically fanlike on a section cut parallel to the direction of solidification (Fig 8); a section cut normal to the direction of solidification displays a honeycomb pattern (Fig 9) In castings not unidirectionally solidified, the honeycomb and the fanlike features identify eutectic colonies (Fig 10)

Fig 9 Colony structure of a lamellar eutectic that was solidified unidirectionally Section taken normal to the

direction of solidification showing the honeycomb pattern of the colonies, which were formed when constitutional undercooling caused the freezing face to be cellular in shape, rather than essentially planar Dark and light layers in each colony are Mg2Al3 phase and aluminum, respectively Etchant not identified 200×

Fig 10 Colony structure of a CuAl2-Al lamellar eutectic in a casting that was not unidirectionally solidified Section shows the honeycomb pattern (where section is normal to direction of solidification) and the fanlike arrangement (where section is parallel to direction of solidification) As-polished 250×

Grain Structure. Eutectic grains are structural features visible on a polished or fractured surface at magnifications close to unity Eutectic grains can be more difficult to define and identify than are grains of a single-phase metal For