Metallurgical Principles - Materials Data Repository Home

CHAPTER 4

Metallurgical PrinciplesMartin Bauser*

4.1 Introduction

THE EXTENSIVE USE of metals over thou-sands of years, but in particular during the in-dustrial age, can be attributed to their specificproperties. In addition to their good electricalconductivity—associated with good thermalconductivity—and visual properties, the mostimportant are the mechanical properties and thegood workability. The use and range of appli-cation of different metals depends on these andmany other properties, including, for example,the corrosion resistance. However, the produc-tion costs also play a decisive role.

These extend from the accessibility and theextraction of the ore, the cost in the refining upto and including the casting and processing. Ex-trusion is suitable as a deformation process indifferent ways for the different materials. Thewell-used expression “extrudability” is a way ofexpressing this.

Knowledge of deformation technology aloneis not sufficient to be able to understand and con-trol the processes taking place during extrusion.The quality of the billet materials, the processestaking place within the extruded material duringextrusion, and the properties of the extruded sec-tion can be understood only from a metallurgicalviewpoint.

This chapter, therefore, explains the basic ter-minology of metallurgy—naturally in the con-text of the aims of this book—to provide thetools needed by those involved in extrusion.More detailed explanations can be found in the

*Casting and Cast Structure, Homogenizing of Aluminum Alloys, Wolfgang Schneider; Casting and Cast Structure of CopperAlloys, Adolf Frei; Deformation, Recovery, Recrystallization, Martin Bauser

literature, e.g., [Hor 67, Guz 70, Ber 83, Sch 81,Dah 93, Hou 93, Blu 93, Alt 94].

This materials science chapter is concernedonly with metals. The atoms are held togetherby the so-called “metallic bond,” which differ-entiates them from the nonmetals (covalentbond, ionic bond). A characteristic of this me-tallic bond is the ease of movement of the outerelectrons, which are no longer attached to indi-vidual atoms but form an electron gas. This mo-bility of the electrons is the reason for the goodelectrical conductivity of metals.

Nonmetallic crystalline materials can, as arule, be extruded only with difficulty, if at all.There is insufficient ductility in the workingrange of temperatures and pressure. Glass withits amorphous structure is, however, an excep-tion because it softens on heating.

Recently, metal and nonmetal composite ma-terials have been extruded. Nonmetallic particlesor fibers are uniformly embedded in a metallicmatrix, usually artificially, to improve the me-chanical properties and often to reduce wear.These materials are also discussed briefly in thischapter.

4.2 Structure

4.2.1 Lattice Structure Single Phase

All metals consist of crystallites (small crys-tals), the arrangement, size, and shape of whichare referred to as the structure. Polishing of

Extrusion: Second Edition M. Bauser, G. Sauer, K. Siegert, editors, p 141-194 DOI:10.1361/exse2006p141

Copyright © 2006 ASM International®All rights reserved.

www.asminternational.org

142 / Extrusion, Second Edition

Fig. 4.1 The structure of brass [Wie 86]

Fig. 4.2 Elementary cell in a simple cubic lattice [Sch 81]

metal specimens and suitable methods of etch-ing are used to reveal the structure—usually un-der the microscope (Fig. 4.1).

The crystallites that form during solidificationfrom the melt change during the cooling. Theyare stretched by the deformation and reform byrecrystallization after annealing a deformed ma-terial. They are, however, always crystalliteswith an ordered lattice structure that constitutethe metal.

If a metal consists of only one type of crys-tallite, it has a single-phase structure.

The regular arrangement of the atoms of eachcrystallite forms the lattice structure. The small-est component of this lattice is referred to as theunit cell. A real crystallite, also called a grain,consists of many unit cells arranged uniformlyin adjacent rows like building blocks.

The unit cells are very small. The edge length,known as the lattice constant, is of the order of10�7 mm so that a crystallite with a mean di-ameter of 0.1 mm contains 1018 elementary lat-tice building blocks.

The location of a lattice—a simple cubic lat-tice is shown in Fig. 4.2—is defined by the di-rection of the edges of the elementary cell. Thisis referred to as the orientation of the lattice.

In a real metal body consisting of numerouscrystallites, the latter can be differentiated by

their orientation, which is usually completelyrandom.

This is shown schematically in Fig. 4.3. Thedifferent types of metal lattices have a signifi-cant effect on their behavior during deformation.The most important lattice structures are shownin Fig. 4.4. They are face-centered cubic (fcc),body-centered cubic (bcc), hexagonal, andbody-centered tetragonal lattices. Figure 4.4 alsoshows the lattice structure of the important met-als at room temperature.

Chapter 4: Metallurgical Principles / 143

Fig. 4.3 Diagram of a polycrystalline structure

Hexagonal close-packedlatticeZinc

Tetragonal body-centeredlatticeTin

Fig. 4.4 The most important types of metal lattice structures [Wie 86]

The description of a space lattice by surfacesand direction is usually given by Miller’s indices(see the Appendix to this chapter).

In the solid state the metal atoms try toachieve the highest packing density. The fcc andspecific hexagonal lattices posses the densestspherical packing. The atoms are assumed to bespheres that preferentially form these arrange-ments when shaken together. A bcc lattice has aspatial filling of 68%. Face-centered cubic lat-tices and the hexagonal close-packed structurehave a spatial filling of 78% of the maximumpossible. It is therefore clear that the density ofa metal depends not only on the atomic weightbut also on the crystal structure.

The size and shape of the crystallites can beinfluenced by control of the solidification fromthe melt and by recrystallisation during the an-nealing of a deformed metal. Extremely fine

crystallites can have diameters well below 1 lmand large grains several mm. It is also possibleto grow single crystals. These are particularlysuited for studying the basics of the deformationprocessed within crystals. Cylindrical specimensseveral centimeters long are produced for thispurpose.

Defect-free ideal crystals do not occur in prac-tice. Indeed, specific defects are necessary forplastic deformation and diffusion processes tooccur at all. The most important types of defectsare shown schematically in Fig. 4.5 and 4.6 andinclude:

● Vacancies: Numerous lattice sites are unoc-cupied. The number of vacancies increasesexponentially with temperature. In the equi-librium state, e.g., in an undeformed alumi-num crystal at room temperature, approxi-mately one in every billion lattice sites isunoccupied. In contrast, one in a thousand isunoccupied just below the melting point.

● Interstitial atoms: In pure metals embeddedatoms of the same type between the regularlattice sites. In alloys specific foreign atomsthat are significantly smaller than the atomsof the base lattice can be embedded into theinterstitial places (e.g., C and N in Fe).

● Foreign atoms that substitute the base metalatoms: Foreign atoms, particularly thosewith a similar atomic radius and with a simi-lar lattice structure can replace atoms of thebase structure.

● Dislocations: Step dislocations are lines atwhich an atomic plane ends. On one side ofthe slip plane—so called because the dislo-cation line can move in this plane—there isone more lattice plane than on the other (Fig.4.6a). In a screw dislocation, the lattice areasare displaced relative to each other so that


VacancyForeign atom(substitutional)

Interlattice atomForeign atom(interstitial)

Fig. 4.5 Point type of lattice defect (vacancies, interlatticeatoms, substitutional and interstitial foreign atoms)

Dislocation line

Slip plane

(a)

Dislocation line

(b)

Fig. 4.6 Dislocation � linear lattice defect. (a) Edge dislocation. (b) Screw dislocation [Wie 86]

the atoms are arranged around the disloca-tion line like a spiral (Fig. 4.6b). A real, usu-ally randomly curved dislocation line in aslip plane can be considered to consist ofedge and screw elements. The dislocationline consists of a closed ring or ends at agrain boundary or particle. In an undeformedstructure, there can be up to 1 km of dislo-cation lines/mm3 (dislocation density 106 to108/cm2.

If the orientation of two adjacent crystallitesdiffers only slightly, the grain boundary betweenthem can be considered to consist of dislocationslocated above each other (low angle grainboundary; subgrain-boundary) (Fig. 4.7a). Sub-grains differ in their orientation from each otherby only a few minutes of arc up to a few degrees[Dah 93, Ber 83]. A grain boundary betweencrystallites with very different orientations(large angle grain boundary) is, in contrast, aseverely distorted region with defects of varioustypes (Fig. 4.7b).Twins represent a special form of crystallites

in contact with each other. Their orientation is amirror image at a plane boundary surface(Fig. 4.8).

A metallic structure that consists of only onetype of atom contains dislocations, vacancies,and interstitial atoms. If a metal contains two ormore different atoms, it is an alloy containingembedded or substituted foreign atoms. These

can accumulate readily in the more or less se-verely distorted grain boundaries. Every type ofdefect naturally causes internal stresses withinthe lattice. Internal stresses are decreased whenforeign atoms migrate to grain boundaries be-cause they are able to fill the voids that are pres-ent.

4.2.2 Multiphase Structure—EquilibriumDiagrams

Alloys, materials that consist of two or moremetals, can have one or more phases (types ofcrystal).

If only one type of crystal is present in a me-tallic material—pure metal or alloy—it has asingle phase referred to as a homogeneous struc-ture. Correspondingly, a heterogeneous struc-ture consists of several types of crystals(phases). The well-known free-cutting brassCuZn39Pb, for example, contains both the fcc �brass as well as the bcc b brass, in addition toundissolved lead inclusions as a chip breaker.Figure 4.1 shows a micrograph of a single phase� brass and a binary phase �-b brass.

The structure, the quantity, and the distribu-tion of the various phases of an alloy depend onthe content of the individual metals, the previoushistory, and the temperature. An alloy can passthrough several states as it cools from the melt.The amounts of the individual phases that formin the thermodynamic equilibrium state, i.e., thestate with the lowest energy, are described bythe equilibrium diagram. The spatial distributionof the crystallites of the phases that occur arestudied using metallographic images and imageanalysis techniques.

4.2.2.1 Binary Systems

The simplest equilibrium diagrams are thosewith only two alloying partners involved. Thisis referred to as a binary alloy. The abscissa


Fig. 4.7 Grain boundary structure. (a) Low angle boundary. (b) High angle boundary [Alt 94, Sch 81]

gives the amount of the two atoms (in weightpercent) and the temperature is plotted along theordinate.

Figure 4.9 shows schematically some of thefrequently occurring basic types of binary phasesystems. The equilibrium diagrams of most bi-nary systems can be constructed from these ex-amples.

The liquidus temperature (at which solidifi-cation commences during cooling from the melt)and the solidus temperature (at which solidifi-cation is complete) can be read as well as theoccurrence of the phases and the melting interval(the temperature difference between the liquidusand the solidus temperature). In the melting in-terval, solidified particles float in the melt.

In pure metals, the liquidus and solidus co-incide because there is only one solidificationtemperature. It is identical to the melting tem-perature as the solid metal transforms to the liq-uid state at the same temperature on heating.

The different basic types of the equilibriumdiagram are shown in Fig. 4.9.

Continuous Solid Solution (Type 1). Thelattice type of the parent metal A is retained over

the full range of solution up to the pure metal B.The atoms B replace to an increasing extent theatoms A in their lattice positions. This is referredto as atoms B dissolving in lattice A, analogousto the processes in liquids. The solution of oneatom type in another is favored where the atomradii differ only slightly from each other, e.g.,copper and nickel, which therefore form a con-tinuous solid solution.

In field a the metal occurs as a liquid, in fieldb both as solidified particles and as liquid, andin field c as 100% solidified solid-solution for-mat.

Because both metals usually have differentmelting temperatures, the liquidus and the soli-dus temperature vary with the alloy addition.

Eutectic System (Type 2). In this case, theatoms of type B are completely insoluble in theatoms of type A—usually when there are verylarge differences between the atom radii. If B isalloyed to A, then B forms its own second phase.The fraction of phase B increases in the equilib-rium diagram uniformly to the right and the frac-tion of phase A decreases accordingly. In the


Fig. 4.8 Twinned grains and twin boundary

Tem

pera

ture

Type 1

Continuous solid solution

(Example: Ni-Cu)

Type 3

Eutectic system with solid solution formation

(Example: Cu-Ag)

Type 2

Eutectic system

(Example: Bi-Cd)

Type 4

Peritectic system

(Example: Cu-Snfor T > 760° C)

B content B content B content B content

Fig. 4.9 Basic types of phase diagram. a, melt; b, melt � solid crystals; c, solid solution (homogeneous); d, two different crystals(heterogeneous); E, eutectic point; P, peritectic point [Wie 86]

eutectic system there is one concentration Ewhere the liquidus and the solidus temperaturecoincide at the minimum melting point. A typ-

ical eutectic structure forms at this concentrationduring solidification, usually consisting of alter-nating layers of atom A and atom B. If, on theother hand, an alloy outside the eutectic concen-tration on the A-rich side cools, the A phase firstsolidifies from the melt when the liquidus lineis crossed. The residual melt from which themetal A is removed increases in concentrationin the metal B as cooling continues. At the eu-tectic temperature the residual melt has reachedthe eutectic composition and solidifies entirelyeutectically. The structure shows the primary so-lidified grains of type A surrounded by the eu-tectoid from the residual melt.

Eutectic System with Solid Solution (Type3). This type of binary system occurs frequentlywhen the atoms of type B dissolve in small con-centrations by substitution of atoms A in the lat-tice (field c). As soon as a limiting concentrationis reached, the internal stresses produced by theinclusion of the second atom type B become sohigh that the formation of a second phase be-comes energetically more favorable (field d). Asmentioned previously, the structure always triesto achieve the state of lowest energy.

The composition of the A-rich phase in fieldd corresponds to the concentration on the leftboundary line to c, the composition of the B-rich phase to the concentration of the rightboundary line c2 at the corresponding tempera-ture. The more the displacement of the compo-sition of the alloy from the left boundary line tothe right, the greater is the number of grains ofthe B-rich phase and, correspondingly, the fewerthe grains of the A-rich phase in this two-phasefield.


Fig. 4.10 Structure of a eutectic binary alloy [Ber 83]

In Fig. 4.10, the structures corresponding todifferent compositions are sketched under theequilibrium diagram of such a eutectic systemwith solid solution formation:

● Concentration �: Crystals of the phase c1 so-lidify here, surrounded by layers of theeutectoid E formed at the eutectic tempera-ture E.

● Concentration b (eutectic composition): Thestructure consists only of the layers of theeutectoid of the phases c1 and c2.

● Concentration c:Grains of phase c2 form ini-tially and are surrounded by the eutectoid onfurther cooling.

As can be seen from the diagram, the solu-bility of the second atom type decreases in thesolution regions (fields c1 and c2) with decreas-

ing temperature, which must result in the pre-cipitation of particles of the second phase evenafter solidification to maintain equilibrium (seealso Fig. 4.13, section 4.2.3).

This type of eutectic system equilibrium dia-gram with solid-solution formation rarely rep-resents the real state of an alloy because the mo-bility of atoms decreases so strongly withdecreasing temperature that the elimination ofthe nonequilibrum state is possible only after along holding period, if at all.

The structural state at high temperatures can,therefore, be “frozen” by quenching from thistemperature so quickly that the atoms have notime to form the room temperature configura-tion.

Peritectic System (Type 4). A peritectic re-action occurs when the melt (field a) forms a


T1 = constant

B = 40%

40%B35%A

S

S

S

S25%C25%C

%C

S

40%B40%B 40%B

35%A 60%A 60%CBA

P

%A

%B

A, B, C: alloy partner

S: meltsolid-solution phases

C

T1

Tem

pera

ture

, T

%C

S

S

S

S

(a) (b)

S

Fig. 4.11 Tertiary system representation. (a) Isothermal section (temperature T1 � constant). (b) Quasi-binary section (alloy com-ponent B, e.g., 40%) [Ber 83]

new solid phase with a primary precipitatedsolid phase. The � grains are then surroundedby a shell of the b phase. In the peritectic con-centration point a peritectoid structure forms thatappears similar in micrographs to the eutectoid.The development of the equilibrium state on fur-ther cooling is very restricted and can beachieved only after long holding periods at thecorresponding temperature.

4.2.2.2 Intermediate Phases

A solid solution consists, as previously men-tioned, of the parent metal with interstitial orsubstitutional foreign atoms that are distributedrandomly or only slightly ordered.

Intermediate phases occur when two or moretypes of atom are bonded together in an ordered,usually complex, structure. Intermediate phaseshave little or no deformability and are very brit-tle.

4.2.2.3 Multicomponent Systems

If an alloy consists of more than two types ofatoms in a significant amount, it is not possibleto depict it completely in a two-dimensional di-

agram. Two different depictions are used forthree-part systems (ternary systems), as shownin Fig. 4.11:

● Isothermal section: The diagram consists ofan equilateral triangle representing the threealloying partners. Each point within the tri-angle can be related to a specific concentra-tion of the alloying elements A, B, C (in theexample point P: 35%A, 40%B, 25%C).This type of diagram naturally shows thestate at a specific temperature. In order torepresent all phase states, therefore, it is nec-essary to have numerous similar isothermalsections that are often depicted within eachother.

● Quasi-binary section: The proportion of twotypes of atoms are fixed (in the exampleshown A:B � 3:2) and the phase states stud-ied for increasing amounts of atom type C.This diagram can be read in the same wayas a binary system.

4.2.3 Diffusion, Precipitation,Nonequilibrium

The phase diagrams show that in many alloysboth the composition of the individual phases of


Foreign atom Vacancy

Activated state

Activation energy

Base state

a b c

a b c

Ene

rgy

stat

e

Fig. 4.12 Transposition through vacancies in diffusion [Ber 83]

a multiphase structure as well as the fraction ofthe phases in the structure change during heatingor cooling. Atoms have to migrate from onephase to another to maintain the new equilibriumwhen the temperature changes. This diffusionoccurs by transposition processes, which, in thecase of substitution solid solutions, requires alarge number of vacancies. The migrating for-eign atom always moves to the nearest vacancyposition (Fig. 4.12).

The atoms of a crystal lattice are stationaryonly at absolute zero. As the temperature in-creases they oscillate more strongly about theirlocation and can then more easily leave their po-sition. When changing places, the migratingatom has to pass through a state of higher en-ergy, known as the activation energy Q, whichis easier in the energy-rich state of a higher tem-perature: diffusion processes occur faster athigher temperatures than at lower.In interstitial solid solutions the signifi-

cantly smaller foreign atoms (e.g., carbon or ni-trogen in iron) migrate from one interstitial lat-tice place to the next without the assistance ofvacancies. The diffusion activation energy islower than with substitution solid solutions.

From diffusion laws, which also apply togases and liquids, the rate of diffusion is not only

temperature dependent but is faster the greaterthe concentration difference between the two re-gions that will be equalized by the diffusion pro-cess.

A precipitation process occurs if a secondphase is formed in the solid state of a homoge-neous alloy during cooling when the solvus lineis crossed. Nuclei first form and continue togrow until phase equilibrium is reached. Typicalexamples are the age-hardening aluminum al-loys (Fig. 4.13).

If a single-phase region is rapidly cooled fromthe solid solution (e.g., the dashed line X2 inFig. 4.13), the foreign atoms initially are re-tained in solution because they are practicallyimmobile at low temperatures. Only when agingoccurs at a temperature that enables significantdiffusion to take place can nuclei of the secondphase form and grow.

Because this process of solution treatmentwith quenching followed by aging is usually as-sociated with an increase in strength, it is re-ferred to as “age hardening.”

Energy also has to be expended for the nucle-ation of precipitates, but this is usually lower atspecific crystal defects or in a grain boundarythan in the interior of the undistorted lattice.With a low precipitation pressure—e.g., slow


Fig. 4.13 Precipitation of a two-phase system [Wie 86]

cooling—the nuclei preferentially form at thesedefects and grain boundaries. The foreign atomshave sufficient mobility at higher temperaturesto cover large distances to these nucleationpoints and grow there to precipitates.

However, after quenching and aging at lowertemperatures, the precipitation pressure is highenough for large quantities of nuclei to form inthe less-distorted zones of the crystal interiors,which can then be reached by atoms in the im-mediate neighborhood in spite of the lower mo-bility at these lower temperatures. The precipi-tates are then so fine that they cannot be detectedunder the microscope.

If a suitable material is hardened by solutiontreatment and aging, the rate of cooling from thesolution region needed for freezing increases thefurther the equilibrium state at low temperaturesis from the solvus line. Accordingly, withAlMgSi0.5 (line X1 in Fig. 4.13) cooling fromthe solution temperature at about 500 �C in sta-tionary or slightly moving air is required,

whereas, with AlMgSi1 (line X2), waterquenching is needed to retain the solid-solutionstate.

This ideal picture can change if foreign pre-cipitates or inclusions are present: manganese inAlMgSi0.5 acts as a nucleating agent so thatrapid cooling has to be used in this case—similarto the higher-alloyed AlMgSi alloys—in orderto retain the solution state [Sca 64].

The change with time (the kinetics) of trans-formation of the solid solution that is supersat-urated by the quenching can be represented inthe time-temperature transformation (TTT) dia-gram. The time is the abscissa and the tempera-ture is the ordinate. The degree of transforma-tion is quantified by curves showing suitableparameters (phase fraction in micrographs, hard-ness, age hardenability).

The isothermal TTT diagram shows the trans-formation progress of rapidly cooled samples ata constant temperature.

In many cases the continuous TTT diagram ismore informative. This shows the precipitationbehavior for different cooling rates from the so-lution temperature. The minimum cooling ratecan be read from the diagram. Above this, nosignificant precipitation of the second phase oc-curs and the full hardening effect is obtainedfrom the subsequent age hardening (Fig. 4.14).This is known as the critical cooling rate.

Figure 4.14(a) also shows the freezing tem-perature TE, below which diffusion processes nolonger occur. It occurs approximately at 0.3 �TS (TS � the melting point) for diffusing sub-stitution atoms. Figure 4.14(b) shows constanthardness curves in the continuous TTT dia-grams.

If the cooling process from the solution tem-perature can be sufficiently accurately charac-terized by a specific parameter—e.g., the cool-ing time ta to 200 �C—it is possible to use thisparameter ta as the abscissa and to plot againstit a parameter that indicates the degree of trans-formation.

In steels and other materials that have differ-ent crystal lattices at different temperatures, thechange in the transformation with time analo-gous to segregation can be depicted by isother-mal or continuous TTT diagrams.

Grain growth follows the nucleation of thesecond phase as additional foreign atoms collectat the nucleus by diffusion. In the case of finelydistributed precipitates, a coarsening process fi-nally occurs after further thermal treatment. Thisis referred to as coagulation (Fig. 4.15).


Fig. 4.14 Time-temperature transformation (TTE) diagram. (a) Principle [Ber 83]. (b) Continuous TTE diagram for AlMgSi0.5. I, solid-solution � (B dissolved in A) in equilibrium; II, solid-solution � (B still dissolved in A) in unstable state; III, second phase

b precipitates to an increasing state from the solid-solution �; IV, equilibrium between phases � and b; V, frozen unstable state of the� phase


Fig. 4.15 Coagulation of precipitates [Ber 83]

During this coagulation process, atoms mi-grate by diffusion from smaller to larger precip-itates. The small precipitates gradually disappearand only the large precipitates remain. Manysmall precipitates have a higher total of bound-ary surfaces between the phases than a few largeones. Because these boundary surfaces are en-ergy rich coagulation reduces the energy levels.

The martensitic transformation is a specialtype of phase transformation in which the atomsadopt new positions without diffusion. Thismovement takes place extremely quickly afterquenching and produces significant stresses be-tween the old surrounding lattice and the newmartensitic structure that is usually needle orplate shaped. The formation of martensite ismainly known in the behavior of steels, but itdoes occur in other materials, including �-bbrasses.

4.2.4 Melting and Casting Processes

The structure in the billet is very importantfor the extrusion process. It has a strong influ-ence on the extrusion load, the maximum pos-sible extrusion temperature and speed, as wellas the structure and properties of the extrudedsection.

The change from chill casting to continuouscasting—now used for most metals—alonebrought a significant quality improvement. Thisprevented the severe billet segregation associ-ated with chill casting. The aim of numerousdevelopments in continuous casting initiallyconcentrated on the production of a fine struc-

*Casting and Cast Structure, Homogenizing of Aluminum Alloys, Wolfgang Schneider

ture with the minimum segregation that solidi-fied from the bottom to the top rather than fromthe outside to the interior.

The formation of the solid phase by coolingof a melt takes place by nucleation and growthwhen the liquids line has been crossed in a simi-lar way to the precipitation of a second phase inthe solid state described previously. Nuclei firstform where the melt is cooler, which is at themold wall in the casting process. As the under-cooling of the melt increases the nucleationwork needed reduces so that the number of nu-clei increases and a finer structure is produced.

Casting and Cast Structure,Homogenizing ofAluminum Alloys

Wolfgang Schneider*

4.3 Development of the ContinuousCast Structure

Figure 4.16 shows the development of thecontinuous cast structure in a simplified man-ner. The heat extraction through the water-cooled mold wall, also referred to as indirect


Fig. 4.16 Schematic of the formation of the continuous cast structure

cooling, causes the solidification of the periph-eral region, i.e., to the formation of the so-called peripheral shell. This shrinks away fromthe mold wall and forms an air gap between thecast surface and the mold wall. This air gaphinders further heat extraction, resulting in re-heating of the peripheral shell up to tempera-tures in the solidification range; i.e., partial oreven complete melting of the peripheral shell ispossible. Peripheral segregation can form dur-ing the dwell time in the air gap region as aresult of the partial melting. The surface of thecast log is characterized by this and by the so-lidification of the peripheral shell itself, particu-larly in the meniscus region. The cast logfinally emerges with the peripheral shell fromthe bottom of the mold and is hit by the coolingwater flowing from the mold to complete solid-ification, referred to as direct cooling. This re-sults in the formation of the cast grain and cellstructure and, finally, the eutectic cast phaseson the grain and cell boundaries as well as acharacteristic profile of the solidification frontacross the cast log cross section, which is alsoknown as the cast sump (Fig. 4.16).

The formation of the structure of the castlog is significantly influenced by the moldtechnology. The parameters of mold technol-ogy, which can be assumed to have a control-

ling influence, are the height of the effectivemold wall and the metal head in the mold, theshape of the metal feed into the mold, and thedirect cooling of the cast log. The direct cool-ing of the cast log is carried out by eitherspray water or mist cooling so that the moldhas to have corresponding nozzles that willproduce these methods of cooling. In alumi-num continuous casting mist cooling predom-inates because aluminum has a high coefficientof thermal conductivity. In this case the appli-cation of a closed water mist below the moldis sufficient for rapid cooling of the cast log,and the use of additional specific cooling zonesis not necessary.

The cooling of the cast log by direct coolingresults in the development of stresses as a resultof the temperature gradient from the interior ofthe cast log to the periphery; these can lead to theformation of hot or cold cracking, depending onthe alloy system. The hot cracks usually form inthe center of extrusion billets. Cold cracks, on theother hand, can occur over the full cross section.Alloys of the type AlMgSi are sensitive to hotcracking, whereas cold cracks are more likely todevelop in AlCuMg and AlZnMgCu alloys.

The formation of the continuous cast structureis influenced by the casting parameters as wellas the mould technology. The most important are


Bill

et e

dge

Allo

ying

ele

men

t con

cent

ratio

n

Distance from the billet surface

Nominal analysis

1 2 3 4 mm250 mm

Fig. 4.17 Typical profile of the alloying element concentra-tion in the peripheral segregation region

the casting speed, the volume of cooling waterand the casting temperature or the temperatureof the melt as it enters the mould. The castingspeed, in particular, is extremely important. Ac-curate selection of the casting speed and themould wall height for the alloy composition isvery important, for example, to avoid hot andcold cracking.

4.3.1 Peripheral Shell andPeripheral Segregation

The peripheral region of the cast log in theform of a peripheral shell is of particular impor-tance in the continuous casting process. Numer-ous casting defects can occur within this periph-eral shell, and these should not finish up in thesemifinished product during subsequent pro-cessing of the cast log. Consequently, the extentand the cast structure of the peripheral shell areimportant and have to be controlled by the cast-ing technology to minimize the peripheral shellthickness and defects.

The extent of the peripheral zone is largelydetermined by the mold wall height and the alloycomposition [Scn 85]. The larger the solidifica-tion interval of the alloy being cast and thehigher the mold wall, the thicker is the periph-eral shell. In comparison, the casting parametersexert only a small influence on the thickness ofthe peripheral shell.

The cast structure of the peripheral shell usu-ally consists of a fine and a coarse cellular re-gion. The fine cellular zone solidifies in directcontact with the mold wall, whereas the forma-tion of the coarse cellular region can be attrib-uted to the insulation effect of the formation ofthe air gap already described.

The mechanical properties of the solidifyingperipheral zone tend toward low values. Theyare influenced by the cast structure formationand the temperature of the peripheral zone, inparticular by the actual temperature in the solid-ification interval and subsequently by the frac-tion of solid phase during the dwell of the pe-ripheral shell in the air gap region [Ohm 88,Ohm 89].

The casting defects that are usually found inthe peripheral zone include peripheral segrega-tion. Zones of peripheral segregation are char-acterized by an above-average high concentra-tion of alloying elements and a sharpconcentration jump to the adjacent structure[Mor 69]. There are also often alloying elementimpoverished zones below the peripheral seg-

regation zones. This can be clearly seen in theexample in Fig. 4.17.

The occurrence of peripheral segregations hasto be considered in relationship to the solidifi-cation of the peripheral shell. Peripheral segre-gation forms during the dwell of the peripheralshell in the air gap region. Transport of residualmelt regions enriched in the alloying elementstowards the surface of the cast log takes place.The main transport mechanism is the metallo-static pressure of the melt before the peripheralshell. This is referred to as air gap segregation.A further mechanism for the formation of pe-ripheral segregation is the residual melt enrichedwith alloying elements overflowing the menis-cus curvature of the peripheral shell. This is re-ferred to as meniscus segregation. Both thesesegregation methods described are shown sche-matically in Fig. 4.18 [Bux 77].

Various segregation formation shapes can bedifferentiated in peripheral segregation. Two ex-amples are shown in Fig. 4.19. The nature of thesegregation depends on the alloy type but alsoto some extent on the mold technology.

4.3.2 Cast Log Surface

The surface structure of the cast log is largelydetermined by the peripheral segregation.


Fig. 4.19 Examples of types of peripheral segregation. (a) Sweating beads. (b) Lap segregations. Casting direction downward. Thewidth is approximately 0.75 mm

Fig. 4.18 Schematic of the peripheral segregation mechanism [Bux 77]

The different forms of peripheral segregationshown in Fig. 4.19 also result in correspondinglydifferent cast log surfaces. Figure 4.20 showsexamples of surfaces caused by different typesof segregation.

Cold shuts are one surface feature that is notrelated to the peripheral segregation. An exam-ple is shown in Fig. 4.20. The formation of coldshuts can be attributed to the freezing of the meltmeniscus in contact with the mold wall followed


Fig. 4.20 Examples of types of cast log surfaces. (a) Surfacewith sweat beads. (b) Surface with segregation

beads. (c) Surface with cold laps. Casting direction downward

Fig. 4.21 Cast structure of a round billet (200 mm Ø) aftergrain-refining treatment

by the melt flowing over the frozen peripheralshell. This surface defect is observed when thecasting speed or the casting temperature is toolow [Alt 65].

One type of surface that appears on the castlog and is best described as a surface defect isthe surface tear that occurs in the form of crackstransverse to the casting direction. These tearsare attributed to the frictional forces between theperipheral shell and the mold wall.

The types of surface appearance just de-scribed are always found on cast logs. The for-mation of the peripheral shell is indeed impor-tant for the formation of the surface. However,a reduction of the peripheral shell thicknessshould not always be associated with an im-provement in the surface quality. In general, theaim is to produce a smooth surface, but in manycases this is merely a purely optical effect. Inother words, a smooth surface is not definitelyassociated with a good peripheral structure qual-ity. This means that more value should be placedon producing a good peripheral structure qualityfor the subsequent processing of the cast logs.

4.3.3 Cast Grain and Cell Structure,Intermetallic Phases

An important requirement of the structure ofthe cast log is a fine and uniform globulitic for-mation of the cast grain structure. The produc-tion of a fine cast grain reduces the hot and coldcrack sensitivity, depending on the alloy system.In addition, the hot-working behavior of the castlogs and the surface appearance of the extrudedproduct after surface treatment also are affected.The globulitic cast grain structure is achievedusing a grain-refining treatment (Fig. 4.21).

Grain-refining alloy additions of the typeAlTiB are used [Los 77]. These are added duringthe casting as a 9 mm thick wire into the meltin the casting launder. The grain-refining actionof the AlTiB alloy is attributable to complex nu-


Fig. 4.22 Example of a cellular/dendritic solidifiedstructure.The width is approximately 1.0 mm

Fig. 4.23 Characteristic variation of the cast cell size overthe billet cross section for different billet diameters

[Bux 77a]

cleation processes in the melt [Rei 80]. The mostimportant constituent of the alloys is insolubleTiB2, which exists as fine particles and plays adirect role in the nucleation processes. Alloy ad-ditions of 0.2 to 1.5 kg/t aluminum are neededto obtain adequate grain refining depending onthe alloy composition.

The majority of commercial wrought materi-als cast by the continuous casting process exhibita cellular to dendritic type of solidification [Fle74]. This, in turn, determines the cast grainstructure. Figure 4.22 shows an example of acellular/dendritic solidified cast structure.

The grain and cell boundaries of the castgrains are lined with intermetallic phases thatoccur as eutectics during the solidification. Thegrain and cell structure is also characterized bygrain segregations, i.e., by nonuniform distri-bution of alloying elements in the structure.These grain segregations, which can be attrib-uted to nonequilibrium solidification with an en-richment of the alloying elements from the graincenter to the grain boundary, can result in theformation of eutectic phases on the grain andcell boundaries even in those alloy systemswhere they cannot occur according to the equi-librium diagram.

The cell structure of a cast log should be fineand uniform [Los 83]. This results in a corre-sponding fine and uniform distribution of thecast phases in the semifinished product after de-formation of the structure. This requirement isparticularly important for semifinished productsthat are given a surface treatment for decorativeapplications. Variations in the cell size can pro-duce a banded surface.

The cast cell structure of aluminum cast logsexhibit a characteristic profile over the cross sec-

tion as shown in Fig. 4.23 for different billetdiameters [Bux 77a].

There is a coarse cellular structure in the pe-ripheral region of the cast log. This is the regionof the peripheral shell. Next to this is a regionwith a very fine cell structure. Toward the cen-ter of the cast log, the cell size increases, reach-ing the maximum in the center. It should alsobe recognized that the quantitative profile of thecell size strongly depends on the diameter ofthe cast log. This relationship can be explainedusing the diagram in Fig. 4.24 [Bux 77a].

The faster the corresponding volume elementof metal is transformed from the liquid to thesolid state, then the finer the cell structure. If itremains too long in the partially solidified statebetween the liquidus and the solidus state,coarse cells form because these have a high “lo-cal solidification time” [Kat 67].

Figure 4.24 shows that the distance betweenthe liquidus and solidus line varies across thecross section of the cast log. The distance is thegreatest in the center of the log where the high-est local solidification time occurs and, accord-ingly, the maximum cell size forms. The cell


Mold

Liquid

Solid

Pasty

Cell size

Thickness of Din the pasty region x

D

D(x)

Fig. 4.24 Relationship between the distance between the liquidus and the solidus line and the cell size across the billet crosssection [Bux 77a]

size minimum coincides with the contact pointof the cast log direct cooling, because this iswhere low local solidification times exist be-cause of the high solidification rates. Corre-spondingly, the distance between the liquidusand the solidus lines is smaller than in the logcenter where the cooling conditions in contin-uous casting result in the lowest solidificationrates.

4.3.4 Homogenizing

The cast structure produced in continuouscasting is neither the optimum for the subse-quent processing nor for any surface treatmentof the semifinished product. As mentioned pre-viously, there is marked grain segregation in thecontinuous cast structure as well as a supersat-uration of alloying elements. The volume frac-tion of eutectic precipitates is also increased.Therefore, a homogenization treatment consist-ing of high temperature annealing at tempera-tures between 450 and 600 �C is carried out toremove grain segregation and supersaturation.

The heating time at the homogenization tem-perature depends on the cell size and the rate ofdiffusion of the specific alloying constituents[Ake 90]. A fine cell structure is therefore ad-vantageous because the concentration equaliza-tion within a cell can occur more quickly be-cause the diffusion path is shorter.

The cooling of the logs from the homogeni-zation temperature is also important in the ho-mogenization process. Depending on the cool-

ing rate, secondary precipitates form in variousquantities and sizes from the �-solid solution asa result of decreasing solubility. The actual pre-cipitation state influences the extrudability of thebillets, as is well known from the AlMgSi alloys[Res 92]. If the cooling rate is too slow, coarseMg2Si phases can preferentially form on grainand cell boundaries in the temperature rangeabove 400 �C. These coarse phases do not dis-solve quickly enough in solution during heatingto the extrusion temperature and can melt duringextrusion if their melting point is exceeded dur-ing the extrusion process. The consequence is areduction in the extrudability of the billet be-cause surface defects can form on the sectioneven at lower extrusion speeds. Mg2Si phasesthat precipitate at temperatures below 400 �C areso fine they can redissolve even during shorttimes at the extrusion temperature.

If the cooling rate from the homogenizingtemperature is too high, fewer and finer secon-dary precipitates form. This is the same as in-creasing the content of dissolved silicon andmagnesium in the �-solid solution. This raisesthe initial extrusion pressure and reduces the ex-trusion speed at a constant extrusion load. Figure4.25 shows the structure of cast logs of Al-MgSi0.5 alloy before and after homogenization.Figure 4.25(a) shows the cast state with needlesof AlFeSi phases on the cell boundaries. Figure4.25(b) shows the structure after a defined an-nealing and cooling. The precipitated secondaryphases (Mg2Si) can be seen in the �-solid so-lution.


Fig. 4.25 Cast structure of AlMgSi0.5 round billet as-cast and after homogenization. (a) As-cast. (b) Homogenized

According to the previous discussion, there-fore, the homogenization including coolingshould be designed to give fast cooling to 400�C to avoid the precipitation of coarse phases.This should be followed by slow cooling to re-duce the content of dissolved alloying constitu-ents [Ake 88].

Casting and CastStructure of Copper Alloys

Adolf Frei*

Similar to aluminum alloys, the crystalgrowth is largely influenced by the temperatureprofile in the mold. The cast structure usuallyconsists of a narrow very fine grain peripheralzone connected to a coarse globulitic or laminastructure extending to the core. Frequently,some form of single crystal is found in the cen-ter of copper billets.

Figure 4.26(a) shows the typical structure ofa copper billet with a large columnar grain frac-tion, and Fig. 4.26(b) shows a brass billet witha globulitic core.

Direct cooling by water impingement is usu-ally situated below a relatively short mold forthe continuous casting generally used today forthe casting of copper alloys (see Chapter 6).The indirect cooling through the mold wall is

*Casting and Cast Structure of Copper Alloys, Adolf Frei

thus reduced and that from the water impinge-ment directly below the casting mold increased(Fig. 4.27).

4.4 Segregation and Gas Porosity

The solidification in the mold is describedusing a copper-tin alloy with 8% tin as an ex-ample (Fig. 4.28). As soon as the liquidus line(at point 1 in Fig. 4.28 at temperature T1) iscrossed, nuclei of the solid phase with compo-sition “a” grow. The residual melt is impover-ished in copper atoms and enriched in tin at-oms. In the equilibrium state the last residue ofthe residual melt at temperature T2 has the com-position “b.” A fully solidified structure haszones with large variations in composition, i.e.,segregation, particularly with slow cooling.

Specific crystallographic directions are char-acterized by high rates of growth leading topreferentially orientated columnar grains thatsuppress others and enable a cast structure todevelop. Preferred growth direction is the rea-son for the formation of fir tree dendrites seen,for example, in cast iron alloys.

The numerous nuclei that form with rapidcooling and simultaneously grow produce a finedistribution of residual melt constituents in thesolid state, which ensures that the segregationis less severe and more finely distributed.

In the solid state, there is both grain segre-gation, which is a variation in concentration inand around the individual grains of the solidi-fied structure, and billet segregation, which pro-


Fig. 4.26 Cast structure of copper alloys. (a) Copper billet. (b) Brass billet

duces differences between the billet peripheryand the core by the growth of a solidificationfront from the cold mold wall toward the inte-rior of the billet. The core is richer in the sec-ond (in the example the tin containing) phasecorresponding to the phase diagram.

If gases are dissolved in a melt (e.g., N2 orH2), their solubility in the solid state is usuallysignificantly lower than in the melt; i.e., the gasprecipitates in the form of pores during solidi-fication. In chill mold casting where the meltslowly grows from the periphery toward thecenter, these gas pores collect like pearl stringsin the center of the cast billet. In continuouscasting they are found in the segregations of theresidual melt, i.e., in the boundaries of thegrains of the cast structure.

Segregations are, as mentioned previously,supersaturated zones outside the equilibriumstate of the solidified structure. The segrega-tions solidified from the residual melt possessa composition that produces crystallites of brit-tle phases, which hinder working. In addition,because these segregation zones, which solidi-fied at lower temperatures, also melt at lowertemperatures than the parent metal during heat-ing, there is a risk of grain-boundary meltingduring hot working.

4.5 Cooling and Casting Defects

In spite of the indisputable advantages ofcontinuous casting compared with the older

processes, there still exists the possibility of de-fects. Billets with the optimum extrudabilitycan be achieved only by careful control of thecasting parameters.

The best billet quality would be obtainedwith a flat solidification front perpendicular tothe casting direction. This would require the en-tire heat extraction to take place in the directionof the log axis. This can be achieved only withuneconomic casting speeds.

Practice is a long way removed from theideal case because both indirect and direct cool-ing remove the heat from the cast log mainlyperpendicular to the axis. The casting sump (re-gion of liquid metal in the cast log) has a U-or V-shaped section.

As a rule of thumb, approximately one-thirdof the heat to be extracted is taken through themold and the rest by direct cooling of the castlog.

If the direct cooling of the log is not exactlysymmetrical, this results in distortion of theemerging log. This bending can quickly be-come so large that it makes direct extrusionvery difficult and indirect extrusion with itslonger billet lengths of up to 1.5 m impossible.

4.5.1 Cast Log Surface

The appearance of the surface of the cast logis largely determined by the processes in themold. The melt transferred into the mold ini-tially solidifies in direct contact with the mold


Fig. 4.27 Direct and indirect cooling in the continuous cast-ing mold for copper alloys

Fig. 4.28 Section of the copper-tin phase diagram [Ray 49]

wall to a thin shell, which pulls away from themold wall after a few centimeters because ofthe solidification shrinkage. This produces ashrinkage gap that restricts further heat flowfrom the cast log. The interaction between thetemperature of the melt, the method of metalfeed, and the cooling conditions can lead to sur-face defects as the cast log passes furtherthrough the mold:● Longitudinal cracks when the shell is

pushed out by the metallostatic pressure ofthe melt in the sump and tears because ofits low strength

● If the shell is pushed outward as an annularbead into the shrinkage gap, this can causeso much friction at the mold wall that trans-verse cracks occur (“cold shuts”).

● In the extreme case, the log shell can locallymelt from the still liquid core. The metalthat emerges into the shrinkage gap solidi-fies as thin strings on the log surface.

● When casting high zinc containing brasseszinc can evaporate from the hot log shelland condense on the cold mold wall. Fromtime to time it is stripped away by the castlog and forms “zinc flakes” on the surfaceof the log.

In the continuous casting of alloys with widesolidification intervals, for example, tinbronzes, “inverse segregation” can occur. Thesolidification shrinkage results in the tin-rich re-


sidual melt, which solidifies last, being enrichedunder the log surface and being partially pushedoutward into the shrinkage gap. Good extrusionresults can be obtained with these billets onlyby turning off the tin-rich peripheral layer andhomogenization. CuSn8 is homogenized at ap-proximately 700 �C for several hours to dis-solve segregations of the brittle, tin-rich low-melting-point d-phases.

The shrinkage gap is reduced in molds thatare tapered toward the bottom, and the effectsdescribed can at least be suppressed to somedegree. These molds are preferred for the cast-ing of copper. The taper has to be carefullymatched to the casting speed and the coolingconditions. The optimization is helped by com-puter programs.

If the metal head in the mold is not constantduring the casting, inclusions of the coveringagent and slag can occur under the surface ofthe cast log.

Excessive direct cooling can cause longitu-dinal cracks in the surface of the cast log, par-ticularly with stress-sensitive alloys.

When casting is carried out with a salt cover,a thin salt film is initially formed on the surfaceof the cast log, and this has to be rinsed off inthe region of direct cooling. If the compositionof the covering salt is unsuitable or the directcooling inadequate, the film can bond firmly tothe cast log. It can then act as a lubricant be-tween the billet and the container during extru-sion and result in parts of the cast skin flowinginto the extruded section, resulting in scrap dueto laminations.

4.5.2 Interior of the Cast Log

Perfect results in extrusion can be achievedonly from billets with smooth, clean surfacesand a dense structure over the complete crosssection. Unsuitable casting conditions not onlyproduce defects on the surface of the extrudedsection, but also within the section.

With a deep V-shaped casting sump there isthe risk that gas bubbles can no longer escapebut are retained in the region of the solidifica-tion point: porosity in the cast. In the region ofthe sump tip small zones of liquid metal canalso be cut from the casting sump. When theysolidify, the shrinkage results in small hollowspaces—solidification blowholes.

*Deformation, Recovery, Recrystallization, Martin Bauser

Solidification with a V-shaped sump resultsautomatically in stresses in the cast log and, inthe worst case, can result in longitudinal cracksin the core of the cast log. With very sensitivematerials such as aluminum containing specialbrasses, these cracks can extend to the surfaceof the cast logs.

Unfortunately, it has to be assumed that suchdefects in the center of the billet along with thesurface defects do not weld together in the ex-trusion press but produce material separation inthe extruded section and thus scrap.

Exception: in the casting of copper billetswith high casting speeds, shrinkage cracks canoccur in the core along the grain boundarieswith limited longitudinal extension. These “spi-ders” disappear during extrusion and do nothave any detrimental effect on the section.

In the sections “Aluminum Alloys” and“Copper Alloys” in Chapter 6, detailed infor-mation on casting plants are given for the dif-ferent materials.

Deformation, Recovery,Recrystallization

Martin Bauser*

As well as the references given at the start ofsection 4.2, the following works are recom-mended for a more detailed discussion of section4.3: [See 85, Haa 84].

4.6 Deformation of PureMetals at Room Temperature

4.6.1 Dislocations

In section 4.2.1, dislocations are described aslinear defects in the crystal lattice as shown inFig. 4.6. In real lattices, they form networks andare unevenly distributed (Fig. 4.29).

If an external force is applied, a crystal latticeinitially deforms elastically following Hook’slaw. The elastic deformation is reversible—i.e.,


Fig. 4.29 Dislocation network in undeformed single crystal(transmission electron micrograph of aluminum

M � 10,000:1) [Alt 94]

the lattice returns to the initial state when theload is removed.

As shown schematically in Fig. 4.30, a plasticnonreversible deformation results in atoms inspecific planes and direction, the slip system, be-ing displaced by one or more atomic spacings.

If all the atoms on a slip plane are simulta-neously moved by one atomic spacing, a theo-retical yield stress a hundred to a thousand timesthe real yield stress is required.

Dislocations make it possible for rows of at-oms individually and successfully to spring tothe next lattice place with only a low appliedload. Figure 4.31 shows schematically themovement of an edge dislocation. As soon as adislocation has passed through a slip plane, theentire crystal part on one side of the slip planehas been displaced by one atomic spacing rela-tive to the other part.

The slip system in which the dislocationmovements occur usually consists of planes witha dense atomic packing but with larger distancesbetween each other. Their slip directions are par-ticularly densely packed atomic rows.

Figure 4.32 shows the possible slip systemsin cubic and hexagonal lattices. The planes and

directions are identified by the Miller’s indicesused in crystallography (see the additional in-formation in the appendix of this Chapter).

There are 12 favorable slip systems in theface-centered cubic (fcc) lattice and eight in thebody-centered cubic (bcc) lattice. The close-packed hexagonal spherical packing on the otherhand has only three. This gives an indication ofthe different degrees of workability of such dif-ferent structures.

If two opposite dislocations meet during theirmigration on the same slip plane, they mutuallycancel each other out. If they are located on twoadjacent slip planes, they form a vacancy chainwhen they meet (Fig. 4.33).

Whereas edge dislocations are tied to theirslip planes and can only slip on these, screwdislocations can also move in other slip planes.When they meet barriers they can move to a par-allel slip plane by “cross slip” and overcome thebarrier (Fig. 4.34).

This cross slip, however, is hindered by stack-ing faults. The slip of a dislocation by oneatomic spacing takes place in two steps whenthis is energetically favorable. The dislocation isthen split into two half dislocations (Fig. 4.35).

The larger the energy needed for the splitting,the “stacking fault energy” of a metal, thesmaller is the separation. Therefore, the stackingfault hinders the cross slip because the separa-tion has to be removed by contraction (Fig.4.34b). The stacking fault energy is therefore aspecific parameter for a metallic material—applicable not only to cross slip. It plays an im-portant role in deformation and annealing pro-cesses at higher temperatures.

4.6.2 Single Crystals

Single crystals are particularly suitable for ex-perimental investigations and for explaining thedeformation processes in metals. If a single crys-tal is loaded in tension a dislocation movingshear stress is at a maximum at 45� to the spec-imen axis. Usually, therefore, only the slip sys-tem closest to this 45� direction is activated(simple slip in the principal slip system).

During the slip, further dislocations form atsources (Fig. 4.36). The dislocations thereforemultiply but also hinder one another in real net-works. During cold deformation the length ofthe dislocation lines can increase from originally1 km/mm3 in the undeformed state by a factorof 1,000 to 10,000 (dislocation density in thecast structure: 106 to 108, after a deformation;up to 1012/cm2) [Hou 93].


Fig. 4.30 Lattice structure during deformation. (a) Elastic deformation. (b) Plastic deformation

a b c

Force

Force

Dislocation

Slip step

Slipplane

Fig. 4.31 Movement of an edge dislocation [Hou 93]

As slip continues the crystal rotates (Fig.4.37). More and more slip systems then moveinto the 45� region and are activated (multipleslip).

In these additional slip systems, the disloca-tions that move have to cross the original onesand are trapped by them. These locations formdefects at which following dislocations pile up.

Strengthening means that dislocation movementis made more difficult by these and other barriersduring the deformation.

4.6.3 Polycrystals

A polycrystal can be considered to consist ofseveral single crystals of different orientations.


a) Face-centered cubic, e.g. Cu, Al

Plane (111)(4 possible planes)

Direction [110](3 in a plane)

Plane (211) (12 possible planes)





Direction [1120]Direction [1120](3 in a plane)

Plane (0001)

Slip system (111)[110]

Slip system (0001)[1120] Slip system (1010)[1120]

Slip system (110)[111]Slip system (211)[111]

Fig. 4.32 Slip systems

Each single crystal will deform as discussedabove in its principle slip system. The bound-aries with the neighboring crystal, however,hinder this free movement.

In order to retain the cohesion between theindividual deforming crystallites, at least five

slip systems have to be active in each grain,which contributes to the strengthening.

Whereas in single crystals dislocation linescan disappear at the surface, they pile up at grainboundaries in polycrystals. If no grain boundaryslip occurs, then the higher the number of grain


Fig. 4.33 Interaction of parallel dislocations of differentsigns. (a) Cancellation. (b) Formation of a vacancy

Fig. 4.34 Cross slip of a screw dislocation [Hou 93]. (a)Nondivided dislocation. (b) Split dislocation

Fig. 4.35 Dislocation split by stacking defects

boundaries, i.e., the finer the structure, thegreater is the strengthening (Hall-Petch relation-ship).Twins can form in individual grains during

working to relieve the severe internal stressesformed during the deformation of polycrystals(Fig. 4.8).

With severe deformation of a polycrystal, theindividual grains extend and align themselves byrotation so that crystal orientations accumulatein the deformed state in specific directions. Thisis referred to as a deformation texture describedby Miller indices and pole figures (see Appen-dix, “Additional Information 2”).

The deformation textures of different materi-als can vary. Also, different methods of defor-mation produce different textures. A rolling tex-ture is not the same as a drawn texture.

Special extrusion textures occur when recrys-tallization during the extrusion process is com-pletely or partially suppressed [Bag 81].

Textures also signify differences in the defor-mation behavior in different directions. In ex-truded bars that have not completely recrystal-lized, e.g., high-strength aluminum alloys, thestrength in the extrusion direction is higher thanin the transverse. This is referred to as the ex-trusion effect.

The stress-strain diagrams of polycrystallinematerials frequently do not exhibit a markedyield point because individual favorably orien-tated grains start to flow earlier than others: thetransition from the purely elastic to the plasticdeformation state is thus smooth. The 0.2% de-formation Rp0.2 is therefore defined as the “yieldstress” at which a permanent deformation of0.2% occurs. The pronounced yield point, whichis seen not only in steels, is described in thefollowing section.


Fig. 4.36 Production of dislocations [Fra 50]

Fig. 4.37 Rotation of single crystal during deformation

4.7 Deformation of Alloys atRoom Temperature

The deformation, i.e., the movement of dis-locations, can be hampered in many ways. For-eign atoms embedded in the solid solution or assecondary phases in the form of inclusions orprecipitates and the self-producing obstructionsfrom secondary slip systems described previ-ously and neighboring grain boundaries.

In solid solution hardening, a substitutedforeign atom in the base lattice causes a localdistortion field the size of which increases thegreater the difference between the atom radii.

However, even with the base and foreign atomshaving approximately the same size, interactionsstill take place because the foreign atoms influ-ence the bonding forces between the atoms.

The movement of dislocations is thereforehindered by the foreign atoms. Because theyhave to cross their stress fields—analogous toslip friction between two solid bodies—this isreferred to as solid-solution friction, which in-creases the deformation resistance.

The higher mechanical properties of the so-called naturally hard aluminum alloys (AlMg,AlSi) are attributable to solid-solution friction.

The action of foreign atoms in solid solutionsis relatively weak compared with the particlehardening described subsequently. The yieldstress and the increase in strength, however, doincrease significantly with the fraction of dis-solved atoms (Fig. 4.38).

Even at slightly increased temperatures, for-eign atoms can accumulate at dislocations bydiffusion particularly where they reduce thestress field of the dislocation. A dislocation isanchored by these foreign metal clouds (Cottrellclouds) and can only be released and set inmovement by higher forces, which produces amarked yield stress.

Interstitial foreign atoms, e.g., carbon in iron,move more easily than substitutional atoms.Carbon clouds around dislocations are the causeof the well-known marked yield point of carbonsteels.

Particle Hardening. The second phases in analloy can take on many different shapes, sizes,and distribution, which determine the effect ofthe particles on the dislocation movement.

Dislocations can pass through so-called softparticles; they are “cut” (Fig. 4.39a).

One part of such a particle is displaced by oneatomic spacing in the same way as the entire


Cu-In

Cu-Au

Cu-Ge

Cu-GaCu-Al

Cu-Zn

Cu-Ni

100

200

0.1 0.2

c2/3

, MP

a

Fig. 4.38 Yield stress of single-phase copper alloys as afunction of the content of foreign atoms

Fig. 4.39 Interaction of dislocations and particles [Ber 83]. (a) Intersection of “soft” particles. (b) Bypassing of “hard particles”

lattice along a slip plane. This requires the ap-plication of a higher force compared with dis-location movement through a single-phase lat-tice. The dislocation in the cutting process istherefore stretched like a rubber band and thenrelaxed after cutting. The greater the number ofparticles, the stronger is the effect. Only smallcoherent or approximately coherent precipitatesin which the matrix and particle lattices havesimilar structures and orientations are cut.

An incoherent particle—which includes mostprecipitates—forms a barrier that cannot beovercome by cutting. The dislocation managesto go around these “hard particles” (Orowan pro-cess, Fig. 4.39b); as soon as a dislocation linehas been stretched to such an extent that thepositive and negative parts move close together,they meet and mutually cancel each other out.A dislocation ring remains around the particle,and this repels the following dislocation. This isbecause dislocations of the same sign mutuallyrepel each other and those of the opposite signattract. Therefore, with increasing deformation


Optimum age hardening

Optimum particle size

Overaging

Smallparticles

Largeparticles

tA, min

Rp

0.2, M

Pa

10 100 1,000 10,000

400

350

300

250

200

Fig. 4.40 Yield stress of age-hardening aluminum alloy as a function of the aging time [Blu 93]

and thus an increasing number of dislocationrings around the particles, an additional type ofhardening takes place. This is referred to as pre-cipitation hardening or dispersion hardening.

The size of the particle does not influence thistype of hardening, but their number and distri-bution do. The higher the number of particlesthat cannot be cut, the greater is the resistanceto dislocation and thus the strengthening.

In hot age hardening, e.g., of a heat treatableAlMgSi alloy, as shown in Fig. 4.40, small andcoherent particles of the second phase first formso that they can be cut. As the aging time in-creases, the particles become incoherent and in-crease in size and can no longer be cut, whichincreases the strength. If coagulation occurs af-ter longer heat treatment, as described in Fig.4.15, large particles grow at the expense ofsmaller ones and the number of particles de-crease. The strength and hardening then reduce.This is known as overaging.

AlCu alloys, for example, exhibit an increasein strength even with cold age hardening (agingafter solution heat treatment at room tempera-ture) by the formation of coherent groups of cop-per atoms in so-called Guinier-Preston (GP)zones. They redissolve, however, during subse-quent hot age hardening before the hot-age-hardening phases form.

Targeted alloy development and heat treat-ment enabled extremely high-strength materialswith specific deformation properties to be de-

veloped. In addition to solid-solution hardeningand precipitation hardening, they involve the in-clusion of foreign particles as dispersions, e.g.,by powder metallurgical processes.

4.8 Higher Temperatureswith Pure Metals

As the temperature increases, the thermalmovement of the atoms in the crystal increases,which increases the ease of transposition pro-cesses by, for example, diffusion.

The climb of edge dislocations also takesplace (Fig. 4.41). In this process, completeatomic rows are removed or added; the dislo-cations are thus moved into another parallel slipplane. The climbing, therefore, naturally is onlyactivated when it results in a reduction in theinternal stresses and thus to a lowering of theinternal energy of the crystal. This can take placeby positive and negative dislocations mutuallycanceling each other and thus reducing the dis-location accumulation.

4.8.1 Annealing of Pure Metals

The first step in stress reduction by annealingof a deformation hardened structure is recovery.The crystallites retain their shape and orienta-tion.

Dislocation freed from their forced locationby climb and cross slip mutually cancel each


Fig. 4.41 Climb of an edge dislocation [Sch 81]

Fig. 4.42 Subgrain formation in Al99.99 (hot strip). Thewidth is approminately 8.4 lm [Grz 93]

other out, as described previously (positive andnegative dislocations), or move and accumulatein low stress subgrain boundaries, which arecomposed of a chain of dislocations of the samesign.

After recovery heat treatment an entire net-work of these subgrain boundaries is found—this is referred to as polygonization of the crys-tallite (Fig. 4.42). The orientation of theresulting subgrains differ only slightly from eachother. The interior of the subgrains is practicallydislocation free.

During recovery the internal stresses are notcompletely eliminated because there is still a

high volume of dislocations in the subgrainboundaries.

Stacking faults hinder not only cross slip butalso the climb of dislocations and thus the re-covery of metals. Low stacking fault energymetals, i.e., large stacking faults such as copper,nickel, and austenitic steels, therefore exhibit alow tendency to recovery. On the other hand,aluminum and alpha iron with high stackingfault energies and thus small stacking faults eas-ily form subgrains during suitable heat treatmentand therefore recover.

If a deformed structure is heat treated at a hightemperature, recrystallization takes place. Newgrains that are largely free of dislocations andthus subgrains are formed during this process.The structure is free of internal stresses andtherefore annealed. The deformation behaviorlargely corresponds with the undeformed caststructure.

Figure 4.43 shows schematically how ahealed recrystallized grain grows into a dislo-cation rich work-hardened lattice.

Recrystallization usually occurs during a suit-able heat treatment of a deformed metal withoutpreceding recovery by nucleation after an incu-bation time.

The nucleation is followed by the growth ofa grain that is sustained as primary recrystalli-zation until all the deformed crystals have dis-appeared. In the subsequent secondary recrys-tallization (large grain formation), large grainsgrow at the expense of small recrystallizedgrains. The internal energy is reduced further bythe shortening of the total grain-boundary sur-face. This is usually undesirable because of theassociated irregular grain size [Alt 94].

Figure 4.44 shows the typical recrystallizationprocess with time. As can be seen, after the in-cubation time there is initially a flat profile dur-ing the nucleation and a steep increase duringthe primary recrystallization. The gradient of thecurve decreases with the subsequent secondaryrecrystallization.

The recrystallization nuclei form mainly at re-gions of high dislocation density. These can oc-cur at grain boundaries between relatively un-deformed original grains. Precipitates can alsoact as recrystallization nuclei [Wei 94, McQ 75].

During primary recrystallization, these nucleigrow by thermally activated migration of thegrain boundaries into the work-hardened struc-ture and consume it. The number of newlyformed largely defect-free grains correspond tothe number of nuclei capable of development.


Dislocationfree lattice(recrystallized)

Dislocationrich lattice(work hardened)

Grainboundary

Fig. 4.43 Growth of recrystallized grain

Rec

ryst

alliz

ed fr

actio

n X

Time, s

350 °C

1.0

0.8

0.6

0.4

0.2

0 2,000 4,000 6,000 8,000 10,000

Fig. 4.44 Variation with time of recrystallization (Al99.5 after 5.1% tensile deformation) [And 48]

The rate of migration of the new grain bound-aries and thus the rate of recrystallization is thegreater the larger the preceding deformation andthus the larger the energy stored in the deformedstructure. Naturally, a higher temperature pro-motes the grain-boundary migration and in-creases the rate of recrystallization.

Recovery preceding recrystallization, e.g., atlow heat treatment temperatures, that has re-duced the deformation energy stored in latticeby sub grain formation hinders recrystallization.In metals with high stacking fault energies like

aluminum, it can even be completely suppressedby recovery—at low heat treatment tempera-tures and small deformation.

The recrystallization temperature abovewhich recrystallization can be expected is usu-ally related to the melting point of a metal:

T � 0.6 • T [�K]recr s

This recrystallization temperature should not beconsidered unequivocal.


Strain

0 0.25 0.5 0.75 1.0

25

20

15

10

5

0

838 K

943 K

1033 K

1103 K

1273 K

She

ar s

tres

s, M

Pa

Fig. 4.45 Work-hardening curves of copper single crystalsas a function of temperature [Got 84]

With deformed brass CuZn37, for example,the softening from recrystallization after highdeformation starts at 400 �C but only at 500 �Cafter low deformation. With pure aluminum after98% deformation recrystallization commencesat 300 �C and at 400 �C after 20% deformation[Alt 94].It is, therefore, more correct to refer toa recrystallization interval instead of a recrys-tallization temperature.

If a structure textured by deformation is givena recrystallization heat treatment, the new struc-ture can also exhibit preferred orientation, a so-called recrystallization texture. One cause can bethat grains of a specific orientation in the de-formed structure are less deformed than others.Nuclei of preferred orientation then grow fromthem into the severely deformed grains.

Deformation and recrystallization textures areclosely related to each other but usually differ.Recrystallization textures are also characteristicfor the deformation process and for the relevantlattice structure (e.g., AlCuMg [Moc 75]).

4.8.2 Deformation of PureMetals at High Temperatures

If a pure metal is deformed at a higher tem-perature, both the yield stress and the work hard-ening are reduced (Fig. 4.45).

As the temperature increases, the dislocationscan start to migrate through their slip systems atlower shear stresses due to thermal activation,which reduces the yield stress. The work hard-ening—that is, the gradient of the strengthcurve—is reduced because the vacancy diffu-sion at higher temperatures is also eased andhence piled up dislocations can move to ener-

getically more favorable positions during the de-formation by cross slip and climb. This meansthat recovery processes can proceed with the for-mation of a substructure during the work-hard-ening—and not only after heat treatment follow-ing cold working [Mug 84]. This is calleddynamic recovery.

If deformation is carried out quickly enoughat a high temperature, the work hardening canbe so high despite the start of recovery that re-crystallization can take place even during the de-formation—this process is frequently observedin the extrusion of specific materials. It is re-ferred to as dynamic recrystallization [Got 84].

The stacking fault energy plays an importantrole in the boundary between recovery and re-crystallization during hot working.

High stacking fault energies and thus fewerdivided dislocations occur in bcc alloys (e.g., �-iron, b-brass) as well as in certain fcc alloys(e.g., aluminum). In these cases, the cross slipand climb needed for subgrain formation and re-covery is easier—recovery is more likely to takeplace in these cases in preference to recrystalli-zation during hot working. Low stacking faultenergies associated with large divided disloca-tions are found with a few other fcc lattices (e.g.,copper and copper alloys). In these cases, recov-ery is more difficult, and dynamic recrystalli-zation processes simplified during hot working.

The strength reducing recovery and recrystal-lization processes during hot working can takeplace so quickly after reaching a specific workhardening that there is equilibrium between soft-ening and work hardening and the hardeningcurve remains horizontal (static case). Naturally,the dynamic recovery processes also result in adelay in the fracture-initiating mechanisms (seesection 4.11), and the elongation to failure intensile tests increases with increasing tempera-ture.

If the events occurring within the structure inhot working are studied, it is clear that these pro-cesses depend not only on the temperature butalso on the rate of deformation.

The slower the deformation rate, the greateris the time for the recovery and recrystallizationprocesses to counteract the work hardening fromthe multiplication and piling up of dislocations.The work-hardening curve is, therefore, flatterand its horizontal leg lower the slower the de-formation rate. At a fast rate of deformation, arelatively high flow stress has to be expected.This explains the high dependence of the defor-mation resistance on the rate of deformation in


Str

ess,

MP

a

100

50

0

Elongation, %

1 2 3

d /dt = 3.96 s–1

Fig. 4.46 Strain-rate dependence in hot working (stress-strain curves of nickel) [San 75]

hot working compared with cold working (Fig.4.46).

The dependence of the flow stress on the tem-perature and the deformation rate in hot workingis discussed in section 3 in the Appendix, “Ad-ditional Information 3,” in this chapter.

The stress-strain curves obtained from torsiontests cover a much larger strain to failure thantensile tests because of the largely tensile-stress-free deformation.

They are therefore more suitable for assessingthe behavior of metals in practical hot working,in particular extrusion, which also does not in-volve any tensile stress (Fig. 4.47) [Fie 57, Ake66].

The hot stress-strain curves obtained from tor-sion tests exhibit after the initial increase, inwhich the number of dislocations and the inter-nal stresses increase significantly, a large sta-tionary region with hardly any additional workhardening. The work-hardening and recoveryprocesses are in balance. The stress-strain curvestarts to fall only when structural defects resultin fracture.

Usually, no recrystallization occurs during thehot deformation of aluminum and many of itsalloys. Subgrains are more likely to develop byrecovery and their shape and size hardly changeduring the stationary region. If the load is re-moved and the temperature maintained or slowcooling permitted, recrystallization can occur—delayed significantly by the low internal energy.One example is the recrystallization during ex-trusion of low-alloyed aluminum alloys afterleaving the die [McQ 90].

4.9 Alloys at Higher Temperatures

4.9.1 Annealing of Alloys

In the annealing of deformed alloys—fromthe point of view of the deformation energy—inwhich there is thermal equilibrium at the an-nealing temperature, recovery and recrystalli-zation processes play the main role as with puremetals. However, with single-phase alloys, thethermal activation of the climb and cross slipprocesses initiating thermal activation is moredifficult because of the presence of foreign at-oms. This also applies to the migration of grainboundaries during recrystallization becausegrain boundaries are frequently enriched withdissolved atoms and have to be torn away fromthem.

If the second phase exists in the form of fineprecipitates (in the range of 0.04–0.4 lm diam),these also tend to retard recrystallization becausethey can exert a significant resistance to the nec-essary migration of grain boundaries, particu-larly where they accumulate. This can occur tosuch an extent that at a moderate annealing tem-perature recrystallization can be completely sup-pressed and only recovery takes place. Thestrength then does not fall completely during an-nealing and the deformation texture remains.

With two-phase alloys with large crystallitesof the second phase, it is possible that only onephase with a low recrystallization temperaturerecrystallizes during the annealing process,whereas the second phase remains unrecrystal-lized.

Still more complicated is the case when a de-formed alloy is in thermal unequilibrium duringannealing. Softening processes then occur by re-covery and recrystallization simultaneously withprecipitation processes (or, in rare cases, disso-lution processes) of second phases [McQ 90,McQ 90a]. The dependence of the mechanicalproperties on the annealing conditions is there-fore complicated and difficult to explain.

There are more vacancies in deformed metalsthan in undeformed metals, which simplifies thediffusion of foreign atoms to precipitates. There-fore, the precipitation processes during anneal-ing take place more quickly in the deformedstructure than in the undeformed state.

It is possible today to construct high-strengthmaterials in which specific alloy composition,heat treatment and deformation precipitationhardening and work hardening are combinedwithout any significant softening processes tak-


Str

ess,

MP

a

Str

ess

k f, MP

a

300

200

100

0

200 °C

300 °C

400 °C

500 °C

600 °C

600 °C

650 °C

700 °C

750 °C

800 °C

850 °C

700 °C

200

100

00 0.5 1.0

Elongation Log strain %g

d /dt = 3s–1

(a) (b),

0 10 20 30 40 50 60 80 70

Fig. 4.47 Hot low curves for CuZn8 obtained from different methods. (a) Hot-tensile tests. (b) Torsion tests [Bau 63, DGM 78]

ing place during the permitted time at the in-tended working temperature [Ber 83].

4.9.2 Deformation of Alloysat Higher Temperatures

In single-phase alloys, the embedded foreignatoms hinder the migration of dislocations as incold working but also impair dynamic recoveryand recrystallization. The hot yield stress, workhardening, and hot strength increase with in-creasing foreign atom content (Fig. 4.48).

Precipitates, which retard static recrystalliza-tion, also naturally hinder dynamic recrystalli-zation processes. Hot-strength curves are there-fore steeper with alloys with precipitates andattain the stationary (horizontal) state later or ata higher temperature.

The high hot strength of high-alloy materialsnaturally makes hot working more difficult andthus materials “more difficult to extrude.”

Because, as described above, a high vacancyconcentration accelerates the process of precipi-tation processes, the rate of precipitation in-creases during the hot working of an alloy notin equilibrium.

The factors that determine whether and atwhat rate recovery and recrystallization takesplace during hot working are the same as thosegoverning annealing after previous cold work-ing.

4.10 Processes in Extrusion

Billet heating before extrusion cannot be con-sidered to be independent of the previous hotprocesses—casting, hot forging, or homogeni-zation. In addition, the temperature and time ofheating are determined by various, often con-flicting, requirements:

● Low extrusion pressure● High extrusion speed● Limitation of the exit temperature to a max-

imum (surface quality) and to a minimum(recrystallization, with aluminum alloys so-lution heat treatment during extrusion).

When determining the preheat temperature,the deformation heat during extrusion (increas-ing the extrusion temperature) and the heat flowinto the tooling (reducing the extrusion tem-perature) have to be taken into account.

If nonhomogenized billets of age-hardeningalloys are used, the homogenizing effect duringbillet heating increases the longer it lasts. In thiscase, slower gas furnace heating is more effec-tive than fast induction heating. Not only cansegregations harmful to extrusion be reduced,but also precipitates dissolved and supersatura-tions reduced.

In contrast, previously homogenized billets ofsuitable aluminum alloys cooled to a specified


Fig. 4.48 Hot tensile strength of different alloys. (a) Alumi-num alloy. (b) Brasses [Alt 94, Wie 86]

Fig. 4.49 Temperature variation in the processing of age-hardening aluminum alloys by extrusion. 1, Heat-

ing the billet; 2, transfer to the press; 3, extrusion � heating fromthe deformation � solution heat treatment; 4, section cooling; 5,elevated temperature age hardening. RT, room temperature[Alt 94]

degree of solubility have to be heated as quicklyas possible and not too high in order not to de-stroy the desired structure. Induction heating isthen recommended.

It is possible with AlMgSi alloys and CuCrand CuCrZr alloys to carry out extrusion at thesolution annealing temperature and thus avoid aseparate solution heat treatment prior to agehardening (Fig. 4.49).

With brass CuZn39Pb, which has an �-bstructure at room temperature, the billet heatingcan develop a structure corresponding to thephase diagram (see the section “Extrusion ofSemifinished Products in Copper Alloys,” Fig.5.51 in Chapter 5), in which the � brass that haspoor hot-working properties is dissolved andonly the readily hot-worked b brass is present.

The surface of the extruded billet quickly de-velops the temperature of the container linerwall, which at approximately 500 �C is coolerthan materials extruded at moderate and hightemperatures. There are materials—such as thefree machining brass CuZn39Pb in which thesurface cooling of the billet increases the contentof the difficult-to-deform � brass—where theflow behavior close to the container wall dras-tically changes. The b-rich core then flows for-ward and the �-rich periphery is held back, re-sulting in the extrusion piping defect (see“Extrusion of Semifinished Products in CopperAlloys,” Fig. 5.56, in Chapter 5).

Dynamic and recrystallization processes takeplace in front of and in the deformation zone ofthe die during extrusion. The stacking fault en-ergy of the extruded metal plays an importantrole here. With a low stacking fault energy (cop-per, �-brass, austenite), dynamic recrystalliza-tion occurs in the extrusion deformation zone ofthe extrusion press at moderate and high extru-sion speeds. On the other hand, aluminum andits alloys with a high stacking fault energy canrepeatedly only recover—the extrusion leavesthe press soft but with a fibrous unrecrystallizedstructure. However, in some cases, recrystalli-zation processes can occur in the hot extrusionafter it leaves the die.

The sketch in Fig. 4.50 schematically showsthe influence of the high and low stacking faultenergy.

In the shear zones (at the liner wall, at thefunnel of the dead metal zones in flat dies andwithin hollow dies) the deformation in alumi-num is up to 60 times larger than in the core[Ake 92].


Fig. 4.50 Dynamic recovery and recrystallization processes in extrusion. (a) High stacking fault energy. (b) Low stacking faultenergy [McQ 75]

If the fibers with higher deformation (e.g.,close to the surface of the extruded section) arevery elongated and thin (width approximatelytwo subgrains), adjacent grains can grow to-gether and the fibers appear shorter [McQ 90].

The grain structure, however, is so severelydistorted in the shear zones that no grain bound-aries are visible under the microscope and theprecipitates present are randomly distributed.AlMgSi0.5 sections recrystallize with fine grainsbut only after leaving the die (static and not dy-namic).

In alloys with recrystallization retarding ad-ditions (e.g., AlMgSi or AlCuMg with additionsof Mn, Cr, or Zr), the recrystallization tempera-ture is raised to such a level that no recrystalli-zation occurs even after leaving the deformationzone (particularly following rapid cooling).

However, subsequent solution heat treatmentreveals that the peripheral zone from the shearzone is more heavily deformed than the coreand, therefore, statically recrystallizes, oftenwith very coarse grains, whereas the moreweakly deformed core remains unrecrystallized.

The unrecrystallized structures usually have astrong texture because of the high deformationin the extrusion press, whereby the strength inthe longitudinal direction is much higher than in

the transverse direction. This is referred to as theextrusion effect.

Recrystallization retarding additions in suit-able aluminum alloys can, however, precipitatecoarsely during annealing at high temperatures(defective homogenization, billet heating in agas furnace). A second phase that has no effecton the recrystallization then forms (e.g., man-ganese in AlMgSi). Then the aluminum materialcan, as described earlier for AlMgSi0.5, fullyrecrystallize after leaving the die—while thesection is still hot.

In the production of aluminum sections witha decorative surface, structural variations andtextures can produce flecks and stripes. Becauseonly the zone immediately below the surface ofthe section plays a role, attention has to be paidnot only to the billet production (see section4.2.4) but also to the die design. The proceduresthat have to be taken into account in the manu-facture of aluminum sections are described indetail in the section “Extrusion of SemifinishedProducts in Aluminum Alloys” in Chapter 5.

Independent of the material, in extrusionwithout lubrication the zones close to the surfaceof the extruded zone are hotter than the core be-cause of the concentration of the deformation inthe shear zone and the friction between the die


Intercrystalline brittle fracture

Cleavagefracture

Plastic growth of voids(transcrystalline) (intercrystalline)

Cracking from necking or shearing

Fig. 4.51 Different failure methods (in tensile testing) [Ash 79]

land and the material. However, at the surfacethis friction also produces tensile stresses as thecore of the extrusion accelerates. It is thereforeno surprise that the billets with residual meltsegregations with low melting points producecrack-sensitive extrusions. If these segregationsare not previously eliminated by homogeniza-tion, the only option is to extrude at low tem-peratures and slow speeds. This prevents thetemperature of the section surface from reachingthe dangerous region of melting, which resultsin transverse cracks (see the section “Extrusionof Semifinished Products in Copper Alloys,”Fig. 5.45, in Chapter 5).

In the extrusion of copper alloys and steels,spontaneous recrystallization usually takes placeduring the deformation as the extrusion tem-perature is well above the recrystallization tem-perature (which is 0.8–0.9 times the meltingpoint TS). It can happen that the start of extru-sion, where the deformation is low and the tem-perature high, has a coarse secondary recrystal-lized structure, but the end of extrusion—wherethe deformation is high and the exit temperaturelower due to cooling during extrusion—is finegrained.

In pure (and low-alloyed) copper so few re-crystallization retarding precipitates are presentthat the grain boundaries migrate relatively eas-ily. After the primary recrystallization in the de-formation zone, a secondary coarse recrystalli-zation has to be expected in the emergingextrusion (at 800–900 �C). Rapid cooling is useddirectly behind the die to prevent this.

4.11 Toughness, Fracture

The toughness of a material is—for the sametemperature and the same deformation rate—

dependent on the deformation conditions. Forexample, in a tensile test the toughness definingelongation to fracture is relatively low, but in thetorsion test in which the deforming elements aresubjected to other forces, it can be higher. Inparticular, in extrusion with its high hydrostaticcontribution to the deforming loads much higherdeformations can be achieved than in the simpletensile test. The phenomena occurring in the ten-sile test up to fracture should be understood (Fig.4.51). A more detailed explanation is found in[Rie 87].

4.11.1 Ductile Fracture

In metal working the most common type offailure is ductile fracture following deformation.As a rule, the deformed metal is not defect free.During the deformation in the tensile test, ini-tially small voids form at internal boundary sur-faces with low adhesion, e.g., at inclusions,coarse precipitates or at existing gas pores,where the stress concentration becomes highenough during the working. Only very pure met-als deform in tensile tests without this void for-mation.

The cross section of the tensile specimen inthe region of the voids is reduced. The tensileforce therefore exerts a higher stress, whichleads to higher elongation in this zone, resultingin a localized necking. Once the necking hasstarted this area continues to reduce in cross sec-tion as the stress increases, whereas the sectionsof the specimen that have not necked elongateonly slightly. Soon the pores in the necked zoneincrease in size and grow together until fractureoccurs between them. The fracture surfaces arerough and consist of craterlike depressions, eachcorresponding to an original void. The ductilefracture usually passes through the grains and istherefore transcrystalline.


At higher temperatures, as discussed previ-ously, the yield stress and the work hardeningare reduced by the onset of the recovery pro-cesses. This means that the stress needed to nu-cleate the voids, e.g., at inclusions, is firstreached after a higher deformation. The elon-gation to fracture can then increase.

However, at higher temperatures the flowstress and the work hardening react more sen-sitively to the deformation rate, and accordinglythe fracture behavior varies at higher speeds. Asa rule, the elongation to fracture decreases as therate of deformation increases.

Grain-boundary decoration by foreign atomsor inclusions and thin films of segregations canweaken the grain boundaries at high tempera-tures to such an extent that the crack propagatesin an intercrystalline manner.

4.11.2 Brittle Fracture and Mixed Types

Pure deformation-free brittle fracture is rare;more common is low-deformation brittle frac-ture. Neither of these fracture types occurs dur-ing extrusion. Their explanation does, however,increase the understanding of the possible frac-ture processes.

At sufficiently low temperatures most metalsfail without or with only slight deformation bybrittle fracture. The starting point is a crack thatis already present or forms after slight plasticdeformation. The tip of this crack acts as anotch; i.e., a high stress concentrates here undera tensile load, which, when the so-called crackpropagation stress is exceeded, results in a de-formation free or low deformation cleavagefracture. The fracture surfaces are smooth.

With increasing temperature, the stress peakthat forms at the initial crack under load is re-duced by local deformation so that the crackpropagation stress is reached only after higherwork hardening, in other words, higher plasticdeformation. The tendency to brittle fracture re-duces with increasing temperature or disappearscompletely.

The cleavage fracture propagates along sur-faces with low cohesion. If these are specificcrystallographic planes, then the fracture istranscrystalline. With weakened grain bound-aries attributed to higher temperatures or accu-mulated foreign atoms, segregations, or precip-itations the fracture propagates between thegrains and is thus intercrystalline. Frequently, afracture shows both transcrystalline and inter-crystalline components.

If the deformation is carried out at such a hightemperature that existing segregations start tomelt, an almost completely brittle fracture oc-curs.

4.11.3 Fracture during Hot-Working

The processes leading to fracture are now ex-plained using hot torsion tests. Torsion tests cor-respond more closely to the processes in extru-sion, but only a few suitable measured resultsare available.

Elongation and necking traverse in hot-tor-sion tests one or two minima. The examples ofE-copper and of phosphorous containing tinbronze illustrate this (Fig. 4.52). With E-copper(Fig. 4.52a), the elongation to fracture and thenecking initially reduce with increasing tem-perature (up to a minimum of approximately 500�C) and then increase. As the temperature in-creases, severe grain-boundary cracks form per-pendicular to the tensile direction. They formand grow by the migration of vacancies to thegrain boundaries associated ultimately withweakening due to grain-boundary accumulation.Above the minimum temperature, where theelongation to fracture increases again, the pro-cesses of recovery and, in particular, recrystal-lization occur. Existing voids are dissolved andthe susceptibility to fracture reduced.

A second minimum similar to that seen with8% tin bronze (Fig. 4.52b) can be related to theeffect of segregations that have a high mobilityand readily melt [Bau 63]. In the example of thetin bronze it is a low melting point phosphoruscontaining phase on the grain boundaries.

4.11.4 Defects in Extrusion

Tensile stresses are usually required for crackformation and propagation. High shear stressescan also result in cracks—referred to as shearcracks.

In extrusion, the material in the die defor-mation zone is under three-dimensional com-pression. Only when it has passed through thedie will the tensile stress developed close to thesurface by the friction on the die bearing havean effect. The magnitude of this tensile stressdepends on the friction and the die design. Inlubricated extrusion (glass lubrication with steeland nickel alloys, hydrostatic extrusion) it islower than in unlubricated (aluminum) or if onlya flat die is lubricated (copper). The tensilestresses at the surface of the extruded section


Elongation to fracture

Elongation to fracture

Tensile strength

Tensile strength

Yield stress

Yield stress

Testing temperature, °C

Testing temperature, °C

Yie

ld s

tres

s Te

nsile

str

engt

h, M

Pa

Yie

ld s

tres

s Te

nsile

str

engt

h, M

Pa

500

400

300

200

100

0

100

80

60

40

20

0

80

60

40

20

0

400

300

200

100

0

100 200 300 400 500 600 7000

100 200 300 400 500 600 7000

(a)

(b)

Fig. 4.52 Elongation to fracture as a function of temperaturein tensile tests. (a) E-copper. (b) Tin-bronze CuSn8

Fig. 4.53 Cracked surface of a CuZn28Sn (longitudinal sec-tion/surface image). (a) The width is approximately

0.66 mm. (b) The width is approximately 1.4 mm [Die 77]

combined with the temperature increase in theperipheral zone result in transverse cracks (firtree defect) if an existing grain-boundary accu-mulation is severely weakened or even melted.Under the microscope these grain-boundary ac-cumulations can be detected along with the pre-dominantly intercrystalline nature of the cracks(Fig. 4.53).

Internal cracks in tubes have the same origin:if the friction between the material and the man-drel is high enough at the extrusion temperatureto destroy the cohesion between grain bound-aries transverse to the direction of extrusion,transverse cracks will form on the internal sur-face of the tube.

At a small extrusion ratio the core acceleratesso quickly during extrusion relative to the sur-face zone that the shear stresses are sufficient toseparate a surface shell from the core. This alsorequires correspondingly weakened zones.

With very small extrusion ratios, conical-shaped transverse cracks can develop becausetensile stresses can occur along the billet axisbefore the die plane. These can result in voidsand cracks at the affected areas.

Both types of defects—the peeling of a sur-face shell from the core and the core cracking—cannot be detected on the surface. Materials atrisk must be tested using nondestructive meth-ods when necessary.


4.12 Solid-State Bonding

Materials can bond together under pressureand at an elevated temperature without macro-scopic shape changes (diffusion bonding). How-ever, at least one of the bonding partners has toundergo a significant compressive deformation(compression, rolling, extrusion).

The bonding processes that occur during de-formation is the reason for the use of the extru-sion process in many cases.

The weld seams in hollow profiles, the pro-duction of composite profiles and the processingof composite materials as well as the extrusionof metal powders are discussed in this section.

4.12.1 Bonding Processes

The most common methods of joining twocomponents are welding and soldering. In bothcases the liquid phase takes the dominant role.In welding, the boundary layer between thecomponents melts. The resolidified melt formsthe “weld seam.” In soldering, both the parts re-main in the solid state, only the solder betweenthem melts and solidifies. The joint is formed byadhesion and diffusion at the boundary surface.

When bonding in the solid state—referred toas pressure welding—the formation of a liquidphase usually is not required. The most commonmethods are roll cladding, friction welding, andultrasonic welding (more information on pres-sure welding is found in [Tyl 68].

If two metallic surfaces meet, they can bondtogether if the surface atoms of both componentsare approximately a lattice spacing apart. Thefaces then act as a grain boundary and the me-tallic bonding forces of the electron gases actacross the boundary surfaces. Experience showsthat this adhesion process rarely occurs withoutthe contribution of pressure and temperature.Absolutely flat and clean boundary surfaces arerequired and it is made easier or possible at allif the two components pressed together haveidentical, or similar, lattice structures.

A higher surface pressure always flattens verysmall irregularities and improves by plastic de-formation of the roughness peaks the contact be-tween the components pressed together over themaximum contact area. With pure aluminum thepressure welding commences after only 10% de-formation of the two surface areas; however, itis only after 70% that the tear resistance of thejoint matches the value of the parent metal.

At a higher temperature, this flattening andsmearing together of the two surfaces is simpli-

fied, atoms changing places with each other canfit with the opposing lattice so well that the bondbetween the atoms at the boundary surface isoptimized. There are hypotheses that the dislo-cations stacked together by pressure welding cantraverse the boundary surface into the other part-ner.

Any existing coatings, e.g., fine oxide skinsor films of lubricant, hinder any adhesive bond.If the areas of the components being bonded areincreased during the bonding process, e.g., inroll cladding or extrusion, brittle coatings suchas oxides cannot follow this expansion. Instead,they break up into individual pieces. Fresh metalforms between the pieces on the surface and thiscan now bond with the opposite side. Pressurewelding requires both high pressure and a largedeformation and thus an increase in the surfacearea. Since, in roll cladding and extrusionthrough a hollow section die there is also rele-vant movement between the two partners andfrequently polyganization or even recrystalliza-tion occurs by recovery processes the grains or-ientate themselves into matching positions. Ifthe bond is very good, it is impossible to detectthe joint under the microscope (see Fig. 4.55).

If the surface coating can follow the increasein the contacting areas, i.e., smear, then bondingis not possible during the contacting process.One example is the damaging effect of lubricantin the extrusion of hollow aluminum alloy sec-tions. This can prevent the formation of a soundpressure weld.

4.12.2 The Formation of PressureWelds in Hollow Sections andduring Billet-on-Billet Extrusion

If hollow sections are produced using bridgeor porthole dies as described in Chapter 5, in thesection on the extrusion of aluminum alloys, thedivided strands bond together in the weldingchamber immediately in front of the die into twoor more longitudinal seams. These are not welds,as sometimes stated, but metallic bonds fromsolid-state adhesion. In these die zones, whichare more correctly referred to as bonding cham-bers, the high transverse forces—almost equalto the hydrostatic pressure in the undeformedbillet—act on the surfaces of the divided strandsproducing an increase in the surface area by thereduction of the cross section in the chamber.The adiabatic heating from the deformation alsoplays a role. The strands that are separated justprior to the bonding—and with the exclusion of


Fig. 4.54 Breakup of the surface layer in longitudinal welds in the production of aluminum hollow sections. (a) Meeting under theleg. (b) Breaking apart. (c) Clean longitudinal weld [Ake 92]

Fig. 4.55 Extrusion weld in aluminum. (a) Different etchantattack. (b) Full recrystallization across the weld

[Ake 92]

air—are free of any oxide or lubricant surfacecoating, assuming a clean die is used, that wouldhinder the bonding.

Oxide is brought into the welding chambersof the die only during the first filling when thefree oxide coated end faces come together (Fig.4.54a) [Ake 92]. The fractured remains of theoxide surface are mostly carried away with thefirst part of the section (Fig. 4.54b). The subse-quent bonding takes place between material re-gions that have initially been compressed infront of the legs of the mandrel and have thenflowed with extreme shearing along the flanksof the leg and combined at the dead metal zonesbeneath the legs. The strands that come togetherare free of any oxide film. The process of weld-ing then consists merely of the initially smallcontact surfaces being extended over the entirelength of the section and thus being increasedby several orders of magnitude. Subgrains thenform continuously even over the original contactsurfaces. This repolygonization is linked to themovement of dislocations but does not requireany atom diffusion.

In addition, the material recrystallizes usuallyin the deformation zone (e.g., copper) or im-mediately on leaving the die (e.g., AlMgSi0.5)so that the newly formed grains grow away fromthe contact faces freshly bonded by adhesion andheal them.

The contact faces that equate to a severelydistorted grain boundary disappear during thisrecrystallization and cannot be identified underthe microscope (Fig. 4.55).

The suitability for producing hollow sectionsvaries from material to material. There are evendifferences between aluminum alloys, which ap-pear to be predestined for hollow section pro-duction. Numerous precipitates or phase segre-gations with foreign lattice structures can impairthe bonding. The hot strength of the high-strength materials can be so high that the large

degree of deformation required to produce suf-ficient pressure in the bonding chamber is notreached and the use of hollow section dies madedifficult or even impossible. Finally, with ma-terials that do not recrystallize at the extrusiontemperature, there is no possibility of healing thevoids in the bonds by recrystallization [Ake92a].

There are, however, cases where defects in thebonds occur even with materials well suited forhollow section extrusion. These problems arediscussed in more detail in Chapter 5.

With copper and brass, which are extruded inthe temperature range 600 to 900�C, the dangerof separating oxides flowing in is greater thanwith aluminum alloys. This is one of the reasonswhy extrusion of these materials through hollowsection dies is usually avoided (see Chapter 7,the section on tooling for copper alloys).

The ability of aluminum alloys to bond bypressure welding is not only utilized in the pro-duction of hollow sections but also in billet-on-


Fig. 4.56 Extrusion of aluminum “billet-on-billet” [Ake 92]. (a) Extrusion of a billet on the discard in the container. (b) With feederdie. (c) In the ports of a porthole die. (d) Pulling off the discard in a bridge die [Ake 92]

billet extrusion where transverse welds form(Fig. 4.56).

There is—however rarely used—the casewhere the billet is extruded onto the discard ofthe previous billet (Fig. 4.56a). This is success-ful only after careful cleaning of the surfaces ofthe end faces and the removal of air to avoidblisters and material separation in the extrusion.This process is used in aluminum cable sheath-ing (see section 3.6 in Chapter 3).

With aluminum alloys, the discard is usuallysheared off (solid section dies and porthole dies,Fig. 4.56c) or pulled off (bridge dies, Fig. 4.56d)before the next billet is extruded. If a solid sec-tion is extruded through a flat die, then a “feederchamber” is normally used (Fig. 4.56b). The res-idue of the first billet remaining in this feederchamber behaves in the same way as the residuein a hollow section die after removal of the dis-card. It can bond by adhesion with the subse-quent billet. This usually occurs because the dif-ferent flow conditions between the regions closeto the surface and the core of the section resultin the material of the next billet flowing as atongue into the material from the preceding bil-let so that the contact surfaces are extensivelyincreased.

4.12.3 Composite Extrusion

Composite extrusion is used to produce sec-tions from two or more different materials, usu-ally with a constant cross-sectional distributionover the entire length. The process and practicalapplications are given in Chapter 5 in the section“Extrusion of Semifinished products from Me-tallic Composite Materials.” This section coversthe mechanism of the bonding of two materialsduring extrusion.

Numerous metals react so severely at the highextrusion temperatures required by adhesionwith the extrusion tooling that the latter is im-mediately destroyed unless protective measuresare taken (See section 4.12.5).

This attack of the extrusion tooling and oxi-dation of the hot billet can be avoided by clad-ding the billet of the reactive material (Be, Ti,Zr, Hf, V, Nb, Ta) with copper or iron. Evenalloy steel can be clad for specific purposes withmild steel, an aluminum alloy with pure alumi-num, or even aluminum with copper to give justa few examples.

To ensure that a solid bond forms between thesheath and the billet by pressure welding, bothcarefully degreased components must undergosuch severe deformation in the deformation zone(at least 50%) that any brittle surface coatingpresent (oxide) breaks up and pure metals con-tact each other.

The diffusion of atoms across the boundarysurface can simplify the bonding by transposi-tion processes. On the other hand, these diffu-sion processes can also result in the formationof undesired alloy layers at the contacting sur-faces. In the worst cases, a low-melting-pointsecondary phase can form from the atoms of thetwo components. Table 4.1 lists these low-melt-ing-point phases in the corresponding binarysystem base material-cladding. The extrusiontemperature must not exceed these limits [Ric69].

If intermetallic brittle phases form by diffu-sion during the pressure welding between thesheath and the base material (e.g., between zir-conium and iron), this is not always detrimentalbecause the bond zone can be kept extremelythin.


Table 4.1 Liquid phases between the cladding and the billet

Lowest temperature for the formation of liquidphases (�C) with

Element Melting point, �C Transformation point, �C Copper Nickel Iron

Beryllium 1280 . . . 850 1157 1165Titanium 1720 882 885 942 1085Zirconium 1860 862 885 961 934Vanadium 1900 . . . . . . 1203 1468Niobium 2410 . . . 1080 1175 1360Tantalum 3000 . . . 1080 1360 1410

Source: [Ric 69]

It is not only the intermetallic phases thatform during extrusion that can damage the sur-face zone of the extruded base material to suchan extent that different properties exist relativeto the core after removal of the cladding, e.g.,corrosion susceptibility. The formation of a solidsolution between the cladding material and thebase metal can also result in the same problems.The selection of the cladding material and theextrusion parameters must, therefore, be care-fully considered.

Other types of composite materials are pro-duced so that a second or third material in theform of bar is added to the billet. Super conduc-tors or contact wires can be produced by placingsuitable bars or wires into holes in a copper bil-let. This is then extruded.

In certain cases extrusion is carried out withthe knowledge that intermetallic phases willform between two adjacent materials, e.g., toachieve a high wear resistance or super con-ducting properties [Web 82].

There is also composite extrusion betweentwo metals that exhibit extremely different prop-erties and that apparently do not react with eachother at the extrusion temperature. One exampleis the production of a composite of an aluminumsection and a steel strip in which the wear resis-tance of the steel and its high hardness are util-ized where the section is subjected to highstresses [Mie 87].

In fiber composite materials, metallic or non-metallic fibers are embedded in a metallic matrixto improve the mechanical properties. The ad-hesive bonding between the fibers and the basematerial is usually low because the lattices areso different and, in the case of nonmetallic fi-bers, a metallic bond cannot be produced. Stifffibers are broken into small pieces during extru-sion by the elongation of the parent metal—thehigher the extrusion ratio, the smaller the pieces.These pieces are mechanically held so tightly inthe base material without adhesion that they can

withstand tensile forces even during elongationof the extruded section and thus increase thestrength. The softer metallic matrix transfers theforces under load into the fibers, which deter-mine the strength [Kel 73].

The behavior of the fiber-reinforced metallicmaterials at different temperatures can only beexplained to a limited extent by the behavior ofthe base material. At low temperatures, the fibersincrease the strength. At high temperatures, incontrast, the base material can shear at the in-terface with the fibers. This can damage the ma-terial in such a way that it fails before a materialwithout fibers [Scu 92].

4.12.4 Powder Extrusion

The process and the application of powder ex-trusion is described in detail in the section “Ex-trusion of Powder Metals” in Chapter 5 (see also[Sch 86]). This section covers the metallurgy ofthis process.

The particles of loose or cold compacted pow-der have a fill density of 50 to 80%; i.e., de-pending on the production process the round orangular particles contact each other, at best, in-completely. After cold compaction the particleshave to some degree linked together and in someplaces even adhesively bonded so that the sta-bilized billet (green compact) can withstand thehandling before extrusion without breaking. Inthe press the particles are initially to some extentplastically deformed so that they contact virtu-ally over the entire surface. A certain amount ofrelative movement between the particles in thisaccommodation process results in the particlesalso rubbing together. This is believed to pro-duce partial friction welding [She 72]. This pro-cess is either stopped in front of or at the startof the deformation zone by the plastic elongationcorresponding to the extrusion ratio. The ap-proximately round powder particles becomeelongated fibers. The surface area of the particles


is increased to such an extent that brittle surfacecoatings, usually oxides, break up to free thepure metal. A good lattice match, combined withtransposition processes across the boundariesbetween the individual grains, develops a goodadhesive bond by pressure welding [Rob 91].

The extruded section has a fine structure cor-responding to the powder grain size that canscarcely be distinguished under the microscopefrom that of an extruded cast billet. However,the structure of a cast (or preforged) billet is usu-ally coarse grained and exhibits in correspond-ing materials coarse primary precipitates thathave a negative influence on the quality of theextruded section. The fine-grain structure ofpowder of the same composition is, in contrast,far more homogeneous (e.g., [Asl 81]).

With perfect bonding, the mechanical prop-erties of a material produced from either a castbillet or a powder metallurgical material of thesame composition are largely identical.

For dispersion hardening, hard particles of anintermetallic phase or a nonmetal (usually ox-ides, nitrides, or carbides) are embedded in thebase material by powder extrusion. The effect ofthese particles corresponds exactly to that of pre-cipitates described in section 4.7. They act asbarriers to dislocation movement, which in-creases the yield strength and the work harden-ing. The number and distribution of the particlesand not the size are important: the finer the dis-tribution, the greater the effect.

The particles added by powder metallurgy areresistant to very high temperatures in contrast toprecipitates in cast alloys. Coagulation and dis-solution are impossible. Therefore, the applica-tion of dispersion-hardened materials is mainlyattributable to the high-temperature properties.They are still effective at very high temperatureswhere other strength-increasing methods fail(work hardening, dissolved foreign atoms, pre-cipitated particles), e.g., [Mat 90] (Fig. 4.57).

Very finely distributed dispersoids also hindergrain-boundary migration, which leads to re-crystallization. Finely distributed aluminum ox-ide particles in a copper matrix can, for example,completely suppress recrystallization at 1000 �C,which is just below the melting point. Thismeans that cold work-hardened materials alsohave a higher strength at high temperatures thanthe original soft material because of the retentionof grain boundaries that hinder dislocationmovement [Zwi 57].

The boundary surfaces matrix-hard particlesdo not usually have good adhesion, similar to

fiber materials, because the adhesive effect isweak with different metallic lattices and com-pletely absent with metallic inclusions. Withtheir small size, however, they are embedded sosecurely in the base material that this is not im-portant. If the hard particles or their materialfraction are too large, tearing at the boundarysurfaces can result in a decrease in strength. Vol-ume fraction, size, and distribution of the hardparticles have to be carefully balanced.

4.12.5 Friction between the ExtrudedMaterial and the Tooling

The friction conditions between the billet andthe container as well as at the end faces and onthe die working faces play a very important rolein extrusion. This section discusses only thephysical and metallurgical aspects of these pro-cesses. The basics of lubrication are alsotouched on (more information on adhesion isgiven in [Tyl 68]).

In the absence of lubrication, there is directcontact between the tooling and the extrudedmaterial. The roughness of the tool surface isdetermined by the machining. The high pressurein extrusion smears the plastically deformed ma-terial into the depressions and as it glides overthe tool, it has to travel over the peaks and intothe valleys. This process results in friction.

The adhesion between the tool and the materialis more important. It occurs where there is metal-to-metal contact. The partial adhesion and break-ing away of the material at the tool increases thefriction between both. The adhesion can, how-ever, be larger than the cohesion within the ma-terial. Particles of the material torn from its sur-face adhere to the tool. It is also possible for thepieces of the tool surface to be torn away and topass out with the extruded section. This increasesthe roughness of the tool. High pressure and tem-perature increase the adhesive tendency.

This tendency to adhesion varies with thematerial/tool combination. It can be assumedthat a slight adhesion tendency and thus a lowfriction force occurs where both partners onlyslightly dissolve in each other. Aluminum andtitanium adhere very strongly with steel tooling.This, naturally, also applies to steel because theadhesion tendency between partners of the samematerial is particularly marked. On the otherhand with copper the adhesion tendency to steeltooling is only half as high as that of aluminum[Czi 72].

The tooling surface is oxidized at the extru-sion temperature. The adhesion tendency will be


1) AlZnMgCu 1.5 (age hardened at elevated temperature)

2) Aluminum with 10% Al2O3 sintered aluminum powder (SAP)

3) AlMg3.5Cr (solid-solution hardened)

4) Al 99

700

600

500

400

300

200

100

0

Tens

ile s

tren

gth,

MP

a

Temperature, °C

0 20 100 200 300 400 500

1

2

3

4

Dispersion hardened with 4% C

Fig. 4.57 Hot strength of normal and dispersion-hardened aluminum alloys [Jan 75]

reduced by this oxide coating on the hot tool. Itseffect has still not been extensively studied.Probably, poorly bonding oxides are partly tornaway and carried out with the extruded sectionso that the clean tool surface comes into contactwith the material.

Even without taking into account the adhesionfrom the relative movement material/tooling, theplastic deformation needed for smearing at thesurface structure of the hard tool produces heat.The catching on and releasing of the materialcaused by the adhesion can also produce localtemperature peaks that can approach the meltingpoint of the material. They are, however, im-mediately reduced by the good conductivity ofthe metal. Both processes are the cause for thetemperature increase due to friction at the ma-terial surface.

The thin but firmly bonded oxide layer on alu-minum is broken up by a large deformation evenat high temperatures. The clean metal comesinto contact with the tooling during extrusion. Athin layer of aluminum quickly forms on thebearing surface of the die, and this bonds so sol-idly because of the adhesion that the profileformed in the bearings largely slides along alu-minum and not steel [Ake 83]. A smooth bearing

surface thus behaves hardly differently from aslightly rougher one. The roughness measuredperpendicular to the section corresponds to theroughness of the aluminum coating on the diebearing surface [The 93] (see also Chapter 6, thesection on extrusion of aluminum alloys).

“Crow’s feet” or “pickup” are particles thathave been torn out within the die opening andhave bonded to the tool. They produce groovesin the surface. They continue to grow until theshear force exerted by the extrusion is sufficientto tear them away and to carry them out with theextruded section [Mer 77].

Nitriding, which is usually used to harden thesurface of extrusion dies for aluminum alloys,has only a slight effect on the adhesion behaviorbut does prevent premature wear of the tool sur-face. Grooves in the tool thus occur later, if atall. The aluminum coating does form later onnitrided tooling during extrusion than on unni-trided tooling. Once it has formed then, as far asthe section surface is concerned, there is nolonger any difference [The 93].

The container liner surface is coated with sol-idly bonded aluminum, even after the first ex-trusion, because of the strong adhesion tendencybetween steel and aluminum already mentioned.


After this, the billet no longer slides or, morecorrectly, shears, over steel during extrusion butover aluminum [Ake 83].

Only extrusion without lubrication has beenconsidered so far, which is usual for aluminum.Lubricants—including certain oxides—shouldseparate the material and the tooling and hinderany friction reaction between them (see [Sce 83]for a more detailed discussion).

The lubricant film has to bond both to thetooling and the material. The slip takes placewithin the lubricant film itself. The selected lu-bricant film has to be thick enough not to breakdown and must have the correct viscosity(toughness) at the extrusion temperature. The lu-bricant on one hand has to be able to follow thesevere increase in the surface area in extrusionand, on the other hand, must not be squeezed outunder the high extrusion pressure.

Only a few metal oxides act as lubricants andonly copper oxide is so ductile at the extrusiontemperature and bonds to the tooling and billetsolidly enough while retaining the sliding effectthat the container does not have to be lubricatedat all and the die only slightly. However, an in-complete coating of the billet surface with oxideresults in irregularities in the slip of the billetalong the container liner wall and to defects inthe extruded section (see the section on copperalloys in Chapter 5 for more information).

If nonlubricating oxides in copper alloysdominate, e.g., zinc oxide in CuZn alloys(brass), the lubricating effect of the copper oxideis lost.

The usual lubricants with copper alloys andother materials with moderate extrusion tem-peratures are graphite containing lubricatingoils, which can be classified under the heading“solid lubrication.” The hexagonal structure ofgraphite, the base surfaces of which exert onlya weak attraction force between each other, isthe reason that graphite plates easily slide overone another. The oil in which the graphite platesare embedded acts as a carrier and ensures thatthe lubricant bonds to the tooling with a suitablethickness.

If the extrusion temperature is higher than ap-proximately 1000 �C, graphite is unsuitable as alubricant because it is destroyed by oxidation, asis the carrier oil. Molybdenum disulfide, whichexhibits at room temperature and slightly ele-vated temperatures outstanding lubricationproperties—attributable like graphite to its crys-tal structure—loses its effect by oxidation above300 �C. It is therefore not suitable for extrusionat higher temperatures.

Glass lubrication in the Sejournet process isthe current practice for high-temperature extru-sion (stainless steels, nickel alloys, etc.). This isdescribed in detail in the section “Extrusion ofIron-Alloy Semifinished Products” in Chapter 5.It should also be mentioned that the thermal in-sulation between the material and the toolingplays an important role in addition to the goodlubrication properties.

The billet and tooling have to be completelywetted by the glass lubricant. Good adhesion isalso a requirement and it should also not attackthe metal. The glass film on the extruded sectionshould be approximately 10 to 30 lm thick.

The relative movement between the materialand the tooling occurs entirely within the glassfilm, the viscosity of which corresponds to a firstapproximation to a Newtonian liquid:

Viscosity g � A • exp(E/RT )

The glass film becomes more fluid as the tem-perature increases. Pressure dependence of theglass viscosity has not been found.

Different glass compositions are recom-mended to give the correct viscosity, whichshould be between 10 and 100 Pa • s, at thedifferent extrusion temperatures and extrusionratios. The table in Fig 4.58 gives the compo-sition of different glasses and shows some vis-cosity curves.

4.13 Extrudability ofMetallic Materials

Good extrudability is a collective term for alow-extrusion load, high extrusion speed, goodsurface finish, freedom from defects, tight di-mensional tolerances, and high tooling through-put. The breadth of the term makes definitionsand deductions difficult. In general, a material isdescribed as having a better extrudability thananother if it has better hot-working characteris-tics in extrusion when specific extrusion param-eters are held constant. Extrusion materials areclassified by this term into high, moderate, andlow extrudability.

One material from an alloy family can thenhave a higher extrudability than another if:

● The maximum extrusion speed before theoccurrence of surface defects (roughness,grooves) under otherwise equal conditions(initial billet temperature, container tempera-ture, extrusion ratio, die) is faster.


No. Type of glass Approximate composition Recommended temperature range, �C

1 Lead-borate 10 B2O3, 82 PbO, 5 SiO2 3 Al2O3 5302 Borate . . . 8703 Potassium-lead-silicate 35 SiO2, 7 K2O, 58 PbO 870–10904 Potassium-sodium-lead-silicate 63 SiO2, 8 Na2O, 6 K2O, 4 MgO, 21 PbO 1090–14305 Boron-silicate 70 SiO2, 1 PbO, 28 B2O3, 1 Al2O3 1260–17306 Boron-silicate 81 SiO2, 4 Na2O, 13 B2O, 2 Al2O3 1540–21007 Silicate (quartz) 96 � SiO2 2210

Fig. 4.58 Dependence of the viscosity of glass lubricants on the temperature [Sce 83]

● The extrusion load under otherwise equalconditions (initial billet temperature, con-tainer temperature, extrusion ratio, die, ex-trusion speed) is lower.

● The initial billet temperature and the con-tainer temperature under otherwise equalconditions (extrusion ratio, die, and extru-sion speed) are lower.

The material and its properties influence theextrudability in many ways:

● The melting point is a controlling factor forthe deformation temperature and thus for thetooling life and for the subsequent process-ing of the extruded product by calibration orfinish drawing.

● Alloying additions (type and content) deter-mine the flow stress, extrusion speed, and thetendency to extrusion defects.

● Large precipitates and inclusions can resultin the formation of surface defects (furrows,

flaking, and cracks) and increase the toolwear.

From the practical point of view it is impor-tant to be able to use simple criteria before ac-cepting an order and before using an unknownmaterial to estimate whether and at what cost anorder can be realized. The size and the type ofpress available, the tooling, and the dimensionsof the extruded product play a role in additionto the material.

The theory of the extrusion load is given inChapter 3. Reference should be made to section3.1.1.4 and the calculation example in Fig. 3.18.

The values of kf or kw and gu cannot be ob-tained from the data obtained from normal ten-sile tests since the strain that can be attained istoo low because of the necking and because therate of deformation in extrusion is several ordersof magnitude higher than the experimental con-ditions on a tensile testing machine.


Fig. 4.59 Experimental die to determine extrudability

Alloy type

Ste

m s

peed

, mm

/s

1000

100

10

10

0

AlZ

nMg1

Al9

9.5

AlC

uMg1

AlZ

nMgC

u1.5

AlC

uMg2

AlM

gSi1

AlM

gSi0

.5

Fig. 4.61 Limiting extrusion speed measured on differentaluminum alloys using the test section in Fig. 4.60

[Cas 78]

Fig. 4.60 Test section to estimate the extrudability of alu-minum alloys using the extrusion speed as the lim-

iting factor [Cas 78]

The logarithmic principle strain rate of 1ugto 10 s�1 reached in torsion tests falls in therange of the average logarithmic principle strainrate in extrusion. In compression testing on fastpresses, ug up to approximately 2 and up toug1000 s�1 are achieved.

The Atlas of Hot-Working Properties of Non-Ferrous Metals (Vol 1, Aluminum Alloys) [DGM78] gathers the known data from the literaturein a standard format (ordinate: flow stress kf, ab-scissa: logarithmic principle strain ug with theparameters temperature and principle strainrate). For the extrusion press specialist it is dif-ficult to find the necessary data for the extrusionratios of V � 10 to 300, corresponding to prin-ciple strains ug between 2 and 6 and the loga-rithmic strain rates between 0.5 and 10 s�1ugnormally used in extrusion.

What helps in practice?One suggestion made after trials with alumi-

num alloys was to define the extrusion behaviorby using a standard two-hole die [Mon 75] (Fig.4.59) as follows: The farther the thicker sectiontravels during extrusion, the higher is the ex-trudability of the material. Tests with this pro-cess found it to be too insensitive. Seven-holedies of a specific design have also been used forexperiments [Mis 91].

Scharf estimated the extrudability of alumi-num alloys in experimental extrusions from theextrusion time at a constant extrusion load. Hetherefore used indirectly the extrusion speed asthe limiting parameter and could thus differen-tiate between alloys and their heat treatment[Sca 64].

Castle and Lang [Cas 78] developed a test dieand test section with which it was possible tomake a meaningful assessment of the materialproperties of continuously cast material with andwithout heat treatment (Fig. 4.60). The limitingspeed for poor shape or tearing away of the ribswas determined. The necessary extrusion loadcan also be used as a measure of the extrudabil-ity of a material.

Many extrusion plants use similar test dies—even dies for hollow sections with correspond-ing ribs.

Figure 4.61 shows the various limiting speedsfor a range of alloys.

The extrusion speed does not play such a largerole with materials that are extruded at moderateand high temperatures (e.g., copper alloys,steels); it is usually very high. In this case the


flow resistance and thus the limiting extrusionload have to be taken into account.

Hot-tearing tests can give a rough comparisonbetween the different materials. The dependenceof the strength of different materials on the tem-perature is in principle correctly reproduced.The kf values obtained, however, cannot be usedto calculate the extrusion load.

The following method is known from the cop-per alloy processing industry: Equations 3.4,3.5, and 3.49 given in Chapter 3 in for the vari-ation of the extrusion load during extrusion withand without lubrication are used.

For simplicity, the relevant formulae for themaximum extrusion load F0 are:

Unlubricated container FSt max

� A0 • kw • ug � D0 • p • (l0 � lR)s(Eq 4.1)

and

Lubricated container FSt max4(l /D ) l 4(l /D ) l• •0 0 c 0 0 c� A {k • u • e � k (1 � e )}0 w g f0

(Eq 4.2)

With glass lubrication the friction factor l0 isapproximately zero. Then:

F � A • k • uSt max 0 w g

where

¯k � k /gw f

and

u � ln A /Ag 0 l

where A0 is the container cross-sectional area;As is the section cross-sectional area; D0 is thecontainer bore; l0 is the upset billet length in thecontainer; is the mean flow stress of the ma-kfterial being deformed in the deformation zone;gu is deformation efficiency; is flow stress inkf0

the billet outside the deformation zone; and, l0is the coefficient of friction for the friction be-tween the billet and the container.

Successive trials are carried out on the extru-sion press in question with a standardized billetsize (diameter, length) at the assumed tempera-ture. Using initially large and then increasinglysmaller extruded bar diameters, extrusions arecarried out up to the maximum press capacity.

A value of kw can then be calculated from Eq4.1 (unlubricated container) or Eq 4.2 (lubri-cated container) using the coefficients g and l0obtained from the practical measurements. It isthen possible using Eq 4.1 or 4.2 to deduce pos-sible billet or section dimensions for the materialstudied on this press [Ret 99].

The very specific test process described to es-timate the extrudability reflects the extent of theproblem and the complexity of this term. At thesame time such tests have become useful aids todetermine optimal deformation conditions.Many extrusion plants also use their own testmethods.

Other criteria for the extrudability, includingfreedom from defects, dimensional tolerances,and tooling life do not lend themselves to modelinvestigations and practical experience has to beused.

Appendix: AdditionalInformation for Chapter 4Metallurgical Principles

Additional Information 1: Miller’sIndices to Describe a Space Lattice

Various planes can be drawn through theatomic points of a crystallite, referred to as crys-tallographic planes. Because there are numerousparallel crystallograhic planes, these are referredto as a family of crystallographic planes. Theorientation of the crystallographic planes is de-scribed by a coordinate system, the axes x, y,and z coincide with the edges of the unit cell(edge length a, b, c) (Fig. 4.62).

The lengths ma, nb, and pc from the origin tothe intersection points of the crystallographicplanes define their location. The reciprocal val-ues of m, n, p, are taken to give the Miller in-dices as follows:

h:k:l � 1/m:1/n:1/p

The Miller indices of closely packed planesconsist of small whole numbers (in the diagram:m � 2, n � 4, p � 1 gives h:k:l � 2:1:4).

The indices are given in ( ): (hkl) � (214).Negative indices are described with ¯ (overbar)


+x

(100)

[100]

[010]

[111

]

[101]

[110][100]

(110) (111)

+y

+z

+x

+x

+z

+y

+z

+x+y

+y

+z

+x

+y

+z

2

4

(a)

(b)

(c)

1

1

2 3

c

b

a

Fig. 4.62 Miller’s indices for space lattices. (a) Designation of lattice planes. (b) Examples of planes in the cubic lattice. (c) Examplesof directions in the cubic lattice

0.5

0.50.5

0.5

0.5

0.5

0.5

0.5

1

1

1

1

1

1

4

4

4

4

4

4

24 16

24 16

8

8

8 (011) [111]

(012) [100]

P

PR

TR

[111]

8

Fig. 4.63 Representation of textures in pole figures. [111] pole figure from the section core of an extruded AlCuMg bar [Moe 75]


over the number: e.g., �1 � . Figure 4.62(b)1shows some examples of crystallographic planesin a cubic lattice.

Directions in the unit cell are described withthe location coordinates of a point on a line go-ing through the origin—also whole numbers butthis time within angular brackets. A surface di-agonal in the cubic lattice is designated with[011], and a spatial diagonal with [111] (Fig.4.62c).

Additional Information 2:Representation ofTextures in Pole Figures

In general, the orientation distribution in atextured structure is shown on a stereographicprojection in which a spherical surface is rep-resented on a plane (known as the projectionplane).

The specimen is located in the center of thesphere, and the normals of the crystallographicplanes penetrate the surface of the sphere. A tex-ture is described by showing in the pole figurethe frequency of the orientations of individualcrystallographic planes (e.g., (111) or (100)).The pole plane coincides with a specific projec-tion plane (e.g., the rolling plane in rolledsheets). In extruded or drawn bar the north-southdirection of the pole figure is usually orientatedalong the bar axis (Fig. 4.63).

Additional Information 3:Temperature and SpeedDependence of the Flow Stress inHot Working [Sel 79, Ake 70, McQ 90]

In the stationary region of the work-hardeningcurve where work hardening and recovery arein equilibrium, the flow stress is dependent onan exponential function of the temperature asdeduced from diffusion. The activation energyis that for self diffusion:

k � f (Q/RT)f

At the same flow stress there is a simple rela-tionship between the strain rate and the absolutetemperature T:

u � A • exp(�Q/RT)

If deformation is carried out at a higher tem-perature and higher speed then, according to thisrelationship, the flow stress is the same as thatat a lower temperature and corresponding lowspeed:

˙k � f (uexp(Q/RT))f

Or, using the Zener-Hollomon parameter Z:

k � f (Z)f

˙Z � uexp(Q/RT)

There are various expressions for the relation-ship between the flow stress kf and the Zener-Hollomon parameter Z that apply depending onthe material and temperature-speed situation:

aZ � a • k Z � a • exp(b • k )f f

REFERENCES

Books and Overviews[Alt 94]: D. Altenpohl, Aluminium von innen

(Aluminum from the Inside), Aluminium-Ver-lag, Dusseldorf, 1994

[Ber 83]: W. Bergmann, Werkstofftechnik, Teil1: Grundlagen, Teil 2: Anwendung (MaterialTechnology, Part 1: Fundamentals, Part 2:Application), Carl Hanser Verlag, Munich,Vienna, 1983

[Blu 93]: W. Blum, H. Mughrabi, and B. Rep-pich, Grundlagen der Werkstoffkunde—Ni-chteisenmetalle in Umformtechnik (Funda-mentals of Material Science), VerlagStahleisen, Dusseldorf, 1993

[Dah 93]: W. Dahl, R. Kopp, and O. Pawelski,Umformtechnik—Plastomechanik und Werk-stoffkunde (Non-Ferrous Metals in Deforma-tion Technology), Verlag Stahleisen, Dussel-dorf, 1993

[Guz 70]: A.G. Guz and G. Petzow, Metallk-unde fur Ingenieure (Metallurgy for Engi-neers), Akad. Verlagsges, Frankfurt, 1970

[Hor 67]: E. Hornbogen and H. Warlimont,Me-tallkunde (Metallurgy), Springer-Verlag, 1967

[Hou 93]: H.P. Hougardy, Grundlagen derWerkstoffkunde (Fundamentals of Metal-lurgy), Stahl in Umformtechnik, Verlag Stah-leisen, Dusseldorf, 1993

[Sce 83]: J.A. Schey, Tribology in Metalwork-ing, ASM International, 1983

[Sch 81]: W. Schatt, Einfuhrung in die Werk-stoffwissenschaft (Introduction to Material


Science), VEB Deutscher Verlag fur die Grun-dindustrie, Leipzig, 4, Auflage, 1981

[Tyl 68]: R.F. Tylecote, The Solid Phase Weld-ing of Metals, Edward Arnold Ltd., London,U.K., 1968

[Wie 86]: A.G. Wielandwerke, Ulm: Wieland-Buch Kupferwerkstoffe (Wieland-Book Cop-per Alloys), Wieland-Werke AG, Ulm, 5, Au-flage, 1986

Special Literature

[Ake 66]: R. Akeret and A. Kunzli, Ermittlungder Formanderungsfestigkeitkf von Alumini-umlegierungen mit Hilfe von Torsionsver-suchen und Vergleich der Er-gebnisse mitdenjenigen von Strangpreßversuchen (Deter-mination of the Flow Stress of Aluminum Al-loys Using Torsion Tests and Comparison ofthe Results with Those from Conclusion Tri-als), Z. Metallkde., Vol 57, 1966, p 7789–7792

[Ake 70]: R. Akeret, “Fließspannung und Struk-tur bei der Warmumformung (von Alumin-ium)” (“Flow Stress and Structure in HotWorking [of Aluminum]”), Alusuisse report220/70

[Ake 83]: R. Akeret, Einfluß der Querschnitt-form und der Werkzeuggestaltung beimStrangpressen von Aluminium, Teil II: DieReibung im Preßkanal (Influence of theCross-Section and Tool Design on the Extru-sion of Aluminum, Part II: Friction in the DieBearing), Aluminium, Vol 59, 1983, p 745–750

[Ake 88]: R. Akeret, A. Ried, and J. Theler, J.Metall, Vol 42, 1988, p 760–769

[Ake 90]: R. Akeret and M. Textor, Strangpres-sen (Extrusion), DGM-Informationsgesells-chaft, Oberursel, Germany, 1990

[Ake 92]: R. Akeret, Strangpreßnahte in Alu-miniumprofilen (Extrusion Welds in Alumi-num Sections), Parts 1 and 2, Aluminium, Vol68, 1992, p 877–886 and 965–973

[Ake 92a]: R. Akeret, “Extrusion Welds—Qual-ity Aspects Are Now Center Stage,” presentedat International Aluminum Technology Sem-inar (Chicago, IL), May 20–22, 1992

[Alt 65]: D. Altenpohl, Aluminium und Alumi-niumlegierungen (Aluminum and AluminumAlloys), Springer-Verlag, 1965

[And 48]:W.A. Anderson and R.F. Mehl, Trans.AIME, Vol 161, 1948, p 140

[Ash 79]: M.F. Ashby, C. Gandhi, and D.M.R.Taplin, Acta Metall., Vol 27, 1979, p 699–729

[Asl 81]: C. Aslund, G. Gemmel, and T. An-dersson, Stranggepreßte Rohre auf pulver-metallurgischer Basis (Extruded Tubes with aPowder Metallurgy Base), Bander BlecheRohre, Vol 9, 1981, p 223–226

[Bag 81]: J. Baumgarten, W. Bunk, and K.Lucke, Strangpreßtexturen bei AlMgSi-Le-gierungen, Teil 1: Rundstangen (ExtrudedTexture in AlMgSi Alloys, Part 1: RoundPoles), Z. Metallkde., Vol 72, 1981, p 75–81

[Bau 59]:M. Bauser, “Einfuhrung in die Physikder plastischen Verformung der Metalle” (“In-troduction to the Physics of the Plastic De-formation of Metals”), unpublished manu-script

[Bau 61]: M. Bauser, Verhalten von Nichteisen-Metallen bei der Warmumformung (Behaviorof Nonferrous Metals in Hot-Working), Vor-trag Metallfachabend, Dusseldorf, 1961

[Bau 63]: M. Bauser, Verformbarkeit undBruchverhalten bei der Warmformgebung(Workability and Fracture Behavior in Hot-Working), Metall., Vol 17, 1963, p 420–429

[Bau 64]: M. Bauser, Zeit-temperatur-Auschar-tungsschaubilder und die Festigkeit strang-gepreßter aushartbarer Aluminiumlegierun-gen (Time Temperature Age-HardeningDiagrams and the Mechanical Properties ofExtruded Age-Hardening Aluminum Alloys),Z. Metallkde,Vol 55, 1964, p 816–821

[Bux 77]: K. Buxmann, Metall., Vol 31, 1977,p 163–170

[Bux 77a]: K. Buxmann, Fachbericht (technicalreports) FB 41/777/100 der Deutschen Ge-sellschaft fur Materialkunde, Oberursel, Ger-many

[Cas 78]: A.F. Castle and G. Lang, ProceedingsAluminium Konferencia 78, Szebesfehervar,Vol 1, 1978

[Czi 72]: H. Czichos, J. Appl. Phys., Vol 5,1972, p 1980

[DGM 78]: Deutsche Gesellschaft fur Metallk-unde: Atlas der Warmformgebungseigen-schaften von Nichteisenmetallen, Alumini-umwerkstoffe (Atlas of Hot-WorkingProperties of Non-Ferrous Metals), Vol 1,Aluminum Alloys, DGM, 1978

[DGM 78a]:Deutsche Gesellschaft fur Metallk-unde: Atlas der Warmformgebungseigen-schaften von Nichteisenmetallen, Band 2,Kupferwerkstoffe (Atlas of Hot-Working Prop-erties of Non-Ferrous Metals), Vol 2, CopperAlloys, DGM, 1978

[Die 77]: O. Diegritz, “Preßfehlerkatalog; Feh-lererscheinungen bein direkten Strangpressen


von Kupferwerkstoffen” (“Defects in the Di-rect Extrusion of Copper Alloys”), manu-script, AK Schwermetall im Strang-preßausschuß, DGM

[Fie 57]: D.S. Fields and W.A. Backofen, De-termination of Strain Hardening Characteris-tics by Torsion Testing, Proc. Amer. Soc. Test.Mat., Vol 57, 1957, p 1259–1272

[Fle 74]:M.C. Flemings, Solidification Process-ing, McGraw-Hill Inc., 1974

[Fra 50]: F.C. Frank and W.T. Read, Phys. Rev.,Vol 79, 1950, p 722–723

[Got 84]: G. Gottstein, Rekrlstallisation metal-lischer Werkstoffe (Recrystallization of Me-tallic Materials), Deutsche Ges., F. Metallk-unde, e.V., 1984

[Grz 93]: B. Grzemba and J. Hasenclever, Non-Ferrous Metals Volume Using Aluminum andCopper as Examples, Umformtechnik, Koppund Pawelski, Ed., Verlag Stahleisen, Dussel-dorf, 1993

[Haa 84]: P. Haasen, PhysikalischeMetallkunde(Physical Metallurgy), Springer-Verlag, Ber-lin, 1984

[Hof 62]: W. Hofmann, Lead and Lead Alloys,Metallurgy and Technology, Springer-Verlag,2, Auflage, 1962

[Jan 75]:G. Jangg, F. Kutner, and G. Korb, Her-stellung und Eigenschaften von dispersions-gehartetem Aluminium (Production and Prop-erties of Dispersion Hardened Aluminum),Aluminium, Vol 51, 1975, p 641–645

[Kat 67]: T.Z. Kattamis, J.C. Coughlin, andM.C. Flemings, Trans. AIME, Vol 239, 1967,p 1504–1511

[Kel 73]: A. Kell, Werksstoffe hoher Festigkeit(High Strength Material), Vol 4, VerlagVieweg, Wiesbaden, Germany, 1973

[Lan 89]: G. Lang, Warmformverhalten von Al-uminiumlegierungen beim direkten Strang-pressen (Hot-Working Characteristics of Alu-minum Alloys in Direct Extrusion),Aluminium, Vol 65, 1989, p 291–294

[Los 77]: E. Lossack, Erzmetall, Vol 30, 1977,p 243

[Los 83]: E. Lossack and W. Schneider, Gal-vanotechnik, Jahrg, Vol 74, 1983, p 787–794

[Mat 90]: K.H. Matucha, K. Drefahl, G. Korb,and M. Ruhle, Neue oxiddispersionsge-har-tete Legierungen fur Hochtemperatur-An-wendungen (New Oxide Dispersion Hard-ened Alloy for High TemperatureApplication), VDI-Berichte (VDI-Reported),1990, Vol 797, p 125–152

[McQ 75]: H.J. McQueen and J.J. Jonas, PlasticDeformation of Materials. Treatise on Mate-

rial Science and Technology, Vol 6, Akad.Press, New York, 1975, p 393–493

[McQ 90]:H.J. McQueen, Micromechanisms ofDynamic Softening in Aluminum Alloys dur-ing Hot Working, Hot Deformation of Alu-minum Alloys Proc. (Detroit, MI), 1990

[McQ 90a]: H.J. McQueen, Effect of Solutesand Precipitates on Hot Working Behavior ofAl Alloys, Hot Deformation of Aluminum Al-loys Proc. (Detroit, MI), 1990, p 105–120

[Mer 77]: G. Merk and S.E. Naess, Pick-UpFormation on Aluminium Extrusion, Z. Me-tallkd., Vol 68, 1977, p 683–687

[Mie 87]: G. Mier, Verbundstromschienen Al-uminium-Stahl fur S-und U-Bahnen (Com-posite Bus-Bars in Aluminum Steel for S andU Trains), Schweizerische Aluminium Rund-schau (Swiss Aluminum Review), 1987,p 12–17

[Mis 91]: W. Misiolek and J. Zasadzinski, Test-Die for Estimation of the Extrudability ofMetals, Aluminium,Vol 67, 1991, p 85–89

[Mol 75]: M. Moller, “Werkstoffluß, Grobkorn-bildung und Preßeffekt bei stranggepreßtemAluminium (AlCuMg)” (“Course Grain For-mation and Extrusion Effect in Extruded Alu-minum”), dissertation, Technical Universityof Clausthal, Germany, 1975

[Mon 75]: L.F. Mondolfo, A.R. Peel, and J.A.Marcantonio,Met. Technol., Sept 1975, p 433

[Mor 69]: G. Moritz, Z. Metallkde., Vol 60,1969, p 742–749

[Mug 84]: H. Mughrabi, Mechanisms and Me-chanics of Plasticity, Rev. Phys. Appl., Vol 23,1984, p 367–379

[Ohm 88]: L. Ohm and S. Engler, Aluminium,Vol 64, 1988, p 513–520

[Ohm 89]: L. Ohm and S. Engler, Metall., Vol43, 1989, p 539–543

[Ray 49]: G.V. Raynor, Ann. Equilibr. Dia-grams, No. 2, 1949

[Rei 80]: W. Reif and W. Schneider,Gießereiforschung, Vol 32, 1980, p 53–60

[Res 92]: O. Reiso, Ph.D. thesis, Norwegian In-stitute of Technology, Trondheim, Norway,1992

[Ret 99]: W. Rethmann, Pers Mitteilung KM-Kabelmetall AG (communication from KM-Kabelmetall AG), Osnabruck, Germany, 1999

[Ric 69]: H. Richter, Die Herstellung von Stan-gen, Rohren und Profilen durch Verbund-strangpressen (The Production of Bars, Tubesand Sections by Composite Extrusion), Z.Me-tallkde., Vol 60, 1969, p 619–622

[Rie 87]: H. Riedel, Fracture at High Tempera-tures, Springer-Verlag, Berlin, 1987


[Rob 91]: Roberts and Ferguson, Strangpressenvon Pulvermetallen (Extrusion of PowderMetals), Int. Mater. Rev., Vol 36, 1991,p 62–79

[San 75]: R. Sandstrom and R. Lagneborg, ActaMetall., Vol 23, 1975, p 387

[Sca 64]:G. Scharf, Einfluß der chemischen Zu-sammensetzung auf Preßbarkeit und Festig-keit von AlMgSi0.5 (Influence of the Chem-ical Composition on the Extrudability andStrength of AlMgSi0.5), Z. Metallkde., Vol55, 1964, p 740–744

[Sch 86]:W. Schatt, Pulvermetallurgie (PowderMetallurgy), Huthig-erlag, Heidelberg, Ger-many, 1986

[Scn 85]: W. Schneider, Aluminium, Vol 61,1985, p 338–345

[Scu 92]: K. Schulte, Metal-Matrix Compositeswith Aluminium Matrix: Advanced AerospaceMaterials, Springer-Verlag, 1992, p 108–118

[See 85]: A. Seeger, Kristallplastizitat. Hand-buch der Physik (Handbook of Physics),Springer-Verlag, Berlin, VII/2, p 1–210

[Sel 79]: C.M. Sellars, The Physical Metallurgyof Hot Working, Proc. Int. Conf. Hot Workingand Forming Processes (Sheffield, U.K.),1979, p 3–15

[She 72]: T. Sheppard and P.J.M. Chare, PowderMetall., Vol 15, 1972, p 17–41

[Stu 69]: H.-P. Stuwe, Einfluß einer plastischenVerformung auf Phasenumwandlungen inMetallen (Influence of a Plastic Deformationon the Phase Transformations in Metals),Hart.-Tech. Mitt., Vol 24, 1969, p 1–5

[The 93]:W.W. Thedja, K. Muller, and D. Rup-pin, Die Vorgange im Preßkanal beim Warms-trangpressen von Aluminium: Reibung imPreßkanal und Matrizenverschleiß (The Pro-cesses in the Die Aperture in the Hot-Extru-sion of Aluminum: Friction in the Die Aper-ture and Die Wear), Aluminium, Vol 69, 1993,p 524–528 and 649–653

[Web 82]: H.R. Weber, Extrusion of TechnicalSuperconductors, Extrusion symposium, DGM,1982, p 277–296

[Wei 94]: H. Weiland, T.N. Rouns, and J. Liu,The Role of Particle Stimulated Nucleationduring Recrystallization of an Aluminum-Manganese Alloy, Z. Metallkde., Vol 85,1994, p 592–597

[Wev 54]: F. Wever and A. Rose, Atlas zur War-mebehandlung der Stahle (Atlas on the HeatTreatment of Steels), Verlag Stahleisen, Dus-seldorf, 1954

[Wie 86]: A.G. Wieland werke, Wieland-BookCopper Alloys, Ulm, 5, Auflage, 1986.

[Zwi 57]: K.M. Zwilsky and N.J. Grant, J. Met.,Vol 9, 1957, p 1197–1201

Documents

Metallurgical Principles - Materials Data Repository Home