SOL~ICATlOr~
PROCESSING
TN
630
,f-53
McGRAW-HILL SERIES IN
MATERIALS SCIENCE AND ENGINEERING
Editorial Board
MICHAEL B. BEVER
M. E. SHANK
CHARLES A. WERT
ROBERT
F.
MEHL, Honorary
Senior
Advisory
Editor
A VITZUR: Metal Forming: Processes and Analysis
AZAROFF: Introduction
to Solids
BARRETTAND MASSALSKI:Structure of Metals
BLATT: Physics of Electronic Conduction in Solids
BRICK, GORDON, AND PHILLIPS: Structure and Properties of Alloys
BUERGER: Contemporary
Crystallography
BUERGER: Introduction
to Crystal Geometry
DE HOFF AND RHINES: Quantitative Microscopy
DRAUGLIS, GRETZ, AND JAFFEE: Molecular Processes on Solid Surfaces
ELLIOTT: Constitution
of Binary Alloys, First Supplement
FLEMINGS: Solidification Processing
GILMAN: Micromechanics
of Flow in Solids
GORDON: Principles of Phase Diagrams in Materials Systems
GUY: Introduction
to Materials Science
HIRTH AND LOTHE: Theory of Dislocations
KANNINEN, ADLER, ROSENFIELD,AND JAFFEE: Inelastic Behavior of Solids
MILLS, ASCHER, AND JAFFEE: Critical Phenomena in Alloys, Magnets, and Super-conductors
MURR: Electron Optical Applications in Materials Science
PAUL AND WARSCHAUER:Solids under Pressure
ROSENFIELD,HAHN, BEMENT,AND JAFFEE: Dislocation Dynamics
ROSENQVIST:Principles of Extractive Metallurgy
RUDMAN, STRINGER, AND JAFFEE: Phase Stability in Metals and Alloys
SHEWMON: Diffusion in Solids
SHEWMON: Transformations
in Metals
SHUNK: Constitution of Binary Alloys, Second Supplement
WERT AND THOMSON: Physics of Solids
McGRAW-HILL
BOOK COMPANY
New York
St. Louis
San Francisco
Dusseldorf
Johannesburg
Kuala Lumpur
London
Mexico
Montreal
New Delhi
Panama
Rio de Janeiro
Singapore
Sydney
Toronto
MERTON C. FLEMINGS
Abex Professor of Metallurgy
Massachusetts Institute of Technology
Solidification
Processing
This book was set in Times Roman.
The editors were B. J. Clark and Michael Gardner;
the production supervisor was Joan M. Oppenheimer.
The drawings were done by John Cordes, J & R Technical Services, Inc.
The printer and binder was The Maple Press Company.
Library of Congress Cataloging in Publication Data
Flemings, Merton C
1929Solidification processing.
(McGraw-Hill series in materials science and
engineering)
Includes bibliographical references.
I. Title.
1. Solidification.
2. Alloys.
73-4261
TN690.F59
669'.9
ISBN 0-07-021283-x
SOLIDIFICA
PROCESSING
nON
Copyright © 1974 by McGraw-HilI, Inc. All rights reserved.
Printed in the United States of America. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,.
without the prior written permission of the publisher.
567 89-MAMM-76
54321
CONTENTS
..
~~ T ~:J.>fJ~~:\"",1
-
,.
''''
~-, ~OTr..
clr.:.H.:J~~1;;;.f
"
V· ..···· ~
Preface
1
(JH)
IX
Heat Flow in Solidification
r;(
2
!:J.J~~
.
l.r
1
5
6
Growth of Single Crystals
Solidification of Castings and Ingots'
Casting Processes Employing Insulating Molds
Casting Processes in which Interface Resistance is Dominant
12
Analytic Solutions for Ingot Casting
Solidification of Alloys
Problems in Multidimensional
Heat Flow
17
21
24
Plane Front Solidification of Single-phase Alloys
31
Introduction
31
Equilibrium Solidification
No Solid Diffusion
33
Limited Liquid Diffusion, No Convection
Effect of Convection
Czochralski
Growth (Crystal Pulling)
34
36
41
44
vi
Cellular
Solidification
Plane Front
Solidification
of Polyphase
Alloys
Solidification
of Castings and
Ingots
CONTENTS
Zone Melting
115
107
167
172
134
154
146
104
127
120
114
135
112
105
93
66
87
141
157
117
160
94
77
58
75
73
85
83
64
46
49
53
51
Nucleation
Kinetics
Fluid
Flow andof Interface
Thermodynamics
Solidification
Polyphase
of Solidification
Alloys: Castings and Ingots
CONTENTS
290
246
244
234
214
252
279
275
286
284
239
229
264
263
274
295
200
215
267
208
224
219
273
272
177
188
183
207
203
193
191
187
180
vii
viii
Tabulation
ofand
Error
Functions
Tables
of Approximate
Thermal Data
Processing
Properties
CONTENTS
Growth
349
335
328
344
347
309
308
341
338
331
305
301
319
318
312
356
359
357
PREFACE
This book has grown largely out of a lecture course given to senior-level and graduate
students at Massachusetts Institute of Technology. It is intended for use in courses
of this type, and also for the practicing engineer and research worker. The essential
aim of the book is to treat the fundamentals of solidification processing and to relate
these fundamentals to practice. Processes considered include crystal growing, shape
casting, ingot casting, growth of composites, and splat-cooling.
The book builds on the fundamentals of heat flow, mass transport, and interface kinetics. Starting from these fundamentals, the basic similarities of the widely
different solidification processes become evident. Problems at the end of each chapter
relate principles to practice, illustrating important differences, as well as similarities
between processes. Two years of college-level mathematics provides ample background for solving the problems given and for adequate comprehension of the text.
In addition, it is desirable, though not necessary, that the student have a previous
course in structure of materials. Emphasis of the book is on metallic alloys, but other
materials are also considered.
An essential element of all solidification processes is heat flow. This subject is
treated in the first chapter, primarily to lend cohesiveness to the material to follow.
It provides an excellent basis for description and comparison of solidification processes,
and it can be treated with rather simple assumptions regarding the solidification mechanism. Chapter 2 deals with mass transport ("solute redistribution") in single-crystal
growth. A quantitative description of transport in this type of solidification is"greatly
simplified by the fact that the liquid-solid interface is single phase and planar.
Equations derived in this chapter are also useful in describing dendritic solidification,
except that they must be applied to tiny regions on the order of the dendrite arm
spacmg.
Chapter 3 deals with the important question of how to maintain a plane front
in crystal growth, and of how solute redistribution occurs when the plane front breaks
down to form "cells." Plane-front solidification is considered again in Chap. 4, this
time for polyphase alloys, such as eutectics and off-eutectic "composites" solidified
with an essentially planar liquid-solid interface. This chapter is the first to utilize the
X
PREFACE
concept that the equilibrium melting point of a solid depends on its radius of curvature.
Solidification as it occurs in usual castings and ingots is considered in Chaps.
5 and 6. More specifically, these chapters consider the microscopic aspects of such
solidification, including dendritic growth, micro segregation, inclusion formation, and
gas-pore formation.
They draw heavily on the heat- and mass-transport
concepts
presented in earlier chapters.
Fluid flow plays a larger role in solidification processes than is generally
recognized. Flow is caused by introducing the metal to a mold, by density differences
due to thermal or solute effects, or by solidification contractions.
Fluid flow, treated
in Chap. 7, has important effects on structure and segregation in solidification
processes; many of these have been only recently recognized. An important part of
this chapter deals with interdendritic fluid flow and its relation to porosity and
segregation in castings and ingots.
The major portions of the first seven chapters, and all quantitative treatments
in these chapters, assume equilibrium at the liquid-solid interface. That is, they
assume that the kinetic driving force necessary to advance a solidifying interface, is
negligibly small. This assumption is not valid when facets form, but it appears to be
an excellent approximation for the many alloys that solidify without facets. Implications of this assumption are considered in Chap. 8, which deals with the thermodynamics of liquid-solid equilibria. A portion of this chapter also deals with what is
possible (and impossible) at the liquid-solid interface when conditions are such that
equilibrium is not maintained.
Kinetic effects at the liquid-solid interface, including nucleation, are discussed
in Chap. 9. An understanding of growth kinetics, however qualitative, provides a
basis for understanding
the faceted growth morphologies observed in many real
systems, and for understanding
such solidification processes as growth of "ribbon
crystals" by a twin-plane, reentrant growth mechanism. The final chapter deals with
relations between the structure and properties of cast materials and with properties of
wrought material produced from cast structures. An essential aim of many solidification processes is to obtain optimum properties in the resultant material. This chapter
gives examples showing how the principles presented in earlier chapters can be
utilized to produce structures with improved mechanical or physical properties.
The book draws heavily on research conducted over the last decade at Massachusetts Institute of Technology by students and associates of the author. A special
note of thanks is due them. Critical comments and suggestions of John Cahn have
been received and acted on with pleasure. The bulk of the book was written while the
author was on sabbatical leave as Overseas Fellow at Churchill College, Cambridge
University, England. He is grateful for the unique combination of stimulation and
relaxation provided by that environment, and by his colleagues there.
MERTON C. FLEMINGS
1
HEAT FLOW IN SOLIDIFICATION
GROWTH OF SINGLE CRYSTALS
A variety of different techniques are employed to produce single crystals from melts.
These can be grouped in three categories as those in which the entire charge is melted
and then solidified from one end, a large charge is melted and a small crystal withdrawn
slowly from it, and only a small zone of the crystal is melted at anyone time. Figure
1-1 shows the methods schematically.
The first category of crystal-growing techniques is termed normal freezing.
A commonly used normal freezing method for low-melting-point metals is growth in
a horizontal boat. Here, a charge of metal is contained within a long crucible of small
cross section open at the top. A seed crystal may be placed at one end of the boat to
obtain a crystal of predetermined orientation.
The charge and part of the seed are
first melted in a suitable furnace. Next, the furnace is withdrawn slowly from the
boat so that growth proceeds from the seed; alternatively, the boat is withdrawn
slowly from the furnace and the solid-liquid interface moves until the whole charge is
solid. In a similar crystal-growing method, the crucible is vertical and open at the
top; this is often termed the Bridgeman method. In a minor modification of these
techniques, neither the furnace nor the crucible moves. The charge is melted and
2
HEAT FLOW IN SOLIDIFICATION
000000000
TO INERT GAS
SOURCE OR
VACUUM
COILS
(a)
TO
t
INERT
VACUUM
SOURCE
TO INERT GAS
SOURCE OR
VACUUM
o
o
o
o
o
o
o
o
o
o
o
o
o
o
GAS
OR
"FLOATING"
ZONE
LIQUID
0
0
0
~HEATING
o
COILS
o
LIQUID
o
o
0--
HEATING
COILS
o
o
FIGURE 1-1
Examples of crystal-growing
(c) floating zone.
methods.
(a) Boat method;
(b) crystal pulling;
equilibrated in a furnace constructed so that one end of the furnace is substantially
colder than the other end (temperature-gradient furnace). The temperature gradient
is maintained constant in the furnace, and crystal growth is obtained by slowly
lowering overall furnace temperature.
In growing single crystals, it is not necessary that the entire charge be molten.
For some purposes, it is desirable to melt initially only a portion of the charge and
move this molten zone slowly through the charge (zone melting and zone freezing).
Many types of heat sources are used for zone melting, including induction, resistance,
electron beam, and laser beam. The zone is moved either by mechanically moving
the power source with respect to the crystal or vice versa. Zone melting is done either
with or without crucible. The latter type, crucibleless zone melting, or floating zone
melting, is widely used for reactive and high-melting-point
materials. The molten
zone is h,eld in place by surface tension forces sometimes aided by a magnetic field.
GROWTH
OF SINGLE CRYSTALS
3
Another single-crystal-growing
technique, used widely for growing single
crystals of silicon, germanium, and nonmetals, is the crystal-pulling, or Czochralski,
technique. In this case, the charge material is placed in a crucible and melted. A seed
crystal is attached to a vertical pull rod, lowered until it touches the melt, allowed to
come to thermal equilibrium, and then raised slowly so that crystallization proceeds
from the seed crystal. The crystal is rotated slowly as it is pulled, and crystal diameter
is controlled by adjusting pull rate and/or heat input to the melt.
Many variations of these crystal-growing
techniques are described in the
literature. Crystals are generally grown in vacuum but may be grown in air or inert
atmosphere.
Highly volatile materials are encapsulated and grown under pressure.
Crystals with volatile species as alloy elements are also encapsulated, or grown under
flux. In one rather old process, the liquid is carried by a plasma arc as small droplets
from an electrode (Verneuil method). These and other techniques for growth of
specific materials have been described. 1
The basic heat-flow objectives of all crystal-growing techniques are to (1) obtain
a thermal gradient across a liquid-solid interface which can be held at equilibrium
(e.g., stable with no interface movement) and (2) subsequently to alter or move this
gradient in such a way that the liquid-solid interface moves at a controlled rate. A
heat balance at a planar liquid-solid interface in crystal growth from the melt is
written
KsGs - KLGL
where Ks
=
PsHR
(1-1)
= thermal conductivity of solid metal, cal/(cm)(DC)(s)
KL
= thermal conductivity of liquid metal, calj(cm)(DC)(s)
Gs
= temperature
gradient
in solid at the liquid-solid
interface,
gradient
in liquid at the liquid-solid
interface,
DC/cm
GL
= temperature
DC/cm
R
= growth velocity, cm/s
Ps
= density of solid metal, g/cm3
H = heat of fusion, cal/g
Note from Eq. (1-1) that growth velocity R is dependent, not on absolute thermal
gradient, but on the difference between KsGs and KLGL. Hence, thermal gradients can
be controlled independently of growth velocity. This is an important attribute of
single-crystal-growing furnaces since growing good crystals of alloys requires that the
temperature gradients be high and growth rate be low. K" KL> H, and Ps are constants
of the materials being solidified; GL is directly proportional to the heat flux in the
liquid at the liquid-solid interface.
4
HEAT FLOW IN SOLIDIFICATION
Growth velocity would be at a maximum when GL becomes negative (undercooled melt); however, good crystals cannot be grown in undercooled liquids, and so
the practical maximum growth velocity occurs when GL -4 0, or from Eq. (1-1)
= Ks~
R
max
(1-2)
PsH
G•• thermal gradient in the solid at the interface, is evaluated by experiment or heatflow calculations. As a simple illustrative example of calculation of solid gradient Gs,
consider the case of floating-zone (crucibleless) crystal growth in which (1) crystal is
of circular cross section, (2) heat transfer from the crystal to surroundings is by convection, (3) growth is at steady state, and (4) temperature gradients within the crystal
transverse to the growth direction are low. Consider a cylindrical element in the solid
crystal dx' in thickness, moving at the velocity R of the liquid-solid interface, Fig. 1-2.
Then, for steady state, the temperature of the moving element remains constant and a
heat balance is written (for unit time)
Net heat change
net heat change from
net heat change from _ 0
from couduction + moving boundary + loss to surroundings as d 2T (pscsna2 dX') - R dT (pscsna2 dX') - h(T - To)(2na dX')
dX'2
dx'
= 0
(1-3)
where x' = distance from liquid-solid interface (negative in solid), cm
= specific heat of solid metal, cal/(g)(°C)
a = radius of crystal, cm
h = heat transfer coefficient for heat loss to surrounding, call
Cs
(cm2Wc)(s)
T = temperature at x', °C
To
= ambient temperature, °C
Ps
= density of the
solid crystal, g/cm3
as = thermal diffusivity of the solid crystal (KslPsc.), cm2
Is
Now, integrating Eq. (1-3) with the boundary conditions that at x' = 0, T = Tm
= - 00, T = To, the temperature in the solidifying
metal is given by
(Tm = melting point of metal), at x'
T - To
~J
(1-4)
Gs = (dTldx')x
=0
= exp {-[~ 2as - J(~)22ets + aKs Xl}
The thermal gradient in the solid at the liquid-solid interface
then
is
SOLIDIFICATION
I SOLID
5
OF CASTINGS AND INGOTS
fS
LIQUID
i i
fu.I
a:
TM
:J
fCI:
a:
w
a.
a5
f-
FIGURE 1-2
Temperature
distribution
growth (schematic).
and when Rj2as
in
To
o
crystal
DISTANCE
FROM .INTERFACE,
x'
«1
Gs::::;
~
(TM-
To)
(1-6)
aKs )1/2
(2h
For crystals of high melting point, where (TM - To) is large and the coefficient
of heat transfer h is increased by radiation heat transfer, thermal gradients attainable
are quite high, 100°Cjcm or more. For lower-melting-point crystals, other cooling is
necessary to attain steep gradients. As an example, Mollard achieved gradients of the
order of 500°Cjcm in i-in-diameter tin crystals by using a thin steel crucible, resistanceheating the crucible at just above the liquid-solid interface and water-cooling it just
below the interface.2 Equation (1-6) is equally applicable to this arrangement, with h
now representing the total resistance to radial heat flow from the metal crystal to the
cooling water. For the arrangement employed, this resistance was primarily at the
metal-crucible interface and was about 0.04 calj(cm2)CC)(s). In Prob. 1-1 at the end
of this chapter we illustrate that calculated thermal gradients obtained using Eq. (1-6)
are about those attained experimentally, 500°Cjcm.
SOLIDIFICATION OF CASTINGS AND INGOTS
In most casting and ingot-making processes, heat flow is not at steady state as in the
above examples. Hot liquid is poured into a cold mold; specific heat and heat of
fusion of the solidifying metal pass through a series of thermal resistances to the cold
mold until solidification is complete. Figure 1-3 shows this process schematically for
solidification of a pure metal. Thermal resistances which, in general, must be considered are those across the liquid, the solidifying metal, and the metal-mold interface
6
HEAT FLOW IN SOLIDIFICATION
AIR
I
LIQUID
SOLID
UJ
a:
::;)
I«
a:
UJ
0..
AT.
METAL-MOLD
INTERFACE
AT,
MOLD-AIR
INTERFACE
:!;
UJ
IFIGURE 1-3
Temperature profile in solidification of a
pure metal.
TO
DISTANCE
and those in the mold itself. The problem is mathematically and physically complex
and becomes even more so when anything other than simple geometries are considered,
when thermal properties are allowed to vary with temperature, or when alloys are
considered. Problems such as these are now usually handled by computer methods,
and some examples will be considered later. There are, however, certain simplifying
approximations that can be made for a number of cases of engineering interest. Some
of these will be examined before considering the general problem in further detail.
CASTING PROCESSES EMPLOYING INSULATING MOLDS
Sand casting and investment casting are two processes for making shaped castings
which employ relatively insulating molds. Both are very old processes, and both are
important commercially today. 3
Figure 1-4 illustrates the sand-casting process used to make a segment of a
household radiator. Three sand-mold segments are made separately and assembled
to produce a mold cavity the shape of the final casting described. Patterns of wood or
metal are employed to make the proper impression in the upper and lower mold
halves; the sand is rammed in place over the pattern. The outer mold segments (cope
and drag) retain the shape of the impression because the sand used contains a few
percent of water and clay and sometimes other binding agents. The internal segment
CASTING PROCESSES EMPLOYING INSULATING
MOLDS
7
COMPLETED
CASTINGBROKEN OPEN
TO REVEAL
INTERIOR
MOLD
SECTION
A
MOLD
SECTION
B
FIGURE 1-4
Sketch of sand-casting process as used in manufacture of a household radiator
(From Taylor, Flemings, and WuljJ.4)
(core) is generally made of baked oil or resin-bonded sand to achieve greater strength
and to reduce the amount of volatile components.
In cope-and-drag investment casting, mold pieces are made, not by ramming a
dry or nearly dry sand mixture, but by pouring a slurry of investment material. One
such material widely used for nonferrous alloys is plaster. For ferrous materials, a
suitable material is mullite bonded with ethyl silicate.
In lost-wax casting, the pattern is originally made of wax and then invested with
a suitable slurry which is subsequently baked at high temperature.
During the hightemperature baking, the wax melts and drips out or volatilizes along with moisture
in the mold. Two types of lost-wax casting are common. In the older process, the
wax is placed in a box, or can, and a slurry poured to fill the box (Fig. 1-5). In the
new shell-investment-casting
process, the pattern is dipped successively in a slurry and
then in a fluidized bed of fine particulate material until a shell of desired thickness is
built up. The basic advantage of both types of lost-wax casting is that the process
allows intricate parts to be made without regard to the problem of pattern removal
from the mold. The major advantages of investment-casting processes as compared
with sand castings are the greater complexity, thinner sections, and better dimensional
accuracy and surface finish that can be obtained. The major disadvantage is their
usually greater cost and size limitation.
8
HEAT FLOW IN SOLIDIFICATION
Preparing
a Mold for
Investment Casting
The "Lost Wax" or Precision Casting Process
u
Wax
n.;
is melted and
injected into a metal
die to form the
In-gate
disposable
Sprue
patterns.
Pattern
Metal die
Pouring
cup -
palleZ:::' ''''
Patterns are "welded" to wax gates
and runners Ita form a "tree".
The "tree"
is
precoated
A metal flask
by
is nexl placed
around the
dipping in a
refractory slurry
and is then dusted
"tree"
and
sealed to the pallet: then
the investment, a coarser
with refractory sand.
refractory
in a more viscous
slurry is poured around the
precoated "tree".
,
Wax
IL.J--;'.
Finally, before casting, the
mold is placed in a furnace
, I .\.
dripping5'---....',
j
When the investment has
and carefully fired to
1300-1900'F.
to rf/1wve
"set",
all wax residue
the mold is placed in
an oven at 200'F. to
investment and melt
wax pattern.
dry the
out the
The
mold
hot )Ready
to is
Pour.
and free of any
trace of wax.
and
reach the temperature at
which it will receive the
molten metal.
FIGURE 1-5
Preparing a mold for investment casting. (From Taylor, Flemings, and WuljJ.4)
From a heat-flow standpoint, the important characteristic of solidification of a
metal in processes such as those discussed above is that the metal is a much better
conductor of heat than the mold. Thus, solidification rate depends primarily on
thermal properties of the mold. The thermal conductivity of the metal has practically
no influence. Also, except in relatively heavy-section shell investment castings, the
mold can be considered to be semi-infinite in extent; i.e., the outside of the mold does
CASTING PROCESSES EMPLOYING
INSULATING
MOLDS
9
not heat up during solidification. The heat-flow problem sketched in Fig. 1-3 is now
very much simplified, especially if we assume further that the metal is poured with no
superheat, that is, exactly at its melting point TM' as shown in Fig. 1-6.
Consider first the problem of unidirectional heat flow. Metal is poured exactly
at its melting point against a thick, flat mold wall initially at room temperature To.
Thus, the mold surface is heated suddenly to TM at time t = O. This is a transient,
one-dimensional heat-flow problem, and the solution must conform with the partial
differential equation
-aT
at =
where
am
ax
(1-7)
= thermal diffusivity of mold, cm2(s
Km
= therl1}.al conductivity
Pm
= density of mold, g(cm3
t
am a2T2
= time,
of mold, cal(cmWC)(s)
s
x = distance from mold wall, cm (negative into the mold)
The solution to this equation for the boundary conditions stated above gives the
temperature T in the mold as a function of time t at distance from the mold surface x:
T - TM
To - TM
= erf
-x
_
2)amt
(1-8)
where erf denotes the error function. The error function of zero is zero, and the error
function of infinity is unity. A list of tabulated error functions is given in Appendix A.
The rate of heat flow into the mold at the mold-metal interface is given by
A x=o = _Km(aT)
ax
(:L)
x=o
(1-9)
where x increases positively from left to right in Fig. 1-6, q is rate of heat flow, and A
is area of the mold-metal interface. By partial differentiation of Eq. (1-8) with respect
to x, letting x = 0 and combining the results with Eq. (1-9), the rate of heat flow
across the mold-metal interface is seen to be
A x=o
(CL)
=
(1-10)
ret
_JKmpmCm
(TM - To)
where em is specific heat of the mold material. Now, the heat entering the mold
comes only from heat of fusion of the solidifying metal since the solid as well as the
liquid metal is exactly TM (Fig. 1-6). Thus,
A x=o
(CL)
at
= _p S H as
(1-11)
- Xem thêm -