CRYSTllllZATION

TECHNOLOGY
HANDBOOK
CRYSTAlllZATION
TECHNOlOGY
HANDBOOK
Second Edition
Revised and Expanded

edited by
A. Mersmann
Technische Universitiit Munchen
Garching, Germany

0 ~~~,~~!~ro,p
Boca Raton London New York

CRC Press is an imprint of the

Taylor & Francis Group, an informa business
Reprinted 2010 by CRC Press
CRC Press
6000 Broken Sound Parkway, NW
Suite 300, Boca Raton, FL 33487
270 Madison Avenue
New York, NY 10016
2 Park Square, Milton Park
Abingdon, Oxon OX14 4RN, UK
ISBN: 0-8247-0528-9

Headquarters
Marcel Dekker, Inc.
270 Madison Avenue, New York, NY 10016
tel: 212---696-9000; fax: 212-685-4540

Eastern Hemisphere Distribution

Marcel Dekker AG
Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland
tel: 41-61-261-8482; fax: 41-61-261 8896

World Wide Web

http://www.dekker.com

The publisher offers discounts on this book when ordered in bulk quantities. For more infor­
mation, write to Special Sales/Professional Marketing at the headquarters address above.

Copyright© 2001 by Marcel Dekker, Inc. All Rights ReseITed.

Neither this book nor any part may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, microfilming, and recording, or
by any information storage and retrieval system, without permission in writing from the
publisher.
Preface

The aim of this book is to provide reliable information not only on the
science of crystallization from solution and from melt but also on the
basic design methods for laboratory and especially industrial crystallizers.
Up to now the niche between scientific results and practical design and
operation of large-scale crystallizers has scarcely been filled. A work devoted
to this objective has to take into account relevant crystallization phenomena
as well as chemical engineering processes such as fluid dynamics, multiphase
flow, and heat and mass transfer. In the design of crystallizers, experiments
are initially performed on laboratory crystallizers to obtain kinetic data. In
this book, information is given on reliable scale-up of such crystallizers. The
selection, design, and operation of large-scale industrial crystallizers based
on fundamentals is the most significant objective of this work. To this end,
an appendix listing important physical properties of a large number of
crystallization systems is included. A selection of design data valid for indus­
trial crystallizers with volumes up to several hundred cubic meters demon­
strates the applicability of the design and scale-up rules.

iii
iv Preface

To date, the design of crystallizers has not been achievable from first
principles. The reason is very simple: a complex variety of different pro­
cesses occur in crystallizers, such as nucleation, crystal growth, attrition and
agglomeration of crystals, fluid dynamics, and heat and mass transfer. Some
of these processes are not yet well understood although the design and
operation of large-scale crystallizers require reliable knowledge of the
most essential processes. The book presents a survey of the state of the
art and stresses the interrelationships of the essential mechanisms in such
an apparatus. Furthermore, with respect to nucleation and crystal growth,
general approaches have been developed to predict the kinetic rates that are
needed for chemical engineering design and a new chapter on agglomeration
has been added.
Supersaturation is the decisive driving force with respect to the kinetics of
crystallization. Optimal supersaturation is a prerequisite for the economical
production of crystals with a desired size, shape, and purity. The book offers
information on the most suitable supersaturation requirements in labora­
tory and industrial crystallizers. Not only are aspects of cooling and
evaporative crystallization considered, but drowning-out and reaction crys­
tallization are also described in detail. In dealing with precipitation the
complex interrelationships between mixing and product quality are dis­
cussed. A special segment is devoted to the problem of how the process
components of an entire crystallization process can be economically fitted
together. The aspects stressed are always those of production quality; size
distribution, coefficient of variation, crystal shapes and purity, and the
problem of encrustation are considered. One chapter is devoted to the con­
trol of crystallizers and another deals with the role of additives and impu­
rities present in the solution. Crystallization from the melt is described in full
detail, and information is given on how to design and operate the corre­
sponding crystallizers. The book describes the most significant devices for
crystallization from the melt and solidification processes. Process develop­
ment such as high-pressure crystallization and freezing are considered and in
this second edition new results on direct contact cooling crystallization have
been added.
My goal has been to edit a work as homogeneous and practical as pos­
sible. The book is therefore not a mere collection of independent articles
written by several contributors but a coordinated handbook with a single list
of symbols and a unified bibliography. It is divided into 15 chapters to make
it easier to find points of interest. Only simple derivations and equations
absolutely necessary for understanding and for calculation are presented.
The book is based on literature that is available worldwide (especially
references from the United States, Europe, and Japan) and on the direct
experience of the contributors. Some of the contributors work in the indus-
Preface V

trial sector, and nearly all have spent some time in industrial plants. A few
fundamental chapters were written by scientific researchers at universities.
Because this volume addresses the theory and practice of crystallization, it
should be valuable in both academia and industry.

A. Mersmann
Contents

Preface iii
Contributors xiii

1. Physical and Chemical Properties of Crystalline Systems 1

A. Mersmann

1. Measures of Solubility and Supersaturation 2

2. Solubility and Phase Diagram 7
3. Heat Effects (Enthalpy-Concentration Diagram) 13
4. Crystalline Structure and Systems 16
5. Polymorphism and Racemism 19
6. Real Crystals (Polycrystals) 22
7. Physical Properties of Real Crystals 27
8. Surface Tension of Crystals 36
References 42

vii
viii Contents

2. Activated Nucleation 45
A. Mersmann, C. Heyer, and A. Eble

1. Homogeneous Nucleation 47
2. Heterogeneous Nucleation 58
3. Surface Nucleation on Crystals 67
4. Comparison of Nucleation Rates 71
References 76
3. Crystal Growth 81
A. Mersmann, A. Eble, and C. Heyer

1. Diffusion-Controlled Crystal Growth 83

2. Integration-Controlled Crystal Growth 86
3. Estimation of Crystal Growth Rates 98
4. Growth in Multicomponent Systems and Solvent Effects 113
5. Influence of Additives and Impurities 119
6. Metastable Zone, Recommended Supersaturation 127
References 137
4. Particle Size Distribution and Population Balance 145
A. Mersmann

1. Particle Size Distribution 145

2. Population Balance 150
3. Clear-Liquor Advance (Growth-Type Crystallizers) 165
4. Fines Destruction with Solute Recycle 167
5. Classified Product Removal with Fines Destruction 170
6. Classified Product Removal 172
7. Deviations from the MSMPR Concept 174
8. Population Balance in Volume Coordinates 182
References 184

5. Attrition and Attrition-Controlled Secondary Nucleation 187

A. Mersmann

1. Attrition and Breakage of Crystals 189

2. Growth of Attrition Fragments 204
3. Impact of Attrition on the Crystal Size Distribution 210
4. Estimation of Attrition-Induced Rates of Secondary
Nucleation 223
References 231
Contents ix

6. Agglomeration 235
A. Mersmann and B. Braun

l. Population Balance 239

2. Interparticle Forces 248
3. Agglomeration Rates 260
4. Avoidance and Promotion of Agglomeration 268
5. Collisions in Multiparticle Systems 272
6. Tensile Strength of Aggregates 278
References 281
7. Quality of Crystalline Products 285
A. Mersmann

l. Median Crystal Size 286

2. Crystal Size Distribution 296
3. Crystal Shape 299
4. Purity of Crystals 310
5. Flowability of Dried Crystals and Caking 315
References 321
8. Design of Crystallizers 323
A. Mersmann

1. Crystallization Apparatus 324

2. Operating Modes 334
3. Mass Balance 345
4. Energy Balance 354
5. Fluidized Bed 357
6. Stirred Vessel (STR) 360
7. Forced Circulation 377
8. Heat Transfer 379
9. Mass Transfer 386
References 390
9. Operation of Crystallizers 393
A. Mersmann and F. W. Rennie

1. Continuously Operated Crystallizers 393

2. Batch Crystallizers 400
3. Seeding 410
4. Crystallizers for Drowning-Out and Precipitation 414
5. Sampling and Size Characterization 426
X Contents

6. Incrustation 437
7. Fitting the Process Parts Together 450
References 471
10. Challenges in and an Overview of the Control of Crystallizers 475
S. Rohani

1. Introduction 475
2. Dynamic Modeling of Crystallization Processes 479
3. Instrumentation in Crystallization Control 484
4. Control of Crystallization Processes 490
5. Conclusion 508
References 509
11. Reaction Crystallization 513
R. David and J. P. Klein

1. Introduction 513
2. Driving Force of Reaction Crystallization 515
3. Reaction Crystallization Kinetics 520
4. Fluid Dynamics, Mixing, and Precipitation 530
5. Conclusion: A General Methodology to Solve a Reaction
Crystallization Problem 555
References 557
12. "Tailor-Made Additives" and Impurities 563
I. Weissbuch, L. Leiserowitz, and M. Lahav

1. Introduction 563
2. Tailor-Made Additives for Crystal Morphology Engineering 564
3. Tailor-Made Additive Molecules for Crystal Dissolution;
Stereospecific Etchants 572
4. Theoretical Modeling 587
5. Mode of Occlusion of Impurities in Crystals 591
6. Tailor-Made Additives for Inhibition and Promotion of
Crystal Nucleation 602
7. Outlook 612
References 612
13. Suspension Crystallization from the Melt 617
K. Toyokura and I. Hirasawa

1. Introduction 617
Contents xi

2. Fundamentals of Crystallization from the Melt 620 ·

3. Suspension Crystallization (Indirect Heat Transfer) 640
4. Direct Contact Cooling Crystallizers 648
References 660

14. Layer Crystallization and Melt Solidification 663

K. Wintermantel and G. Wellinghoff

l. Layer Crystallization 663

2. Melt Solidification 683
References 700

15. Thermal Analysis and Economics of Processes 703

A. Mersmann

1. Capital Costs of Crystallizers and Operating Parameters 704

2. Role of Incrustation for Economics 707
3. Model of Solids Production Processes 708
4. Energy of the Evaporation Step 710
5. Energy of the Drying Step 714
6. Solid-Liquid Separation 717
7. Crystallization or Precipitation Step 718
8. Thermal Analysis of the Entire Process 720
9. Overall Economics 721
References 723

Appendix 725
Notation 787
Bibliography 799
Substance Index 807
Subject Index 819
Contributors

B. BRAUND Technische Universitat Miinchen, Garching, Germany

R. DAVID Ecole des Mines d'Albi-Carmaux-CNRS, Albi, France
A. EBLE Technische Universitat Miinchen, Garching, Germany
C. HEYER Technische Universitat Miinchen, Garching, Germany
I. HIRASAWA Waseda University, Tokyo, Japan
J. KLEIN Universite Claude Bernard-CNRS, Villeurbanne, France
M. LAHAY The Weizmann Institute of Science, Rehovot, Israel
L. LEISEROWITZ · The Weizmann Institute of Science, Rehovot, Israel
A. MERSMANN Technische Universitat Miinchen, Garching, Germany
F. W. RENNIE Du Pont de Nemours & Co., Wilmington, Delaware
S. ROHANI The University of Western Ontario, London, Ontario,
Canada
xiii
xiv Contributors

K. TOYOKURA Waseda University, Tokyo, Japan

I. WEISSBUCH The Weizmann Institute of Science, Rehovot, Israel
G. WELLINGHOFF BASF AG, Ludwigshafen, Germany
K. WINTERMANTEL BASF AG, Ludwigshafen, Germany
1
Physical and Chemical Properties of.
Crystalline Systems
A. MERSMANN Technische Universitat Miinchen, Garching, Germany

Unlike crystallization from melting and freezing, heat transfer is not the
decisive process in cystallization from solution because the properties of
the crystalline product depend primarily on supersaturation generated by
cooling, evaporation, drowning out, or a reaction. The quality components
of the product (i.e., its crystal size distribution, its median crystal size, its
purity, and its crystal shape) are strongly influenced by (a) the geometry and
type of the crystallizer, (b) the operating conditions, and (c) the properties of
the liquid and solid phases.
The requirements of the product rather than the method of creating
supersaturation are decisive when selecting a crystallizer. In an industrial
crystallizer, there is such a variety of complex processes that it is difficult for
the chemical engineer to decide on a suitable procedure for the design of a
full-scale apparatus. Therefore, let us consider some general guidelines.
If the aim is to obtain a certain product having a specific crystal size
distribution, grain size, and purity instead of a random product, it is
necessary to control the local and mean supersaturation as well as the
residence time of the solid in the supersaturated solution. Supersaturation
1
2 Mersmann

is a prerequisite for nucleation and growth, which are decisive not only for
the formation of a solid phase but also for its occurrence (i.e., size distribu­
tion of crystals and their shape). The degree of supersaturation is deter­
mined by the flows of materials and energies, on the one hand, and by
crystallization kinetics, such as nucleation and growth, on the other hand.
In addition to the conservation laws of materials and energy, the population
balance is very important because in most cases, a crystalline product with a
certain size distribution is required. With respect to this fact, it is necessary
to take into account processes that influence the population balance (i.e.,
mainly agglomeration for very small crystals and attrition for large crystals).
As a rule, crystals have a higher density than that of the surrounding
liquid, which results in settling. Therefore, a certain upflow current is needed
in all crystallizers to compensate settling. As a result, crystallizers are
equipped with rotors, such as stirrers or pump impellers, which can cause
attrition of large crystals. In addition, the quality of a crystalline product
may depend on the fluid dynamics of the slurry in the crystallizer. As a rule,
the distribution of supersaturation and solid material in a crystallizer
depends on the process of macromixing. This may be important for coarse
crystalline products. With drowning-out and reaction crystallization, the
local supersaturation is influenced by the process of micromixing (i.e., mix­
ing on a molecular scale). Finally, the production rate of an industrial
crystallizer may be high or low. As a rule, continuous_ crystallizers are
used in the case of high production rates because this operation is more
economical with respect to investment, energy, and labor costs. If several
products are to be crystallized in the same crystallizer, a batch crystallier is
chosen. Sometimes, crystallizers with circulating slurry or with fluidized
beds are used instead of stirred vessels. As will be shown later, the fluid
dynamics and, consequently, phenomena such as mixing and attrition differ
substantially in such apparatus. Therefore, it is necesary to provide detailed
information on the various crystallizers and their flow behavior. As stated
earlier, the most important process parameter is supersaturation, or the
difference between the actual concentration and the equilibrium concentra­
tion of the liquid.

1. MEASURES OF SOLUBILITY AND

SUPERSATURATION
First, information on the phase equilibrium of solid-liquid systems and
solubility and melt diagrams will be presented. Crystallization from the
melt is described in detail in Chapters 13 and 14. However, because there
is no distinct boundary between crystallization from solution and crystal-
Physical and Chemical Properties 3

lization from the melt, it is reasonable to regard phase equilibria diagrams in

a general way. It will also be shown how crystallization processes can best be
represented in enthalpy-concentration diagrams. Important basic principles
of thermodynamics are followed by an explanation of the essential pro­
cesses: the rate of nucleation and crystal growth. It is precisely these kinetic
parameters that determine the crystal size distribution of a product having a
large number of crystals. First, however, let us take a look at concentration
and supersaturation measures.
The number of collisions of elementary units (atoms, ions, molecules)
with those in the fluid phase or at the phase interface of the crystalline
phase depends on the number of units per unit volume of the fluid phase:
Number of units = nNA = CN (1.1)
Volume of fluid phase V A

where C (mol/L or kmol/m3) is the molar concentration and NA is

Avogadro's number. For reasons of practicality, mass concentration c is
often used:

c - CM
- [kmol
m3 •
kg ] [kg
kmol or m 3 -
g g]
dm 3 - I (1.2)

In addition to these volumetric concentrations, it is convenient to use mole

or mass fraction y or was well as mole or mass ratios Y or W (generally, Y
and Y should be used in the fluid phase and X or x in the solid phase). A
scale is often used to determine the mass that can be converted to the
amount of substance if the molar mass is known. Table 1.1 provides infor­
mation on definition and conversions.
A saturated fluid phase having concentration C* or c* is in thermody­
namic equilibrium with the solid phase at the relevant temperature. If the
solution is liquid, the saturation concentration often depends strongly on
temperature but only slightly on pressure. If a fluid phase has more units
than C* NA, it is said to be supersaturated. Crystallization processes can take
place only in supersaturated phases, and the rate of crystallization is often
determined by the degree of supersaturation. Supersaturation is expressed
either as a difference in concentration
l::!.C =C - C* or l::!.c =c- c* (1.3)
or as relative supersaturation
C
S =-
C*
= -c*C or u =S - I (1.4)

Generally, differences or ratios of molar and mass fractions can also be

used; however, precise and detailed information is always required when
Table 1.1. Definitions and Conversion of Concentration Units "'"
Referred to mass [kg] Referred to amount of substance [kmol]
M;=C;V n;=M;/M;

Two components k components Two components k components

k k
Total mass M =Ma+Mh M = LMi Total amount of n = na + nb n= Lni
)=a substance J=a

Ma M; X = na
Mass fraction U' =- W·=- Mole fraction Xj=~
" M I M n n

W =Ma n;
Mass ratio W-=_!!.J._ Mole ratio X =:.:E. X:-= -
a Mb 1 a nb 1
Mcarrier ncarrier

Conversion Mass fraction w; +----------➔ Mole fraction x;

~ -I ~ -I
Mb 1-x") X;M; x = ( I + -~-h _I_-_w_a) wi/ M;
Mass fraction from Wa- _(I +------ wi= k
Mole fraction from X; = k
mole fraction Ma Xa mass fraction a Ma Wa
L(x;Mj) L(wifMi)
;-11 )=a

Conversion Mass ratio Mass fraction Mole ratio ~-----> Mole fraction
IV IV X
IV=-- w=--- X= X
w I- l+IV 1-x X= l+X

~ Af (IV l-w 0 ~ k ( k (
wj ]-1
Mean molar mass: Mm= - [
= --d!- +--_- )-I ; general: Mm= L [ ·y )]-I Mean density: Pm = L
m A1a Mb i=a Mi i=a PJ ~
~
;;i
3
~
=
=
Physical and Chemical Properties 5

the crystalline phase integrates (e.g., solvents, as is the case with all
hydrates). The ratios kg anhydrate/kg solvent and kg hydrate/kg solvent
are always different and, therefore, so are the supersaturation values. The
dimensionless supersaturations S and a also often vary considerably
depending on whether, for example, values of a 1 = flC/C*, a2 = fly/y*,
a 3 = fl Y / Y", a4 = fl w/w*, or a 5 = fl W / W* are involved. These statements
also show that the expression of supersaturation as a percentage is entirely
insufficient.
If the phase involved is a vapor or gaseous supersaturated phase, it is
often useful to use partial pressures instead of concentrations. If the ideal
law of gases applies, the following is valid:

C*=L or C* =poM (l.5)

RT fflT
or
p C=pM
C=- or (l.6)
RT fflT
This gives the following relative supersaturation for isothermal systems:

S=.E_=f!_ (1.7)
C* Po

and
p-po
a =po
-­ (1.8)

Generally, it can be said that the difference in chemical potential flµ, = µ,F
-µ,c between the fluid (index F) and the cryal (index C) phases is the kinetic
driving force, which can be described via the relationship among the che­
mical potential µ,, standard potential µ,0 , the activity a, and the heat of
crystallization flHcL:
µ, = Jl,o + fflTln a (l .9)
or

flµ, * = -flµ,
91T
= v In (a)
-
a*
= v ln Sa (solution) (l. lOa)

flµ,* = ~ = flT (melt) (l.lOb)

flHh T

with the undercooling flT.

6 Mersmann

Table 1.2. Driving Forces for Nucleation and Growth

Kinetic parameter Driving force

Growth AC[kmol/m 3]or

Diffusion controlled
Ac [kg/m 3]
Surface-integration
controlled
<T = AC/C*
Nucleation Primary (homogeneous,
S=l+a=l+AC/C*
heterogeneous)
Secondary
= AC/C* or
<T
AC= Ac/M

Equation (1.10a) is especially important in drowning-out or precipitation

crystallization when a component crystallizes out of an electrolytic solution
and decomposes into ions. For example, if an ionic crystal consists of vA = x
ions of species A and va = y ions of species B with v = vA + va = x + y, it
can dissociate into anions and cations:

(1.11)

In this equation, z is the valency of the ions. The solubility product can be
formulated with activities a = y • y (y is the activity coefficient and Y; the
mole fraction of component i) or concentration C:

(1.12a)

or

(1.12b)

The supersaturation Sa based on activities is given by

(1.13a)

and the supersaturation is then for S = 1 + a = C/ C* is then for x =y = 1

(1.13b)

See Tables 1.2 and 1.3.

Physical and Chemical Properties 7

Table 1.3. Definitions of Supersaturation in Crystallizing Systems

Solubility product
With activity Ka= (alr(ajJf = bAYAr(,BYBY
With concentration Kc= (c;.r(csY
C* = ( Kc ) 1/(x+y)
Solubility for binary electrolytes xxyy
or with x =y = 1
c·=~
Sa= ((aA):~aBY) l/(x+y)

Supersaturation S = ((CA):~CB)Yy/(x+y)

or with x =y = I and 'YA = 'YB = 1

S=~= ,.;c;:c;
c· ~

2. SOLUBILITY AND PHASE DIAGRAM

The saturation concentration of a substance in a solvent is obtained experi­
mentally by determining the maximum amount that is soluble at a given
temperature. Figure 2.1 illustrates several solubility curves for anhydrates;
and Figure 2.2 shows solubility curves for hydrates. Solubility often increases
with temperature, but there are also other systems in which the saturation
concentration remains approximately constant or decreases with increasing
temperature. The solubility curve for hydrates has a kink where the number
of solvent molecules per molecule of dissolved substance changes.
The dissolution of two substances in one solvent can be represented in an
equilateral triangular coordinate network. Mole or mass fractions of the
three components are plotted along the lateral lines. Figure 2.3 illustrates
the ternary system of sodium carbonate and sodium sulfate in water. The
area above the saturation isotherm toward the water corner is the under­
saturated field, which contains a clear solution.
Each corner of the triangle represents a pure substance [i.e., solvent
(water) and the two dissolved substances, in this case Na 2CO 3 and
Na 2SO4]. The three possible binary mixtures are represented by a lateral
8 Mersmann

~ 0.8
.£
.a
::J
0 0.4
Cl)
ll~
0
0 20 40 60 100
Temperature a [°C]
Figure 2.1. Solubility curves for several anhydrates.

Temperature a [°C]
Figure 2.2. Solubility curves for several hydrates.

line and every ternary mixture by a point inside the triangle. The area above
a saturation isotherm toward the solvent corner represents undersaturated
solutions. This type of triangular network is suitable for drawing·up mate­
rial and mass balances for evaporative crystallization, as will be shown later.
As there is no clear boundary line between crystallizations from solutions
and from melts, such difficulties in distinction also arise in the discussion of
phase-state diagrams. When only one component precipitates in a pure state
from a binary, real liquid mixture, we speak of crystallization from solution.
However, when a binary mixture exhibits virtually ideal behavior, mixed
crystals precipitate during crystallization and this is known as melt crystal­
lization. The terms eutectic (Latin for "well shaped") and peritectic (Latin
for "covered") will be explained in more detail by means of equilibrium
diagrams. According to Ref. 2.1, a distinction can be made between two
Physical and Chemical Properties 9

0.1 0.2 0.3 0.4

Mass fraction w~.,so, -

Figure 2.3. Triangular diagram of the sodium carbonate-sodium sulfate­

water system.

main groups of binary mixtures (see fig. 2.4): (a) eutectic systems and (b)
systems with mixed-crystal or solid-solution formation.
According to Roozeboom [2.2], noneutectic systems can be divided into
five subgroups:
Type I. A (x = 0) and B (x = 1) form a continuous series of mixed crystals
(e.g., anthracene/carbozole) [2.3].
Type II. A and B form a continuous series of mixed crystals, which is, how­
ever, modified by a maximum value (e.g., o-carvoxim/L-carvoxim) [2.4].
Type III. A and B form a continuous series of mixed crystals, which is,
however, modified by a minimum value (e.g., m-chloronitrobenzene/m­
fluoronitrobenzene) [2.5].
Type IV. A and B form a series of mixed crystals interrupted by peritectic
(e.g. eikosanol/hexakosanol) [2.6].
Type V. A and B form a series of mixed crystals interrupted by an eutectic
(e.g. azobenzene/azoxybenzene) [2.7].
A mixture with an eutectic concentration crystallizes when the tempera­
ture falls below the eutectic temperature, forming uniform mixed crystals;
that is, the two types of mixed crystals crystallize simultaneously. During
10 Mersmann

0)

l e--------i
~

E solid
i!l!~---~--'
0 Mole fraction x
@ System with Eutecticum

&~
y .s • I
Xs &~•XX::, X
y'
X

O x,y· 0 x,y· 1

0 x,y· O x,y·
@ Systems with Mixed Crystal Formation

Figure 2.4. Eutectic (a) and noneutectic (b) systems.

peritectic solidification, first one phase crystallizes, then the other. In prac­
tice, eutectic systems are most common, followed by pairs of substances
exhibiting continuous formation of mixed crystals (type I). Figure 2.5
explains the systems in more detail by giving some concrete examples.
The overview panel in the figure contains some liquid-solid equilibria. The
left-hand side of the figure shows the melting-point diagram with respect to
the molar or mass fraction, and the right-hand side shows the equilibria: the
relationship between the concentration y* in the liquid phase and the con­
centration x in the solid phase. The top line represents a virtually ideal
mixture exhibiting partial solubility, and the bottom line a mixture exhibiting
no solubility of its components. The solubility diagrams on the left contain a
melting-point line corresponding to concentration y* and a solidification line
corresponding to concentration x. As a system in equilibrium must have the
same temperature in both phases, the corresponding x value can be read off
for every y* value and be represented in the equilibrium diagram on the right.
If a homogeneous solution is cooled, the left diagram can be used, for exam­
ple, to determine the temperature at which the first crystal is formed and the
concentration of the crystal. It is then possible to determine the concentra-
Physical and Chemical Properties 11

80 Homogenous solution
·c

40 Homogenous Solid
0 Mole fraction x,y*
Solidx
2,4 Dinitro- 2,4 Dinitro-
brombenzene chlorbenzene

100
·c
IC>
I!!
80
.3 ·=--
~
(1)
C.
60 §
E ~
i5! 40 bMixed•
0 crystals
crys

O Mole fraction x,y* I

Solid x
2,4 Dinitro- 2,4 Dinitro-
anisol chlorbenzene

Homogenous I
1
solution
150
,c,
I!!
~ 100
(1)
·c
/ XL
§
crystals

~ ~
C.
E
i5! E

oo~----------'
Solidw
Ice KCI

Figure 2.5. Melting and solidification curves of several systems.

tion of the liquid phase that is in equilibrium with the solid phase, the per­
centages of liquid and solid, and the composition of the last drop of liquid
before it solidifies. These diagrams not only resemble boiling-point and dew­
point curves of vapor and liquid equilibria but also allow analogous applica­
tions (law of mixtures, lever principle).
12 Mersmann

If phase splitting exists, solidification and melting-point curves

also occur, meeting at the eutectic point. At a specific temperature, a solid
of composition x and a liquid of composition y* are in equilibrium.
The eutectic point divides the entire concentration range into two sections.
In the first section, a certain substance in the liquid phase is at a higher
concentration than in the solid phase, whereas in the second section,
the reverse applies. At the temperature of the eutectic point, a liquid
exhibiting the concentration at this point exists. The small symbols
in Figure 2.5 indicate when a single-phase liquid or solid system is involved,
when a two-phase solid-liquid system exists, and whether mixed crystals
occur.
Finally, the bottom line represents the system of potassium chloride in
water. In this case, the phase splitting extends over the entire concentration
range. This system also has an eutectic point. For example, if an aqueous
KCl solution having a mass fraction of less than 0.2 is cooled, crystals of
x ~ 0 (i.e., virtually pure water crystals) are formed below the melting-point
curve. This behavior is used, for example, in the desalination of seawater by
freezing out water. The equilibrium diagram on the right shows that the
miscibility gap extends across the entire concentration range. With regard to
thermal separation, it should be noted that only one component precipitates
in a pure state in an eutectic system, whereas two components cannot gen­
erally be obtained in a pure state in systems where mixed crystals are con­
tinuously formed. In the case of an eutectic mixture, one component can be
obtained in each side of the eutectic point (cf. Fig. 2.4) in a single theoretical
separation step.
Two or more components of a binary solution or melt may combine
to form one or more different compounds. When water is the solvent
such a compound is called a hydrate, and for nonaqueous systems, the
term solvate is used. If such a compound is in stable equilibrium with
the surrounding liquid phase, the compound formed has a congruent melt­
ing point. However, if the solid-liquid system does not assume a stable or
time-invariant equilibrium, the melting point of the compound is termed
incongruent.
Sometimes, a third substance or another solvent is added to a solution or
melt in order to separate the components of the binary system and/or to
obtain a product of high purity. For example, the addition of urea to a
mixture of n- and i-paraffin leads to solid urea-n-paraffin adducts which
can be separated mechanically from the liquid phase, and the adducts can be
destroyed in water. In general, adductive crystallization is applied to separate
mixtures that cannot be isolated economically by other separation processes.
As a rule, the adducts are destroyed by heating or dissolution in a suitable
liquid.
Physical and Chemical Properties 13

If a third component is added to a mixture to alter its solid-liquid equili­

brium, the process is called extractive crystallization. For instance, it is
possible to separate mixtures of different isomers of cresol by the addition
of acetic acid.

3. HEAT EFFECTS (ENTHALPY-CONCENTRATION

DIAGRAM)
The solidification of solute molecules on a crystal is accompanied by heat
effects. The heat of crystallization is the negative value of the heat dissolu­
tion, which can be derived from the solubility curve for ideal systems. The
heating effect involved when a solid i dissolves in liquids or when a sub­
stance i crystallizes out of solutions can be calculated if the activity a; or the
activity coefficient JI; = a;/y; is known in relation to temperature. The
expression

l::!..H;* = -m [alnai]
a(l/T) P fior a;-+ a;
* (3.1)

represents the differential heat of solution, which describes the heating effect
when this solid i dissolves in a solution that is almost saturated. In such
solutions, the solid and fluid phases are practically in equilibrium, which
explains why the values t::..Hi and ai have an asterisk. As the value t::..H;
represents the heating effect of the last soluble molecules, this is known as
the last differential heat of solution. The negative value of t::..Hi represents
the heat of crystallization, which is either given off or taken up during
crystallization. First differential heat of solution is the term that applies
when a substance i is dissolved in a pure solvent or in an infinitely diluted
solution:

t::..H;* = -m [alna;]
a(l/T) P 1or
t'
a;-+ o (3.2)

The entire heat ofsolution is the energy released or absorbed when a solute is
added isobarically/isothermally to a pure solvent until a specific concentra­
tion is obtained. As illustrated in Table 3.1, the first heat of solution of
different substances can vary considerably.
If the mixtures are ideal, a; = Y; or JI;= 1. Equation (3.1) then becomes

t::..H'!" = -m[alnyi] (3.3)

I a(l/T) p
14 Mersmann

Table 3.1. First Heat of Solution of Salts in Water

at 20°C
Salt Heat of solution 6.HcL [kJ/mol]
A}i(S04)3 -500
Na2S04 -1.18
NcCI +3.9
Na2S04·lOH20 +78

In this case, the heat of solution is equal to the heat of fusion of a pure
substance (cf. Chapter 13). Equation (3.3) can be written in the form

(3.4)

or
a(ln Ci) ~ (- t:,.Hi) (3.5)
a(ln T) ffiT
It will be shown later that the slope a(ln Ci)/a(ln T) of the solubility curve or
the heat of crystallization is very important for the decision on the mode of
crystallization and also for the problem of encrustation. Principally speak­
ing, the slope of the solubility curve can be positive, zero, or negative
(inverse solubility dependence). Further information is given in Chapter 9.
Heat balances can easily be established in enthalpy-<:oncentration dia­
grams for binary systems. In such diagrams, the molar enthalpy H (or the
specific enthalpy h) is plotted against the mole fraction y (or the mass frac­
tion w) with isotherms as a parameter, see Figure 3.1. The molar enthalpies
of the pure substances are the product of the molar heat capacities and the
temperature: H = CPT. Most diagrams are based on the Celsius tempera­
ture {}, but any convenient temperature can be chosen. If the binary systems
behave ideally, the isotherms in such enthalpy-<:oncentration diagrams are
straight lines. This means that mixing the two components that have the
same temperature does not lead to a change in temperature and the heat of
mixing is zero. However, nearly all aqueous systems show real behavior,
with the result that the heat of mixing is negative (exothermic) or positive
(endothermic). Because enthalpy-<:oncentration diagrams for real systems
can only be established by measuring the heat of mixing, only a few dia­
grams have hen published in the literature.
As an example, Figure 3.1 shows the temperature-<:oncentration diagram
for a magnesium sulfate-water system. Line EB is the melting-point line (or
Physical and Chemical Properties 15

80 l
·c
60
/J
IA L

I
0 j__ _E Va H~ ____!_
I
C
0 0.1 02 03 0.4 05
kg MgSO4 / kg mixture
Mass fraction w*
100
kJ/kg
0B

·100

·200
.c.
>-
.9- ·300 L
<ti
C
~(I)
J
,g ·(00 K
·o(I)
a.
(/J -500

0 0.1 0.2 0.3 0.4 05

kg MgSO4 / kg mixture
Mass fraction w*

Figure 3.1.Solubility and enthalpy-concentration diagram of an aqueous

magnesium sulfate solution.

liquidus), and the straight line EC is the solidification line (or solidus) in the
range up to the concentration of the eutectic point E. The melting-point line
of different hydrates starts at point E and continues through DAF. In the
EBC field, solid water (ice) and magnesium sulfate solutions are in equili­
brium. The isothermal ECI triangle of -3.89°C represents a three-phase
system with a magnesium sulfate solution of composition E consisting of
16 Mersmann

solid water (ice) and magnesium sulfate crystals that contain 12 molecules of
water for every molecule of MgSO 4. The DHJ triangle also represents a
three-phase system consisting of solid MgSO4·12H 2O and MgSO4•7H 2O
and a saturated magnesium sulfate solution with a mass fraction of
w* = 0.21. The EDHI and DJLA areas contain two-phase systems in
which MgSO4-12 H 2O or MgSO4-7H 2O crystals exist in saturated solutions.
More information is provided in [3.1].

4. CRYSTALLINE STRUCTURE AND SYSTEMS

Crystals are solids with a three-dimensional periodic arrangement of units
(atoms, ions, molecules) in a spatial lattice. Crystals differ from amorphous
solids by their highly organized structure resulting from the varying bond
forces. Table 4.1 shows typical properties and some examples of various
types of crystals. The crystal lattice of an ideal crystal is made up of
elementary cells whose corners or even surfaces and spatial centers have
lattice components arranged on them to form stuctures that are completely
regular in shape. The elementary cell determines a coordinate system with x,
y, and z axes and a, {3, and y angles. Crystals of different substances vary in
their elementary length a, b, and c and in the size of the angle. Figure 4.1
shows this type of elementary cell. A distinction is made among seven dif­
ferent crystal systems (illustrated in Table 4.2), depending on the spatial
periodic arrangement of the components. The 14 possible Bravais lattices
can be classified into these 7 crystal systems by their symmetry.
The external shape of a regularly formed crystal is not fully determined
by the type of lattice. Information is required on the interfacial surfaces,
which are represented by network planes with a high packing density of
elementary components. Growth conditions can characterize the overall
external impression or habit of crystals of the same substance in different
ways, even in the formation of the same peripheries. A distinction is made
among prismatic, acicular, dendritic, plateletlike, or, in the case of uniform
growth in all spatial directions, isometric habits.
If substances crystallize into different but chemically identical forms, this
behavior is called polymorphism. For example, carbon is a dimorphous
material: graphite (hexagonal) and diamond (regular). The three modifica­
tions of calcium carbonate are aragonite (orthorhombic, Pc = 2930 kg/m3),
calcite (hexagonal, rhombohedral, Pc= 2710 kg/m 3), and vaterite (hexago­
nal, Pc= 2560 kg/m 3). Because the molar mass is always M = 100 kg/kmol,
the molecule diameters are different. The modifications can be determined
by x-ray diffraction.
 ~
~
n·
Table 4.1. Crystal Types and Bonding Forces

Examples
-=5.=
(bond energy n
Crystal type Component Lattice forces Properties in kJ/mol) [
n·
Metallic lattice Atomic residue with Metallic bond Not very volatile, high electro­ Fe (400) e.
free outer electrons conductivity and thermal Na (110)
conductivity Bronze
Ionic lattice Ions Ionic bond Not very volatile, nonconductor, NaCl (750) ;·
i
r,J

(Coulomb forces) conductive in melt, usually LiF (1000)

soluble CaO (3440)
Atomic lattice Atoms Atomic bond= Not very volatile, nonconductor, Diamond (710)
valency bond insoluble, extremely hard SiC (1190)
(common electron Si, BN
pair)
Molecular lattice Molecules Van der Waals Not very volatile, nonconductor CH 4 (10)
forces (induced Ji, SiCl4
dipole), Ice (50)
permanent dipoles HF (29)
(e.g. hydrogen
bridges)

'""'
...:a
18 Mersmann

1) triclinic 2) monoclinic

(a) uiJaJ rxJ mbic

[llJ
4) tetragonal 5) hexagon.

(bl

6) rhombohedral 7) cubic
(cl

Figure 4.1. (a) Elementary cell (upper left); (b) explanation of Miller's
index; (c) different crystal systems.

Table 4.2. Crystal Systems

Crystal system Elementary length Angle of axis
Triclinic a=fb=/.c o:=/./3=/.1
Monoclinic af.b=/.c 0: = 'Y = 90° =/. /3
(Ortho)rhombic ·a=/.b=/.c 0: /3 = 'Y = 90°
=
Tetragonal a=b=/.c 0: /3 = 'Y = 90°
=
Hexagonal a=b=/.c 0: = /3 = 90°; 'Y = 120°
Trigonal rhombohedral a=b=c 0: = /3 = 'Y =/. 90°
Cubic a=b=c 0: = /3 = 'Y = 90°
Physical and Chemical Properties 19

5. POLYMORPHISM AND RACEMISM

Crystalline material can be composed of isomorphs or polymorphs.
Isomorphs crystallize in almost identical forms and are chemically similar.
Many arsenates, phosphates, selenates, and sulfates are isomorphous.
Polymorphs are chemically identical, but they are capable of crystallizing
into different crystalline forms. In Table 5.1, some inorganic and organic
polymorphs together with distinct typical properties are listed. Such
crystalline modifications can be subdivided into polymorphs, hydrates,
and solvates. Obviously, a certain substance can crystallize in different crys­
tallographic forms with different physical properties, for instance different
densities. In hydrates and solvates, the number of solvent molecules per
molecule solute can differ in forms which are more or less stable.
Sometimes, the stable modification is obtained only after a fairly long period
of phase transformation. In rapid crystallization or precipitation, at first a
metastable phase (either amorphous or crystalline) may appear, and this
intermediate is transformed into another polymorph, hydrate, or solvate.
This process of transformation can be very slow and is sometimes hindered.
The different polymorphs of a substance have different physical properties
and solubilities in a solvent at a given temperature. During crystallization or
precipitation, the supersaturation of the polymorphs differs, with the result
that the polymorph with the highest supersaturation is generated more read­
ily than the other. This polymorph that is formed first is often not the more
stable form. A general prediction seems to be difficult, but Ostwald's rules of
stages can be very useful in many cases. According to Ostwald, "If the
supersaturated state has been spontaneously removed, then, instead of a
solid phase which under given conditions is thermodynamically stable, a
less stable phase will be formed." This means that the process of solid
production is at first controlled by kinetics, which leads to a less stable
intermediate. This can happen very quickly, but the transformation process
of the unstable polymorph to the stable form can take a long time. With
respect to high supersaturation, kinetics overrides thermodynamics. It is
necessary to describe the chemical reactions and the crystallization kinetics
(e.g., nucleation and growth as a function of supersaturation) in order to
predict the polymorph that is formed first. Ostwald's law of stages has been
confirmed by many authors. For example, during precipitation of CaCO3
from strongly supersaturated aqueous solutions at 25°C, amorphous car­
bonate is formed first. This amorphous substance is converted to the more
stable modification, vaterite, which then changes to calcite.
Sometimes, the progress of the chemical reaction and the subsequent
supersaturation are controlled by mixing phenomena. The occurrence of
an intermediate may depend on fluid dynamics. This can be seen from
 ~

Table 5.1. Physical Properties of Some Inorganic Polymorphs

Crystal density Mohs Refractive
Substance Polymorphs Pc [kg/m3] hardness H index n Crystal system
C Diamond 3510 IO 2.42 Regular
Graphite 2160 1.5-2 - Hexagonal
Al(OHh Gibbsite 2530 2.5-3 1.57 Monoclinic
Bayerite 2350 - 1.58 Monoclinic
Nordstrandite 2420 3 1.58 Triclinic
AIOOH Diaspore 3440 6.5-7 l.72 Orthorhombic
Boehmite 3010 3.5-4 1.65 Orthorhombic
SiO2 Crystobalite 2320 6.5 1.48 Regular
Tridymite ~ 2300 6.5-7 ~ 1.48 Monoclinic or
Orthorhombic
Quartz 2650 7 1.54 Hexagonal
CaCO3 Aragonite 2930 3.5-4 1.68 Orthorhombic
Calcite 2710 3 1.66 Hexagonal/
rhombohedral
Vaterite 2560 3 1.65 Hexagonal
~
fD

i
to;
=
=
Physical and Chemical Properties 21

I,'

~0 mean specific power input

E 0.8
;:!!:. 1111111111 i:=
(.)
0
Vaterite 0.7 W/kg
C 0.6 0.1 W/kg
0

1Q)
0
0.4
C
0 0.2
0
0:::, ragonite:
"C 0
LU
0 900 1800 2700 3600
Time't [s]

Figure 5.1. Polymorphism of calcium carbonate with residence time,

concentrations of reactants, and mean specific power input as parameters.

Figure 5.1, in which, besides the parameters concentration and time, the
specific power input in a stirred vessel decides whether vaterite or aragonite
is formed. Another example is the precipitation of aluminum hydroxide,
which occurs in five different polymorphs and two different stoichiometries
as AlOOH or as Al(OH)J. Temperature, interfacial tension, and Zeldovich
factors are the parameters that determine which polymorph predominates;
the latter two have to be seen as specific to each polymorph and with respect
to specific adsorption (see Chapter 8) [5.1].
With respect to polymorphism, it is not only the kinetics of the chemical
reaction, nucleation, crystal growth, and phase transformation that can be
complex but also the determination of thermodynamically stable states or
phase diagrams. Very long periods of contact between a solid phase and a
liquid phase are often necessary to obtain the equilibrium state. When
equilibrium is approached, changes in phase properties become smaller
and smaller and it may be difficult to decide on the final equilibrium
state. Sometimes, slow crystallization with seeding may finally lead to
a reliable point in a phase diagram, but even this procedure can fail.
The term "dynamic phase diagram" is used to characterize this behavior.
This is a major drawback because supersaturation as the driving force for
crystallization kinetics always refers to the equilibrium state.
When crystallizing organic substances, enantiomorphs and racemates can
play an important role. Two crystals of the same substance are said to be
enantiomorphous when they are the mirror images of each other. This can
be seen in Figure 5.2 in which the L- and the n-form of amino acids are
depicted.
22 Mersmann

H~,I
Mirror plane

=~"Mi,

L.........,

+H
OH

CH,

L-Alanin D-Alanin

Figure 5.2. The L-form and the D-form of the amino acid alanine.

Enantiomorphous crystals are often, but not always, optically active.

Racemism is a property of many natural substances with asymmetric
carbon atoms. When crystallizing a racemic mixture with two optical
antipodes, two cases can be distinguished: Either the antipodes are
separated into different crystals (conglomerate) or the two antipodes are
united in the same crystal (racemate). A conglomerate is an equimolal
mixture of two pure enantiomorphs and can be separated into the two
forms. A racemate is an equimolal mixture of two enantiomers that are
homogeneously distributed throughout the crystal lattice, with the
consequence that the two forms cannot be separated. Separation is some­
times possible when seed crystals of only one enantiomer are added to a
solution that is supersaturated with respect to this component and the
crystallization of the other component can be avoided. Racemates and
conglomerates can change into a common form when heated to a certain
temperature. Above this transformation, a racemate rather than a conglom­
erate crystallizes.

6. REAL CRYSTALS (POLYCRYSTALS)

As a rule, crystals produced in the laboratory or in industry are polycrystals
that are composed of a great number of small crystals whose lattices show
Physical and Chemical Properties 23

many imperfections [6.1]. In Figures 6.1 and 6.2, polycrystals are illustrated,
and Figure 6.3 shows the formation of polycrystals: At first, a large number
of nuclei are generated.
After the growth period 't' (distance of nuclei/growth rate), the crystals
touch each other and then grow together. Because growth is hindered, the
polyhedral shape of the single crystal is abandoned and a polycrystal with
grain boundaries is obtained. It will be shown later that the growth of such
crystals and the attrition behavior depend on these grain boundaries and on
the mosaic structure of the polycrystal. X-ray investigations of 500 µm
KAI(S04h•l2H 20 polycrystals have shown that the size of a mosaic
block is approximately 200 nm. This means that the polycrystal may consist
of 10 10 mosaic blocks. The energy Yee per unit area of grain-boundary area
can be calculated from

Yee= 4 rr(t: Ve) 0[1+ In (i!rJ -In 0] (6.1)

In this equation, µ is the shear module of the polycrystal, a is the intermo­

lecular distance, and Ve the Poisson ratio. The angle 0 explained in Figure

00000 °00 ooo ooo ooo 00 0 00

g:
00000°00 <l,ooo Ooo80 oooo 00 ogo 0 oo 0
~
0000
ooo 0
0000000000000000 0 00~0
00000000000oa;>oo&,ooo o 0 ° o 0000
ooo ooo ooooi:, 00 0
0000000ooogo oooo 00 ooo 0 gog 0 00 000000000008
0
g
000000000 8000000 000008000°0°0000 00000000000
000000000 0000 0 00 oQ) 0 o oo g000oooooooo
0000000000000000 00 O 0000000000000
888goooooooooooooioooggJgggoo000000000 0 g
0 00 000000000 00 00 00 000 ..,,. 000 g0o 00
00 0 00
00 0 00000000000000o 0 .,..
o 00 ~ 0o 0oooooooo 000000 00 ggg °
° 000 00
8
oo oooooogoo
00000 oo
oooooooo
oav ooooooooooooo ooo 0 0 0 00 0 00ooooooog
\\t88&8i88gg888888888888888888&&8ggg§§§:o
00i 0c°c°88888ooggggggggggggggoooooooooo~gggg
0000000000000000000000000 oooooooooeeo:ooo
ooooooooooooooooooooooooo8gggggoooooooooo
osooooooooooooooooooooooooo 0000000000
0000 oooooogoooo ooooooooooooo8888°0 o0 oooooooo
0000000 000 0 00000000000000 000000
gi 00 0000000000ooooooooooooooggggooooo 00 00 gg
000000000000000000000000000000000000000

iifill §flli!illiiii!ll!l!lff
0§o
OOOV'V" 0 0 0 00000000000000 0
0 0°ggogooggol!!~g
00000000000000000000000808ggggggoooooogeo
0000000
000000000000000000000000000000 00000000000
8 00000000000000000000000000000000000000000000
0000000000000000000 0 00000
00000000000000000
000000000000 00
0
°
0000000000000000
oooooooooooogooooo 00 0 00000000000000000oooo
0 000000000000000000000
0000000000000 00 ~ oggooooooooooooooooooo 0 o
0000000000000 000=0 ooooOOoooooooooooooo o
0000000 0 0000 0000 gt>o 00~00000ooooooooooogooo
oooooo 0000 oao 0000 oo -.. ooo 000000000000 ooo
o 0000 00 0 000 o oooo oo 00 0000 ooo oooooooooooog
0 ooo 0000 ooo 000000000000

Figure 6.1. Polycrystals.

24 Mersmann

dislocation
Figure 6.2. Grain-boundary angle between two mosaic blocks.

~ 0 [!]
®

0 0 I!]
~

~ 0 ~
~
a) b) c)

Figure 6.3. Formation of polycrystals.

6.4 is equivalent to a line of edge dislocations. The small-angle or low-angle

tilt can be expressed by the ratio of the Burgers vector b and the distance D
or the radius of the dislocation point. The Burgers vector can be one or
several intermolecular distances in length. For a simple tilt boundary, the
relationship between the angle of tilt 0, the size L of a polycrystal, the size /0
of a mosaic block, and the mosaic spread 'f/ is given by

'f/=0
f/Li;; (6.2)

In the grain boundaries, the energy YccAo, with A 0 as the surface of all
mosaic blocks, is stored, which results in an enlarged energy of the poly­
crystals or chemical potential JJ,c with the consequence that the supersatura­
tion b..µ, = JJ,c - µ,L is reduced. Here, the potential µ,L represents the
Physical and Chemical Properties 25

•• • • •••• •••

.. . .. .. .........·:1-
•• • ••••••••

...........
••
••
•
•
••• •••• •
• ••• • • •

•• •• •• •• •• •• •• •• •• • □ ==E.0
••••••••• : I
••••••••• ;__i_
.• .• .• .• •. •. •. •. •.
•••••••••

••• • • •• ••
••• •• ••••
•••
.. .
. . . .• .•. •••
••• •• •

-
•••
. •••
b 1}

Figure 6.4. Grain boundary due to a small angle.

potential of the solute in the solution. In Figure 6.5, the effective relative
supersaturation O'eff is plotted against the crystal size L with the mosaic
spread 1J as the parameter. The smaller the crystal (attrition fragments)
and the larger the mosaic spread TJ, the more the supersaturation is reduced
and can become negative (dissolution) [6.2].

:C 0.05

J
§ 0.03

~
~ 0.01
~ 0
~ -0.01
::::,
en
Cl> Potash alum
>
:;:::; -0.03 8=25°C

55
u
-0.05 +--------~~---~~~-,....../
10 100
cr=0.05

Crystal size L [µm]

Figure 6.5. Effective relative supersaturation against the crystal size for
different mosaic spreads.
26 Mersmann

In general, three main types of lattice imperfection exist: point or zero

dimensional, line or one dimensional, and surface or two dimensional.
Surface imperfections, or mismatch boundaries, are mostly caused by ir­
regular and fast growth or by mechanical and thermal stresses. It will be
shown later that lattice deformation energy is stored in particular in attrition
fragments generated by collisions with a rotor with a circumferential velo­
city of several meters per second.
Line defects play an important role not only for crystal growth but also
for slip or shearing in crystals. The two main types are screw and edge
dislocations; see figures 6.6 and 6.7 [6.1]. The degree of deformation or
the displacement of molecules can be described by the Burgers vector,

screw dislocation

Figure 6.6. A screw dislocation and a created spiral step .

...

1 ...
0

Figure 6.7. A simple tilt boundary.

Physical and Chemical Properties 27

which indicates the direction and the magnitude of slip. With respect to
these defects and the grain boundaries, the tensile strength a of polycrystals
is much lower than it would be for a perfect crystal. This strength for defect­
free crystals is
a~ 0.1 E (6.3)

but for polycrystals, the factor is approximately 0.005-0.01 instead of 0.1.

Point defects in the lattice occur when one lattice site ·is either vacant
(vacancy) or occupied by a substitutional impurity. A small atom or mole­
cule may also be trapped in the interstice of a lattice. This is called an
interstitial impurity. Such point defects are a prerequisite for starting crystal
growth in a supersaturated environment when the crystal surface is very
smooth. Because the number, type, and extent of point, line, and surface
imperfections in real polycrystals differ from crystal to crystal, it is virtually
impossible to predict the behavior (such as growth, microattrition, and
breakage) of an individual crystal.

7. PHYSICAL PROPERTIES OF REAL CRYSTALS

It will be shown later in more detail that the kinetics of nucleation and
growth of crystals depends on the deformation of the crystal lattice,
which is the result of strain caused by mechanical processes. The degree
of deformation is a function of the shear modulusµ, Young's modulus E,
the Poisson ratio ve, and the fracture resistance r. An isotropic material has
two independent elastic constants (i.e., the shear modulus µ and Poisson
ratio ve) from which all other possible constants can be calculated.
Anisotropic crystals may have between three (in the cubic system) and 21
(in the triclinic system) elastic constants, depending on their properties of
symmetry. From their values, average quasi-isotropic constants can be cal­
culated according to Voigt, Reuss, Hill [7.1, 7.2]. For many substances the
elastic constants are tabulated (see [7.3]). Assuming that impacts on a crystal
are distributed statistically in all axis directions, the effective (isotropic)
elastic properties can be estimated from the constants according to Voigt,
Reuss, and Hill. The Poisson ratio is given by
E
Ve =--1 (7.1)
2µ
written with the shear modulus C44 ; see [7.4].
The attrition of crystals as the result of a mechanical impact depends on
the brittleness of the solid material, which can be described by the ratio of
28 Mersmann

Young's modulus E and the hardness H. Materials with a high E/H ratio
( > 180) exhibit ductile behavior (chorides, iodides, bromides, and fluorides),
whereas solids with E/H < 100 are brittle. Sulfates, phosphates, and nitrates
are often brittle. The hardness of a solid is considered to be its resistance to
local plastic deformation, which may be caused by the indentation of a
Vickers pyramid. Dividing the load applied by the area of the indentation
formed gives the Vickers hardness:

F
Hv = 1.854 did2 (7.2)

Here, Fis the force applied and d1 and d2 are the two diagonals of the plastic
indentation. The factor 1.854 results from the geometry of the indenter,
which is a pyramid having an included angle of l 36°. When the change in
geometry of the indentation during unloading is neglected, the hardness is
equivalent to the contact pressure during loading. Assuming that the contact
pressure of a plastically deformed cone and the Vickers hardness are the
same, we can use the appropriate Vickers hardness of a substance in equa­
tions presented later. However, the quasi-static hardness measured in this
way can be significantly lower than the hardness measured under impact
conditions. This is because plastic flow requires the movement of disloca­
tions in the crystal, which depends on the strain rate.
Because the required pressure necessary for plastic deformation is high, it
is assumed that ductility is restricted to a region limited by the area of
contact and that the material is elastic elsewhere. This regtion is taken to
be a spherical segment of the radius a below the circle of contact. The
plastically deformed region is assumed to be under hydrostatic pressure,
which coresponds to the contact pressure. This pressure is equivalent to
the dynamic hardness after complete loading. The average strain energy
per unit solid volume Wv depends on the hardness Hv, the shear modulus
µ,, and the radius a based on the radius r (see Fig. 7.1 [7.4]):

hydrostatic core

Figure 7.1. Indentation test (Vickers hardness) (left); definition of a and r

(right).
Physical and Chemical Properties 29

w ~3H
V
2
-v -
16µ,
(;)4
r
(7.3)

When Rittinger's law applies, this energy can be transformed into surface
energy r or fracture resistance after a crack, and the characteristic size of a
fragment in a given direction L; will depend on the strain energy per unit
volume according to

w =2-
r (7.4)
v L;K
K is an efficiency constant that is dependent on the nature of the stresses
(tension or shear or both). Up to now, it has been difficult to predict the
fracture resistance (r / K) exactly. A simple equation has been proposed by
Orowan [7.5]:

(r) = (
K l.1E nCcNA
1 ) 1 3
/
= 1.1Elo· (7.5)

In this equation, n is the number of atoms in a molecule, Cc is the molar

density of the crystal, and NA is Avogadro's constant. According to Gahn
[7.6], the fracture resistance (r / K) can be calculated from the characteristic
deformation work We, which depends on the indentation force Fe and the
hardness:

(7.6)

Fe is the critical force necessary to form cracks. In Table 7.1, data for Hv,
Fe, and We are given for 10 solid substances that were investigated by
indentation tests. When the energy We required for cracks is determined
experimentally, the fracture resistance (r / K) can be calculated from

(-r) =
K
0.1
wI/3 HS/3
e
µ,
V (7.7)

As has been shown, all mechanical material properties important for crystal
attrition and deformation can be obtained from indentation tests. Gahn
[7 .6) investigated 10 inorganic and organic substances. Indentation tests
were carried out on a total of 27 different crystallographic faces; see
Table 7.2. The results are summarized in Table 7.1, where data for the
hardness Hv, the shear modulus µ,, the ratio E/Hv, the critical work We,
the density Pc, the molar mass M, the number of elements n, the length /0 ,
and the fracture resistance (r / K) are given. By comparing the hardness as
 ~

Table 7.1. Material Properties of Brittle, Semibrittle, and Ductile Systems

Material properties

Hv µ E lo r/K, r/K,
[xl0-6 [x10- 9 [xl0- 9 w;, Pc M [x 10 10 [Eq. (7.5)] [Eq. (7.7)]
Substance N/m2] N/m 2] N/m 2] Ve E / H v [ x IO 10 J] [kg/m3] [kg/kmol] n m] [J/m2] [J/m2]

L( +)-Tartaric acid 1030 9.58 25.4 0.32 25 36 1759 150.1 16 2.1 8.95 16.8
Potash alum 754 7.96 20.3 0.27 27 7 1760 474.4 48 2.1 7.26 7.0
Citric acid 454 4.71 12.6 0.35 28 35 1542 210.1 24 2.1 4.54 8.7
Potassium sulfate 1502 17.4 44.1 0.27 29 30 2662 174.3 7 2.5 18.7 16.4
Magnesium sulfate 649 9.06 23.6 0.31 36 48 1680 246.5 27 2.1 8.37 9.1
Thiourea 137 3.08 8.09 0.33 59 1800 1405 76.12 8 2.2 3.08 6.7
Ammonium sulfate 355 8.90 23.4 0.32 66 41 1769 132.1 15 2.0 8.06 3.2
Potassium nitrate 265 7.17 18.9 0.32 71 59 2109 101.l 5 2.5 8.09 2.8
Sodium chloride 166 14.7 36.9 0.25 222 - 2163 58.44 2 2.8 17.7
Potassium chloride 91 9.44 24.l 0.28 265 - 1984 74.56 2 3.1 12.9

[
a
=
5
Physical and Chemical Properties 31

Table 7.2. Indentation Tests on Different Crystallographic Faces of

Different Materials
Formula Examined faces
Ammonium sulfate (NH4)iSO4 {001}, {010}, {110}
Citric acid COH -COOH(CH2COOH2)·H2O {011}, {101}, {110}
L( + )-Tartaric acid (CHOH 2)(COOHh {100}, {110}
Magnesium sulfate MgSO 4•7H2O {010}, {ll0}, {lll}
Potash alum KAl(SO4)z-12H2O {100}, {110}, {111}
Potassium chloride KCI {100}
Potassium nitrate KNO3 {010}, {011}, {012},
{110}
Potassium sulfate K2SO4 {010}, {021}, {110},
{111}
Sodium chloride NaCl {100}
Thiourea CS(NHh {011}, {110}, {110}

well as the crack patterns formed by indentation tests, the material behavior
with respect to attrition can be grouped into three classes:

I. Brittle crystals. Covalently bonded crystals are generally hard and

brittle. For these crystals, the following results are typically obtained
from an indentation experiment:

• Brittle crystals have a low ratio of E / H v and generally a high value

of Hv.
• The crack patterns observed on these crystals are radial as well as
lateral (Fig. 7.2).
• The crack patterns are very similar on different faces.

------- -$-
side view top view

radial cracks

lateral cracks

Figure 7.2.
-------
--- ~
Schematic of the observed radial and lateral cracks.
32 Mersmann

Of the 10 substances investigated, 5 can be attributed to this group: citric

acid, magnesium sulfate, potash alum, potassium sulfate and tartaric
acid. They have a relatively high hardness (Hv between 454 x 10 6 and
1502 x 106 Pa) and a ratio of E/Hv between 25 and 36. Typical crack
patterns on these crystals are shown in Figure 7.3 for the {111} face of
potash alum and the {101} face of citric acid.
2. Semibrittle crystals. Semibrittle crystals are neither typical ionic nor
covalent crystals. As a consequence, larger values of E / H v are generally
observed for these crystals. Three substances that can be assigned to this
class are ammonium sulfate, potassium nitrate, and thiourea. They have
a hardness H v between 137 x 106 and 355 x 106 Pa and a ratio of E / H v
between 59 and 71. When the indentation and the types of cracks
formed are considered, pronounced anisotropic behavior can be
observed for these crystals. Typical indentations and crack patterns
on a potassium nitrate crystal are shown in Figure 7.4. The indentation
diagonals are very different on the {110} and {010} faces. Because of the
anisotropic plastic flow, lines (corresponding to regions of surface
pileup) can be observed. Radial cracks are formed on these faces,
regardless of the orientation of the Vickers diagonals. The {011} and
{021} faces of potassium nitrate exhibit a different type of behavior.
Only lateral cracks form on these faces; this is understandable, as these
cracks run roughly along the cleavage plane {011} [7.7]. It is important

Figure 7.3. Indentations on the {11 l} face of potash alum (left) and the
{101} face of citric acid (right) at a load of 0.1 N.
Physical and Chemical Properties 33

lateral
crack

{110} {010}

Figure 7.4. Illustration of crack patterns and surface pileup for indenta­
tions on potassium nitrate.

to note that cleavage is most frequently observed in semibrittle materials

[7.8, 7.9]. On substances that are brittle, this tendency is generally much
less pronounced, as their (covalent) bonds are highly directional and
hinder the movement of dislocations and, consequently, plastic flow.
Although the indentations on semibrittle substances may be strongly
anisotropic, strong variations in the hardness values on different crystal­
lographic faces (by more than 20%) cannot be observed.
3. Ductile crystals. Ionic bonds are nondirectional, which is why disloca­
tions can glide easily, resulting in hardness values that are lower than
the elastic constants. The alkali halides sodium chloride and potassium
chloride have a hardness Hv of 166 x 106 and 91 x 106 Pa and a ratio of
E / H v of 222 and 265. On these crystals, no cracks can be formed by a
Vickers indentation even at loads of 30 N. A Vickers indentation test
should, however, not lead to the misleading conclusion that these sub­
stances will not form attrition fragments under impact conditions.
The ability of ionic crystals to deform plastically strongly depends on
the strain rate. In a conventional Vickers identation experiment, the
load is applied slowly in a quasi-static manner. Under impact conditions
in a crystallizer, however, the strain rate is significantly higher, and
dislocations have less time to move, resulting in a smaller ability to
deform plastically [7.10]. A dynamic measurement will therefore result
in significantly higher values of the hardness for ionic crystals (up to a
factor of 5 [7 .11 ]). For these substances, a quasi-static indentation test is
insufficient for determining the relevant mechanical material properties
that affect the attrition process.
34 Mersmann

From the above-described observation it can be expected that it will be

possible to extract the following information from indentation experiments:

• An indentation test is a good method of determining whether a sub­

stance behaves in a brittle, semibrittle, or ductile manner under quasi­
static conditions. The quantitative measure of this behavior is the ratio
of Young's modulus and hardness; the qualitative measure is the char­
acteristic crack pattern.
• Indentation fracture is a helpful experimental technique for the observa­
tion of anisotropic material behavior.

One of the important material properties is the shear modulus µ or C44 ,

which depends mainly on the interatomic distance and the kind of bonding.
In Figure 7.5, the shear modulus C44 is plotted against the interatomic
distance for covalent and ionic solids [7.12]. Because the hardness Hv in
combination with Young's modulus determines the brittleness of a solid
material, the question arises as to whether there is a relationship between
E and Hv. According to Tabor [7.13], the following equations for the hard­
ness can be derived with a 1 as the tensile strength:

For a,IE::: O.Ql

Hv
a,=-
3
(7.8a)

for 0.01::: a,IE::: 0.1

1012

IJ 1011
Cl)
::,
:5
'tJ
0
E 1010
roQl
.c
(/)
109
0.1 0.15 0.2 0.3 0.4
lnteratomic distance [nm]

Figure 7.5. Shear modulus C44 against interatomic distance.

Physical and Chemical Properties 35

Hv
a1
=~ +
3
[1 ln(0.4£)]
3a 1
(7.8b)

for a 1/ E ::: 0.Ql

Hv
u1 = Li (7.8c)

The hardness can be calculated by measuring the tensile strength. Plastic

sliding of neighboring network planes occurs at the maximum theoretical
shear stress, •max, which can be estimated accoding to Chin [7.12].
•max~ Hv ~ 0.1µ ~ 0.lC44 (7.9)
for crystals with weak ionic bonding. In Figure 7.6, the shear modulus C44 is
plotted against the Vickers hardness H v for covalent and ionic bonding.
Materials with strong ionic bonding can be described by the relationship
Hv ~ 0.01µ ~ 0.0lC44 (7.10)
for materials withµ < 2 x 10 10 N/m 2 • In Chapter 5, it will be shown that the
expression

.!!}__(r)
H~ K
3
(7.11)

which is proportional to the minimum energy necessary for the generation

of attrition fragments is important for attrition-controlled secondary
nucleation. The hardness in particular, as well as the shear modulus µ
and the fracture resistance, are important material properties for the abra-

Mohs hardness number

1d2---.,........-~1~2-r-3,_4._s....s__.,1,..a~s---110

I
O; 10
11
+---1------4
ti)
:::,
"3

~ 1010 _ _ _...., ---t----t----i

109 +---,i..-----,~----1----+...
107 108 109 1010
Vickers hardness Hv [Nim']

Figure 7.6. Shear modulus C44 against Vickers hardness.

36 Mersmann

,, plastic
10
crystals
.s
ti;
1010
Q)
0
+AgC~
C: •~~~~o"nite)

~ 109 KCI caco,

+ NaCl (Calcite)
·.;;
~ KNO+ •.NaNO.., : • BaSO4
108 3 K2 so4
C:
0
.[@ + KAl(S0 4 ) 2 12H f)
Ni(NO3 ), 6NH 3 \
.0 107
<( brittle
crystals
106
107 108 109 1010

Vickers hardness Hv [N/m2]

Figure 7.7. Resistance against attrition. (From Ref. 7.8.)

sion of solids. In Figure 7.7, the abrasion resistance as the reciprocal of the
abraded volume is plotted against the Vickers hardness for many crystalline
solids. Most of the crystals are in the semibrittle region. The figure clearly
shows that the abraded volume is not a function of the hardness alone. Even
the morphology plays a role, which can be seen from the different behavior
of calcite and aragonites as modifications of calcium carbonate.

8. SURFACE TENSION OF CRYSTALS

The interfacial tension YCL between the solid crystal phase and the sur­
rounding mother liquor is of great importance for crystallization kinetics,
especially for primary nucleation and for integration limited growth. In both
cases, the kinetics is limited by a thermodynamic balance; the free enthalpy,
which is gained proportional to the crystallized volume, must extend the free
enthalpy necessary to build the new surface. In principle, this energy per unit
surface area YCL is comparable to the surface tension of a liquid phase in
equilibrium with its vapor. Nevertheless, because of the restricted mobility
of the molecules in the solid phase, the surface is not usually in the stae of
minimum free energy. Dislocations and any strain the crystal has undergone
will contribute to the surface stress, 1:CL, so a distribution of surface stress
along the surface area A and in a collective of crystals is to be expected
([8.1]; see also Chapter 7), which may partly explain the dispersion of experi­
mentally determined interfacial tensions given in the literature. The total
energy per unit area is given by
Physical and Chemical Properties 37

8TCL
•CL= 'YCL+A 8A (8.1)

For the first term in equation (8.1), which contributes to the surface stress,
the interfacial tension, 'YCL, the solid molecules can be assumed to be fully
mobile, with every deviation from this assumption being described in the
second term. 'YCL is defined as a material-specific property; however, it is still
fixed to a single face of the crystal and the overall or global interfacial
tension of the crystal collective will also depend on the morphology and
the structure of the crystals. A material will exhibit a different global inter­
facial tension when it crystallizes in needles then it will exhibit when it
crystallizes in cubes or plates. This is because of the varying extension of
the different crystal faces, where each of the faces has a specific interfacial
tension. The question now arises as to what the physical understanding of a
global interfacial tension should be, which also depends on the geometry of
the crystals. However, it has to be kept in mind that, so far, it has not been
possible to measure the interfacial tension of a single crystal face directly
and reliably. Almost all experimental data for 'YCL result from measurements
of rates of primary nucleation, expressing, in fact, a global interfacial ten­
sion of a crystal collective. On the other hand, for the application in models
to describe the kinetics of the crystal collective in an industrial crystallizer, a

C*/Cc
f 40 ..---o~.a_ _o~.5_ _~10_·1_1~0_·2 _~10_·•-10~·10....,,
~ ;i
cJ
S2"

·~
II 10

C 5
0
"iii
C
~
Iii
~
~ 0.5
K=0.333
""" ,'
,,,,',,'
/

,,,,'
fil ,,
c / / "K = 0.414
.Q
rn
C
Q) o.1..,.__ _- , . - ~ - - - ~ - , . - - - - 1
E 0.1 0.5 1 5 10 40
i5
Concentration Ratio In (CJC*)

Figure 8.1. Dimensionless interfacial tension 'YcL against the concentration

ratio ln(Cc/C*) for different K values.
38 Mersmann

global interfacial tension will be useful. However, it can be assumed that

there are few morphological effects in interfacial tensions deduced from
nucleation rates, as the shape of nuclei may be assumed to be spheres
because of the principle of minimum free energy in the system. Nielsen
and Sohnel [8.2-8.4] plotted the surface tension YCL, gained from rates of
primary nucleation, against the solubility C* for a total of 58 systems.
There have only been a few attempts to take a theoretical approach to the
interfacial tension. According to Mersmann [8.5], the interfacial tension YCL
of ideal systems (concentrations instead of activities or activity coefficients
equal to unity) can be described by

YCLd;, = Kln(Cc) (8.2)

kT C*
Despite the fact that the above equation corresponds well with the experi­
mental data obtained by Nielsen and Sohnel (see Fig. 8.1), there is some
vagueness in the constant K that cannot be clarified because the scattering in
the experimental data is too high. The constant K = 0.414 was derived from
=
an estimate of the interfacial geometry [8.5], whereas K 0.333 is suggested
from a comparison with experimental data [8.6]. Nielsen [8.7] plotted the
edge work (i.e., the surface tension YCL multiplied by the square of a growth
unit against the logarithm of the solubility, In C*. The results determined
experimentally via nucleation were described by a straight line in a semi­
logarithmic plot with slope -0.272. In equation (8.2), the solubility C* is
based on the molar density Cc of the crystals. Most of the densities Cc of
nonhydrate crystals are in the range

20 kmol < Cc < 40 kmol

m3 m3

or 3 < In Cc< 3.69.

Assuming a crystal density of 39.5 kmol/m 3 or In 39.5 = 3.67 or
1/3.67 = 0.272, the empirical equation of Nielsen can be written

YCLd;,
kT
= 0.272 . In (Cc)
C* (8.3)

With Cc= 20kmol/m 3, the constant K would be K = 0.333 instead of

0.272. As can be seen from these considerations, the comparison with
experimental data may not definitively resolve the vagueness in the constant
K.
Another theoretical approach to the constant K can be derived from
classicial nucleation theory (see Chapter 2). It has been shown that the
diameter of a critical nucleus, L;rit• is
Physical and Chemical Properties 39

L• YCLVm
crit = 4 kTlnS (8.4)

which is valid for nondissociating molecules. The smallest critical diameter

of a "nucleus" is the size dm,s of a spherical molecule. The relationship
between this size and the size dm = (CcNA)- 113 of a cubic molecule is
given by

dm
3
= 67r dm,s
3
(8.5)

or dm,s = (6/rr) 113 dm. Combining these equations leads to

L~rit _ (~ 1/3ln(~:) _ (~ l/3ln Cc - ln C*

(8.6)
dm,s - 4 K 61 ln S - 4K 61 ln C - ln C*
In the case of a sparingly soluble substance (C* « Cc) and a very high
supersaturation S(C* « C) or C-+ Cc, the critical nucleus cannot be smal­
ler than the diameter of a spherical molecule. These considerations lead to
the result that 4K(rr/6) 113 is unity or K = 0.31, whereby it should be noted
that in this theoretical consideration, the activity coefficients were assumed
to be unity despite the high concentrations, which might be critical. The
=
value K 0.31 is between the values that were deduced above by comparing
them with the data obtained by Nielsen and Sohnel from experimental
nucleation rate measurements.
To sum up the above considerations, there are different approaches lead­
ing to different values for K, which do not deviate from each other to any
great extent. Due to the wide scatter of data, the experimental data allow no
clarification of the value of K. We, therefore, recommend using one's own
experimentally determined data whenever possible; care should be taken
when determining the data obtained and with regard to their consistency.
Experimental values may be taken from Nielsen and Sohnel when the con­
ditions in the application do not differ greatly from their experimental con­
ditions. For a predictive calculation of interfacial tensions, we recommend a
=
factor of K 0.414, which might be slightly too high but has the strongest
consistency in its derivation. It may thus be seen as the most material
independent and the most consistent over a wide range of solubilities and
densities of materials.
· The accuracy of equation (8.2) may also be improved with a more
detailed model, avoiding some strong assumptions, such as neglecting the
dissociation of the lattice ion and neglecting any adsorption of ions other
than those belonging to the crystal structure. There is experimental evidence
that the specific adsorption of potential-determining ions, such as lattice
ions or H + and OH-, diminishes the interfacial tension of oxides in electro-
40 Mersmann

lyte solutions [8.8, 8.9]. This can be applied more generally to all crystal
systems, taking into account that with the dissociation of the lattice mole­
cules, the interface is not necessarily in a stoichiometric composition.
Depending on the concentration of the dissociated ions in the mother liquor,
there is generally a surplus adsorption of one lattice ion above the other. In
addition, other substituting ions may be adsorbed directly on a surface
lattice site and thus affect the interfacial tension. This may be included in
a physical model derived from Eble and Mersmann [8.10]. The specific
adsorption of ions leads to an excess charge, which results in a Coulomb
potential that, in turn, attracts oppositely charged ions form the bulk of the
solution. The surface charge density, a, as the excess adsorption of potential
determining ions of valence, z, in the electrical double layer is
a= zF(f'A+ - f'B-) (8.7)
with Fas the Faraday constant. f' A+ and f' B- are the surface densities of the
lattice ions A+ and B-, respectively, in the adsorption layer. The maximum
loading and thus the maximum surface charge may be assumed to be equal
to the total number of lattice sites at the interface, given by the molar crystal
density Cc with the length of the elementary cell a0 orthogonal to the inter­
face:
(8.8)
With the Gibbs adsorption equation, the decrease in interfacial tension with
the specific adsorption (taking into account the nonstoichiometric case
f' A+ =I= f'B-, at constant temperature and pressure and in the case of equili­
brium throughout the two-phase system) is given by
(8.9)
According to the considerations of Gilbert and Rideal [8.11 ], the differential
of the electrochemical potential aµ,A+ can be split up into an electrostatic
part, which is necessary for making an ion approach the charged surface,
and a chemical part that is due to the sorption described here using the
Langmuir equation. Combined with equation (8.7), the decrease in the
interfacial tension due to the specific adsorption of an ion of valence z
then reads

(8.10)

Assuming that the potential-determining ions are adsorbed directly on the

surface, we may approximately reduce the electrical double layer to the
diffuse layer from the Gouy-Chapman theory. Therefore, the relationship
Physical and Chemical Properties 41

between the surface charge a, the ionic strength/, the Debye length 1/K, and
the electostatic potential % is given by [8.12, 8.13]

a= 4 FI sinh(F%) (8.11)
K 291T
The integration of equation (8.10) thus leads to the general equation for the
interfacial tension YCL,,,, which is valid for a surface having the surface

I
charge density a:

!RT { ( Umax ) 8F/ [ Ka ) 2

( ill
YCL,a = YCL - zF Umaxln Umax - lal +~ 1+ -1
]

sorption of ions approach of ions

(8.12)
or
(8.13)
There are two important parameters that determine whether the actual sur­
face charge density a, which is due to a surplus of lattice ions or a specific
adsorption of substituting ions, significantly contributes to a decrease in the
interfacial tension: the maximum surface charge density amax and the ionic
force I of the solution. As shown in Figure 8.2, where l:!.yCL is plotted
against the surface charge density a with the ionic force / as the parameter,

Surface charge density a [C/m2]

-
N""
E 0.0 0.2 0.4 0.6 0.8 1.0
3 ., 0.00
~
<l
Cl) -0.02 ,10•,
10-•,

~0- <ll'J~'
Ill
(II
~
u -0.04
Cl)
"C ,o~o
tt>
C:
0 ~v
'iii -0.06 ,o
.s
C:

i5
·u -0.08
~ calculated for:
.s
.E -0.10
a== 1C/m2

Figure 8.2. Reduction in the interfacial tension l:!.yCL with the surface
charge density. Calculated from equation (8.12), valid for amax = l C/m2 •
(From Ref. 8.10.)
42 Mersmann

a low ionic force makes the interfacial tension sensitive to the sorption of
potential determining ions, the effect being essential at high ratios of a/amax
together with high values of amax.

REFERENCES
[2.1] S. Rittner and R. Steiner, Die Schmelzkristallisation von organischen
Stoffen und ihre grosstechnische Anwendung, Chem. Ing. Techn.,
57(2): 91 (1985).
[2.2] H. W. B. Roozeboom, Die Loslichkeitskurve fiir Salzpaare, welche
sowohl Doppelsalz als Mischkrystalle bilden, speziell fiir Salmiak mit
Eisenchlorid, Z. Phys. Chem., JO: 145 (1892).
[2.3] E. Funakubo and S. Nakada, Jpn. Coal Tar., 201 (1926).
[2.4] H. Reinhold and M. Kircheisen, Eine Methode zur Untersuchung
biniirer Systeme. 4. Mitteilung: Das "Auftauschmelzdiagramm" als
Mikromethode, J. Prakr. Chem., 113: 203 (1926).
[2.5] M. Hasselblatt, Uber die lineare Kristallisationsgeschwindigkeit iso­
morpher Mischungen, Z. Phys. Chem., 83: 1 (1913).
[2.6] H. Schildknecht, Zonenschmelzen, Verlag Chemie, Weinheim, p. 27
(1964).
[2.7] W. Polaczkowa, Preparatyka organiczna, Warsaw, p. 51 (1954).
[3.1] International Critical Tables (E. W. Washburn, ed.), McGraw-Hill,
New York, Vol. IV, (1933)
[4.1] 0. Sohnel and J. Garside, Precipitation, Butterworth-Heinemann,
Oxford (1992).
[5. l] A. Eble and A. Mersmann, The kinetics of boehmite (y-AIOOH)
crystal growth, nucleation and aluminium hydroxide polymorphism,
J. Cryst. Growth (in press).
[5.2] J. Franke, Uber den Einfluss der Prozessparameter auf die
Fiillungskristallisation am Beispiel von Calciumcarbonat und
Calcium sulfat Dihydrat, Thesis, TU Miinchen, 1994.
[6.1] J. Mullin, Crystallization, Butterworth-Heinemann, Oxford (1993).
[6.2] U. Zacher, Die Kristallwachstumsdispersion in einem kontinuierli­
chen Suspensionskristallisator, Thesis, TU Miinchen (1995).
[7.1] A. Reuss, Berechnung der Fliepgrenze von Mischkristallen auf
Grund der Plastizitiitsbedingung fiir Einkristalle, Z. Angew. Math.
Mech., 9: 49 (1929).
[7.2] R. Hill, The elastic behaviour of a crystalline aggregate, Proc. Phys.
Soc. London., A65: 349 (1952).
[7.3] Landolt Btknstein, New Series, Springer-Verlag, Berlin (1966/1977).
Physical and Chemical Properties 43

[7.4] C. Gahn and A. Mersmann, Theoretical prediction and experimental

determination of attrition rates, Trans. I. Chem. E., 75A: 125 (1997).
[7.5] E. Orowan, Fracture and strength of solids, Rep. Prog. Phys., 12: 185
(1949).
[7.6] C. Gahn, Die Festigkeit von Kristallen und ihr Einfluss auf die
Kinetik in Suspensionskristallisatoren, Thesis, TU Miinchen (1997).
[7.7] C. Klein and C. S. Hurlbut, Manual on Mineralogy, 20th ed., John
Wiley & Sons, New York (1985).
[7.8] W. von Engelhardt and S. Haussiihl, Festigkeit und Harte von
Kristallen, Fortschr. Miner., 42: 5 (1965).
[7.9] B. R. Lawn and T. R. Wilshaw, Fracture of Brittle Solids, Cambridge
University Press, Cambridge (1975).
[7 .10] E. Nadgornyi, Dislocation Dynamics and Mechanical Properties of
Crystals (J. W. Christian, P. Haasen, and T. B. Massalaski, eds.),
Progress in Material Science, vol. 31, Pergamon Press, Oxford (1988).
[7.11] M. M. Chaudri, J. K. Wells, and A. Stephens, Dynamic hardness,
deformation and fracture of simple ionic crystals at very high rates of
strain, Phil. Mag. A., 43: 643 (1981).
[7.12] G. Y. Chin, Strong and hard solids, Trans. Am. Crystal/ogr., 11: 11
(1975).
[7.13] D. Tabor, Indentation hardness and its measurement: some caution­
ary comments, in Microindentation Techniques in Material Science
and Engineering (P. J. Blan and B. R. Lawn, eds.). ASTM STP 889,
ASTM, Philadelphia, pp. 129-159 (1986).
[8.1] A. W. Adamson and A. P. Gast, Physical Chemistry of Surfaces, 6th
ed., John Wiley & Sons, New York (1997).
[8.2] A. E. Nielsen and 0. Sohnel, Interfacial tensions electrolyte crystal­
aqueous solution from nucleation data, J. Cryst. Growth., 11: 233
(1971).
[8.3] 0. Sohnel, Electrolyte crystal-aqueous solution interfacial tensions
from crystallization data, J. Cryst. Growth., 57: 101 (1982).
[8.4] 0. Sohnel, Estimation of electrolyte-crystal-aqueous-solution inter­
facial tension, J. Cryst. Growth., 63: 174 (1983).
[8.5] A. Mersmann, Calculation of interfacial tensions, J. Cryst. Growth.,
102: 841 (1990).
[8.6] A. Mersmann, General prediction of statistically mean growth rates
of a crystal collective, J. Cryst. Growth., 147: 181 (1995).
[8.7] A. E. Nielsen, Electrolyte crystal growth mechanisms, J. Cryst.
Growth., 67: 289 (1984).
[8.8] R, J. Stol and P. L. de Bruyn, Thermodynamic stabilization of col­
loids, J. Colloid Interf Sci., 75(1): 185 (1980).
44 Mersmann

[8.9] S. M. Ahmed, Studies of the double layer at oxide-solution interface,

J. Phys. Chem., 73: 3546 (1969).
[8.10] A. Eble and A. Mersmann, Interaction of kinetics governing the
precipitation of Nanoparticles, in Proc. 14th Int. Symp. Industrial
Crystallization (J. Garside and M. Hounslow, eds.), IChemE (1999).
[8.11] G. A. Gilbert and E. K. Rideal, The combination of fibrous proteins
with acids, Proc. Roy. Soc. A., 182: 335 (1944).
[8.12] D. C. Grahame, The electrical double layer and the theory of electro­
capillarity, Chem. Rev., 41: 441 (1947).
[8.13] W. Stumm, H. Hohl, and F. Dalang, Interaction of metal ions with
hydrous oxide surfaces, CCACAA, 48: 491 (1976).
2
Activated Nucleation
A. MERSMANN, C. HEYER, AND A. EBLE Technische Universitiit
Miinchen, Garching, Germany

Crystals are created when nuclei are formed and then grow. The kinetic
processes of nucleation and crystal growth require supersaturation, which
can generally be obtained by a change in temperature (cooling in the case of
a positive gradient dC* / dfJ of the solubility curve or heating in the case of a
negative gradient), by removing the solvent (usually by evaporation), or by
adding a drowning-out agent or reaction partners. The system then attempts
to achieve thermodynamic equilibrium through nucleation and the growth
of nuclei. If a solution contains neither solid foreign particles nor crystals of
its own type, nuclei can be formed only by homogeneous nucleation. If for­
eign particles are present, nucleation is facilitated and the process is known
as heterogeneous nucleation. Both homogeneous and heterogeneous nuclea­
tion take place in the absence of solution-own crystals and are collectively
known as primary nucleation. This occurs when a specific supersaturation,
known as the metastable supersaturation Ll.Cmet, is obtained in the system.
However, in semicommercial and industrial crystallizers, it has often been
observed that nuclei occur even at a very low supersaturation Ll.C < 6.Cmet
when solution-own crystals are present (e.g., in the form of attrition
45
46 Mersmann, Heyer, and Eble

Homogeneous
Primary

Heterogeneous

Nucleation Contact

Shear

Secondary Fracture

Attrition

Needle

Figure 0.1. Various kinds of nucleation. (From Ref. 0.1.)

.u /
// I
I I
1

/ // /
c5 ,/
;/'
/
,,~ /
/
/
/
/

C:
0 ,./ ,/' , / Metastable zone
,,-· ,,, ., width for nucleation
~C: - secondary
~ -primary,
C: heterogeneous
0
u -primary,
homogeneous

Temperature T

Figure 0.2. Metastable supersaturation against temperature for several

types of nucleation process.

fragments or added seed crystals). Such nuclei are known as secondary

nuclei. However, it should be noted that a distinction is made between
nucleation resulting from contact, shearing action, breakage, abrasion,
and needle fraction (see Fig. 0.1). Figure 0.2 illustrates the dependence of
supersaturation on several types of nucleation process plotted against solu­
bility.
In the following sections, the three mechanisms of activated nucleation
will be discussed in more detail: homogeneous, heterogeneous, and activated
secondary nucleation. All these mechanisms have in common the fact that a
free-energy barrier must be passed in order to form clusters of a critical size,
beyond which the new phase grows spontaneously. The height of this barrier
Activated Nucleation 47

or, equivalently, the extent of penetration into the metastable zone is dif­
ferent for each process due to different physical mechanisms. Homogeneous
nucleation is treated in Sec. l, including discussions on classical theory (Sec.
I.I) and kinetic theory (Sec. 1.2), and rules for industrial application are
given in Sec. 1.3. In Secs. 2 and 3, heterogeneous and secondary nucleation
are presented, and, finally, all three mechanisms are compared in Sec. 4.

1. HOMOGENEOUS NUCLEATION
1.1. Classical Theory
The classical nucleation theory dates back to the work of Volmer and Weber
[l.l, l.2], who were the first to argue that the nucleation rate should depend
exponentially on the reversible work of the formation of a critical cluster
and was later extended by authors such as Becker and Doring [l.3], Farkas
[1.4], Zeldovich [1.5], Frenkel [1.6], and others [1.7]. In order for a new phase
to appear, an interface must be formed, which (in the absence of impurities
or suspended foreign material) occurs by small embryos in the new phase
being formed within the bulk metastable phase. These embryos are formed
due to spontaneous density or composition fluctuations, which may also
result in the spontaneous destruction of such an embryo. The creation of
nuclei can, therefore, be described by a successive addition of units A
according to the formation scheme
k,4
Ai+ A = A2; A2 +A= A3; ... ; An+ A<=> An+I
kD
(1.1)

Here, it is assumed that there is no molecular association in the metastable

solution and that the concentration of embryos is small. Under these con­
ditions, embryos can only grow or shrink as a result of single-molecule
events, which can be described by the rate constants kA and kD. The value
kA is the rate constant of addition and kD that of decay of units from a
cluster. Because addition is a random process-if supersaturation is suffi­
ciently high-more and more elementary units can join together and create
increasingly large nuclei known as clusters. The reversible work necessary to
form such a cluster is given by a balance of the free enthalpy l:!.Gv, that is
gained (being proportional to the condensed matter and, thus, to the volume
of the cluster) and the free-surface enthalpy l:!.G A needed to build the new
surface. The change in positive free-surface enthalpy l:!.GA increases with the
interfacial tension YCL between the solid crystal surface and the surrounding
solution, as well as with the surface of the nucleus. The enthalpy change is to
be added to the system and is therefore positive. On the other hand, the
change in free-volume enthalpy l:!.Gv during solid phase formation is set free
48 Mersmann, Heyer, and Eble

Figure 1.1. Free enthalpy t::,,.G against nucleus size L.

and is thus negative. The magnitude !::,,.Gv of this enthalpy is proportional to

the volume of the nucleus and increases with increasing energy ffiTlnS,
where S = a/a* or in ideal systems, S = C/C*, when the concentration C
of the elementary units changes to the lower equilibrium concentration
C* = C-t::,,.C.
According to Figure I.I, in which the free enthalpies !::,,.GA and t::,,.Gv and
the total enthalpy !::,,.G =!::,,.GA+ !::,,.Gv are plotted against the nucleus radius,
the following is obtained with nucleus surface An, nucleus volume Vn, the
degree of dissociation Ot, and the number of ions v:
(l.2)
The change in total enthalpy !::,,.G with respect to the nucleus size L passes
through a maximum value. This is because the free-volume enthalpy t::,,.Gv is
a function of the volume of a cluster and therefore
(1.3)
whereas the free-surface enthalpy t::,,.G A is related to the size of the nucleus in
the following manner:
!::,,.GA ~ An ~ L 2 (1.4)
A thermodynamically stable nucleus exists when the total enthalpy !::,,.G does
not change when elementary units are added or removed; in other words, for
the case of

(l.5)
Activated Nucleation 49

a= 0.001 0.01 0.005 1.72 2.210'

S=1+a= 1.001 1.01 1.005 2.72 2.2001 10'

10 5 ,------'---'---'----,
Ye, = 0.1 J/m2
1Q 4 '\._-- Yet. = 0.01 J/m'

'\.'\.
103 '\. '\. dm =0.7 nm
-..] 10
E
'O

2
0
~
a:
10
_,O O 8-=20DC
10·3 10·2 10-1 100 101
Supersaturation In S
Figure 1.2. Ratio Le/dm with respect to the natural logarithm of super­
saturation S for different interfacial tensions and molecule diameters.

In the case of such nuclei, the rate constant kA of addition is as great as that
of decay kn. Therefore, neither growth nor decomposition takes place. The
last four equations yield the following relationship for the critical nucleus
diameter L~rit in the case of spherical nuclei:

L*· - 4YCL (16)

cnt - (1 - ex+ vcx)iRTCc In S ·
With the Boltzmann constant k = in/NA and
1 M (1.7)
Vm=--=--
CcNA PcNA
we obtain for ex = 0 and v = I (i.e., molecules)
4YcL
Le= kTlnS Vm (l.8)

Because the free enthalpy !:::,.G for nucleus sizes L > Le decreases with the
nucleus size, the addition reaction takes place by itself due to the laws for
disturbed equilibria (i.e., the nucleus can continue growing). On the other
hand, in the range L < Le, the change in free enthalpy increases with
50 Mersmann, Heyer, and Eble

increasing nucleus size. This means that the rate constant of decay is greater
than that of growth, and the nucleus disintegrates.
The rate of primary homogeneous nucleation according to the classical
nucleation theory can be obtained by calculating the number of critical
clusters that cross the nucleation barrier described earlier and is then
given by
(1.9)
It can be seen that the rate is a product of three factors:
• ne, the number concentration of critical clusters
• k, the rate at which clusters cross the barrier
• Z, the imbalance factor
all of which will be explained in more detail.
Assuming that the cluster size distribution n; is caused by random
collisions of molecules and can be described by a Boltzmann distribution,
we obtain

n; ky'
= n0 exp ( - l),.G.) (1.10)

or for critical clusters,

(1.11)

where n0 is the number concentration of monomers in the supersaturated

solution.
The value !),.Ge in equation (1.11) is the free-nucleation enthalpy of a
critical cluster. According to Volmer and Weber [1.1], this enthalpy is
given by
(1.12)

which can be achieved by inserting Le [Eq. (1.8)] into the expression for l),.G
[Eq. (1.2)]. Equation (1.6) for the critical cluster diameter Le = J Ae/rr gives
the following expression for the distribution of critical clusters with a = 1:

(1.13)

Having determined the number of critical clusters, it is now necessary to

obtain the rate at which they cross the nucleation barrier (i.e., grow by one
monomer).
Activated Nucleation 51

For nucleation from the gas phase, the rate at which monomers impinge
successfully on the surface of a nucleus of size L is given by
w
k=Anq- (1.14)
4
Here, A is the surface area of a nucleus, n is the number concentration of
monomers, w is the mean velocity of monomers, and q the "condensation"
coefficient, which is the fraction of monomers hitting the surface that actu­
ally stick. Assuming that every monomer that hits the surface is incorpo­
rated (i.e., q = l) and that the number concentration of monomers is hardly
not reduced by the nucleation process, for an ideal gas

k=A Cn0 (l.15)

is obtained, where m1 is the mass of a monomer.

Using the same assumptions as above, the following form of the impact
coefficient k was derived by Kind and Mersmann [l.8, 1.9] for nucleation in
condensed fluids:
k = 4no
3 4/3D A
AB c (l.16)
So far, all these calculations have been made assuming an equilibrium. In
order to relate the number of critical nuclei in the equilibrium distribution- to
the number in the steady-state distribution, the Zeldovic factor is intro­
duced. It is given by

z = ✓2rr~T (l.17)

where

(l.18)

is the second derivative of the free energy with respect to cluster number i at
the top of the barrier. Using the expression of the classical nucleation theory
[Eq. (l.2)] for !).G and assuming that the nucleation process is volume con­
serving, leading to

(l.19)

as the number of elementary units of a cluster with diameter Le, for the
imbalance factor Z
52 Mersmann, Heyer, and Eble

(1.20)

is obtained. Inserting the values of nc [Eq. (1.13)], Z [Eq. (1.20)], and k [Eq.
(1.16)] into the expression for Bhom [Eq. (1.9)] finally gives the following
form of Bhom:
~ 1
Bhom -
_ (
l.5DAB CNA
)7/3
VkT CcNA
(1.21)
16 YCL 3 1 1 ]
( ) 2
x exp [ -3n(kT) CcNA (vlnS) 2

This equation provides a simple way of predicting nucleation rates, espe­

cially because only easily accessible substance properties are needed: the
diffusion coefficient DAB• the actual and equilibrium concentrations, C
and C*, or activities, a and a*, the molar density of the solid Cc and the
surface tension, YCL, which can be calculated according to Mersmann [1.10]
in the following manner:

kT = K(CC N A )2/3 ln(Cc)

YCL
C*
(1.22)

The derivation of this surface tension and especially of the factor K are
elucidated in Chapter I.
It should be noted that for the derivation of the enthalpy of a critical
nucleus according to equations (1.2) and (1.12) and thus for the rate of
homogeneous nucleation, it was assumed that the nucleus consists of only
one component and that no solvent molecules are incorporated into the
lattice; that is, the solvent is insoluble in the solid phase. This, however, is
not necessarily the case in every nucleation process and there might be some
two-component or multicomponent nucleating systems. Although it is
beyond the scope of this book to go into more detail with regard to these
special processes, we wish to draw the reader's attention to the fact that
there are some differences for multicomponent systems. The main difference
is the evaluation of the free enthalpy of formation because one has to take
into account the ratio in which the different molecules are incorporated into
the cluster and how they may affect each other. In addition, the preexpo­
nential term of the rate of homogeneous nucleation may be altered due to
changes in the diffusivities of the different types of molecules, which not
only act on each other within the solid phase but also in the solution.
Besides the nucleation of multicomponent systems, the classical nuclea­
tion theory presented is strictly speaking not valid for the case of nucleation
in the presence of temperature changes, because its derivation is based on an
 Activated Nucleation 53

isothermal process. For its practical application, this, however, is usually no

hindrance. Most processes take place under isothermal conditions, and in
the few cases in which high temperature changes occur (e.g., due to the heat
of mixing in acid pasting processes), the equations can at least be used for
the estimation of a range of possible nucleation rates by using the highest
and lowest temperatures occurring. In this way, it is possible to obtain an
idea about the nucleation rates a priori and to control the process.
It should also be noted that the classical nucleation theory is based on
two approximations:
• The treatment of small embryos as though they have bulk properties
• The need to invoke an equilibrium distribution of embryos in order to
calculate kD, the dissociation or decay coefficient in equation (1.1)
The first approximation, also known as the capillarity approximation, cir­
cumvents statistical mechanics in favor of heuristics, whereas the latter
transforms a kinetic theory into a thermodynamic one. The first-principle
treatments of energetics in nucleation, such as the idea of a physically con­
sistent cluster according to Reiss et al. [I.I l-1.19] or the application of
density functional theory to nucleation according to Oxtoby [l.20-1.27],
require extensive computer simulations, and they are, therefore, not dis­
cussed any further here. For more information on these topics, the reader
is referred to the articles by Reiss and Oxtoby. The kinetic approach for
determining nucleation rates, however, seems to be a promising, straightfor­
ward idea, and will be explained in more detail.

1.2. Kinetic Approach

The aim of nucleation theory is to predict a priori the rate at which nuclei
are formed, pass a critical threshold size, and grow spontaneously. So far,
this task has been treated merely thermodynamically. The reason for this
was that the dissociation constant KD in equation (I.I) is not known at all,
whereas the arrival rate kA can be obtained more easily. Therefore, kD was
· expressed in terms of kA by applying equilibrium conditions and using
detailed balancing. The main problem to be solved was then the calculation
of the equilibrium nuclei distribution or the energy of formation of critical
nuclei, respectively.
A purely kinetic approach to nucleation in liquids was proposed for the
first time in 1989 by Ruckenstein et al. [l.28-1.31]. They avoided the
assumption of equilibrium and calculated the rate a(n) at which single
molecules left a cluster of n molecules. This was done assuming that a
molecule that moves in a thin layer surrounding the cluster leaves this
cluster when it acquires enough energy to overcome the potential energy
54 Mersmann, Heyer, and Eble

barrier due to the cluster and its surrounding fluid. Using the mean disso­
ciation time r(n), the rate at which molecules abandon the cluster can be
written
Ns(n)
ot.=-=- (1.23)
r(n)
where Ns is the number of molecules in the surface layer (i.e., within a
distance R < r < R + dm of the center of the cluster). Here, R is the radius
of the cluster and dm is the diameter of a molecule. If r > R + dm, the
molecule dissociates from the cluster.
If the cluster is a uniform spherical solid, the number of molecules within
it is given by
(1.24)
which is equivalent to equation (1.19). The number of surface molecules is
then, according to Ruckenstein and Nowakowski [1.30],

(1.25)

Here, it should be noted that Ns is proportional to the surface of the cluster

only for large particles.
i is the mean dissociation time averaged over all initial positions of a
molecule within the surface layer. It is, therefore, a function of the mean
dissociation rate of one molecule initially at a certain point~ in the surface
layer integrated over the initial distribution function of surface molecules.
Because the derivation of this expression would be beyond the scope of this
textbook, the interested reader is referred to an article by Ruckenstein and
Nowasowski [1.30] and we will discuss only the results of this kinetic nuclea­
tion theory. Ruckenstein and Nowasowski showed that the rate a is only a
function of the intermolecular potential and the size of the cluster. Using a
supersaturation-dependent rate of condensation, {J, they obtained the
following equation for the nucleation rate:

J~
g {J(g) - a(g)
= Jo {J(g) + a(g) dg
{J(l)n0
w(g) (1.26)
Bhom,kin =2 exp[-2w(g)] dg'

where {J(l) is the rate of single-molecule coagulation. Equation (1.26) can be

evaluated by the method of steepest descent, because the function w(g)
exhibits a fairly sharp minimum at the critical cluster size g*. The nucleation
rate, Bhom,kin, can then be expressed by
Activated Nucleation 55

Bhom,kin = (0.5)tJ(l)no (w
"(: *))0.5 exp[2w(g*)] (1.27)

which looks similar in structure to the result of the classical nucleation

theory; see equation (1.21). However, this approach has the advantage
that it does not rely on the assumption of an equilibrium or on the applica­
tion of continuum thermodynamics to very small objects. Although there
are still some disadvantages, such as the following:
• Numerical calculations are needed in order to estimate a and therefore
w(g)
• Accurate knowledge of the intermolecular potential is essential
the work of Ruckenstein et al. provides an important step toward the devel­
opment of a kinetic theory of homogeneous nucleation. There is now a need
for calculations using realistic intermolecular potential and for the evalua­
tion of those results, not only in comparison with classical theory but also
using experimental data.

1.3. Industrial Application

So far, the theoretical background of homogeneous nucleation has been
discussed. In industrial crystallizers, however, homogeneous nucleation is
usually not desired at all, and for the production of large crystals in parti­
cular, it has to be avoided. Only for the crystallization of very fine or even
nano-sized materials may this mechanism be useful. In each case, it is essen­
tial to know the supersaturation 11.Cmet,hom that must be obtained in order to
produce a certain number of homogeneous nuclei. This supersaturation
11.Cmet,hom is known as the metastable zone width in homogeneous nuclea­
tion. Although homogeneous nuclei are not formed in the Ostwald-Miers
range of0 < fi.C < 11.Cmet,hom, crystals can grow at ti.C > 0. The curve C* +
11.Cmet,hom as a function of temperature iJ is known as the supersolubility
curve and depends on substance values and generally on the concentration
C. Relationship (1.21) can be used to calculate the metastable supersatura~
tion 11.Cmet,hom for given nucleation rates. Combining equations (1.21) and
(1.22) leads to

(1.28)
56 Mersmann, Heyer, and Eble

or, in general,

(1.29)

The value C* /Cc now represents dimensionless solubility and the expression
Bhom = Bhom/ DAB(NACc) 513 is the dimensionless rate of primary, homo­
geneous nucleation. If a certain dimensionless rate Bhom is given, the dimen­
sionless metastable supersaturation (ll.Cmet)/Cc of homogeneous primary
nucleation can be calculated for a system with C* /Cc:

ll.Cmet (C* Bhom ) (1.30)

~ = gl Cc; v; D AB(~~NA)513
horn

Figure 1.3 illustrates this dimensionless metastable supersaturation ll.Cmet!

Cc against the dimensionless solubility C* /Cc for various rates of homo­
geneous primary nucleation, which are valid for K = 0.414. The curves are
valid for DAB = 1 x 10-9 m 2 /s and Cc = 13.3 kmol/m3 • These are mean

- ' ...~ ...~ ...~ ~"'

<.)
(.) .._1:1 .._<S .._cs .._c;!;
10° 6
(.)
<'.I
// / / / /
C:
0
:;:.
/ /S
~
.._1:1

.
-
~
::::, .._1:1
10--
<ti
~
Q)
/ / ~

a.
::::, / / .._1:1

en v= ;/ / / / / •
o// .._1:1
(/J
(/J 10"'
Q)
C: / / / / ~
.._1:1
0
"ii)
C:
/ / / / //
E
Q)
10.. /
i5 10.. 10.. 10"' 10-- 10°
Dimensionless Solubility C*/C 0

Figure 1.3. Dimensionless metastable supersaturation against dimension­

less solubility for homogeneous nucleation for v = 1 (molecules) and v = 2.
Activated Nucleation 57

values for Cc and DAB• Note the strong influence of the stoichiometric
coefficient v for sparingly soluble systems [1.32].
It is somewhat surprising that all curves pass through a maximum for a
dimensionless solubility of approximately C* /Cc:::::: 0.16. The reason for this
may be that equation (1.22) is a simplified version of the general equation
(see Ref. 1.10):

(1.31)

Unfortunately, neither the activity coefficient Y; of the dissolved substance

nor the integration constant riL is known [1.10]. It is assumed that both
riL and the term 0.414kT(CcNA)213 ln (yflyf} are small compared to YCL·
The values yf and jif are the activity coefficients (determined by means
of the concentration) of the dissolved substance i in the crystalline or liquid
phase. The greater the solubility C* of a substance in a solvent, the
greater the activity coefficient depends on the concentration and the less
accurately the interfacial tension YCL can be calculated according to the
simplified equation (1.22). The metastable zone width l:!t.Cmet!Cc of homo­
geneous primary nucleation depends on the nucleation rate and the di­
mensionless solubility C* /Cc, The following simplified equations can be
given (compare Fig. 1.3) for a dimensionless nucleation rate of 10-26
(valid for v = 1):

For C* /Cc< 10- 5

l:!t.Cmet:::::: 0.02(C*)l/8 (1.32)

Cc Cc

for 10-5 < C* I Cc < 1

l:!t.Cmet:::::: O.l (C*) 114 (1.33)

Cc Cc

However, these relationships are valid only when the solution is abso­
lutely free of foreign particles, which is never the case in industrial crystal­
lizers. Moreover, heterogeneous nucleation of all apparatus and machine
parts wetted by the solution must also be avoided, which is very difficult
and, again, not the case in practice. Besides, one also has to keep in mind
that other effects such as mixing might have a huge influence on nucleation
rates in industrial crystallizers. So far, mixing has not been considered at all,
58 Mersmann, Heyer, and Eble

because, theoretically, it should not influence the physical mechanism of

nucleation. However, the buildup of supersaturation can be influenced con­
siderably by the mixing process; hence, the nucleation rate changes and,
thus, the emerging particle size distribution may be altered significantly.
For the application of the equations of homogeneous nucleation rates
deduced, it is therefore absolutely necessary to know the correct local super­
saturation that can only be achieved if the local composition and therefore
mixing are well understood, which is, unfortunately, not the case for most
types of apparatus.
The last descriptions already suggest why it is so difficult to verify the
theory of nucleation and why experimentally determined nucleation rates
differ so strongly from theoretically predicted ones. In addition to the prob­
lems stated, it must be borne in mind that in order to determine nucleation
rates experimentally, a number of particles are needed that are produced in a
specific volume over a certain amount of time. As long as it is not possible to
count particles in the nanometer or even smaller size range in-line, there will
always be certain assumptions to get to the desired rate: The time is not
known exactly, the actual supersaturation usually cannot be monitored in­
line and particle sizes have to be measured off-line and recalculated to
particle numbers. Therefore, it is not surprising that all nucleation rates
measured so far are much lower than the theoretically predicted ones.
With improving measurement devices and standard sampling methods,
this should be strongly improved and it is expected that measured values
approach theoretically predicted ones in the near future.

2. HETEROGENEOUS NUCLEATION
2.1. Theoretical Approach
Up to now, it has been assumed that collisions between units in a solution
lead to clusters of varying size and that nuclei result from the fact that
clusters above the critical nucleus size L~rit can continue growing into crys­
tals, assuming that the original solution is entirely pure (i.e., free of solid
particles). In reality, it is impossible to remove completely all solid matter
from a solvent or solution. In bidistilled water, many Si02 particles in the
size range of only a few nanometers can be found. Schubert [2. I] came to the
conclusion that the volumetric surface of these foreign particles is approxi­
mately aror ~ 2.5 x 103 m2 /m 3; but how do such small foreign particles
(such as sand, rust, etc.) affect nucleation? This is illustrated in Figure
2.1, which shows a foreign particle in a supersaturated solution.
According to Blander and Katz [2.2] and also Schubert [2.1], the rate of
heterogeneous nucleation is proportional to the volumetric surface aror of
Activated Nucleation 59

Figure 2.1. Nucleation on a foreign particle for different wetting angles.

foreign particles and not to the number of those particles. The model pro­
posed by Schubert [2.1] describes the rate of heterogeneous nucleation Bhet
as a product of the volumetric surface aror of foreign particles and the rate
Bhet,surf of heterogeneous nucleation based on the surface of foreign particles
according to equation (2.1):
(2.1)
The rate Bhet,surf can be described as being equivalent to the rate of homo­
geneous nucleation by
(2.2)
Again, the factor Z takes into account the difference between the equili­
brium and stationary state, having the same definition as in homogeneous
nucleation [see Eqs. (1.17) and (1.18)]. The number of critical nuclei, nc, is
now based on the surface of foreign particles and not on the volume, as in
homogeneous nucleation, and the rate at which clusters cross the barrier, k,
takes into account not only impingements from the volume but also those
from molecules adsorbed on the surface of the foreign particle. Because the
two factors in equation (2.2), k and nc, are different than those in classical
homogeneous nucleation theory, they will be discussed in more detail.
The number nc of critical clusters depends on the number of molecules
CadNA adsorbed on the foreign surface and the nucleation energy .6.Gc,het of
heterogeneous nucleation according to a Boltzmann distribution:

(2.3)

Assuming now that the thickness of the adsorption layer does not exceed the
diameter dm of a molecule, which is valid for precipitation because of very
low concentrations in the solution, the surface-based concentration Cad can
be expressed in terms of the volume-related concentration Cad,v by
(2.4)
60 Mersmann, Heyer, and Eble

Taking into consideration the fact that the adsorption isotherm, which
describes the relation between the concentration Cad,v of units adsorbed
on the foreign surface and the concentration C of these units in the bulk
of the solution, can be described by a linear relationship due to the low
concentrations, expression
(2.5)

is obtained, with the adsorption constant Head· The surface-based concen­

tration Cad can then be expressed in terms of the bulk concentration C and
the adsorption constant Head:
(2.6)

The free-nucleation enthalpy of heterogeneous nucleation, ~Gc,het> is smal­

ler than the nucleation enthalpy needed for homogeneous nucleation and is
given by

~Gc,het =f ~Gc,hom =f 3 (16) rr kT

l )2kT
(YCL)3 Vm2(vlnSa (2.7)

Here,! is a geometric correction factor, which is the ratio of the volume of

the truncated nucleus to that of a sphere with the same radius. Thus, it
depends on the wetting or contact angle 0 as follows, assuming a flat foreign
surface:

f = (2 + cos0)(1 - cos0)2
(2.8)
4
If the foreign surface is not flat, as in the case of nano-sized foreign particles,
the size of the foreign material has to be taken into account too, because it
influences the ratio of the truncated nucleus volume to that of a sphere of
equal size.
In Figure 2.2, the factor f is plotted against the angle 0 for the case of
a smooth, flat foreign surface. This angle ranges from 0° to 180°, depend­
ing on the "wetting" of the foreign particle by the units. A contact angle
of 180° (point contact) corresponds to nonwettability or homogeneous
nucleation with f = 1. When the angle 0 is between 0° and 180°, the
nucleation work is reduced by the "wetting" surface of the foreign
particle. If 0 ➔ 0°, the foreign particle is completely wetted and both the
nucleation work and the formation of heterogeneous nuclei tend toward
zero. This shows that the better the wetting, the smaller the volume of the
truncated nucleus and the smaller the free-energy cost associated with its
formation.
Activated Nucleation 61

0 0
0 90° 180°
Contact angle 8
Figure 2.2. Factor f as a function of the contact angle.

A.=21t r.2(1-cos 0)

U0 =21t r0 sin 0
Figure 2.3. Schematic diagram of a nucleus on a solid surface.

The impact coefficient k in heterogeneous nucleation has to account for

impingements from the bulk of the solution, which are proportional to the
surface of the nucleus, A, as well as for impingements of units already
adsorbed on the foreign surface, which increase with the circumference
Usurf of the nucleus; see Figure 2.3. The total number of impacts is

(2.9)
The impact rate of units arriving from the volume can be described by a
model already presented for units hitting a homogeneous nucleus:
(2.10)

This rate is proportional to the diffusivity DAB of units in the volume. On

the other hand, surface diffusion is the decisive transport mechanism for
units that arrive from the surface of the foreign particle, and the impact rate
62 Mersmann, Heyer, and Eble

,isurf can be derived from the laws of statistical thermodynamics for a two­
dimensional problem [2.l]:

(2.11)

In this equation, Dsurf is the surface-diffusion coefficient of units moving

on the foreign surface. According to Suzuki [2.3] and Schubert [2.1], it is
given by

1 ✓11:kTNA (2.12)
Dsurr = 4dm Cad N A 2M-

Although this equation predicts physically impossible, nonfinite values of

the surface diffusion for Cad ~ 0, it gives plausible results for common
values of Cad•
Next, expressions are needed for the surface Ac and the circumference Uc
of a critical nucleus with radius re and the contact angle 0. The different
geometries are depicted in Figure 2.3. The surface Ac of a critical nucleus
can be calculated to be

Ac= 211:~(1 - cos0) = 211:(PcVR

2MifL S
n a
)\1 -
cos0) (2.13)

and the circumference Uc of such a nucleus can be expressed by

. ( 2MyCL ) .
Uc= 211:rcSlil 0 = 217: PcVRTlnSa sm 0 (2.14)

Combining equations (2.1)-(2.3), (2.6), and (2.9)-(2.14) leads to the follow­

ing expression for the rate of heterogeneous nucleation:

Bhet = 2~ llrordm Head(CNA)713 /lj.vm

x esin~DSurfHeadd,;;'2(CNA) 116 +311:DAB(l-cos0)) (2.15)

x exp[-1G)1r~~ ~]

The preexponential term can be divided into a contribution of impacts

coming from the volume and another contribution that describes the
impacts from the surface of the foreign particle on the circumference of
the nucleus. The higher the concentration C, the more dominant is the
volume contribution in comparison to that coming from the surface. This
Activated Nucleation 63

is also the case if the adsorption constant Head or the surface diffusion
coefficient Dsurf is small.
The rate of heterogeneous nucleation depends on the same parameters as
the rate of homogeneous nucleation (i.e., the supersaturation Sa, the con­
centration C, the surface tension YCL, the diffusivity DAB, the equilibrium
concentration C*, and the molar density Cc) and also on the following,
additional parameters:
• The volumetric surface aror of foreign particles present in the solution
• The adsorption constant Head
• The contact angle 0 or the factor f, which is a function of the contact
angle
• The surface diffusion coefficient Dsurr
We are not yet able to predict the contact angle 0 and the adsorption con­
stant Head, but some experiments have shown that the contact angle in
aqueous systems is often in the range 40° < 0 < 53° or 0.038 <f < 0.1,
respectively. Schubert [2.1] carried out precipitation experiments with the
reactants Ba(OH)i and H 2SO4 in the presence of SiO2, Al 2O3, and TiO 2 in
the nanometer range. The particles were added in order to investigate their
influence on the rate of heterogeneous nucleation. The volumetric surface
aror of the foreign particles was varied between 5 x 103 and 2.5 x 105 m2 /m 3 •
In all cases, the rate of heterogeneous nucleation was proportional to the
surface aror of foreign particles. When aror is smaller than 104 m2 /m3 , it is
necessary to take into account the basic level of Si02 f.articles that is always
present, even in ultrapure water (aror ~ 2500 m2 /m ). It is important to
mention that it is not the number of foreign particles but their volumetric
surface aror that is the decisive parameter for heterogeneous nucleation.

2.2. Application
In Sec. 2.1, the theoretical background of heterogeneous nucleation was
discussed. Comparison with experimental results is a good tool not only
for evaluating this theory but also for providing a set of rules on how to
control nucleation and, therefore, the final product quality.
When carrying out nucleation experiments, it is expedient to plot the
nucleation rate against 1/(ln Saf In Figure 2.4, the rate of heterogeneous
nucleation, Bhet> is plotted versus 1/(ln Sa)2 for BaSO4 and SiO2 particles
(Fig. 2.4a), TiO2 particles (Fig. 2.4b), and AI2 O 3 particles (Fig. 2.4c). In
all cases, the volumetric surface of the added foreign particles was
aror = 1.5 x 105 m2 /m 3 • Additional experiments were carried out without
the addition of foreign particles. The figures show that the nucleation rate
increases with the surface aror for a given supersaturation. When the critical
64 Mersmann, Heyer, and Eble

~ [1/(m's)]
&C:
10"
0
~
Q)
0
::::, 10"
z a)
0.02 0.04 0.06 0.08
1/(ln S.)'
Supersaturation S,
10" 1000
~~~~-~-~-~
200 100 50

cc [1/(m's))

*
0:::
C:
.Q
10"

iii
Q)
0::::,
10"
z
1012 1--~~--~~----- b)
0.02 0.04 0.06 0.08
1/(ln S.)'
Supersaturation S"-
1000 200 100 ::,0
10" , - - ~ ~ ~ - ~ - ~ - - ,
cc [1/(m's)]

iC: 1015
0
~
Q)
0::::,
10"
z
1012 +------~--------- c)
0.02 0.04 0.06 0.08
1/(ln s.)'

Figure 2.4. Experimentally determined rate of heterogeneous nucleation

for BaS04 and Si02 (a), Ti0 2 (b), and Al 20 3 (c). o = without the addition
of foreign particles.

supersaturation Sa,crit is exceeded, the rate of homogeneous nucleation is

higher than the rate of the heterogeneous nucleation mechanism. This cri­
tical supersaturation depends inter alia on the volumetric surface aror of
foreign particles. In Table 2.1, results of the parameter f, 0, and Sa,crit are
given for different substances of the foreign particles. The adsorption con­
stant of BaS04 on Si02 was determined to be Head= 1.1 X 10-s at 25°C.
This constant decreases with increasing temperature, as can be seen in
Figure 2.5, in which Head is plotted against the temperature.
Activated Nucleation 65

1.2-10·8 -------------~
-0
co'°
:::c 1·10"8
0

C:
0

e-0
~

8•10·9 0
1/)

:lt
-
0
C:

~C:
0 0
(.)
0
S
0 = 138
10 40 70 100
Temperature 3 [ C] 0

Figure 2.5. Experimentally determined adsorption constant of BaSO4 on

Si02 as a function of temperature.

Table 2.1. Experimentally Determined Values of the

Parameters f, 0, and Sa,crit
Solid substance f 0 Sa,crit

0.102 52.83° 322

0.034 38.85° 263
0.041 40.88° 330

It has been shown that the rate of heterogeneous nucleation depends on

the following heterogenity parameters:
• Volumetric surface aror of foreign particles
• Contact angle 0 or the factor f
• Adsorption constant Head
In Figures 2.6, 2.7, and 2.8 the supersaturation tl.C/Cc is plotted against
the dimensionless solubility for a certain set of parameters: Whereas in
Figure 2.6, the volumetric surface area is varied, Figure 2. 7 shows plots
for different adsorption constants, and in Figure 2.8, one may see a varia­
tion of the f factor (i.e., a change in contact angle). All figures are valid for
the stoichiometric factor v = 1.
66 Mersmann, Heyer, and Eble

Dimensionless Solubility C*/Cc

Figure 2.6. Metastable zone width of heterogeneous nucleation, valid for

Bhet = 10 12 m- 3s- 1, / = 0.1, Head= 10-9 , llfor = 103 , 105 , and 107 m2 /m3.

(.) " ...~ ...~ ..,<S ..,<S G ~ ...~ (/

u<l /" /"
10'
/ /
C: / / / / / ...~
0
/ / / /
/ .
~

- /
~
:::,
ro
10-2 / / .._<::I

~
Q) .._<S
0..
:::,
en / /
Cl)
Cl)
10.. / / / ►
.._<::I
/ / / /
-
Q)
-6
"E / / / / /
~

0 ad ~
·0
C: / / / / / / / /
Q)
E 10•
i5 10.. 10.. 10.. 10.. 10°
Dimensionless Solubility C*/Cc

Figure 2.7. Metastable zone width of heterogeneous nucleation, valid for

Bhet = l0 12 m- 3s- 1,f =0.1,aror = 105 m2 /m 3,Head = 10-6 , 10-9 ,and 10- 12 .

In the range of high solubility (C* / Cc > 0.01 ), the metastable zone width
for the rate Bhet = 10 12 m- 3s- 1 is scarcely affected by the parameter aror and
Head· The smaller the factor f or the contact angle 0, the more the
metastable zone width is reduced. With sparingly soluble substances
Activated Nucleation 67

(.)
(J
....~ ,s .._,s ....~ cs ~ ....~ ~"
u<I 10° "' "' "' 6

// // / ~

/
C .._c::i
0

-
~::::, '
.._c::i
( ti
104
~ ~
Q) ....'s)
C.
::::,
en ►
fl)
10.. .._c::i
fl)
Q)
"E
0
/ ~
.._c::i
"in
C: /
Q)
E 10.. /
i5 10.. 10.. 10.. 10.. 10°

Dimensionless Solubility C*/Cc

Figure 2.8. Metastable zone width of heterogeneous nucleation, valid for
Bhet = 10 12 m- 3s- 1, Oror = 105 m2 /m 3, Head = 10-9 , J = 0.5, 0.1, and 0.0 l.

(C* /Cc< 10-5), the parameters aror, Head, and 0 have a strong influence on
!),.Cmet,het!Cc. For the nucleation rate Bhet = 10 12 m- 3s- 1, the metastable
zone width decrease with increasing values of aror and Head but decreasing
factors f or contact angle 0. Because these parameters are not known for a
real solution, the metastable zone width is unpredictable for substances with
a very low solubility. On the other hand, the equations may be useful for the
seeding of solutions in order to obtain controlled nucleation. This can be
done by the addition of foreign particles which can be characterized by the
parameters aror, Head, and/. When these added foreign particles are domi­
nant with respect to nucleation in comparison to particles that are already in
the solution, the nucleation process can be controlled. In combination with
controlled growth, it would be possible to produce crystals with a desired
specification.

3. SURFACE NUCLEATION ON CRYSTALS

In seeded batch crystallizers and in continuously operated crystallizers, a
great number of crystals with the volumetric surface ac[m2 crystal surface/
m3 crystals] are present. The surface of such crystals is very smooth in
molecular terms if the supersaturation is low (a < 0.01) and the crystals
are small in size in the absence of particle collisions. With increasing
68 Mersmann, Heyer, and Eble

supersaturation the probability of surface nucleation rises. There is some

evidence in the literature that secondary nuclei that are not attrition frag­
ments are formed either as preordered species or as clusters in the immediate
solution vicinity of the crystal surface or on the crystal surface by dendritic
growth and dendritic coarsening [3.1-3.8]. These mechanisms can lead to the
formation of nuclei and/or the detachment of small dendrites from the
crystal surface without any fluid dynamics or mechanical shear [3.1-3.4].
Surface nuclei are formed when the relative supersaturation Smet,s is
exceeded. According to Nielsen [3.9], this supersaturation depends on the
edge energy, Ye, and the maximum difference in the free energy, ~Gmax,s:

y;a2
V In Smet,s = T( kT~G
max,s
(3.1)

With the approximation a~ dm ~ (CcNA)- 113 and the interfacial energy

YcL ~ Ye! dm, the following equation is obtained:

~Gmax s =
---·- Y~Ld!
T(-----',,2...:.=.-"-'--- (3.2)
kT (kT) v In Smet,s
Again, the expression

YCLd! = Kln(Cc) (3.3)

kT C*

is introduced for the interfacial tension YCL· Combining the last two equa­
tions gives

~Gmax s [Kln(Cc/C*)] 2
--~· ~ 1 f - - - - - - (3.4)
kT vlnSmet,s

The rate of surface nucleation, Bs, in nuclei per square meter crystal surface
and per second depends on the diffusivity, DAB, and the nucleation energy,
~Gmax,s, according to [3.9, 3.10] (compare with page 73)

B =~
D exp ( - ~Gmax,s) (3.5)
s d! kT

or

DAB
B ~--exp [Kln(CcfC*)J2)
( -T (------ (3.6)
s d! V In Smet,s

In general, the following relationship is obtained: