Retrospective Theses and Dissertations
1982
Effects of NaCl and MgCl2 on the vacuum
crystallization of KCl
Salaheddin Abdallah Sayed
Iowa State University
Follow this and additional works at: http://lib.dr.iastate.edu/rtd
Part of the Chemical Engineering Commons
Recommended Citation
Sayed, Salaheddin Abdallah, "Effects of NaCl and MgCl2 on the vacuum crystallization of KCl " (1982). Retrospective Theses and
Dissertations. Paper 7479.
This Dissertation is brought to you for free and open access by Digital Repository @ Iowa State University. It has been accepted for inclusion in
Retrospective Theses and Dissertations by an authorized administrator of Digital Repository @ Iowa State University. For more information, please
contact [email protected].
INFORMATION TO USERS
This was produced from a copy of a document sent to us for microfilming. While the
most advanced technological means to photograph and reproduce this document
have been used, the quality is heavily dependent upon the quality of the material
submitted.
The following explanation of techniques is provided to help you understand
markings or notations which may appear on this reproduction.
1.The sign or "target" for pages apparently lacking from the document
photographed is "Missing Page(s)". If it was possible to obtain the missing
page(s) or section, they are spliced into the film along with adjacent pages.
This may have necessitated cutting through an image and duplicating
adjacent pages to assure you of complete continuity.
2. When an image on the film is obliterated with a round black mark it is an
indication that the film inspector noticed either blurred copy because of
movement during exposure, or duplicate copy. Unless we meant to delete
copyrighted materials that should not have been filmed, you will find a good
image of the page in the adjacent frame. If copyrighted materials were
deleted you will find a target note listing the pages in the adjacent frame.
3. When a map, drawing cr chart, etc., is part of the material being photo­
graphed the photographer has followed a definite method ir. "sectioning"
the material. It is customary to begin flrning zt the upper left hand corner of
a large sheet and to continue from left to right in equal sections with smai:
••
» i
y y
L # V * «11 *^
Q^CXH I —
:I J I' I I ^
L / C lV V T
LI I C
first row and continuing on until complete.
4. For any illustrations that cannot be reproduced satisfactorily by xerography,
photographic prints can be purchased at additional cost ana ripped into your
xerographic copy. Requests can be made to cur Dissertations Customer
Servîtes Department.
5. Some pages in any document may have indistinct print. In all cases we have
filmed the best available copy.
University
Microfîims
Internationa!
300 N. /ELs KU . 4XN ARKOH.,'/! jRiOG
8221224
Sayed, Salaheddin Abdallah
EFFECTS OF SODIUM-CHLORIDE AND MAGNESIUM-CHLORIDE ON THE
VACUUM CRYSTALLIZATION OF POTASSIUM-CHLORIDE
Iowa Stale University
PH.D. 1982
University
Microfilms
Int©rnati0nâ!
300 N. Zeeb Road, Arm Arbcr. MI 48106
Copyright 1982
by
Sayed, Salaiieddin Abdallali
All Rights Reserved
PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy.
Problems encountered with this document have been identified here with a check mark V
.
1.
Glossy photographs or pages
2.
Colored illustrations, paper or print
3.
Photographs with dark background ^
4.
Illustrations are poor copy
5.
Pages with black marks, not original copy
6.
Print shows through as there is text on both sides of page
7.
Indistinct, broken or small print on several pages ^
8.
Print exceeds margin requirements
9.
Tightly bound copy with print lost in spine
10.
Computer printout pages with indistinct print
11.
Page(s)
author.
lacking when material received, and not available from school or
12.
Page(s)
seem to be missing in numbering only as text follows.
13.
Two pages numbered
14.
Curling and wrinkled pages
15.
Other
. Text follows.
University
Microfilms
International
Effects of NaCl and MgCl2 on
the vacuum crystallizacion of KCl
by
Salaheddin Abdallah Sayed
A Dissertation Submitted to the
Graduate Faculty in Partial Fulfillment of the
Requirements for the Degree of
DOCTOR OF PHILOSOPHY
Major:
Chemical Engineering
Approved:
Signature was redacted for privacy.
In Charge of Major work
Signature was redacted for privacy.
For the Major Department
Signature was redacted for privacy.
For the Gradcate
Le College
uoiiê^e
Iowa State University
Ames, Iowa
1982
Copyright^ Salaheddin Abdallah Sayed, 1982.
All rights reserved.
ii
TABLE OF CONTENTS
Page
NOMENCLATURE
DEDICATION
vi
viii
INTRODUCTION
1
LITERATURE REVIEW
3
Recovery of KCl from Brines
3
Nucleation
4
Primary nucleation
Secondary nucleation
Crystal Growth
Two-dimensional nucleation models
Continuous step model
6
8
11
13
14
Crystal Population Balance
16
Effect of Impurities
18
Continuous Evaporative Crystallizers
21
Crystallization of KCl
23
Structure of Electrolyte Solutions
25
Water structure
Ion-water interactions
Ion-ion interactions
EXPERIMENTAL
Equipment
Crystallizer
Auxiliary equipment
Procedure
Feed preparation
Operation of the crystallization system
25
26
30
37
37
37
40
44
44
44
iii
Page
Sampling and filtration
Separation of crystals
RESULTS
45
46
47
Treatment of Data
47
Determination of Growth and Nucleation Rates
47
Chemical Analysis
63
DISCUSSION
68
CONCLUSIONS
75
RECOMMENDATIONS
77
BIBLIOGRAPHY
78
ACKNOVJLEDGMENTS
86
APPENDIX A
87
APPENDIX B
88
APPENDIX C
89
APPENDIX D
90
Selected Physical Properties
90
iv
LIST OF TABLES
Page
Table 1.
Brine composition in percentage by weight
Table 2.
Chemical composition of KCl crystals and feed
solutions
3
66
Table Dl. Physical properties of the salt t 25°C
90
Table D2. Physical properties of the ions at 25°C
90
V
LIST OF FIGURES
Page
Figure
1.
Flowsheet of the recovery of KCl from the Dead Sea
(from Epstein (38))
5
Figure
2.
Schematic diagram of the crystallizer vessel
39
Figure
3.
Equipment and flow diagram for crystallization
system
42
Typical plot of crystal size distribution for KCl
crystallized from solution containing KCl only
51
Figure
4.
Figure
5.
Photomicrograph of KCl crystals
53
Figure
6.
Photomicrograph of KCl crystal
53
Figure
7.
Photomicrograph of a particle noticed on KCl crystal
55
Figure
8.
Photomicrograph of single KCl crystal
55
Figure
9.
Typical plot of crystal size distribution of KCl
crystallized from solution containing 32g KCl and
4.06 g NaCl per 100 ml solution
58
Typical plot of crystal size distribution for KCl
crystallized from a solution containing 32.6g KCl
and 3.3e MaCl^-ôH^O oer 100 ml solution
60
A plot of the crystal size distribution for KCl
crystallized from solution containing 27.3 g KCl,
8.3 g NaCl, and 1.7 g MgCl2*6H20 per 100 ml
solution
62
A plot of the crystal size distribution for KCl
crystallized from a solution containing 27.0 g
KCl, 6.4 g NaCl, and 30.5 g MgCl?'6H20 per 100 ml
solution
55
Hypothetical ordered solution configurations of
a) KCl, b) KCl + MgCl2, c) KCl + NaCl, and
d) KCl + NaCl + NgCl?
71
Figure 10.
-
Figure 11.
Figure 12.
Figure 13.
-
z
z.
vi
NOMENCLATURE
A
area
B
nucleation rate
C
concentration
equilibrium concentration
E
energy
G
linear growth rate
G^
growth rate at zero size
AG
overall excess free energy
AG .
crit
overall excess free energy of a particle with radius r
^
c
AGg
surface excess free energy
AG^
volume excess free energy
K
Boltzmann constant
volumetric shape factor
L
crystal size, equivalent diameter
M
molecular weignt
suspension density
n
population density
n°
population density at zero size
N
number of crystals
^i
volumetric flow rate in
volumetric flow rate out
particle radius
r
c
critical radius
C-C^, supersaturation
vii
t
time
T
temperature
V
volume
W
weight of crystals
a
surface characteristics
S
driving force related to chemical potential
p
density
T
residence time
a
surface energy
constant described by BCF equation
viii
DEDICATION
To my mother and the memory of my father
1
INTRODUCTION
Crystallization of a soluble solute from an aqueous solution is
widely used in industry for both separation and purification purposes.
Crystal size distribution (CSD) is one of the most important character­
istics of the product of industrial crystallizers. The control of CSD
has long been a major problem in industry because of further processing
and customer specifications.
Crystal growth rate and nucleation rate (crystallization kinetics)
are the principal physical phenomena affecting CSD.
The pH of the
solution, degree of agitation, supersaturation level, temperature of the
crystallizer, and presence of impurities are the factors which have the
most important influence on the crystallization kinetics and hence, on
the CSD.
The effect of these factors on the CSD must be studied for a
proper crystallizer design.
The effect of impurities on the CSD and as a habit modifier has
been studied in Lerms of addicives being added no che solution of the
pure materials to be crystallized.
Studies available for the effect of
impurities which accompany these materials in their natural resources
are rare.
Mullin (80) reported the lack of design information for the
processing of complex brines, which is a subject of considerable
commercial interest at the present time.
In this work, the crystallization of KCl from saturated solutions
containing various concentrations of NaCl and MgCl^ separately and com­
bined was investigated.
The investigation was carried out in a vacuum
2
cooling crystaliizer which satisfies the constraints and assumptions
incorporated in the continuous, mixed suspension, mixed product removal
crystaliizer.
This type of crystaliizer was chosen because the solu­
bility of KCl increases with temperature while that of NaCl is almost
independent of temperature.
A considerable amount of theoretical work has been done in the area
of continuous, mixed suspension, mixed product removal crystallization.
The analysis technique previously developed by Randolph and Larson (90)
for studying the kinetics of nucleation and growth in a mixed-suspension
crystaliizer was used here.
This method of analysis provided a useful
tool for studying the effect of impurities on the crystal size distribu­
tion and the crystallization kinetics.
3
LITERATURE REVIEW
Recovery of KCl from Brines
Potassium chloride can be recovered from carnallite (KCl-MgCl2*6H20)
or from sylvinite (KCl-NaCl) by solution mining or it can be recovered
from brines.
The major production lakes of KCl are the Dead Sea,
Bonneville Lake, and Great Salt Lake.
The brine composition of these
lakes and sea water is shown in Table 1 (Hadzeriga (54)).
Table 1.
Brine composition in percentage by weight
Bonneville
Dead Sea
Great
Salt Lake
Sea water
bittern
KCl
1.2
1.0
2.1
1.9
MgClg
1.2
12.1
2.7
8.7
MgSO^
0.3
5.4
6.1
18.4
12.1
2.8
CaClg
NaCl
24.0
7.0
From Table 1, it is clear that KCl concentration in sea water bitterns is twice that of the Dead Sea and Bonnevi1le.
However, potash
recovery from sea water bittern is not economical.
The recovery of KCl from the Dead Sea is more economical than in
other processes due to absence of sulphate which, if present, compli­
cates the recovery of KCl and results in low grade products and because
it contains very little slime which, if present in large amounts, would
complicate separation of KCl from the brines as Sauchelli (98) suggested.
4
Flotation was the method of recovery of KCl from mixed salts in the
past, but crystallization is in greater use at the present time.
Noyes
(83) attributed the shift to crystallization to the possibility of get­
ting more pure products, more control of product size, and to get better
handling properties of KCl.
A flowsheet of the recovery of KCl from the
Dead Sea brines by crystallization as presented by Epstein (38) is shown
in Figure 1.
A knowledge of the components present and quantitative data on the
phase equilibria for the production of KCl from brines is required.
D'Ans (33) reported such data for a large number of salt mixtures.
Braitsch (23) suggested that not only the solubility of one salt will
decrease in the presence of other salts, but that the temperature coef­
ficients of solubility of KCl is decreased from 0.34 to 0.31 mole
K2CI2/IOO mole H2O per degree in the presence of NaCl and is reduced
more by the addition of MgCl2.
The temperature coefficient of solubility
of NaCl becomes negative in the presence of KCl which suggests that NaCl
can be precipitated from KCl-NaCl solution upon heating; and becomes
negative, also, at a concentration of about 20 mole MgCl2 per 1000 mole
H^O.
This suggests that in a crystallization process MgCl2 concentra­
tion should never attain this value to prevent precipitation of NaCl.
Nucleauion
The formation of a nucleus capable of further growth is the first
step in crystallization.
Mullin (79) divided all nucleation into
a) primary nucleation which refers to all manifestations of homogeneous
and heterogeneous nucleation and b) secondary nucleation.
5
Dead Sea brines
MgCl2, NaCl, KCl
H^O
HgO
Solar Evaporation
Solar Evaporation
End Brine
NaCl
Carnallite + NaCl
Decomposition j
Sylvinite
Hot Leach
NaCl
Crystallization
Î
KCl
End Brine:
Will be used for the production of Br?, MgO, HCl, MgCl^
Sylvinite:
KCl - NaCl
Carnallite: KCl*MgCl2*6H20
Figure 1.
Flowsheet of the recovery of KCl from the Dead Sea (from
Epstein (38))
6
Primary nucleation
Homogeneous nucleation
Homogeneous nucleation is the spontaneous
nucleation which occurs in the labile region of the temperature concen­
tration diagram.
Mullin (79) discussed the free energy changes associ­
ated with the process of homogeneous nucleation.
maximum value,
He concluded that the
corresponds to the critical nucleus, r^, and for
a spherical cluster is given by
4irar_^
AGcrit =---3"^-
(1)
where a is the surface energy of the nuclei per unit area.
From equation 1, it is clear that a newly created nucleus in a super­
saturated solution depends on its size; it can either grow or dissolve,
but the process which it undergoes should result in the decrease in the
free energy of the particle.
The critical size r^, therefore, repre­
sents the minimum size of a stable nucleus.
Volmer (108) was the first to propose that the rate of spontaneous
nucleation follows an Arrhenius type relationship:
B° = C exp(-AG/KT)
where
(2)
is the nucleation rate, C is a constant, and AG is the free
energy of formation of a nucleus.
Nielsen (82) estimated the pre-
exponential factor in equation 2 and expressed AG as a function of
supersaturation and other system properties:
o
30
B =10
exp [-
SoSyZ
_ „
r]
K^T^dn s)2
(3)
7
Here 6 is a geometric factor, a is the surface energy, V is the molecular
volume, K is Boltzmann's constant.
T is the absolute temperature and S
is the ratio of the supersaturated solution concentration to the
equilibrium concentration.
Heterogeneous nucleation
Generally, heterogeneous nucleation
refers to nucleation resulting from the presence of foreign insoluble
material.
Since it is very difficult to produce an impurity free system
which is necessary for homogeneous nucleation, heterogeneous nucleation
is probably the important type of primary nucleation.
Mullin (79) showed
that the overall free energy change associated with the formation of a
critical nucleus is less in heterogeneous nucleation than that of the
homogeneous nucleation.
Miers and Isaac (78) proposed an equation of the power-law form for
nucleation rate which accounts for the observed fact that nucleation
doesn't occur at very low supersaturation.
This equation is based on
the concept of the metastafale region and is given as:
B° = K (C-C )^
HI
(4)
where C is the solute concentration and C is a concentration greater
m
than the saturation concentration but below which nucleation doesn't
occur.
systems
Randolph and Larson (90) pointed out that in most inorganic
approaches the saturation concentration
and equation 4
becomes
B° = K (C-Cg)^ = KS^
.
(5)
8
Secondary nucleation
Botsaris (19) defined secondary nucleation as that which takes
place only because of the prior presence of crystals of the material
being crystallized.
Strickland-Constable and Mason (103) suggested
several possible mechanisms for secondary nucleation such as:
initial
breeding, needle breeding, polycrystalline breeding, and collision
breeding.
Initial breeding (103) occurs when dry crystals are first intro­
duced into a solution.
Small crystalline fragments present on the
crystal surface are washed from the surface by the solution and form
nuclei.
Needle breeding (103) sometimes arises at high supersaturation
where dendrites may develop or needles grow from existing crystals.
In stirred solution, these crystals break and may produce secondary
nuclei of visible size.
Polycrystalline breeding (103) occurs when a polycrystalline
aggregate is broken off to give new crystals.
Collision breeding or contact nucleation (103) can occur when a
crystal collides with another crystal or other solid objects such as
agitator or crystallizer walls.
Garside and Davey (48) suggested that the most important mechanism
of secondary nucleation in industrial crystallizers would appear to be
contact nucleation because initial breeding could be important only when
crystallizers are seeded with fresh material and needle breeding would
occur only at supersaturation far above normal operating levels.
9
Melia and Moffitt (76) found that the number of secondary nuclei
produced in the solution of KCl is dependent on the degree of agitation,
the rate of cooling, and the degree of supercooling of the solution.
Lai et al. (69), while working on the collision breeding of magnesium
sulphate, showed that the number of nuclei produced strongly depended
on supersaturation and suggested that the predominant mechanism for
collision breeding must be by microattrition.
In order to explain the
supersaturation dependence, it was suggested (69) that only nuclei
larger than a certain critical size survive.
This survival theory
assumes that the number of fragments produced during the contact process
are not affected by supersaturation, but that supersaturation affects
the number of fragments that survive.
This general behavior was sup­
ported by the work of Garabedian and Strickland-Constable (45).
Clontz and McCabe (31) and Johnson et al. (62) studied the contact
nucleation of magnesium sulphate and confirmed that increasing supersaturation increased the number of nuclei.
It was also shown that
increasing contact energy at first increased the production rate although
a plateau value was eventually reached, and different faces produced
different nucleation rates.
These authors found that nucleation rate
is related to area and energy of contact by:
K(Og - 1)E
(6a)
where N is the number of nuclei per contact, A is the area of contact,
E is the energy of contact,
empirical constant.
is the supersaturation ratio, and K is
10
Bauer et al. (6) and Larson and Bendig (70) investigated the
mechanism of contact nucleation and the parameters affecting it.
A
strong dependence on supersaturation was found, and nucleation rate
increased with an increase in residence time.
Farther, as the contact
frequency was increased, a critical value was reached above which the
number of nuclei per contact decreased.
The decrease was very rapid
and enabled surface regeneration time to be defined.
Powers (85) suggested that nucleation can occur by fluid shear.
He suggested that the crystal surface acts as a gathering place for
submicroscopic sized solute particles which are awaiting their turn to
be incorporated into the crystal lattice.
These particles are at times
swept away by fluid shear and hence become nuclei.
Sung et al. (104)
demonstrated secondary nucleation by fluid shear for the MgS04*7H20
system.
Garside and Davey (48) suggested that the contribution of such
a mechanism to secondary nucleation in crystallizers is uncertain.
An experimental study by Denk and Botsaris (36) making use of rightand left-handed optical properties of sodium chlorate crystals showed
that nuclei originated from the crystal surface rather than the solution
at low supersaturation and at very low supersaturation, the solution
again appeared to become a source of nuclei.
Garside and Larson (49), Garside et al. (50), and Rusli et al.
(95), using microscopic observations and the Coulter Counter, showed
that attrition of the crystal surface can readily produce secondary
nuclei.
The size distribution of the fragments was shown to be dependent
on the solution supersaturation.
11
Botsaris (19) suggested six mechanisms for the removal of nuclei
from the following sources:
a) spontaneous removal, b) removal by fluid
shearing forces, c) removal by collision of a crystal with an external
surface, d) removal by collision of crystals with one another, e) removal
of nuclei by increasing the effective supersaturation around the crystal
and f) removal of the nuclei by activating existing embryos in the solu­
tion.
Larson et al. (72) found that secondary nucleation is related to
the magma density.
Randolph and Cise (89) showed that nucleation rate
increases with supersaturation, mass and exposed surface area of the
crystal suspension as well as the level of agitation.
Many investi­
gators (11, 89, 109) studied the effect of the speed of the stirrer on
nucleation.
All the above investigators have expressed nucleation as
a power law model of the form
B° = f(T, RPM, S^, M^^)
(6b)
where RPM, S, and M^ are the stirrer speed, the supersaturation and the
suspension density respectively, and a, b, and c are empirical constants.
Crystal Growth
A unified crystal growth theory has not yet been developed.
Instead
a set of complementary crystal growth theories exists, each dealing with
several aspects of the crystal growth process.
These general theories
of crystal growth have been applied and perhaps modified for growth
from solution (47).
12
Mullin (79) classified the many theories of crystal growth under
three general headings of surface energy, adsorption layer, and diffusion
theories.
The surface energy theories are based on the postulation that
the shape a growing crystal assumes is that which has a minimum surface
energy.
The diffusional theories assume that matter is deposited con­
tinuously on a crystal face at a rate proportional to the difference in
concentration between the point of deposition and the bulk of the solu­
tion.
The adsorption layer theories assume that crystal growth was a
discontinuous process, taking place by adsorption, layer by layer, on
the crystal surfaces.
Several modifications of this adsorption theory
have been proposed in recent years.
Bennema (9, 10) discussed many of
these modifications.
Garside (47) discussed in his review, the nature of crystal/solution
interface.
When a growth unit reached the crystal surface, it must find
a position such as a kink site where it is strongly bound, or it is
likely to return to the fluid.
The concentration of kinks and hence,
ulic ùieci'iâi'iisïïi and growtli rate depend on the structure of the crystal
surface.
Molecularly rough surfaces have many kinks and continuous
growth can occur.
As the crystal surface becomes smoother, growth be­
comes more difficult and kink sites are only found on the edge of two
dimensional nuclei or in the edges of steps formed by screw dislocations.
Jackson (60) introduced a dimensionless factor to describe surface
characteristics defined as:
a = 4e/KT
(7)
13
where £ is the bonding energy.
Defining the driving force as the dif­
ference in chemical potential of growth units in the fluid and the
solid states.
S = A;j/KT
(8)
In general, low values of a correspond to a molecularly rough surface
in which many kinks are available.
High values of S, which is almost
equal to the relative supersaturation, also correspond to molecularly
rough surfaces.
For most cases of growth from solution, it appears
that either a nucleation or a continuous step growth mechanism best
fits the available experimental data (47).
In the following, these
models will be discussed.
Two-dimensional nucleation models
Two-dimensional nucleation models are based on the formation of
a stable two-dimensional nucleus surface.
Once these nuclei are formed,
otucr growth uiilcs can aci-ach Luemselves co the edge of the nucleus
and growth across the face takes place.
Based on the spreading velocity
of these nuclei as compared to the rate of formation of new nuclei,
Ohara and Reid (84) discussed three models which all result in the rela­
tion between growth rate R and supersaturation a which in general can
be put in the form of
P
R = Ma exp(-N/o)
(9)
The mononuclear model is used when the spreading velocity of the nucleus
14
is infinitely fast, P = 1/2.
The polynucleus model is used when the
spreading velocity is essentially zero; thus, the only way the crystal
can grow is by accumulating many nuclei to cover the entire surface,
P = -3/2.
The third model known as the birth and spread model allows
for formation of critical nuclei and their subsequent growth of a
finite rate, P = 5/6.
Continuous step model
The two-dimensional nucleation model requires a reasonably high
degree of local supersaturation.
For low supersaturation, Frank (40)
suggested the idea that the growth of a crystal can be explained by
the presence of screw dislocations which will produce steps necessary
for growth.
As a result, Burton et al. (28) developed a step model
theory of crystal growth.
This theory is based on the assumption that
far from the center, the steps are considered to be parallel and equi­
distant.
The authors derived a well-known BCF equation:
G = C — tanh (-f')
where C and c
c
(10)
are the characteristic constants.
It should be noted
that
CO?
for
a << a
G = ——
for
G >> Gg
G = Co
(11)
and
.
(12)
Bennema (8) has reworked the BCF theory with particular reference to
15
growth from solution and his calculations confirm the importance of
surface diffusion.
The integration of growth units from the bulk solu­
tion directly into kink sites appears to be a more difficult process
than integration via surface diffusion.
Berthoud (14) suggested that there were two steps for crystal
growth, a diffusional process followed by first order surface reaction
Based on this assumption and for crystals having size-independent
step.
growth, Randolph and Larson (90) suggested a simple power law equation
of the form
(13)
This is the most useful and applicable growth expression even it
presumes that growth rate is not a function of size.
McCabe (75) showed that growth rate was independent of size under
most conditions.
Later many investigators (1, 24, 29) showed that
growth is size dependent.
The most famous equation which represents
this situation was proposed by Abegg et al. (1) and is of the form
G = G
where
(1 + yL)°
(14)
is the growth rate at zero size and is a function of super-
saturation, and Y and b are the two adjustable empirical constants
defining the size dependence of the growth rate.
Janse and de Jong
(61) suggested, instead of size dependent growth, that dispersion,
that is, crystals of a given size growing under the same conditions
exhibit a distribution of growth rates, better explains what appears
16
to be size dependent growth.
Garside and Davey (48) suggested, however,
that microscopic evidence supports the idea of growth dispersion while
data obtained by Coulter counter analysis show size dependent growth
rates.
Crystal Population Balance
Randolph and Larson (90) developed a general theoretical descrip­
tion of a crystallization system.
The following representation is
based on their work.
To formulate the population balance, it is necessary to define a
continuous variable to represent the discrete distribution.
To do
this, it is necessary to define the population density function which
is a representation of the number of the crystals in a given size range.
The population density, n, is defined by
n =
lim ^
(15)
AL->O ^
where AN is the number of crystals in tbe size range
to
per unit
volume of suspension.
The population balance can be applied to a mixed suspension, mixed
product removal crystallizer (MSMPR) operating at steady state.
It is
assumed in the analysis that the MSMPR crystallizer operates at conscant cemperacure, conscant working volume (v), constant feed flow race
and composition, perfect mixing, no classification at withdrawal, and
no breakage of crystals.
The population balance requires that the number
of crystals in a given size range must be conserved with no accumulation
17
if the system is to operate at steady state.
The population balance in a system of volume (V) for an arbitrary
size range
to
and time interval It is
Q^n^ALAt + VG^n^ At = QnALAt + VG2n2 At
where Q is the volumetric flow rate; n^ and
at sizes
and L^, respectively;
densities in the range
respectively.
respectively.
to
(16)
are population densities
and n are the average population
in the feed and in the product streams,
and G^ are the linear growth rates at sizes
and L^,
Rearranging, dividing by AL and At and taking limit as AL
approaches zero, equation 16 becomes
^
= Qi%i -
•
(17)
Note that as AL approaches zero, the average values of n become point
values.
Assuming there are no crystals in the feed and McCabe law
holds, i.e. G
G(L), the equation becomes
IE + 7^ = 0
(18)
where T = V/0 is the mean retention time of crystals in the crystallizer.
The solution of equation 18 is:
n =
exp (-L/GT)
(19)
where n° is the population density of zero size particles.
Equation 19 is the fundamental relationship between L and n char­
acterizing the crystal size distribution obtained from a MSMPR
18
crystallizer.
A plot of In n versus L results in a straight line.
growth rate can be obtained from the slope of the line (the intercept of the line at L = 0.
The
; n° is
The parameter n° is related to the
nucleation rate by
B° = n°G
.
(20)
Effect of Impurities
Every component of the system except the one being crystallized
may be considered as an impurity, including the solvent. Impurities
may affect crystallization kinetics, crystal habit, and crystal caking
properties.
Garrett (46), Mullin (79), and Khamskii (66) summarized
the possible mechanisms of these effects as follows:
(1) Change in solubility.
(2) Its role in heterogeneous nucleation.
(3) Change in the characteristics of the adsorbed layer at the
crystal interface and subsequently the integration of the
growth unit.
(4) Lattice structure similarity which enhances heterogeneous
nucleation.
(5) Complex formation between the crystallizing material and the
impurity.
Boistelle (16) supported the interpretations based on changing the
characteristics of the adsorption layer.
Bunn and Emmet (27) found that the thickness of the adsorption
layer is affected by the dielectric constant of the aqueous solvent
which can be reduced by the addition of alcohol.
Amsler (2) measured
nucleation rates of KCl in water and water-ethanol mixtures and found
19
that crystal-solution interfacial tension decreased with increasing the
concentration of ethanol.
Bourne and Davey (21) studied the solvent
effect on the growth kinetics of hexamelhylene tetramine crystal and
concluded that for growth from ethanol a dislocation or nucleation
mechanism may occur and for growth from aqueous solution growth may be
continuous.
Khamskii (66) found that for potash alum, impurities with common
ions accelerate the growth process by reducing the solubility while
inorganic salts with no common ion effect accelerate the growth at low
concentration and reduce it at high concentrations because it acts as a
diffusion retarder.
For the case of potassium nitrate, Khamskii (67)
found that the dielectric constant was higher in the presence of impur­
ities; in addition, the dielectric constant increased with increasing
the ion charge.
+3
Davey and Mullin (34) found that Cr
growth inhibiting
+3
power on ammonium dihydrogen phosphate is greater than that of Fe
and
+3
+3
A1 . They (34) attributed that to the adsorption of Cr
on the steps
while Afand Fe'"" adsorption is on the ledges between the growth steps.
Mullin (79) attributed that to the hydration of ions.
Shor and Larson
+3
+2
(100) studied the effect of Cr
and CO
on the crystallization kinetics
of KNO^.
In both cases, they found that nucleation rate decreased and
growth rate increased with increasing concentration of impurities.
They
concluded that these highly charged impurities adsorbed on the nucleat­
ing sites, thus inhibiting nucleation.
The growth rate increased due
to the increase in supersaturation, which was caused by the constraint
on the system which required equal production rates at various impurity
20
+3
concentrations. Larson and Mullin (71) studied the effect of Cr
the crystallization kinetics of ammonium sulphate crystals.
on
Their
results paralleled those of Shor and Larson (100) and a similar explana­
tion was given for this behavior. Khambaty and Larson (65) studied the
+3
effect of Cr
on the crystallization kinetics of MgSO^'YHgO. The
+3
growth rate decreased with increasing Cr
concentration.
The growth
rate decreased because the chromium ion at higher concentrations poisons
the crystal surface, thereby retarding the rates of incorporation of
solute molecules into the crystal lattice.
With this decrease in growth
rate of crystals, fewer nuclei survive, resulting in a decrease in the
net nucleation rate.
Cooke (32) studied the effect of MnCl^, CdCl2, and ZnCl2 on the
habit of NaCl in a vacuum crystallizer.
He found that the cubical
habit of NaCl had changed to octahedral and attributed that to lattice
structure similarity between these inorganic additives and NaCl.
Liu
and Botsaris (74) studied the crystallization of NaCl in MSMPR in the
presence of Pb^^.
+2
of Pb
They suggested the following model for the effect
on the crystallization kinetics.
f" =
^2
2
(7^)'^
""12
(21)
É. _
3% "
(22)
'Ciz'
where C^ is the impurity concentration; P, q, and q' are independent
of impurity concentration; 1, 2 refer to different experimental condi­
tions. Boistelle and Simon (17) attributed the habit change of NaCl
21
from cubic to octahedral in the presence of CdCl2 to the epitaxial
formation of Na,CdCl^'3H,0 which is compatible only with the (111)
face of NaCl.
Continuous Evaporative Crystallizers
There are three basic types of industrial crystallizers.
Cooling
crystallizers are those in which cooling is used to reduce the solute
solubility by lowering crystallizer temperature.
In precipitation
crystallizers, supersaturation is induced by the addition of a third
component in which the solute is insoluble or by carrying on a chemical
reaction in which the product is insoluble.
The third type is the
evaporative crystallizers in which supersaturation is produced by removal
of the solvent, usually under vacuum.
Randolph and Larson (90) suggested
that the precipitation crystallizer is the easiest to do on the labora­
tory scale because it can be carried out isothermally and at atmospheric
pressure; on the other hand, severe problems occur in removing a represciiuauxvc
yxuu.uv.u
aamy
u.ii
uiic
vaL-LiLua
ujyc
w x.
interest in this work.
Bamforth (5) discussed in detail the features and design of the
different types of industrial crystallizers.
may be any one of the above types.
Industrial crystallizers
Newman and Bennett (81) reviewed
the types of crystallizers and reported the advantages of using a
draft tube baffle crystallizer (DTB).
These advantages are a) bringing
the growing crystals to the boiling surface, b) maintaining a large
internal circulation to reduce supersaturation of the boiling surface.
22
c) uniformity of the product, and d) achieving the desired degree of
mixing with minimum power.
Randolph and Larson (90) discussed the physical features of the
MSMPR crystallizer which meet the requirements made by the assumptions
inherent in the population balance derivation.
Crystallizers with
draft tubes must be used to achieve good mixing with minimum power.
It is also necessary to achieve isokinetic withdrawal of slurry to pre­
vent classification.
This can be done by use of intermittent product
withdrawal to eliminate classification.
The crystallizer volume must
be small enough so that minimum feed material is required but large
enough so that sampling does not result in appreciable disturbance of
the system.
Two studies using continuous, laboratory scale, evaporative
crystallizers have been reported in the literature.
(39) used a plexi-glass crystallizer.
Epstein and Sowul
It had 2.2 to 2.5 liters work­
ing volume, and an operating temperature of 70°C.
The slurry was
pumped out of the crystallizer continuously by a metering pump through
an cutlet located at the bottom of the crystallizer, and baffles were
used to prevent vortexing.
Rosen and Hulburt (94) and Rosen (93) used
an eight liter pyrex glass cylinder closed at either end by brass plates.
The feed was added intermittently so that the level of 8000 ml was held
constant.
The slurry is removed from the crystallizer intermittently
by a vacuum system through an outlet located at the bottom of the
crystallizer.
23
Crystallization of KCl
Botsaris et al. (20) studied the crystallization of KCl in the
+2
presence of Pb . They suggested that the impurity suppresses primary
nucleation but secondary nucleation can occur if the uptake of impurity
by growing crystals is significant.
The seed crystal creates an
impurity concentration gradient about itself and the concentration of
impurity near the crystal surface becomes lower than that in the bulk
solution.
If it is reduced low enough, nucleation can occur.
The
distribution coefficient is usually less than 1, but may become greater
than 1 at very low solution impurity concentration.
This led to postu­
lation of a model for nonequilibrium capture of impurities in growing
crystals.
Metchis (77) reported that PbCl2 prevents agglomeration of
KCl crystals but changes habit.
Sariag et al. (97) found that PbClg
added to KCl solution gives single cubic crystals which decrease in
size with increasing concentration of PbCl^-
They attributed that to
poisoning the active growth sites or to the role of a PbCl2 complex in
KCl nucleation.
Genck (51) studied the effect of temperature on the crystallization
kinetics of KCl.
The order of nucleation was 2.6 and independent of
temperature range 12-30°C. The range of the growth rate was 11.7x10 ^
to 20.7x10
-8
m/s.
The KCl crystals obtained were a collection of
contiguous cublets instead of a single cubic crystals.
Randolph et al. (88) reported excessive agglomeration of KCl in
a classified crystallizer experiment.
Noticeable improvement in crystal
habit was found by adding magnesium sulphate (0.75 g Mg/100 g water)
24
and sodium chloride (near saturation) to the feed solution.
Single
cubic crystals were present in large amounts in sizes up to approximately
0.4 mm. They attributed this habit change to the amount of sodium
chloride added.
The order of nucleation obtained was 4.99 at 32°C for
the range 10 to 150 mm.
m/s.
The range of growth rate was 2x10
to 12x10
Garrett (46) reported the agglomeration of KCl crystals in the
presence of NaCl.
It is difficult to obtain pure KCl when NaCl is also
crystallizing, and usually this impurity is partially in mixed crystal
form.
Epstein (38) reported that in marine type sylvinite, formed from
the incongruent dissolution of carnallite, the KCl crystals are small
and tend to stick together, and agglomerate with NaCl crystals, probably
by electrostatic attraction.
He suggested that the agglomeration of
KCl and NaCl crystals may be reduced considerably by adding small
quantities of crystal habit modifiers for sodium chloride, during the
decomposition of carnallite, e.g. sodium ferrocyanide and nitrilotriacetamide.
Sarig et al. (97) reported agglomeration of KCl crystals in the
+2
prsence of Pb
.
Also in the production of potash particles larger
than 1 mm, in industrial crystallizers excessive agglomeration does
occur, even when the mother liquor is saturated with respect to sodium
chloride and contains 20-60 g per liter MgCl2.
Thay suggested that the
higher the degree of solvation of the crystal surface, the lesser is the
tendency to aggregation; for example, sodium chloride tends to aggre­
gate less than potassium chloride.
This is because the affinity of
sodium to solvation by water is higher than that of potassium.
Under
similar conditions, they found that the sites of the single crystals
25
which compose the agglomerates is inversely proportional to the
degree of undercooling, which is also proportional to the concentration
of lead.
Svoronos (105) studied the effect of CaCl2, KNO^ and MgCl^ on the
stability of supersaturated solution of KCl.
They were able to determine
the path of the spontaneous nucleation of KCl by measuring the intensity
of a diffusing light through the solution.
They concluded that spontane­
ous crystallization of KCl occurred at lower supersaturation in the
presence of these salts than in the pure KCl solution.
Furthermore,
they found MgCl2 and CaCl2 have the same effect and required lower
supersaturation than that of KNO^.
Structure of Electrolyte Solutions
The structure of electrolyte solution is related to solvent
structure, ion-solvent interactions, and ion-ion interactions.
Since
water is the most used solvent, the discussion below will be focused
on the structure of aqueous electrolyte solutions.
Water structure
The structure of water has been discussed extensively by Kavanau
(64), Bokris and Reddy (18), and Hunt (58).
The properties of ice and water are determined by the structure of
the water molecule.
The gaseous water molecule is a bent molecule
with the HOH angle being 105°.
In ice, the angle HOH is practically
the same as the tetrahedral angle.
X-ray scattering measurements on
ordinary ice showed a tetrahedral arrangement of oxygen atoms which
26
can be understood in terms of a tetrahedral arrangement of lone pairs
of electrons and two hydrogen atoms about the oxygen atom.
In liquid water, the structure is described as somewhat broken
down, slightly expanded form of the ice lattice (18).
X-ray, Raman
spectra and infrared studies have been made on water.
They indicated
that there is considerable degree of short-range order characteristic
of tetrahedral bonding in ice.
In addition to the water molecules which
are part of the network, there can be a certain fraction of structurally
free, unassociated water molecules.
Hunt (58) estimated the fraction
of this free water molecules to be about 15%.
Ion-water interactions
Bokris and Reddy (18) suggested that when an ion enters water,
the structure of water is disturbed to an extent which depends on the
distance from the ion.
They (18) described the structure of water
near an ion by referring to three regions.
In the primary, or structure
enhanced region next to the ion, the water molecules are immobilized
and oriented by the ionic field; they move as and where the ion moves.
Then, there is a secondary, or structure broken region, in which the
normal bulk structure of water is broken down to varying degrees.
The
in-between water molecules, however, do not partake of the translational
motion of the ion.
Finally, at sufficient distance from the ion, the
water structure is unaffected by the ion and displays the tetrahedrally
bonded networks characteristic of bulk water.
The structural changes in
the primary and secondary regions are generally referred to as hydration.
27
The above descriptions of the solvent surrounding an ion have been
used as a basis of a structural treatment of ion-water interactions,
initiated by Bemal and Fowler (13).
To separate out the various aspects
of the total interaction, Eley and Evans (37) proposed the following
cycle:
1. Tetrahedral group of n primary solvent molecules plus one more
to make room for the bare ion is removed from the solvent to
the gas phase.
2.
This leaves behind a hole in the water.
The gaseous tetrahedral group is dissociated into n+1
separate water molecules.
3.
lon-dipole bonds are formed between the ion and n of the
n+1 water dipoles, thus, the primary hydrated sheath is
formed; n is referred to as the hydration number.
4.
The ion together with its primary hydrated sheath is transferred
from vacuum into the hole in the water.
5.
The introduction of the primary hydrated ion into the hole
does lead to some disturbance of the structure of the surrounding
water.
In fact, this is the structure breaking that has been
referred to in the secondary region.
6.
The remaining water molecules return to the water by condensation.
The heat of hydration calculated using this model is lower than that
obtained experimentally.
Hunt (58) attributed that to the neglecting of
all forces but the simple coulombic ones.
Also, this model does not
predict a difference in hydration enthalpy of negative and positive ions
of the same size (18).
28
Buckingham (26) modified the preceding model by taking into account
the important lone-pair electrons in water molecule, induced dipole
moment, London forces, and the effect of ion charge on the dielectric
constant.
Accordingly, he concluded that the heat of ion-dipole inter­
actions, in step 3 above, has to be replaced by the heat of ionquadrupole interactions.
Bokris and Reddy (18) found that there is
agreement between heat of hydration calculated by this model and the
experimental value, with an average disagreement of 5%.
Also, the model
predicts that opposite charged ions of equal size have different hydra­
tion enthalpies.
On the other hand, Eley and Evans (37) suggested that
positive and negative ions of the same size have different hydration
entropies, that is, in contrast with their calculation of enthalpies
in which no asymmetry was found.
Frank (41) noticed that an inert gas in water, in addition to
loosing freedom of motion causes an ordering of nearby water that leads
to a larger negative hydration entropy term, he (41) called that structure
formation.
In addition to the structure formation, there will be a
structure breaking effect on more distant water molecules that will give
an entropy increase.
The structure breaking is due to the fact that the
new arrangement will not easily fit into the usual water structure.
He (41) also noted that inert gas atoms in water have more negative
entropy than gaseous ions of the same size.
This suggests a greater
structure breaking effect for the ions than for the inert gas.
Based
on the entropy of hydration, Frank (41) described the ions as follows:
29
1.
Cations smaller or more highly hydrated than k"^ are net
structure-formers.
2.
k"^ is slightly structure breaking, and the effect incrases
from K
3.
F
to CS .
is structure-former, but CI , Br , I are increasingly
structure-breakers.
X-ray studies on KCl solutions by Brady (22) and Raman spectral studies
on KI by Schultz and Hornig (99) indicated a lessening or breakdown
of the water structure which confirms the findings of Frank (41).
The picture of ion hydration as presented above is a static one.
In reality, individual molecules in the primary hydration shell have
a finite residence time before they exchange with molecules outside
this shell.
Thus, the hydration number expresses only an effective
average number.
Based on this dynamic picture of ion hydration,
Kruus (68) and Samoilov (96) proposed a somewhat rather different
approach to the properties of aqueous solutions.
Kruus (68) indicated that the mean time that a water molecule is
a nearest neighbor of the ion, "gQ2.v' depends on the energy required
by a molecule to escape from the ion.
This energy can be greater or
smaller than that necessary for a water molecule to replace its neighbor
in the bulk water.
If the difference between these two energies, AE, is
negative, then
= «P #
(23)
30
is less than one, i.e. the water molecule next to an ion is in less rigid
environment than in the bulk water.
residence time in bulk water.
t
w
in equation 23 is the mean
Cases where
T
solv
/T
w
less than one are
usually referred to as negative hydration or structure breaking.
Negative hydration would be expected for large ions where ion-water
interactions are smaller.
The so called negative hydration phenomena have been treated
extensively by Samoilov (96).
He suggested that water molecules in the
vicinity of ions, which show negative hydration, undergo exchange more
frequently than water molecules in the vicinity of other water molecules
in water.
He concluded that ions which show negative hydration,
e.g., K , CS , I , increase the translational motion of the nearest
water molecules.
The opposite is true for ions which show effective
binding to the nearest water molecules in solution, e.g., MG
2+
, Ca
2+
, and
Lit
The existence of negative hydration is related to the inevitable
distortion of water structure by the ions.
This distortion obviously
corresponds with an increase in the potential energy of the water mole­
cules lying next to the ion (98).
Ion-ion interactions
There are many models which describe the ion-ion interactions.
None of these proposed models as yet represent the behavior of electrolyte
solutions over the entire range between anhydrous fused salt and pure
water.
This is maybe due to the fact that ion-ion interactions are
concentration dependent.
The most common models which describe the
31
properties of electrolyte solutions are a) Debye-Huckel model, b) Ionpairing, and c) lattice models.
Debye- Huckel model
Debye-Huckel (35) proposed a model which was
a spectacular advance in the understanding of ionic solutions.
In this
model, dilute aqueous solutions are described as a mixture of hard
spheres in a continuous dielectric medium.
Bokris and Reddy (18)
summarized the basic postulates of this model as follows:
1.
The central ion sees the surrounding ions in the form of a
smoothed-out charge density and not as discrete charges.
2.
All the ions in the electrolytic solution are free to contribute
to the average density, and there is, for instance, no pairing
up of positive and negative ions to form any electrically
neutral couples.
3.
Only long-range coulombic forces are relevant to ion-ion
interactions; and short-range noncoulombic forces, such as
dispersion forces, play a negligible role.
4.
The solution is sufficiently dilute.
5.
Ions are regarded as point charges.
6.
The only role of the solvent is to provide a dielectric medium
for the operation of interionic forces.
The final result obtained for the mean activity coefficient by using
this model is
logf, = - A(Z Z ) 1^/2
(24)
32
where: f^= the mean activity coefficient
A
= a constant depends on water properties and the ionic
charge
Z
= the valency of the ion
I = the ionic strength and is given by
1=1/2 1
(25)
= concentration in moles per liter of ion i
Bokris and Reddy (18) discussed the validity of the above assumptions.
They (18) concluded that this model is good for very dilute solutions
(0.01 M) and the activity coefficient calculated by equation 24 does
not depend on the particular electrolyte (NaCl or KBr) but only on its
valence type.
An obvious improvement of the model consisted in removing the
assumption of point charge ions and taking into account their finite size
Dy
introducing an ion-size parameter, a, to be determined experimentally
(18).
The expression for the mean ionic activity coefficient becomes
"
-1
where x
- A(Z
1
<26)
is the radius of the ionic cloud surrounding the central ion.
This model yielded agreement with experiment at concentration up toward
0.1 M.
It also introduced through the value of a, the ion-size param­
eter, a specifity to the electrolyte which makes NaCl different from
KCl.
33
Various extensions to Debye-Huckel model have been proposed by
Robinson and Stokes (92), Earned and Owen (55), Rasaiah and Friedman (91),
Vaslow (107), and many others.
Frank and Thompson (42) indicated that
all these extensions employ parameters which are not only empirical but
also take on numerical values which are often peculiar.
Frank and
Thompson (42) and Braunstein (25) suggested that the Debye-Huckel
extensions succeeded in predicting electrolyte solutions behaviour at
concentration up toward 1 M at the most.
Ion pairing
In Debye-Huckel model, the possibility was not con­
sidered that some negative ions in the cloud would get sufficiently close
to the central positive ion, and vice versa, so that their thermal
translational energy would not be sufficient for them to continue their
independent movements in the solution.
Bjerram (15) suggested that a
pair of oppositely charged ions may get trapped in each other's
coulombic field and an ion pair may be formed.
The ions together
form an ionic dipole on which the net charge is zero.
On the average,
a certain fraction of the ions will be present in the solution in the
form of ion pairs.
Fuoss (44) suggested that two ions will form an ion
pair only if there is no solvent molecules between them.
Activity
coefficient can be calculated from equation 26 by using the free ion
fraction only.
Levine (73) and Bokris and Reddy (18) concluded that ion associa­
tion in pairs scarcely occurs for 1:1 electrolyte in aqueous media.
Kruus (68) and Bokris and Reddy (18) suggested that introducing the ionpair concept to the Debye-Huckel models allows the extension of this
34
model to higher concentrations.
Braunstein (25) discussed the new modi­
fications on this model and found it difficult to justify the applica­
tion of these models to concentrated electrolyte solutions where ionsolvent molecules interactions may be masked in the parameters of
the theory.
A more exact treatment of the Debye-Huckel model by
Friedman (43) gave reasonable agreement with the experimental results
up to 1 M.
Lattice theories
From the preceding discussion, we concluded
that Debye-Huckel model fits experimental data at low concentration.
Bokris and Reddy (18) estimated the concentration at which Debye-Huckel
model breaks down to be 0.01 M. This estimate is based on the mean
distance £ between ions in solution of concentration c in moles per
liter.
By imagining ions to be arranged in a periodic manner, one can
obtain £ as
(=
where
<")
is Avogadro's number.
In such regular arrangement, % can be
thought of as the lattice parameter.
Since £ is proportional to C ~
jo
,
it would be expected that, if the parameter I were of any importance,
there should be a dependence of the properties of an electrolyte solution
1/3
(e.g. log f^) on c
.
Bokris and Reddy (18), Robinson and Stokes (92),
and Harned and Owen (55) have commented that when numerical values of
1/3
log f^ in aqueous solutions are plotted against c
, long linear regions
35
can be noted at concentrations above 0,01 M.
This strengthens the
argument for thinking in terms of a lattice parameter, rather than an
ionic cloud, for all but very dilute solutions which satisfy the
Debye-Huckel model as presented in equation 1, i.e., a linear region
is obtained by plotting log f_^ versus c
1/2
for concentration below
0.01 M.
Ghosh (52) was the first to suggest that when an ionic crystal
is dissolved in water, it exists in the form of an expanded lattice.
1/3
He obtained a dependence of log f^ on c
by using the idea of lattice
energy.
Braunstein (25) proposed a lattice model based on an analogy
between concentrated electrolyte solutions and molten salts which have
quasi-lattice structure.
Based on statistical thermodynamics, Frank and Thompson (42),
Glueckauf (53), Bahe (3), and Prigogine (86) suggested the presence
of lattice arrangement in concentrated ionic solutions and that log f_^
1/3
is proportional to c
.
~
As proposed by Frank and Thompson (42), the
mean activity coefficient is given by
- log f^ = a -
bc^/3 + sC^/S
where a,b, and s are constants dependent on temperature.
(28)
Priogine (86)
suggested chat the degree of order of ionic lattices in solution should
depend not only on temperature but also on the concentration.
X-ray studies by Prins and Fonteyne (87) on thorium nitrate. Beck
(7) on LiCl and LiBr, Skryschewski (101) on alkali chlorides, and by
other studies reported by Samoilov (96) on H^SO^ and HNO^ indicated a
36
lattice like structure.
Raman spectral studies on LiNO^, Ca(N02)2j
and Zn(N02)2> tiy Irish and Davis (59), Hester and Plane (56),
and Hester and Scaife (57) indicated that, with decreasing water
content, nitrate ions that have been separated from cations by water
molecules displace water from the coordination spheres of the cations.
Braunstein (25) interpreted this in terms of a quasi-lattice model in
which water and nitrate ions compete for positions adjacent to the
cations.
Bokris and Reddy (18) discussed the basic characteristics of a
quasi-lattice model which accounts for the behavior of concentrated
electrolytic solutions.
The contribution of ion-ion interactions to the
chemical potential of an ionic species must be proportional to the
average interaction energy for a pair of ions.
The proportionality
constant must express the latticeness of the electrolytic solution and
the quasiness of the lattice.
For the latticeness , a Madelung like
number M* can be inserted into the proportionality constant as an
ion-arrangement factor.
For the quasiness of the lattice, it is necessary
to insert an inverse function of temperature and to include terms for
short-range interactions which produce the ordering.
Thus, the quasiness
factor will represent the decrease in the degree of order of the lattice,
-1/2
this factor may be included by inserting a function of T
portionality constant.
to the pro­
Kruus (68) suggested that the lattice in
solution is more expanded and ions have more translational energy than
that of the solid crystals.
37
EXPERIMENTAL
Equipment
Crystallizer
A continuous MSMPR vacuum crystallizer was used.
gram of the crystallizer is shown in Figure 2.
A schematic dia­
A Sargent-Welch resin
reaction vessel of two liters capacity was used as the crystallizer.
The cover of the crystallizer has three No. 24/40 grindings and one No.
34/35 grinding in the center of the cover.
This large opening was con­
verted to No. 24/40 grinding by a ground glass adapter.
The flask and
the cover were fastened together by a cast bronze ring with three clamps.
A rubber gasket was used to prevent any liquid or air leakage.
The
crystallizer was heated by a heating mantle connected to a Powerstat
variac.
Agitation was provided by a two inch diameter three blade propeller
fastened at the bottom end of the stirrer shaft.
The shaft extended
through the stirring assembly which provides an air-tight system.
A
Sargent-Welch stirring assembly consisting of 24/40 grinding, a Teflon
bushing, 0 ring, and teflon adjusting screw locking nut was mounted in
the central opening of the cover.
The propeller was driven by an
electrical motor connected to Cole Parmer constant speed control unit
by which the stirrer speed could be adjusted to achieve good mixing.
The direction of flow was down the center and up the outside annular
space between the draft tube and the vessel walls.
One of the other three openings was used to introduce the feed
Figure 2.
Schematic diagram of the crystallizer vessel
REPEAT CYCLE
TIMER
50 0 08
w
(TOP VIEW)
DRAFT TUBE
WATER
BOTTLE
HEAl
MANTLE
CRYSTALLIZER
40
through 1/4 inch glass tube directly to the center of the draft tube
below liquid level in the crystallizer.
The thermometer was housed in
the same opening as shown in Figure 2.
The second opening was used for
product removal.
A 1/4 inch glass tube was placed parallel to the circu­
lation flow in the annular space between the draft tube and the vessel
walls.
The withdrawal was made in the direction of flow.
This arrange­
ment insured isokinetic withdrawal of the slurry out from the crystal­
lizer.
The remaining opening was used for vapor removal and was con­
nected by 2 inch diameter tube to an automatic reflux control distilling
head used by Baliga (4).
Reflux or take-off was achieved by moving a
swing funnel in the distilling head by an electromagnet which was con­
nected to a percentage, repeat cycle timer with a 30 second time cycle.
The on-time could be varied between 1 percent and 100 percent of the
total cycle.
The "take-off" condensate was collected in a round bottom
flask connected to a vacuum system.
Auxiliary eg uipmeni:
A schematic diagram of the entire crystallization system is shown in
Figure 3.
The feed tank consisted of 13 gal narrow mouth poly bottle
placed in 30 gal stainless steel drum.
The drum was filled with water
which was heated by means of stainless steel immersion heater with thermoO U CL u a. V_
U 1. V./a.•
l. w j.
cue
OtlVl
L-IIC.
J
b.
W 0.0
M J. w
vided by two 'Lightnin' mixers, one mounted in the poly bottle and the
other mounted in the drum.
A constant feed flow rate throughout the run
was assured by Cole Parmer masterflex tubing pump connected to a speed
controller.
The feed was pumped through a heat exchanger and 0.2 micron
Figure 3.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Equipment and flow diagram for crystallization system
Feed immersion heater
Poly-bottle
Stainless steel drum
Lightnin mix(?r
Feed pump
Heat exchanger
Filter
Thermometer
Agitator
Automatic reflux control distil]ing head
Heat exchanger
Distilled water bottle
Distilled water bottle
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
4 llteis receiving flask
Vacuum gauge
.Sampling flask
] liter vacuum flask
Condensate collecting flask
Vacuum gauge
Pressure regulator
Water bath
Water bath
Water bath
Coolant refrigeration unit
Coolant tank
Coolant pump
10
i-.20
m
TO WATER
-*•
ASPIRATOR
<S)19
12
-<—
w—
Z
t
23
M
nnmnnmr
15
»
4>ro
'nrmrinnnnr'
IL
T
TO VACUUM PUMP
25
sr*'
26
21
43
Pall Disposable Filter, respectively, before entering the crystallizer.
Intermittent product withdrawal was used to avoid size classificaion.
This was done manually by using three-way valves because a level control­
ler arrangement did not work properly due to continuous splashing of the
solution with boiling. The slurry was withdrawn from the crystallizer
by vacuum.
The withdrawn slurry passed through a heat exchanger before
it entered in the 4 liter receiving flask or in the 250 ml sampling
flask.
The receiving flask was placed in cold water bath to minimize
evaporation.
The sampling flask was placed in constant temperature bath
at the same temperature of the crystallizer.
pump was connected to 1 liter vacuum flask.
A Sargent DUO-Seal vacuum
This flask was placed in a
cold water bath to minimize evaporation as much as possible.
The receiv­
ing flask and the sampling flask were connected to the 1 liter flask.
Two distilled water bottles were connected to the withdrawal line for
flushing purposes.
The vacuum in the crystallizer was attained by using a water aspira­
tor which was connected to a bleeding valve and then to the distillate
collection flask.
The Emil Greiner Company Cartesion Monostat was
used to regulate the pressure in the crystallizer.
The U.S. Gauge Com­
pany vacuum gauges were used to measure the vacuum in the crystallizer
and in the withdrawing vacuum system.
The cooling system consisted of a Blue M refrigeration unit with
the cooling coils immersed in the two cooling baths.
The coolant used
was a mixture of ethylene glycol and water mixture stored in 30 gal
stainless steel drum.
A Cole Parmer masterflex tubing pump was used
44
for pumping the coolant to the two baths.
Procedure
Feed preparation
Analytical grade KCl, NaCl, and MgCl2'6H20 were used in preparation
of the feed solution.
glass tank.
The feed solution was prepared in a 50 liter
The proper amount of water, extra amount of KCl, and
predetermined amount of NaCl and MgClg were mixed together in the tank.
The solution was stirred and heated by an immersed heater to the required
temperature and was kept at that temperature for 12 to 24 hours.
30 liters were prepared for each run.
Usually,
One hour before starting the run,
the clear solution from the glass tank was transferred to the preheated
poly bottle.
Meanwhile, a 1 liter feed solution was charged to the
crystallizer.
The feed solution was then heated at least 5 degrees
above saturation temperature.
Operation of the crystallization system
The solution in the crystallizer was heated till all solid particles
dissolved.
The two vacuum systems were turned on.
The crystallizer
vacuum was adjusted according to the operating temperature of each run.
The automatic reflux control distilling head was then turned on.
This
semibatch operation was continued till the first crystals appeared in
the crystallizer.
The feed was then pumped at high flow rate to warm up
the feed line, then decreased to a specified flow rate.
The heat load,
the reflux/condensate ratio, and the stirrer speed were adjusted.
mixing was achieved by 750 RPM stirrer speed for all runs.
The
Good
45
withdrawal valve was opened when the liquid level in the crystallizer
was 11/2 inches above the draft tube and was closed when the liquid
level was just above the draft tube.
After all the necessary adjust­
ments were made, eight residence times were enough to reach the steady
state.
Sampling and filtration
Before sampling, the bleed valve was opened to the atmosphere and
the vacuum to the crystallizer was turned off so that the pressure in
the crystallizer was increased to atmospheric.
as fast as possible by one of two methods:
Samples were taken then
a) through the three-way
valve to the sampling flask, or b) directly from the crystallizer by
using tygon tubing connected to the sampling flask.
The second method
was preferred because it eliminated the possibility of plugging of the
line with large crystals.
The volume of the sample was recorded and
the sample was filtered by using a buchner type funnel connected to a
vacuum Larough a vacuum flask.
As soon as all mother liquor had been
filtered off, the filtrate was transferred to a measuring cylinder con­
taining 200 ml distilled water.
All crystals were washed with a mixture
of acetone and water saturated with KCl.
volume) acetone to water.
The mixture was 3 to 1 (by
Suction was then continued until the crystals
were dry enough for easy removal.
The crystals were then removed and
spread on a watch glass for further drying.
The filtrate and a sample
from the feed were analyzed by the analytical laboratory of the Engineer­
ing Research Institute.
46
Separation of crystals
The sample of dry crystals was sieved in a nest of 3 inch U.S.
Standard Sieves.
Shaking was provided by a Rotap testing sieve shaker.
Each sample was shaken for one hour. Sieves of the following mesh
and aperture sizes were employed; the aperture sizes expressed in milli­
meters are in parentheses:
18 (1.00), 20 (0.841), 25 (0.710), 30
(0.600), 35 (0.500), 40 (0.420), 45 (0.355), 60 (0.250), 80 (0.180),
100 (0.150), 140 (0.106), 400 (0.037), and the pan.
Each size fraction
was then removed from its screen by brushing the back of the screen
and the samples were then placed in previously weighed bottles.
bottles were weighed again using a Mettler balance.
The
From the various
weights, the population density of each fraction was calculated.
47
RESULTS
Treatment of Data
The population densities were obtained by sieve analysis.
The
equation to calculate population density on a basis of 100 ml suspen­
sion is:
n =(
C-^)
K^pL^AL
V
(29)
where W is the total weight of a given size fraction,
L is the arithmetic average diameter of the aperture for the
sieve the fraction was collected on and the aperture of the
sieve directly above in the nest,
AL is the difference in aperture diameter between successive
sieves,
V is the suspension sample volume,
p
K
V
is the density of solid crystals, and
is the volumetric shape factor.
For this case, one should note that K
'
due CO its cubic habit.
V
for KCl is equal to 1.0
Also, AL's for the different sieve fractions
are not equal due to the choice of sieves employed.
For a continuous MSMPR crystallizer operating at steady state,
equation 30 is applicable:
n = n° exp (-L/GT)
.
(30)
48
A semilog plot of the population density (n) versus the size (L) gives
a straight line with a slope of -1/GT and an intercept equal to the
nuclei population density (n°).
A semilog plot of population density
versus size was made for each set of data.
In order to obtain the best
straight line through the data, a least square analysis was done for
each plot.
The intercept and the slope obtained from such plots were
used to calculate n°, G, and the nucleation rate, B°, using equation
31:
B° = n°G
.
(31)
The kinetic order, i, of KCl was calculated for each system using equa­
tion 32:
n° = K G^~^
.
(32)
As was mentioned earlier, the nucleation rate is a function of the
suspension density.
Therefore, for the nucleation rates of the differ­
ent runs to be comparable it will be necessary to put them on a common
basis.
This was done by using a suspension density of 1.0 g/100 ml.
This was achieved by dividing all nuclei population density, n \ by
the suspension density.
T
The suspension density used is given by:
^ Total He^^ht of crystals ^ ^00
Sample volume
The same procedure was followed for all systems studied.
(33)
The experi­
mental results of all the runs are tabulated in Appendix A.
The crystallization of KCl from solutions containing a) KCl only.
49
b) KCl and NaCl, c) KCl and NgClg, and d) KCl, NaCl, and MgClg in dif­
ferent concentrations were studied. For all the runs, the solution
fed to the crystallizer was saturated at 42°C, the crystallizer tempera­
ture was at 40°C, and the residence time was in the range of 17 to 22
minutes.
A typical plot of In n versus L, for solutions containing KCl only,
is shown in Figure 4. This particular plot displays the experimental
data for a 19 minute residence time.
This plot shows a noticeable
deviation from the linear relationship predicted by equation 30.
The
deviation is more pronounced in the region where L is larger than 460
microns.
This deviation is due to polycrystalline formation instead
of a single cubic crystal.
Figures 5, 6, 7, and 8 show typical pic­
tures of different sizes of KCl crystals taken under the electronic
microscope.
These pictures suggest that the polycrystals were formed
and grown together.
The pronounced deviation in the region where L is
larger than about 460 microns in Figure 5 may be due to the larger
surface area of the crystals exposed to the solution and, hence, faster
growth rate.
Genck (51), Randolph et al. (88), and Sarig et al. (97)
observed similar polycrystalline formation when attempting to crystal­
lize KCl from solution.
Polycrystalline formation was noticed in other systems studied in
this work but to a lesser extent, depending on the system.
Accordingly,
as a criterion for comparison between the crystallization kinetics of
KCl in the different systems, another least squares fit was done on part
of the data which corresponds to sizes smaller than the largest size at
Figure 4.
Typical plot of crystal size distribution for KCl
crystallized from solution containing KCl only
Upper line:
Lower line:
Best fit for sizes <0.5 mm
Best fir for complete size range
OO'B
QO'2
OD'9
-oo r— 3D
-CO ft
•23
- Q
O
O
* Nnu
-ta
L_
15
Figure 5. Photomicrograph of KCl crystals
Magnification: 30X
Average size: 180-250 micron
Figure o.
Photomicrograph of KCl crystal
Magnification: 400X
Size range: 180-250 micron
Figure 7.
Photomicrograph of a particle noticed on KCl crystal
Magnification: 3000X
Size range: 37-106 micron
Figure 8.
Photomicrograph of single KCl crystal
Magnification: 2000X
Size range: 37-106 microns
56
which the pronounced deviation starts occurring.
The best straight line
obtained using this criterion is shown on the same graph which shows
the best straight line fitted to each set of data.
This procedure was
followed whenever it was needed for all the systems under consideration.
The growth and nucleation rates for all the systems obtained from such
graphs are tabulated in Appendix B.
A typical plot of In n versus L, for KCl crystals obtained from
solution containing NaCl and KCl, is shown in Figure 9.
This particular
graph displays the experimental data for a 19 minute residence time.
A little improvement in the crystal size distribution can be noticed
in Figure 9 compared to that shown in Figure 4.
This may suggest that
the presence of NaCl reduces the formation of KCl polycrystals.
Con­
centration of NaCl in the solutions used in the different runs was in
the range of 3.2 to 13.5 g/100 ml solution.
A typical plot of In n versus L, for KCl crystals obtained from
solution containing MgCl^ and KCl, is shown in Figure 10.
This particu­
lar graph displays the experimental data for a 17 minute residence time.
Comparing Figure 10 and Figure 4, one can conclude that the presence of
MgCl^ reduced polycrystal formation as was the case with NaCl.
The
concentration of MgCl^'ôH^O used in the different runs was in the range
of 0.9 to 3.3 g per 100 ml solution.
The crystal size distribution of KCl crystallized from a solution
containing NaCl, MgCl2, and KCl together was markedly different from
those obtained in the other three systems.
That can be seen from plots
of In n versus L in Figures 11 and 12. Figure 11 displays experimental
Figure 9. Typical plot of crystal size distribution of KCl crystallized
from solution containing 32g KCl and 4.06 g NaCl per 100
ml solution
Upner line:
Lower line:
Best fit for sizes < 0.5 mm.
Best fit for complete size range.
58
RUN * 29
O
o
M
z
z
LU
O
z
o
V—
(X
_l
D
CL.
D
n
0.00
G.20
O.UO
0.60
CRTSTfiL DIAMETER L,MM.
0.80
Figure 10. Typical plot of crystal size distribution for KCl crystallized
from a solution containing 32.6g KCl and 3.3g MgCl„.6H 0
per 100 ml solution
Upper line:
Lower line:
Best fit for sizes < 0.5 mm.
Best fit for complete size range.
60
_1
o
o
s:
s:
H
z
z
D
k-
cr
—I
I Ûo
0-
cr
0.00
0.20
0.60
CRYSTAL DIAMETER L,MM.
0.80
Figure 11.
A plot of the crystal size distribution for KCl crystallized
from solution containing 27.3 g KCl, 8.3 g NaCl and 1.7 g
MgCl2.6H20 per 100 ml solution
Upper line:
Lower line:
Best fit for sizes <0.7 mm.
Best fit for complete size range.
62
dcE
RUN « 31
a5I
o
CO —
z ke
LU
-
Ûi Û-
in_
I
I
I
I
0.00
0.20
O.UO
O.SO
ChTSlHL Û Ï H M t i t » L,MM.
0.80
63
data for a 19.5 minute residence time and MgCl2"6H20 concentration of
1.7 g/100 ml while those corresponding to Figure 12 are 18.2 minutes
and 3.05 g/100 ml, respectively.
Also, it was observed that, on the
average, 50 percent of the total solid obtained in the different runs
of this system were for crystals of size smaller than 550 microns com­
pared to a 10 percent average in the other three systems.
This sug­
gests strongly that polycrystals formation of KCl was markedly reduced
in the presence of these two components.
Also, it was observed that
reduction of polycrystals formation increased with increased concentra+2
tion of Mg
in the solutions of this system.
The phenomena of poly-
crystal formation will be discussed later.
Growth and nucleation rates of KCl obtained from the different sys­
tems are tabulated in Appendices A and B.
The addition of NaCl and/or
MgCl2 did not show any noticeable effect on the growth and nucleation
rates of KCl.
This may be due to the formation of polycrystals which
makes the comparison even more unjustified.
The kinetic data of KCl
obtained from solutions containing MgCl2, NaCl, and KCl may be the most
useful.
This is because of the polycrystals formation reduction and
the industrial recovery of KCl from solutions containing both components.
Chemical Analysis
Composition of the feed solutions used in many runs and thai: of
some KCl crystals was determined by the Analytical Laboratory of the
Engineering Research Institute.
The composition of feed solutions
is tabulated in Appendix C and that of KCL crystals with the correspond­
ing feed composition is shown in Table 2.
From this table, we can see
Figure 12.
A plot of the crystal size distribution for KCl crystallized from a solution containing 27.0 g KCl, 6.4 g NaCl,
and 3.05 g MgCl2*6H20 per 100 ml solution
65
RUN
LU
Q
314
H
O •—
t—
_
—
cr QO
-A
I n
cn
X
0.00
1
1
O.IAO
0.60
CRYSTAL DIAMETER L,MM.
0.20
r
0.80
66
Table 2.
Run
Chemical composition of KCl crystals and feed solutions
% Impurity content in KCl crystals
% Sodium
25
0.12
29
0.31
32
0.14
33
34
% Magnesium
—
. . a
Feed solution composition
i
KCL
24.6
NaCl
NgCl^-ôH^O
5
32.2
13.5
<0.007
27.3
8.5
1.71
0.089
<0.007
27.0
7.6
2.47
0.062
<0.007
27.0
6.4
3.04
41
<0.007
32.0
43
<0.007
32.6
^Composition based on 100 ml solution.
—
1.61
3.28
67
that magnesium concentration in all solid samples analyzed, if any is
present, is less than 0.007 percent which is negligible, bearing in
mind that 0.007 percent is the minimum amount that can be detected by
the atomic absorption spectrophotometer used in the analysis.
On the
other hand, concentration of sodium in the solid samples increased
linearly with increasing the concentration of NaCl in the feed solu­
tion.
Comparing sodium content of KCl crystals in run 25 with that of
run 34 and that of run 29 with runs 32 and 33 suggests that the presence
of MgCl2 reduced the sodium content of KCl crystals.
Also, increasing
the concentration of magnesium chloride in the feed solution reduced
the sodium content more for the same concentration of NaCl in the feed
solution.
This can be seen by comparing runs 32 and 33.
of this will be given in the discussion section.
An explanation
68
DISCUSSION
Polycrystal formation of KCl, observed here, was observed by Genck
(51), Randolph et al. (88), and Sarig et al. (97).
Randolph et al. (88)
reported that addition of NaCl and MgSO^ reduced the formation of these
polycrystals and attributed that to the presence of NaCl, neglecting
+2
the effect of Mg
which was observed in this work.
The reason behind
this may be that they (88) used only one concentration of MgSO^ in their
experiment.
Potassium chloride is not the only compound which has exhibited
polycrystalline behavior.
Timm (106) obtained polycrystals of NaCl
from a salting-out crystallizer.
Strickland-Constable (102) obtained
polycrystals of KBr while attempting to grow clear cubic crystals.
The
KBr crystals obtained appeared to consist of a collection of contiguous
cublets; differences of alignments between these cublets was shown by
x-ray to be up to 3°.
LiciuxccLc.
uy
The pictures obtained while studying secondary
j. v ou
9
own \'tyj
potash alum, Kaufman (63) for potassium nitrate, and Berglund (12) for
citric acid show polycrystals formation, the extent of which depends on
the system used.
These observations may suggest that polycrystals are
not unique to KCl only and the cause of polycrystal formation may result
from the nature of primary or secondary nucleation.
Strickland-Constable (102) suggested that polycrystal formation in
the bulk solution may be due to one of the following:
a) a mechanism
may exist by which a crystal can catalyze the nucleation of a new crystal
with a new lattice in contact with its surface but he didn't specify any
69
mechanism, b) it may be due to high supersaturation, but he added that
supersaturation almost certainly is not the only condition which
determines polycrystal formation, and c) it may be due to nucleation
in the bulk of the solution with agglomeration of the resulting nuclei.
The restriction to nucleation occurring in the bulk of the solution is
not applicable because polycrystals have been formed while studying
contact nucleation in a contacting cell.
In this cell, the possibility
of bulk nucleation is minimal because crystals are produced by collision
or sliding.
Mullin (79), on the other hand, suggested that polycrystal
formation (agglomeration) depends on the thickness of the adsorption
layer surrounding a growing crystal.
At very low supersaturation the
adsorption layer will be thin and strongly attracted to the crystal
face, and the growth units will be readily incorporated at the growth
sites.
At very high supersaturation the adsorption layer will be thick
and this will lead to secondary nucleation.
However, at some inter­
mediate supersaturation the adsorption layer may be neither thick nor
thin and contact between two crystals in the same condition could rapidly
result in a permanent bond through their integrated adsorption layers.
It must be noted, however, that he did not specify the nature of the
growth units.
Also, his assumption of the existence of the two crystals
before a permanent bond can be formed is questionnable because the photo­
graph shown in Figures 6 and 8 strongly suggest that these crystals
were formed and grown together.
Accordingly, a better understanding of
70
nucleation and growth mechanisms may lead to an explanation to the
phenomena of polycrystal formation.
Mullin (79) suggested that a stable nucleus containing n molecules
could occur by the molecular addition of molecules till a critical sized
cluster would result in primary nucleation.
then grow by the addition of growth units.
The nuclei produced would
The molecular addition
suggested above is unlikely to occur in aqueous electrolyte solutions
due to the presence of ions rather then molecules in such solutions.
Accordingly, primary nucleation may be due to successive collisions
between the quasi-lattices previously referred to, present in concentrated
solutions which are most probably the growth units suggested by Mullin
(79). For convenience, these quasi-lattices will be called, from now
on, sub-nuclei. The presence of these sub-nuclei in the adsorption layer
of a crystal will lead to secondary nucleation on disturbance of such
layer.
This is because such disturbance will cause successive collisions
between the sub-nuclei present in this layer causing enlargement of the
cluster eventually appearing as secondary nuclei.
If the alignment of
the subnuclei which form the stable nucleus is different, then each
sub-nuclei in this stable nucleus will grow somewhat independently,
meanwhile the stable nucleus appears to grow as a single unit.
This
unit may be a single crystal or a polycrystal depending on the additional
sub-nuclei incorporated into the unit during the course of its growth.
Some possible alignments of KCl (cubical habit) are shown in Figure 13.
The reduction of polycrystal formation of KCl in the presence of
+
Na
+2
and Mg
may be attributed to the difference in the physical prop-
71
•o •
o•o
•o•
(a)
KCl SUB-NUCLEI SURROUNDED
BY WATER CLUSTERS
o
(b)
KCl SUB-NUCLEI SURROUNDED BY
+2
WATER CLUSTERS AND HYDRATED Mg
u®®®u
U
o
®n
o
• o•
•o n
(c)
KCl SUB-NUCLEI SURROUNDED BY
WATER CLUSTERS AND HYDRATED Na
n
Q
(d)
KCl SUB-NUCLEIC SURROUNDED BY
WATER CLUSTERS, HYDRATED Na+ AND
uvnoATcn mo+2
KCl SUB-NUCLEI
WATER CLUSTER
HYDRATED Na"^
+2
HYDRATED Mg
Figure 13. Hypothetical ordered solution configurations of
a) KCl, b) KCl + MgCl_, c) KCl + NaCl, and
d) KCl + NaCl + MgCl^
72
erties of these ions in aqueous solution as shown in Appendix D.
As
was mentioned earlier, k "^ ion is a structure-breaker (negative hydration)
while Na^ and
are structure-formers in aqueous solutions.
This can
be seen from the entropies of hydration of these ions which are
-15.3, -8.1, and -50.9 (cal/mol. deg) for Na^,
and Mg"*"^ respectively.
+2
These values suggest a strong ordering effect for Mg
and less ordering
effect for Na"^ and still somewhat less ordering effect, perhaps structurebreaking effect, for K . The average hydration number of K , Na , and
+2
Mg
are 4, 4, 14 respectively. The water molecules are strongly bound
+2
to Mg
+
while they are less tightly bound to Na
bound for
and still less tightly
as can be seen from the heat of hydration.
Based on these physical properties, KCl solution only will be the
least ordered solution of the four systems studied.
Accordingly,
we expect more random collisions between the sub-nuclei in the solution
and hence, more mismatch between the faces of these sub-nuclei which
upon nucleation and growth will lead to polycrystal formation.
Adding
Na' ions causes more ordering in the solution, reduces the amount of
+
water affected by the structure breaking effect of K , and it may
separate KCl sub-nuclei from each other.
These factors ?i=y lead to less
but more ordered collisions between KCl sub-nuclei and hence, less polycrystals formation.
+2
Mg
The same behavior can be anticipated by adding
+
to KCl solution but at lower concentrations than Na . A simple
+2
calculation, based on the hydration numbers of Mg
+2
Mg
+
concentration of 30% of that of Na
KCl solution by both ions.
+
and Na , shows that a
could give the same effect on
Considering the strong binding of water
73
molecules to
compared to that of Na"*", then Mg"^^ of concentration
less than 30 percent of that of Na"*" would produce the same effect on
KCl solution as that of Na"^.
~i~
Na
H~2
and Mg
This explains the comparable effect of
on reducing the polycrystals formation of KCl at lower con+2
centrations of Mg
+
as compared to that of Na
+
It was observed that the presence of Na
as was discussed earlier.
+2
and Mg
together in KCl
solution reduced the polycrystals formation more than the presence of
each one separately with the same ionic strength.
This is likely due
to different structural changes in the solution by the presence of these
ions together from that existing when they are in solution alone.
In
other words, Mg"'"^-H20-Na"'" interactions are different from Mg"'"^-H20-Mg^^
interactions and Na"'"-H20-Na"'" interactions.
The change in these ion-
water interactions with the other factors mentioned in the case of the
presence of each ion separately may lead to a more ordered solution
structure in which KCl sub-nuclei can fit easily.
Such an arrangement
suggests a cubical solution structure as opposed to a tetrahedral
structure of water.
This highly ordered solution will lead to more and
more ordered collisions and, hence, less polycrystal formation.
While
these arguments are related to nucleation from bulk solution, the same
arguments obtain when considering the nature of the adsorption layer at
the surface of a growing crystal where secondary nucleation occurs.
The decrease in the sodium content of KCl crystals in the presence
"H2
H~
of Mg
ion can be attributed to the change in Na -water interactions
and to blocking the Na"*" from reaching KCl sub-nuclei by the hydrated
74
which is consistent with the separation of KCl sub-nuclei suggested
earlier.
75
CONCLUSIONS
The analysis of steady state crystal size distributions obtained
from an evaporative vacuum MSMPR crystallizer indicated that such
system can be used to study the effect of impurities on the crystal­
lization of KCl.
The data from the experiment showed polycrystal formation in
crystallization of KCl from solution containing KCl only.
Addition o
NaCl or MgCl2 to KCl solution slightly reduced polycrystal formation
and polycrystal formation was reduced markedly when both NaCl and
MgCl^ were added to KCl solution.
The sodium content of the KCl crystals obtained was very small and
increased linearly with increasing the concentration of NaCl in the
solution.
The magnesium content in KCl crystals, if any, was less
•;han 0,007%.
The Na"^ content of KCl crystals, however, was reduced
almost 50% in the presence of some concentrations of MgClg.
The photograph taken under the electronic microscope, with magnifica­
tion ranges from 30-3000 X, showed that these polycrystals were
formed together and had grown together.
The above observations with other previous observations led to the
conclusion that sub-nuclei of KCl were present in the solution and
these were probably the growth units by which the crystals grew.
*4"
-jThe interaction of the different physical properties of Na , K ,
+2
Mg
in aqueous solutions was the reason behind the reduction in
polycrystal formation when all were present as well as the reduction
of the sodium content of KCl crystals obtained.
76
7.
The comparison between the growth and the nucleation rate obtained
from the different systems was unjustified due to polycrystal
formation.
The growth and nucleation rates obtained from KCl
solutions containing NaCl and MgCl2 are the most useful due to
noticeable reduction in polycrystals and because KCl is produced
industrially from solutions containing these compounds.
For this
system, the growth rate is 7.5 mm/min and the nucleation rate is
of the order of 1x10
3
no./minrlOO ml.
77
RECOMMENDATIONS
1.
The problems involved in the construction and operation of a vacuuûi
evaporative crystallizer can be minimized by employing a level
controller which is not affected by splashes of liquid and by using
a better cooling system.
2.
The investigations which deal with the effect of impurities can be
extended to include the study of the effect of impurities which
accompany many materials in their natural mineral states.
3.
The relation between the structure of the solution and the crystal­
lization kinetics should be investigated.
X-ray and Raman spectra
techniques provide handy tool in performing such investigation.
4.
The effect of anions, cation, transition metal ions and complex
ions on the structure of electrolyte solutions should be studied.
Such studies may be helpful in predicting the effect of such ions
on the crystallization kinetics of the different materials.
78
BIBLIOGRAPHY
1.
Abegg, C., J. Stevens and M. A. Larson. 1968. Crystal size
distribution in continuous crystallizers when growth rate is size
dependent. AIChE J. 14:118-122.
2.
Amsler, J. 1942. Versuche liber die Keimbildung in ubersattigten
Losungen. Helv. Phys. Acta 15:699-732.
3.
Bahe, L. W. 1972. Structure in concentrated solutions of electro­
lytes. J. Chem. Phys. 76:1062-1071.
4.
Baliga, J. B. 1970. Crystal nucleation and growth kinetics in
batch evaporative crystallization. Ph.D. Thesis. Library, Iowa
State University of Science and Technology, Ames, Iowa.
5.
Bamforth, A. W.
Hill, London.
6.
Bauer, L. G., M. A. Larson and V. J. Dallons. 1974. Contact
nucleation of MgSO^'TH^O in a continuous MSMPR crystallizer.
Chem. Eng. Sci. 29:1253-1261.
7.
Beck, I. 1939. Konstruktion einer 10-kW-Derhanoden-Rontgenrohre
USW. Phys. Z. 40:474-483.
8.
Bennema, P. 1969. The importance of surface diffusion for crystal
growth from solution. J. Crystal Growth 5:29-43.
9.
Bennema, P. 1974.
-T Tnô*> +-
T
1965.
o *• o 1
Industrial crystallization.
Leonard
Crystal growth from solution-theory and experV» 0^. /
10.
Bennema, P. 1979. Progress in crystal growth from solution.
Pages 115-134 in E. J. deJong and S. J. Jancic, eds. Industrial
crystallization. North-Holland Publishing Co., New York.
11.
Bennet, R. C., H. Fiedelman and A. D. Randolph. 1973.
influenced nucleation. Chem. Eng. Prog. 69(7):86-93.
12.
Berglund, K. A. 1981. Formation and growth of contact nuclei.
Ph.D. Thesis. Library, Iowa State University of Science and
Technology, Ames, Iowa.
13.
Bemal, J. D. and R. H. Fowler. 1933. A theory of water and
ionic solution. J. Chem. Phys. 1:515-548.
14.
Berthoud, A. 1912. Theorie de la formation des faces d'un
crystal. J. de Chimie Physique 10:624-636.
Crystallizer
79
15.
Bjerrum, N. 1926. Untersuchungen iiber loneassoziation. Kgl
danske videnskab. Selskab. 7:1-9.
16.
Boistelle, R. 1976. Survey of crystal habit modification in
solution. Pages 203-214 ^J. W. Mullin, ed. Industrial crystalliza­
tion. Plenum Press, New York.
17.
Boistelle, R. and B. Simon. 1974. Epitaxies de CdCl2.2NaCl.3H2O
sur les faces (100). (110) et (111) des cristaux de chlorure de
sodium. J. Crystal Srowth 26:140-146.
18.
Bokris, J. 0. and A. K. Reddy.
Plenum Press, New York.
19.
Botsaris, G. D. 1976. Secondary nucleation—A review. Pages
3-22 ^J. W. Mullin, ed. Industrial crystallization. Plenum
Press, New York.
20.
Botsaris, G. D., E. A. Mason and R. C. Reid. 1967. Incorporation
of ionic impurities in crystals growing from solution. AIChE J.
13:764.
21.
Bourne, J. R. and R. J. Davey. 1976. Solvent effect in the
growth of hexamethylene tetramine. Pages 232-238^ J. W. Mullin,
ed. Industrial crystallization. Plenum Press, New York.
22.
Brady, G. W. 1958.
Phys. 28:464-469.
23.
Braitsch, 0. 1971. Salt deposits - Their origin and composition.
Springer-Verlag, New York.
24.
Bransom, S. H.
crystallizers.
25.
Braunstein, J. 1971. Statistical thermodynamics of molten
salts and concentrated aqueous electrolytes. Pages 179-26
S. Petrucci, ed. Ionic Interactions, Vol. I. Academic Press,
New York.
26.
Buckingham. A. D. 1957. A theory of ion-solvent interaction.
Disc. Faraday Soc. 24:151-157.
27.
Bunn, C. W. and H. Emmet. 1949.
Disc. Faraday Soc. 5:119-114.
28.
Burton, W. K., N. Caberra and F. C. Frank. 1951. The growth
of crystal and equilibrium structure of their surfaces. Phil.
Trans. Roy. Soc. London, A 243:299.
1970.
Modem electrochemistry.
Structure in ionic solutions.
J. Chem.
1960. Factors in the design of continuous
Br. Chem. Eng. 5:838.
Crystal growth from solutions
80
29.
Canning, T. P. and A. D. Randolph. 1967. Some aspect of
crystallization theory: System that violates McCabe Delta L
Law. AIChE J. 13:5-9.
30.
Cise, M. D. and A. D. Randolph. 1972. Secondary nucleation of
potassium sulphate in a continuous-flow seeded crystallizer.
AIChE Symp. Ser. .68(121):42-56.
31.
Clontz, N. A. and W. L. McCabe. 1971. Contact nucleation of
magnesium sulphate. Chem. Eng. Prog. Symp. Ser. 67:6-17.
32.
Cooke, E. G. 1966. The effect of additives on the crystal form
of sodium chloride. Pages 259-268^ the Second Symp. on Salt.
The Northern Ohio Geological Society, Inc. , Cleveland, Ohio.
33.
D'Ans, J. 1933. Die Losungsgleichgewichte der Systeme der Salze
ozeanischer Salzablagerungen. Verlagsgesell-Schaft fur Ackerbau
M.B.H., Berlin.
34.
Davey, R. J. and J. W. Mullin. 1974. Growth of the faces of
ammonium dihyrogen phosphate crystals in the presence of ionic
species. J. Crystal Growth 23:89.
35.
Debye, P. and E. Huckel.
Phys. Z. 24:185-208.
36.
Denk, E. G. and G. D. Botsaris. 1972. Fundamental studies in
secondary nucleation in solution. J. Crystal Growth 13/14:493.
37.
Eley, D. D. and M. G. Evans. 1938. Heat and entropy changes
accompanying the solution of ions in water. Trans. Faraday Soc.
34:1098-1112.
38.
Epstein, J. A. 1977. The recovery of potash from the Dead Sea.
Chemistry and Industry. July: 572-570.
39.
Epstein, Marry Ann, F. and L. Sowul. 1980. Continuous sucrose
crystallization with a laboratory-scale evaporative crystallizer.
AIChE Symp. Ser. 76(193): 73-81.
40.
Frank, F. C. 1949. The influence of dislocations on crystal
growth. Disc. Faraday Soc. 5:48-54.
41.
Frank, H. S. 1966. Solvent models and the interpretation of
ionization and solvation phenomena. Pages 53-67 in B.E.
Conway and R. G. Barrads, eds. Chemical physics of ionic
solution. John Wiley and Sons, Inc., New York.
1923.
Zur Theorie clektrolyte.
81
42.
Frank, H. S. and P. T. Thompson. 1959. A point of view on ion
clouds. Pages 118-135^ W. Hammer, ed. The structure of
electrolyte solutions. John Wiley and Sons, Inc., New York.
43.
Friedman, W. L. 1962.
Sons, Inc., New York.
Ionic solution theory. John Wiley and
44. Fuoss, R. M. 1935. Properties of electrolytic solutions.
Rev. 17:27-42.
Chem.
45.
Garabedian, H. and R. F. Strickland-Constable, 1972. Collision
breeding of crystal nuclei: Sodium Chlorate II. J. Crystal
Growth 12:53-56.
46.
Garrett, D. E. 1966. Overgrowth during industrial crystalliza­
tion. Chem. Eng. Prog. 62(12):74-80.
47.
Garside, J. 1976. Kinetics of crystallization from solution.
Paper presented at the first European Conference on crystal
growth, Zurich, 12-18 September.
48.
Garside, J. and R. J. Davey. 1980.
Chem. Eng. Commun. 4:393-424.
49.
Garside J. and M. A. Larson. 1978. Direct observation of secondary
nuclei production. J. Crystal Growth 43:694-704.
50.
Garside, J., I. T. Rusli and M. A. Larson. 1979. Origin and size
distribution of secondary nuclei. AIChE J. 25:57-63.
51.
Genck, W. J. 1969. Temperature effects on growth and nucleation
rates in mixed suspension crystallization. Ph.D. Thesis.
Library, Iowa State University of Science and Technology, Ames,
Iowa.
52.
Ghosh, J. C. 1918. The abnormality of strong electrolytes.
J. Chem. Soc. 113:449-458.
53.
Glucckauf, E. 1969. Electrostatic interactions in electrolyte
solutions. Proc. Royal Soc. Ser, 310:449-462.
54.
Hadzeriga, P. 1966. Considerations on the recovery of potash
from sea water bitterns. Pages 204-211
Second Symp. on Salt.
The Northern Ohio Geological Soc. Inc., Cleveland, Ohio.
Secondary contact nucleation.
55. Hamed, H. S. and B. B. Owen. 1958. The physical chemistry of
electrolytic solutions. Reinhold, New York.
82
56.
Hester, R. E. and R. A. Plane. 1964. Raman spectrophotometric
study of complex formation in aqueous solutions of calcium
nitrate. J. Chem. Phys. 40:411-414.
57.
Hester, R. E. and C. W. Scaife. 1967. Vibrational spectra of
molten salts. J. Chem. Phys. 47:5253-5258.
58.
Hunt, J. P. 1965.
Inc., New York.
59.
Irish, D. E. and A. R. Davis. 1968. Interactions in aqueous
alkali metal nitrate solutions. Can. J. Chem. 46:943-953.
60.
Jackson, K. A. 1958. Mechanisms of growth. Page 174 in Liquid
metals and solidification. Am. Soc. for Metals, Cleveland, Ohio.
61.
Janse, A. M. and E. J. De Jong. 1976. The occurrence of growth
dispersion and its consequences. Pages 145-154
W. Mullin,
ed. Industrial crystallization. Plenum Press, New York.
62.
Johnson, R. T., R. W. Rouseau and W. L. McCabe. 1972. Factors
affecting contact nucleation and growth rates for crystals in
supersaturated solutions. AIChE Sym. Ser. 68(121):31.
63.
Kaufman, E. L. 1980. Photographic observation of secondary
nucleation. M. S. Thesis. Library, Iowa State University of
Science and Technology, Ames, Iowa.
64.
Kavanau, J. L. 1964. Water and solute-water interactions.
Holden-Day, San Francisco, California.
55.
IChambaty, S. and M. A. Larson. IS73. Crystal regeueraclon and
growth of small crystals in contact nucleation. Ind. Eng. Chem.
Fundam. 17(3):160-165.
66.
Khamskii, E. V. 1969. Crystallization from solutions.
sultant Bureau, New York.
67.
Khamskii, E. V. 1979. The role of impurities in crystallization.
Pages 105-114 ^ E. J. de Jong and S. J. Jancic, eds. Industrial
crystallization. North-Holland Publishing Co., New York.
68.
Kruus, P. 1977. Liquid and solutions, structure and dynamics.
Marcel Dekker, Inc., New York.
69.
Lai, D. P., R.E.A. Mason and R. F. Strickland-Constable. 1969.
Collison breeding of crystal nuclei. J. of Crystal Growth 5:1-8.
70.
Larson, M. A. and L. L. Bendig. 1976. Nuclei generation from
repetitive contacting. AIChE Symp. Ser. 72(153):21-27.
Metal ions in aqueous solution.
W. A. Benjamin
Con­
83
71.
Larson, M. A. and J. VJ. Mullin. 1973. Crystallization kinetics
of ammonium sulphate. J. Crystal Growth 20:182-191.
72.
Larson, M. A., D. C. Timm and P. R. Wolff. 1968.
suspension density on crystal size distributions.
14:448-452.
73.
Levine, I. N.
New York.
74.
Liu, Y. A. and G. D. Botsaris. 1973. Impurities effects in con­
tinuous flow mixed suspension crystallizer. AIChE J. 19:
510-516.
75.
McCabe, W. L. 1929.
Eng. Chem. 21:30-33.
76.
Melia and W. P. Moffitt. 1964. Secondary nucleation from aqueous
solution. Ind. Eng. Chem. Fund. 3:313-317.
77.
Metchis, S. G. 1974. Nucleation kinetics of the potassium
chloride-water system with predictions of crystal-size distributions
in crystallizers of complex configuration. M. S. Thesis. Library,
University of Arizona, Tucson, Arizona.
78.
Miers, H. A. and F. Isaac. 1906. Refractive indices of
crystallizing solutions. J. Chem. Soc. 89:413.
79.
Mullin, J. W.
80.
Mullin, J. W. 1979. Crystal growth in pure and inpure systems.
Pages 93-104^ E. J. de Jong and S. J. Jancic, eds. Industrial
crystallization. North-Holland Publishing Co., New York.
81.
Newman, H. M. and R. C. Bennett. 1959. Circulating magma
Crystallizers. Chem. Eng. Prog. 55(3):65-70.
82.
Nielsen, A. E. 1964.
Co., New York.
83.
Moyes, R. 1966. Potash and potassium fertilizers. Chemical
process monograph 15. Noyes Data Corporation, Park Ridge, NJ.
84.
Ohara, M. and R. C. Reid. 1973. Modeling crystal growth rates
from solution. Prentice-Hall Inc., Englewood Cliffs, NJ.
85.
Powers, H. E. C. 1963.
Ind. Chem. 39:351-355.
1978.
1972.
Effect of
AIChE J.
Physical chemistry. McGraw-Hill Book Co.,
Crystal growth in aqueous solutions.
Crystallization.
Ind.
Butterworths, London.
Kinetics of precipitation.
The Macmillan
Nucleation and early crystal growth.
84
36.
Prigogine, I. 1939. Contributions à la Theorie des electrolytes
forts. Gauthier-Villars, Paris.
87.
Prins, J. A. and R. Fonteyne. 1935. X-ray evidence on the
ionic arrangement in thorium nitrate solution. Physica 2:
570-574.
88.
Randolph A. D., Beckraan, J. R. and Z. I. Kraljevich. 1977.
Crystal size distribution dynamics in a classified crystallizer.
Part I. AIChE J. 32:500-510,
89.
Randolph, A. D. and M. D. Cise. 1972. Nucleation kinetics of
the potassium sulphat-water system. AIChE J. 18:798-807.
90.
Randolph, A. D. and M. A. Larson. 1971.
processes. Academic Press, New York.
91.
Rasiah, J. C. and H. L. Friedman. 1968. Integral equation
methods in the computation of equilibrium properties of ionic
solutions. J. Chem. Phys. 48:2742-2752.
92.
Robinson, R. A. and R. H. Stokes.
Butterworths, London.
93.
Rosen, H. N. 1969. The crystallization of potassium sulphate:
nucleation and growth. Ph.D. Thesis. Northwestern University,
Evanston, Illinois.
94.
Rosen, H. N. and H. M. Hulburt. 1971.
crystallization of potassium sulphate.
Ser. 67(110):18.
95.
Rusli, I. T., M. A. Larson and J. Garside. 1980. Initial growth
of secondary nuclei produced by contact nucleation. Chem. Eng.
Symp. Ser. 76(193):52-58.
96.
Samoilov, 0. Y. 1965. Structure of aqueous electrolyte solutions
and the hydration of ions. Consultants Bureau, New York.
97.
Sarig, S., C. L. Leci and N. Eidelman. 1979. Agglomeration of
potassium chloride crystals from supersaturated aqueous solution,
J. Crystal Growth 47:365-372.
98.
Sauchelli, V. 1960. Chemistry and technology of fertilizers.
Reinhold Publishing Corp., New York.
99.
Schultz, J. W. and D. F. Hornig. 1961. The effect of alkali
halides on the Raman spectrum of water. J. Phys. Chem. 65:
2131-2138.
1968.
Theory of particulate
Electrolyte solutions.
Continuous vacuum
Chem. Eng. Prog. Symp.
85
100.
Shor, S. M. and M. A. Larson. 1971. Effect of additives on
crystallization kinetics. Cham. Eng. Prog. Symp. Ser. 67(110).
32-42.
101.
Skryschewski, A. F. 1954. The application of integral analysis
of intensity curves to the study of the structure of binary
solutions. Dissertation. Institute of Metal Physics, Kiew.
102.
Strickland-Constable, R. F. 1968. Kinetic and mechanism of
crystallization. Academic Press, London.
103.
Strickland-Constable, R. F. and R. E. A. Mason. 1966.
of crystal nuclei. Trans. Faraday Soc. 62:455.
104.
Sung, C. Y., J. Estrin and G. R. Youngquist. 1973. Secondary
nucleation of magnesium sulphate by fluid shear. AIChE J. 19:
957-962.
105.
Svoronos, D. M. 1969. Sur la cristallization du chlorure de
potassium en solution aqueuse. C. R. Acad. Sci. Paris, Series
C 269:133-136.
106.
Timm, D. C. 1967. Effect of residence time and suspension
density on crystal size distribution in continuous crystallizers.
Ph.D. Thesis. Library, Iowa State University of Science and
Tehcnology, Ames, Iowa.
107.
Vaslow, F. 1972. Thermodynamics of solutions of electrolytes.
Pages 465-518
A. Home, ed. Water and aqueous solutions.
Wiley-Interscience, New York.
105.
Volmer. K. 19 39. Kinetik der Phasenbildung. Sreinkopff.
Dresden and Leipzig.
109.
Youngquist, G. R. and A. 0. Randolph. 1972. Secondary nuclea­
tion in Class II System. AIChE J. 18:421-429.
Breedings
86
ACKNOWLEDGMENTS
The author wishes to thank Dr. M. A. Larson for his guidance and
assistance throughout the course of this work.
A deep feeling of gratitude is extended to my wife for her patience,
understanding, and help during the course of this work.
The author wishes to thank all those who contributed to his
education, and especially to Dr. D. L. Ulrichson.
Discussion with Dr. J. H. Espenson, Dr. R. S. Hansen, and Dr.
D. K. Hoffman were useful and appreciated.
Appreciation is also extended to Iowa State Engineering Research
Institute and University of Jordan for their financial support.
87
APPENDIX A
Experimental results for the runs of which the feed solution had been
analyzed. These results are based on the best straight line -vfhich fits
each set of data
Run
T
•"
/ . \
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
n° X 10 4
G X 10^^
, g
\
nOO ml^
.
no
^nnn-lOO ml-'
.im.
18.7
17.8
18.8
18.5
19.3
18.8
1.2620
2.1205
3.6514
0.8609
1.4712
2.9388
19.737
11.743
1.8053
22.716
3.5337
2.5897
22.6
0.9707
0.6868
0.8324
0.4909
1.4781
1.3316
17.813
6.3624
44.206
10.501
13.836
17.322
12.773
6.4898
287
8.8295
927
990
1375
959
19.3
19.6
20.5
18.2
20.0
Bad run
18.7
17.0
18.1
20.1
18.3
21.8
16.5
=
2.009
1.1991
2.1110
2.4788
1.4521
5.0963
3.3041
.7431
0.6415
0.8362
1.1271
2.0002
0.4987
1.2113
6.9494
3.4421
13.5309
6.0882
11.866
7.2237
7.159
7.9385
7.4337
15.8423
12.0555
14.888
14.4742
10.69
9.5941
10.0592
based on total solid content of the sample.
B°
^
no
^
min -100 ml
1372
404
244
219
420
187
1133
178
77
125
163
213
48
122
88
APPENDIX B
Experimental results for the runs of which the feed solution had been
analyzed. These results are based on the best straight line which fits
the data for crystal size less than 500 micron
Run
T
JL
#
/
\
(mm)
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
18.7
17.8
18.8
18.5
19.3
18.8
22.6
19.3
19.6
20.4
18.2
20.0
Bad Run
18.7
17.0
18.1
20.1
18.3
21.8
16.5
n° x 10~^
/
ê
(loo
\
/
^
^mm-lOO ml^
G x 10^
mm
^min^
1.1458
0.3161
0.5911
0.3383
0.4506
0.1620
0.5306
0.4384
0.3608
0.2384
0.7519
0.4672
138.0400
62.1120
6.8508
387.9200
1.8267
9.5265
4.5665
14.9950
3.3869
3.1511
3.9248
4.2076
2.0175
8.0075
5.0491
4.9860
4.5568
7.1485
6.5083
7.3018
6.2487
5.5548
0.2366
0.0276
0.1755
0.156
0.2105
0.0778
n 1 01 ç
1.8418
9.1898
2.2575
1.2579
1.2788
6.4706
6.0090
7.2038
4.1116
11.765
2.1195
7.11^5
6.5724
2.6603
3.4282
3.6243
B° x 10 ^
^
no
^
^min -100 ml^
36.7228
21.379
2.8825
7.8263
1.4628
4.8222
2.2768
7.2879
2.4211
2.0508
2.8658
3.4710
1.1917
5.5222
1.6263
5.1718
1.5045
7.6815
4.6752
= based on solid content for crystals smaller than 600 micron.
89
APPENDIX C
Feed composition in gm/lOO ml saturated solution as analyzed by
the analytical laboratory of the Engineering Research Institute
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
KCl
NaCl
32.2
24.6
17.7
4.06
5.00
3.2
MgCl2.6H20
Comments
No feed sample
14.8
32.2
27
27.3
27.3
27.0
27.0
28.2
36.6
37.18
34.9
33.94
33.7
32.6
32.6
32.6
9.5
13.5
8.3
8.3
8.5
7.6
6.4
6.3
0.81
1.71
1.71
2,47
3.04
3.70
0.87
1.61
2.35
3.29
90
APPENDIX D
Selected Physical Properties
Physical properties of NaCl, KCl, and MgCl- or their ions as reported
by many investigators ((18), (55), (58), (92))
Table Dl.
Physical properties of the salt t 25°C
Physical Property
KCl
NaCl
Dielectric constant for 1 molar
solution
69+2
70+2
Lattice Energies (KCal)
219
168
MgClg
50+2
600
-1
Heat of hydration (KCal mole )
-183
•163.8
Heat of formation (KCal mole )
- 98.2
-104.2
-153.2
Free energy of solution (KCal mole )
-
-
- 30,1
Table D2.
2.2
456
1.2
Physical properties of the ions at 25°C
Physical Property
Na"*"
K+
CI
M
g
Ionic radii (A°)
0.95
1.33
.61
Ionization Potential (ev)
5.138
4.339
7.644
Hydration numbers
(Average)
4
4
14
Heat of hydration
(K:Cal mole"^)
-95
75
-456
-90
Entropies of hydration
(Cal/mol. deg)
-23
14
- 67
-21 (gaseous)
-15.3
• 8.1
- 50.9
-10.,9 (aq.)
1.81
13.01 (1st)
1.0