Crystal Structure Theory and Applications, 2013, 2, 16-27
http://dx.doi.org/10.4236/csta.2013.21003 Published Online March 2013 (http://www.scirp.org/journal/csta)
Kinetics and Mechanism of Spontaneous Crystallization of
Potassium Nitrate from Its Supersaturated Aqueous
Solutions
Oleg D. Linnikov*, Irina V. Rodina, Igor G. Grigorov, Evgeniy V. Polyakov
Institute of Solid State Chemistry, Ural Branch of the Russian Academy of Sciences, Ekaterinburg, Russia
Email: *linnikov@mail.ru
Received October 19, 2012; revised December 3, 2012; accepted December 14, 2012
ABSTRACT
Kinetics of spontaneous crystallization of potassium nitrate from its supersaturated aqueous solutions has been studied
simultaneously by electrical conductance and optical transmittance methods. It was found that spontaneous crystalliza-
tion of potassium nitrate was accompanied by aggregation of crystals. Growth of salt crystals was in the kinetic mode of
the growth process, and was described by the equation of the first order regarding supersaturation of solution. The
mechanism of aggregation and intergrowth of crystals during bulk crystallization via formation of nucleus-bridges be-
tween crystals found earlier for several salts was confirmed. Specific surface energy of potassium nitrate was evaluated
on the basis of the above mechanism of aggregation and intergrowth of crystals. The established value of the specific
surface energy was reasonable and agreed satisfactorily with the available literature data. Examination of crystal deposit
after completion of crystallization allowed detecting crystal agglomerates of freakish and irregular forms, which may be
considered as the direct confirmation of the above mechanism of intergrowth of crystals. Kinetics of crystallization,
aggregation and size distribution of salt crystals after completion of crystallization have been satisfactory described by
the earlier proposed model of the crystallization process. An excellent agreement was established between the experi-
mental data on potassium nitrate solubility in aqueous solutions found in the present work and those available in the
literature.
Keywords: Potassium Nitrate; Aggregation; Crystallization; Solubility; Activation Energy; Crystal Growth;
Intergrowth; Agglomerates
1. Introduction
Very scarce information about crystallization kinetics of
potassium nitrate in aqueous solutions is available in the
literature. There are only a few works where crystalliza-
tion of potassium nitrate from aqueous solutions was
studied [1-3]. So, Helt and Larson [1] investigated the
crystallization of KNO3 using MSMPR crystallizer. The
operation temperature was from 10˚C to 25˚C, and a de-
gree of supersaturation was varied within 0.005 - 0.01.
The order of the crystal growth process was found to be 1
and the activation energy of the growth process was 31
kJ·mol1 [1]. The crystal growth rate ranged in the inter-
val . Later, Graber et al. [2] stud-
ied the growth of seed crystals of sodium and potassium
nitrates from aqueous solutions in a perfectly stirred
batch crystallizer. It was also found that the order of the
crystal growth process was 1 for both salts. The maxi-
mum crystal growth rate was directly proportional to the
seed crystal size. The activation energy of the growth

8
3.5 - 10.710ms

process for potassium nitrate was 24.512 kJ·mol1, and
the process was under diffusion control [2].
The growth rate of different faces of single potassium
nitrate crystals from aqueous solution was studied by
Lacmann et al. with the aid of optical microscopical
method [3]. For some crystals, the surface morphology
was determined by means of light microscopy, scanning
electron microscopy, Michelson interferomtry or scan-
ning force microscopy [3]. Crystals were grown under
constant supersaturation at 15˚C - 20˚C. The relative su-
persaturation was in the range from 0.3% to 1.8%. In-
tense dispersion of the crystal growth rates was observed.
For all crystal faces, the order of the crystal growth
process was between 1.15 and 1.60 [3]. Examination of
the surface morphology showed that crystal faces had a
flat structure; no significant objects were found.
1
The present work deals with spontaneous crystalliza-
tion of potassium nitrate from its supersaturated aqueous
solutions. The objective of this work was to study (in situ)
kinetics of spontaneous crystallization of potassium ni-
trate from its supersaturated aqueous solutions using si-
*Corresponding author.
C
opyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL. 17
multaneously two methods of measurement: 1) electrical
conductivity; and 2) optical transmittance of a super-
saturated solution in salt crystallization.
2. Experimental
2.1. Experimental Setup
The experiments were carried out on a laboratory setup
with a glass crystallization vessel of 250 cm3 capacity.
The volume of a supersaturated solution of KNO3 was
165 cm3. The laboratory setup was analogous to that used
in [4], and it allowed to conduct simultaneously and
automatically computer records of electrical conductivity
W and optical transmittance I of the supersaturated solu-
tion during salt crystallization. The concentration of the
salt during crystallization was determined from electrical
conductivity of the solution using calibration curves. The
principal diagram of the laboratory setup is shown in
Figure 1.
2.2. Experimental Procedure
Supersaturation in investigated salt solution was created
by thermic method. For this, supersaturated solution of
KNO3 in the crystallization vessel was heated on 5˚C -
10˚C above the saturation temperature in order to dis-
solve small salt crystals, and then it was stayed at this
temperature about 0.5 h. After that the solution was
cooled to the temperature below the saturation tempera-
ture. When the required temperature was achieved, a
magnet stirrer was turned on, and spontaneous crystalli-
zation of KNO3 started after a certain time (induction
period). Simultaneously, the electrical conductance and
optical transmittance of the solution during salt crystalli-
zation were recorded automatically.
After the crystallization was completed, the stirrer was
switched off and a variation in optical transmittance of
the solution was registered during sedimentation of
formed potassium nitrate crystals. Analogous curves of
variation of optical transmittance of solutions during
sedimentation are used in photo-sedimentation method
for determination of particle-size distributions of pow-
ders [5]. Therefore, the obtained curve also was used to
calculate the particle-size distribution of the formed salt
crystals, and for determination of the optical absorption
constant Kc of the salt [5].
After sedimentation, the most part of the solution was
decanted, and salt crystals were rapidly (during 3 - 5 s)
filtered under vacuum-suction using a ceramic filter, and
then on the filter immediately were washed with ethanol
saturated by KNO3. Then salt crystals on the filter were
air-dried under vacuum-suction. These operations made
it possible to remove remainders of the solution from
crystals and prevent their intergrowth during air-drying.
The crystalline product formed in each run was imaged
with optical and SE microscopy.
Two liquid samples (5 ml) were collected from the
solution. The first probe was taken from the supersatu-
rated solution at the beginning of the crystallization proc-
ess, and the second one-after the end of the crystalliza-
tion. The samples were evaporated and dried at 80˚C, and
then weighed. The salt concentrations in the solution
were subsequently calculated from the mass of anhy-
drous solid and the volume (5 ml) of samples.
In the process of crystallization, the solution was kept
(thermostated) in the crystallization vessel at constant
temperature (0.1˚C) and vigorously agitated with a
magnetic stirrer. The rotation frequency of the stirrer was
450 rpm. The error in optical transmittance measure-
ments was 1.5%; the root-mean-square deviation of the
measured electrical conductivity was 0.2%.
The solutions were prepared using analytical grade salt
and distilled water. For additional purification, the solu-
tions were recrystallized by thermal evaporation. Ex-
periments were carried out at temperature 16.3˚C - 39.6˚C
in the range of 01. 0 11. 0 6CC
.
The experimental data were mathematically treated
using the following equations [4]:
0
d
d
z
T
S
CKC C
v
 (1)
WabС
(2)
0
0log c
I
DKS
I
 (3)

2
3
0
1
π6
T
ci
D
N
K
CCv



(4)

13
6
π
i
T
CCv
dN

(5)
where d is the average diameter of salt crystals formed
in the solution (m); τ is the time (s); ρT is the density of
salt crystals
3
kg m
; S is the total surface area of
growing salt crystals (m2); v is the volume of the solution
3
m; C0 is the concentration of the saturated salt solu-
tion
3
kg m
; C is the concentration of the supersatu-
rated salt solution at the moment of time
; Ci
is the initial concentration (before crystallization, at

= 0)
of the supersaturated salt solution ; K is the
average coefficient of the growth rate of salt crystals; z is
the order of the salt crystal growth process (usually z = 1 -
2); W is the electrical conductivity of the salt solution in
a relative scale of measurement (mV); a, b are the re-
gression coefficients; D0 is the optical density of the salt
solution in the course of crystallization; I0, I are the opti-
cal transmittance of the salt solution at the initial moment
of time and at moment of time
, respectively (%); Kc is
the optical absorption constant of the salt (m2); N is the

3
kg m
3
kg m
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL.
18
total number of growing salt crystals in the solution; Ψ is
the spherical coefficient of salt crystals [6].
Formula (1) is the well-known empirical equation,
which describes the rate of salt crystallization from su-
persaturated solutions after the end of the induction pe-
riod when nucleation of salt crystals in the solution is
usually completed [2,4,7-9]. Equation (2) describes elec-
trical conductivity of salt solutions in the examined con-
centration range. Formula (3) represents the Wagner’s
law [5], according to which the optical density of a sus-
pension with a small content of the solid phase is directly
proportional to the total surface area S of solid particles
(here, the total surface area of growing salt crystals) con-
tained in the solution. This law is valid when the size of
solid particles exceeds 2 μm. In every run of this study,
the Wagner’s law was valid. This was verified in special
runs. The optical absorption constant Kc in Equation (3)
was determined at sedimentation of formed salt crystals
after the crystallization was completed.
Equations (4) and (5) make it possible to calculate the
total number and the average diameter of the crystals
formed during crystallization [4]. So, if m is the mass of
salt precipitated from solution and N is the total
number of salt crystals in the solution, then the average
mass mc of one salt crystal is

kg
c
mmN and the aver-
age volume of one salt crystal vc is ccT
vm
. There-
fore, under assumption that the shape of the crystal is
nearly spherical, its average diameter is determined by
the equation

13
c
6πdv (6)
From this, the average surface area of one salt crystal
S is

23
2
ππ6π
c
Sd v (7)
Expressing vc in terms of a change in the solution
concentration during crystallization and the total number
N of formed salt crystals, we obtain formula (5).
Often in the description of processes occurring in solu-
tions (crystallization, sedimentation, etc.) it is convenient
to consider crystals as ball-shaped particles. However,
the shape of most crystals differs sharply from a sphere.
In this case a correction factor (a spherical coefficient)
should be introduced [6]:
s
c
S
S
 (8)
where Sc is the surface area of a crystal (m2); Ss is the
surface area of a sphere calculated under condition that
the volume of the sphere equals to the volume of the
crystal (m2).
The spherical coefficient Ψ makes it possible to con-
sider crystals of different shapes as ball-shaped particles.
With account of the fact that SSN, Equation (3)
may be transformed by using formulas (5)-(8) into

23
2
0
0
6
ππ
log π
i
cc c
T
CCv
INdN
DKSKK
IN

 


(9)
Then we derive Equation (4) from expression (9).
Integration of Equation (1) for the simplest case of z =
1 and the subsequent joint solution with Equations (2)
and (3) gives the average coefficient of the growth rate of
salt crystals

0
00 0
ln
c
Ti
K
vC
KDCC


C
(10)
where τ0 is the time, after which only crystal growth
process occurs in the solution (nucleation of crystals is
completed or negligible); iis the concentration of the
supersaturated salt solution at the moment of time τ0
(kg·m3).
C
As was mentioned above, the crystal-size distribution
was found when the crystallization was completed using
curves of variation of optical transmittance of solutions
during sedimentation of formed salt crystals. The diame-
ters of the formed crystals during their sedimentation
were calculated using the equation [5,6]:

18
0.843log 0.065
i
Ti
H
dg


(11)
where μ is the dynamic viscosity of the saturated salt
solution (Pa·s); ρ is the density of the saturated salt solu-
tion (kg·m3); H is the sedimentation height (see Figure
1) of crystals (m); τi is the time required for sedimenta-
tion of crystals having the diameter . The viscosity
and density of the saturated salt solutions were measured
in special runs using an efflux viscometer and a set of
areometers.

s
i
d
Figure 1. Principal diagram of a laboratory setup: 1—
crystallization vessel containing supersaturated salt solution;
2—electric heater; 3—thermometer probe; 4—thermo re-
gulator; 5—magnetic stirrer; 6—platinum conductometric
electrodes; 7—electrical conductance meter; 8—light; 9—
light-sensitive device; 10—amplifier; 11—analog-digital con-
verter; 12—computer; 13—thermometer; 14—cover; H-sedi-
mentation height.
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL. 19
3. Results and Discussion
3.1. Spherical Coefficient
To determine the spherical coefficient, 514 images of
potassium nitrate crystals formed in the present experi-
ments were measured. We could not make microphoto-
graphs of crystals in all dimensions and, therefore, we
had no possibility to measure the size (height) of formed
potassium nitrate crystals perpendicularly to a plane of
photos. So for its evaluation we considered the parame-
ters of the unit cell for a modification of KNO3 with or-
thorhombic structure. There is a following basis for this.
First, although in reality the crystallographic cell does
not define the crystal shape of crystal, but relations be-
tween dimensions of a crystal often are approximately
like in the cell (consider, for example, such salts like
NaCl or KCl). Second, at aggregation of crystals at bulk
crystallization there is always a morphological selection
of formed agglomerates of crystals. Most of aggregates
have a correct form with low hydrodynamic resistance,
and what is more, the most agglomerates look like mono-
crystals although in actual fact they are pseudo-mono-
crystals. Hence, for these agglomerates the relation be-
tween their dimensions is the same like for individual
crystal.
At measurements the crystals were imagined as paral-
lelepipeds, and a ratio between length and width of them
was calculated. The ratio found was 1.73 and was close
to its theoretical value 1.43 for the parameters of the unit
cell for modification of KNO3 with orthorhombic struc-
ture (а = 0.54119 nm, b = 0.91567 nm, с = 0.64213 nm
[3], 1.43bc). As was mentioned above, it was not
possible to measure the height of formed potassium ni-
trate crystals, so for calculations of the spherical coeffi-
cient, the ratio was used, it was
close to the above structural parameters. The calculated
spherical coefficient was 0.77. This value was used in all
further calculations. It should be noted, however, that,
taking into account accepted above assumption about
relations between dimensions of salt crystal, this calcu-
lated value of the spherical coefficient was approximate.
::0.8 :1.73 :1abc
3.2. Initial Period of Spontaneous Crystallization
As was mentioned above, the initial supersaturation of
KNO3 solutions was very small. Therefore, spontaneous
crystallization of potassium nitrate was accompanied by
a long induction period. Typical experimental curves
describing the variation of the optical transmittance I and
electrical conductivity W of the salt solution in the course
of the crystallization are shown in Figure 2.
As is seen from Figure 2, in the initial period of crys-
tallization, the optical transmittance of the solution
slightly changes, whereas the electrical conductivity re-
mains unchanged during about 70 s. Then some dramatic
change in the optical transmittance and electrical con-
ductivity of the solution occurs after about 80 and 70 s,
respectively. This corresponds to an induction period of
the salt crystallization. Different duration of the induc-
tion period for I (about 80 s) and W (about 70 s) reflects
different sensitivities of the employed measurement
techniques.
3.3. Variation of the Average Diameter and the
Total Number of Salt Crystals during
Crystallization
Typical curves of the variation of the average diameter
and the total number of salt crystals in solution during
crystallization are shown in Figure 3. The curves were
calculated from experimental data using Equations (4)
and (5).
It is seen from Figure 3 that the calculated depend-
ences of the average diameter and the total number of
salt crystals in the solution change non-monotonously
with time during crystallization. According to these cal-
culations in the initial period of the crystallization, the
salt crystals are large, and their average diameter is about
140 μm. And there are only a few salt crystals in the so-
lution. Then a drastic reduction in the average diameter
of salt crystals takes place (in 100 s), and the total num-
ber of salt crystals in the solution sharply increases. But
after about 120 s the average diameter of salt crystals
Figure 2. Electrical conductivity W (in a relative scale of
measurement) and optical transmittance I of the salt so-
lution during spontaneous crystallization (run 1) vs. time.
Figure 3. Calculated dependences of the average diameter
d and the total number of salt crystals N during crystal-
lization (run 1) vs. time.
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL.
20
begins to increase and the total number of salt crystals in
es in Figure 3 (to
12
ter 120 s calculations
w
3.4. Solubility of Potassium Nitrate
us to ascertain
4 it is seen that the experimental results
fo
tt (12)
the solution monotonously decreases.
Obviously, the initial parts of curv
0 s) are in contradiction with the modern experimental
and theoretical data on crystallization, and this is a result
of incorrect calculations with formulas (4) and (5). In
fact, salt crystals cannot be so large at the beginning of
the crystallization process and then to decrease during
crystallization. The obtained dependences may be ex-
plained by understated values of optical transmittance of
solution in the initial stage of salt crystallization (to 120
s). As was mentioned above, Equation (3) is valid when
the size of solid particles in solution exceeds 2 μm. In
this study, the size of salt crystals formed in the initial
period of crystallization was very small and, as a result of
diffraction of light, the optical transmittance of solutions
was also very small (understated). Therefore, the calcula-
tions by means of expressions (4) and (5) were incorrect.
A similar effect during spontaneous crystallization of
salts was observed also in [4,10].
From Figure 3 it is seen that af
ith formulas (4) and (5) give “true” values of d and N.
From this moment, the average diameter of salt crystals
increases monotonously, and their total number in the
solution decreases. This indicates that after 120 s only
crystal growth process takes place in the solution (nu-
cleation of crystals is completed or negligible). Therefore,
these experimental points may be used to determine the
growth parameters of salt crystals. Moreover, the reduc-
tion in the total number of salt crystals formed in the so-
lution demonstrates that aggregation of crystals takes
place during crystallization. N is largest at the beginning
of the crystallization (after 120 s, Figure 3). Earlier, the
aggregation of crystals was observed during spontaneous
crystallization of other salts [4,10,11], but their induction
periods were short and the largest total number of salt
crystals was in solutions immediately after the comple-
tion of the induction period. Aggregation of potassium
nitrate crystals during bulk crystallization was also found
in [12].
The obtained experimental results allow
the data on the solubility of potassium nitrate in aqueous
solutions. All obtained experimental results on the solu-
bility and the data of other authors are visualized in Fig-
ure 4.
From Figure
und in the present work agree excellently with the data
of [13]. In the range 0˚C - 40˚C the solubility of KNO3 in
water and the density of the saturated salt solution may
be described by the empirical equations (correlation co-
efficients 0.999):
C
0132.17 2.52ln
Figure 4. Effect of temperature on solubility of potassium
nitrate in aqueous solution.
1045.75.8297 t
 (13)
where t is the temperature (˚C).
z = 1 gives
3.5. Growth of Crystals
Integration of Equation (1) at
0
ln T
CC SK
0
0
iv
CC
(14)
and at z = 2


0
00
11T
i
SK
СС v
CC
 
(15)
These two equations are usually used at dete
of th
rmination
e order of crystal growth process at bulk crystalliza-
tion. Examination of the obtained experimental data on
crystallization of KNO3 showed that in all runs the crys-
tallization process gives best possible fit with Equation
(14). The typical kinetic dependences of one and the same
run in the coordinates of Equations (14) and (15) are dis-
played in Figure 5.
It is seen that the experimental points are well plotted
on a straight line only in the coordinates of Equation (14)
(see Figure 5(a)). In the coordinates of Equation (15) a
curvilinear relation takes place (see Figure 5(b)). Calcu-
lation of the average coefficient of the growth rate of salt
crystals with Equation (10) showed that K was constant
during the crystallization process. This also confirms the
first order
1z
of Equation (1). The linear depend-
ence of natural-logarithmic fitting for the obtained aver-
age coefficients K of the growth rate of KNO3 crystals
versus the inverse temperature 1T according to the
well-known Arrhenius equation is graphically presented
in Figure 6.
It is seen a great scattering of experimental points is
observed in Figure 6. Probably this is a result of some
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL. 21
Figure 5. Experimental data on crystallization of one and
the same run (run 1) in the coordinates of Equation (1) after
integration at z = 1 - 2; a z = 1 [Equation (14)]; b z = 2
[Equation (15)].
10
3
/T, K-
1
Figure 6. Natural-logarithmic fitting of the obtained values
of the average coefficients K of the growth rate ofNO3
ld be noted also that only the
verage coefficients K of the growth rate of KNO3 crys-
K
crystals against the inverse temperature 1/T according to
the Arrhenius equation.
experimental errors. It shou
a
tals for every run are presented in Figure 6. But the total
number of experimental points in all experiments was
4134. From Figure 6 one may find that variations with
temperature of the average coefficient of the growth rate
of KNO3 crystals is described by equation:

3
63.5 3.210
3547.2 expKRT





(16)
where R is the universal gas constant (J·K1mol1); Т is
the temperature (K).
The correlation factor of Equation (16) is
Fisher test showed that Equation (16) adequately de-
tion energy of the growth process
of
0.52. The
scribed the experimental data. From Equation (16) one
can see that the activa
KNO3 crystals is significantly higher than those de-
rived in [1,2], it is quite high and has reasonable value
1
63.5 kJmol
(the standard deviation of the activation
energy is 1
1.61 kJmol
). Such high value of the activa-
tion energy is characteristic of the kinetic mode of the
growth process of salt crystals [8]. In work [2] it was
ystal growth process was under diffusion
control. Thay be concluded that growth of
KNO3 crystals in the experimental conditions of the pre-
sent work was in the kinetic mode.
3.6. Kinetics of Aggregation and Mechanism of
Intergrowth of Salt Crystals
The kinetic of aggregation of crystals
found that cr
erefore, it m
during spontaneous
partif the equation for
crystallization of KNO3 may be formally described by
the Smoluchowski equation for coagulation of colloidal
cles [14]. The differential form o
aggregation of crystals looks as
2
dd
a
NK
N
 (17)
where Ka is the aggregation constant.
A typical curve of variation of the total number of
crystals in solution during the spo
of potassium nitrate in one of the runs is shown in Figure
rdinates of trans-
fo
ap
ntaneous crystallization
7. The data are presented in the coo
rmed Equation (17) after integration (the integration
was performed under the condition that Ka was constant).
As can be seen, at the initial period of the crystallization
the dependence is almost linear. This confirms that Equa-
tion (17) is valid during this stage of the crystallization.
But further a curvilinear dependence takes place. Appar-
ently, the reason of this is consecutive decreasing of the
aggregation constant Ka due to falling of the supersatura-
tion in solution. The analogous curves were also found in
other runs and at crystallization of other salts [4,10,11].
The presented results in Figure 7 allow to find the
values of the aggregation constant Ka at the different time
moments of the crystallization process. So, the experi-
mental points in Figure 7 with high accuracy may be
proximated by the following empirical regression for-
mula:
9
1 64497.476
10 594.09434
N




(18)
Differentiation of expression (18) allows to find the
aggregation constant Ka at any moment of time. I
be noted, however, Equation (18) is an emp
function
t should
irical fitting
YfX, where Y is 1N, and X is τ. So,
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL.
22
pa
me
r-
w
e. So, formation of nucleus-bridges in
th
rameters of expression (18) have no physical sense,
and they should not be compared with Equation (17).
In works [4,15] was found that aggregation and inter-
growth of salt crystals into agglorates during bulk
crystallization occur owing to the formation of nu-cleus-
bridges between crystals according to the theory put fo
ard by Polak [16-18]. These nucleus-bridges connect
crystals in aggregates and form chemical bonds between
them (Figure 8).
From Polak’s theory follows that nucleation of nu-
cleus-bridges is more advantageous in comparison with
nucleation in the volume of solution and even nucleation
on its own substrat
e narrow gap between two crystals takes place even in
absence of nucleation in the volume of solution and on
the surface of a crystal. For example, according to classic
nucleation theory the work of formation of a nucleus on
its own substrate is [8]
0
100
200
300
400
500
600
0 100 200 300 400 500 600
, s
- experimen tal
- regression
(1/N)·10
Figure 7. Aggregation kinetics of potassium nitrate crystals
during spontaneous crystallization (run 1) in the coordi-
nates of Equation (17) after integration.
9
2
0
π
ln
s
T
Ma
AC
RT C



(19)
where As is the work of formation of a nucleus in the
form of a disc on its own substrate (J); М is the molar
weight of a salt (kg·mol 1); σ is the specific surface en-
ergy (J·m2); a is the parameter of crystal lattice (m).
Polak found that the work Ab of formation of a nu-
cleus-bridge between the nearest faces of connivent
crystals is [16-18]
22
0
4
ln 2
b
T
h
ART C
hMC



(20)
From this one can evaluate relations between the work
of formation of a nucleus on its own substrate As and the
work Ab of formation of a nucleus-bridge between the
nearest faces of connivent crystals:
Figure 8. Scheme of intergrowth of two crystals proposed
by Polak [16-18], where h is the height of a nucleus-bridge
connecting two connivent crystals (m).
0
22 ln
b
Ah C
hR
TC
π1
s
Aa




From Equation (21) it follows that relation
(21)
s
b
A
A is
increasing when supersaturation of solution is decreasing.
Let’s make some evaluations. For example, for
CaSO4·2H2O in [15] was found 9
110mh
 , hence at
32
1710Jm,
 3 10
2330 kgm,4.7510ma
 ,
T
0
298K, 3.05TCC
one can calculate 2.34.
sb
AA
narrow gap Thus, formation of nucleus-bridges in the
between two crystals takes place even in absence of nu-
cleation on surface of crystal.
According to results of [4,15] variation of the aggre-
gation constant Ka during crystallization may be de-
scribed by the expressions:
1
13
2
0
C
1
13
2
0
exp
ln
D
KD
D
C
K
1exp exp
ln
a
D
KD CD
C




















 (22)
2
122
4A
T
hNM
DR
(23)
T
2
2
T
M
DhRT
(24)
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL. 23
30k
DI
(25)
where K1, D1, D2, D3, I0 are constants; h is the height of a
nucleus-bridge connecting two connivent crystals (m);
NA is the Avogadro constant

1
mol;k
is the time
during which crystals are in conllision in
solution (s). It should be noted that the aggregation con-
stant Ka is a function of the aggregate size. But earlier it
was shown that the dependence of Ka on the sizes of
crystals should not be taken into consideration if a stirrer
rate is constant and does not change during the crystalli-
zation process [4].
Values of the aggregation constant Ka, estab
from the experimental data, allow to
So, in Equation (22) there are three u
D1, D2 and
tact at their co
lished
verify Equation (22).
nknown coefficients:
13
·
K
D
D1 and
es no
nce, on
. But at crystallization in isothermal
conditions If
stant and douri
process, he
2con s tD.
t change d
e can take
agitation rate is con-
ng the crystallization
const
k
. Therefore, D3
tant value, and K1·D3
will be also approximately cons
const. After taking the log of expression (22) one can get
the equation useful for following calculations:

1
13
2
0
ln ln
ln
a
D
KKDCD
C

(26)
The known computer program Table Curve 2D v5.00
allows to find these unknown coefficients D1, D2 and
13
K
D for the linear dependence
02
ln1 ln
a
K
CC D



[please see Equation (26)]. Linearity of above equation is
possible only for one found value of D2. Therefore, this
linearity may be considered like some evidence of valid-
ity of Equation (22). Then if from found values of D1 and
D2 will be calculated reasonable values of h and
, and
will be near to the published data for this salt, then it will
be additional evidence of validity of Equation (22).
Results of calculations according to Equation (26) are
shown in Figure 9, where two typical obtained
depend-
ences are presented.
As can be seen, displayed in Figure 9 l
dences (correlation coefficients are in the range 0.75-0.99)
confirm validity of Equation (22). Calculations with the
he
2.11 ints-4202) and
for the height of nucleus-bris (standard
0.67 × 109 number of calculated points-4202).
inear depen-
lp of Equations (23) and (24) give values of 2.19
mJ·m2 for the specific surface energy (standard deviation
mJ·m2, number of calculated po
9
2.46 10m
deviation,
dge
Obviously, if one will take into account all procedures of
these calculations, it is difficult to expect more precise
values. However, the values obtained are reasonable and
agree well with those found in [4,15,19]. For example, in
two runs of [15] for spontaneous crystallization of KNO3
Figure 9. Aggregation constant versus degree of supersa-
turation of solution 0
CC in the coordinates of Equation
(26): a—run 1 and b—run 3.
from supersaturated aqueous-ethanol solutions
was
found to be 2.03 and 2
0.96mJ m
and the height of
nucleus-bridges was 9
1.83 10
and 9
2.34 10m
, re-
spectively. It is interesting to note that respective substi-
tutions in Equation (21) give 7
b
As A.
The validity of Polak’s mechanism of crystal inter-
growth was also confirmed in spontaneous crystallization
of KCl from supersaturated aqueous and aqueous-ethanol
solutions [4], in bulk crystallization of CaC2O4ּH2O from
supersaturated aqueous solutions (recalculation in [15]
data of [20,21]), and in spontaneous crystallization of
different salts from supersaturated aqueous-ethanol solu-
tions: NaCl [19], NaNO3, CaSO42H2O, KNO3, NaNO2,
K2SO4, Ba(NO3)2 [15].
m nitrat
rystals at bulk crystallization.
e of
near to
0.05 is not enough for intergrowth of
into contact for intergrowth.
3.7. Forms of Intergrowth of Crystals in
Agglomerates
To obtain additional information about mechanism of
intergrowth of crystals during bulk crystallization, the
agglomerates of potassiue crystals formed in the
present runs were analyzed by means of optical and SE
microscopy. So, the mechanism of intergrowth of crys-
tals proposed by Polak [16-18 its realizatio
] in case ofn at
bulk crystallization assumes that formation of nucleus-
bridges connecting crystals occurs at the moment of col-
lision of the crystals in solution (formation of nucleus-
bridges is nucleation process, therefore, there is always
some probability that formation of nucleus-bridges be-
tween crystals may take place immediately at their colli-
sion or during very short time interval when the crystals
are near each other). Only this, in our opinion, may ex-
plain intergrowth of c
Really, according to results of [22] the contact tim
crystals at their collision in a solution must be
s. It is clear that it
crystals by means of crystal growth, especially, for spar-
ingly soluble salts having low growth rates of crystals.
Besides, a thin layer of a solution always exists between
approaching crystals. Therefore, when two nearest faces
of the crystals grow towards each other, wedging pres-
sure between these faces does not allow them to come
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL.
24
If intergrowth of crystals occurs almost instantly, then
some agglomerates of crystals of very freakish form
should be found in a crystal product after finish of the
crystallization process. Of course, the number of them
will be small, as hydrodynamic conditions of bulk crys-
tallization process lead to the morphological selection of
formed agglomerates of crystals and prevalence among
them agglomerates of the correct form with low hydro-
dynamic resistance, right up to pseudo-monocrystals [23,
24]. Hence, detection of agglomerates of crystals of the
irregular form will serve as additional evidence that the
mechanism of intergrowth of crystals proposed by Polak
takes place in the conditions of bulk crystallization from
solutions.
In Figure 10(a) a few images of agglomerates of po-
tassium nitrate crystals of the irregular shape, which had
been found in a deposit after the end of crystallization,
are shown. More detailed optical and SE microscopy
images of formed large agglomerates of two crystals are
sh
wth of potassium nitrate crystals during bulk
cr
own in Figures 10(b)-(d).
It is clear that such agglomerates of irregular shape
might be formed only, if the intergrowth of two crystals
took place immediately at their collision in agitated solu-
tion. These images have been obtained after slight etch-
ing of a surface of crystals in 94% ethanol using ultra-
sound. This procedure allowed to destroy crystal pseudo-
agglomerates, in which were only Van der Waals forces
between crystals and there was no intergrowth due to
chemical bonds.
Examination of published data showed that there are
also many other works, where were found crystal ag-
glomerates of irregular and freakish forms. For example,
the intergro
ystallization was also found in [12].
It is interesting that in the photos presented in [12] one
may see almost the same crystal agglomerates of potas-
sium nitrate of irregular shape like in Figure 10(a) of the
present work. In the photos of crystals of another salt

42
6
KFeCN 3HO


studied in [12] is also discovered
netics of the crys-
tallization process after the end of the induction period of
agglomerates of irregular shapes. The agglomerates of
freakish forms formed from relatively large crystals of
potassium chloride may be found in Figures 3 and 4 of
work [25].
3.8. Model of Spontaneous Crystallization
Process
The great interest for theory and practic of bulk crystal-
lization present question about a possibility to predict
variation of salt concentration and concentration of salt
crystals, and also particle-size distribution of crystal
product during and at the end of a crystallization process.
Earlier in [26,27] a model of spontaneous crystallization
was proposed. The model describes ki
Figure 10. Optical and SE microscopy images (a)-(d) of po-
tassium nitrate crystal agglomerates of irregular shape
formed in runs of the present work.
crystallization, when aggregation of salt crystals started
and nucleation of new crystals was finished or was neg-
ligible. It is supposed that intergrowth of salt crystals
during crystallization proceeds via the formation of nu-
cleus-bridges between these crystals according to the
above mechanism of crystal intergrowth. The model is
described by a system of differential equations [26,27]:
 
13 23
113
0
d3.3 π
d
z
Ti
CnKCC CC
  (27)
2
d
da
nnvK
 (28)
1
1
d2
da
nvKn n
 (29)

2
2
12
d20.5
da
nvKnn n
  (30)

3
1
d2
da
nvKn nnn
23
  (31)

112 23 311
d
d
2
k
ak kkk
ak
nvKnnnnnnnn
vKn n
 

 
(32)
where n is the total concentration of growing salt crystals
in the solution
3
m, 1
nNv
; n1 is the concentra-
tion of initial salt crystals in the solution (m-3), i. e. those
crystals which were in the solution immediately after the
beginning of th, in particular, at
moment
e aggregation process this
1
nn
; k is th
23
,,,
k
n
her crystal agg
e numb
are the c
regate
er of crystals in a c
aggreg oncentrations of salt
crystals (os
rystal
ate; nn
r rat) consisting of two
, three
and
3
n
k
kn in
2
nitial salt crystals, respec-
tively
3
m.
rst equatThe fiion of systems (27)-(32)
tion of a salt concentration in solution during crystalliza-
describes varia-
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL. 25
tion. The equation takes into account the o
growth process and the total surface area of
crystals. The second equation of the system is trans-
formed Equation (17). It describes the kineti
gation of salt crystals during bulk crystallization. The
aggregation constant Ka in this formula may be calcu-
lated with Equation (22). The others equations of sy
(27)-(32) are transformed equations of the Smoluchow
g
rder of the
growing salt
cs of aggre-
stems
ski
theory [14]. They determine concentrations of crystal
aggregates consistinof two
, three

3
nand more

k
n initial salt crystals. The detailed derivation of sys-
tems (27)-(32) was given in [26]. It should be noted that
here we introduce into the model some correction factor
taking into account that most of salt crystals are not
ball-shapedis correction factor is the sphere-
cal coefficient already used above.
find a crystal size distribution during or at the end
of crystllization, Equa22) and all equations of sys-
tem
2
n
particles. Th
To
ation (
s (27)-(32) must be solved simultaneously. Then, in
the first approximation, diameters and weight fractions of
formed crystal agglomerates may be found using equa-
tions [26,27]:

13
23
23
6
π
i
k
Tk
k
kCC
dnk





(33)
23
23
100 k
k
k
k
nk
Qnk
(34)
k
k
QQ (35)
where dk is the average diameter of crystal agglomerates
consisting of k initial salt crystals (m); Qk is the weight
fraction of crystal agglomerates consisting of k initial salt
crystals in a crystal deposit of salt (%); Q is the total
weight fraction of all crystal agglomerates consisting
from 1 to k initial salt crystals in a deposit of salt (%). To
estimate the crystal size distribution at the end of crystal-
lization, C0 instead of C must be used in Equation (33).
The typical results of calculations according to the
above model fo
r spontaneous crystallization of potassium
nitrate studied in the present work are given in Figures
11 and 12. To establish the variation
tion C of the salt and the total number
in the solution during crystallization, the first two equa-
tions of systems (27)-(32) together with Equation (22)
were solved simultaneously by a
this the Runge-Kutta method and the found average
crystal growth rate coefficients K, and crystal aggrega-
tion parameters D1, D2, K1·D3 we
As is seen, Figures 11 and 12 exhibit a satisfactory
agreement between the corresponding experimental and
sy
ers of crystals
fo
cleation and breaking of salt crystals
du
s in the concentra-
N of salt crystals
numerical method. For
re used.
theoretical data.
In order to estimate a crystal-size distribution at the
end of the crystallization process, Equation (22) and
Equations (33)-(35) together with full system of Equa-
tions (27)-(32) were also solved using the special com-
puter program. The theoretical calculations were carried
out for k = 150, i.e. for crystal agglomerates consisting of
150 initial salt crystals. Therefore, 2 + 150 equations of
stems (27)-(32) and Equation (22) were solved simul-
taneously. Typical results of these calculations are dis-
played in Figure 13, where arrows point out the begin-
ning of the theoretical curves.
From Figure 13 it is seen that there are some differ-
ences between the experimental points and the theoretical
curves. For large d this discrepancy is within experiment-
tal errors, whereas for small d the deviations are signifi-
cant (see experimental points for small d, and arrows in
Figure 13). In particular, the proposed model of sponta-
neous crystallization gives larger diamet
r small d in comparison with those found experimen-
tally (see arrows in Figure 13). This may be explained
by secondary nu
ring crystallization. Nevertheless, the theoretical curves
for the size distribution of potassium nitrate crystals, on
the whole, are in satisfactory agreement with the experi-
mental data.
Figure 11. Variation of concentration C of potassium ni-
trate in solution during spontaneous crystallization: a—run
1; b—run 4; —experimental points; —theoretical cal-
culation according to systems (27)-(32) and Equation (22).
Figure 12. Variation of the total number N of salt crystals in
solution during spontaneous crystallization of potassium
nitrate: a—run 1; b—run 4; —experimental points;
theoretical calculation according to systems (27)-(32) and
Equation (22).
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL.
26
Figure 13. Size distribution of potassium nitrate crystals
after the completion of spontaneous crystallization: a—run
1; b—run 4; —experimental points; —theoretical cal-
culations according to systems (27)-(32), Equation (22) and
Equations (33)-(35), where arrows point out the beginning
of the theoretical curves; d is the diameter of potassium
nitrate crystals (agglomerates).
4. Conclusion
Thus, the experiments performed showed that spontane-
e 63.5 kJ·mol-1 of the activation energy
points out on the kinetic mode of the growth process.
Analysis of the experimental data on the aggregation
kinetics showed that the intergrowth of crystals pro-
ceeded via formation of nucleus-bridges between crystals.
This was confirmed by examination of equations of
above mechanism on the obtained experimental data and
by detection of crystal agglomerates of freakish and ir-
regular shapes, which we consider like the direct confir-
mation of the mentioned mechanism of intergrowth of
crystals. Besides, the specific surface energy of potas-
sium nitrate evaluated on the basis of the above mecha
of spontaneous crystalliza-
tion was also tested. It wasctory coinci-
e experimental data and theoret
of the salt and the total number
the experimental data on solubility of potassium nitrate
ous crystallization of potassium nitrate from its super-
saturated aqueous solutions is accompanied by aggrega-
tion of salt crystals. The growth of crystals during crys-
tallization was described by the equation of the first order.
The high valu
-
nism of aggregation and intergrowth of crystals had rea-
sonable value, and agreed satisfactorily with the avail-
able literature data. This is additional evidence of validity
of mechanism of intergrowth of crystals during bulk
crystallization via formation of nucleus-bridges between
crystals.
The earlier proposed model
found a satisfa
dence between th
curves on concentration
ical
of salt crystals in solution during spontaneous crystalli-
zation, and also between experimental data and theoreti-
cal curves of particle-size distribution of formed crystals
after completion of the crystallization.
An excellent agreement was also established between
in aqueous solutions found in the present work and those
available in the literature.
5. Acknowledgements
This work was financially supported by the Russian Foun-
dation for Basic Research, Project No 05-08-01289-a.
REFERENCES
[1] J. E. Helt and M. A. Larson, “Effects of Temperature on
the Crystallization of Potassium Nitrate by Direct Meas-
urement of Supersaturation,” AIChE Journal, Vol. 23, No.
6, 1977, pp. 826-828.
[2] T. A. Graber, M. E. Taboada, M. N. Alvarez and E. H.
Schmidt, “Determination of Mass Transfer Coefficients
for Crystal Growth of Nitrate Salts,” Crystal Research
and Technology, Vol. 34, No. 10, 1999, pp. 1269-1277.
doi:10.1002/(SICI)1521-4079(199912)34:10<1269::AID-
CRAT1269>3.0.CO;2-3
[3] J. Rolfs, R. Lacmann and S. Kipp, “Crystallization of
Potassium Nitrate (KNO3) in Aqueous Solution I,” Jour-
nal of Crystal Growth, Vol. 171, No. 1-2, 1997, pp. 174-
182. doi:10.1016/S0022-0248(96)00430-7
[4] O. D. Linnikov, “Spontaneous Crystallization of Potas-
sium Chloride from Aqueous and Aqueous-Ethanol Solu-
tions. Part 2: Mechanism of Aggregation and Coalescence
of Crystals,” Crystal Research and Technology, Vol. 39,
No. 6, 2004, pp. 529-539. doi:10.1002/crat.200310221
[5] P. A. Kouzov, “Fundamentals of Analysis of the Disper-
sion Composition of Industrial Dusts and Powdered Ma-
terials,” Chemistry, Leningrad, 1974.
[6] A. N. Verigin, I. A. Shchupliak and M. F. Mihalev, “Crys-
ems,” Chemistry, Leningrad,
e
Science, Vol. 37, No. 3, 1971, pp. 612-618.
doi:10.1016/002
tallization in Dispersible Syst
1986, pp. 60-61.
[7] B. R. Smith and F. Sweet, “The Crystallization of Cal-
cium Sulfate Dihydrate,” Journal of Colloid and Interfac
1-9797(71)90339-0
, “Modern Crystallogra[8] A. A. Chernovphy,” Nauka, Mos-
cs of Bulk Crystallization of Salts
cow, 1980.
[9] J. W. Mullin, “Crystallization,” 4th Edition, Butterworth-
Heinemann, Oxford, 2001.
[10] O. D. Linnikov, “Kineti
from Aqueous Solutions of Ethanol,” Russian Journal of
Applied Chemistry, Vol. 74, No. 7, 2001, pp. 1106-1111.
doi:10.1023/A:1013002531012
[11] O. D. Linnikov, “Spontaneous Crystallization of Sodium
Chloride from Aqueous-Ethanol Solutions. Part 1. Kinet-
ics and Mechanism of the
Crystallization Process,” Crys-
tal Research and Technology, Vol. 41, No. 1, 2006, pp.
10-17. doi:10.1002/crat.200410522
[12] L. N. Matusevich and K. N. Shabalin, “The Intensity of
Stirring of a Solution and Size of Received Crystals,”
Zhurnal Prikladnoi Khimii, Vol. 25, No. 11, 1952, pp.
1157-1164.
[13] M. Tovbin and S. Krasnova, “The Stability of Supersatu-
rated Salt Solutions,” The Journal of Physical Chemistry,
Copyright © 2013 SciRes. CSTA
O. D. LINNIKOV ET AL.
Copyright © 2013 SciRes. CSTA
27
2, 1917, pp. 129-
during Bulk Crystallization from So-
Vol. 23, No. 7, 1949, pp. 863-870.
[14] M. Smoluchowski, “Versuch Einer Mathematischen The-
orier der Koagulationskinetik Kolloider Lösungen,” Zeits-
chrift Fur Physikalische Chemie, Vol. 9
168.
[15] O. D. Linnikov, “Mechanism of Aggregation and Inter-
growth of Crystals
lutions,” Crystal Research and Technology, Vol. 43, No
12, 2008, pp. 1268-1277. doi:10.1002/crat.200800176
[16] A. F. Polak, “Hardening of Mono-Mineral Ast
stances,” Publishers of literatures on
ringent S
construction, Mos-
in, E. A. Amelina, R. K. Yusupov, V. P. Va-
ub-
cow, 1966.
[17] E. D. Shchuk
ganov and P. A. Rebinder, “The Experimental Study of
Influence of Solution Supersaturation and the Contact
Time of Crystals on Intergrowth of Separate Crystals,”
Soviet Physics—Doklady, Vol. 213, No. 1, 1973, pp. 155-
158.
[18] V. P. Vaganov, E. A. Amelina, R. K. Yusupov, E. D .
Shchukin and P. A. Rebinder, “The Experimental Study
of Formation of Crystal Contacts at Intergrowth of Sepa-
rate Crystals,” Colloid Journal of the USSR, Vol. 36, No.
3, 1974, pp. 436-442.
[19] O. D. Linnikov, “Spontaneous Crystallization of Sodium
Chloride from Aqueous-Ethanol solutions. Part 2. Mecha-
nism of Aggregation and Coalescence of Crystals,” Crys-
tal Research and Technology, Vol. 41, No. 2, 2006, pp.
138-144. doi:10.1002/crat.200510545
[20] A. S. Bramley, M. J. Hounslow and R. L. Ryall, “Aggre-
gation during Precipitation from Solution. Kinetics for
Calcium Oxalate Monohydrate,” Chemical Engineering
Science, Vol. 52, No. 5, 1997, pp. 747-757.
doi:10.1016/S0009-2509(96)00447-2
[21] M. J. Hounslow, A. S. Bramley and W. R. Paterson, “Ag-
gregation during Precipitation from Solution. A Pore Dif-
fusion-Reaction Model for Calcium Oxalate Monohy-
drate,” Journal of Colloid and Interface Science, Vol. 203,
No. 2, 1998, pp. 383-391. doi:10.1006/jcis.1998.5501
gulation Growth
udin, “The
hydrate of Calcium
of
[22] D. M. Levins and J. R. Glastonbury, “Particle-Liquid
Hydrodynamics and Mass Transfer in a Stirred Vessel.
Part 1,” Transactions of the Institution of Chemical En-
gineers, Vol. 50, 1972, pp. 32-41.
[23] I. V. Melikhov and A. S. Kelebeev, “Coa
of Crystals of Barium Sulphate from Strong Supersatu-
rated Solution,” Crystallography, Vol. 24, No. 2, 1979,
pp. 410-412.
[24] I. V. Melikhov, I. V. Miheeva and V. N. R
Mechanism of Crystallization of Semi
Sulphate in the Conditions Modelling Production of Phos-
phoric Acid by Hemihydrate Method,” Theoretical Foun-
dation of Chemical engineering, Vol. 19, No. 6, 1985, pp.
742-748.
[25] S. Sarig, C. L. Leci and N. Eidelman, “Agglomeration
Potassium Chloride Crystals from Supersaturated Aque-
ous Solutions,” Journal of Crystal Growth, Vol. 47, No. 3,
1979, pp. 365-372. doi:10.1016/0022-0248(79)90200-8
[26] O. D. Linnikov, “Spontaneous Crystallization of Potas-
.1002/crat.200610709
sium Chloride from Aqueous and Aqueous-Ethanol Solu-
tions. Part 3: Model of the Crystallization Process,” Cry-
stal Research and Technology, Vol. 41, No. 10, 2006, pp.
988-996. doi:10
[27] O. D. Linnikov, “Spontaneous Crystallization of Sodium
Chloride from Aqueous-Ethanol Solutions. Part 3: Ex-
amination of a Model of the Crystallization Process,”
Crystal Research and Technology, Vol. 42, No. 8, 2007,
pp. 758-765.