south african journal of chemical engineering 24 (2017) 1e7

Contents lists available at ScienceDirect

South African Journal of Chemical Engineering

journal homepage: http://www.journals.elsevier.com/
south-african-journal-of-chemical-engineering

Analysis of potassium nitrate purification with

recovery of solvent through single effect
mechanical vapor compression

Kiprotich E. Kosgey*, Sammy L. Kiambi, Peter T. Cherop

Durban University of Technology, Department of Chemical Engineering, 1334, Durban, 4000, South Africa

article info abstract

Article history: Analysis of purification of potassium nitrate with incorporation of single effect mechanical
Received 16 November 2016 vapor compressor for solvent recovery was done. Analysis focused on the effect of concen-
Received in revised form tration and temperature of mother liquor on the energy efficiency of the process and the
24 May 2017 amount of recovered solvent. Performance coefficient of mechanical vapor compressor
Accepted 25 May 2017 ranged between 1.5 and 7.5 depending primarily on the temperature of mother liquor. It was
found that with increase in temperature of mother liquor through pre-heating, the power of
Keywords: the compressor, compression ratio and amount of heat supplied to the evaporator decrease.
Performance coefficient For a 40% concentrated feed solution and mother liquor temperature above 80
C, perfor-
Mother liquor mance coefficient is higher than 4. It is therefore concluded that preheating mother liquor
Concentrated solution and reduction of the effect of concentration of both mother liquor and concentrated waste
Recovered solvent stream through other methods reduces the power consumption of purification process.
Boiling point elevation © 2017 The Authors. Published by Elsevier B.V. on behalf of Institution of Chemical Engi-
Mechanical vapor compressor neers. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
Potassium nitrate (KNO3) is used as a food preservative, fer-
tilizer and heat transfer agent in chemical industries. KNO3 is
also essential in the production of explosives, glass and steel
(Abidaud, 1991; Freilich, 2005; Jaroszek et al., 2016). The only
known ore of KNO3 is caliche mined in Chile (Freilich, 2005;
Velasco, 1992). Potassium chloride (KCl), also known as pot-
ash, is another mined compound of potassium. To meet the
high demand of KNO3, the mineral deposits are supplemented
with the manufactured compound. The following chemical
processes have been developed for its production:
(i) Reaction of potassium chloride with nitric acid at elevated
temperatures (Eq. (1)) (Abidaud, 1991; Freilich, 2005).
(ii) Potassium chloride react with hot aqueous sodium ni-
trate to produce sodium chloride (NaCl) and potassium
nitrate (Eq. (2)) (Jaroszek et al., 2016). NaCl is less soluble
at elevated temperatures and thus crystallizes out of
the hot solution rich in KNO3 (Abidaud, 1991; Freilich,
2005; Jaroszek et al., 2016).
NaNO3 ðaqÞ þ KCl ðsÞ/NaCl ðsÞ þ KNO3 ðaqÞ (2)
(iii) Nitric acid is reacted with potassium chloride to pro-
duce potassium nitrate and hydrochloric acid which is
extracted with an organic solvent (Eq. (3)) (Freilich, 2005;
Jurisova
 et al., 2013).
HNO3 þ KCl /HCl þ KNO3 (3)

100
C
3KCl þ 4HNO3 ƒƒ! 3KNO3 þ Cl2 þ NOCl þ 2H2 O (1) (iv) Electrodialysis e cation exchange membranes (CEM)
and anion exchange membranes (AEM) alternate in a

* Corresponding author.
E-mail address: kippserico86@gmail.com (K.E. Kosgey).
http://dx.doi.org/10.1016/j.sajce.2017.05.004
1026-9185/© 2017 The Authors. Published by Elsevier B.V. on behalf of Institution of Chemical Engineers. This is an open access article
under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
2 south african journal of chemical engineering 24 (2017) 1e7

NaNO3 þ K2 SO4 /Na2 SO4 þ KNO3 (4)

Nomenclature
(v) Ion exchange e nitric acid is passed through cationic
Roman letters and abbreviations exchange resin loaded with potassium leading to the
P Solvent, kg/s production of potassium nitrate (Eq. (5)). Potassium
E evaporator hydroxide (KOH) solution is added to neutralize
W concentrated solution, kg/s unreacted nitric acid (0.5% by weight) subsequently
t Temperature,
C producing pure potassium nitrate (Nagarale et al., 2006).
S crystals, kg/s
L mother liquor, kg/s
HNO3 þ RK /HR þ KNO3 (5)
p Pressure, kPa
h Enthalpy, kJ/Kg Each method of KNO3 production has its advantages and
D Dissolution disadvantages and the choice of one method over another is
SP separator influenced by the availability of raw materials, purity of KNO3,
H heat exchanger waste treatment and cost of production (Jurisova  et al., 2013).
EV Expansion valve In Freilich (2005) it was reported that sodium chloride is the
Cr crystallizer main contaminant in all the ores of potassium. Other con-
K Ebullioscopic constant,
C kg/mol taminants include clay, silica, kieserite (MgSO4$H2O), anhy-
MVC Mechanical vapor compressor drite (CaSO4) (Freilich, 2005), perchlorates (Urbansky et al.,
Q Quantity of heat, kJ 2001) and iodine (Velasco, 1992). The ores are purified
F Feed, kg/s through froth flotation, heavy media separation processes,
N Power of compressor, kW screening and fractional crystallization (Freilich, 2005).
x Concentration, % w/w Jaroszek et al. (2016) produced KNO3 through electrodialysis of
C Concentration, mol/m3 the highest reported purity of 99.9% but transportation of ions
c Specific heat capacity, kJ/kg-
C of impurities with potassium and nitrate ions still presented a
n Compression ratio challenge. Machuca and Cernı́n (2014) reported molar per-
r Latent heat, kJ/Kg centage composition of chlorides at 2% of KNO3 produced
i Van't Hoff factor through electrodialysis.
Impurities in KNO3 especially for use in agriculture and
Subscripts
food industry can be hazardous. Caliche deposits are natural
v1 Inlet vapor
sources of perchlorates (Urbansky et al., 2001) and it is highly
v2 Outlet vapor
likely that mined deposits are contaminated with it. Perchlo-
o Fresh solvent
rates are known to affect the uptake of iodine in the thyroid
cr Fractionation temperature
gland (Urbansky et al., 2001). Other contaminants could also
eb Ebullioscopic
present other health issues and there is therefore need to
T Solid feed
purify KNO3. One of the ways of its purification is
compr compressor
recrystallization.
adi adiabatic
Crystallization techniques are advantageous over other
A solute
purification methods because they are cheap, highly efficient
mex mechanical
and simple (Lu et al., 2017; Mullin, 2001). The possibility of
cond condensation
solvent recirculation and self-heat recuperation greatly im-
sol solvent
proves the prospects of the process. Multi stage flash distil-
Greek symbols lation, single and multi-effect evaporation, mechanical vapor
h efficiency compression (MVC), thermal vapor compression (Alasfour and
r Density, kg/m3 Abdulrahim, 2011; Han et al., 2015), membrane technology (Lu
d Boiling point elevation,
C et al., 2017; Han et al., 2015) ion exchange technology and
ε Performance coefficient nanotechnology (Lu et al., 2017) have been widely researched
on for water recovery. Compression of vapors has made
thermal processes of water recovery common especially
because of low cost, flexibility, high purity of distillate and
compartment with external electric field. The ion ex- thermodynamic efficiency (Alasfour and Abdulrahim, 2011).
change membranes keep ions of the same charge but Of the thermal processes, MVC is highly efficient (Alasfour and
allow ions of opposite charge to pass through (Jaroszek Abdulrahim, 2011; Zhou et al., 2014).
et al., 2016; Nagarale et al., 2006). Jaroszek et al. (2016) Extensive research in the use of mechanical vapor com-
fed K2SO4 into the two outermost cells in a four cell pressors in water recovery from waste water and saline water
compartment and NaNO3 in the central cell leading to has been done. Zhou et al. (2014) investigated power con-
separation of Naþ and NO3  ions and subsequent sumption through reduction of temperature difference be-
production of sodium sulfate and potassium nitrate tween the compressed steam in the evaporator tubes and the
(Eq. (4)). Potassium and sodium ions easily pass saline wastewater. They concluded that power consumption
through negatively charged CEM but are retained by increase linearly with temperature difference. These findings
positively charged AEM while sulfate and nitrate an- were in line with (Alasfour and Abdulrahim, 2011) conclusions
ions easily pass through positively charged AEM but on their study in temperature drop in MVC unit. Treatment of
are retained by negatively charged CEM (Nagarale high strength wastewater in double effect mechanical vapor
et al., 2006). recompression unit was studied by Liang et al. (2013) with
 south african journal of chemical engineering 24 (2017) 1e7 3

PO, xO, tO

FT, xT, tT F, xF, tF S+L S, xS, tS

D Cr tcr SP

L, xL, tL
P, x2, t2

H1

L, xL, tL1
V, hv1, p1 tv1

H2 W, xW, tW2

Compr
E L, xL, tL2
tv2, p2 ,hv2

EV
h3, p3
P, h4, t1, p4

W, xW, tw1

Fig. 1 e Schematic diagram of re-crystallization in which mother liquor L is pre-heated by recovered solvent P and
concentrated solution W.

focus on decrease in total heat transfer area with decrease in
temperature difference between condensing steam and
boiling solution. Han et al. (2014) applied self-heat recovery
theory in evaporative crystallization of ammonium sulfate.
Han et al. (2017) demonstrated that in double effect mechan-
ical vapor compression, the total power consumption is less
than in a single effect mechanical vapor compression system
but compression stages higher than two do not guarantee
reduction in total power consumption with increase in the
number of stages.
In this paper, design for purification process incorporating
single effect mechanical vapor compression for solvent re-
covery is developed with analysis of performance based on
energy efficiency. This is because purification of KNO3 is
necessary to remove harmful contaminants like perchlorates.
In addition, water recovery is essential especially with the
common problem of water scarcity and strict regulations
crystals S, is fed into evaporator E where most of the solvent is
recovered, leaving concentrated solution W. Vapors P of the
solvent are compressed from pressure p1 to pressure p2 raising
the enthalpy of the vapors from hv1 to hv2. The compressed
vapors are then fed into the heating chamber of evaporator E
where upon condensation boils the mother liquor L. The
condensate P with enthalpy h3 and the concentrated solution
120
100
80
60
tF , оС

concerning wastewater treatment and disposal.

2. Principle of operation 40

The feed, is dissolved in the recovered solvent P with the 20

addition of a small amount of fresh solvent PO (Fig. 1). The
amount of PO depends on the loss of solvent with concen-
0
trated solution W. Energy required for dissolution in the pro-
cess can be supplied by solvent P coming from evaporator E.
Solution F of KNO3 is fed to crystallizer Cr where it is cooled to -20
temperature tcr which will change with the concentration of 0 20 40 60 80
хF , % КNО3
KNO3 in the solution (Fig. 2). As a result of crystallization,
magma S þ L is produced which is fed to separator SP (Fig. 1). Fig. 2 e T-X diagram for KNO3 solution in water (Pavlov
Mother liquor L at temperature tL, after its separation from et al., 1987).
4 south african journal of chemical engineering 24 (2017) 1e7

W which is rich in impurities exit the evaporator. Pressure of solution respectively; rD is the heat of dissolution of the solid
the condensate P drops from p3 to p4 on passing through feed.
expansion valve EV. Recovered solvent P from the evaporator From Eq. (9), it is possible to define temperature t2 at which
E is hot and if its heat content is higher than the heat needed it is necessary to return the recovered solvent to stage D for
in dissolution stage then it is necessary to cool to the required dissolution of the solid feed so that solution F exits at
temperature t2. For this purpose, the hot solvent is fed to heat temperature tF:
exchanger H1 to pre-heat mother liquor to temperature tL1.
FcF tF þ FT ðrD  cT tT Þ  PO xO tO
The temperature of mother liquor can further be raised to tL2 t2 ¼ (10)
PCP
by passing it through heat exchanger H2 which is heated by
hot concentrated solution W from the evaporator. Heating The amount of energy QE that need to be supplied to
mother liquor L before feeding to the evaporator, can signifi- evaporator E can be determined from Eq. (11).
cantly reduce the amount of heat needed in the evaporator.
LcL tL þ QE ¼ WcW tW1 þ Phv1 (11)

where cL and cW are the specific heats of mother liquor and

concentrated solution respectively; tL and tW are temperatures
2.1. Basis for calculation
of mother liquor at entry into the evaporator and concen-
trated solution at exit from the evaporator respectively.
2.1.1. Material balance
QE transferred from compressed vapors P to the boiling
The material balance for stage D (dissolution) in Fig. 1 can be
solution P can also be determined as:
determined as:
QE ¼ Pðhv2  h3 Þ ¼ Prcond (12)
P þ P O þ FT ¼ F (1a)
where rcond is the latent heat of condensation of solvent vapors
Px2 þ PO xO þFT xT ¼ FxF (2a) in the evaporator.
The specific heat capacity of solutions L and W can be
where P, PO, FT and F are the mass flow rates of recovered
determined using the method of mixtures as (Qiao et al., 2017):
solvent, fresh solvent, feed (solid) and solution of KNO3
respectively; x2, xO, xT and xF are the solute concentrations in cL ¼ cA xL þ cP ð1  xL Þ (13)
the recovered solvent, fresh solvent, solid feed and solution
respectively. cW ¼ cA xW þ cP ð1  xW Þ (14)
From Eqs. (1a) and (2a), solute concentration in solution F
can be determined as: where cA and cP are specific heat capacities of the solute
(KNO3 ¼ 1.39 kJ/kg-
C (Takahashi et al., 1988)) and solvent
Px2 þ PO xO þ FT xT (water ¼ 4.18 kJ/kg-
C (Han et al., 2014)) respectively.
xF ¼ (3a)
P þ PO þ FT The heat content of the recovered solvent may be more
Without a trace of the solute in the recovered solvent P and than the heat necessary to dissolve the solid feed. In this case,
PO (x2 ¼ 0 and xO ¼ 0), Eq. (3a) simplifies to: the recovered solvent is cooled down to temperature t2 in heat
exchanger H1 and in the process transferring heat Q1 to
FT xT mother liquor L.
xF ¼ (4a)
P þ PO þ FT
Q1 ¼ PcP ðt2  t1 Þ ¼ LcL ðtL1  tL Þ (15)
Material balance for evaporator E is:
where t1 and t2 are the temperature of the solvent before and
L¼WþP (5a) after passing through the heat exchanger H1.
If mother liquor L is fed to the evaporator at the boiling
LxL ¼ WxW þ Px2 (6) temperature, tL2 can be determined from Eq. (16) (Alasfour and
where xL, xW and x2 are solute concentrations in mother li- Abdulrahim, 2011; R. N. G. Pavlov and Noskov, 1987).
quor, concentrated solution and recovered solvent
tL2 ¼ tP þ dL (16)
respectively.
From Eqs. (5a) and (6), P can be determined as: where tP is the boiling temperature of pure solvent at pressure
p1 and dL is boiling point elevation of mother liquor with con-
xW  xL
P¼L (7) centration xL. dL can be determined as (Gazagnes et al., 2007):
xW  x2
And since there is no solute in the recovered solvent i:CKNO3
dL ¼ Keb (17)
(x2 ¼ 0), Eq. (7) becomes: rsol

  where Keb is ebullioscopic constant of water (0.51
C.kg/mol), i

xL
P¼L 1 (8) is the Van't Hoff factor of 2 for KNO3, CKNO3 is the concentration
xW
of KNO3 (mol/m3) and rsol is the density of water (kg/m3)
(Gazagnes et al., 2007).
2.1.2. Energy balance
In heat exchanger H2, Q2 transferred from concentrated
The amount of heat required for dissolution of the feed can be
solution W to mother liquor L can be determined as:
determined as:
Q2 ¼ LcL ðtL2  tL1 Þ ¼ WcW ðtw1  tw2 Þ (18)
FT cT tT þ PcP t2 þ PO cO tO ¼ FT rD þ FcF tF (9)
where tW1 and tW2 are the inlet and exit temperature of
where cT, cP, cO and cF are the specific heats of solid feed,
concentrated solution before and after passing through the
recovered solvent, fresh solvent and solution respectively; t2,
heat exchanger.
tT and tF are temperatures of recovered solvent, solid feed and
 south african journal of chemical engineering 24 (2017) 1e7 5

Boiling temperature of W can be determined as (Alasfour 0.6

and Abdulrahim, 2011; Pavlov et al., 1987):
3 a
tW ¼ tP þ dW (19) 0.5

FRPF=P/F
where dW is boiling point elevation of solution W with con-
2
centration xW and can be calculated using Eq. (17). 0.4
Heat exchange between streams P, W and L can reduce 1
energy demand for solvent recovery. If the energy supplied
0.3
with recovered solvent is not enough for dissolution of solid 10 20 30 40
feed then fresh solvent should be heated up before supplying хL , % KNO3
for dissolution. Heat transfer area for heat exchangers and
2
evaporator can be calculated as described in Ettouney (2006).
1
2.1.3. Performance coefficient of the heat pump
1.5 b

FRWFW/F
The theoretical power of compressor can be determined from
Eq. (20) (Zhou et al., 2014).
1 2

Ncompr ¼ Pðhv2  hv1 Þ (20) 0.5 3

Accordingly, the actual compressor capacity is (Zhou et al.,
0
2014; Nosov et al., 2009): 10 20 30 40
хL , % KNO3
Pðhv2  hv1 Þ
Ncompr ¼ (21)
hadi hmex Fig. 3 e Change in flow rate ratio of recovered solvent to
where hadi and hmex are adiabatic and mechanical efficiencies flow rate of KNO3 solution (a) and flow rate ratio of
of the compressor respectively. concentrated solution to concentration of KNO3 in mother
Performance coefficient ε of mechanical vapor compressor liquor (b) tL ¼ 90
С: xF ¼ 1 e 20%; 2 e 30%; 3 e 40%.
is (Zou and Xie, 2017):

QE h h ðhv2  h3 Þ increase in NaCl concentration, the performance of a desali-

ε¼ ¼ ad mex (22)
Ncompr hv2  hv1
2.1.4. Process parameters
Process parameters were based on data from previous study
(Nosov et al., 2009): solid feed rate FT ¼ 1 kg/s, KNO3 concen-
tration in the solution xF ¼ 20e50%, range of fractionation
temperature tcr ¼ 0e30
C. The following assumptions have
been made:
1) There is no heat loss in the process apparatuses
2) The effect of impurities on process thermal parameters
will be negligible compared to the effect of KNO3 in mother
liquor
3) Only pure crystals of KNO3 will be produced
3. Results and discussion
Calculated results for elevation of boiling point of the solvent
are presented in Table 1:
With increase in concentration of KNO3, boiling point of
the solution rises. This directly affects the amount of vapors
generated in the evaporator and subsequently flow rate of the
recovered solvent (Fig. 3(a)). More solvent will be recovered in
the evaporator from a more concentrated feed solution to the
nation unit diminished because less water vaporized at the
membrane surface. The amount of exiting concentrated so-
lution, however, increases with elevation of boiling point as
less water evaporates from mother liquor with increase in its
concentration (Fig. 3(b)).
Recycled solvent P supplies the required heat for dissolu-
tion in stage D and as a result, reduces demand for additional
heat at this stage. Mathematical analysis has shown that
concentration xF depends primarily on temperature t2 (Fig. 4).
This is because the recovered solvent supplies heat of disso-
lution of the fresh solid. Thus, with increasing xF, temperature
t2 at which recovered solvent is supplied to stage D increases.
A low fractionation temperature leads to a lower t2 than a high
fractionation temperature. t2, also increases with concentra-
tion xL of mother liquor and xF of solution (Fig. 5). This is due to
the elevation of boiling point of the solvent.
Another important factor influencing the process of
recrystallization of KNO3, is energy consumption at the re-
covery of solvent. In this case, thermal energy consumption
100
3
80 2
1
t2, °C

crystallizer than a less concentrated feed solution to the 60

crystallizer. Gazagnes et al. (2007) established that with

40

20
20 30 40 50
Table 1 e Boiling point elevation of water at different xF, %
concentrations of potassium nitrate.
Fig. 4 e Change in temperature of recovered solvent fed
Concentration, % 20 30 40 50
to stage D with the concentration of KNO3 solution:
Boiling point elevation,
C 2.02 3.03 4.04 5.05
tcr ¼ 1 e 0
С; 2 e 10
С; 3 e 20
С.
6 south african journal of chemical engineering 24 (2017) 1e7

100 50

80 4 a
40

60
3
3
t2 , °С

n = P2 /P1
30
40
1 20 2
20
10
1
0
10 20 30 40 0
хL , % KNO3 0 10 20 30 40 50 60 70 80

Fig. 5 e Variation of temperature of recovered solvent tL , °C

supplied for dissolution of solid feed with concentration of
mother liquor xF ¼ 1 e 20%; 2 e 30%; 3 e 40%; 4 e 50%.
800
1 b
2

Ncompr /F, KJ/Kg

600
QE in the evaporator, depends on parameters of dissolution of
3
solids and subsequent crystallization. It was found that QE is
400
mainly dependent on the concentration xL of mother liquor
and its temperature tL (Fig. 6(a) and (b)). On increasing con- 200
centration xL there is a reduction in the amount of heat QE per
unit mass of the solution F since generation of vapors P de- 0
0 10 20 30 40 50 60 70 80
creases and mass flow rate of concentrated solution W in-
tL , °C
creases. An increase in temperature tL through preheating
with recovered solvent P and waste stream W also leads to a
Fig. 7 e Relationship between the level of compression of
reduction of QE.
vapors (a) and power of compressor (b) with temperature of
If mother liquor L is supplied to evaporator E without pre-
mother liquor. xF ¼ 40%, xL ¼ 12%: xW ¼ 1 e 75%; 2 e 65%;
heating, that is at tL ¼ tcr ¼ 0
C, consumption of heat in the
3 e 55%.
evaporator is high and the compression ratio of the
compressor reaches very high values of n ¼ 40e45 (Fig. 7a).
Failure to preheat mother liquor leads to a significant increase increase in temperature is small. Aybar (2002) concluded that
in the power of the compressor per unit mass of solution the minimum temperature at which water should be fed into
(Fig. 7b). Lower than 70
C, compression ratio rapidly de- the evaporator is 75
C as latent heat of condensation of steam
creases with increase in temperature of mother liquor and at does not give enough heat for evaporation at lower tempera-
higher temperatures the change in compression ratio with tures. Lower temperatures lead to a larger temperature dif-
ference between compressed superheated steam and mother
liquor in the evaporator. According to Zhou et al. (2014), larger
temperature difference leads to a higher compression ratio
1600 and subsequently higher energy consumption by the
1 a compressor. High power consumption reduces the energy
QE /F, kJ/Kg

1500 efficiency of the process.

2 At tL below 10
C, performance coefficient is extremely low,
in the range ε ¼ 1.5e1.7 (Fig. 8). Performance coefficient at
1400
3
temperatures of mother liquor above 70
C rises rapidly with
1300 increase in temperature. For a 55% w/w concentrated solution,
performance coefficient is double that of 75% w/w concen-
1200 trated solution with tL above 80
C. An increase in concen-
0 10 20 30 40 50 60 70 80 tration of concentrated waste steam leads to an increase in
tL , °C the pressure ratio. This is as a result of the effect of boiling
point elevation (Han et al., 2015).
2000

1 b
1800
QE /F,KJ/Kg

2 8
1600
6 3
1400
3 4
2
ε

1200 1
2

1000 0
10 20 30 40 0 10 20 30 40 50 60 70 80
xL , % tL , °C

Fig. 6 e Variation of heat QE in evaporator E with Fig. 8 e Variation of performance coefficient with
temperature of mother liquor (a) and its concentration temperature of mother liquor xF ¼ 40%: xW ¼ 1 e 75%;
(b): xF ¼ 1 e 20% KNO3; 2 e 30%; 3 e 40%. 2 e 65%; 3 e 55%.
 south african journal of chemical engineering 24 (2017) 1e7
4. Conclusion
Production of KNO3 through conventional methods with
subsequent purification through recrystallization with sol-
vent recovery and incorporation of mechanical vapor
compressor is energy efficient and it presents good economic
prospects. Analysis has shown that in the considered design,
the performance coefficient, depending on the temperature of
mother liquor, is between 1.5 and 7.5. Without pre-heating
mother liquor, COP is very low. Above 80
C, power con-
sumption by the compressor is low and COP is higher than 4.
Concentration of mother liquor influences solvent recovery
through elevation of boiling point. A more concentrated
mother liquor boils at a higher temperature than a pure sol-
vent and thus consumes more power. As a result, other than
preheating mother liquor, the rise in power consumption as a
result of boiling point elevation should be minimized through
other methods. In instances where KNO3 concentration is
high and impurities are low, recirculation of mother liquor can
be considered whereas if concentration of impurities is high,
an additional crystallizer and separator to reduce concentra-
tion of impurities before solvent recovery can be installed.
Acknowledgement
The authors acknowledge the support given in sourcing some
of the articles given by Jane Avenal Finlayson of Durban Uni-
versity of Technology Steve Biko campus library.
References
Abidaud, A., 1991. Continuous production of potassium nitrate
via ion exchange.
Alasfour, F.N., Abdulrahim, H.K., 2011. The effect of stage
temperature drop on MVC thermal performance. Desalination
265, 213e221, 1/15/.
Aybar, H.S., 2002. Analysis of a mechanical vapor compression
desalination system. Desalination 142, 181e186, 2/1/.
Ettouney, H., 2006. Design of single-effect mechanical vapor
compression. Desalination 190, 1e15, 2006/04/15.
Freilich, R.L.P.M.B., 2005. Kirk-othmer encyclopedia of chemical
technology. In: Potassium Compounds, vol. 19. John Wiley &
Sons. Posted.
Gazagnes, L., Cerneaux, S., Persin, M., Prouzet, E., Larbot, A., 2007.
Desalination of sodium chloride solutions and seawater with
hydrophobic ceramic membranes. Desalination 217, 260e266,
11/5/.
Han, D., He, W., Yue, C., Pu, W., Liang, L., 2014. Analysis of energy
saving for ammonium sulfate solution processing with self-heat
recuperation principle. Appl. Therm. Eng. 73, 641e649, 12/5/.
Han, D., Yue, C., He, W., Liang, L., Pu, W., 2015. Energy saving
analysis for a solution evaporation system with high boiling
point elevation based on self-heat recuperation theory.
Desalination 355, 197e203, 1/1/.
7
Han, D., He, W.F., Yue, C., Pu, W.H., 2017. Study on desalination of
zero-emission system based on mechanical vapor
compression. Appl. Energy 185 (Part 2), 1490e1496, 1/1/.
Jaroszek, H., Lis, A., Dydo, P., 2016. Transport of impurities and
water during potassium nitrate synthesis by electrodialysis
metathesis. Sep. Purif. Technol. 158, 87e93, 1/28/.
Juri
sova , J., Fellner, P., Danielik, V., Lencse
s, M., Kra
 lik, M.,

Sı́po 
s, R., 2013. Preparation of potassium nitrate from
potassium chloride and magnesium nitrate in a laboratory
scale using industrial raw materials. In: Acta Chimica Slovaca,
vol. 6, p. 15.
Liang, L., Han, D., Ma, R., Peng, T., 2013. Treatment of high-
concentration wastewater using double-effect mechanical
vapor recompression. Desalination 314, 139e146, 4/2/.
Lu, H., Wang, J., Wang, T., Wang, N., Bao, Y., Hao, H., 2017.
Crystallization techniques in wastewater treatment: an
overview of applications. Chemosphere 173, 474e484, 4//.
Machuca, L., Cernı́n, A., 2014. A Method of Production of
Potassium Nitrate by Electrodialysis and Apparatus for Making
the Same.
Mullin, J.W., 2001. Crystallization. Available: https://books.google.
co.za/books?hl¼en&lr¼&id¼Et0EtojQmvsC&oi¼fnd&pg¼
PP2&dq¼purification%2Bcrystallization%
2Bsubstances&ots¼VreA_VNzDM&sig¼w_V4__kxS-
Ta0pyyE6Jy5mWthws#v¼onepage&q¼purification%
2Bcrystallization%2Bsubstances&f¼false.
Nagarale, R.K., Gohil, G.S., Shahi, V.K., 2006. Recent developments
on ion-exchange membranes and electro-membrane
processes. Adv. Colloid Interface Sci. 119, 97e130, 2/28/.
Nosov, G.A., Uvarov, M.E., Myasaedenkov, V.M., 2009.
Recrystallization of Substances from Solutions with
Application of Heat Pumps [Online]. Available: http://elibrary.
ru/item.asp?id¼12981816& (in Russian).
Pavlov, K.F., Romankov, P.G., Noskov, A.A., 1987. Examples and
Problems in Industrial Unit Operations [Online]. Available:
http://bookree.org/reader?file=541321&pg=1 (in Russian).
Qiao, G., Lasfargues, M., Alexiadis, A., Ding, Y., 2017. Simulation
and experimental study of the specific heat capacity of
molten salt based nanofluids. Appl. Therm. Eng. 111,
1517e1522, 1/25/.
Takahashi, Y., Sakamoto, R., Kamimoto, M., 1988. Heat capacities
and latent heats of LiNO3, NaNO3, and KNO3. Int. J.
Thermophys. 9, 1081e1090, 1988//.
Urbansky, E.T., Brown, S.K., Magnuson, M.L., Kelty, C.A., 2001.
Perchlorate levels in samples of sodium nitrate fertilizer
derived from Chilean caliche. Environ. Pollut. 112, 299e302, 5//.
Velasco, P., 1992. Minerals Yearbook, 01/02/2017 (Volume 3 ed.).
Available: http://digicoll.library.wisc.edu/cgi-bin/EcoNatRes/
EcoNatRes-idx?type¼turn&id¼EcoNatRes.
MinYB1992v3Lat&entity¼EcoNatRes.MinYB1992v3Lat.
p0164&q1¼caliche.
Zhou, Y., Shi, C., Dong, G., 2014. Analysis of a mechanical vapor
recompression wastewater distillation system. Desalination
353, 91e97, 11/17/.
Zou, S., Xie, X., 2017. Simplified model for coefficient of
performance calculation of surface water source heat pump.
Appl. Therm. Eng. 112, 201e207, 2/5/.