Electrical Conductivity of Lithium Chloride, Lithium Bromide, and Lithium Iodide Electrolytes in Methanol, Water, and Their Binary Mixtures

Full text

Article
Electric conductivity and electric convertibility of potassium acetate in
water, ethanol, 2,2,2–trifluoroethanol, 2–propanol and their
binary blends☆
Xi Wu ⁎, Shiming Xu, Debing Wu, Huan Liu
Key Laboratory of Ocean Energy Utilization and Energy Conservation of Ministry of Education, School of Energy and Power Engineering, Dalian University of Technology, Dalian 116024, China
abstractarticle info
Article history:
Received 26 February 2018
Received in revised form 20 May 2018
Accepted 9 June 2018
Available online 19 June 2018
Salinity gradient energy between the concentrated and diluted electrolyte solutions can be converted to electric
energy by using reverse electrodialysis (RED) technology. Electrolyte solution is a vital factor that impacts the
energy conversion efficiency. Potassium acetate (KAc) was chosen as solute, and water, ethanol, 2,2,2-
trifluoroethanol (TFE), 2-propanol (IPA) and several of their binary mixtures were selected as solvents. Electric
conductivity of these solutions were measured under varying conditions. KAc was easily ionized in water and
possessed the maximum electric conductivity, following by KAc–H
2
O–TFE and KAc–H
2
O–ethanol, and then
KAc in pure TFE, ethanol, and IPA respectively. For electric convertibility of these solutions working in a RED
power generation system, it was found that the KAc–H
2
O possessed the maximum power density, and the
KAc–ethanol–H
2
O possessed the larger open circuit voltage than aqueous KAc solution under the same working
condition. Besides, it was observed that both the electric conductivity and electric convertibility were
significantly influenced by the concentration and temperature of solution. With the increasing of concentration,
electric conductivity of these solutions increased firstly and then reached to the peak, but later it decreased.
Solution temperature took a positiveimpact role to the electric conductivity. Electric conductivity of these solu-
tions can be estimated by using a modified amplitude version of Gaussian peak function.
© 2018 The Chemical Industry and Engineering Society of China, and Chemical Industry Press. All rights reserved.
Keywords:
Conductivity
Alcohol
Electr ochemist ry
Solution
Mixtures
1. Introduction
1.1. A closed type RED power generation system driven by thermal energy
Salinity gradient energy exists between the concentrated and diluted
electrolyte solutions, and which can be converted to electric energy by
using reverse electrodialysis (RED) power generation technology. The
RED power generation technology is firstly proposed by Pattle in 1954
[1], and attracting increasing researchers in energy field recently with
the rapid development of ion exchange membrane (IEM) [2, 3].Asimple
RED apparatus consists of many alternately distributed cation exchange
membranes (CEM, only permeable for positive ions) and anion exchange
membranes (AEM, only permeable for negative ions), sandwiched
between two polar plates with electrodes at each side respectively, as
well as some feeding pumps and tubes, as Fig. 1 shown.
When pumping the concentrated and diluted electrolyte solutions
into the alternating compartments respectively, anion ions in the
concentrated solution compartments can diffuse spontaneously across
the AEM and transport to the adjacent diluted solution compartments.
In contrast, cation ions move towards the opposite direction by passing
through the CEM. At this time, a net diffusion current is generated, and
the chemical potential difference between the concentrated and diluted
solutions generates a voltage across each IEM [4]. Oxidizing reaction
and reduction reaction are ongoing respectively near each electrode
with the looping of electrode rinse solution. If connecting an external
load to the electrodes, then the inner ionic current can be converted
into the external electrical current [5].
The outgoing electrolyte solutions from the RED apparatus are
collected and regenerated in the distiller by using the multi-effect distil-
lation (MED) technology, driven by the low temperature thermal energy
[6], such as solar energy, geothermal energy, engine waste heat, etc.
Then, the mixed electrolyte solution is separated to the concentrated so-
lution and diluted solution again with the harvestable chemical potential
difference. By means of integrating the above two processes together, a
closed type thermal energy driven power generation cycle is available
[6–8], and the schematic of which has been drawn as Fig. 2. This is a
novel technology in electrochemical energy engineering fields [9].
Chinese Journal of Chemical Engineering 26 (2018) 2581–2591
☆Supported by the National Natural Science Foundations of Chin a (51606024,
51776029), and the Fundamental Research Funds for Central Universities (DUT17JC31),
the China Scholarship Council (iCET2017 Program).
⁎Corresponding author.
E-mail address: xiwu@dlut. edu.cn (X. Wu).
https://doi.org/10.1016/j.cjche.2018.06.004
1004-9541/© 2018 The Chemical Industry and Engineering Society of China, and Chemical Industry Press. All rights reserved.
Contents lists available at ScienceDirect
Chinese Journal of Chemical Engineering
journal homepage: www.elsevier.com/locate/CJChE

1.2. Potential working solutions for the RED power generation system
Further study results indicate that the electrolyte solution looping
in this power generation system is one of the vital factors that directly
impact the energy conversion efficiency (ECE). And the ECE can be calcu-
lated by using Eq. (1) [5, 7],
ECE ¼Power Output
Heat input þpump power ¼ηS E ηH S ð1Þ
where η
s_E
is the efficiency of process of converting salinity gradient
energy to electric energy;
η
H_S
is the efficiency of process of converting thermal energy to
salinity gradient energy.
The η
H_S
can be calculated by using Eq. (2), thus it can be seen that
the absorbent of working solution should be easily evaporated by
absorbing thermal energy as little as possible [5, 7].
ηH S ¼ϕC=D=hHOV ð2Þ
where, ϕ
C/D
is the chemical potential difference between the concen-
trated and diluted solution;
h
HOV
is the latent heat of vaporization of mixed solution.
Most of the current attention in this field has been paid to the
sodium chloride (NaCl) aqueous solution and natural sea–river water
Fig. 1. Schematic of salinity gradient energy powergeneration system with RED method.
Fig. 2. Schematic of a closed type thermal energy driven power generation cycle with RED method.
2582 X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

[1–4, 10, 11], while some researchers preferred to the ammonium bicar-
bonate (NH
3
HCO
3
)aqueoussolution[12, 13] and aqueous solutions of
lithium chloride (LiCl) and lithium bromide (LiBr) [5]. As a matter of
fact, the pure water is not an ideal absorbent here due to its relative
huge latent heat of vaporization and relative high operating temperature
(NH
3
HCO
3
aqueous solution is an exception), resulting in the severe
efficiency reduction during the process of solution regeneration. An
ideal working solution, for the energy conversion system shown in
Fig. 2, needs to possess various practical characteristics in the aspects
of thermophysical and transportation properties, electrochemistry
characteristics, environmental influence, safety, cost, etc. Methodology
of assessing the available working solutions can be found in the previous
report, and it has been suggested that the binary or ternary solutions
consisted of some monovalent salts and appropriate organic solvents or
blends solvents may be workable [8, 14], due to their relative higher sol-
ubility and electric conductivity, relative lower latent heat of vaporization,
acceptable electric convertibility, and suitable boiling point temperature.
Recently, working solutions consist of potassium acetate (CH
3
COOK, or
KAc), water and organic solvents have been taken into consideration.
1.3. Physical and chemical properties of KAc in water and organic solvents
Potassium acetate, with molecular weight of 98.1423 g·mol
−1
,isone
of food additives and medicine ingredients, and also used as anti-freeze
and de-icing products [15]. The normal physical form of KAc is white
flakes or crystalline powder, while it is hygroscopic [16]. Solubility of
KAc in water is 256 g·(100 ml H
2
O)
−1
at 20°C [16], 269 g·(100 ml
H
2
O)
−1
at 25°C [16], 283 g·(100 ml H
2
O)
−1
at 30°C [17], and
324 g·(100 ml H
2
O)
−1
at 40°C [17], respectively. The saturated vapor
pressure of potassium acetate aqueous is 0.581 kPa at 25°C, and the
molar enthalpies of vaporization is 4560 J·mol
−1
at the same tempera-
ture condition [18]. KAc can be ionized in water, and the approximate
effective ionic radii is 0.3 nm for K
+
and 0.45 nm for CH
3
COO
—
respectively at 25°C [17]. KAc is less soluble in ethanol (C
2
H
5
OH) than
in water, and which tends to form association complexes in the liquid
phase preferentially with the water molecules over those of the alcohol
[19]. Molar electric conductivity at infinite dilution for aqueous KAc solu-
tion is 40.9 Ω
−1
·mol
−1
·cm
2
at 25°C [16]. The calculated value of limiting
conductance of KAc in heavy water (D
2
O) was 94.7 K.U. (Kohlrausch
units) [20].
Some efforts have been made on the conductivity of KAc in different
organic solvents. Ref. [21] reported the electric conductance behavior of
KAc in mixed solvent (20% acetic acid and 80% acetonitrile, wt.%), and
also calculated the values of limiting conductance at 35°C according
to the Fuoss–Hsia equation, Fuoss–Kraus extrapolation technique
and Shedlovsky technique, respectively. Sah et al. studied the molar
conductance of KAc in aqueous 2-butanol solutions with an alcohol
mass fraction of 0.7, 0.8 and 0.9 at 25°C, 30°C, and 35°C respectively,
and the limiting molar conductivities, ion-association constants were
also estimated by using Fuoss conductance–concentration equation
[22]. More practical information on properties of KAc in some other or-
ganic solvents (with the lower boiling temperature and lower latent
heat of vaporization than water) are desiderated, so as to estimate
their usability for the closed type RED power generation system driven
by thermal energy.
The purpose of this work is to explore some new available working
solutions for above introduced power generation system by means of
measuring and analyzing two key parameters, the electric conductivity
and electric convertibility. Here, KAc is selected as the solute, and the
following substances are selected as the solvents, including water,
ethanol (C
2
H
5
OH), 2,2,2–trifluoroethanol (C
2
H
3
F
3
O, also named TFE),
2–propanol (C
3
H
8
O, also named IPA) as well as their several binary
blends. The experimental temperature and concentration conditions
are variable, and which are set carefully according to the practical
operating requirements of the working solutions looping in the power
generation system as Fig. 2 shown.
2. Electric Conductivity of Solutions
2.1. Experimentation
2.1.1. Experimental materials and apparatus
Purity of NaCl is ≥99.5% (Sinopharm Chemical Reagent, China); KAc
is analytical purity (Tianjin DaMao Chemical Reagent, China); Purity of
TFE is ≥99.5% (Jiangsu Blue-Star Green Tech., China) and purity of IPA
is ≥99.7% (Tianjin DaMao Chemical Reagent, China). Anhydrous ethanol
is better than 99.7% (Tianjin Guangfu Tech., China). Deionized water
was produced in laboratory, and the electric conductivity of which is
less than 1 μS·cm
−1
. Both NaCl and KAc were dried in an electrical
oven at 423.15 K for more than 3 h with periodic grinding. All experi-
mental materials were weighed by using an electronic analytical
balance (Ohaus, U.S.) with accuracy of 0.001 g. Some key thermophysical
properties of these solvents are listed in Table 1 [23].
Fig. 3 showed the test rig schematic of electric conductivity measure-
ment. A jacketed glass reaction vessel was designed to contain the
solution hermetically. A thermal resistance temperature sensor (with
accuracy of ±0.15°C) and a corrosion resistance electrode (Model: inlab
710) were inserted into the solution directly. The tail of electrode was
connected with a conductivity meter (Model: FE38-Standard; available
range: 10
−5
mS·cm
−1
to 500 mS·cm
−1
), both of which were produced
by Mettler Toledo (Switzerland) with accuracy of ±0.5%. The reaction
vessel was placed on the platform of a stirrer (Tianjin Honour Instrument,
China) with rotating speed variation of 0–2000 r·min
−1
. The stirring rod
Table 1
Some key thermophysical properties of these four solvents [23]
Solvents Molecular
mass
Normal boiling
temperature/ °C
Latent heat of vaporization
@20°C/kJ·kg
−1
Ethanol 46.07 78.42 926.01
TFE 100.04 74.85 458.82
IPA 60.01 82.35 762.73
Water 18.015 99.97 2453.5
Fig. 3. Experimental appara tus of conductivity meas urement. 1 Glass reaction v essel;
2 stirrer; 3 stirri ng rod; 4 thermost atic bath; 5 cycling water pump; 6 temperature
sensor; 7 electrode; 8 conductivity meter; 9 data acquisition system.
2583X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

was covered by polytetrafluoroethylene outside to guard against reaction
with the solution. Temperature of solution was maintained by looping
water supplied from thermostatic bath (Julabo Instrument, Germany)
with temperature stability of ± 0.03°C. The maximal temperature
fluctuation of solution was 0.2°C during the whole experimental pro-
cess. Computer, programmable logic controller (Siemens S7–200,
Germany) and some matched modules were used to acquire data and
control testing conditions.
2.1.2. Reliability test and error analysis
Reliability tests were carried out to confirm the availability of this
experimental method and apparatus. Electric conductivity of NaCl
aqueous were tested out under different concentrations at 18°C firstly
and then compared with the literature [24],seeninFig. 4. The measured
results and the reference data were closed to each other with an average
relative error of 2.32%.
Uncertainty of measured conductivity data has also been analyzed
and drawn along with the experimental results in the figures. The
experimental error can be calculated by Eq. (3).
Δκ¼XEþXTþXcþXIð3Þ
where Δκis the error of displayed value of conductivity meter; X
E
comes
from the error of repetition test; X
T
comes from the error of solution
temperature condition; X
C
comes from the error of standard solution;
X
I
is the inherent error of testing system.
The first part of uncertainty of measured data caused by error of
repetition tests that can be calculated by Eqs. (4) and (5) [25, 26],
uX
E
ðÞ
¼Si=nð4Þ
Si¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X
n
i¼1
κi−κave
ðÞ
2
!,n−1ðÞ
v
u
u
tð5Þ
where S
i
is the standard deviation; nis the total times of tests; κ
i
is
conductivity value under the test of sequence number i, mS·cm
−1
;
κ
ave
is the average conductivity value, mS·cm
−1
.
The second part of uncertainty of measured data caused by solution
temperature condition can be calculated by Eqs. (6) and (7) [26].
uX
T
ðÞ¼κRαTΔT=ffiffiffi
3
pð6Þ
αT¼κT−κR
κRtT−tR
ðÞ ð7Þ
where κ
R
and κ
T
are conductivity value at reference temperature and
current solution temperature, mS·cm
−1
;α
T
is temperature coefficient
of conductivity measurement; △Tis the uncertainty of solution tempera-
ture, °C; tand t
R
are the current temperature and reference temperature
of solutions, °C.
The third part of uncertainty of measured data caused by standard
solution, can be calculated by [25].
uX
C
ðÞ¼κRuE=ffiffiffi
3
pð8Þ
where u
E
is the expanded uncertainty of standard solution. In this test,
standard solutions certificated by Mettler Toledo Co. were used, with an
expanded uncertainty of 1.5% at 25°C.
The forth part of uncertainty of measured data is inherent error of
testing system that can be predicted by Eq. (9) [25].
uX
I
ðÞ¼κRuT=ffiffiffi
3
pð9Þ
where u
T
is the accuracy of apparatus, given as ± 0.5% of the measured
value in this work.
The compound uncertainty u(Δκ) can be estimated as Eq. (10).
uΔκðÞ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
∂Δκ
∂XE

2
u2XE
ðÞþ∂Δκ
∂XT

2
u2XT
ðÞþ∂Δκ
∂XC

2
u2XC
ðÞþ∂Δκ
∂XI

2
u2XI
ðÞ
s
ð10Þ
The expanded uncertainty U(Δκ) of measurements can be estimated
by Eq. (11).
UΔκðÞ¼2uΔκðÞ ð11Þ
2.2. Results and discussions on electric conductivity
2.2.1. Electric conductivity of KAc in water
Electric conductivity of aqueous KAc solution have been tested out
and listed in Table 2 under different temperature (20°C, 40°C and
60°C) and mass concentrations conditions (from 0.5% to 60%), with the
unit of mS·cm
−1
(1 mS·cm
−1
= 1000 μS·cm
−1
). The maximum tested
value is 273.2 ± 5.451 mS·cm
−1
at the concentration of 40.11 wt.%
when the solution temperature is 60°C. The tested peak values of electric
conductivity are (133.8 ± 2.544) mS·cm
−1
and (202.3 ±
3.957) mS·cm
−1
at 20°C and 40°C respectively. The relationships of the
concentration, temperature and electric conductivity of solutions are
drawn in Fig. 5, with the experimental uncertainty.
01234
0
50
100
150
200
y
t
i
v
i
tcud
n
o
C / mS·cm
-1
Concentration /mol
·
L
-1
literature
this work
Fig. 4. Conductivity data comparison of NaCl aqueous solution at 18 °C.
Table 2
Conductivity data of aqueous KAc solution under different conditions
20°C 40°C 60°C
C/ wt.% κ/ mS·cm
−1
C/ wt.% κ/mS·cm
−1
C/ wt.% κ/ mS·cm
−1
0.50 4.67 0.68 9.092 0.53 9.521
1.03 9.13 1.01 13.12 1.00 17.34
2.04 17.11 1.98 24.29 2.13 34.50
4.86 37.59 5.17 57.01 5.00 73.17
9.73 68.40 10.06 99.55 10.23 133.3
15.04 94.68 15.16 136.1 15.03 178.8
20.03 113.3 20.02 163.3 20.20 217.4
25.06 125.1 25.12 184.7 25.12 245.0
30.00 133.4 30.10 197.7 30.10 262.6
35.04 133.8 35.04 202.3 35.04 273.0
40.11 129.5 40.11 200.4 40.11 273.2
45.10 120.7 45.10 191.0 45.10 264.5
50.02 108.5 50.02 175.0 50.02 247.0
54.99 91.60 54.99 153.2 54.99 222.2
60.00 72.88 60.00 127.8 60.00 190.8
2584 X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

As Fig. 5 shown, firstly, temperature of solution is a positive impact
factor on the electric conductivity under the testing conditions. The
most powerful influence region of temperature on the electric conduc-
tivity of aqueous KAc solution is Region I in Fig. 5, where the peak values
of the electric conductivity appear. In contrast, the influence of solution
temperature is quite feeble on the electric conductivityof the extremely
diluted concentration solutions, seen Region II in Fig. 5.
Secondly, with the increasing of solution concentration, electric
conductivity increases at the first half time and then reaches the peak
value, but later it decreases gradually, as Fig. 5 displayed. The reason
can be explained by using the free ions theory. Solvent can be regarded
as a continuous molecular medium in solution, theoretically, which
provides the spaces for motions of free ions. Electric conductivity of
solution usually depends on the amount of free ions and their transport
velocities [27]. The amount of free ions is found to be affected signifi-
cantly by three factors, the ionic concentration, ionic association
and solvation [28].Theinfluence of ionic association and ionic solvation
are negligible when the solution is in low concentration, since that
the distances are large enough among the free cation ions and anion
ions. At this time electric conductivity can be enlarged if the amount
of free ions is increased by adding more crystal KAc into its aqueous
solution. While, the curve slopes of electric conductivity turn down to
be negative for the extremely concentrated solutions. At this moment,
cation and anion ions (also including hydrated ions) are closed to each
other, and ionic association may occur between the ions with opposite
charges.
Ion pairing in aqueous KAc solutions exists, as described by
equilibrium Eq. (12) of Bjerrum type. Products of ionic association
are electroneutral, which have no positive contribution to electric
conductivity of solutions.
KþþCH3COO−⇌CH3COOK ð12Þ
Influence of ionic solvation also become obvious when solution is in
high concentration. At this situation, the relaxation effect and electro-
phoretic effect increase with the reduction of interionic distances [27],
and the drift of ions is retarded, thus resulting in the decrease of electric
conductivity of solution. Besides, different with the hydration process
of normal alkalis salt aqueous solution, as description in Eq. (13),
CH
3
COO
—
ions (proton acceptors) exist in solution that can be connected
with H
+
, according to Brönsted–Lowry's acid–base theory and Robinson–
Harned's localized hydrolysis hypothesis [29]. Due to this connection, as
Eqs. (14) and (15) show, the number of free ions are further reduced,
consequently leading to the electric conductivity reduction.
KþþH2O→Kþ⋯OH−
⋯Hþð13Þ
Kþ⋯OH−
⋯HþþCH3COO−→Kþ⋯OH−
⋯Hþ⋯CH3COO−ð14Þ
KþþH2OþCH3COO−⇌Kþ⋯OH−
⋯Hþ⋯CH3COO−ð15Þ
2.2.2. Electric conductivity of KAc in ethanol, TFE and IPA
Electric conductivity of KAc in organic solvents (ethanol, TFE, IPA)
was measured under varying temperature (20 to 60°C) and concentra-
tions (up to 15.09%) conditions, and the results are listed in Table 3.KAc
still can be ionized in these organic solvents, while the effective trans-
port amount of cation ions (K
+
) and anion ions (CH
3
COO
—
) are not as
abundant as that of aqueous KAc solution. The measured electric
conductivity of KAc–ethanol and KAc–TFE solutions are both smaller
0 102030405060
0
50
100
150
200
250
300
Region II
Mass concentration / %
Sm/ytivitcudnoccirtcelE
g
cm
-1
Tested @20
ć
Tested @40
ć
Tested @60
ć
Fitted @20
ć
Fitted @40
ć
Fitted @60
ć
I Uncertainty
Region I
Fig. 5. Electric conductivity of aqueous KAc solution under varying conditions.
Table 3
Electric conductivity of KAc in three organic solvents
Solutions 20°C 40°C 60°C
C/
wt.%
κ/
mS·cm
−1
C/
wt.%
κ/
mS·cm
−1
C/
wt.%
κ/
mS·cm
−1
KAc–TFE 0.14 0.232 0.14 0.361 0.10 0.321
0.37 0.515 0.37 0.636 0.28 0.697
0.46 0.601 0.46 0.811 0.51 1.019
0.71 0.819 0.71 1.060 0.71 1.287
0.96 1.008 0.96 1.365 0.95 1.585
1.98 1.626 1.98 2.275 1.97 2.762
2.90 2.100 2.90 3.053 2.91 3.770
3.84 2.610 3.84 3.742 3.75 4.635
4.66 2.965 4.66 4.310 4.66 5.5306
5.71 3.323 5.71 4.920 5.64 6.430
6.45 3.536 6.45 5.313 6.47 7.118
7.23 3.733 7.23 5.677 7.27 7.711
8.20 3.860 8.20 6.022 8.21 8.292
8.99 3.949 8.99 6.230 8.95 8.660
9.82 3.986 9.82 6.392 9.82 8.987
10.57 3.982 10.57 6.444 10.56 9.213
11.31 3.936 11.31 6.450 11.32 9.333
12.11 3.871 12.11 6.409 12.12 9.385
12.82 3.792 12.82 6.371 12.85 9.402
13.59 3.688 13.59 6.279 13.63 9.402
14.25 3.599 14.25 6.181 14.31 9.358
15.05 3.448 15.05 6.030 15.09 9.273
KAc–ethanol 0.12 0.175 0.17 0.274 0.31 0.489
0.31 0.371 0.29 0.413 0.55 0.704
0.49 0.498 0.50 0.593 0.70 0.833
0.70 0.627 0.60 0.680 0.97 1.028
0.97 0.767 0.96 0.914 1.98 1.639
1.94 1.197 1.90 1.422 2.85 2.104
2.83 1.491 3.03 1.926 3.68 2.535
3.87 1.742 4.00 2.303 4.69 2.996
4.71 2.010 4.72 2.588 5.56 3.363
5.63 2.252 5.57 2.882 6.48 3.740
6.50 2.432 6.46 3.160 7.29 4.049
7.29 2.582 7.24 3.388 8.18 4.373
8.17 2.733 8.33 3.681 9.04 4.671
8.93 2.858 8.99 3.843 9.77 4.908
9.77 2.974 9.74 4.009 10.57 5.165
10.56 3.085 10.73 4.220 11.34 5.388
11.31 3.159 11.30 4.329 12.17 5.618
–– 12.15 4.483 12.82 5.795
–– 12.82 4.596 13.59 5.985
–– 13.56 4.708 14.30 6.150
–– –– 15.02 6.310
KAc–IPA 0.30 0.021 –– 0.30 0.024
0.49 0.028 –– 0.49 0.033
0.68 0.035 –– 0.68 0.041
1.00 0.044 –– 1.00 0.052
1.43 0.054 –– 1.43 0.065
1.91 0.065 –– 1.91 0.081
2585X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

than that of aqueous KAc solution, and most of which are less than
10 mS·cm
−1
under the testing conditions. The maximum electric
conductivity of KAc–TFE solution is (9.402 ± 0.189) mS·cm
−1
at 60°C,
and it is (6.310 ± 0.126) mS·cm
−1
for KAc–ethanol solution under
the same temperature. The solubility of KAc in IPA is less than 1.91% in
weight percentage, and the electric conductivity of KAc–IPA solution is
quite small, with the maximum value of (80.8 ± 1.63) μS·cm
−1
at
60°C under this experimental conditions. KAc seems to prefer water
molecules to ethanol or TFE molecules in their solvation, and resist dis-
solving in IPA. The extremely weak ability of ionization, dissolution and
conduction of KAc–IPA solution would lead to the quite low output volt-
age, power density, and energy efficiency of the RED power generation
system, thus the KAc–IPA solution should be excluded.
The relationships of the concentration, temperature and electric
conductivity of these solutions are drawn in Figs. 6–8,withuncertainty
analysis. Solution temperature is still observed to take a positive impact
role for electric conductivity of solutions regardless of whether the
solvent is water or organics. The influence mechanism can be explained
as [27, 28]: with the increase of the temperature of solution, (I) the vis-
cosity of solution decreases, bringing down the friction and irreversible
loss during the process of ions transport; (II) the solubility of some
electroactive species from the environment decreases, promoting the
reduction of external disadvantageous interference; (III) the average
kinetic energy of ionsincreases, leadingto the reinforcement of thermo-
dynamic movement of ions; (IV) electrophoretic effect becomeweaken,
reducing the obstruction of ions transport; (V) effects of hydrolysis and
solvolysis become weaken, shrinking the effective interactive radius
and retarding interionic constraint.
From Figs. 5-8, it can be seen that the electric conductivity of these
four binary solutions is in the ranking of KAc–H
2
ONKAc–TFE NKAc–
ethanol NKAc–IPA, which is coincident with the relationship of their
solubility. The results can be analyzed by introducing a parameter
Bjerrum length (β). βrepresent the separation at which the electrostatic
interaction between two elementary charges is comparable in magnitude
to the thermal energy scale [30], which is described as Eq. (16),
β¼e2=4πε0εrKBTðÞ ð16Þ
where K
B
is the Boltzmann constant; eis the elementary charge; ε
0
is the
vacuum permittivity; and ε
r
is the relative dielectric constant. The value
of ε
r
can be found from references [17, 31–33], and the βof solvents
(water, TFE, ethanol, IPA as well as the mixture of 90%water and 10%
TFE) are calculated and shown in Fig. 9 under the temperature from
20°C to 60°C. Compared Fig. 9 with Table 3, it can be concluded that
solution consist of KAc and some a solvent that with the relative small β
value may possess the better electroconductibility. This conclusion
is helpful for selecting new suitable solvents for the closed type RED
power generation system driven by thermal energy.
The phenomena of ionic association and solvation still exist for the
solutions consist of KAc and organic solvents. Thermal motion and
interionic forces establish a steady state, and each ion in solution are
in one of three categories, the unpaired ion, the solvent separated pair
or the contact pair, basing on Fuoss theory [34],asEq.(17) described.
KþþCH3COO−⇌Kþ⋯CH3COO−⇌KþCH3COO−ð17Þ
Accordingto the hypothesis of localized hydrolysis [29, 35], the water
molecules in cation hydration shell can be polarized, and then a part of
0 2 4 6 8 10121416
0
2
4
6
8
10
1 mS
g
cm
-1
=10
3
µS
g
cm
-1
Mass concentration / %
S
m
/y
t
i
v
itcudnocc
i
rt
c
e
lE
g
cm
-1
Tested @20
ć
Tested @40
ć
Tested @60
ć
Fitted @20
ć
Fitted @40
ć
Fitted @60
ć
I uncentainty
Fig. 6. Electric conductivity of KAc–TFE solution under varying conditions.
0 2 4 6 8 10121416
0
1
2
3
4
5
6
7
1 mSgcm-1=103 µSgcm-1
Mass concentration / %
Sm/ytivitcudnoccirtcelE
g
cm-1
Tested @
ć
Tested @
ć
Tested @
ć
Fitted @
ć
Fitted @40
ć
Fitted @60
ć
I Uncertainty
Fig. 7. Electric conductivity of KAc–ethanol solution under varying conditions.
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Mass concentration / %
1 mSgcm-1=103 µSgcm-1
S
m/yt
i
v
i
tcu
dn
occir
tce
lE
g
cm
-1
Tested @20ć
Tested @60ć
Fitted @20ć
Fitted @60ć
I uncertainty
Fig. 8. Electric conductivity of KAc–IPA solution under different conditions.
2586 X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

anions can interact with hydrogen atom of the polarized water. This
leads to the formation of a solvent separated ionic pair as follows [36].
ð18Þ
The hypothesis of localized hydrolysis can be considered as a particular
case for aqueous solutions of the general pattern of ionic association,
named localized solvolysis [37]. Ionic association, taking place by localized
solvolysis patternasbothcationandanionoftheionpairs,areconnected
with different parts of the separated solvent molecules by donor-acceptor
bonding. The higher the concentration of solution is, the more remarkable
the connections are [37]. In solvents of ethanol and TFE, the ionic pair are
likely to be in the structures as Eqs. (19) and (20) shown respectively.
ð19Þ
ð20Þ
Consequently, for the extremely concentrated solutions, the amount
of free ions with opposite charges are decreased due to the connections
of ion pairs with the separated parts of the solvent molecules, and
resulting in the reduction of electric conductivity. In order to increase
the amount of free ions and electric conductivity, it is necessary to take
action to create the protective solvating shells around ions to prevent
the interionic attraction and connection. Two examples of the effective
actions are to add surfactants and use the solvents with small βvalue.
Next section, blending solvents will be tried, due to their appropriate β
value and possibility of facilitating a protective construction.
2.2.3. Electric conductivity of KAc in blending solvents
Electric conductivity of KAc in four bending solvents (90% H
2
O–10%
TFE, 80% H
2
O–20%TFE, 90% H
2
O–10% ethanol, and 80% H
2
O–20% ethanol)
have been measured, and the results are listed in Table 4.
From Fig. 10, it can be seen that the goal of improving the electric
conductivity of KAc in organic solvents can be really achieved by means
of adding a certain proportion of water, and the more the proportion of
water in blending solvent, the larger the electric conductivity of the mea-
sured solution. Here it must be emphasized again, the pure water is not an
ideal absorbent for the closed type RED power generation system driven
by thermal energy, due to its relative huge latent heat of vaporization
20 25 30 35 40 45 50 55 60
0.5
1.0
1.5
2.0
2.5
3.0
3.5
H2O
TFE
Ethanol
IPA
TFEH
2
O
htgnelmurrejB
Temperature/
Fig. 9. Bjerrum length of several solvents.
Table 4
Electric conductivity results of KAc–H
2
O–ethanol and KAc–H
2
O–TFE
KAc–H
2
O–TFE KAc–H
2
O–ethanol
90%H
2
O–10%TFE 80%H
2
O–20%TFE 90%H
2
O–10%
ethanol
80%H
2
O–20%
ethanol
C/wt.% κ/
mS·cm
−1
C/ wt.% κ/
mS·cm
−1
C/ wt.% κ/
mS·cm
−1
C/ wt.% κ/
mS·cm
−1
0.32 2.734 0.16 1.226 0.11 0.755 0.09 0.418
0.50 4.115 0.31 2.314 0.31 2.129 0.25 1.215
0.70 5.635 0.48 3.570 0.49 3.224 0.56 2.644
0.98 7.731 0.68 4.920 0.73 4.789 0.82 3.755
1.94 14.33 1.06 7.415 1.04 6.603 1.21 5.382
2.86 20.35 1.93 13.01 1.96 11.89 2.00 8.506
3.82 26.36 2.83 18.18 2.91 17.08 2.88 11.86
4.67 31.48 3.81 23.69 3.75 21.33 3.93 15.63
5.58 36.77 4.75 28.53 4.71 26.13 4.86 18.94
6.42 41.22 5.64 33.26 5.57 30.34 5.84 22.28
7.32 46.09 6.39 36.88 6.46 34.46 6.58 24.72
8.12 50.21 7.27 41.11 7.47 38.92 7.42 27.46
8.93 54.19 8.21 45.27 8.85 44.92 8.47 30.78
9.80 58.41 8.94 48.62 9.69 48.39 9.22 33.07
10.60 62.12 9.77 51.93 10.52 51.75 10.09 35.71
11.41 65.56 10.54 55.23 11.26 54.59 10.76 37.68
12.08 68.46 11.33 58.19 12.09 57.67 11.64 40.25
12.82 71.58 12.08 61.17 12.91 60.55 12.49 42.59
13.57 74.66 12.84 63.79 13.61 63.23 13.44 45.22
14.35 77.70 13.58 66.34 14.29 65.55 14.66 48.42
15.01 80.10 14.31 68.51 14.91 67.64 15.38 50.23
16.41 85.23 15.02 70.97 16.37 72.38 16.85 53.88
18.09 90.64 16.43 75.19 18.00 77.16 18.38 57.44
19.70 95.65 18.09 79.82 19.66 81.59 20.13 61.23
21.25 100.3 19.67 83.71 21.26 85.40 21.72 64.52
22.72 103.8 21.41 87.53 22.69 88.59 23.29 67.39
24.17 106.8 22.79 90.27 24.09 91.43 24.67 69.76
25.57 109.4 24.21 92.95 25.40 93.86 26.06 71.89
28.22 113.3 25.55 94.87 28.04 97.70 28.71 75.57
30.63 115.4 28.17 97.78 30.55 100.6 31.24 78.24
32.91 116.5 30.65 99.93 32.80 102.3 33.54 80.01
35.06 116.6 32.95 100.5 34.93 102.8 35.61 81.25
36.07 116.3 35.03 100.2 –– ––
37.08 115.7 –– –– ––
0 5 10 15 20 25 30 35 40 45 50
0
20
40
60
80
100
120
140
↑ KAc
—
ethanol
↓ KAc
—
TFE
KAc aqueous solution →
S
m
/
ytiv
i
tc
u
dnocci
r
tce
lE
g
cm
-1
Mass concentration / %
KAc
—
water(90%)
—
TFE(10%)
KAc
—
water (80%)
—
TFE(20%)
KAc
—
water (90%)
—
ethanol(10%)
KAc
—
water (80%)
—
ethanol(20%)
Fitted for water (90%)
—
TFE(10%)
Fitted for water(90%)
—
ethanol(10%)
Fitted for water (80%)
—
TFE(20%)
Fitted for water(80%)
—
ethanol(20%)
IUncertainty
Fig. 10. Electric conductivity of KAc in blending solvents under vary conditions.
2587X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

(needing much more heat input during the process of solution regener-
ation) and its relative high operating temperature (requiring the heat
source with a higher temperature or operating the system under the
negative pressure state), seen in Table 1.
Besides, as illustrated in Fig. 10 visibly, electric conductivity of solu-
tion KAc–H
2
O–TFE is larger than solution of KAc–H
2
O–ethanol at the
same experimental condition. In addition, the electric conductivity
curves of KAc–H
2
O–TFE are closer to the aqueous KAc solution than
that of the KAc–H
2
O–ethanol solution, and far above than the solution
of KAc in either pure TFE or pure ethanol. If using 20% ethanol (or
TFE) to replace 20% water, then the maximum electric conductivity of
KAc–H
2
O–ethanol solution reduces about 40% relative to the aqueous
KAc solution, and as for solution KAc–H
2
O–TFE, it is only about a quarter
reduction in its electric conductivity. While at this situation the good
news is that the thermal energy consumptions during the solution
regeneration process (supposing at 80°C) are only a half for of a RED
power generation system working with KAc–H
2
O–ethanol, and only a
third for KAc–H
2
O–TFE, compared to the aqueous KAc solution. Another
benefit of using these blending solvents rather than the pure water is
that the whole system can be operated in positive pressure state so
that it is not necessary to particularly enhance the leakproofness of
the whole system.
2.3. Electric conductivity estimation
Electric conductivity of diluted solutions can be described by using
Kohlrausch equation [28], Onsager limiting equation [28],Fuoss
conductance–concentration equation [22], Fuoss–Hsia equation [21],
Foss–Chen–Justice equation [38],etc. with the accepted accuracy.
However, most of those equations are no longer available for estimating
the electric conductivity of concentrated solutions. Researchers in this
field made a major improvement over the classical theories by means
of formulating a linear response theory in which Onsager continuity
equations were combined either with the mean spherical approxima-
tion (MSA) or the hypernetted chain equations (HNC) [39, 40].This
method was applied by Bernard et al. to study the self–diffusion [40]
and electrical conductance [41] for 1–1 binary electrolyte solutions
(such as aqueous NaCl solution or aqueous KBr solution), with the
concentration below 1 mol·L
−1
. Dufrêche et al. found the combined
Smoluchowski–MSA theory of the primitive model was able to describe
simultaneously the different transport and equilibrium properties of
aqueous alkali chloride (LiCl, NaCl and KCl) solutions with higher
concentrations 1–2mol·L
−1
[42].Shiet al. further improved the tradi-
tional Brownian dynamics simulation method for electrolyte solution
estimation on self-diffusion coefficientand molar conductivity by taking
the hydrodynamic interaction effect into account, and resulting a good
agreement between the simulation values and the tested data for both
aqueous NaCl solution and aqueous KCl solution up to 3 mol·L
−1
[43, 44]. Gao et al. extended the Dufrêche et al.’s work by adding the
parameter of effective cationic diameter (was a function of total ionic
strength), and by means of which, the mutual diffusion coefficients
of 18 uni-valence electrolyte solutions were investigated under a wider
concentration ranges (0–4mol·L
−1
)[39].
Some other efforts were made on developing the electric conductivity
theories of mixtures of electrolytes in aqueous solutions. By following Wu
et al.'swork[45] and Young's rule [46], Miller found that the simple linear
approximations to the specific conductance of a mixture solution (NaCl–
MgCl
2
–H
2
O, 1–2mol·L
−1
) could be written in terms of various solute
fractions (molar, equivalent, or ionic strength) and the specific conduc-
tance of its constituent blending systems [47]. Young's rule was also
developed by Chen et al. to predict the conductivity of ternary solutions
of [PP
1,6
]Br (N-hexyl, methylpyrrolidinium bromide) –[PP
1,4
]Br (N-butyl,
methylpyrrolidinium bromide) –H
2
O[48].Huet al. proposed a simple
equation for predicting the electric conductivity of ternary electrolyte
solution based on the semi-ideal solution theory and Eyring absolute
rate theory, and the accuracy of which was verified by comparing
the predicted results with the tested data on NaCl–LaCl
3
–H
2
Osolution
[49],KCl–CdCl
2
–H
2
O solutions [47],and[C
6
mim][Cl] (1–hexyl–3–
methylimidazolium chloride) –[C
6
mim][BF
4
](1–hexyl–3–
methylimidazolium tetrafluoroborate) –H
2
Osolution[50].
Here, a modified amplitude version of Gaussian peak function
is used to estimate the electric conductivity of binary and ternary
solutions under the wide range of solution concertation. Besides, the for-
mat of Arrhenius equation is referred due to its advantage on indicating
the impact of temperature factor on chemical reaction rate of solution.
The electric conductivity estimation equation is expressed as the follows,
κT¼A1TþA2Texp −1
2
Cx−Cm
A3T

2
! ð21Þ
where κ
T
is conductivity value at current solution temperature T(K),
mS·cm
−1
;A
1
,A
2
and A
3
is offset coefficient, amplitude coefficient and
width coefficient; C
x
is the current solution concentration, by mass
percentage; C
m
, is the solution concentration at maximum conductivity,
by mass percentage. The coefficients of Eq. (21) are listed in Table 5.
The calculated results of electric conductivity of solutions have been
drawn in lines and compared with the measured results, which can be
seen in Figs. 5–8 and 10. Almost all of the measured data are located
nearby the estimated lines under different temperature and concentra-
tion conditions, and most of the average relative errors are less than
Table 5
Coefficients of equation for estimating electric conductivity of KAc solutions
Solvents Conditions Coefficients of Eq. (21) Average relative errors
A
1
A
2
A
3
C
m
In H
2
O 20°C −1672.2408 1672.7010 5.1566 35.0543 3.48%
40°C −1488.4386 1489.0883 4.0108 36.7771 1.81%
60°C −1961.9588 1962.7859 3.9825 38.1504 1.65%
In ethanol 20°C −15.7167 15.7275 1.2264 12.6671 2.10%
40°C −17.9276 17.9427 1.2890 15.9847 2.93%
60°C −13.0889 13.1090 1.1410 20.4263 2.50%
In TFE 20°C −51.6037 51.6175 1.5989 10.4394 2.22%
40°C −24.0013 24.0221 0.9058 11.5239 1.28%
60°C −0.14270 0.17122 0.0648 12.7898 1.30%
In IPA 20°C −0.06614 0.06641 0.1414 3.48928 1.48%
60°C −0.04341 0.04378 0.1225 5.01793 2.78%
In H
2
O–TFE 9:1 −183.7645 184.1622 1.7796 34.0819 1.42%
8:2 −367.0622 367.4033 2.6047 32.6760 3.24%
In H
2
O–ethanol 9:1 −134.2850 134.6358 1.7095 36.0483 1.32%
8:2 −14.63472 14.91774 0.71996 41.17836 1.38%
2588 X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

3% (kick off the first test point), and the total average relative error
between the measured results and estimated data is about 2.06%.
3. Electric Convertibility of Solutions
3.1. Experimentation
In order to test out the electric convertibility (from salinity gradient
energy to electric energy) of the solutions consist of KAc (as solute)and
H
2
O, TFE, ethanol as well as their binary blends (as solvents), a test rig
has been designed and built, as Fig. 11 illustrated. The RED cells was
the key component in this experimental system. Totally 11 CEMs and
10 AEMs (Asahi glass, Japan) were used, and each IEM was in size of
20 cm × 10 cm and with effective area of 78.75 cm
2
.IEMwasseparated
by polyamide woven spacer (Tianwei membrane Tech., China) and
silicone gasket combination with the total thickness of 0.35 mm.
Endplates and electrodes were made of polymethyl methacrylate and
titanium plates coated with Ru/Ir mixed metal oxide. Peristaltic
pumps (Longer Pump, China) were used to feed the weak and strong
solutions into the RED cells with the velocity of 0.1 cm·s
−1
,andto
cycle the electrode rinse solutions with flow rate of 100 ml·min
−1
.
Electrode rinse solution was mixed by aqueous solution of KAc, K
4
Fe
(CN)
6
and K
3
Fe(CN)
6
. Three solutions tanks were placed into a thermo-
static bath (Julabo, Germany) and controlled at293.15 K. Power output
of the RED cells was measured by the potentiostat (CHI 660E, Chenhua,
China). The chronopotentiometry (CP) was applied to measure the
internal resistance (R
stack
), open circuit voltage (E
OCV
), and terminal
voltage (E
Vact
)[51]. The gross output power (P) and power density
(P
d
) can be estimated as follows:
I¼EOCV=Rstack þRext
ðÞ ð22Þ
P¼EVact I¼EOCV−RstackIðÞIð23Þ
Pd¼P=NIP Ae
ðÞ ð24Þ
where R
ext
is external load resistance, Ω;N
IP
is number of IEM pairs; A
e
is the effective area of IEM.
Several pre-experiments were carried out by using aqueous NaCl
solution or aqueous LiCl solution to test the usability of this experimental
system as well as to reduce the inherenterrorsaslittleaspossible.Much
more details about the experimentation can be seen in the previous
introductions [52].
3.2. Results and discussions
3.2.1. Variations of voltage and power density
The output voltage and power density of the RED cells working with
six working solution pairs have been tested out under the above intro-
duced experimental conditions, and the six working solution pairs are
(1) aqueous KAc solutions (0.05 mol·kg
−1
as weak solution and
2.5 mol·kg
−1
as strong solution); (2) KAc in blending solvents consist
of 90% water and 10% ethanol (0.05 mol·kg
−1
as weak solution and
2.5 mol·kg
−1
as strong solution); (3) KAc in blending solvents consist
of 80% water and 20% ethanol (0.05 mol·kg
−1
as weak solution and
2.5 mol·kg
−1
as strong solution); (4) aqueous KAc solutions
(0.05 mol·kg
−1
as weak solution and 5 mol·kg
−1
as strong solution);
(5) KAc in blending solvents consist of 90% water and 10% TFE
(0.05 mol·kg
−1
as weak solution and 5 mol·kg
−1
as strong solution);
(6) KAc in blending solvents consist of 80%water and 20% TFE
(0.05 mol·kg
−1
as weak solution and 5 mol·kg
−1
as strong solution).
The test results have been shown in Figs. 12 and 13.

Electrode
Rinse
Solution
Tan k
Wea k
Solution
Tan k
Strong
Solution
Tan k
Brackish
Solution
Tan k
Peristaltic
Pump
RED Cells
Potentiostat
Thermostatic
bath
Fig. 11. Schematic diagram of test rig to measure electric convertibility of solutions.
0.00 0.04 0.08 0.12 0.16 0.20 0.24
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
KAc in water (0.05:2.5, mol · kg
-1
)
KAc
—
water(90%)
—
ethanol(10%)
KAc
—
water (80%)
—
ethanol (20%)
KAc in water (0.05:5, mol · kg
-1
)
KAc
—
water (90%)
—
TFE (10%)
KAc
—
water (80%)
—
TFE (20%)
E / V
I / A
Fig. 12. Variation relationships of voltage with current for different solutions.
0.00 0.05 0.10 0.15 0.20 0.25 0.30
-0.50
-0.25
0.00
0.25
0.50
0.75
1.00
1.25
1.50
KAc in water (0.05:2.5, mol · kg
-1
)
KAc
—
water(90%)
—
ethanol(10%)
KAc
—
water(80%)
—
ethanol (20%)
KAc in water (0.05:5, mol · kg
-1
)
KAc
—
water(90%)
—
TFE(10%)
KAc
—
water(90%)
—
TFE (20%)
Pdm·W/ -2
I / A
Fig. 13. Variation relationships of power density with current for different solutions.
2589X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

First of all, as Fig. 12 drawn, the output voltage of this RED cells is
decreased with the increase of current, and their variation relationships
are almost linear. Secondly, the output voltage is observably influenced
by the solvents. RED cells working with solution pair (4) (0.05 mol·kg
−1
as weak solution and 5 mol·kg
−1
as strong solution) possesses the
highest output voltage than the other five solution pairs. With the
increase of percentage of the TFE (or ethanol) in the blending solvents,
the output voltage decreases gradually under this tests. Thirdly, the
output voltage is distinctly influenced by the concentration difference be-
tween the feeding week solution and the feeding strong solution. Taking
the results gained from the tests under working pair (4) (aqueous KAc
solutions, 0.05 mol·kg
−1
: 5 mol·kg
−1
) and working pair (1) (aqueous
KAc solutions, 0.05 mol·kg
−1
:2.5mol·kg
−1
) into to comparison, it can
be found that the average output voltage of former (symbol: open square
in Fig. 12) is about twice than that of the latter (symbol: solid square in
Fig. 12). It is worth mentioning of that the predicted output voltage
cannot be always enlarged with the increase of concentration difference
between the feeding solutions.
The output power density of this RED cells is found to be increased
with the increase of current firstly, and then reach to the peak value,
but later it goes down, seen in Fig. 13. RED cells working with aqueous
KAc solution possess the highest power density than either the KAc–
H
2
O–TFE solution or the KAc–H
2
O–ethanol solution. The more the TFE
(or ethanol) percentage in bending solvents is, the smaller the output
power density of this RED cells. Besides, it is one of ways to enlarge
the power density by extending the concentration difference between
the feeding strong and weak solutions within a certain range. Finally,
the convex parabolic curve relationship in Fig. 13 indicates that the
maximum power density exists.
3.2.2. Maximum power density and open circuit voltage
The maximum power density (P
dmax
) and open circuit voltage
(E
OCVmax
) of this experimental system running with the working
solutions consist of KAc, water, TFE, ethanol, or their blends have been
tested out and listed in Table 6. The following results are gained: (1) It
can be seen that electric convertibility of KAc solutions is equivalent in
order of magnitudes with that of NaCl aqueous reported in Ref.
[53–55], and better than that of NH
3
HCO
3
solution reported in Ref.
[15]. (2) RED cells working with aqueous KAc solution possesses the big-
gest P
dmax
than both of KAc–H
2
O–TFE and KAc–H
2
O–ethanol solutions.
(3) RED cells working with KAc–H
2
O–ethanol solution has a bigger
E
OCVmax
than the aqueous KAc solution when other testing conditions
are the same. (3) When keeping the weak solution of KAc–H
2
Oasa
concentration of 0.05 mol·kg
−1
, but concentrating its strong solution
from 0.25 mol·kg
−1
to 0.5 mol·kg
−1
, the measured E
OCVmax
and P
dmax
are found to be increased 17.9% and 66.2% respectively. The concentration
different between the strong solution and weak solution that flow sepa-
rately in the neighboring compartments of the RED cells is an important
factor on electric convertibility. Usually, the concentration different
between the feeding solutions can be enlarged by further concentrating
the strong solution or further diluting the weak solution. However, their
variable range are not unlimited, since that for one hand electric resistiv-
ity must be significant for the excessive weak solution, and for another
hand electric conductivity will be decreased for the excessive concen-
trated solution, thus leading to the reduction of E
OCVmax
and P
dmax
.
To further enlarge the E
OCVmax
and P
dmax
, the following actions may be
workable. For instances, (1) reducing the thicknesses of IEM and gasket
(even integrating them together by 3D print technology); (2) reducing
the effect of concentration polarization around the IEM; (3) improving
the ions permselectivity; (4) increasing solution temperature (but cannot
exceed the thermostability of IEM); (5) using practicable feeding solu-
tions with suitable concentrations; (6) adding more pairs of IEMs (such
as fifty pairs); (7) adjusting the matched electrode rinse solution with
thesuitablevelocity,andsoon.
4. Conclusions
This work reports the measured electric conductivity and electric con-
vertibility of potassium acetate in water, ethanol, 2,2,2–trifluoroethanol,
2–propanol, water–ethanol blends and water–TFE blends within the
wide ranges of solution concentration (up to about 60% in mass concen-
tration) and solution temperature (from 20°C to 60°C) conditions. The
experimental temperature and concentration conditions are selected
carefully according to the practical operating requirements of working so-
lutions in a closed type RED power generation system that is developed to
convert the low grade thermal energy to electric energy. The following
conclusions can be gained:
(1) KAc can be ionized in water, ethanol, and TFE, as well as their mix-
tures, and electric conductibility ranking of these solutions are in
the following order of KAc–H
2
ONKAc–90%H
2
O–10%TFE NKAc–
90%H
2
O–10% ethanol NKAc–TFE NKAc–ethanol NKAc–IPA under
the same temperature and concentration conditions. The maxi-
mum tested electric conductivity is (273.2 ± 5.451) mS·cm
−1
gained at the concentration of 40.11% and the temperature of 60°C.
(2) Electric conductivity is influenced by both temperature and con-
centration factors of solutions. With the increasing of concentra-
tion, electric conductivity of solution increases firstly and then
reaches the peak value, but later it decreases. Solution temperature
plays a positive influence role to the electric conductivity, and the
most powerful influence region of temperature on the electric
conductivity of solution is located at the place where the peak elec-
tric conductivity appears.
(3) Electric conductivity of binary solutions (KAc–H
2
O, KAc–TFE, KAc–
ethanol and KAc–IPA) and ternary solutions (KAc–H
2
O–ethanol
and KAc–H
2
O–TFE) are verified to be estimable by using a modi-
fied amplitude version of Gaussian peak function with a similar
format as the Arrhenius exponential equation. The total average
relative error between the measured results and estimated data
of fifteen solutions is about 2.06%.
Table 6
Measured maximum power density and open circuit voltage
Weak solution Strong solution Percentage of solvents/ wt.% N
IP
E
OCVmax
/
V
R/
Ω
P
dmax
/
W·m
−2
Source
KAc–H
2
O–TFE This work
0.05 mol·kg
−1
5.0 mol·kg
−1
0:100% 10 1.714 6.077 1.428
10%:90% 10 1.676 10.170 0.855
20%:80% 10 1.689 15.071 0.610
KAc–H
2
O–ethanol
0.05 mol·kg
−1
2.5 mol·kg
−1
0:100% 10 1.44 7.536 0.859
10%:90% 10 1.513 10.414 0.696
20%:80% 10 1.541 12.559 0.570
0.56 g·L
−1
111 g·L
−1
NaCl–H
2
O 8 0.64 –0.72 [53]
1 g·L
−1
30 g·L
−1
NaCl–H
2
O50N6–0.93 [54]
0.017 mol·L
−1
0.507 mol·L
−1
NaCl–H
2
O5––2.2 [55]
0.02 mol·L
−1
1.5 mol·L
−1
NH
3
HCO
3
–H
2
O 20 1.5 –0.33 [13]
2590 X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591

(4) Electric convertibility of binary solutions (KAc–H
2
O) and ternary
solutions (KAc–H
2
O–ethanol and KAc–H
2
O–TFE) are measured.
RED cells working with KAc–H
2
O solution possesses the biggest
output power density (1.428 W·m
−2
) than both of KAc–H
2
O–
TFE and KAc–H
2
O–ethanol solutions, while as for maximum
open circuit voltage, KAc–H
2
O working pair is medium in this
test, which is larger than KAc–H
2
O–TFE solutions but smaller
than KAc–H
2
O–ethanol solutions under the same experimental
condition. The concentration difference between the feeding
strong solution and weak solution is found to be an important
impact factor on electric convertibility.
(5) Due to its extremely weak ability of ionization, dissolution and
conduction, the KAc–IPA solution is suggested to be excluded
for RED power generation system. The solutions of KAc–TFE,
KAc–H
2
O–TFE, and KAc–H
2
O–ethanol are recommended due to
their relative higher solubility and electric conductivity, acceptable
electric convertibility, as well as their relative lower latent heat of
vaporization and suitable boiling point temperature.
References
[1] R.E. Pattle, Production of electricpower by mixing fresh andsalt water in the hydro-
electric pile, Nature 174 (1954) 660-660.
[2] P. Długołecki, K. Nymeijer, S. Metz, M. Wessling, Current status of ion exchang e
membranes for power generation from salinity gradients, J. Membr. Sci. 319 (1–2)
(2008) 214–222.
[3] E. Güler, W.V. Baak, M. Saakes, K. Nijmeijer, Monovalent-ion-selective membranes
for reverse electrodialysis, J. Membr. Sci. 55(4)(2014)254–270.
[4] H.I. Jeong,H.J. Kim, D.K. Kim, Numerical analysis of transport phenomena in reverse
electrodialysis for system design and optimization, Energy 68 (4) (2014) 229–237.
[5] A. Cipollina, G. Micale, Sustainable Energy from Salinity Gradients, Woodhead
Publishing, Duxford, 2016.
[6] M. Rahimi, A. D'Angelo, C.A. Gorski, O. Scialdone, B.E. Logan, Electrical power pro-
duction from low-gradewaste heat using a thermally regenerative ethylenediamine
battery, J. Power Sources 351 (2017) 45–50.
[7] Xu S.M., Wu X., Wu D.B., A novel generation method and device powered by low
grade heat energy. Chinese Pat. ZL 201510694726.4 (2017).
[8] X. Wu, S.M. Xu, D.B. Wu, H. Liu , S.Q. Chen, Methodology of assessing working
mediums availability for a novel heat -power conversion system with reverse
electrodialysis technology, J. CIESC 57 (S2) (2016) 326–332 (in Chinese).
[9] Y.Y. Xie, W.X.Zhang, S. Gu, Y.S. Yan, Z.F. Ma, Process engineering in electrochemical
energy devices innovation, Chin. J. Chem. Eng. 24 (1) (2016) 39–47.
[10] J. Veerman, M. Saakes, S.J. Metz, G.J. Harmsen, A.Daniilidis, D.A. Vermaas, R. Herber,
K. Nijmeijer, Experimentally obtainable energy from mixingriver water, seawater or
brines with reverse electrodialysis, Renew. Energy 64 (2) (2014) 123–131.
[11] N.Y. Yip, D.A. Vermaas, K. Nijmeijer, M. Elimelech, Thermodynamic, energy efficiency,
and power density analysis of reverse electrodialysis power generation with natural
salinity gradients, Environ. Sci. Technol. 48 (9) (2014) 4925–4936.
[12] M. Bevacqua, A. Tamburini, M. Papapetrou, A. Cipollina, G. Micale, A. Piacentino,
Reverse electrodialysis with NH
4
HCO
3
-water systems for heat-to-power conversion,
Energy 137 (2017) 1293–1307.
[13] X. Luo, X. Cao, Y. Mo, K. Xiao, X. Zh ang, P. Liang, Power gene ration by coupling
reverse electrodialysis and ammonium bicarbonate: implication for recovery of
waste heat, Electrochem. Commun. 19(1)(2012)25–28.
[14] X. Wu, S.M. Xu, D.B. Wu,Ternary working fluids for a heat-power conversionsystem
with reverse electrodialysis technology. Chinese Pat.ZL 201610126962.0(2018).
[15] PubChem Database, Compound summary for CID 517044: potassium acetate.U.S.
National Library of Medicine. [2017-9-15], https://pubchem.ncbi.nlm.nih.gov/com-
pound/Potassium_acetate#section=Top.
[16] D.R. Lide, Handbook of Chemistry and Physics, 84th ed.CRC Press, Florida, 2003.
[17] J.G. Speiht, LANGE's Handbook of Chemistry, 15th ed. Mcgraw-Hill, New York, 2005.
[18] A. Apelblat, E. Manzurola, The vapour pressures over saturated aqueous solutions of
sodium and potassium acetates, chlorates, and perchlorates, J. Chem. Thermodyn. 39
(8) (2007) 1176–1181.
[19] D. Meranda, W.F. Furter, Vapor-liquid equilibrium data for system: ethanol-water
saturated with potassium acetate, Can. J. Chem. Eng. 44 (5) (1966) 298–300.
[20] V.K.L. Mer, J.P. Chittum, The conductance of salts (potassium acetate) and the
dissociation constant of acetic acid in deuterium oxide, J. Am. Chem. Soc. 58 (9)
(2002) 1642–1644.
[21] T.S. Murthy, K. Lakshminarayana, Conductance studies of alkali metal acetates in
acetic acid-acetonitrile mixtures, Phys. Chem. Liq. 20 (2–3) (1989) 167–176.
[22] R.S. Sah, B. Sinha, M.N. Roy, Ion association and solvation behavior of some alkali
metal acetates in aqueous 2-butanol solutions at T= 298.15, 303.15 and
308.15 K, Fluid Phase Equilib. 307 (2) (2011) 216–221.
[23] National Institute of Standards and Technology, NIST chemistry webbook, SRD 69,
[2018-05-18] http://webbook.nist.gov/chemistry/.
[24] A.D. John, Lange's Handbook of Chemistry, 15th ed. McGraw-Hill, New York, 1998.
[25] L. Wu, Q. Zhang, Error Analysis and Data Processing, Tsinghua University Press,
Beijing, 2010 (in Chinese).
[26] National Institute of Metrology, Electrolytic conductivity meters, Chinese Metrology
Verification Regulation JJG 2007, pp. 376–2007 , (in Chinese).
[27] Y.G. Li,J.F. Lu, The Theory of ElectrolyteSolutions, Tsinghua University Press, Beijing,
2005 (in Chinese).
[28] C.H. Hammann, A. Ha mnett, W. Vielstich, Electroche mistry, 2rd ed. Wiley-VCH
Press, Weinheim, 2007.
[29] R.A. Robinson, H.S. Harned, Some aspects of the thermodynamics of strong electro-
lytes from electromotive force and vapor pressure measurements, Chem. Rev. 28 (3)
(1941) 419–476.
[30] Wikipedia, Bjerrum length, [2017-02-26] https://en.wikipedia.org/wiki/Bjerrum_
length.
[31] Y. Tanaka, Y.F. Xiao, S. Matsuo, Relative permittivity of fluoroalcohols at tempera-
tures from 293 to 323 K and pressures up to 50 MPa, Fluid Phase Equilib. 170 (1)
(2000) 139–149.
[32] G. Gente, C.L. Mesa, Water–trifluo roethanol mixtures: some physicochemical
properties, J. Solut. Chem. 29 (11) (2000) 1159–1172.
[33] G. Gente, C.L. Mesa, R. Muzzalupo, G.A. Ranieri, Micelle formation and phase equilibria
in a water-trifluoroethanol-fluorocarbon surfactant system, Langmuir 16 (21) (2000)
7914–7919.
[34] R.M. Fuoss, Conductance-concentration function for the paired ion model, J. Phys.
Chem. 82 (22) (1978) 2427–2440.
[35] R.M. Diamond, The activity coefficients of strong electrolytes.The halide salts, J. Am.
Chem. Soc. 80 (18) (1958) 4808–4812.
[36] B.S. Krumgalz, Ion-solvent interactions and ionic association in ethanol solutions,
J. Solut. Chem. 11 (4) (1982) 283–293.
[37] B.S.Krumgalz, Interaction of tetraalkylammonium and some other organic ionswith
solvent molecules, J. Gen. Chem. USSR 4 (1974) 1585–1588.
[38] H. Chen, L.S. Wang, B. Jiang, M.Y. Li, Measurements of conductivity for low concen-
tration strongelectrolytes in organic solvents (I) LiBr, LiCl,and LiNO
3
in alcohols,
Chin. J. Chem. Eng. 20 (5) (2012) 1024–1033.
[39] G.H. Gao, H.B. Shi, Y.X. Yu, Mutual diffusion coefficients of concentrated 1:1 electrolyte
from the modified mean spherical approximation, Fluid Phase Equilib. 256 (1) (2007)
105–111.
[40] O. Bernard, W. Kunz,P. Turq, L. Blum, Self-diffusion in electrolytesolutions using the
mean spherical approximation, J. Phys. Chem. 96 (1) (1992) 398–403.
[41] O. Bernard, W. Kunz, P. Turq, L. Blum, Conductance in electrolyte solutions usingthe
mean spherical approximation, J. Phys. Chem. 96 (9) (1992) 3833–3840.
[42] J.F. Dufrêche, O. Bernard, P. Turq, Transport equations for concentrated electrolyte solu-
tions: reference frame,mutual diffusion, J. Chem. Phys. 116 (5) (2002) 2085–2097.
[43] H.B. Shi, Y.X. Yu, G .H. Gao, Brownian dynamics simulatio n of self-diffus ion
coefficients of electrolyte solutions, Chem. J. Chin. Univ. 25 (6) (2004) 2317–2321
(In Chinese).
[44] H.B. Shi, Y.X. Yu, G.H. Gao, Study on the transport properties of aqueous electrolyte
solution by brownian dynamics simulation, Acta Chim. Sin. 63 (5) (2005) 358–362
(In Chinese).
[45] Y.C. Wu,W.F. Koch, E.C. Zhong,H.L. Friedman, The cross-square rule for transport in
electrolyte mixtures, J. Phys. Chem. 92 (6) (1988) 1692–1695.
[46] T.F. Young, M.B. Smith, Thermodynamic properties of mixtures of electrolytes in
aqueous solutions, J. Phys. Chem. 58 (9) (1954) 716–724.
[47] D.G. Miller, Binary mixing approximations and relations between specific conductance,
molar conductance, equivalent conductance, and ionar conductance for mixtures,
J. Phys. Chem. 100 (4) (1996) 1220–1226.
[48] Y.F. Chen, Y.F. Hu, J.G. Qi, Y. Sun, Z.Y. Li, Densities, conductivities, and viscosities
of aqueous solutions of N-hexyl, methylpyrrol idinium bromide and N-butyl,
methylpyrrolidinium bromide at different temperatures, Chin. J. Chem. Eng. 23 (1)
(2015) 213–218.
[49] Y.F. Hu, X.M. Zhang, J.G. Li, Q.Q. Liang, Semi-ideal solution theory. 2. Extension to
conductivity of mi xed electrolyte s olutions, J. Phys. Chem. B 112 (48) (2 008)
15376–15381.
[50] Q.Q. Liang, Y.F. Hu, W.J. Yue, Electrical conductivities for four ternary electrolyte
aqueous solutions with one or two ionic liquid components at ambient tempera-
tures and pressure, Chin. J. Chem. Eng. 23 (6) (2015) 873–879.
[51] D.A. Vermaas, M. Saakes, K. Nijmeijer, Early detection of preferential channeling in
reverse electrodialysis, Electrochim. Acta 117 (2014) 9–17.
[52] S.M. Xu, D.B. Wu, X. Wu, J.Y. Hu, H. Liu, H.J. Zhang, S.Q. Chen, J. Chen, Experimental
study on power generated by solution concen tration difference changing with
lithium chloride aqueous, J. Dalian Univ. Tech. 57 (4) (2017) 337–344 (in Chinese).
[53] M. Turek, B. Bandura, P. Dydo, Power production from coal-mine brine utilizing
reversed electrodialysis, Desalination 221 (1–3) (2008) 462–466.
[54] J. Veerman, M. Saakes, S.J. Metz, G.J. Harmsen, Reverse electrodialysis: performance
of a stack with 50 cells on the mixing of sea and river water, J. Membr. Sci. 327 (1)
(2009) 136–144.
[55] D.A. Vermaas, M. Saakes, K. Nijmeijer, Doubled power density from salinity gradients at
reduced intermembrane distance, Environ. Sci. Technol. 45 (16) (2011) 7089–7095.
2591X. Wu et al. / Chinese Journal of Chemical Engineering 26 (2018) 2581–2591