S
1
Supporting Information
Tuning
Interfacial
Electrical Field
of Bipolar Membranes
with
Temperature and Electrolyte Concentration
for Enhanced Water
Dissociation
H
uanlei Z
hang
1,2,
†
, Dongbo Cheng
1,2,
†
, Chengxiang Xiang
3
, Meng Lin
1,2,*
1
Department of
Mechanical and Energy Engineering, Southern University of Science
and Technology, Shenzhen 518055, China
2
SUSTech Energy Institute for Carbon Neutrality, Southern University of Science and
Technology, Shenzhen 518055, China
3
Liquid Sunlight Alliance, Dep
artment of Applied Physics and Material Science,
California Institute of Technology, Pasadena, California 91125, United States
† contributed equally to this work
Corresponding Author:
linm@sustech.edu.cn
This file includes:
Total
–
17
pages,
13
figures, and
1
table
Figure S1
-
S
13
, Table S1
S
2
Figure
S1
.
IV curves
comparison with various literature data.
S
3
The dielectric constant was estimated using
0
sol
W
sol
c
c
−
=
(eq. 21), in which
β
c
is solution dependent.
4
In the reference
5
,
β
c
is 12.5% lower for K
+
compared to Na
+
based electrolyte. Note that the electr
olyte discussed in reference
4
was KCl and NaCl
and we assumed the dielectric constant difference was induced by the cations. We
performed additional calculations with different diffusion coefficients and dielectric
constants (see Figure S2). We observe no
difference when using two different
electrolytes.
Figure S2.
Comparison between Na
2
SO
4
and K
2
SO
4
with different diffusion
coefficients and dielectric constants.
S
4
Experimental
Methods for 4
-
Probe Measurement
BPM voltage measurements were performed in a H type cell. The
Keithley 2601B
(V/I range 100 fA
-
10A, 100 nV
-
40 V, accuracy 0.015%) was used as the power supply.
T
he BPM (Fumatech) voltage
drop
was determined by measuring the voltage difference
between t
wo Ag/AgCl reference electrodes (4 M KCl). T
he temperature of electrolyte
was measured through two K
-
type thermocouples (
range 0 ~ 1000
o
C
, ±2 K
).
The
Keithley DAQ6510 (V/I range 100 pA
-
3A, 100 nV
-
1000 V, accuracy < 0.015%) was
used for the
data acquis
ition of voltage and temperatures.
Na
2
SO
4
(Aladdin, AR, 99%)
used as
the catholyte and the anolyte was flowed to the
H type cell at a flow rate of 25.5
mL min
-
1
with the control of peristaltic pump (
Longer Pump, BT100
-
2J, 0.1rpm
–
100rpm, accuracy = 0.1rpm
), and the temperature of the electrolyte with the control of
the right water bath. The
BPM voltages were measured in the multistep
chronopotentiometry mode and the applied current density was swept from low current
density to high current density. The vol
tage at each applied current density was recorded
once the voltage stabilized
(typical after
5
mins)
.
The hot water with the control of left
water bath was flowed to the insulation layer of
H type cell at a rate of 1090 mL min
-
1
using peristaltic pump (
Kam
oer, KKDD
-
24B17 A, flow rate >840 ml/min
). The
temperature of electrolyte was heated by hot water and stabilized at 25
o
C
, 50
o
C
, and
80
o
C
(see Figure S
3
).
S
5
Figure S
3
.
Schematic illustration of the
Experimental setup consisting of water bath,
Platinum (Pt) electrodes, catholyte compartment, BPM, and anolyte compartment.
The
anolyte and catholyte were 1 M
Na
2
SO
4
and BPM with an active area of 1.0 cm
2
.
S
6
Figure S
4
.
The temperature of electrolyte wi
th concentration of
(a)
0.6 M and
(b)
1.0
M stabilized at 25
o
C, 50
o
C, and 80
o
C.
S
7
Higher IEC value leads to better performing BPM due to enhanced junction electrical
field. However, the IEC value do not change the trend of the BPM performance and
hence not changing the conclusions of this study. The detailed effect of IEC can be
found in our previous work.
6
Figure
S5
.
IV curves for different temperature with IEC is 1 mmol g
-
1
(solid lines)
and
2 mmol g
-
1
(dash lines).
S
8
Table S
1
. Water Transport
Parameters in BPM
Property
Value
Unit
D
W
0
<
λ
<
4
(
)
(
)
2 79 6
.
5
23
10
8
K
1
10
2
K
2436
exp
44
.
1
exp
10
41
.
5
0.12
exp
10
32
.
7
+
−
+
−
−
−
T
T
m
2
s
-
1
4
≤
λ
≤
22
(
)
(
)
(
)
2 79 6
.
5
23
7
8
K
1
10
2
K
2436
exp
.04
0
1
exp
10
45
.
1
66
.
4
-
exp
10
58
.
1
+
−
+
−
−
−
T
T
ξ
W
25
o
C
-
1.4
1
50
o
C
-
2.4
80
o
C
-
4.1
Figure S
6
.
(a)
IV
curves for different
B
ruggm
a
n coefficients. (b)
IV
curves for different
diffusion coefficients of water. (c) Potential drop of junction (
V
junction
) as a function of
current density at different temperatures and different
c
sol
.
The effective diffusion coefficient of each ion species (
D
i
eff
) depends on the phase.
In the aqueous electrolyte phase, these diffusion coefficients are equal to their values in
pure water. The Bruggman relation was used within the bipolar membrane
(
). Here
f
mem
is the porosity of the membrane (0.15)
，
q
is Brugg
m
a
n
coefficient in the model. Regarding the effect of electrolyte concentration on the
q
0
2
4
6
8
10
0
300
600
900
1200
1500
0
2
4
6
8
10
0
300
600
900
1200
1500
0
600
0.0
0.5
1.0
1.5
0
600
0
600
Current density
(
mA cm
-2
)
Potential (V)
1.5
1.8
2
q
-
Bruggmen coefficient
O
E
A
B
C
D
Current density
(
mA cm
-2
)
Potential (V)
0.5x
1x
2x
D
W
a
)
b
)
c
)
c
sol
-
concentration of solution
0.6 M
1 M
V
junction
(V)
25 °C
50 °C
80 °C
Current density
(
mA cm
-2
)
q
i
i
f
D
D
mem
eff
=
S
9
during BPM operation, the current research data is lacking, and the internal mechanism
is not clear. This study did not explore this in depth.
q
is corrected by passin
g through
our experimental data. In this study, it is set to 1.8 when the electrolyte concentration is
1 M, and it is set to 1.6 when the electrolyte concentration is 0.6 M.
Figure S
6
a shows the
IV
curves for different
q
. It can be shown that the change of
q
affects the slope of the B
-
C region (ohmic region). Figure 2b shows the B
-
C region
(ohmic region) of the
IV
curve of simulation and experiment, which can be completely
corresponded. At the same time, the change of
q
will also affect the D
-
E region (Wate
r
-
limiting region). Compared with
q
, the change of
D
W
does not affect the
IV
curve of B
-
C region (ohmic region). But it will affect D
-
E region (Water
-
limiting region).
In order to fit the experimental data, the
q
was adjusted from 1.8 to 1.6
when the
elect
rolyte concentration changed
.
Figure 5a shows that the experimental data of
c
sol
=
0.6
M and
c
sol
=
1
M
at different temperatures are basically consistent with the
simulated data, which proves that the model developed in this study has universal
applicability of temperature and concentration.
In addition, the decrease of the solution
concentration has an ef
fect on the ohmic limit stage, which is the decrease of the
V
junction
voltage (see
Figure S
6
c
).
S
10
Figure S
7
.
(a)
Fixed charge
concentration as a function of location over the entire
calculation
domain
.
(b) Zoom in
f
ixed charge
concentration profiles and
catalyst
concentration profiles (c) for the junction layer with its adjacent AEL and CEL regions
(10 nm for each) at different
L
char
.
S
11
Figure S
8
.
(a)
IV
curves for different relative tolerance of solver. (b)
IV
curves for
different mesh elements.
The IV curves are almost the same for different relative tolerances and different mesh
elements
. Decreasing the relative tolerance or increasing the
elements
of
meshes has
little effect on the accuracy of the results, justifying our choice of relative tolerance and
number of meshes.
0
2
4
6
8
10
0
300
600
900
1200
1500
0
2
4
6
8
10
0
300
600
900
1200
1500
Current density
(
mA cm
-2
)
Potential (V)
Relative tolerance
0.001
0.0005
0.0001
Current density
(
mA cm
-2
)
Potential (V)
Mesh
5300
26528
53029
a
)
b
)
S
12
Figure S
9
. (a) The slope of the
IV
curve at different temperatures. (b) Ions diffusion
coefficient and WD equilibrium constant as
a function of temperature. (c) Average
water concentration in the bipolar membrane as a function of current density at different
temperatures. (d) Conductivity of a 1 M Na
2
SO
4
solution as a function of water
concentration.
0
2
4
6
8
10
0
300
600
900
1200
1500
290
300
310
320
330
340
350
4
8
12
16
0
300
600
900
1200
1500
4000
5000
6000
7000
8000
9000
0
2000
4000
6000
8000
0.0
0.5
1.0
1.5
2.0
2.5
Slope
(
mA cm
-2
V
-1
)
Voltage (V)
25 °C
50 °C
80 °C
Diffusion COEF
(
10
-5
cm
2
s
-1
)
T (K)
H
3
O
+
OH
-
10
-18
10
-17
10
-16
K
1
K
1
(-)
H
2
O concentration
(
mol m
-3
)
Current density
(
mA cm
-2
)
25 °C
50 °C
80 °C
C
onductive
(
S m
-1
)
H
2
O
concentration
(
mol m
-3
)
25 °C
50 °C
80 °C
a
)
b
)
c
)
d
)
S
13
0
10000
20000
30000
40000
2.2
2.4
2.6
2.8
3.0
3.2
Potential (V)
Time (s)
1 M Na
2
SO
4
25 °C 620 m
A/cm
2
Fig
ure S
10
.
Long time (>10 hours) test of BPM performance at 25
o
C
, current density
620
mA cm
-
2
, 1
M Na
2
SO
4
.
S
14
Figure S
11
.
(a) WDH
+
concentration profiles and (b) OH
-
concentration profiles at the
junction layer for various junction layer thicknesses and t
emperatu
re
s at 600 mA cm
-
2
.
Solid lines are for 25
o
C and dashed lines for 80
o
C.
0
20
40
60
0
300
600
900
1200
1500
1800
2100
0
20
40
60
0
300
600
900
1200
1500
1800
2100
c
WDH
+
(
mol m
-3
)
Location (nm)
600 mA cm
-2
25 °C
80 °C
c
OH
-
(
mol m
-3
)
Location (nm)
600 mA cm
-2
25 °C
80 °C
L
JL
=5 nm
L
JL
=10 nm
L
JL
=50 nm
a
)
b
)
S
15
Figure
S12
.
(a) The enhancement factor
f
electrical,F
for reaction 1 due to electric field
increase as a function of
current density
at various
L
char
.
(b) Zoom in sub
figure
for the
current density in the range of 500
-
700 mA cm
-
2
.
S
16
Figure S
13
.
(a) The simulation contribution of catalyzed pathway and uncatalyzed
pathway of WD as a function of voltage across the BPM at different junction layer
thickness
, (b) different abruptness. (c) different solution concent
r
ation.