Professional Documents
Culture Documents
1 s2.0 S0376738822002964 Main
1 s2.0 S0376738822002964 Main
A R T I C L E I N F O A B S T R A C T
Keywords: Salt permeability and ionic conductivity are critical properties for membranes in water purification and energy
Ion exchange membrane applications. Both properties depend on individual ion sorption and diffusion coefficients, which are significantly
Salt permeability influenced by polymer chemical and physical parameters such as fixed charge concentration and membrane
Ionic conductivity
water content. However, systematic studies connecting polymer structure to ion transport properties are still
Salt diffusion
lacking. In this study, a series of uncharged and charged membranes were synthesized using poly(ethylene
Ion diffusion
glycol) diacrylate (PEGDA) as a cross-linker and 2-acrylamido-2-methyl-1-propanesulfonic acid (AMPS) as a
charged monomer. Membrane fixed charge concentration and water uptake were systematically varied by
adjusting AMPS content in the pre-polymerization mixture. Salt sorption and permeability coefficients and ionic
conductivity of these membranes were measured as a function of NaCl solution concentration (0.01–1 M).
Combining the solution-diffusion model and Nernst-Planck equation, individual ion diffusion coefficients were
calculated. Experimental Na+ diffusion coefficients for all materials were well described by the Mackie and
Meares tortuosity model, highlighting the strong influence of water content on ion diffusivity in both uncharged
and charged polymers. Model predictions for Cl− diffusion coefficients agree reasonably well with experimental
values, with some deviations occurring in more highly charged membranes. This discrepancy might result from
interactions not captured by the Mackie and Meares model (e.g., fixed charge-ion interactions).
1. Introduction diffusion model [17,18]. Ions from the external solution first partition
into the membrane at the upstream side, diffuse down an electro
Ion exchange membranes (IEMs) are actively used and explored for chemical potential gradient, and desorb into the contiguous solution at
applications such as reverse osmosis, forward osmosis, electrodialysis the downstream side [12,18,19]. Ion sorption in highly charged IEMs is
(ED), reverse electrodialysis (RED), and fuel cells due to their ion se strongly influenced by the fixed charge groups via electrostatic in
lective properties [1–9]. To enhance separation efficiency and reduce teractions between the fixed charges and ions [20–22]. Typically,
energy costs, these technologies often have specific requirements for counter-ions (i.e., ions bearing a charge opposite to that of the fixed
membrane properties (e.g., ionic conductivity, permselectivity) [10]. charge groups) are attracted to the membrane, sorbing in large quanti
Meeting these specifications requires rational control of ion transport ties and co-ions (i.e., ions bearing the same charge as that of the fixed
properties via tailoring polymer chemical and physical parameters [10, charge groups) are largely excluded from the membrane (i.e., Donnan
11]. IEMs contain charged functional groups covalently bound to the exclusion) [12]. This exclusion effect is stronger when the fixed charge
polymer backbone (i.e., fixed charge groups). These ionogenic groups concentration in the membrane is higher than the electrolyte concen
can ionize and retain substantial amounts of water in the polymer upon tration in solution [12]. At equilibrium, sorbed counter-ions must
exposure to aqueous electrolyte solutions [12]. The concentration of electrically balance both the fixed charge groups on the polymer back
fixed charge groups and water content of the membrane significantly bone and any co-ions sorbed in the membrane, so there are more
influence ion transport [13–16]. counter-ions in a charged membrane than co-ions, if the counter-ions
Ion transport in nonporous membranes is described by the solution- and co-ions have the same valence.
* Corresponding author.
E-mail address: freeman@che.utexas.edu (B.D. Freeman).
https://doi.org/10.1016/j.memsci.2022.120549
Received 12 January 2022; Received in revised form 1 April 2022; Accepted 5 April 2022
Available online 8 April 2022
0376-7388/© 2022 Elsevier B.V. All rights reserved.
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
Ion diffusion is strongly influenced by the water content in a mem where ξ is the Manning parameter and can be determined based on
brane [13,23]. Water can plasticize polymer chains, thereby facilitating polymer structure [25,33]. Condensation does not occur in the polymers
ion diffusion [24]. In addition, fixed charges may affect ion diffusion via considered in this study, and in this case Cm,w can be solved numerically
electrostatic interactions with sorbed ions [25,26]. However, the
−
by combining Eqns. (1) and (4) [33]:
methods used to determine ion transport properties in IEMs vary greatly ⎡ ⎤
in the literature, and inconsistent results across these studies make
fundamental relations between polymer structure and ion transport ( ) ⎢
C−m,w C−m,w + CAm,w exp⎢
ξ ⎥ ( s )2 ( s )2
⎥= γ Cs (5)
⎣− 1 + 2 CC−m,w ⎦
properties difficult to discern [27–31]. Moreover, studies unifying ion
m,w ±
2
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
while diffusion coefficients predicted using the Mackie Meares model do coefficient, Dm
m*
s (which ≈< Ds > in this study as calculated from < Ps >
not. Frame of reference effects can be accounted for by including con and Ks ). For example, once Dm
− is calculated from Eqn. (13), Eqn. (11) or
vection in Fick’s law [41]. Non-ideal thermodynamic effects can be (12) can be used to calculate Dm+.
accounted for using ion activity coefficient gradients in the membrane,
which can be obtained from experimental ion sorption results [41,42]. 3. Experimental methods
After accounting for these effects, the local salt diffusion coefficient, Dm
s ,
can be obtained [41]. In this study, frame of reference and non-ideal 3.1. Materials
thermodynamic effects nearly offset each other for highly charged
m*
samples considered, such that Dm s ≈< Ds >, (cf. Supporting Informa A detailed description of the polymer synthesis procedure is reported
tion), similar to findings from a previous report on salt diffusion in ion elsewhere [42]. Fig. 1 presents the structures of XLPEGDA homopoly
exchange membranes [41]. Dm s values for the uncharged sample are mers and XL(AMPS-PEGDA) copolymers, which were prepared as
m*
slightly higher (<5%) than < Ds > values, since both the frame of free-standing films by UV irradiation of a pre-polymerization mixture
reference and non-ideal thermodynamic effects increased salt diffusion [44–47]. This mixture contained the cross-linker, poly(ethylene glycol)
coefficients (cf. Supporting Information). However, the difference be diacrylate (PEGDA) [n = 13, n = 10, and n = 4, where n represents the
m* average number of ethylene oxide (EO) groups per cross-linker mole
tween Dm s and Ds is within the experimental uncertainty. Thus, for the cule], the ionic monomer, 2-acrylamido-2-methyl-1-propanesulfonic
materials considered in this study, using the apparent salt diffusion co
acid (AMPS), and the photoinitiator, 2,2-dimethoxy-2-phenylacetophe
efficients is a reasonable approximation for the local salt diffusion co
none (DMPA). Water and, in some cases, water/methanol (MeOH)
efficients for further data analysis and comparison to model.
mixtures served as the solvent for the free radical polymerization.
Table 1 presents compositions of the pre-polymerization mixtures. After
2.4. Nernst-Planck equation UV-polymerization, the solid hydrogel films were soaked in DI water for
at least one day to extract residual components not bound to the
Ion transport in membranes can be described by the Nernst-Planck network. Water was changed three times daily. The as-synthesized acid
equation, which accounts for concentration and electric potential gra form (H+) XL(AMPS-PEGDA) films were converted to the Na+ form
dients across the membrane [43]: using procedures described elsewhere [42]. All reagents were used as
[ m,p ] received from Sigma-Aldrich (Milwaukee, WI). De-ionized (DI) water
dCi zi FCim,p dψ (18.2 MΩ cm, 1.2 ppb total organic carbon) was generated by a Millipore
Jim = − Dmi + (10)
dx RT dx Milli-Q Advantage A10 water purification system (Millipore Corpora
tion, Bedford, MA).
where Jim and Dm i are the flux and diffusion coefficient of ion i in the
membrane. ψ is the electric potential, F is Faraday’s constant, R is the 3.2. Salt permeability
gas constant, and T is absolute temperature.
Ion transport through membranes varies with the type of driving Prior to salt permeability tests, films of XLPEGDA and Na+ form XL
force applied [32]. For example, when a membrane separates two so (AMPS-PEGDA) were equilibrated in DI water for at least one day.
lutions with different salt concentrations and no electric potential Films were then clamped between a pair of jacketed glass diffusion cells
gradient is imposed across the membrane, a chemical potential (i.e., (PermeGear Side-bi-Side, Hellertown, PA) [47,48]. The donor (up
concentration) gradient across the membrane drives cations and anions stream) and receiver (downstream) chambers were filled with equal
to diffuse in the same direction to maintain electroneutrality [32]. This volumes of NaCl solution at a desired concentration (i.e., 0.01–1 M) and
coupled diffusion of cations and anions across the membrane can be DI water, respectively. The solutions in both chambers were vigorously
described by the salt diffusion coefficient of the membrane [41]: stirred, and the temperature was maintained at 25 ± 0.1 ◦ C. The con
(
Dm Dm Cm,p + C+m,p
) ductivity change in the downstream solution was monitored as a func
Dms = + m− m − m m
(11) tion of time using a conductivity meter (WTW inoLab Cond 730,
D+ C+ + D− C−
Woburn, MA). The downstream salt (i.e., NaCl) concentration was
where Dm computed from conductivity data using a calibration curve. As described
s is the local salt diffusion coefficient of the membrane
elsewhere, at pseudo-steady state, the integral salt permeability coeffi
mentioned above, Cm,p+ and C−
m,p
are the cation and anion concentrations
cient, < Ps >, is determined as follows [47,48]:
in the membrane (mol of ions per liter of swollen membrane), and Dm +
[ ] ( )
and Dm− are the cation and anion diffusion coefficients in the membrane 2Cs [t] 2A < Ps >
ln 1 − s sl =− t (14)
[41]. Cs0 [0] VL
When a membrane separates two electrolyte solutions at the same
concentration, and an electric potential gradient is applied across the
membrane, cations and anions migrate in opposite directions towards
their respective electrodes, inducing an electric current [12,32,43]. In
this case, ion transport is described by the ionic conductivity, κ, which
for a monovalent electrolyte is given by Ref. [43]:
F 2 ( m m,p )
κ= D C + Dm− C−m,p (12)
RT + +
Combining Eqns. (11) and (12) yields:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
κRT
( κRT )2 Cm,p
m,p m
− C + Ds
F2
± F2
− 4 κRT F 2 C−m,p +Cm,p
(13)
+
m
D− =
2C−m
can be calculated using data for ionic conductivity, κ, the concentrations Fig. 1. Chemical structures of: (a) XL(AMPS-PEGDA) copolymers and (b)
of sorbed ions in the membrane, Cm,p
− and Cm,p
+ , and the local salt diffusion XLPEGDA homopolymers.
3
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
where Css0 [0] is the initial upstream salt concentration, and Cssl [t] is the
Methanol [g]
downstream salt concentration at time t. A is the active sample area
(1.77 cm2), V is the liquid volume in both cells (35 mL), and L is the
2.20
hydrated membrane thickness measured immediately after the experi
–
–
–
–
–
–
–
–
ment was completed.
For low upstream salt concentrations, dissolution of CO2 from air in
the solutions of the diffusion cells can artificially elevate the down
DI water [g]
Theoretical IEC values of the acid form (H+) XL(AMPS-PEGDA) films were determined as milliequivalents (i.e., mmol) of AMPS/(g of AMPS + g of PEGDA)42 or meq/g (dry polymer).
when salt permeation tests are performed with highly charged mem
5.50
5.77
5.00
5.50
5.00
5.50
5.00
3.30
5.00
branes at low upstream salt concentrations (i.e., 0.01 M) [49]. To
mitigate this effect for the highly charged membranes (e.g., samples
with IEC values of 0.44–1.93 meq/g) used in this study, ultra-high purity
N2 (Airgas, Austin, TX) was bubbled through the upstream (i.e., 0.01 M
Initiator (DMPA) [g]
where [49].
1.5
3.0
5.0
6.0
1.5
6.0
trapped between the solution and cell wall [41]. Two capillary tubes
The samples were labeled as x-y, where x the number of EO units in PEGDA, n, and y is the theoretical IEC value [33].
–
were then attached to the top of each cell, each containing the same
solution as in their respective cells. At pseudo-steady state, the decrease
of the solution level in the downstream tube (or the solution level in
Cross-linker (PEGDA) [g]
20
30
40
40
9
0
9
0.44
0.97
1.46
1.93
0.44
1.93
0
10–0.44
10–0.97
10–1.46
10–1.93
4–0.44
4–1.93
Table 1
about 200 mL of feed solution (i.e., NaCl solution having the same
10–0
4
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
which was then pumped through the cell chambers using a peristaltic
pump (150 mL/min, Masterflex 7528–10, 77292–50). In the measure
ments conducted at high salt concentrations (0.1–1 M), a DC current of
5 mA and an AC current of 1 mA were used [50]. At lower salt con
centrations (0.01–0.03 M), an AC current of 0.3 mA was used with no DC
current [50]. The frequency sweep was between 40 and 0.5 kHz, and the
step interval for frequency change was 10 Hz. The total electrolyte
resistance of the membrane loaded cell (i.e., membrane + solution
resistance), Rm+s , and a blank cell containing no membrane (i.e., solu
tion resistance), Rs , were recorded. The membrane resistance, Rm , was
calculated as follows [50]:
Rm = Rm+s − Rs (16)
In uncharged and weakly charged membranes, Rm+s , was 2–10 times
higher than Rs over the entire range of salt concentrations considered.
This large difference ensures that the Rm value can be accurately
determined using the difference method. The ionic conductivity of the
membrane was calculated as follows [50]:
L
κ= (17)
Rm A
Table 2
Polymer water uptake, water volume fraction, and fixed charge concentration [42].
Sample (n-IEC) wu a [g (water)/g (dry polymer)] ϕw b [L (water)/L (swollen membrane)] Cm,w
A
c
[mol/L (water sorbed)]
5
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
6
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
concentrations in the membranes via Eqn. (7). Fig. 3 (b) presents NaCl non-electrostatic interactions between polymer segments, water, and
sorption coefficients, Ks , as a function of external NaCl concentration. ions in such materials [22,33,35,42].
The dependence of Ks on Css in uncharged and highly charged mem
branes is qualitatively different. Ks values in the most highly charged 4.2.2. Na + sorption
membrane (IEC = 1.93 meq/g) exhibit significant variation with Css , Fig. 5 presents sorbed Na+ concentrations (i.e., the counter-ion
while those of the uncharged membrane (IEC = 0) change relatively concentration in a CEM) as a function of external NaCl concentration,
little with Css [20]. Css . For the uncharged membranes, the Css dependence of Cm,p + resembles
Fig. 4 presents the dependence of experimental NaCl sorption co that of Cm,p
− in Fig. 3 (a), because the membrane contains equal numbers
efficients on IEC at the lowest and highest external NaCl concentrations of Na+ and Cl− at any external NaCl concentration to ensure electro
considered, 0.01 M and 1 M. For comparison, theoretical NaCl sorption neutrality in the membrane [20]. In contrast, the Na+ concentration in
coefficients calculated using the Donnan/Manning model [i.e., Eqn. (5)] the membrane of IEC = 1.93 meq/g changes little with external NaCl
are also shown in Fig. 4. At the lowest salt concentration considered (i.e., concentration. This weak dependence of Cm,p s
+ on Cs is typical for highly
Css = 0.01 M), NaCl sorption coefficients initially decrease by about one charged cation exchange membranes [21], where the majority of the
order of magnitude as IEC increases from 0 to 0.44 meq/g, presumably sorbed cations are electrically balancing fixed charges on the polymer
due to enhanced Donnan exclusion induced by increased fixed charge backbone (i.e., Cm,p m,p
+ = CA + C−
m,p
≈ Cm,p
A ), such that the concentration of
concentration in the membrane (cf. Cm,w A in Table 2). Then, Ks values are sodium ions is relatively independent of Css [33,42]. For samples in be
relatively constant at high IEC values (i.e., IEC >0.44 meq/g). These tween these extremes (i.e., IEC of 0.01–1.46 meq/g), Cm,p exhibits
+
trends are qualitatively consistent with the model predictions [53]. That behavior intermediate to these two limiting cases (i.e., the uncharged
is, the effects of Donnan exclusion are very sensitive to the change in and most highly charged samples considered).
fixed charge concentration when only low levels of fixed charges are
introduced to the polymer network [42]. Further addition of fixed
charges to an already highly charged polymer promotes co-ion exclusion 4.3. Ion diffusion coefficients
to a lesser extent [42]. At the highest salt concentration considered (i.e.,
Css = 1 M), both experimental and theoretical Ks values depend relatively 4.3.1. Na + diffusion coefficients
weakly on IEC and asymptotically approach that of the uncharged The Na+ diffusion coefficients were computed using the measured
membrane (IEC = 0). This phenomenon indicates the Donnan exclusion ion concentrations, salt permeability coefficient, and membrane ionic
effect in charged membranes is greatly reduced at high Css , such that the conductivity values using the Nernst-Planck framework [i.e., Eqn. (11)
and (12)]. Fig. 6 (a) presents Na+ diffusion coefficients as a function of
charged membranes sorb co-ions in a similar fashion to the uncharged
upstream NaCl concentration. For each membrane, Na+ diffusion co
membranes [20]. The dependence of experimental NaCl sorption co
efficients are fairly constant at low NaCl concentrations (Css < 0.1 M).
efficients on fixed charge concentration (i.e., mol of fixed charge groups
per liter of water sorbed in the membrane) can be found in the Sup Then, Dm + values decrease by about 25–35% at high NaCl concentrations
The quantitative agreement between the experimental values and trend of polymer water volume fraction (ϕw ) shown in Fig. 2. Addi
theoretical predictions are good for membranes with higher IEC values tionally, Dm
+ values increase as IEC increases, similar to the increase in ϕw
(i.e., IEC >0.44 meq/g), suggesting that thermodynamic non-idealities with increasing IEC (cf. Fig. 2). The similarities in the dependence of
in these membranes are primarily governed by strong electrostatic in Na+ diffusion coefficients and water volume fraction on Css and IEC can
teractions between fixed charges and ions captured by Manning’s be qualitatively rationalized by the Mackie and Meares model [cf. Eqn.
counter-ion condensation theory [35]. For membranes with lower IEC (8)], which predicts that ions diffuse faster in samples with higher water
values (i.e., IEC <0.44 meq/g), the experimental Ks values are below the content.
model predictions, likely due to additional thermodynamic To quantatitively compare experimental and theoretical results,
non-idealities arising from dielectric exclusion and unfavorable Fig. 6 (b) presents the experimental Na+ diffusion coefficients as a
function of water volume fraction. The experimental results are
7
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
predicted remarkably well by the Mackie and Meares model over a wide solutions. This behavior is qualitatively consistent with the Mackie and
range of ϕw values, regardless of whether ϕw was increased by adding Meares model [cf. Eqn. (8)]. As shown in Fig. 7 (a), Dm − values in the
fixed charges to the polymer backbone (i.e., varying IEC) or decreased uncharged membrane depend somewhat more strongly on Css relative to
by osmotic de-swelling caused by the increase in external NaCl con the weakly charged samples (i.e., IEC = 0.01–0.13 meq/g). The mo
centration. Thus, membrane hydration largely controls counter-ion lecular basis for this behavior is not fully understood.
diffusion in both uncharged and charged polymers where counter-ion Fig. 7 (b) presents experimental and theoretical Dm − values as a
condensation does not occur. Similar agreement was also found in function of water volume fraction. Given the relatively simple nature of
other polymers prepared with PEGDA of n = 13 and n = 4 (cf. Sup the Mackie and Meares model, which contains no adjustable parameters,
porting Information). this agreement is reasonable. However, some deviation is observed in
more highly charged samples for reasons that are not fully understood at
4.3.2. Cl- diffusion coefficients this time. Recently, the effect of electrostatic interactions between fixed
The Cl− diffusion coefficients were determined via Eqn. (13) and charges and ions on ion diffusion coefficients, as described by Manning’s
presented in Fig. 7 (a) as a function of upstream NaCl concentration. For model [26], was found to be small compared to the effect of tortuosity
all membranes considered, decreasing IEC and increasing upstream on ion diffusion, as captured by the Mackie and Meares model [25].
NaCl concentration led to decreasing Cl− diffusion coefficients, pri Similar results were also found for the membranes in this study. Recent
marily due to reduced water content in membranes with lower IEC simulation studies have observed reduced co-ion diffusion coefficients in
values and in membranes equilibrated with more concentrated salt charged matrices without obstructive polymer chains relative to those in
m*
Fig. 8. Apparent NaCl diffusion coefficients, < Ds >, as a function of: (a) upstream NaCl concentration and (b) polymer water volume fraction in membranes
prepared with PEGDA of n = 10. The dashed lines were drawn to guide the eye, and the solid line represents the Mackie and Meares model predictions [cf. Eqn. (8)].
m* m*
The uncertainty of < Ds >, determined using propagation of error analysis [54], was less than 15% of the average < Ds > value.
8
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
(18)
m*
Dms ≈ Ds ≈ Dm−
Fig. 10. The influence of upstream NaCl concentration on salt permeability coefficients of: (a) uncharged and weakly charged membranes (i.e., IEC <0.44 meq/g)
and (b) highly charged membranes (i.e., IEC >0.44 meq/g). These membranes were prepared with PEGDA of n = 10. The uncertainty, determined as the standard
deviation from measurements made on at least six separate samples, was less than 15% of the average of these measurements. The dashed lines were drawn to guide
the eye. The measured < Ps > value at Css = 0.1 M agrees with previous results reported for XLPEGDA membranes [7, 12].
9
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
Fig. 12. The influence of: (a) external NaCl concentration and (b) sorbed Na+ concentration on membrane ionic conductivity in membranes prepared with PEGDA of
n = 10. The uncertainty of κ, determined as the standard deviation from measurements made on at least six separate samples, was less than 25% of the average of
these measurements. The dashed lines were drawn to guide the eye.
10
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
F2 m m,p Fig. 13. Experimental and theoretical membrane ionic conductivity in mem
κ≈ D C (19) branes prepared with PEGDA of n = 10. The dashed line was drawn to guide the
RT + +
eye. The uncertainty of κ was determined as the standard deviation from
Thus, κ primarily depends on the counter-ion concentrations and measurements made on at least six separate samples.
diffusion coefficients. The counter-ions are more numerous in the sam
ples than the co-ions, so they carry most of the current [12]. As Css in
weakly charged samples (i.e., IEC <0.05), since the Cl− concentrations
creases from 0.01 to 1 M, Cm,p+ is relatively constant in the most highly in such materials cannot be accurately predicted by the Donnan/Man
charged membrane (i.e., IEC = 1.93 meq/g), as shown in Fig. 5. Dm + of ning model, which was designed for highly charged materials [33].
this sample also changes little with Css due to relatively constant water Minor discrepancies are also found in highly charged samples (i.e., IEC
content in the membrane [cf. Fig. 7 (a)]. Consequently, κ is essentially >1.46), which may be due to the effect of polymer-ion interactions on
independent of Css in the most highly charged membrane considered, as co-ion diffusion that is not captured by the Mackie and Meares model.
suggested by Eqn. (19).
Unlike the highly charged membrane, the uncharged membrane’s κ 5. Conclusions
value increases by orders of magnitude as Css increases, which is quali
tatively similar to the trend of Cm,p s
+ with increasing Cs in this sample (cf. Ion sorption, salt permeability, and ionic conductivity in a series of
Fig. 5). As mentioned earlier, equal numbers of Na+ and Cl− sorb in the uncharged and charged samples were characterized. Ion diffusion co
uncharged membrane to maintain electroneutrality (i.e., Cm m
+ = C− ). efficients were calculated from these data using the solution-diffusion
Thus, Eqn. (12) can be rewritten as follows: model and Nernst-Planck framework. Diffusion coefficients were
mainly affected by the sample’s water volume fraction, which is
F 2 m,p ( m )
κ= C D+ + Dm− (20) consistent with the Mackie and Meares model. Model predictions for
RT s
Na+ diffusion coefficients agree remarkably well with the experimental
where the sorbed salt concentration (Cm,p = Cm,p m,p results, using no adjustable parameters. Reasonable agreement was also
+ = C− ) and both Na
+
s
and Cl diffusion coefficients contribute to the membrane conductivity.
− found between experimental and theoretical Cl− diffusion coefficients.
Since Cm Minor deviations were observed in highly charged membranes, pre
s of the uncharged membrane has a relatively strong dependence
on Css , as shown in Fig. 5, and ion diffusion coefficients are weakly sumably due to interactions between polymer and ions that are not
dependent on Css [cf. Fig. 6 (a) and Fig. 7 (a)], the strong dependence of κ captured by the relatively simple Mackie and Meares model.
Salt permeability coefficients depend on salt sorption and diffusion
on Css of the uncharged membrane is largely dictated by the behavior of
coefficients. The salt concentration dependence of the salt permeability
Cm,p
s .
in uncharged membranes was mainly influenced by salt diffusivity,
Motivated by Eqn. (19) and (20), experimental κ values are presented
while the behavior in charged membranes was dominated by salt solu
in Fig. 12 (b) as a function of Na+ (i.e., counter-ion) concentration in the
bility. Increases in polymer IEC are often accompanied by a simulta
membrane. Fig. 12 (b) shows a strong correlation between ionic con
neous increase in fixed charge concentration and water content. When a
ductivity and Na+ concentration in almost all the membranes studied.
membrane is equilibrated with a dilute electrolyte solution, salt
Ionic conductivity values can be predicted from Eqn. (12), where Na+
permeability initially decreases as IEC increases due to the strong
and Cl− diffusion coefficients were estimated using the Mackie and
depression in salt solubility caused by Donnan exclusion. As IEC con
Meares model, Cl− concentrations were estimated using the Donnan/
tinues to increase, salt permeability increases slightly due to increased
Manning model, and Na+ concentrations were computed using Eqn. (3).
salt diffusivity induced by higher water content in the membrane at
Membrane fixed charge concentration at each external NaCl concen
higher IEC values. When a membrane is equilibrated with a concen
tration was estimated using the IEC and water uptake values, as reported
trated electrolyte solution, salt permeability increases monotonically
elsewhere [42]. Fig. 13 shows the predicted and measured ionic con
with increasing IEC, which is mainly governed by the increase of salt
ductivity results in a parity plot. Reasonably good agreement is found
diffusion coefficient as IEC increases.
between theoretical and experimental κ values, given that no adjustable
Membrane ionic conductivity depends on sorbed ion concentrations
parameters are used. Some deviations are evident in uncharged and
11
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
and diffusion coefficients in the membrane. The dependence of mem [13] H. Yasuda, C.E. Lamaze, L.D. Ikenberry, Permeability of solutes through hydrated
polymer membranes. Part I. Diffusion of sodium chloride, Makromol. Chem. 118
brane ionic conductivity on external NaCl concentration correlates well
(1) (1968) 19–35.
with the sorbed Na+ concentration, both for charged and uncharged [14] P. Meares, Ion exchange membranes: principles, production and processes, in: Ion
membranes. Reasonably good agreement was found between experi Exchange: Science and Technology, Springer, 1986, pp. 529–558.
mental and predicted ionic conductivity values. Some discrepancy was [15] E.Y. Safronova, D.V. Golubenko, N.V. Shevlyakova, M.G. D’yakova, V.A. Tverskoi,
L. Dammak, D. Grande, A.B. Yaroslavtsev, New cation-exchange membranes based
observed in uncharged and weakly charged samples, since the Donnan/ on cross-linked sulfonated polystyrene and polyethylene for power generation
Manning model only describes ion sorption well in highly charged systems, J. Membr. Sci. 515 (2016) 196–203.
sample, but fails in samples that are uncharged or weakly charged. [16] I.A. Prikhno, E.Y. Safronova, I.A. Stenina, P.A. Yurova, A.B. Yaroslavtsev,
Dependence of the transport properties of perfluorinated sulfonated cation-
exchange membranes on ion-exchange capacity, Membr. Membr. Technol. 2
Author contributions (2020) 265–271.
[17] J.G. Wijmans, R.W. Baker, The solution-diffusion model: a review, J. Membr. Sci.
107 (1995) 1–21.
Ni Yan: Conceptualization, Methodology, Validation, Formal Anal [18] G.M. Geise, H.S. Lee, D.J. Miller, B.D. Freeman, J.E. McGrath, D.R. Paul, Water
ysis, Investigation, Data Curation, Writing – Original Draft, Writing – purification by membranes: the role of polymer science, J. Polym. Sci. B Polym.
Review and Editing, Visualization, Project Administration, Rahul Suja Phys. 48 (2010) 1685–1718.
[19] G.S. Park, J. Crank, Diffusion in Polymers, Academic Press, New York and London,
nani: Visualization, Writing – Original Draft, Writing – Review and 1968.
Editing, Jovan Kamcev: Methodology, Writing – Original Draft, Writing [20] G.M. Geise, L.P. Falcon, B.D. Freeman, D.R. Paul, Sodium chloride sorption in
– Review and Editing, Kentaro Kobayashi: Data Curation, Writing – sulfonated polymers for membrane applications, J. Membr. Sci. 423–424 (2012)
195–208.
Original Draft, Writing – Review and Editing, Eui-soung Jang: Investi [21] J. Kamcev, M. Galizia, F.M. Benedetti, E.-S. Jang, D.R. Paul, B.D. Freeman, G.
gation, Writing – Original Draft, Writing – Review and Editing, Donald S. Manning, Partitioning of mobile ions between ion exchange polymers and
R. Paul: Writing – Conceptualization, Validation, Original Draft, Writing aqueous salt solutions: importance of counter-ion condensation, Phys. Chem.
Chem. Phys. 18 (8) (2016) 6021–6031.
– Review and Editing, Supervision, Benny D. Freeman: Conceptualiza [22] M. Galizia, F.M. Benedetti, D.R. Paul, B.D. Freeman, Monovalent and divalent ion
tion, Validation, Resources, Writing – Original Draft, Writing – Review sorption in a cation exchange membrane based on cross-linked poly (p-styrene
and Editing, Supervision, Project Administration, Funding Acquisition. sulfonate-co-divinylbenzene), J. Membr. Sci. 535 (2017) 132–142.
[23] J.S. Mackie, P. Meares, The diffusion of electrolytes in a cation-exchange resin
membrane. I. Theoretical, Proc. Roy. Soc. Lond.: Math. Phys. Eng. Sci. 232 (1191)
Declaration of competing interest (1955) 498–509.
[24] P. Meares, Ion-exchange membranes, in: Mass Transfer and Kinetics of Ion
Exchange, Springer, 1983, pp. 329–366.
The authors declare that they have no known competing financial [25] J. Kamcev, D.R. Paul, G.S. Manning, B.D. Freeman, Predicting salt permeability
interests or personal relationships that could have appeared to influence coefficients in highly swollen, highly charged ion exchange membranes, ACS Appl.
Mater. Interfaces 9 (4) (2017) 4044–4056.
the work reported in this paper. [26] G.S. Manning, Limiting laws and counterion condensation in polyelectrolyte
solutions II. Self-diffusion of the small ions, J. Chem. Phys. 51 (3) (1969) 934–938.
Acknowledgements [27] J. Mackie, P. Meares, The diffusion of electrolytes in a cation-exchange resin
membrane. II. Experimental, in: Proceedings of the Royal Society of London A:
Mathematical, Physical and Engineering Sciences, The Royal Society, 1955,
This work was supported as part of the Center for Materials for Water pp. 510–518.
and Energy Systems (M-WET), an Energy Frontier Research Center [28] P. Meares, The conductivity of a cation-exchange resin, J. Polym. Sci. 20 (96)
(1956) 507–514.
funded by the U.S. Department of Energy, Office of Science, Basic Energy
[29] P. Meares, Self-diffusion coefficients of anions and cations in a cation-exchange
Sciences under Award #DE-SC0019272. resin, J. Chem. Phys. 55 (1958) 273–279.
[30] P. Meares, D. Dawson, A. Sutton, J. Thain, Diffusion, conduction and convection in
synthetic polymer membranes, Ber. Bunsen Ges. Phys. Chem. 71 (8) (1967)
Appendix A. Supplementary data 765–775.
[31] W. McHardy, P. Meares, A. Sutton, J. Thain, Electrical transport phenomena in a
Supplementary data to this article can be found online at https://doi. cation-exchange membrane II. Conductance and electroosmosis, J. Colloid
Interface Sci. 29 (1) (1969) 116–128.
org/10.1016/j.memsci.2022.120549.
[32] N. Lakshminarayanaisah, Transport Phenomena in Membranes, Academic Press,
New York and London, 1969.
References [33] N. Yan, R. Sujanani, J. Kamcev, M. Galizia, E.-S. Jang, D.R. Paul, B.D. Freeman,
Influence of fixed charge concentration and water uptake on ion sorption in AMPS/
PEGDA membranes, J. Membr. Sci. 644 (2022), 120171.
[1] P. Długołęcki, K. Nymeijer, S.J. Metz, M. Wessling, Current status of ion exchange
[34] K.S. Pitzer, G. Mayorga, Thermodynamics of electrolytes. II. Activity and osmotic
membranes for power generation from salinity gradients, J. Membr. Sci. 319 (1)
coefficients for strong electrolytes with one or both ions univalent, J. Phys. Chem.
(2008) 214–222.
77 (19) (1973) 2300–2308.
[2] G.M. Geise, D.R. Paul, B.D. Freeman, Fundamental water and salt transport
[35] J. Kamcev, D.R. Paul, B.D. Freeman, Ion activity coefficients in ion exchange
properties of polymeric materials, Prog. Polym. Sci. 39 (1) (2013) 1–42.
polymers: applicability of Manning’s counterion condensation theory,
[3] E. Güler, R. Elizen, D.A. Vermaas, M. Saakes, K. Nijmeijer, Performance-
Macromolecules 48 (21) (2015) 8011–8024.
determining membrane properties in reverse electrodialysis, J. Membr. Sci. 446
[36] Q. Lei, K. Li, D. Bhattacharya, J. Xiao, S. Kole, Q. Zhang, J. Strzalka, J. Lawrence,
(2013) 266–276.
R. Kumar, C.G. Arges, Counterion condensation or lack of solvation?
[4] S. Maurya, S.-H. Shin, Y. Kim, S.-H. Moon, A review on recent developments of
Understanding the activity of ions in thin film block copolymer electrolytes,
anion exchange membranes for fuel cells and redox flow batteries, RSC Adv. 5 (47)
J. Mater. Chem. 8 (31) (2020) 15962–15975.
(2015) 37206–37230.
[37] R. Sujanani, L.E. Katz, D.R. Paul, B.D. Freeman, Aqueous ion partitioning in Nafion:
[5] A. Galama, M. Saakes, H. Bruning, H. Rijnaarts, J. Post, Seawater predesalination
applicability of Manning’s counter-ion condensation theory, J. Membr. Sci. (2021)
with electrodialysis, Desalination 342 (2014) 61–69.
638.
[6] J. Veerman, R.M. De Jong, M. Saakes, S.J. Metz, G.J. Harmsen, Reverse
[38] G.S. Manning, Limiting laws and counterion condensation in polyelectrolyte
electrodialysis: comparison of six commercial membrane pairs on the
solutions I. Colligative properties, J. Chem. Phys. 51 (3) (1969) 924–933.
thermodynamic efficiency and power density, J. Membr. Sci. 343 (1) (2009) 7–15.
[39] K.S. Pitzer, Thermodynamics of electrolytes. I. Theoretical basis and general
[7] L.F. Greenlee, D.F. Lawler, B.D. Freeman, B. Marrot, P. Moulin, Reverse osmosis
equations, J. Phys. Chem. 77 (2) (1973) 268–277.
desalination: water sources, technology, and today’s challenges, Water Res. 43 (9)
[40] J.H. Wang, S. Miller, Tracer-diffusion in liquids. II. The self-diffusion as sodium ion
(2009) 2317–2348.
in aqueous sodium chloride Solutions1, J. Am. Chem. Soc. 74 (6) (1952)
[8] T.Y. Cath, A.E. Childress, M. Elimelech, Forward osmosis: principles, applications,
1611–1612.
and recent development, J. Membr. Sci. 281 (2006) 70–87.
[41] J. Kamcev, D.R. Paul, G.S. Manning, B.D. Freeman, Accounting for frame of
[9] T.S. Chung, S. Zhang, K.Y. Wang, J. Su, M.M. Ling, Forward osmosis processes:
reference and thermodynamic non-idealities when calculating salt diffusion
yesterday, today and tomorrow, Desalination 287 (2012) 78–81.
coefficients in ion exchange membranes, J. Membr. Sci. 537 (2017) 396–406.
[10] T. Sata, Ion Exchange Membranes: Preparation, Characterization, Modification and
[42] N. Yan, D.R. Paul, B.D. Freeman, Water and ion sorption in a series of cross-linked
Application, Tokuyama Research: Tokuyama City, Japan, 2002.
AMPS/PEGDA hydrogel membranes, Polymer 146 (2018) 196–208.
[11] J. Kamcev, B.D. Freeman, Charged polymer membranes for environmental/energy
[43] A.J. Bard, L.R. Faulkner, J. Leddy, C.G. Zoski, Electrochemical Methods:
applications, Annu. Rev. Chem. Biomol. Eng. 7 (2016) 111–133.
Fundamentals and Applications, vol. 2, Wiley, New York, 1980.
[12] F. Helfferich, Ion Exchange, McGraw-Hill Book Co., Inc., New York, 1962.
12
N. Yan et al. Journal of Membrane Science 653 (2022) 120549
[44] H. Lin, B.D. Freeman, S. Kalakkunnath, D.S. Kalika, Effect of copolymer [51] J.C. Díaz, J. Kamcev, Ionic conductivity of ion-exchange membranes: measurement
composition, temperature, and carbon dioxide fugacity on pure-and mixed-gas techniques and salt concentration dependence, J. Membr. Sci. (2021) 618.
permeability in poly (ethylene glycol)-based materials: free volume interpretation, [52] P.J. Flory, Principles of Polymer Chemistry, Cornell University Press, 1953.
J. Membr. Sci. 291 (1) (2007) 131–139. [53] F.G. Donnan, Theory of membrane equilibria and membrane potentials in the
[45] H. Lin, E.V. Wagner, J.S. Swinnea, B.D. Freeman, S.J. Pas, A.J. Hill, presence of non-dialysing electrolytes. A contribution to physical-chemical
S. Kalakkunnath, D.S. Kalika, Transport and structural characteristics of physiology, J. Membr. Sci. 100 (1995) 11.
crosslinked poly (ethylene oxide) rubbers, J. Membr. Sci. 276 (1) (2006) 145–161. [54] P.R. Bevington, D.K. Robinson, J.M. Blair, A.J. Mallinckrodt, S. McKay, Data
[46] A.C. Sagle, E.M. Van Wagner, H. Ju, B.D. McCloskey, B.D. Freeman, M.M. Sharma, reduction and error analysis for the physical sciences, Comput. Phys. 7 (4) (1993)
PEG-coated reverse osmosis membranes: desalination properties and fouling 415–416.
resistance, J. Membr. Sci. 340 (1) (2009) 92–108. [55] L. Yuan-Hui, S. Gregory, Diffusion of ions in sea water and in deep-sea sediments,
[47] H. Ju, A.C. Sagle, B.D. Freeman, J.I. Mardel, A.J. Hill, Characterization of sodium Geochem. Cosmochim. Acta 38 (5) (1974) 703–714.
chloride and water transport in crosslinked poly (ethylene oxide) hydrogels, [56] M. Jardat, B. Hribar-Lee, V. Vlachy, Self-diffusion coefficients of ions in the
J. Membr. Sci. 358 (1) (2010) 131–141. presence of charged obstacles, Phys. Chem. Chem. Phys. 10 (3) (2008) 449–457.
[48] G.M. Geise, B.D. Freeman, D.R. Paul, Characterization of a novel sulfonated [57] M. Jardat, B. Hribar-Lee, V. Dahirel, V. Vlachy, Self-diffusion and activity
pentablock copolymer for desalination applications, J. Membr. Sci. 24 (51) (2010) coefficients of ions in charged disordered media, J. Chem. Phys. 137 (11) (2012),
5815–5822. 114507.
[49] J. Kamcev, E.-S. Jang, N. Yan, D.R. Paul, B.D. Freeman, Effect of ambient carbon [58] R.A. Robinson, R.H. Stokes, Electrolyte Solutions, Courier Corporation, 2002.
dioxide on salt permeability and sorption measurements in ion-exchange [59] G.M. Geise, B.D. Freeman, D.R. Paul, Sodium chloride diffusion in sulfonated
membranes, J. Membr. Sci. 479 (2015) 55–56. polymers for membrane applications, J. Membr. Sci. 427 (2013) 186–196.
[50] J. Kamcev, R. Sujanani, E.-S. Jang, N. Yan, N. Moe, D. Paul, B. Freeman, Salt
concentration dependence of ionic conductivity in ion exchange membranes,
J. Membr. Sci. 547 (2018) 123–133.
13