ArticlePDF Available

Nickel Hexacyanoferrate Electrodes for Continuous Cation Intercalation Desalination of Brackish Water

Authors:

Abstract and Figures

Using porous electrodes containing redox-active nickel hexacyanoferrate (NiHCF) nanoparticles, we construct and test a device for capacitive deionization in a two flow-channel device where the intercalation electrodes are in direct contact with an anion-exchange membrane. Upon reduction of NiHCF, cations intercalate into it and the water in its vicinity is desalinated; at the same time water in the opposing electrode becomes more saline upon oxidation of NiHCF in that electrode. In a cyclic process of charge and discharge, fresh water is continuously produced, alternating between the two channels in sync with the direction of applied current. We present proof-of-principle experiments of this technology for single salt solutions, where we analyze various levels of current and cycle durations. We analyze salt removal rate and energy consumption. In desalination experiments with salt mixtures we find a threefold enhancement for K+ over Na+-adsorption, which shows the potential of NiHCF intercalation electrodes for selective ion separation from mixed ionic solutions.
Content may be subject to copyright.
1
Nickel Hexacyanoferrate Electrodes for Cation Intercalation Desalination
Slawomir Porada,1 Pamela Bukowska,1 A. Shrivastava,2 P.M. Biesheuvel,1 and Kyle C. Smith3,4,5,*
1Wetsus, European Centre of Excellence for Sustainable Water Technology, Oostergoweg 9, 8911 MA
Leeuwarden, The Netherlands. 2Department of Materials Science and Engineering, University of Illinois at
Urbana-Champaign, Urbana, IL 61801. 3Department of Mechanical Science and Engineering, University of
Illinois at Urbana-Champaign, Urbana, IL 61801. 4Computational Science and Engineering Program,
University of Illinois at Urbana-Champaign, Urbana, IL 61801. 5Beckman Institute, University of Illinois at
Urbana-Champaign, Urbana, IL 61801. *Corresponding author’s email: kcsmith@illinois.edu
Abstract
Using porous electrodes containing redox-active nickel hexacyanoferrate (NiHCF) nanoparticles,
we construct and test a device for electrochemical water desalination in a two flow-channel device
where the electrodes are in direct contact with an anion-exchange membrane. Upon reduction of
NiHCF, cations intercalate into it and the water in its vicinity is desalinated; at the same time water
in the opposing electrode becomes more saline upon oxidation of NiHCF in that electrode. In a
cyclic process of charge and discharge, fresh water is continuously produced, alternating between
the two channels in sync with the direction of applied current. Compared to capacitive deionization
using porous carbon electrodes, a higher salt adsorption capacity per cycle is achieved, much lower
cell voltages are needed, and the energy costs of desalination can be significantly reduced.
----------
Electrochemical water desalination with porous electrodes can make use of two fundamentally
different mechanisms for salt storage. The first mechanism is capacitive electrosorption, where ions
are held in electrical double layers (EDLs) formed in the micropores of porous electrodes comprised
of ideally polarizable material (e.g., carbon) [1]. In the second mechanism, which has recently begun
research exploration [26], intercalation electrodes are used where ions are stored within the sites
of a solid-state host compound.
The first mechanism, capacitive electrosorption, is used in Capacitive Deionization (CDI), a
process in which ions are held near the carbon surface in the diffuse part of the EDL. CDI
electrodes are made of carbon (carbon nanotubes, graphene, activated carbon powder, etc.) which
can be processed into porous, ion- and electron-conducting, thin electrode films, suspensions, or
fluidized beds [7]. CDI based on capacitive EDL charging is a promising method, but to reach a
certain salt adsorption capacity (SAC; a typical number being of the order of SAC=5-15 mg/g,
referring to mass of NaCl removed, per total mass of carbon in a two-electrode cell, measured at a
standard cell voltage of Vcell=1.2 V), the energy input is not insignificant [810], while the current
2
efficiency (quantifying the fraction of current input that results in salt adsorption) of CDI cells can
be well below unity, implying that in the charging process not only counterions adsorb but also
coions desorb from the electrode [11]. In CDI with membranes, or using improved charging
schemes, can be close to unity [10].
Like capacitive carbons, intercalation host compounds (IHCs) can be incorporated into porous
electrode films and can adsorb charge, but the ion storage mechanism of IHCs is fundamentally
different from EDL charging. In intercalation electrodes, ions are stored in the crystallographic sites
of the IHC as a result of its redox activity. Water desalination using IHCs, which is currently much
less developed and utilized than CDI, has the advantage that to reach a certain SAC a much lower
voltage and energy is needed than using capacitive electrosorption, because the change in
electrode potential with electrode charge can be much lower. Also, IHCs have the potential to
selectively remove one ion (e.g., Na+) out of a multi-ion mixture with other ions of the same valence
and charge.
Figure 1. Schematic diagram of the present CID flow cell with a single anion exchange membrane
(AEM) and two adjacent and identical electrodes containing NiHCF IHC. A flow channel (“spacer”) is
attached to each current collector and two of these structures sandwich the cell. In one half of the
cycle (A), sodium ions de-intercalate from the left IHC while chloride ions migrate through the
membrane, to the left, producing water of increased salinity on the left-hand side. During this stage,
electronic current runs to this electrode through the external circuit. At the same time, on the right-
hand side, sodium ions intercalate into the PBA, and desalinated water is produced. After some
time, the maximum amount of intercalated sodium is reached in the right-hand electrode and the
current direction must be reversed (panel B).
AEM
Cl-
Cl-
Cl-
Na+
Na+
Na+
Na+
Na+
Na+
Na+
Na+
Na+
Cl-
Cl-
Cl-
Feed water
Concentrated water Desalinated water
AEM
Cl-
Cl-
Cl-
Na+
Na+
Na+
Na+
Na+
Na+
Na+
Na+
Na+
Cl-
Cl-
Cl-
Feed water
Desalinated water Concentrated water
A B
Intercalated
sodium
AEM
Na+
Sodium ion
Spacer and
current collector
Cl-
Concentrated water Desalinated water Desalinated water Concentrated water
time
3
The use of IHCs for water desalination has been reported using several novel cell architectures.
The “Desalination Battery,a cell consisting of one Na2Mn5O10 (NMO) electrode and one Ag/AgCl
electrode, was used for water desalination by adsorption of cations within NMO and adsorption of
Cl- ions by conversion of Ag to AgCl [3]. In “Hybrid CDI [2], an IHC electrode (either NMO [2] or
Na2FeP2O7 [12]) was combined with a carbon capacitive electrode for anion adsorption. In both of
these desalination devices a single cation-intercalating electrode was paired with a different
electrode that adsorbs anions. Later, Smith and Dmello (SD) [4,13] proposed desalination with Na-
ion IHC electrodes having identical chemical composition, originally referred to as Na-Ion
Desalination (NID) [4]. Presently, we refer to this technology as Cation Intercalation Desalination
(CID) to emphasize the charge adsorption mechanism by which desalination is achieved. While the
use of the same IHC in both electrodes is impractical for battery use, SD showed that desalination is
possible, in theory, with electrodes of identical composition. To achieve this, an innovative cell
design was proposed where porous IHC electrode films are placed on either side of a separator
layer with feed water in a “flow-through” mode directed through the electrode, along the separator
layer. SD showed that for an IHC that is cation-adsorbing, the separator layer must be cation-
blocking to achieve high salt removal, i.e., an anion-exchange membrane (AEM) must be used. In
this way a CID device operates where, during one half of a charging-discharging cycle, one channel
produces desalted water and the other brine, and, during the other half of the cycle, the situation is
reversed. Recently, Srimuk et al. experimentally demonstrate the use of identical intercalation
electrodes for desalination in a membraneless cell, where a MXene intercalation compound is used
which is host to both anions and cations. Xing et al. recently used a hybrid cell with MoS2 as a
cathode, combined with a capacitive anode [6].
The recent development of IHCs for aqueous rechargeable Na-ion batteries (ARSBs) has been
stimulated by the low cost compared with non-aqueous Li-ion systems and by the inherent safety of
water-based electrolytes [14]. By leveraging the CID device concept these developments can be
applied to electrochemical water desalination. Also, the plethora of Na-ion IHCs developed as
cathodes for ARSBs can be applied as both cathode and anode in CID, eliminating the need to
search for low-potential IHCs, which, for ARSBs, are less numerous compared to high-potential
IHCs. Theoretical predictions of CID performance were originally obtained [4] using symmetric cells
with either NMO, a common ASRB cathode [1517], or NaTi2(PO4)3 (NTP), a common ASRB anode
[16,18,19]. These respective materials have high specific capacities of 45 mAh/g-NMO [15] and 120
mAh/g-NTP [16], but other IHCs exist that also show promise for use in CID. Among the various
cathode compounds developed, nanoparticulate Prussian Blue Analogues (PBAs) stand apart for
their high cycle life and facile cation intercalation kinetics that have enabled their demonstration use
in Na+ [2023], K+ [20,21], Mg2+ [24], Ca2+ [24,25], and Zn2+ [26] ion batteries. Here, we use a
particular PBA, nickel hexacyanoferrate (NiHCF), to exchange Na ions reversibly with saline
4
solution flowing through a CID cell. Its average reduction potential of ~0.6 V vs. SHE [20,23], though
modest as a cathode in ARSBs, is 200 mV below the O2 evolution potential at neutral pH. As a
result, minimal electrolyte decomposition is expected even at large overpotentials applied to CID
cells using these IHCs. Also, Smith recently predicted that NiHCF has sufficient capacity to
desalinate seawater [27]. Based on the results in our present work, we calculate that this material
can potentially achieve in brackish water SACs of ~40 mg/g-electrode (assuming 59 mAh/g-NiHCF
capacity [20,23]; current efficiency 92%; charge capacity utilization 85%; IHC 80 wt% of electrode
mass; parameters to be explained later on).
In this work we test and demonstrate desalination within a CID cell for the first time
experimentally. Here, desalination is achieved by using redox-active NiHCF nanoparticles to
intercalate Na+ ions from NaCl solution. We synthesize Na-rich NiHCF using precipitation chemistry
and fabricate porous electrode films from them. Prior to CID cell testing electrochemical titration is
used to characterize the equilibrium relationship between potential and stored charge in a half-cell
using non-flowing, concentrated Na2SO4 electrolyte. These porous electrode films are then
integrated within a CID cell adjacent to an AEM (Fig. 1). Feed water is flowed behind the porous
electrodes through flow channels adjoined to each of the cell’s current collectors, enabling
continuous desalination of feed water. The dynamic, cyclic operation of this CID cell is subsequently
characterized at a salinity that is representative of brackish water. A systematic study of desalination
performance is then conducted by varying cycle time with the same current in each case,
demonstrating the range of electrode charge, salt adsorption capacity, and energy consumption
achievable using NiHCF nanoparticles in a CID cell.
NiHCF has previously been synthesized for Na-ion batteries in both Na-deficient [20] and Na-rich
[23] forms, the latter form enabling batteries to be assembled and charged without pre-sodiation
when it is used as a positive electrode. In this work, we synthesize NiHCF in a Na-rich state, with
Na2NiFe(CN)6 being the ideal composition [19]. Synthesis of Na-rich NiHCF (with Na4Fe(CN)6
precursor) eliminates introduction of additional cationic species that are present in the typical
synthesis of Na-deficient NiHCF, which, for example, used K3Fe(CN)6 precursor in ref. [20] and,
thus, resulted in K+ within the synthesized NiHCF structure. The present Na-rich NiHCF material is
later converted to intermediate compositions via electrochemical oxidation in a three-electrode cell.
NiHCF particles undergo nucleation and growth as the precursors are combined, and the manner in
which this is done will affect the final particle size distribution as well as the crystallinity of the
synthesized particles. We use a solution/precipitation method with consistent reaction conditions to
synthesize NiHCF (see Methods).
Based on ex situ elemental analysis Na-rich (as-prepared) NiHCF samples were found to have a
composition of Na1.20Ni1.28Fe1.22C6N5.75.4.41H2O (see Methods). This composition includes
5
significantly less Na per formula unit CN than that of the theoretical composition (Na2NiFe(CN)6),
also with a substantial amount of hydrated H2O within its lattice. In light of these factors, the
presence of H2O trapped within the interstitial sites of the NiHCF structure may play a role in limiting
the incorporation of Na+ ions. Despite their non-ideal composition, we note that the present NiHCF
possesses more than double the degree of alkali cation intercalation and similar nickel stoichiometry
per CN formula unit compared with those of Na/K-deficient NiHCF synthesis
(K0.6Ni1.2Fe(CN)6·3.6H2O, ref. [20]). These observations demonstrate that we have successfully
synthesized Na-rich NiHCF. The crystal structure of as-synthesized NiHCF was confirmed by X-ray
diffraction (see Methods) to belong to the Fm3m space group with a lattice parameter of 10.25 , in
good agreement with literature [20]. Peak positions match well with previous reports in literature for
Prussian Blue [28] and NiHCF [20,23] (see Supplementary Information, SI). The crystallinity and
physical morphology of NiHCF particles were analyzed using transmission (TEM) and scanning
electron microscopy (SEM, see SI). TEM images show nanocrystallites between 2 to 7 nm, with
mean crystallite size of approximately 4 nm. Nanocrystallite shapes appear as rounded cubes (in
contrast with previous reports of highly faceted PBA nanoparticles [29]), likely as a result of crystal
defects introduced during nanocrystal growth. SEM images of the NiHCF particles showed a range
of primary and secondary particle sizes. While secondary particles were observed with diameters on
the order of 10 µm, detailed examination at higher magnification reveals the presence of
nanoparticles with diameters on the order of 100 nm, consistent with prior reports on NiHCF [20,23].
The nanoparticulate morphology of NiHCF was further corroborated by N2 gas adsorption
measurements. This measurement was also used to assess the size distribution of pores within the
electrodes (see Methods). As shown in the SI, in the range up to 30 nm pore size, porosity analysis
of synthesized NiHCF powder resulted in a pore volume of 0.028 mL/g, a BET area of 15 m2/g, and
a surface area using non-linear NLDFT theory of 12 m2/g [30]. Assuming a density for NiHCF of 2.0
g/mL (based on its ideal stoichiometry and XRD analysis), we find a corresponding pore volume per
unit NiHCF volume of 5.6%, revealing ultra-low microporosity indicative of nanoparticle aggregation.
Accordingly, this powder was ball milled to separate nanoparticle aggregates (see Methods).
Electrodes were subsequently prepared using a procedure similar to that reported in ref. [31],
combining 80 wt% NiHCF with 10 wt% PTFE binder and 10 wt% conductive carbon black, after
which the electrodes were calendered to increase electronic conductivity (see Methods). After this
process the BET area of the electrode was 103 m2/g, and pore volume (in the range to 30 nm pore
size) was 0.12 mL/g or 29 vol% (assuming 2.0 g/mL density, as before), confirming the separation
of nanoparticulate aggregates that is necessary for high electrochemical activity. SEM
characterization of the electrode showed a homogeneous microstructure consisting of NiHCF,
carbon black, and macroporosity (see SI). The macropores within the electrode (which, during cell
operation, is filled with saline water) are estimated to have a porosity of 40 to 50 vol%.
6
In the context of desalination by CID, electrode charge translates to salt removal, as Na ions are
the primary intercalant species storing charge within the electrodes, and, therefore, understanding
the cycling characteristics of each electrode is essential to constructing a high-performance CID
cell. To illustrate this principle, consider a dimensionless number between 0 and 1 denoting the
fraction of cation-filled sites within the IHC, hereafter referred to as the intercalation degree θ. For
an electrode with θ=0 the IHC contains no exchangeable intercalant, while at θ=1 the IHC is
saturated with intercalant. To convert from intercalation degree θ, for which 0<θ<1, to electrode
charge Q in mAh/g, we multiply by the maximum charge Qmax at full intercalation, i.e., Q=Qmaxθ. For
the half-cell cycling results discussed later, equivalent spans of these two scales are shown in
Fig. 2B. In desalination with Na-ion IHCs (such as NiHCF), the intercalation degree within opposing
electrodes evolves (on average) in a symmetrical fashion when side reactions are negligible,
switching in the fraction of intercalated Na, θ, between two states of θ (see ref. [4,32]). During one
half of the cycle the composition in one electrode changes from θ1 to θ2 21), while in the other
electrode, simultaneously, the composition changes from θ4 to θ3 43). Such operation is referred
to as the “rocking-chair” cycling mechanism utilized commonly in Li-ion [33] and many other
rechargeable batteries using IHCs. The importance of this principle to the CID concept is exactly
this: cations are intercalated from the electrolyte in one electrode, while at identical rate cations are
deintercalated into the electrolyte in the opposing electrode [4]. When the IHCs in both electrodes
are identical and are of equal mass, this action produces an equal and opposite change in
composition, i.e., θ2-θ14-θ3. Thus, based on the use of IHCs an innovative cell design is proposed
with only one ion-exchange membrane and two chemically identical IHC electrodes.
Half-cells of NiHCF electrodes were characterized in 1 M Na2SO4 solution (often used in ARSBs)
using the galvanostatic intermittent titration technique (GITT, see Methods). Typically, when IHCs
are used for energy storage in batteries, the principal metric of consideration is the maximum
electrode charge in mAh/g. The equilibrium potential versus electrode charge in Fig. 2B, extracted
from GITT, shows a maximum charge of 59 mAh/g (similar to previous reports for NiHCF [20,23])
with a region in which potential varies in a roughly linear manner between 5 and 55 mAh/g (per
gram NiHCF in one electrode). From half-cell characterization we also find that the relationship
between equilibrium potential and the intercalation degree within NiHCF particles can be correlated
by the Temkin isotherm [37],
E=Eref-RT/F*ln(cNa,/c0 θ/(1-θ))-gθ, (1)
where at room temperature RT/F=25.6 mV and the interaction parameter g is positive when
intercalated Na-ions in the IHC repel one another, and therefore do not phase-separate into Na-rich
and Na-deficient regions (which is predicted when g<-100 mV). Here, cNa, is the Na+ concentration
within the electrolyte surrounding IHC particles, and c0=1 M is a reference concentration. A best fit
7
to the data is obtained using Eref=510 mV vs. Ag/AgCl and g=+90 mV. Such a correlation is also an
essential input for porous-electrode modeling of Na-ion batteries and CID, which can affect the
dynamic of ions within electrodes during operation [4,38].
Figure 2. Electrochemical titration of a single NiHCF IHC in 1 M Na2SO4. In this experiment, a
constant current of 60 mA/g of NiHCF is applied for 118 s, after which current is set to zero for 60 s.
Around 1.5 hr, current direction is reversed. From the time-dependent trace of voltage (panel A), we
take the voltage at the end of the “rest”-period and construct panel B, which shows the equilibrium
potential vs charge of the NiHCF-material, with excellent reversibility between de- and intercalation
of Na+. The Temkin-equation fits data according to Eq. (1) well.
To compare CID performance for the present electrodes to that of typical capacitive electrodes in
CDI, we calculated the differential capacitance, -dQ/dE. Over the linear range of potential variation,
capacitance exceeds 1000 F/g, which is much higher than values for purely capacitive electrodes
which are typically not higher than 100 F/g defined per single electrode voltage and single electrode
mass. This difference in capacitance implies that, to store the same charge, a much lower voltage is
needed for intercalation electrodes than for capacitive electrodes. As an example, to achieve the
same salt adsorption capacity, with C ten-fold higher (as for IHC electrodes compared to capacitive
ones) and assuming the same current efficiency, we require a final cell voltage that is ten-fold lower,
and thus, with the energy proportional to CV2, the energy input can be ten-fold lower.
Desalination experiments, as reported in Fig. 3, are performed at a salt concentration in the feed
water of 20 mM and a current density of 2.8 A/m2 (see Methods). A constant current (CC) mode was
used because (1) simulation has predicted that such operation produces nearly constant effluent
concentration after an initial transient period [4,39] and (2) recent analysis of CDI cycling modes has
shown that the CC mode is more energy efficient than the constant potential mode [9]. This
particular current density is small in comparison with typical CDI experiments, but large cell
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.00
0.25
0.50
0.75
1.00
Voltage (V) vs. Ag/AgCl
Time (hr)
Time 103 (s)
010 20 30 40 50 60
-0.25
0.00
0.25
0.50
0.75
1.00
Voltage (V vs. Ag/AgCl)
Electrode charge (mAh/gNiHCF)
0.0 0.2 0.4 0.6 0.8 1.0
Intercalation degree
OCV
A
Bintercalation
deintercalation
1 M Na2SO4, 60 mA/gNiHCF
theory
+ 5 C + 5 C + 5 C
= 0.0
= 0.5
= 1.0
8
polarization at high currents prevented further increases in current density at the present feed water
salinity level. The large cell voltages that developed at higher currents may be due to an ion
transport rate limitation in the cell. Specifically, Na+ ions must transport between the porous
electrode and the flow channel, and Cl- ions must transport between the porous electrodes on
opposing sides of the cell through the membrane. Thus, the microstructure and thickness of the
porous electrodes, as well as those of the flow channels, can be optimized to enable efficient
operation at higher rates, as has been predicted previously [4]. In addition, because NiHCF is a type
of metal-organic framework whose CN- ligands are poorly conducting, electronic conduction is
expected to be slow within these particles. As such, it is possible that the inhomogeneous
distribution of conductive carbon black could limit the transmission of electrons to NiHCF particles.
Solid-state diffusion of Na+ ions within NiHCF particles may also have played a role, depending on
the degree of size dispersity among them. Membrane resistance is not expected to affect
performance substantially due to the high concentration of membrane fixed charge in comparison
with influent salinity. The exact reason for the high resistance remains to be identified and solved.
Figure 3. Desalination cycles using NiHCF IHCs at a constant current of 2.8 A/m2 in 20 mM NaCl,
for sets of cycles where the charging time is reduced after each 2 or 3 cycles, starting at a situation
where one electrode is almost completely intercalated with Na+, and the other is in an almost
completely deintercalated state. Panel A presents one cycle from each set of 2 or 3 similar cycles,
showing both desalination (concentration change <0) and salt release (concentration change > 0).
Note that here only results for the effluent salinity from one of the two channels is shown. In panel B
profiles of cell voltage versus time are shown during the same desalination cycle. Here, to show the
symmetry in the process, the voltage trace during the second half of the cycle is vertically switched
and moved left, to overlap with the voltage trace in the earlier half-cycle.
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Cell voltage (V)
-1.0
-0.5
0.0
0.5
1.0
Concentration change (mM)
A
B
2.5 hr
90 C/gNiHCF 80 70 60 50 40 30 20
2.5 hr
90 C/gNiHCF 80 70 60 50 40 30 20
9
To characterize desalination dynamics, we measure the time variation of effluent salinity and cell
voltage in cycles of different duration but at the same current. Data in Fig. 3 were obtained in cell
where initially one electrode is almost completely de-intercalated and the other almost fully
intercalated following the approach proposed by SD [4]. Here, the total charge transferred between
electrodes was reduced from 90 to 20 C/g-NiHCF in steps of 10 C/g every 2-3 cycles (Fig 3 shows
results of one cycle for each condition). In the SI we present results of a different experiment where
the cell starts with two electrodes that initially have a similar intercalation degree θ (in the range 0.4-
0.6). Here, the total charge transferred during each cycle was varied in a similar manner to that of
the first routine, but in increasing order from 20 to 90 C/g-NiHCF. Results of this experiment match
very closely with those presented in Fig. 3. In Fig. 3, panel A shows the effluent salinity (minus
inflow salinity) during one cycle, as measured in the effluent of one of the two channels. It is
expected that the other channel has exactly opposite (“symmetric”) behavior, i.e., a salinity-versus-
time trace that is the mirror image (reflection about the horizontal dashed line) of the one measured.
This expectation is based on the fact that both channels and electrodes are chemically identical
(except for the fact that at any moment in time the value of θ will be different), and that voltage-
versus-time traces during charging and discharging overlap quite well, as shown in panel B. Thus,
though we did not measure the salt effluent salinity from both channels simultaneously, it is difficult
to envision a scenario where one channel exhibits the behavior presented in Fig. 3, while the
effluent salinity-versus-time trace in the opposing channel would be very different. Nevertheless, in
future work it is relevant to measure the effluent salinity of both channels simultaneously.
In Fig. 3, panel A shows that effluent salinity change (i.e., the difference between effluent and inlet
salinity levels) is fairly constant, in line with the expectation for operation at constant current. With
decreasing charge transfer, the half-cycle time of charging and discharging (desalination and salt
release) decreases, from around 5 hr in the longest experiments (i.e., for a charge of 90 C/g,
defined per unit NiHCF mass in both electrodes) to 1 hr at 20 C/g-NiHCF, but for all cycles the
influent/effluent salinity change is constant. The cell voltage during charge and discharge is plotted
in panel B. Here, to show overlap between the two half-cycles, the voltage-time trace during one
half-cycle is vertically mirrored and shifted left, to overlap with the other voltage-time trace. In
general we observe excellent agreement, and thus symmetric behavior in the two halfs of the cycle,
as expected. Some differences are observed between the two traces for the lower values of charge
transfer, for which we do not have an explanation. Otherwise, the two voltage-time traces overlap
well. Fig. 3B shows how, at the start of each half-cycle after the reversal of current direction, the cell
voltage rises rapidly in time. Such a voltage spike can typically be ascribed to the effect of a linear
resistance (e.g. in wires or across a separator). This effect is indeed largely responsible for the
observed spike, but cannot explain it entirely, as its magnitude depends on the total charge in
addition to the applied current. Thus, other aspects related to the state-of-charge of the NiHCF IHC
must also have played a role.
10
In general, the results presented in Fig. 3 are consistent with our expectations that cell voltage
gradually increases within a half-cycle, and thus for longer cycles it reaches a higher value. We also
observe that neglecting the short initial periods of negative cell voltage in all instances the cell
responds with a positive voltage when a charging current is applied (i.e., charging needs energy
input). Furthermore, in none of the cycles do we see a linear increase in cell voltage versus time (or,
equivalently, versus charge), as would be expected on the basis of the linear behavior of the
“plateau” region of the voltage-charge curve in Fig. 2B. This deviation is linked to rate limitations
mentioned before. This conclusion is by a comparison of the large cell voltages applied, to those
expected based on equilibrium cycling: for a maximum charge transferred of 90 C/g-NiHCF (two
electrode mass), which translates to 50 mAh/g-NiHCF (single electrode mass), each electrode will
change its equilibrium potential by approximately 175 mV, and thus in the ideal case, without ion or
electron transfer limitations, the potential of each electrode (i.e., with respect to a stationary
reference) varies between -175 to +175 mV (neglecting the effect of concentration polarization),
manifesting as a maximum cell voltage of 0.35 V. Instead, for the cycle with the highest charge, we
observe the cell voltage approaching 1.0 to 1.5 V at the end of a half-cycle. Thus, we believe that
transport limitations must have played a role in limiting the rate capability of the present CID
experiments, as a result of sluggish pore-scale or solid-state transport processes. We reiterate that
the charging behavior in 1 M Na2SO4 with higher current densities, as discussed in Fig. 2, did not
show such strong rate limitations, but note that in such an electrolyte the Na-concentration is
hundred times larger than in the CID experiments. This aspect of our electrode design requires
further study and optimization.
Figure 4. A) Salt adsorption capacity of NiHCF IHCs, in mg NaCl per cycle per gram of both
electrodes vs. electrode charge (per gram NiHCF in one electrode) for the experiments of Fig. 3
together with data at a 2x lower current density. B) Current efficiency, which is the ratio of salt
adsorption rate over current density, see Eq. (2) (~0.8 for I=2.8 A/m2, and ~0.95 for I=1.4 A/m2).
010 20 30 40 50
0.0
0.2
0.4
0.6
0.8
1.0
Current efficiency
Electrode charge (mAh/gNiHCF)
036 72 108 144 180
Electrode charge (C/gNiHCF)
010 20 30 40 50
0
10
20
30
40
Salt adsorption (mg/gelec)
Electrode charge (mAh/gNiHCF)
036 72 108 144 180
Electrode charge (C/gNiHCF)
A B
20 mM NaCl, 2.8 A/m2
20 mM NaCl, 1.4 A/m2
11
Figure 4 summarizes all data presented in Fig. 3 for a current density of 2.8 A/m2 as function of
electrode charge and plots salt adsorption capacity, SAC, which is expressed in mg NaCl adsorption
per mass of both electrodes (not per mass of active component, which in our case is 80 wt% of the
total electrode mass). The highest value of SAC here, 34 mg/g, is about 2.5 times a reference value
for CDI with capacitive activated carbon electrodes of SAC=12.5 mg/g [1,34]. Though SAC in CDI
can be increased to higher values when using higher cell voltages, membranes, or chemically
modified electrodes [35,36], on this metric CID using NiHCF IHCs quite well competes with state-of-
the-art CDI technology. Fig. 4A also presents a limited number of data at a twice lower current
density (1.4 A/m2), related to one voltage-time trace that is shown in Fig. 5A. For both current
densities, Fig 4B presents data for current efficiency , which is the ratio of average salt adsorption
rate, Jsalt,avg (which is SAC multiplied by two times the electrode mass, Mel, divided by half-cycle
time, HCT, and NaCl mass, Mw=58.44 g/mol), over current, I, in A, divided by F=96485 C/mol, and
thus is given by
=(2SACMel/(HCTMw))/(I/F). (2)
For current efficiency, , both data sets show a nearly perfect independence of on electrode
charge, with ~0.8 at 2.8 A/m2 and ~0.95 at 1.4 A/m2, see Fig. 4. Ideally, for perfect membranes
(i.e., only counterion transport is permitted) and perfect cation intercalation, the current efficiency
must be unity. Various factors may be responsible for the value of being less than unity. One
factor contributing to this effect, at least in part, is the loss of charge through undesired side
reactions, such as O2 evolution. Such a reaction is expected to occur to a greater degree when
cycling with large overpotentials that are experienced at higher current densities. Additionally,
though unreported for previous PBAs using Na2SO4 electrolyte, it is possible that NiHCF not only
intercalates Na+ ions, but also anions. Based on a simple model we determine that salt back-
diffusion through the AEM seems an unlikely explanation under the present experimental conditions
(see SI). The influence of current density on current efficiency deserves further attention and the
CID process and materials should be improved to increase to a value much closer to unity, also at
higher currents than presently employed.
Finally, we analyze the energy invested in desalination by integrating the cell voltage versus
charge in one half-cycle (see Fig. 5A) and divide by salt removal (given by the salt concentration
difference between inflow and outflow, multiplied by the water flow rate). Making this analysis for the
data at 30 C/g-NiHCF from Fig. 3B, which results in a SAC of around 12.5 mg/g (similar to many
CDI data), we obtain for 2.8 A/m2 current density a value of ~12 kT/ion removed, and even as low
as ~3 kT/ion at a current of 1.4 A/m2 (Note that energies in kT/ion can be multiplied by 5 to obtain
the energy consumption in kJ per mol salt removed), which are clearly below those for MCDI
reported in refs. [1,8]. Figure 5B analyzes energy consumption in these experiments, and compares
12
the “energy per ion” for these two cases of CID operation, with membrane-CDI data from ref. [8]. We
clearly observe that, based on this metric, CID data at the lowest current (1.4 A/m2) show more than
a five-fold reduction in energy consumption compared to MCDI. This five-fold reduction is half the
factor of 10 that we calculated earlier on the basis of the difference in capacitance between
capacitive and intercalation electrodes, and thus a larger gain in energy consumption is still feasible.
In conclusion, though salt adsorption rates are for now very slow, the results in Fig. 4 and 5
provide evidence that there is significant potential for an enhancement in salt adsorption capacity
and a reduction in energy use in electrochemical water desalination when employing intercalation
host electrodes.
Figure 5. A) Cell voltage increase in time, for two values of the current density, in a cycle where the
charge transfer is 30 C/g-NiHCF. B) Energy consumption per amount of salt removal, where CID
cell performance with NiHCF is compared with data for membrane-CDI [8].
Methods
Synthesis. Disodium nickel hexacynanoferrate was synthesized by a solution/precipitation reaction
with aqueous reagents using the following synthesis reaction:
NiCl2+Na4Fe(CN)6Na2NiFe(CN)6 + 2NaCl. A previous recipe used to synthesize Na-rich NiHCF
[23] was employed with several modifications by incorporating findings from a recipe used to
synthesize Na-rich manganese hexacyanoferrate [40]. In this procedure, 20 mM of
Na4Fe(CN)6·10H2O (Sigma Aldrich) and 2.40 M of NaCl (Sigma Aldrich) were dissolved in a mixture
of pure water with 25 wt% pure ethanol (compared to pure water). Subsequently, 40 mM of aqueous
solution of NiCl2·6H2O (Alfa Aesar) was added dropwise to the mixture of Na4Fe(CN)6, NaCl, water
and ethanol. The solid NiHCF precipitate was filtered, washed with pure water, dried in a vacuum
oven at 50°C to remove residual moisture, and ground using a ball mill.
0
5
10
15
20
Energy per ion removed (kT)
B
-0.2
0.0
0.2
0.4
0.6
0.8
Cell Voltage (V)
1.0 hr
A30 C/gNiHCF
current
density
2.8 A/m2
Capacitive electrosorption - MCDI
I = 2.8 A/m2
Intercalation electrodes
1.4 A/m2
13
Material Characterization. Elemental analysis was conducted with NiHCF decomposed into
elemental species in acid using a PerkinElmer 2400 Series II CHN/O Elemental Analyzer and a
PerkinElmer 2000DV ICP-OES instrument. X-ray diffraction was conducted using a Bruker D8
Venture instrument with Cu-Kα radiation (λ=1.54 , 0.02˚ angular resolution, and < < 100˚).
Both samples showed slight peak broadening indicative of their nanocrystalline microstructure, and
several high-angle peaks were not apparent due to background noise. Also, it should be noted that
fluorescent interaction between Fe and X-Ray radiation emitted from Cu can produce mild
inaccuracy in the measured intensity. N2 adsorption measurements were performed at -196°C using
a TriStar 3000 gas adsorption analyzer (Micromeritics).
SEM samples were deposited and pressed gently onto carbon tape attached to an aluminum
sample holder. To minimize charging the samples were coated with approximately 4 nm of gold-
palladium film using a Denton (Moorestown NJ) Desk-II turbo sputter coater, and imaged at 5 kV in
HiVac mode using a Quanta 450 FEG environmental scanning electron microscope (FEI Company,
Hillsboro OR). TEM samples were prepared by dispersion in deionized water by sonication in a
water bath for approximately 5 minutes, followed by vibration using a Fisher S56 Miniroto touch
shaker. Dried droplets were then glow-discharged on wax film and subsequently transferred to
carbon-stabilized Formvar-coated TEM grids (cat. no. 01811; Ted Pella, Inc., Redding CA). TEM
images were obtained using a Philips/FEI CM200 instrument at 160 and 200 kV.
Electrode Preparation. 80 wt% of the NiHCF powder was mixed with 10 wt% carbon black (Vulcan
XC72R, Cabot Corp., Boston, MA) and 10 wt% polytetrafluoroethylene binder (PTFE, 60 wt%
solution in water from Sigma Aldrich, USA) and pure ethanol to obtain an homogeneous slurry. The
final electrodes were calendered using a rolling machine (MTI HR01, MIT Corp.) to produce a
thickness of 200 m. Prior to CID experiments, electrodes were cut into 6x6 cm squares and dried
in a vacuum oven at 50°C to remove residual moisture, which is necessary to measure electrode
dry mass.
Half-Cell Characterization. Electrochemical characterization of NiHCF material was performed by
constructing a three-electrode half-cell with the working electrode containing 0.98 g of NiHCF in
1 M Na2SO4, cycled against a titanium mesh electrode coated with Ir/Ru (Magneto Special Anodes
B.V., the Netherlands) as counter electrode and Ag/AgCl as a reference electrode placed in the
vicinity of the working electrode. In “bursts” of 5 C (at a current of 60 mA/g-NiHCF in the electrode)
the working electrode was charged, after which a “rest period” (“OCV” for “open circuit voltage”) of
60 s follows in which the potential relaxes (see inset in Fig 2A). We continued this process until the
electrode potential diverged strongly, implying full (de-)intercalation of Na ions was reached. The
electrode potential at the end of each OCV period was taken as equilibrium potential (versus
Ag/AgCl) and plotted as a data point in Fig 2B.
14
Cell Design and Testing. The experimental CID flow cell was constructed in the following way. Two
end plates on the outer sides of the cell were used to sandwich meshed current collectors, NiHCF
electrodes, and an AEM. In this work we used a commercial AMX Neosepta membrane with
thickness of 140 m. Meshed current collector also served as a spacer material to allow feed water
to flow along the NiHCF electrode surface. The cell contained an electrode of Mel=0.99 g (of which
~0.8 g is NiHCF) on each side, and the membrane area available for ion transfer was 36 cm2.
Desalination experiments were performed at a constant current of 10 mA, thus 2.77 A/m2, and a
water flow rate of 10 mL/min directed into each channel. Constant current was applied until a certain
value of charge transfer was reached (total number of Coulombs per gram of material transferred,
as shown in Figs. 3 and 4). Subsequently, the current direction was reversed. After 2 or 3 similar
cycles with the same total charge transfer, the charge transfer was reduced stepwise. Because all
experiments in Fig. 3 were done at the same current density (which can only switch sign), the cycles
become shorter in duration.
Acknowledgments
KCS acknowledges support from the Department of Mechanical Science and Engineering and the
College of Engineering at the University of Illinois at Urbana-Champaign. AS acknowledges support
from the Department of Materials Science and Engineering at the University of Illinois at Urbana-
Champaign. Part of this work was performed in the cooperation framework of Wetsus, European
Centre of Excellence for Sustainable Water Technology (www.wetsus.eu). Wetsus is co-funded by
the Dutch Ministry of Economic Affairs and Ministry of Infrastructure and Environment, the Province
of Fryslân, and the Northern Netherlands Provinces.
References
[1] M.E. Suss, S. Porada, X. Sun, P.M. Biesheuvel, J. Yoon, V. Presser, Water desalination via capacitive
deionization: what is it and what can we expect from it?, Energy Environ. Sci. 8 (2015) 22962319.
doi:10.1039/C5EE00519A.
[2] J. Lee, S. Kim, C. Kim, J. Yoon, Hybrid capacitive deionization to enhance the desalination performance
of capacitive techniques, Energy {&} Environ. Sci. 7 (2014) 36833689. doi:10.1039/C4EE02378A.
[3] M. Pasta, C.D. Wessells, Y. Cui, F. La Mantia, A Desalination Battery, Nano Lett. 12 (2012) 839843.
doi:10.1021/nl203889e.
[4] K.C. Smith, R. Dmello, Na-Ion Desalination (NID) Enabled by Na-Blocking Membranes and Symmetric
Na-Intercalation: Porous-Electrode Modeling, J. Electrochem. Soc. 163 (2016) A530A539.
doi:10.1149/2.0761603jes.
[5] P. Srimuk, F. Kaasik, B. Krüner, A. Tolosa, S. Fleischmann, N. Jäckel, M.C. Tekeli, M. Aslan, M.E. Suss, V.
Presser, MXene as a novel intercalation-type pseudocapacitive cathode and anode for capacitive
deionization, J. Mater. Chem. A. 4 (2016) 1826518271. doi:10.1039/C6TA07833H.
[6] F. Xing, T. Li, J. Li, H. Zhu, N. Wang, X. Cao, Chemically exfoliated MoS2 for capacitive deionization of
saline water, Nano Energy. 31 (2017) 590595. doi:10.1016/j.nanoen.2016.12.012.
[7] G.J. Doornbusch, J.E. Dykstra, P.M. Biesheuvel, M.E. Suss, Fluidized bed electrodes with high carbon
15
loading for water desalination by capacitive deionization, J. Mater. Chem. A. 4 (2016) 36423647.
doi:10.1039/C5TA10316A.
[8] J.E. Dykstra, R. Zhao, P.M. Biesheuvel, A. van der Wal, Resistance identification and rational process
design in Capacitive Deionization, Water Res. 88 (2016) 358370. doi:10.1016/j.watres.2015.10.006.
[9] Y. Qu, P.G. Campbell, L. Gu, J.M. Knipe, E. Dzenitis, J.G. Santiago, M. Stadermann, Energy
consumption analysis of constant voltage and constant current operations in capacitive deionization,
Desalination. 400 (2016) 1824. doi:10.1016/j.desal.2016.09.014.
[10] T. Kim, J.E. Dykstra, S. Porada, A. van der Wal, J. Yoon, P.M. Biesheuvel, Enhanced charge efficiency
and reduced energy use in capacitive deionization by increasing the discharge voltage, J. Colloid
Interface Sci. 446 (2015) 317326. doi:10.1016/j.jcis.2014.08.041.
[11] C. Prehal, D. Weingarth, E. Perre, R.T. Lechner, H. Amenitsch, O. Paris, V. Presser, Tracking the
structural arrangement of ions in carbon supercapacitor nanopores using in situ small-angle X-ray
scattering, Energy Environ. Sci. 8 (2015) 17251735. doi:10.1039/C5EE00488H.
[12] S. Kim, J. Lee, C. Kim, J. Yoon, Na2FeP2O7 as a Novel Material for Hybrid Capacitive Deionization,
Electrochim. Acta. 203 (2016) 265271. doi:10.1016/j.electacta.2016.04.056.
[13] K.C. Smith, R. Dmello, Desalination Devices, 62/427,502, 2016.
[14] H. Kim, J. Hong, K.-Y. Park, H. Kim, S.-W. Kim, K. Kang, Aqueous Rechargeable Li and Na Ion Batteries,
Chem. Rev. 114 (2014) 1178811827. doi:10.1021/cr500232y.
[15] J.F. Whitacre, A. Tevar, S. Sharma, Na4Mn9O18 as a positive electrode material for an aqueous
electrolyte sodium-ion energy storage device, Electrochem. Commun. 12 (2010) 463466.
doi:10.1016/j.elecom.2010.01.020.
[16] Z. Li, D. Young, K. Xiang, W.C. Carter, Y.-M. Chiang, Towards High Power High Energy Aqueous
Sodium-Ion Batteries: The NaTi2(PO4)3/Na0.44MnO2 System, Adv. Energy Mater. 3 (2013) 290294.
doi:10.1002/aenm.201200598.
[17] J.F. Whitacre, S. Shanbhag, A. Mohamed, A. Polonsky, K. Carlisle, J. Gulakowski, W. Wu, C. Smith, L.
Cooney, D. Blackwood, J.C. Dandrea, C. Truchot, A Polyionic, Large-Format Energy Storage Device
Using an Aqueous Electrolyte and Thick-Format Composite NaTi2(PO4)3/Activated Carbon Negative
Electrodes, Energy Technol. 3 (2015) 2031. doi:10.1002/ente.201402127.
[18] Z. Li, D.B. Ravnsbæk, K. Xiang, Y.-M. Chiang, Na3Ti2(PO4)3 as a sodium-bearing anode for
rechargeable aqueous sodium-ion batteries, Electrochem. Commun. 44 (2014) 1215.
doi:10.1016/j.elecom.2014.04.003.
[19] X. Wu, Y. Cao, X. Ai, J. Qian, H. Yang, A low-cost and environmentally benign aqueous rechargeable
sodium-ion battery based on NaTi2(PO4)3Na2NiFe(CN)6 intercalation chemistry, Electrochem.
Commun. 31 (2013) 145148. doi:10.1016/j.elecom.2013.03.013.
[20] C.D. Wessells, S. V. Peddada, R.A. Huggins, Y. Cui, Nickel hexacyanoferrate nanoparticle electrodes for
aqueous sodium and potassium ion batteries, Nano Lett. 11 (2011) 54215425.
doi:10.1021/nl203193q.
[21] C.D. Wessells, R.A. Huggins, Y. Cui, Copper hexacyanoferrate battery electrodes with long cycle life
and high power, Nat. Commun. 2 (2011) 550.
[22] X. Wu, M. Sun, Y. Shen, J. Qian, Y. Cao, X. Ai, H. Yang, Energetic Aqueous Rechargeable Sodium-Ion
Battery Based on Na2CuFe (CN) 6NaTi2 (PO4) 3 Intercalation Chemistry, ChemSusChem. 7 (2014)
407411.
[23] X. Wu, Y. Cao, X. Ai, J. Qian, H. Yang, A low-cost and environmentally benign aqueous rechargeable
sodium-ion battery based on NaTi2(PO4)3-Na 2NiFe(CN)6 intercalation chemistry, Electrochem.
Commun. 31 (2013) 145148. doi:10.1016/j.elecom.2013.03.013.
[24] R.Y. Wang, C.D. Wessells, R.A. Huggins, Y. Cui, Highly reversible open framework nanoscale electrodes
for divalent ion batteries, Nano Lett. 13 (2013) 57485752. doi:10.1021/nl403669a.
[25] A.L. Lipson, B. Pan, S.H. Lapidus, C. Liao, J.T. Vaughey, B.J. Ingram, Rechargeable Ca-Ion Batteries: A
New Energy Storage System, Chem. Mater. 27 (2015) 84428447.
doi:10.1021/acs.chemmater.5b04027.
[26] L. Zhang, L. Chen, X. Zhou, Z. Liu, Morphology-Dependent Electrochemical Performance of Zinc
16
Hexacyanoferrate Cathode for Zinc-Ion Battery, Sci. Rep. 5 (2015) 18263. doi:10.1038/srep18263.
[27] K.C. Smith, Theoretical evaluation of electrochemical-cell architectures using cation intercalation
electrodes for desalination, Submitted. (2016).
[28] H.J. Buser, D. Schwarzenbach, W. Petter, A. Ludi, The crystal structure of Prussian Blue:
Fe4[Fe(CN)6]3.xH2O, Inorg. Chem. 16 (1977) 27042710. doi:10.1021/ic50177a008.
[29] Y. Yue, Z. Zhang, A.J. Binder, J. Chen, X. Jin, S.H. Overbury, S. Dai, Hierarchically superstructured
Prussian blue analogues: Spontaneous assembly synthesis and applications as pseudocapacitive
materials, ChemSusChem. 8 (2015) 177183. doi:10.1002/cssc.201402520.
[30] J. Jagiello, J.P. Olivier, 2D-NLDFT adsorption models for carbon slit-shaped pores with surface
energetical heterogeneity and geometrical corrugation, Carbon N. Y. 55 (2013) 7080.
doi:10.1016/j.carbon.2012.12.011.
[31] S. Porada, F. Schipper, M. Aslan, M. Antonietti, V. Presser, T.P. Fellinger, Capacitive Deionization using
Biomass-based Microporous Salt-Templated Heteroatom-Doped Carbons, ChemSusChem. 8 (2015)
1823. doi:10.1002/cssc.201500696.
[32] K.C. Smith, R. Dmello, Comment on: “Na-Ion Desalination (NID) Enabled by Na-Blocking Membranes
and Symmetric Na-Intercalation: Porous-Electrode Modeling” [J. Electrochem. Soc., 163, A530
(2016)], J. Electrochem. Soc. 163 (2016) Y17Y17. doi:10.1149/2.0601610jes.
[33] B. Scroscatti, Lithium Rocking Chair Batteries: An Old Concept?, J. Electrochem. Soc. 139 (1995) 2776
2781.
[34] S. Porada, P.M. Biesheuvel, V. Presser, Comment on Sponge-Templated Preparation of High Surface
Area Graphene with Ultrahigh Capacitive Deionization Performance, Adv. Funct. Mater. 25 (2015)
179181. doi:10.1002/adfm.201401101.
[35] P.M. Biesheuvel, H.V.M. Hamelers, M.E. Suss, Theory of Water Desalination by Porous Electrodes with
Immobile Chemical Charge, Colloids Interface Sci. Comm. 9 (2016) 15.
doi:10.1016/j.colcom.2015.12.001.
[36] X. Gao, S. Porada, A. Omosebi, K. Liu, P.M. Biesheuvel, J. Landon, Complementary surface charge for
enhanced capacitive deionization, Water Res. 92 (2016) 275282. doi:10.1016/j.watres.2016.01.048.
[37] B.E. Conway, Electrochemical Supercapacitors: Scientific Fundamentals and Technological
Applications, 1991. doi:10.1017/CBO9781107415324.004.
[38] K.C. Smith, Simulation of Na-Ion Intercalation Electrodialysis Using Prussian-Blue Analogues,
Submitted. (2016).
[39] K.C. Smith, Theoretical evaluation of electrochemical cell architectures using cation intercalation
electrodes for desalination, Submitted. (2016).
[40] J. Song, L. Wang, Y. Lu, J. Liu, B. Guo, P. Xiao, J.J. Lee, X.Q. Yang, G. Henkelman, J.B. Goodenough,
Removal of interstitial H<inf>2</inf>O in hexacyanometallates for a superior cathode of a sodium-ion
battery, J. Am. Chem. Soc. 137 (2015) 26582664. doi:10.1021/ja512383b.
... 127 In the case of electrodes that selectively capture cations (e.g., NMO, metal phosphates, PBAs), the membrane must be an AEM, which functions as the anion selective surface in the device. 116,127,742 The AEM in this symmetric design partitions the feed into two channels in a way that resembles ED, except the electric field is still cyclic. This mode of operation was mathematically modeled by Singh et al. 88 Recent advances in electrosorption with intercalation materials have provided opportunities for combined theoretical and experimental studies, 86,96,127,634,895 which build on earlier models of ion intercalation in porous electrodes of lithium-ion batteries. ...
... 117,954 Initial studies showed that CDI with activated carbon permits an applied current greater than 1 mA cm −2 , 115 whereas intercalation electrodes limit the applied current to ∼100 μA cm −2 . 115,718, 742 Lee et al. recently proposed a multichannel desalination battery, as shown in Figure 25f, to overcome the mass transfer limitations of traditional desalination batteries. 946 The multichannel system was designed to have two independent channels (one for both electrodes and one for the feed) by placing an IEM at each interface. ...
Article
Full-text available
Agricultural development, extensive industrialization, and rapid growth of the global population have inadvertently been accompanied by environmental pollution. Water pollution is exacerbated by the decreasing ability of traditional treatment methods to comply with tightening environmental standards. This review provides a comprehensive description of the principles and applications of electrochemical methods for water purification, ion separations, and energy conversion. Electrochemical methods have attractive features such as compact size, chemical selectivity, broad applicability, and reduced generation of secondary waste. Perhaps the greatest advantage of electrochemical methods, however, is that they remove contaminants directly from the water, while other technologies extract the water from the contaminants, which enables efficient removal of trace pollutants. The review begins with an overview of conventional electrochemical methods, which drive chemical or physical transformations via Faradaic reactions at electrodes, and proceeds to a detailed examination of the two primary mechanisms by which contaminants are separated in nondestructive electrochemical processes, namely electrokinetics and electrosorption. In these sections, special attention is given to emerging methods, such as shock electrodialysis and Faradaic electrosorption. Given the importance of generating clean, renewable energy, which may sometimes be combined with water purification, the review also discusses inverse methods of electrochemical energy conversion based on reverse electrosorption, electrowetting, and electrokinetic phenomena. The review concludes with a discussion of technology comparisons, remaining challenges, and potential innovations for the field such as process intensification and technoeconomic optimization.
... Traditional carbonaceous electrodes have low desalination capacities (< 15 mg g −1 ) because of the electrical double-layer (EDL) behavior of charge storage limited to the surface area [9,10]. In the quest for higher deionization performance, battery electrodes have been employed in recent research owing to their larger electrochemical capacities, such as sodium-ion battery materials (sodium transition metal oxides, transition metal oxides, sodium super ion conductor materials, and Prussian blue analogues) and chloride battery materials (Ag-based or Bi-based materials, etc.) [11][12][13][14][15][16][17][18][19]. Although quite a few studies have initiated the use of battery materials applied in deionization systems, these materials are still difficult to scale up for real applications due to the low energy density and instability of the battery materials in aqueous-based electrolytes [20,21]. ...
Article
Full-text available
First report of a lithium-ion battery cathode as a deionization electrode for desalination.A novel approach to suppress manganese dissolution by exposing the (111) facet is proposed.Excellent desalination performance by the LiMn2O4/C cathode. The material achieves an ultrahigh desalination capacity of 117.3 mg g−1 at 1.0 V and a longer cycle life (200 cycles without capacity decay) with minor manganese dissolution during the cycling test in 10 mM aqueous LiCl solution. First report of a lithium-ion battery cathode as a deionization electrode for desalination. A novel approach to suppress manganese dissolution by exposing the (111) facet is proposed. Excellent desalination performance by the LiMn2O4/C cathode. The material achieves an ultrahigh desalination capacity of 117.3 mg g−1 at 1.0 V and a longer cycle life (200 cycles without capacity decay) with minor manganese dissolution during the cycling test in 10 mM aqueous LiCl solution. Battery materials as emerging capacitive deionization electrodes for desalination have better salt removal capacities than traditional carbon-based materials. LiMn2O4, a widely used cathode material, is difficult to utilize as a deionization electrode due to its structural instability upon cycling and Mn dissolution in aqueous-based electrolytes. Herein, a facile and low-cost ball-milling routine was proposed to prepare a LiMn2O4 material with highly exposed (111) facets. The prepared electrode exhibited relatively low dissolution of Mn during cycling, which shows its long cycle stability. In the hybrid capacitive deionization system, the LiMn2O4/C electrode delivered a high desalination capacity of 117.3 mg g−1 without obvious capacity decay at a voltage of 1.0 V with a 20 mM initial salt concentration. In addition, the exposed (111) facets significantly alleviated Mn ion dissolution, which also enhanced the structural steadiness.
... These family materials facilitate the uptake of ions with their unique structure, resulting in improving the ion adsorption. For the CDI desalination application, those materials also show higher desalination performance compared to the conventional electrode materials [141][142][143][144]. Recently, there have been some efforts to apply intercalation materials in FCDI desalination. ...
Article
Flow-electrode capacitive deionization (FCDI) is an emerging desalination technology that overcomes the drawbacks of traditional capacitive deionization (CDI) by providing larger salt removal capacity and continuous desalination operation. Various approaches to the choice of cell configurations and electrode materials allow FCDI to increase the salt removal performance and extend its potential applications. In particular, Faradaic electrode materials for FCDI have recently emerged due to their various advantages over capacitive material, which include higher salt removal rate, lower energy consumption, and the ability to selectively remove specific ions. In this review, we summarize the background technology and mechanism of the FCDI system with an emphasis on the development of electrode materials, including capacitive and redox active electrodes, as well as their role in the various applications for FCDI and its future direction.
Article
Full-text available
Capacitive deionization (CDI) is an emerging eco-friendly desalination technology with mild operation conditions. However, the energy consumption of CDI has not yet been comprehensively summarized, which is closely related to the economic cost. Hence, this study aims to review the energy consumption performances and mechanisms in the literature of CDI, and to reveal a future direction for optimizing the consumed energy. The energy consumption of CDI could be influenced by a variety of internal and external factors. Ion-exchange membrane incorporation, flow-by configuration, constant current charging mode, lower electric field intensity and flowrate, electrode material with a semi-selective surface or high wettability, and redox electrolyte are the preferred elements for low energy consumption. In addition, the consumed energy in CDI could be reduced to be even lower by energy regeneration. By combining the favorable factors, the optimization of energy consumption (down to 0.0089 Wh·gNaCl−1) could be achieved. As redox flow desalination has the benefits of a high energy efficiency and long lifespan (~20,000 cycles), together with the incorporation of energy recovery (over 80%), a robust future tendency of energy-efficient CDI desalination is expected.
Article
Cesium (¹³⁷Cs) is one of the representative radionuclides which must be eliminated from nuclear waste. Here, we designed a zinc hexacyanoferrate (ZnHCF) composite using ZIF-8 derived carbon (ZDC) and utilized it as an electrode to selectively remove cesium ions. Specifically, we focused on how the ZIF-8 pyrolysis temperature affected the composite formation and non-radioactive cesium removal performance. With an optimized temperature of 700 °C, a highly conductive and uniform composite with well-distributed ZnHCF was produced, and it exhibited a large cesium uptake capacity (204.9 mg g⁻¹). The composite electrode also retained high selectivity in Na-rich environments (Na/Cs = 1330, Kd = 1.04 × 10⁵), K-rich environments (K/Cs = 133, Kd = 7.20 × 10⁴), and groundwater conditions (95 % removal). Moreover, the reversible uptake and release of cesium over 5 cycles were feasible in our system without any chemical additives, which can be reached 100 % regeneration at the fourth cycle. Using in-depth characterizations including XRD and XPS, we investigated the faradaic behavior, phase transition, and structural stability of the ZnHCF-ZDC composite over 5 cycles. This study formed a composite electrosorbent with a ZIF-derived carbon support and applied it to cesium removal for the first time. This electro-mediated cesium removal process is expected to serve as a green technology for the future nuclear industry and environmental remediation.
Article
A highly efficient water purification device is designed by a combination of electrochemical deionization and microfluidic techniques. The device is constructed with seven layers, in which MnO2 and polypyrrole are employed as positive and negative electrodes, respectively. The advantages of high sensitivity, large surface-to-volume ratio, and short ion diffusion path provided by such a microfluidic device dramatically increase the salt removal capacity (SRC) and salt removal rate (SRR) to 132 mg g⁻¹ and 30 mg g⁻¹ min⁻¹ in the desalination test using a 600 mM NaCl solution. The salt removal percentage in a single-pass test of this microfluidic device using an 8 mM NaCl solution can reach 88 %, 33 times higher than 2.6 % obtained from a macro-scale experiment setup. The SRC remains over 75 % after a 180-cycle stability test. The “memory effect” provides both deionization and concentration ability without using a membrane. Over 160 mg g⁻¹ of NaCl can be transferred from one solution to another under a 120-min deionization/concentration test. This innovative device in a semi-automatic system has been tested in real salinized underground water and seawater with the SRC of 199.4 mg g⁻¹ and 456.6 mg g⁻¹ for removing several ions such as Na⁺, K⁺, Mg²⁺, and Ca²⁺.
Article
In this study, the electrosorption selectivity of porous activated carbon (AC) and nickel hexacyanoferrate (NiHCF), which represent two working mechanisms of capacitive electrosorption and redox intercalation, was investigated to separate cations in capacitive deionization (CDI). The cyclic voltammetry diagrams of AC showed the rectangular shape of double-layer charging, while that of NiHCF showed separated peaks associated with redox reactions. The specific capacitance of NiHCF was 143.6 F/g in 1 M NaCl, which was almost two times higher than that of AC. Cation selectivity experiments were conducted in single-pass CDI for a multi-cation solution. The electrosorption preference of the AC cathode was determined by a counterbalance between the ionic charge and hydrated size, reflecting the selectivity coefficient of different cations over Na⁺ in the range of 0.86–2.63. For the NiHCF cathode, the cation selectivity was mainly dominated by the hydrated radius and redox activity. Notably, high selectivities of K⁺/Na⁺ ≈ 3.57, Na⁺/Ca²⁺ ≈ 9.97, and Na⁺/Mg²⁺ ≈ 18.92 were obtained. A significant improvement in the electrosorption capacity and monovalent ion selectivity can be achieved by utilizing the NiHCF electrode. The study demonstrates the fundamental aspects and promising opportunities of CDI in regard to ion selectivity.
Article
NASICON (sodium superionic conductor) materials are promising host compounds for the reversible capture of Na+ ions, finding prior application in batteries as solid-state electrolytes and cathodes/anodes. Given their affinity for Na+ ions, these materials can be used in Faradaic deionization (FDI) for the selective removal of sodium over other competing ions. Here, we investigate the selective removal of sodium over other alkali and alkaline-earth metal cations from aqueous electrolytes when using a NASICON-based mixed Ti-V phase as an intercalation electrode, namely, sodium titanium vanadium phosphate (NTVP). Galvanostatic cycling experiments in three-electrode cells with electrolytes containing Na+, K+, Mg2+, Ca2+, and Li+ reveal that only Na+ and Li+ can intercalate into the NTVP crystal structure, while other cations show capacitive response, leading to a material-intrinsic selectivity factor of 56 for Na+ over K+, Mg2+, and Ca2+. Furthermore, electrochemical titration experiments together with modeling show that an intercalation mechanism with a limited miscibility gap for Na+ in NTVP mitigates the state-of-charge gradients to which phase-separating intercalation electrodes are prone when operated under electrolyte flow. NTVP electrodes are then incorporated into an FDI cell with automated fluid recirculation to demonstrate up to 94% removal of sodium in streams with competing alkali/alkaline-earth cations with 10-fold higher concentration, showing process selectivity factors of 3-6 for Na+ over cations other than Li+. Decreasing the current density can improve selectivity up to 25% and reduce energy consumption by as much as ∼50%, depending on the competing ion. The results also indicate the utility of NTVP for selective lithium recovery.
Article
Cation intercalation desalination (CID) has gained great popularity in the field of water desalination because of its excellent desalination performance. Due to lack of understanding of the relationship between electrode microstructures and reactive adsorption processes in CIDs, the further development of high-performance CID electrodes has been hindered. To this end, the influences of electrode microstructures on the desalination performance of CID cells were investigated from the pore-scale level in this work. The three-dimensional microstructures of porous electrodes are first reconstructed by an improved random generation method. Based on the reconstructed porous electrodes, the flow, mass transfer and intercalation reaction processes under the constant current are simulated using the lattice Boltzmann method. The effects of applied current density, porosity and size of active particles on the change of liquid-phase sodium concentration and Na-intercalated degree in the solid phase are evaluated. The simulation results show that increasing the current density can accelerate the ion intercalation and desalination rates. During the desalination, the ion concentration is unevenly distributed in the pores of porous electrodes, and the intercalation degrees are different along the electrode thickness direction and in the particle radial direction. The increase of porosity can alleviate the concentration polarization of liquid-phase sodium ion and reduce the difference of particle intercalation degree between the front- and back-sides of the electrode. In addition, the increase of porosity can accelerate the desalination process, but cannot improve the total salt removal. Reducing particle size can shorten the time for particles to reach the sodium-rich state, but it can aggravate the polarization of sodium ion concentration in the liquid phase at both ends of the electrode. This work reveals the ion intercalation behavior in the CID and its relationship with the electrode microstructures, and may provide useful information for the design and research of CID.
Article
Full-text available
The demand for fresh water has been increasing, caused by the growing population and industrialization throughout the world. In this study, we report a capacitive-based desalination system using Prussian blue materials in a rocking chair desalination battery, which is composed of sodium nickel hexacyanoferrate (NaNiHCF) and sodium iron HCF (NaFeHCF) electrodes. In this system, ions are removed not only by charging steps but also by discharging steps, and it is possible to treat actual seawater with this system because the Prussian blue material has a high charge capacity with a reversible reaction of alkaline cations. Here, we demonstrate a rocking chair desalination battery to desalt seawater, and the results show that this system has a high desalination capacity (59.9 mg/g) with efficient energy consumption (0.34 Wh/L for 40% Na ion removal efficiency).
Article
Full-text available
Water scarcity is a dilemma facing much of the global population. Cation intercalation desalination (CID) cells, which use intercalation host compounds (IHCs) in combination with ion-exchange membranes (IEMs), could aid in addressing this challenge by treating saline water sources. Originally, the performance of such cells was predicted utilizing continuous flow of saline water through porous IHC electrodes. Here, we use two-dimensional porous-electrode theory with concentrated solution transport to evaluate the performance of various cell architectures where flow occurs through open flow channels (OFCs) when two IHC electrodes comprised of nickel hexacyanoferrate (NiHCF) are used to store Na+ ions. We show that, when two OFCs are used, cation exchange membranes (CEMs) are adjoined at flow-channel/electrode interfaces, and an anion exchange membrane (AEM) is arranged between flow channels, salt removal increases relative to the original design with flow-through (FT) electrodes. The IEM stacking sequence within such a membrane flow-by (MFB) cell is the fundamental repeat unit for electrodialysis (ED) stacks using many IEMs (CEM/AEM/.../CEM/AEM/CEM) with many diluate streams. Accordingly, we simulate the performance of such ED stacks using NiHCF IHCs, and we predict that salt adsorption capacity (per unit NiHCF mass) is amplified by twenty-fold relative to MFB and FT cells, while simultaneously decreasing 0.7 M NaCl feed water to 0.2-0.3 M within diluate streams. The generality of these findings is further supported by simulations using Na0.44MnO2 IHC instead of NiHCF. Thus, we propose the use of cation IHCs as alternatives to the gas-evolution reactions used in conventional ED.
Article
Full-text available
In this proof-of-concept study, we introduce and demonstrate MXene as a novel type of intercalation electrode for desalination via capacitive deionization (CDI). Traditional CDI cells employ nanoporous carbon electrodes with significant pore volume to achieve a large desalination capacity via ion electrosorption. By contrast, MXene stores charge by ion intercalation between the sheets of its two-dimensional nanolamellar structure. By this virtue, it behaves as an ideal pseudocapacitor, that is, showing capacitive electric response while intercalating both anions and cations. We synthesized Ti3C2-MXene by the conventional process of etching ternary titanium aluminum carbide i.e., the MAX phase (Ti3AlC2) with hydrofluoric acid. The MXene material was cast directly onto the porous separator of the CDI cell without added binder, and exhibited very stable performance over 30 CDI cycles with an average salt adsorption capacity of 13 ± 2 mg g⁻¹.
Article
New electrochemical technologies that use capacitive or battery electrodes are being developed to minimize energy requirements for desalinating brackish waters. When a pair of electrodes is charged in capacitive deionization (CDI) systems, cations bind to the cathode and anions bind to the anode, but high applied voltages (>1.2 V) result in parasitic reactions and irreversible electrode oxidation. In the battery electrode deionization (BDI) system developed here, two identical copper hexacyanoferrate (CuHCF) battery electrodes were used that release and bind cations, with anion separation occurring via an anion exchange membrane. The system used an applied voltage of 0.6 V, which avoided parasitic reactions, achieved high electrode desalination capacities (up to 100 mg-NaCl/g-electrode, 50 mM NaCl influent), and consumed less energy than CDI. Simultaneous production of desalinated and concentrated solutions in two channels avoided a two-cycle approach needed for CDI. Stacking additional membranes between CuHCF electrodes (up to three anion and two cation exchange membranes) reduced energy consumption to only 0.02 kWh/m³ (approximately an order of magnitude lower than values reported for CDI), for an influent desalination similar to CDI (25 mM decreased to 17 mM). These results show that BDI could be effective as a very low energy method for brackish water desalination.
Article
Desalination is a sustainable technology that removes sodium and chloride ions from seawater. Herein, we demonstrate a faradic mechanism to promote the capacity of capacitive deionization in high concertation salty water via an electrochemical deionization device. In this system, ions removal is achieved by the faradic mechanism via a constant current operation mode, which is improved based on the constant voltage operation mode used in the conventional CDI operation. Benefiting from the high capacity and excellent rate performance of Prussian blue as an active electrochemical reaction material, the designed unit has revealed a superior removal capacity with ultrafast ion removal rate. A high removal capacity of 101.7 mg g-1 has been obtained with proper flow rate and current density. To further improve the performance of the EDI, a reduced graphene oxide with nanopores and Prussian blue composite has been synthesized. The PB@NPG have demonstrated a high salt removal capacity of 120.0 mg g-1 at 1 C with an energy consumption of 6.76 kT per ion removed, which is much lower than most CDI methods. A particularly high rate performance of 0.5430 mg g-1 s-1 has been achieved at 40 C. The faradic mechanism promoted EDI has provided a new insight into the design and selection of host materials for highly salty water desalination.
Article
Materials that can selectively store Na and Cl ions in the bulk of their structures and release these ions with good cycle stability can enable the construction of a high capacity, rechargeable desalination cell for use in seawater desalination. In this study, the ability of a nanocrystalline Bi foam electrode to serve as an efficient and high capacity Cl-storage electrode using its conversion to BiOCl was investigated. When Bi as a Cl-storage electrode was coupled with NaTi2(PO4)3 as a Na-storage electrode, a new type of rechargeable desalination cell, which is charged during desalination and discharged during salination, was constructed. The resulting Bi-NaTi2(PO4)3 cell was tested under various salination and desalination conditions to investigate advantages and potential limitations of using Bi as a Cl-storage electrode. Slow Cl– release kinetics of BiOCl in neutral conditions and an imbalance in Cl and Na storage (i.e., Cl storage requires three electrons/Cl, while Na storage requires one electron/Na) were identified as possible drawbacks, but strategies to address these issues were developed. On the basis of these investigations, optimum desalination and salination conditions were identified where the Bi/NaTi2(PO4)3 cell achieved a desalination/salination cycle at ±1 mA cm–2 with a net potential input of only 0.20 V. The kinetics of Cl– release from BiOCl was significantly improved by the use of an acidic solution, and therefore, a divided cell was used for the salination process. We believe that with further optimizations the Bi/BiOCl electrode will enable efficient and practical desalination applications.
Article
Electrochemical selective ion separation via capacitive deionization, for example, separation of lithium resource from brine, using lithium ion batteries is proposed and demonstrated to have the potential for separating specific ions selectively from a solution containing diverse ions. This separation method is of great industrial concern because of applicability in various fields such as deionization, water softening, purification, heavy metal removal, and resource recovery. Nevertheless, besides the selectivity of materials for lithium ion batteries toward Li⁺, there is very little investigation on the selectivity of the materials for sodium ion batteries toward Na⁺. Here, the electrochemical selectivity of sodium manganese oxide (Na0.44MnO2), one of the most widely used material in sodium ion batteries, for Na⁺ and other cations (K⁺, Mg²⁺, and Ca²⁺) is investigated. Selective Na⁺ separation using the system consisting of Na0.44MnO2 and a Ag/AgCl electrode is successfully demonstrated from a solution containing diverse cations (K⁺, Mg²⁺, and Ca²⁺) via a two-step process that involves a capturing step (charging process) and a releasing step (discharging process). The results showed that Na0.44-xMnO2 has over 13 times higher selectivity for Na⁺ than for K⁺ and 6–8 times higher selectivity for Na⁺ than for Mg²⁺ and Ca²⁺ in the electrolyte containing equal concentrations of the respective ions. Additionally, as a practical demonstration, Na⁺ was successfully separated from an industrial raw material used for pure KOH production (estimated ratio of Na⁺:K⁺ = 1:200).
Article
Seawater desalination is a leading way for tackling the global freshwater shortage challenge. However, existing desalination technologies have their own limits, including high energy consumption or low ion removal capacity, which is not sufficient for desalting high-concentration saline water with low cost. Herein, we present a concept of dual-ions electrochemical deionization technology, which consists of BiOCl for chloride ion Faradaic electrode on the negative side, sodium manganese oxide (Na0.44MnO2) as sodium ion Faradaic electrode on the positive side. It utilizes a redox reaction to individually absorb chloride ions and sodium ions concurrently. Under positive electric current operations, the two ions are released to flow NaCl water electrolyte. Upon switch to negative electric current, the chloride ions are extracted into the negative electrode from flowing NaCl solution while sodium ions are electrochemically captured into the positive electrode. The novel dual-ions Faradaic deionization delivers a stable and reversible salt absorption/desorption capacity of 68.5 mg g-1 when operated at a current density 100 mA g-1, which makes over twice salt absorption of the previous reported best performance (31.2 mg g-1) obtained by a hybrid capacitive deionization system. The electric charge efficiency is up to 0.977 during salt desorption process and 0.958 during absorption process. Owing to ion intercalation process in the salt removal process, energy will be recovered during discharge process, therefore, the current system is called “desalination generator”. Our research will supply a new method for desalination flow through system.
Article
Reversible mixed ion intercalation in non-selective host structures has promising applications in desalination, in mixed-ion batteries, in waste-water treatment and in lithium recovery by electrochemical ion pumping. One class of host compound that possesses many of the requirements needed for such applications (cost effectiveness, fast ion kinetics and stability in aqueous medium) includes the Prussian blue derivatives. Here, we study the fundamental process of intercalation of multiple cations at the thermodynamic level, by means of galvanostatic cycling. We focus on nickel hexacyanoferrate because of its stability and low potential for electrochemical process compared to other hexacyanoferrates. Various cations can be intercalated; large cations (K+ and NH4+) are intercalated at higher potentials than smaller cations (Na+). When mixtures of cations are present in solution, the potential profile is not qualitatively altered with respect to the single-salt solutions, but the potentials of (de-)intercalation is shifted; we introduce a simple thermodynamic model able to predict the potential and distribution at which the intercalation takes place.
Article
Capacitive deionization (CDI) is a promising technology for removal of ions from saline water upon applying a voltage between two electrodes. In this work, chemically exfoliated MoS2 (ce-MoS2) has been explored as electrode material for CDI. The ce-MoS2 nanosheets demonstrate a good cycling stability, high ion quality removal capacity of 8.81 mg/g, and ion volume removal capacity of 16.51 mg/cm³ at 1.2 V applied voltage in 400 mM NaCl solution. Comparison with that of carbon materials and bulk MoS2, the ce-MoS2 nanosheets remain good removal capacity even at a low concentration of 50 mM salt. These results suggest that ce-MoS2 can be a promising candidate for CDI. The sound performance of ce-MoS2 in CDI can be attributed to the unique two-dimensional (2D) thin sheet structure of 1 T phase. The large layer-to-layer space and sound electric conductivity help ion intercalation and ion transportation, and the abundant negative charges on the surface of ce-MoS2 sheets enhance the electrostatic attraction of ion.