Journal of Water Process Engineering 26 (2018) 257–263

Contents lists available at ScienceDirect

Journal of Water Process Engineering

journal homepage: www.elsevier.com/locate/jwpe

A comparative study on electrosorptive rates of metal ions in capacitive T

deionization
⁎
Yanting Lia, Ty C. Stewartb, Hao L. Tangb,
a
School of Environment, Harbin Institute of Technology, Harbin, Heilongjiang 150090, China
b
Department of Chemistry, Indiana University of Pennsylvania, Indiana, PA, 15705, USA

A R T I C LE I N FO A B S T R A C T

Keywords: A comparative study on electrosorptive rates was conducted to explore the affinity preference of Na+, K+,
Electrosorption Mg2+, Ca2+, Cd2+, Pb2+ and Cu2+ to activated carbon cloth electrodes of capacitive deionization (CDI) in
Heavy metal mono-component solutions. Results showed the electrosorptive kinetics of these metal ions followed the
Ionic hydrated radius Lagergren equation. For alkali and alkali-earth metal ions, the order of electrosorptive rates followed
Kinetics
K+ > Na+ > Ca2+ > Mg2+, which revealed an inversely proportional relationship with their ionic hydrated
Capacitive deionization
radii. The relationship between electrosorptive rates and ionic hydrated radii, however, did not hold for heavy
metal ions, and the order was Cu2+ > Pb2+ > Cd2+. The abnormal electrosorptive behaviors of Cu2+ and Pb2+
could be attributed to electrodeposition and formation of OH− complexes during electrosorption. Besides, Cu2+
and Pb2+ showed significantly low desorption efficiencies during the electrode regeneration processes, in-
dicating serious electrode fouling during the electrosorptive process. This work provides fundamental aspects on
individual removal of metal ions and the results can be beneficially used as a reference for treatment and
recovery of metal ions by CDI technology.

1. Introduction metal laden waters [4]. It has a simple structure using juxtaposed pairs
of charged electrodes and is highly energy-efficient using a low voltage
Metal pollution of fresh water resources by industrial discharges has (0.6–1.2 VDC) to attract ions from water by electrostatic force. After the
received significant worldwide concerns for the last few decades. Waste adsorption capacity is reached, the ions are desorbed into brine water
streams containing one or more toxic metals are often encountered in by removing the electric field and the regenerated electrodes are ready
metallurgical, tannery, chemical pharmaceutical, mining, battery to use for the next cycle [5]. Therefore, the adsorption cycles generate
manufacturing, fertilizers, and many other industries [1]. Since the an effluent with reduced concentrations of anions [6–9] and metal ca-
substantial build-up of toxic metals in food chains poses a high health tions [10–14] such as sodium (Na+), potassium (K+), calcium (Ca2+),
risk to human consumers, it is necessary to remove these metals from magnesium (Mg2+), and heavy metals such as lead (Pb2+), cadmium
the waste streams before discharging to the fresh water environment. (Cd2+) and copper (Cu2+), resulting in water deionization and pur-
There are many methods for metal removal, such as chemical pre- ification, while the desorption cycles produce a brine stream with
cipitation, membrane filtration, ion exchange, adsorption, etc. concentrated such ions, which offers easier recovery of these metals
However, some disadvantages of these methods make them unsuitable with other terminal processes (e.g. evaporation).
for all scenarios of industrial wastewater treatment [2]. For example, Although many studies have indicated that metal ions can be re-
conventional chemical precipitation is known to have poor perfor- moved by CDI and a variety of electrode materials have been proposed
mance in low strength heavy metal removal and its operation is de- to enhance CDI performances [15–17], the affinity preference of these
pendent on many factors such as pH and chemical doses. It is also costly metal ions to the electrodes remains obscure and literatures on CDI
and generates a lot of sludge that poses environmental disposal pro- research using multi-ionic solutions sometimes presents contradictory
blems [3]. In recent years, capacitive deionization (CDI) is studied as a results [18]. The first study reporting this obscurity was by Gabelich
cost-effective technology with no sludge production and could compete et al. [11], who observed the adsorption in molarity followed a trend of
with other commonly used alternatives (e.g., ion exchange, reverse Na+ > Ca2+ > Mg2+ > K+ in surface water with a low total organic
osmosis, electrodialysis, etc) for the treatment of various strengths of carbon (TOC) concentration while the trend changed to

⁎
Corresponding author.
E-mail address: [email protected] (H.L. Tang).

https://doi.org/10.1016/j.jwpe.2018.10.021
Received 12 April 2018; Received in revised form 9 October 2018; Accepted 29 October 2018
2214-7144/ © 2018 Elsevier Ltd. All rights reserved.
Y. Li et al. Journal of Water Process Engineering 26 (2018) 257–263

Na+ > Mg2+ > Ca2+ > K+ in water with a high TOC. Results on 2.2. Metal ions and pretreatment of CDI cells
produced water treatment by Xu et al. [12] supported the former trend.
The discrepancy was thought to be attributable to different water types Seven analytical grade salts (NaCl, KCl, CaCl2, MgCl2, PbCl2, CdCl2,
with various initial concentrations of metal ions in a competitive multi- and CuCl2) purchased from Sinopharm Chemical Reagent Co.
ionic environment. The inconsistent uses of units for description of (Shanghai, China) were used to make stock mono-component solutions
adsorption (e.g., milligram salt per gram electrode (mg/g), millimole with distilled water. Each salt solution was made at an initial con-
ion per gram electrode (mmol/g), removal percentage (%), etc) further centration of 7.5 mM by direct dilution of the corresponding stock so-
added to this obscurity. For example, Li et al. [13] reported an order of lution. Prior to the electrosorptive experiment on each salt, a CDI cell
Fe3+ (0.62 mg FeCl3/g) > Ca2+ (0.55 mg CaCl2/g) > Mg2+ (0.52 mg was washed with distilled water to remove any impurities and was then
MgCl2/g) > Na+ (0.45 mg NaCl/g) in mono-component solutions while immersed in the target salt solution for at least 2 h while stirring. After
the order became Na+ > Mg2+ > Ca2+ > Fe3+ in unit of mmol/g. physical adsorption of the salt was completed, the cell was transferred
Besides, removal of certain heavy metals by CDI was complicated, and to a new target salt solution where the electrosorptive experiment was
an observed trend of Cr2+ > Pb2+ > Cd2+ in percentage removal by conducted. The pre-treatment procedures of each CDI cell reduced the
Huang et al. [14] was under the influence of physical adsorption, which physical adsorption influence of the target metal ion and all electro-
introduced uncertainties to the quantification of the sole electrosorp- sorptive experiments for different metal ions started with the same
tion amount. Because the existing results reporting the electrosorption concentration at 7.5 mM.
of metal ions in multi-ionic conditions are limited due to incomparable
experimental details from one research to another, there is a need to 2.3. The electrosorptive experiments
explore the affinity of different metal ions to CDI electrodes in mono-
component solutions with comparable settings. The electrosorptive experiments were carried out at room tem-
In this research, mono-component solutions of metal chloride (NaCl, perature in batch series depicted in Fig. 1. The system consisted of a
KCl, CaCl2, MgCl2, PbCl2, CdCl2, and CuCl2) were used for a series of beaker with 85 mL single-electrolyte solution, a CDI cell with virgin
batch CDI treatment. Electrosorption was started with the same initial ACC electrodes, a conductivity meter, a potentiostat as external power
mole concentrations of these metal ions in a non-competitive environ- supply, and a magnetic stirring plate with a stirrer for mixing. A voltage
ment using CDI apparatuses with virgin electrodes. The influence of of 1.2 V was consistently applied to the cell until desorption by short-
physical adsorption was reduced by equilibrating the CDI system with circuiting at equilibrium. Solution pH values before and after the
target salt solutions prior to each electrosorptive experiment. The ob- electrosorptive experiments were recorded. The pH values of the NaCl,
jective of this study was to obtain the electrosorptive rates of different KCl, CaCl2, MgCl2, and CdCl2 solutions were fluctuating around a base
metal ions so that the relative affinity of these metal ions to CDI elec- level close to neutral, while the starting pH values of the PbCl2 and
trodes could be compared. The study aids in gaining a better under- CuCl2 were 3.26 and 5.19, respectively, and a complete electrosorp-
standing of metal removal during the CDI process. tion/desorption cycle caused pH fluctuations around the two starting
levels. Conductivity data was continuously monitored at a 30-second
interval, and was converted to mole concentrations of the salt based on
2. Experimental calibration curves. The electrosorptive amount of metal ion per unit
mass of ACC (q in mmol/g) at any time was calculated as
2.1. Assembly of CDI cells
C0 − Ct
q= V
m (1)
Multiple CDI cells measured as 4 cm × 4 cm were assembled with
layers of titanium wire mesh (the current collector), 0.55 g virgin ac- where C0 (mM) and Ct (mM) are the concentrations of the metal ion at
tivated carbon cloth (ACC, the electrode), and 2-mm-thick rubber slices time zero and time t during the CDI treatment, respectively, m (g) is the
(the separator to prevent short-circuiting). The layers in the order mass of ACC, and V (L) is the volume of the solution.
shown in Fig. 1 were firmly pressed together with a cable tie. The The concentration of the metal ion after desorption (Cd) was also
electrode material was commercially available ACC from Chemviron measured, and the desorption efficiency (D) [20] was calculated as
UK (model no. Zorflex FM10), which had a specific surface area of
Cd − Ce
1000 m2/g and a thickness of 0.5 mm. The pore size distribution and D= × 100%
C0 − Ce (2)
surface chemistry of the ACC were described elsewhere [19]. The ACC
was cut into multiple 4 cm × 4 cm square-shaped pieces weighing ap- where Ce and Cd are the equilibrium concentrations (mM) of the metal
proximately 0.55 g for the CDI cells. ion after CDI electrosorption and desorption, respectively.
After one complete electrosorption/desorption cycle of each target
metal ion, the used ACC electrodes were removed from the CDI cell and
the microstructure of the electrode surface was examined by field
emission scanning electron microscopy (SEM).

3. Results and discussion

3.1. Electrosorptive kinetics

In order to investigate the electrosorptive rates of metal ions by CDI,

the pseudo-first order (Lagergren equation, Eq. (3)) and pseudo-second-
order (Eq. (4)) equations were applied to fit the experimental data using
the method of least squares [21,22].
ln qe−ln(qe − q) = k1 t (3)

t 1 1
= + t
Fig. 1. Schematic diagram of reactor setup and CDI configuration. q k2 qe2 qe (4)

258
Y. Li et al. Journal of Water Process Engineering 26 (2018) 257–263

Table 1
R2 values of Lagergren equation and pseudo-second-order equation obtained by
least-squares fitting to the experimental data for each metal ion.
Ion R2 of Lagergren equation R2 of pseudo-second-order equation

Na+ 0.9562 0.3695

K+ 0.9690 0.2831
Mg2+ 0.9607 0.7323
Ca2+ 0.9636 0.7811
Cd2+ 0.9411 0.4220
Pb2+ 0.8873 0.6815
Cu2+ 0.9797 0.8598

where k1 is the rate constant (min−1) of the Lagergren equation, k2 is

the rate constant (g mmol−1 min−1) of the pseudo-second-order equa-
tion, qe and q are the amount of electrosorbed metal ion (mmol/g) at
equilibrium and any time during the CDI treatment, respectively.
It was found that the Lagergren equation fit the experimental data
for the metal ions better than the pseudo-second-order equation due to
the higher R2 values (Table 1). The electrosorptive kinetics of these
metal ions was well presented by the Lagergren equation, which agreed
with Mossad et al. [23] on CDI removal of salinity and hardness from
brackish groundwater. The pseudo-second-order kinetics, on the other
hand, was reported to fit better on physical adsorption of some heavy
metals according to Wang et al [24]. The less preferred application of
the pseudo-second-order kinetics in the present study implied the lump
sum influences from physical adsorption was minimal and the mono-
layer electrosorption via electrical double-layer (EDL) was the pre-
dominant mechanism during the CDI treatment.

3.2. Effect of metal carryover on electrosorptive kinetics

The use of a single CDI cell for tests of multiple metal ions was found
limited due to the carryover of metal ions from the previous run. Fig. 2 Fig. 2. Electrosorption of K+ in mono-component solutions during the initial
shows the electrosorptive kinetics of K+ using a same CDI cell in two use of a CDI cell with virgin activated carbon electrodes and the use of the CDI
cell after one complete electrosorption/desorption cycle of Pb2+: (a) electro-
scenarios: (1) the use of virgin electrodes for the electrosorption of K+
sorptive amount; (b) electrosorptive kinetics.
and (2) the use of the same electrodes for the electrosorption of K+ after
one complete electrosorption/desorption cycle of Pb2+. The initial use
of the CDI cell reported electrosorption of K+ with a Lagergren rate significant change in electrosorptive kinetics [30]. To address any
constant of 0.1232 min−1. However, after one complete electrosorp- possible discrepancy among the CDI cells during the assembly process,
tion/desorption cycle of Pb2+, the CDI cell reported much slower the electrosorptive kinetics of metal ions were compared against the
electrosorption of K+ with a Lagergren rate constant of 0.0859 min−1. Na+ and a normalization process was presented in Eq. (5) as follows:
The 30% reduction on the electrosorptive rate constant of K+ was at- The Lagergren rate constant of the target ion was divided by that of the
tributed to inorganic fouling of the ACC electrodes by formation of lead reference ion
hydroxide complexes on the electrodes [25]. The phenomenon reduced km
the effective surface area for ion electrosorption. Similar observations Normalized electrosorptive rate =
kref (5)
were reported by Mossad and Zou [26], who found when ferric ion was
added, the salt removal efficiency started to decline. The results in- where km and kref are the Lagargren rate constants of the target metal
dicate if a single CDI cell is used for the tests of multiple metal ions, the ion and reference metal ion, respectively, using a same CDI cell. The
influence from the carryover of other metal ions would make the ratios representing the normalized electrosorptive rate for each metal
electrosorptive results difficult to reproduce and thus invalidate any ion were compared and discussed in the following section.
comparisons. Acid treatment is known capable of controlling the ac-
cumulation of inorganic salts retained on the electrode surface. Con- 3.3. Comparative study of the normalized electrosorptive rates
sidering the externally introduced H+ ions may become another vari-
able, the electrosorption could otherwise be affected by competition 3.3.1. Electrosorption of alkali metal ions
between the externally introduced H+ and the metal ion [27,28], ir- Comparison of two monovalent alkali metal ions is illustrated in
reversibile portion of fouling, and potential change of functional groups Fig. 3a. K+ showed faster electrosorption as evidenced by a higher
on the electrode surface [29]. Lagergren rate constant (0.0897 min−1) than that of Na+
Multiple CDI cells made with the same mass of virgin ACC elec- (0.0742 min−1). The normalized electrosorptive rate of K+ was 1.21.
trodes allowed each set of electrosorptive experiment using its own The trend of K+ > Na+ was opposite to those of Gabelich et al. [11]
apparatus so as to avoid errors due to the carryover of miscellaneous and Xu et al. [12] whose results were obtained from multi-ionic solu-
metals from the used electrodes. Na+ is known to have no noticeable tions with different initial molar concentrations of metal ions. Since the
effect on long term operation of CDI process [26]. In the present study, present experiment started with the same initial molar concentrations
Na+ was used as the reference ion, and each CDI cell was used to obtain in a mono-component system, the phenomenon could be attributed to
the electrosorptive kinetics for the reference ion and the target ion. It the difference in their ionic hydrated radii. It is known that for two ions
has been noted that a tiny difference in spacer width would lead to a with the same charge, the one with larger ionic hydrated radius was

259
Y. Li et al. Journal of Water Process Engineering 26 (2018) 257–263

Fig. 3. Electrosorptive kinetics of (a) K+, (b) Mg2+, (c) Ca2+, (d) Cd2+, (e) Pb2+, and (f) Cu2+ in mono-component solutions.

Table 2 electrosorption of K+ is more favorable to the CDI electrode.

Chemical data of tested ions.
Ion Charge Mass [amu] Ionic radius [pm] Ionic hydrated radius [pm] 3.3.2. Electrosorption of alkali-earth metal ions
Na +
+1 22.99 116 358
Fig. 3b and c show the electrosorptive kinetics of Mg2+ and Ca2+ in
K+ +1 39.098 152 331 reference to Na+, respectively. The sequence of the normalized elec-
Mg2+ +2 24.305 86 428 trosorptive rates was Ca2+ (0.66) > Mg2+ (0.58). Again, the result
Ca2+ +2 40.08 114 412 could be explained by the difference in their ionic hydrated radii: Ca2+
Cd2+ +2 112.411 97 426
(412 pm) < Mg2+ (428 pm) and thus the electrosorption of Ca2+ is
Pb2+ +2 207.2 112 401
Cu2+ +2 63.546 72 419 more favorable compared to that of Mg2+. It is important to note that
the normalized electrosorptive rates were both less than 1, indicating
less preferred electrosorption of divalent cations compared to the
encountered with a stronger EDL overlapping, which caused restrictions monovalent cations. It is proposed that although the divalent cations
for the ion moving onto the electrode surface [31]. Since K+ is smaller are associated with stronger Columbic interactions between the charged
in hydrated radius (331 pm) than Na+ (358 pm) (Table 2) [32], the surfaces, they may have more restrictions to penetrate the pores of the
electrode due to stronger EDL overlapping. Summing up the above, the

260
Y. Li et al. Journal of Water Process Engineering 26 (2018) 257–263

Fig. 4. SEM images of (a) virgin activated carbon cloth (ACC) electrode, (b) ACC electrode after initial CDI desorption of Cd2+, (c) ACC electrode after initial CDI
desorption of Cu2+, and (d) ACC electrode after initial CDI desorption of Pb2+.

sequence of normalized electrosorptive rates for alkali and alkali-earth cathode was observed by Liu et al. [27] at a voltage of 1.5 V. Because
metal ions was as follows: K+ (1.21) > Na+ (1.00) > Ca2+ the standard reduction potential of Pb2+ was higher than that of Cu2+,
(0.66) > Mg2+ (0.58). This was partially contradictory with literatures the electrodeposition effect of Pb2+ could be smaller than that of Cu2+.
reporting both mass and percentage removals in competitive multi- Moreover, in addition to the EDL charging, oxygen could be reduced to
ionic solutions [11,12,33], but largely agreed with electrosorptive be- OH− at the cathode [39], which might facilitate the formation of hy-
haviors in unit of mmol/g in single-electrolyte solutions [13]. drolyzed lead complexes on the electrode [25] and reduce the acces-
sible pores for ion storage. On the other hand, the ACC electrode ma-
3.3.3. Electrosorption of heavy metal ions terial had been quantified with 0.4 meq/g lactonic groups, 0.2 meq/g
Fig. 3d–f show the electrosorptive kinetics of Cd2+, Pb2+ and Cu2+ phenolic groups, and 0.1 meq/g total basic groups [19]. The pH value at
in reference to Na+, respectively. All normalized electrosorptive rates the point of zero charge (pH-PZC) was 6.80. For the heavy metal so-
of these heavy metal ions were less than 1, indicating slower electro- lutions that started with a lower pH than pH-PZC, the surface charge of
sorption compared to the monovalent Na+. The order of the normalized ACC electrodes and the degree of protonation of the functional groups
electrosorptive rates followed Cu2+ (0.72) > Pb2+ (0.62) > Cd2+ were affected, suggesting heavy metals being adsorbed by chelation
(0.59). Because the trend of ionic hydrated radii for the three metal ions instead of EDL capacitors-based electrosorption [40].
was Pb2+ (401 pm) < Cu2+ (419 pm) < Cd2+ (426 pm), an inversely It has been noted that the solutions became more acidic after elec-
proportional relationship between electrosorptive rates and ionic hy- trosorption of the three heavy metal ions, and electrodeposition of Cu
drated radii did not hold for these heavy metal ions. The result was also was observed on the cathode. Huang et al. [14] discussed smaller sta-
partially contrary to the affinity order of Pb2+ > Cu2+ > Cd2+ during bility constant (pKa) values for stronger metal ligands to OH− and
a competitive physical adsorption [34]. It is commonly believed that lower OH− stability. Although the pH conditions affect the OH−
the chemical interaction between the metal ions and the surface func- complex formation, such concentrations of OH− complexes were not
tional groups of adsorbent is the major sorption mechanism during quantified in this study. The present research agreed with Huang et al.
physical adsorption [35]. The electrosorption mechanisms of these [14] on the discovery of slower electrosorption of Cd2+ than Pb2+.
heavy metal ions, on the other hand, could be more complex, which are Since Cd2+ had been reported to demonstrate less physical adsorption
discussed as follows. in mono-component solutions [41], the slower kinetics of both elec-
Cu2+ showed faster electrosorption than other tested heavy metal trosorption and physical adsorption for Cd2+ could be attributed to its
ions. The mechanism of enhanced Cu2+ removal by CDI was two-fold: larger ionic hydrated radius (426 pm). It was proposed that EDL char-
(1) electrosorption via formation of EDL upon an applied voltage. ging was the predominant mechanism for Cd2+ removal by CDI. The
Increased Cu2+ removal was observed with application of a voltage as effect of OH− complex formation, however, might be insignificant,
low as 0.3 V compared to that of an open circuit [36,37]; and (2) because the pKa1 of Cd2+ (10.08) is higher than that of Pb2+ (7.71).
electrodeposition of Cu2+ via electrochemical reactions. Although there The SEM images shown in Fig. 4 further confirm that the ACC elec-
were no evident redox reactions in the voltage range of 0-0.9 V [38], trodes were less influenced by Cd2+ (Fig. 4b) compared to the heavily
the present study used a cell voltage of 1.2 V, and Cu2+ could be re- influenced electrodes by Pb2+ and Cu2+ (Fig. 4c and d)
duced to cuprous oxide (Cu2O) or/and copper metal by electrodeposi-
tion [36]. 3.3.4. Relationship between the normalized electrosorptive rates and ionic
The mechanism for the removal of Pb2+ by CDI was more complex, hydrated radii
as evidenced by the relatively poor R2 value (0.8873) of Lagergren Transport of anions and cations to their respective electrode is by
equation (Fig. 3e). Reduction of Pb2+ and deposition on the surface of convection, diffusion, and electromigration [3]. Since there is evidence

261
Y. Li et al. Journal of Water Process Engineering 26 (2018) 257–263

Fig. 5. Comparison of tested metal ions regarding their normalized electrosorptive rates and ionic hydrated radii.

Table 3 undermined, which affected their continuous uses. The results indicate
Desorption performance of metal ions after one electrosorption cycle of CDI that the CDI electrodes are susceptible to serious inorganic fouling by
with virgin activated carbon electrodes. Cu2+ and Pb2+ during the CDI treatment. Frequent acid washes are
Ion Desorption efficiency Duration of desorption [h] needed to ensure long-term operational efficiency in practical en-
gineering applications.
+
Na 100% 0.5
K+ 100% 0.5
4. Conclusions
Mg2+ 94% 1.0
Ca2+ 91% 1.0
Cd2+ 94% 2.0 The electrosorptive behaviors of Na+, K+, Mg2+, Ca2+, Cd2+, Pb2+
Cu2+ 62% 3.0 and Cu2+ in mono-component solutions during the CDI treatment were
Pb2+ 24% 5.0 investigated. It was found that the electrosorptive kinetics of all tested
metal ions followed the Lagergren equation. By comparing the nor-
malized electrosorptive rates of tested ions in reference to Na+, the
that counterions also played an important role in an individual ion’s
study revealed an affinity preference order of
sorption [11], the present research used chloride as the only external
K+ > Na+ > Ca2+ > Mg2+ for alkali and alkali-earth metals and
anion and thus, it can be deduced that the differences in observation of
Cu2+ > Pb2+ > Cd2+ for heavy metals during the CDI treatment. It
electrosorptive rates solely depend on the affinity of cations to the CDI
was deduced that the ionic hydrated radii generally dictated the affinity
electrode. Comparison of all tested metal ions regarding their normal-
preference of tested ions except Cu2+ and Pb2+. The abnormal elec-
ized electrosorptive rates and ionic hydrated radii is presented in Fig. 5.
trosorptive behaviors of Cu2+ and Pb2+ were attributed to the complex
It was found that except the abnormal behaviors of Cu2+ and Pb2+ due
mechanisms occurring during electrosorption, such as electrodeposition
to their complex mechanisms during electrosorption, all other metal
and OH− complexes formation beside EDL charging. Moreover, Cu2+
ions followed a trend of inversely proportional relationship. It was
and Pb2+ showed significantly low desorption efficiency during elec-
deduced that the ionic hydrated radii generally dictated the removal
trode regeneration, which implied serious fouling potential during the
phenomena of metal ions in mono-component solutions, which was in
CDI treatment. This work provides fundamental aspects on individual
some aspects similar to the ion exchange and membrane separation
removal of metal ions, and the findings can be beneficially used as a
processes [11,42].
reference for the treatment of industrial effluents by CDI.

3.4. Regeneration of CDI electrodes Acknowledgement

Table 3 shows the desorption performance of CDI electrodes after The authors acknowledge the start-up fund provided by College of
one electrosorption cycle of metal ions. It was found that the alkali Natural Sciences and Mathematics at Indiana University of
metal ions including Na+ and K+ had no noticeable effect on electrode Pennsylvania to support this research.
regeneration, since the electrosorbed ions could be 100% desorbed by
short-circuiting in 0.5 h. The alkali-earth metal ions including Mg2+ References
and Ca2+, however, took longer time for desorption, and there were
slight losses of electrosorption capacity due to scaling [26]. The deso- [1] A.K. Meena, G. Mishra, P. Rai, C. Rajagopal, P. Nagar, Removal of heavy metal ions
rption of heavy metal ions were found most time-consuming. The from aqueous solutions using carbon aerogel as an adsorbent, J. Hazard. Mater. 122
(2005) 161–170, https://doi.org/10.1016/j.jhazmat.2005.03.024.
desorption efficiencies of Cu2+ and Pb2+ were only 62% and 24%, [2] P. Rana, N. Mohan, C. Rajagopal, Electrochemical removal of chromium from
respectively. The occurrence of electrodeposition, formation of OH− wastewater by using carbon aerogel electrodes, Water Res. 38 (2004) 2811–2820,
complexes, and slow diffusion of heavy metal ions due to their large https://doi.org/10.1016/j.watres.2004.02.029.
[3] P. Rana-Madaria, M. Nagarajan, C. Rajagopal, B.S. Garg, Removal of chromium
atomic masses may contribute to the poor desorption efficiencies. The from aqueous solutions by treatment with carbon aerogel electrodes using response
surface analysis of the ACC by SEM also reveals loss of active electro- surface methodology, Ind. Eng. Chem. Res. 44 (2005) 6549–6559, https://doi.org/
sorption sites by Cu2+ and Pb2+ (Fig. 4). During two additional CDI 10.1021/ie050321p.
[4] M.E. Suss, S. Porada, X. Sun, P.M. Biesheuvel, J. Yoon, V. Presser, Water desali-
cycles, the desorption efficiencies of alkali, alkali-earth metals and nation via capacitive deionization: what is it and what can we expect from it?
Cd2+ were consistently as high as their first reported values. Con- Energy Environ. Sci. 8 (2015) 2296–2319, https://doi.org/10.1039/C5EE00519A.
sidering the desorption of Cu2+ and Pb2+ could not proceed to com- [5] Q. Yao, H.L. Tang, Occurrence of re-adsorption in desorption cycles of capacitive
deionization, J. Ind. Eng. Chem. 34 (2016) 180–185, https://doi.org/10.1016/j.
pletion, the salt adsorption capacity of electrodes had been

262
Y. Li et al. Journal of Water Process Engineering 26 (2018) 257–263

jiec.2015.11.004. acid on competitive adsorption of Pb(II), Cd(II) and Cr(II) onto titanate nanotubes,
[6] W. Tang, P. Kovalsky, D. He, T.D. Waite, Fluoride and nitrate removal from brackish Chem. Eng. J. 215-216 (2013) 366–374, https://doi.org/10.1016/j.cej.2012.11.
groundwaters by batch-mode capacitive deionization, Water Res. 84 (2015) 029.
342–349, https://doi.org/10.1016/j.watres.2015.08.012. [25] A.A. Ihsanullah, A.M. Al-Amer, T. Laoui, M.J. Al-Marri, M.S. Nasser, M. Khraisheh,
[7] L. Pan, X. Wang, Y. Gao, Y. Zhang, Y. Chen, Z. Suo, Electrosorption of anions with M.A. Atieh, Heavy metal removal from aqueous solution by advanced carbon na-
carbon nanotube and nanofibre composite film electrodes, Desalination 244 (2009) notubes: critical review of adsorption applications, Sep. Purif. Technol. 157 (2016)
139–143, https://doi.org/10.1016/j.desal.2008.05.019. 141–161, https://doi.org/10.1016/j.seppur.2015.11.039.
[8] Z. Chen, H. Zhang, C. Xu, Y. Wang, W. Li, A study of electrosorption selectivity of [26] M. Mossad, L. Zou, Study of fouling and scaling in capacitive deionisation by using
anions by activated carbon electrodes in capacitive deionization, Desalination 369 dissolved organic and inorganic salts, J. Hazard. Mater. 244-245 (2013) 387–393,
(2015) 46–50, https://doi.org/10.1016/j.desal.2015.04.022. https://doi.org/10.1016/j.jhazmat.2012.11.062.
[9] J.-H. Yeo, J.-H. Choi, Enhancement of nitrate removal from a solution of mixed [27] Y. Liu, J. Yan, D. Yuan, Q. Li, X. Wu, The study of lead removal from aqueous
nitrate, chloride and sulfate ions using a nitrate-selective carbon electrode, solution using an electro-chemical method with a stainless steel net electrode
Desalination 320 (2013) 10–16, https://doi.org/10.1016/j.desal.2013.04.013. coated with single wall carbon nanotubes, Chem. Eng. J. 218 (2013) 81–88, https://
[10] Y. Gao, L. Pan, H. Li, Y. Zhang, Z. Zhang, Y. Chen, Z. Sun, Electrosorption behavior doi.org/10.1016/j.cej.2012.12.020.
of cations with carbon nanotubes and carbon nanofibres composite film electrodes, [28] I. Agbozu, F. Emoruwa, Batch adsorption of heavy metals (Cu, Pb, Fe, Cr and Cd)
Thin Solid Films 517 (2009) 1616–1619, https://doi.org/10.1016/j.tsf.2008.09. from aqueous solutions using coconut husk, Afr. J. Environ. Sci. Technol. 8 (2014)
065. 239–246, https://doi.org/10.5897/AJEST2013.1577.
[11] C.J. Gabelich, T.D. Tran, I.H. Suffet, Electrosorption of inorganic salts from aqueous [29] L. Zou, G. Morris, D. Qi, Using activated carbon electrode in electrosorptive deio-
solution using carbon aerogels, Environ. Sci. Technol. 36 (2002) 3010–3019, nisation of brackish water, Desalination 225 (2008) 329–340, https://doi.org/10.
https://doi.org/10.1021/es0112745. 1016/j.desal.2007.07.014.
[12] P. Xu, J.E. Drewes, D. Heil, G. Wang, Treatment of brackish produced water using [30] C. Wang, H. Song, Q. Zhang, B. Wang, A. Li, Parameter optimization based on ca-
carbon aerogel-based capacitive deionization technology, Water Res. 42 (2008) pacitive deionization for highly efficient desalination of domestic wastewater bio-
2605–2617, https://doi.org/10.1016/j.watres.2008.01.011. treated effluent and the fouled electrode regeneration, Desalination 365 (2015)
[13] H. Li, L. Zou, L. Pan, Z. Sun, Using graphene nano-flakes as electrodes to remove 407–415, https://doi.org/10.1016/j.desal.2015.03.025.
ferric ions by capacitive deionization, Sep. Purif. Technol. 75 (2010) 8–14, https:// [31] C.-H. Hou, C.-Y. Huang, A comparative study of electrosorption selectivity of ions
doi.org/10.1016/j.seppur.2010.07.003. by activated carbon electrodes in capacitive deionization, Desalination 314 (2013)
[14] Z. Huang, L. Lu, Z. Cai, Z.J. Ren, Individual and competitive removal of heavy 124–129, https://doi.org/10.1016/j.desal.2012.12.029.
metals using capacitive deionization, J. Hazard. Mater. 302 (2016) 323–331, [32] E. Nightingale Jr, Phenomenological theory of ion salvation, Effective radii of hy-
https://doi.org/10.1016/j.jhazmat.2015.09.064. drated ions, J. Phys. Chem. 63 (1959) 1381–1387.
[15] H. Duan, T. Yan, G. Chen, J. Zhang, L. Shi, D. Zhang, A facile strategy for the fast [33] M. Mossad, L. Zou, A study of the capacitive deionisation performance under var-
construction of porous graphene frameworks and their enhanced electrosorption ious operational conditions, J. Hazard. Mater. 213–214 (2012) 491–497, https://
performance, Chem. Commun. 53 (2017) 7465–7468, https://doi.org/10.1039/ doi.org/10.1016/j.jhazmat.2012.02.036.
C7CC03424E. [34] Y.-H. Li, J. Ding, Z. Luan, Z. Di, Y. Zhu, C. Xu, D. Wu, B. Wei, Competitive ad-
[16] Z.U. Khan, T. Yan, L. Shi, D. Zhang, Improved capacitive deionization by using 3D sorption of Pb2+, Cu2+ and Cd2+ ions from aqueous solutions by multiwalled
intercalated graphene sheet-sphere nanocomposite architectures, Environ. Sci. carbon nanotubes, Carbon 41 (2003) 2787–2792, https://doi.org/10.1016/S0008-
Nano 5 (2018) 980–991, https://doi.org/10.1039/C7EN01246B. 6223(03)00392-0.
[17] J. Zhang, J. Fang, J. Han, T. Yan, L. Shi, D. Zhang, N, P, S co-doped hollow carbon [35] G.P. Rao, C. Lu, F. Su, Sorption of divalent metal ions from aqueous solution by
polyhedron derived from MOFs-based core-shell nanocomposites for capacitive carbon nanotubes: a review, Sep. Purif. Technol. 58 (2007) 224–231, https://doi.
deionization, J. Mater. Chem. A 6 (2018) 15245–15252, https://doi.org/10.1039/ org/10.1016/j.seppur.2006.12.006.
C8TA04813D. [36] S.-Y. Huang, C.-S. Fan, C.-H. Hou, Electro-enhanced removal of copper ions from
[18] R. Zhao, M. van Soestbergen, H.H.M. Rijnaarts, A. van der Wal, M.Z. Bazant, aqueous solutions by capacitive deionization, J. Hazard. Mater. 278 (2014) 8–15,
P.M. Biesheuvel, Time-dependent ion selectivity in capacitive charging of porous https://doi.org/10.1016/j.jhazmat.2014.05.074.
electrodes, J. Colloid Interface Sci. 384 (2012) 38–44, https://doi.org/10.1016/j. [37] C.C. -Huang, Y.-J. Su, Removal of copper ions from wastewater by adsorption/
jcis.2012.06.022. electrosorption on modified activated carbon cloths, J. Hazard. Mater. 175 (2010)
[19] L. Donnaperna, L. Duclaux, M.P. Hirn, R. Gadiou, C. Merli, Adsorption of anionic 477–483, https://doi.org/10.1016/j.jhazmat.2009.10.030.
azo dyes onto microporous carbon felt and mesoporous carbon cloth, International [38] C.-C. Huang, J.-C. He, Electrosorptive removal of copper ions from wastewater by
Conference on Carbon (2009) p. ID785. using ordered mesoporous carbon electrodes, Chem. Eng. J. 221 (2013) 469–475,
[20] W. Liu, T. Wang, A.G.L. Borthwick, Y. Wang, X. Yin, X. Li, J. Ni, Adsorption of Pb2+, https://doi.org/10.1016/j.cej.2013.02.028.
Cd2+, Cu2+ and Cr3+onto titanate nanotubes: competition and effect of inorganic [39] J.-H. Lee, W.-S. Bae, J.-H. Choi, Electrode reactions and adsorption/desorption
ions, Sci. Total Environ. 456-457 (2013) 171–180, https://doi.org/10.1016/j. performance related to the applied potential in a capacitive deionization process,
scitotenv.2013.03.082. Desalination 258 (2010) 159–163, https://doi.org/10.1016/j.desal.2010.03.020.
[21] R.L. Zornitta, L.A.M. Ruotolo, Simultaneous analysis of electrosorption capacity and [40] P. Liu, T. Yan, J. Zhang, L. Shi, D. Zhang, Separation and recovery of heavy metal
kinetics for CDI desalination using different electrode configurations, Chem. Eng. J. ions and salty ions from wastewater by 3D graphene-based asymmetric electrodes
332 (2018) 33–41, https://doi.org/10.1016/j.cej.2017.09.067. via capacitive deionization, J. Mater. Chem. A 5 (2017) 14748–14757, https://doi.
[22] D. Liu, K. Huang, L. Xie, H.L. Tang, Relation between operating parameters and org/10.1039/C7TA03515B.
desalination performance of capacitive deionization with activated carbon elec- [41] K. Kadirvelu, J. Goel, C. Rajagopal, Sorption of lead, mercury and cadmium ions in
trodes, Environ. Sci.: Water Res. Technol. 1 (2015) 516–522, https://doi.org/10. multi-component system using carbon aerogel as adsorbent, J. Hazard. Mater. 153
1039/C5EW00102A. (2008) 502–507, https://doi.org/10.1016/j.jhazmat.2007.08.082.
[23] M. Mossad, W. Zhang, L. Zou, Using capacitive deionisation for inland brackish [42] S.-J. Seo, H. Jeon, J.K. Lee, G.-Y. Kim, D. Park, H. Nojima, J. Lee, S.-H. Moon,
groundwater desalination in a remote location, Desalination 308 (2013) 154–160, Investigation on removal of hardness ions by capacitive deionization (CDI) for
https://doi.org/10.1016/j.desal.2012.05.021. water softening applications, Water Res. 44 (2010) 2267–2275, https://doi.org/10.
[24] T. Wang, W. Liu, L. Xiong, N. Xu, J. Ni, Influence of pH, ionic strength and humic 1016/j.watres.2009.10.020.

263