ARTICLE

pubs.acs.org/est

Mesoporous Carbon for Capacitive Deionization of Saline Water

C. Tsouris,†,* R. Mayes,† J. Kiggans,† K. Sharma,‡ S. Yiacoumi,‡ D. DePaoli,† and S. Dai†
†
Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6181, United States
‡
Georgia Institute of Technology, Atlanta, Georgia 30332-0373, United States

bS Supporting Information
ABSTRACT: Self-assembled mesoporous carbon (MC) materials have
been synthesized and tested for application in capacitive deionization
(CDI) of saline water. MC was prepared by self-assembly of a triblock
copolymer with hydrogen-bonded chains via a phenolic resin, such as
resorcinol or phloroglucinol in acidic conditions, followed by carboniza-
tion and, in some cases, activation by KOH. Carbon synthesized in this
way was ground into powder, from which activated MC sheets were
produced. In a variation of this process, after the reaction of triblock
copolymer with resorcinol or phloroglucinol, the gel that was formed was
used to coat a graphite plate and then carbonized. The coated graphite
plate in this case was not activated and was tested to serve as current
collector during the CDI process. The performance of these MC materials
was compared to that of carbon aerogel for salt concentrations ranging
between 1000 ppm and 35,000 ppm. Resorcinol-based MC removed up to 15.2 mg salt per gram of carbon, while carbon aerogel
removed 5.8 mg salt per gram of carbon. Phloroglucinol-based MC-coated graphite exhibited the highest ion removal capacity at
21 mg of salt per gram of carbon for 35,000 ppm salt concentration.

’ INTRODUCTION
Capacitive deionization (CDI) is being developed as a poten-
tial method for removing salts from aqueous solutions. The
mechanism of CDI is similar to that of energy storage in super-
capacitors. It involves application of an electric field between two
electrodes to force ionic species toward oppositely charged
electrodes. The ions are held within the electrical double layer
(EDL) formed near the electrode surfaces. Electrodes of low
electrical resistivity and high surface area contribute to a higher
ion removal capacity. This process has several potential advan-
tages over conventional desalination methods including reversi-
bility, operation at low voltages, potential for low energy require-
ments, and reduction of secondary wastes. CDI has been investigated
for the removal of sodium, chloride, and heavy metal ions using
carbon aerogels.1 5 The potential for commercial application of
CDI as an energy efficient technology for water treatment has
been discussed in recent literature.6,7
Because of their high conductivity and high surface area per
unit mass, porous, nanostructured carbon aerogels are acceptable
materials for CDI. Carbon aerogel has a low electrical resistivity
(<40 mΩ) and high surface area (∼400 m2/g),8 10 and the
removal efficiency of ions by these materials can be measured
in terms of electrosorption capacitance. Previous studies have
reported that the electrosorption efficiency is dependent on the
size distribution of pores and the presence of surface functional
groups.8,11 13 Carbon aerogels can be specified as mesoporous
(20 Å < pore diameter <500 Å) and microporous (<20 Å). The
presence of significant microporosity results in greater surface
area; however, micropores inhibit ion penetration, and the mate-
rial is more susceptible to collapse at higher temperatures.12,14 16
Ying et al.6 showed that a higher surface area associated with
microporosity does not contribute to higher ion capacity because
of electrical double-layer overlapping.
Johnson and Newman17 discussed a macroscopic porous elec-
trode model to simulate the electrosorption mechanism, while
Farmer et al.2 proposed that self-shielding effects might explain
the discrepancy between their CDI experimental results and the
Gouy Chapman theory predictions for simple planar electro-
des. Lin et al.15 conducted experiments suggesting that 8 Å is the
cutoff pore width, below which ions cannot penetrate micro-
pores, while micropores of 8 20 Å in width account for the
major part of electrosorption capacitance. Yang et al.18 developed
an EDL model taking into account EDL overlapping effects on
electrosorption capacity. When the overlapping effect was ignored,
model results overestimated the electrosorption capacitance.
To achieve a high capacity, electrode materials of high specific
area, electrical conductivity, and stability are required. Although
previous works have mainly focused on carbon aerogel as the
electrosorption material,18,19 the majority of surface area of this
material is in pores of <10 Å, which do not contribute significantly
to the electrosorption capacity.12 Thus, even though carbon
Received: May 6, 2011
Accepted: October 27, 2011
Revised: October 15, 2011
Published: October 27, 2011

r 2011 American Chemical Society 10243 dx.doi.org/10.1021/es201551e | Environ. Sci. Technol. 2011, 45, 10243–10249
Environmental Science & Technology
aerogel has a relatively high surface area, its capacitance may not
be as high as that of mesoporous materials. Furthermore, as
microporosity limits mass transport rates, there may be some
kinetic advantage in utilizing materials in which the surface area
arises from mesopores. Therefore, we have been investigating
the use in CDI of materials having high surface area in the meso-
porous range.
The soft-template method for preparing mesoporous carbons
has been discussed in detail elsewhere.14 The phenolic precur-
sors, i.e. resorcinol or phloroglucinol, incorporate into the hydro-
philic PEO arms of the structure-directing template, i.e. a copolymer
such as Pluronic F127. This is followed by an aldehyde con-
densation involving formaldehyde or glyoxal, thus cross-linking
the phenolics together while “locked” in place by the copolymer
structure-directing template.14 During carbonization, the copo-
lymer decomposes leaving void spaces in the phenolic carbon
polymer, which become the pores. The narrow mesopore size
distribution in soft-templated carbon materials is directly related
to the use of the copolymer template, while the pore size is re-
lated to the hydrophobic block length in the copolymer template.
Carbon aerogel does not have this level of pore control and
therefore does not have a narrow mesopore size distribution due
to the mesoporosity being introduced through a solvent extrac-
tion step prior to curing. The narrow pore size distribution found
in templated carbons has the benefit of producing a higher sur-
face area with similar pore volumes found in aerogels but with
pores predominantly in the mesopore regime.
A number of recent studies have been focused on the applica-
tion of various carbon materials such as carbon aerogel,20 ordered
mesoporous carbon,21,22 graphene,23,24 carbon nanotubes,25 29
carbon fiber webs,30 etc. in the CDI process. Zou et al. (2008)
concluded that the presence of ordered structure in mesoporous
carbon and control of micropores lead to a greater electrosorption
capacity in mesoporous carbon. Li et al. (2009) investigated a sol
gel process to synthesize ordered mesoporous carbon (OMC) and
deduced that OMC obtained by addition of NiSO4 3 6H2O
exhibited highest electrosorption capacity. Previously, it was
shown that hierarchical structures are good for increased mass-
transport throughout the carbon material.31 A new ordered
mesoporous carbon material was developed in our laboratories
using glyoxal, which exhibits a hierarchical structure arising from
the cross-linking reagent with pore sizes up to 200 nm. It was also
shown that mesoporous carbon prepared by self-assembly of block
copolymer and phloroglucinol-formaldehyde resin via the soft-
template method can be activated by CO 2 and potassium
hydroxide.32 Activated mesoporous carbons exhibited high spe-
cific surface areas (up to 2000 m2 g 1) and large pore volumes (up
to 1.6 cm3 g 1), while maintaining the mesoporous structure.
In the current study, self-assembled mesoporous carbon
(MC) electrodes suitable for CDI have been synthesized and
characterized to determine their properties, including pore size
distribution and specific capacitance. The main purpose of the
study is to compare the CDI performance of developed MC elec-
trode materials and commercial carbon aerogel. The effects of
applied voltage and solution concentration on the electrosorp-
tion capacity of mesoporous carbon have also been investigated.
’ MATERIALS AND METHODS
Materials Synthesis and Characterization. For gel-based
electrodes, resorcinol (8.00 g) and Pluronic F-127 (8.00 g)
(BASF) were dissolved in 34 mL of ethanol (200 proof, Decon
10244
ARTICLE
Figure 1. Schematic of the procedure for producing graphite plates
coated with mesoporous carbon (RMCG and PMCG materials).
Laboratories, Inc.) and 34 mL of 3 M hydrochloric acid. After the
addition of formaldehyde, the mixture was stirred for 50 min,
with phase separation occurring at 20 min after formaldehyde
addition. The mixture was allowed to set and the gel separated
followed by the addition of chopped carbon fibers (Toho Tenax
344). The gel-fiber composite was spread onto the roughened
surface of graphite current collector plates and covered with a
carbon fiber sheet (Hollingsworth and Vose) that was subse-
quently pressed into the gel lightly. The composite electrodes
were allowed to dry and then cured at 353 K for 24 h. The plates
were carbonized at 1123 K under argon.31 33 Details about the
templating mechanism can be found in Liang and Dai (2006).34
The mesoporous carbon used to produce activated carbon (AC)
is described elsewhere.32 Briefly, phloroglucinol (26 g) and Plu-
ronic F127 (52 g) were refluxed in acidic ethanol (1.2 L of 200
proof ethanol and 10 mL of 12 M HCl) followed by addition of
formaldehyde (26 g). The solution was allowed to reflux 3 h
followed by filtration and carbonization in a nitrogen atmosphere
at 1123 K. The resulting mesoporous carbon was physically
mixed with potassium hydroxide in a 1:4 ratio respectively and
heated at 10 K/min to 1123 K for 2 h. The carbon was then rinsed
with deionized water, stirred in warm dilute hydrochloric acid,
and then washed to neutral pH with deionized water. The meso-
porous carbon was incorporated into two types of CDI electrodes
activated carbon sheets (ACS), and mesoporous carbon coated
graphite plates as described below.
ACS Preparation. The activated MC was ball milled with high-
density zirconia media (Glenn Mills) in isopropanol to produce
MC powder with an average particle size of 10 μm. Slurries
of activated MC powder were made of 7.5 wt % PVDF binder
(polyvinylidene fluoride) in dimethylformamide (DMF). The
slurry was heated to decrease the viscosity of the solution and
spread onto a mold for drying. Nonwoven carbon fiber paper
(Hollingsworth and Vose) and bonded carbon fiber paper, EC-
TP2-030 (Toray Industries) were used as both a nonporous layer
for backing the carbon sheet and as a nonstick mold liner,
respectively. ACS samples were generated by drying the slurry
at 100 °C under rough vacuum. Samples were cut into small por-
tions for characterization and into 0.102 by 0.203 m sheets for
CDI experiments.
MC-Coated Graphite. The steps for generating MC-coated
graphite plates are shown in Figure 1. Prior to carbonization, the
carbon polymer gel was mixed with chopped carbon fibers, cast
onto the roughened surface of graphite plates, and covered with a
carbon fiber sheet. The graphite plates must have a rough surface
for good adhesion while the carbon fibers and fiber sheet assist
adhesion by breaking surface tension, allowing smaller domains
of mesoporous carbon on the graphite plate. The graphite plates
with carbon polymer were then carbonized under nitrogen at
dx.doi.org/10.1021/es201551e |Environ. Sci. Technol. 2011, 45, 10243–10249
Environmental Science & Technology
1123 K to produce resorcinol-based (R-type) MC-coated gra-
phite (RMCG) planar CDI electrodes. The mesoporous carbon
adheres well to the graphite and does not flake off during cutting
with an abrasive disk. This strong adhesion suggests the meso-
porous carbon may withstand exposure to liquid shear flows. The
gel can be deposited on both sides of the graphite, enabling double-
sided electrodes in CDI stacks. A macroporous-mesoporous hier-
archical gel-coated graphite plate is produced via an identical
method as described in ref 32.
Nitrogen sorption analysis was performed using a Micromeri-
tics Gemini Analyzer. Surface area was calculated using the
Brunnauer-Emmet-Teller (BET) method or a comparative method.35
Pore size distributions were calculated in the 0.05 0.2 relative
pressure region using the Barrett Joyner Halenda (BJH) method.36
Surface area analysis for the ACS sheets was performed on the
composite material, i.e., the activated carbon with binder and the
carbon fiber backing, while the analysis of the MC-coated gra-
phite electrodes was restricted to the mesoporous carbon
carbon fiber composite scraped off the graphite current collector.
The resistivity of the carbon aerogel, activated carbon sheets, and
MC coated graphite plates was measured by a Surface Resistivity
Meter (Guardian Manufacturing, Model SRM-232-2000). Multi-
ple measurements were taken at several points across the sheets,
and the average resistivity was calculated.
The synthesized mesoporous carbon sheets (ACS) and MC-
coated graphite plates were tested in the present work for their
electrosorption capacity and deionization rate at various ranges
of salt concentrations to determine the most effective material.
Cyclic Voltammetry. Cyclic voltammetry experiments were
performed with a Bioanalytical Systems (BAS, West Lafayette, IN)
voltammetric analyzer (CV-50W Version 2) connected to a BAS
C2 cell stand. The experiments employed three types of electro-
des: (1) the working electrode consisted of a platinum wire with a
small monolithic carbon sample clipped at the end; (2) the
counter electrode was a platinum wire attached to a large piece of
carbon aerogel, serving as a current collector; and (3) the BAS
Model RE-1 (Ag/AgCl electrode immersed in a 3 M NaCl
electrolyte solution) was used as the reference electrode. The
carbon samples were immersed in the cell solution such that pro-
per wetting conditions were maintained during the experiment. All
experiments were carried out under a flow of nitrogen gas.
Capacitive Deionization. The laboratory-scale CDI reactor
consisted of a single pair of carbon electrodes and a pair of cur-
rent collectors (Figure 2). For carbon aerogel and activated MC
sheets, the current collectors were made of titanium plates, and
the carbon sheets were taped at the edge on the current col-
lectors. For the MC-coated graphite, the graphite plates played
the role of both supporting material and current collectors. The
electrodes were held at a spacing of ∼6 mm by a hollow piece
of Plexiglas. The power was supplied to the cell by connecting
the current collectors to a direct-current (dc) HP 3632A power
supply (Hewlett-Packard, Loveland, CO) with a voltage range of
0 to 15 V and a current range of 0 to 7 Å. The conductivity was
measured at the cell exit stream by using a conductivity meter
(model 835 Multi Cell (Au) with Model 3082 EC Meter, Amber
Science, Inc., Eugene, OR). The active area of each plate was
∼110 cm2 and the approximate volume of the solution in the cell
was 50 mL. The aqueous solution that was made from Instant
Ocean (Aquarium Systems, content listed in Table S1) was con-
tinuously recycled through the system using a micropump until
steady state conditions were obtained (Figure 2). A solution
volume of 100 mL was maintained in the system. The temperature
10245
ARTICLE
Figure 2. Single-cell capacitive deionization reactor with two half cells
each of which containing (1) a Plexiglas cover, (2) a current collector,
(3) a carbon electrode, and (4) a middle hollow plate. Right: STEM
images of the P-type mesoporous carbon utilized to produce the MC-
coated graphite electrode materials showing uniform pores of ∼8 nm
(bottom right) with a combined hierarchical meso- and macroporosity
(top right).
of the solution was maintained at ∼298 K and the flow rate was
held constant at 30 mL/min. Regeneration was carried out by
shorting the electrodes and discharging the cell at 0 V.
’ RESULTS AND DISCUSSION
Templated carbon was employed in electrode materials in two
ways: powdered mesoporous carbon and gel-based mesoporous
carbon. Surface areas are comparable for the two materials prior
to activation of the powdered mesoporous carbon. Chemical
activation at high temperature with potassium hydroxide (KOH)
results in a significant increase in surface area based on the KOH
to carbon ratio during activation. At a 4:1 ratio, the surface area
increases from ∼440 m2/g to 1700 m2/g. This increase is due to
the introduction of microporosity without degradation of the
mesopores.32
Figure 3 shows the nitrogen adsorption isotherms and BJH
pore size distributions for the RMCG-FS (MC-coated graphite
with a fiber sheet), PMCG-FS (MC-coated graphite with a fiber
sheet), and commercial carbon aerogel. The commercial carbon
aerogel used in these tests is a high surface area material, 289 m2/g,
with a broad pore size distribution and a total pore volume of
0.70 cc/g. This is in contrast to the two templated carbon
samples, where slightly lower specific pore volume is achieved
with narrower pore size distributions. The surface areas for the
RMCG-FS and PMCG-FS active materials on the graphite plates
are 488 m2/g and 610 m2/g, respectively. Both templated carbon
materials have total pore volumes of ∼0.54 cc/g. The difference
in the surface areas for the templated carbons is a result of several
parameters during synthesis, including reaction time and ambi-
ent humidity. Stopping the reaction early can result in a lower
surface area as well as a less humid environment, which influences
hydrogen bonding during self-assembly. The primary difference
between the RMCG-FS and PMCG-FS is the hierarchical poro-
sity found in the PMCG-FS material. The PMCG-FS material
has ∼8 nm mesopores within a macroporous carbon matrix
where pores from 50 to 300 nm can be observed (Figure 2).31
This hierarchical porosity is believed to enable higher active sur-
face area and increased transport rates of ions through the mate-
rial. Without the hierarchical structure, the RMCG-FS material is
believed to have slower ion sorption kinetics than the PMCG-FS
dx.doi.org/10.1021/es201551e |Environ. Sci. Technol. 2011, 45, 10243–10249
Environmental Science & Technology
Figure 3. Nitrogen sorption isotherms and BJH pore size distributions
for RMCG, PMCG, and carbon aerogel materials.
carbon. The ACS electrode suffers a dramatic loss if surface area
is compared to the starting material due to the use of PVDF as a
binder from pore blockage or partial pore filling. The starting
surface area of 1700 m2/g is reduced to 364 m2/g upon gene-
ration of the ACS electrode. This negatively affects the sorption
capacity, as the capacity is directly related to the surface area of
the porous electrode material.
The average resistivity of carbon aerogel, activated carbon
sheets, and MC-coated graphite plates measured by the surface
resistivity meter was 0.0 ohm/sq, 15.8 ohm/sq, and 14.6 ohm/sq,
respectively. The reason mesoporous carbon had a higher resis-
tivity than carbon aerogel was because it was not isotropic like
carbon aerogel. While the resistance was 0 in one direction, it had
a finite value in another direction, so that the average value was
not zero. It is not expected that carbon aerogel, which is isotropic
with respect to resistivity, will perform better than mesoporous
carbon because current will preferentially flow to the current
collector along the path of minimum resistance, which is near
zero for both materials.
CDI Electrode Development. The activated carbon sheets
(ACS) produced in this work are macroscopically similar to the
carbon aerogel sheet material available commercially. Utilization
of the PVDF polymeric binder with the activated carbon resulted
in a decrease in porosity. This is believed to be due to the block-
ing or filling of micropores in the activated carbon by the
PVDF binder.37 The mesoporous carbon produced a smooth
self-supporting sheet when sandwiched between two loose-
weave carbon fiber sheets. The carbon fiber sheets are nonporous
and only serve to provide structural support. Figure 4a is an SEM
image of carbon aerogel. Notice the low density of carbon on the
carbon fibers while the ACS material has a significantly higher
amount of carbon on the carbon fibers (Figure 4b). With the
significant pore blocking/filling by polymeric binders, the ad-
vantages of activated carbons are muted and thus the ion capacity
is limited. Another limitation of the sheet material involves the
need to affix the carbon sheets to a current collector, such as
titanium plates, while maintaining a high conductivity of the
material.
The MC-on-graphite CDI electrodes circumvent the need for
polymeric binders by casting the MC in its precarbonized gel
state directly onto graphite plates. Combining the current col-
lector and porous carbon into a single monolithic electrode
offers additional benefits in the replacement or cleaning of elec-
trodes in a multicell CDI process. Furthermore, no secondary
10246
ARTICLE
Figure 4. SEM images of carbon aerogel (a) and ACS material (b).
Figure 5. Cyclic voltammetry results with activated MC sheets (ACS)
and commercial carbon aerogel (with 0.1 M NaCl) using 10 mV/s
scan rate.
conductive adhesives are needed as the carbon directly adheres
to the graphite.
Cyclic Voltammetry Experiments. For the analysis of CV
data, the measured electrical current is divided by the scan rate
and weight of the sample to yield the specific capacitance. The
results are presented by plotting specific capacitance (F/g)
versus applied potential E (with respect to the reference elec-
trode Ag/AgCl). The specific capacitance of carbon aerogel in
CV experiments has been reported in previous studies.12,38
Results in Figure 5 for a voltage range of (0.5 V, a scan rate of
1 mV/s, and NaCl and CaCl2 at 0.1 and 1 M electrolyte solutions
show that the activated MC sheets have approximately 50%
higher specific capacitance than commercial carbon aerogel sheets.
The capacitance for the high ionic-strength solutions is slightly
higher, because more ions participate in the electrical double
layer formation in solutions of higher ionic-strength. The results
obtained indicate that the MC material has the potential to
exceed the performance of current carbon aerogel sheets.
Capacitive Deionization Experiments. Experiments were
run to compare the carbon aerogel and the newly synthesized
activated MC sheets and MC-coated graphite plates. Mesoporous
carbon performed much better than carbon aerogel under the
conditions of the experiment. It was associated with 25% drop in
conductivity as compared to 2% drop for carbon aerogel
(Figure 6a). Another set of experiments was conducted to
compare the performance of MC-coated graphite and carbon
dx.doi.org/10.1021/es201551e |Environ. Sci. Technol. 2011, 45, 10243–10249
Environmental Science & Technology ARTICLE

Table 1. Capacitive Deionization Measurements on Meso-

porous Carbon Sheets
initial conc. final conc. voltage conc. change mg of salt
(ppm) (ppm) (V) (ppm) removed/g carbona

Activated MC Sheets
1000 650 1.2 350 3.5
1000 650 1.4 350 3.5
1000 770 1.0 230 2.3
3000 2450 1.2 550 5.5
3000 2470 1.2 530 5.3
5000 4740 1.2 260 2.6
5000 4870 1.2 130 1.3
RMCG-FS
1230 650 1.2 580 6.1
3970 2570 1.2 1400 14.6
9570 8890 1.2 680 7.1
7340 6380 1.2 960 10.0
4460 3070 1.2 1400 14.5
a
Significant removal capacity observed as compared to the values
Figure 6. Representative capacitive deionization test results: (a) carbon reported in Wang et al. (2010).
aerogel sheets and activated MC sheets . Conditions: 1000 ppm initial
concentration of Instant Ocean, at 1.2 V applied voltage. The electrical
conductivity of the solution dropped by 2% for carbon aerogel and 25%
for activated MC sheets. After ∼2,500 s, the conductivity increased
because the applied potential was turned to 0 and the electrodes were
shorted. (b) Comparison of RMCG and PMCG performance with
carbon aerogel performance for 5,000 ppm Instant Ocean initial
concentration. A voltage of 1.2 V was applied at 600 s in all the
experiments. Mass of carbon in each electrode: ACS 5.0 g MC;
RMCG-FS 4.8 g MC; aerogel 1.1 g.

aerogel at ∼5000 ppm Instant Ocean concentration. Results in

Figure 6b demonstrate that the MC-coated graphite, both R- and
P-type, have a higher capacity for ions than carbon aerogel. The
salt removal capacity of RMCG-FS and PMCG-FS at 5000 ppm
Instant Ocean concentration was 13.8 and 12.4 mg/g of carbon,
respectively.
Further CDI experiments were conducted using the activated
MC sheets and the MC-coated graphite plates to determine the
effects of applied voltage and salt concentration and compare the
different types of MC materials with respect to their performance
in deionization. Results in the first three lines in Table 1 show
that the decrease in conductivity and the mass of salt removed
were higher for the 1.2 V experiment as compared to 1 and 1.4 V
experiments. This behavior indicates that 1.2 V is an optimum
value for the applied voltage and that increasing the voltage above
this value does not result in a greater removal capacity for ions.
Farmer et al.2 also observed optimal performance at 1.2 V. The
rate of the conductivity decrease was observed to be only 13%
faster for the 1.4-V experiment as compared to the 1.2-V experi-
ment. Due to the relatively small increase in the deionization rate
and the smaller removal capacity for 1.4 V applied voltage, sub-
sequent experiments were run at 1.2 V.
The effect of salt concentration on the CDI performance of
activated MC sheets is also examined in Table 1 for initial salt
concentration of 1000, 3000, and 5000 ppm Instant Ocean solu-
tions. The highest removal efficiency was observed at 3000 ppm
concentration. Increasing the initial concentration to 5000 ppm
resulted in a sharp drop in the ion removal capacity. This implies
10247
Figure 7. Capacitive deionization results for carbon aerogel sheets and
R-type mesoporous-carbon-coated graphite at 40 °C. A voltage of 1.2 V
was applied at 600 s in the experiments. RMCG-FS and carbon aerogel
exhibit an ion removal capacity of 15.2 mg/g and 5.8 mg/g of carbon,
respectively.
that activated MC sheets are most suitable for use at intermediate
salt concentrations.
Further experiments run with the RMCG-FS electrodes
exhibited a higher removal capacity than activated MC sheets.
This result was observed for all salt concentrations used in the
experiments. The RMCG-FS electrodes showed the highest
mass of salt removed at salt concentrations in the range of
4000 ppm. This result indicates that the MC-coated graphite
electrodes can be used to deionize solutions of high con-
centration. Further experiments were carried out to compare
the performance of RMCG-FS and carbon aerogel and ob-
serve the influence of temperature on the performance of both
the materials. RMCG-FS exhibited a 34% higher ion removal
capacity over carbon aerogel, along with more efficient re-
moval at a higher temperature. The capacity of RMCG-FS
increased by 10.5% at 40 °C compared to the capacity at room
temperature (Figure 7).
dx.doi.org/10.1021/es201551e |Environ. Sci. Technol. 2011, 45, 10243–10249
Environmental Science & Technology
In a recent publication by Mayes et al.,31 RMCG and PMCG
were compared for potential application in the CDI process. The
main difference between these two materials is that the RMCG-
FS has a narrow pore size distribution in the range of 6 10 nm,
whereas the PMCG-FS has the narrow mesopore size distribu-
tion with a maximum at 7.5 nm, coupled with macropores with
sizes between 60 and 300 nm. Because of the relatively wide pore
size distribution, the PMCG showed a faster deionization rate
than that of the RMCG.31 Further experiments were conducted
here to compare the capacity of RMCG and PMCG at increasing
salt concentrations up to 35000 ppm, which corresponds to the
salinity of seawater. It was observed that the capacity of the
PMCG material increases with concentration and is higher than
the capacity of the RMCG material at similar concentrations. The
capacity of the RMCG material decreased slightly with concen-
tration. The highest capacity and ion-removal rate from the solu-
tion were 21 mg of salt and 6 mg/min, respectively, and were
observed for the PMCG electrodes at the highest salt concentra-
tion (i.e., 35000 ppm). For comparison, the capacity for carbon
aerogel was found to be 5.8 mg for a salt concentration of 5000
ppm Instant Ocean. Li et al. have reported salt removal capacity
of 0.93 mg/g of carbon for OMC-S, 0.54 mg/g of carbon for
OMC-N, 0.602 mg/g of carbon for OMC-W, and 0.275 mg/g of
carbon for activated carbon (AC).21 The materials prepared by
the addition of NiSO4 3 6H2O and Ni(NO3)2 3 6H2O are denoted
as OMC-S and OMC-N, respectively. MC obtained without
incorporating nickel has been denoted as OMC-W. The salt
removal capacity reported in this study is significant as compared
to the values reported by Li et al. (2009). The salt adsorption
efficiency reported by Zhao et al. (2010) in terms of salt adsorbed
per gram of carbon is 0.2 mmol/g or 11.68 mg/g of carbon at
Vcell = 1.4 V.39 This value is lower than the salt removal capacity
reported in the present study.
The results obtained in this study suggest that mesoporous
carbon materials are superior to carbon aerogel with respect to
capacitive deionization. This result is attributed to the fact that a
large fraction of the carbon aerogel surface area comes from
micropores that cannot be accessed by ions. MC-coated graphite
materials exhibit a higher capacity at higher concentrations of
salts. Thus, these materials are a more suitable choice for carrying
out desalination of salt water by the CDI process. Among the
MC-coated graphite materials, PMCG is most suited for high
concentration of dissolved salt and faster deionization rates.
Experimental results from MC-coated graphite electrodes dis-
cussed in this work indicate good promise for the implementa-
tion of mesoporous carbon for CDI.
’ ASSOCIATED CONTENT
bS Supporting Information. Table S1 gives information on
the constituents of synthetic seawater. Figure S2 contains
capacitive deionization results for RMCG-FS with NaCl solution
to show the electrodes are reversible. Figure S3 illustrates the
regeneration of PMCG-FS electrodes. This material is available
free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: 865-241-3246. Fax: 865-241-4829. E-mail: tsourisc@
ornl.gov.
10248
ARTICLE
’ ACKNOWLEDGMENT
This research was conducted at the Oak Ridge National Labora-
tory and supported by the U.S. DOE Office of Energy Efficiency
and Renewable Energy (EERE), under Contract DE-AC05-0096-
OR22725 with Oak Ridge National Laboratory, managed by UT-
Battelle, LLC. Partial support to S. Yiacoumi and K. Sharma was
provided by the National Science Foundation, under Grant No.
CBET-0651683. The authors are also thankful to Bob Campbell,
Tom Dorow, Sunita Kaushik, Bill Bourcier, and Fred Seamon of
Campbell Applied Physics, Inc., for frequent discussions on capa-
citive deionization.
’ REFERENCES
(1) Farmer, J. C.; Bahowick, S. M.; Harrar, J. E.; Fix, D. V.; Martinelli,
R. E.; Vu, A. K.; Carroll, K. L. Electrosorption of Chromium Ions on
Carbon Aerogel Electrodes as a Means of Remediating Ground Water.
Energy Fuels 1997, 11 (2), 337–347.
(2) Farmer, J. C.; Fix, D. V.; Mack, G. V.; Pekala, R. W.; Poco, J. F.
Capacitive Deionization of NaCl and NaNO3 with Carbon Aerogel
Electrodes. J. Electrochem. Soc. 1996, 143 (1), 159–169.
(3) Miller, J. M.; Dunn, B. Morphology and Electrochemistry of
Ruthenium/Carbon Aerogel Nanostructures. Langmuir 1999, 15 (3),
799–806.
(4) Oren, Y.; Soffer, A. Electrochemical Parametric Pumping.
J. Electrochem. Soc. 1978, 125 (6), 869–875.
(5) Ying, T.-Y.; Yang, K.-L.; Yiacoumi, S.; Tsouris, C. Electrosorp-
tion of Ions from Aqueous Solutions by Nanostructured Carbon
Aerogel. J. Colloid Interface Sci. 2002, 250 (1), 18–27.
(6) Anderson, M. A.; Cudero, A. L.; Palma, J. Capacitive deioniza-
tion as an electrochemical means of saving energy and delivering clean
water. Comparison to present desalination practices: Will it compete?
Electrochim. Acta 2010, 55 (12), 3845–3856.
(7) Oren, Y. Capacitive deionization (CDI) for desalination and
water treatment -- past, present and future (a review). Desalination 2008,
228 (1 3), 10–29.
(8) Pekala, R. W.; Farmer, J. C.; Alviso, C. T.; Tran, T. D.; Mayer,
S. T.; Miller, J. M.; Dunn, B. Carbon aerogels for electrochemical
applications. J. Non-Cryst. Solids 1998, 225 (1), 74–80.
(9) Wang, J.; Angnes, L.; Tobias, H.; Roesner, R. A.; Hong, K. C.;
Glass, R. S.; Kong, F. M.; Pekala, R. W. Carbon aerogel composite
electrodes. Anal. Chem. 1993, 65, 17.
(10) Zhang, S. Q.; Wang, J.; Shen, J.; Deng, Z. S.; Lai, Z. Q.; Zhou, B.;
Attia, S. M.; Chen, L. Y. The investigation of the adsorption character of
carbon aerogels. Nonstruct. Mater. 1999, 11 (3), 375.
(11) Gabelich, C. J.; Tran, T. D.; Suffet, I. H. M. Electrosorption of
Inorganic Salts from Aqueous Solution Using Carbon Aerogels. Environ.
Sci. Technol. 2002, 36 (13), 3010–3019.
(12) Hou, C.-H.; Liang, C.; Yiacoumi, S.; Dai, S.; Tsouris, C.
Electrosorption capacitance of nanostructured carbon-based materials.
J. Colloid Interface Sci. 2006, 302 (1), 54–61.
(13) Yang, K.-L.; Yiacoumi, S.; Tsouris, C. Monte Carlo simulations
of electrical double-layer formation in nanopores. J. Chem. Phys. 2002,
117 (18), 8499–8507.
(14) Liang, C.; Li, Z.; Dai, S. Mesoporous Carbon Materials:
Synthesis and Modification. Angew. Chem., Int. Ed. 2008, 47 (20),
3696–3717.
(15) Lin, C.; Ritter, J. A.; Popov, B. N. Correlation of Double-Layer
Capacitance with the Pore Structure of Sol-Gel Derived Carbon
Xerogels. J. Electrochem. Soc. 1999, 146 (10), 3639–3643.
(16) Yoon, S.; Lee, J.; Hyeon, T.; Oh, S. M. Electric Double-Layer
Capacitor Performance of a New Mesoporous Carbon. J. Electrochem.
Soc. 2000, 147 (7), 2507–2512.
(17) Johnson, A. M.; Newman, J. Desalting by Means of Porous
Carbon Electrodes. J. Electrochem. Soc. 1971, 118 (3), 510–517.
(18) Yang, K.-L.; Ying, T.-Y.; Yiacoumi, S.; Tsouris, C.; Vittoratos,
E. S. Electrosorption of Ions from Aqueous Solutions by Carbon
dx.doi.org/10.1021/es201551e |Environ. Sci. Technol. 2011, 45, 10243–10249
Environmental Science & Technology ARTICLE

Aerogel: An Electrical Double-Layer Model. Langmuir 2001, 17 (6), Structure Inside of Porous Electrodes and Application in the Modeling
1961–1969. of Capacitive Deionization. J. Phys. Chem. Lett. 2009, 1 (1), 205–210.
(19) Gabelich, C. J.; Tran, T. D.; Suffet, I. H. ^a. !. Electrosorption of
Inorganic Salts from Aqueous Solution Using Carbon Aerogels. Environ.
Sci. Technol. 2002, 36 (13), 3010–3019.
(20) Xu, P.; Drewes, J. E.; Heil, D.; Wang, G. Treatment of brackish
produced water using carbon aerogel-based capacitive deionization
technology. Water Res. 2008, 42 (10 11), 2605–2617.
(21) Li, L.; Zou, L.; Song, H.; Morris, G. Ordered mesoporous
carbons synthesized by a modified sol-gel process for electrosorptive
removal of sodium chloride. Carbon 2009, 47 (3), 775–781.
(22) Zou, L.; Li, L.; Song, H.; Morris, G. Using mesoporous carbon
electrodes for brackish water desalination. Water Res. 2008, 42 (8 9),
2340–2348.
(23) Li, H.; Zou, L.; Pan, L.; Sun, Z. Novel Graphene-Like Electro-
des for Capacitive Deionization. Environ. Sci. Technol. 2010, 44 (22),
8692–8697.
(24) Li, H.; Zou, L.; Pan, L.; Sun, Z. Using graphene nano-flakes as
electrodes to remove ferric ions by capacitive deionization. Sep. Purif.
Technol. 2010, 75 (1), 8–14.
(25) Chen, W.; Fan, Z.; Gu, L.; Bao, X.; Wang, C. Enhanced capa-
citance of manganese oxide via confinement inside carbon nanotubes.
Chem. Commun. 2010, 46 (22), 3905–3907.
(26) Gao, Y.; Pan, L.; Li, H.; Zhang, Y.; Zhang, Z.; Chen, Y.; Sun, Z.
Electrosorption behavior of cations with carbon nanotubes and carbon
nanofibres composite film electrodes. Thin Solid Films 2009, 517 (5),
1616–1619.
(27) Li, H.; Pan, L.; Lu, T.; Zhan, Y.; Nie, C.; Sun, Z. A comparative
study on electrosorptive behavior of carbon nanotubes and graphene for
capacitive deionization. J. Electroanal. Chem. 2011, 653 (1 2), 40–44.
(28) Pan, L.; Wang, X.; Gao, Y.; Zhang, Y.; Chen, Y.; Sun, Z. Electro-
sorption of anions with carbon nanotube and nanofibre composite film
electrodes. Desalination 2009, 244 (1 3), 139–143.
(29) Wang, X. Z.; Li, M. G.; Chen, Y. W.; Cheng, R. M.; Huang,
S. M.; Pan, L. K.; Sun, Z. Electrosorption of NaCl Solutions with Carbon
Nanotubes and Nanofibers Composite Film Electrodes. Electrochem.
Solid-State Lett. 2006, 9 (9), E23–E26.
(30) Wang, M.; Huang, Z.-H.; Wang, L.; Wang, M.-X.; Kang, F.;
Hou, H. Electrospun ultrafine carbon fiber webs for electrochemical
capacitive desalination. New J. Chem. 2010, 34 (9), 1843–1845.
(31) Mayes, R. T.; Tsouris, C.; Kiggans, J. O., Jr.; Mahurin, S. M.; De
Paoli, D. W.; Dai, S. Hierarchical ordered mesoporous carbon from
phloroglucinol-glyoxal and its application in capacitive deionization of
brackish water. J. Mater. Chem. 2010, 20 (39), 8674–8678.
(32) Wang, X. Q.; Lee, J. S.; Tsouris, C.; DePaoli, D. W.; Dai, S.
Preparation of activated mesoporous carbons for electrosorption of ions
from aqueous solutions. J. Mater. Chem. 2010, 20 (22), 4602–4608.
(33) Wang, X. Q.; Liang, C. D.; Dai, S. Facile synthesis of ordered
mesoporous carbons with high thermal stability by self-assembly of
resorcinol-formaldehyde and block copolymers under highly acidic
conditions. Langmuir 2008, 24 (14), 7500–7505.
(34) Liang, C.; Dai, S. Synthesis of Mesoporous Carbon Materials
via Enhanced Hydrogen-Bonding Interaction. J. Am. Chem. Soc. 2006,
128 (16), 5316–5317.
(35) Kruk, M.; Jaroniec, M. Gas adsorption characterization of
ordered organic-inorganic nanocomposite materials. Chem. Mater. 2001,
13 (10), 3169–3183.
(36) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. The determination of
pore volume and area distributions in porous substances. I. Computa-
tions from nitrogen isotherms. J. Am. Chem. Soc. 1951, 73, 373–80.
(37) Ruiz, V.; Blanco, C.; Granda, M.; Menendez, R.; Santamaria, R.
Influence of electrode preparation on the electrochemical behaviour of
carbon-based supercapacitors. J. Appl. Electrochem. 2007, 37 (6), 717–721.
(38) Yang, K.-L.; Yiacoumi, S.; Tsouris, C. Electrosorption capaci-
tance of nanostructured carbon aerogel obtained by cyclic voltammetry.
J. Electroanal. Chem. 2003, 540, 159–167.
(39) Zhao, R.; Biesheuvel, P. M.; Miedema, H.; Bruning, H.; van der
Wal, A. Charge Efficiency: A Functional Tool to Probe the Double-Layer

10249 dx.doi.org/10.1021/es201551e |Environ. Sci. Technol. 2011, 45, 10243–10249