RSC Advances

View Article Online

PAPER View Journal | View Issue

Activated Luffa derived biowaste carbon for

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Cite this: RSC Adv., 2019, 9, 14884

enhanced desalination performance in brackish
water†
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

Deepa Sriramulu, Sareh Vafakhah and Hui Ying Yang *

Membrane capacitive deionization (MCDI) is an effective process to remove salt ions from brackish water. In
this work, a systematic investigation was carried out to study the effects of applied potential and salt
concentration on salt adsorption capacity (SAC), charge efficiency (L) and energy consumption in an
MCDI system using Luffa biowaste derived carbon as electrodes. We studied the comparative MCDI
performance of Luffa derived carbon as electrodes before and after activation. Furthermore, the
desalination capacities of the electrodes were quantified by batch-mode experiments in a 2500 mg L
1
NaCl solution at 0.8–1.2 V. Activated Luffa carbon showed a high SAC of 38 mg g
1 at 1.2 V in
a 2500 mg L
1 NaCl solution with a low energy consumption of 132 kJ mol
1 salt as compared to non-
Received 12th March 2019
Accepted 3rd May 2019
activated samples (22 mg g
1, 143 kJ mol
1). The adsorption mechanisms were investigated using kinetic
models and isotherms under various applied potentials. Consequently, the excellent SAC of activated
DOI: 10.1039/c9ra01872g
Luffa carbon can be attributed to the presence of micro/mesoporous network structure formed due to
[Link]/rsc-advances the activation process for the propagation of the salt ions.

infrastructure and long operational lifetimes.1,2,6 However,

1 Introduction industrial adoption of CDI is hampered by low salt adsorption
Currently, we are facing a surge in global demand for freshwater capacity (SAC) of its electrode materials and low charge effi-
and this has put pressure on our limited water supplies.1,2 In ciency of operation. To alleviate these problems, ion-exchange
response, many countries have adopted desalination as membranes are placed in front of electrodes in a format
a means to supplement water supplies. Technologies such as called membrane capacitive deionization (MCDI).10 An anion-
reverse osmosis, thermal distillation and electrodialysis have all exchange membrane placed in front of the anode forces the
been thoroughly developed to effectively desalinate water at an anode to only accept or reject anions whereas a cation-exchange
industrial scale.3–5 Despite this, the high costs of production, membrane placed in front of the cathode does the opposite.
energy requirements and hazardous by-products prevent these This prevents the expulsion of co-ions during electrode
technologies from being fully sustainable.6,7 As such, there is charging11 and improves overall SAC and charge efficiency.
a need to investigate alternative methods of desalination for an Furthermore, a reverse potential can be used to regenerate the
environmentally viable solution. electrodes. This is not possible in traditional CDI since ions
One of the most promising solutions is capacitive deioniza- would be re-adsorbed onto the opposing electrode.
tion (CDI). In a typical CDI process, an external potential is To complement the advantages of MCDI, much research has
imposed across two electrodes as water is allowed to pass been devoted to developing electrode materials with high
through a channel between them. Charged species in water are surface area, better pore size distribution and superior electrical
extracted and stored within electrical double layers (EDLs) at conductivity.12 Among them, carbon aerogels,13,14 graphene
the surfaces of electrodes while de-ionized water is produced. A composites,15–17 carbon nanotubes,18,19 nanobers,20,21 meso-
short-circuit is then applied to regenerate the electrodes for the porous carbons,22,23 activated carbons,24,25 have been studied as
next adsorption phase.8,9 CDI can potentially rival conventional promising electrode materials for MCDI. Unfortunately, most of
desalination methods due to low energy costs, simple these materials were synthesized using petroleum-derived
chemicals, which are limited and non-renewable involving
tedious synthetic process thus enhancing the cost of production
Pillar of Engineering Product Development, Singapore University of Technology and applications.26 Therefore, developing low-cost, eco-friendly
Design, Singapore 487372. E-mail: yanghuiying@[Link] carbon-based electrode materials for high-performance CDI is
† Electronic supplementary information (ESI) available: MCDI setup, full scans
of high importance.
XPS spectrum, high-resolution spectra of O 1s for SDL-A and SDL-C, CV of
SDL-C electrode, adsorption kinetics and adsorption isotherm curve. See DOI:
Biowaste carbons refer to porous carbon materials derived
10.1039/c9ra01872g from naturally biodegradable waste products. These carbon

14884 | RSC Adv., 2019, 9, 14884–14892 This journal is © The Royal Society of Chemistry 2019
 View Article Online

Paper RSC Advances

materials usually inherit the unique porous structures of their (l ¼ 0.154 nm) radiation at 40 kV. Nitrogen adsorption–
biowaste parents and have been shown to be effective for desorption isotherms were recorded at 77 K using a Quantach-
CDI.27–29 For instance, watermelon peel derived carbon28 was rome Autosorb-IQ gas sorption analyzer. The resulting specic
able to deliver a SAC of about 17.38 mg g
1 at 1.2 V in surface area and pore size distribution were determined using
a 500 mg L
1 NaCl solution while pomelo melon derived carbon multipoint Brunauer–Emmett–Teller (BET) and Barrett–Joyner–
delivered a SAC 20.78 mg g
1 at 1.4 V in a 1000 mg L
1 NaCl Halenda (BJH) methods. Raman spectra were obtained using
solution.27 Luffa was also used as biowaste for carbon and a WITec confocal Raman system with 532 nm laser excitation. X-
exhibited a SAC of 20 mg g
1 in a 500 mg L
1 NaCl solution at ray photoelectron spectroscopy (XPS) analysis was performed
1.0 V using an MCDI system. However, the said work did not using PHI-5400 equipment with Al Ka beam source (250 W) and
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

adequately study the effects of potential and salt concentration position-sensitive detector (PSD).
on electrosorption performance. Furthermore, the work did not
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

provide energy consumption results by the system during

adsorption.30 2.4 Electrochemical characterization
Herein, a systematic investigation was carried out to identify The capacitive performance of Luffa derived carbon was studied
the effects of applied potential and salt concentration on SAC, using cyclic voltammetry (CV) and electrochemical impedance
charge efficiency (L) and energy consumption (kJ mol
1 salt spectroscopy (EIS) in a three-electrode conguration in a 1 M
removed) in an MCDI system using Luffa derived biowaste NaCl solution. The three-electrode system consists of a working
carbon. We compared the electrosorption performances of Luffa electrode, a Pt foil as a counter electrode and a standard calomel
derived carbon before (SDL-C) and aer chemical activation electrode (SCE) as a reference. The working electrode was
using KOH31 (SDL-A). Activated materials are known to possess prepared by rst mixing a slurry of 80 wt% active material,
high surface areas suitable for ion adsorption and functional 10 wt% carbon black and 10 wt% polyvinylidene uoride (PVDF)
groups on activated surfaces can improve wettability. Indeed, with N-methyl-2-pyrrolidone (NMP) as a solvent. This slurry was
our experiments show better electrochemical and deionization then coated onto a graphite current collector and dried at 80  C
performances for SDL-A over SDL-C. Low energy consumption overnight to form the working electrode (1  1 cm2, 1.2 mg).
was also recorded for SDL-A when brackish salinities were used. The specic capacitance (C, F g
1) was calculated using CV
Our experimental results highlight the feasibility of using bio- curves obtained from eqn (1) 32 as follows:
waste carbons such as Luffa as electrodes materials for desali- ð
IdV
nation of brackish water. C¼ (1)
2vDVm
where I is the current density (A), DV is the potential window, v
2 Experimental is the scan rate (V s
1) and m is the mass of the active material
2.1 Materials (g). All electrochemical experiments were conducted using
Luffa sponge was sourced from commercial markets. Potassium a Bio-logic VMP3 electrochemical workstation.
hydroxide (KOH) and concentrated hydrochloric acid (HCl,
36%) were purchased from Sigma-Aldrich. All chemicals were of
2.5 Electrosorption experiments
AR grade and used without further purication.
The MCDI cell was constructed as shown in the schematic of
2.2 Preparation of Luffa derived biowaste carbon (SDL-C and Fig. 1. A photograph of the cell and the experimental setup is
SDL-A) also provided in Fig. S1.† A typical MCDI setup consists of a pair
of ion-exchange membranes (Hangzhou Iontech Environmental
In a typical synthesis, Luffa sponge was cut into pieces and
Technology Co. Ltd, China) placed in front of Luffa derived
carbonized under Ar atmosphere at 800  C for 1 h. This sample
carbon electrodes sandwiching a nylon spacer. Electrodes were
was denoted as SDL-C. To activate SDL-C, SDL-C and KOH were
prepared in the same way as the ones used for electrochemical
rst mixed in a mass ratio of 1 : 2 in 30 mL deionized water for
characterization but were of 4  4 cm2 and 200 mm thick. The
10 h then dried in an oven at 80  C. The mixture was then
total mass of active material, carbon black and PVDF was
heated to 800  C at a rate of 5  C min
1 and kept at for 800  C
around 100 mg.
1 h. The product was nally cooled to room temperature before
Batch mode experiments were conducted to evaluate the
it was washed thoroughly with 1 M HCl to remove any inorganic
desalination performance of Luffa derived carbon electrodes. A
salts. The sample was washed for a second time with deionized
NaCl solution (50 mL) was circulated between the MCDI cell
water to remove the acid and dried in an oven at 80  C over-
and a reservoir tank using a peristaltic pump (BT100S, Lead
night. The activated sample was denoted as SDL-A.
Fluid) at a ow rate of 50 mL min
1. The electrical conductivity
was monitored every 10 s at the outlet using a conductivity
2.3 Physical and chemical characterization meter (DDSJ-308F, Leici). Electrosorption experiments were
The morphology of the carbonized samples was observed by conducted by either varying the potential from 0.8 to 1.2 V at
eld-emission scanning electron microscopy (FESEM, JEOL increments of 0.2 V or by changing the NaCl concentration from
JSM-7600F). Powder X-ray diffraction (XRD) patterns were 500 to 2500 mg L
1 at increments of 500 mg L
1. The SAC (G, mg
recorded using a Bruker D8 Advance diffractometer with Cu Ka g
1), specic charge stored during the adsorption (S, C g
1) and

This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 14884–14892 | 14885
 View Article Online

RSC Advances Paper

3 Results and discussion

3.1 Synthesis, structural and morphological
characterization of Luffa derived carbon (SDL-A, SDL-C)
Pristine Luffa sponge consists of an interconnected network of
bers34 (Fig. 1). Aer carbonization and KOH activation, a large
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

porous structure was observed in SEM images of SDL-C and

SDL-A (Fig. 2) due to the inherent brous structure of the Luffa
sponge. These porous well-interconnected structures can serve
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

as efficient pathways for rapid diffusion of ion species.

The surface area and pore structure of SDL carbon before
and aer activation were further investigated using N2 adsorp-
tion–desorption curves as shown in Fig. 3a. SDL-C exhibited
a very low nitrogen uptake and the BET surface area was only 69
m2 g
1 while the pore size distribution ranged between 2–10 nm
Fig. 1 Schematic illustration of the preparation process of SDL-A from which implied the presence of mesopores. On the contrary,
Luffa sponge for MCDI application. SDL-A exhibited a high surface area of 2062 m2 g
1 and a broad
pore size distribution ranging from 0.5–4 nm. N2 isotherm of
SDL-A was a combination of type I and type II isotherms and
charge efficiency (L) were calculated according to eqn (2), (3) featured an increase in N2 adsorption at P/Po < 0.05 caused by
and (4), respectively:33 capillary lling of micropores and a wide knee at higher pres-
ðCo
Ce ÞV sure P/Po > 0.05 which implied the existence of narrow meso-
G¼ (2)
m pores (Fig. 3b). Judging from the high surface area and
Ð hierarchical pore size distribution, we expect SDL-A to provide
Idt more active sites for the adsorption of ions.35,36
S¼ (3)
m The crystallographic structure was determined based on
powder XRD patterns and Raman spectra (Fig. 3c and d). From
G  F
L¼ (4) Fig. 3c, we observed two broad diffraction peaks representing
M  S
(002) and (100) reections of amorphous graphitic carbon in
where Co is the initial NaCl concentration (mg L
1), Ce is the SDL-C and SDL-A.37 There was a shi in (002) peak aer acti-
nal equilibrated NaCl concentration (mg L
1), V is the volume vation from 2q ¼ 23.8 to 28.5 which indicated a decrease in
of NaCl solution (L), and m refers to the total mass of the two the interlayer distance (d002) and an increase in crystallinity.
electrodes (g), I refers to the current during the adsorption Fig. 3d shows the Raman spectra of SDL-C and SDL-A. Both
process (A), F is the Faraday's constant (96 500 C mol
1) and M samples exhibited bands at 1357 cm
1 (D band) and 1601 cm
1
is the molar mass of NaCl (58.5 g mol
1). Another metric, the (G band). D band corresponds to defects and disorders in
salt adsorption rate (SAR) is calculated according to the eqn (5). graphitic structures whereas G band refers to in-plane vibra-
SAC tions of sp2 bonded carbon structures.38 Further, the intensity
SAR ¼ (5) ratio (ID/IG) of D and G bands was used to estimate the graph-
t
itization degree of carbon. ID/IG peak ratios were 0.98 for SDL-A

Fig. 2 SEM images of (a) SDL-C and (b) SDL-A Luffa carbon.

14886 | RSC Adv., 2019, 9, 14884–14892 This journal is © The Royal Society of Chemistry 2019
 View Article Online

Paper RSC Advances

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

Fig. 3 (a) Nitrogen adsorption–desorption isotherms, (b) pore size distribution (inset is the pore size distribution of SDL-C), (c) XRD pattern, (d)
Raman spectra, (e) and (f) are C 1s XPS spectra of SDL-C and SDL-A materials.

and 0.89 for SDL-C which suggested a greater degree of disorder a binding energies of 284.4, 285.7, 286.6 and 288.6 eV corre-
and defects in SDL-A. This result can be attributed to the highly sponding to C–C, C]C, C–O and C]O respectively.23,39 Upon
porous structure of SDL-A. activation with KOH and carbonization, the carbon content
XPS measurements were employed to investigate the chem- SDL-A increased as compared to SDL-C samples. Furthermore,
ical composition of SDL-C and SDL-A carbon. As shown in the high resolution O 1s spectrum shown in Fig. S2† contained
Fig. S2a,† the full spectrum of both SDL-C and SDL-A exhibited three deconvoluted peaks at 531.5 eV (C]O), 532.9 eV (C–O)
two peaks at 284.8, and 351.9 eV corresponding to C 1s and O 1s and 533.6 eV (C–O–C) respectively for both SDL-C and SDL-A.40
respectively. High-resolution C 1s XPS spectrum was also ob- Accordingly, the presence of higher oxygen-containing groups
tained (Fig. 3e and f) and can be deconvoluted into four peaks at on the surface of SDL-A electrodes enhances the wettability of

This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 14884–14892 | 14887
 View Article Online

RSC Advances Paper

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

Fig. 4 Electrochemical performance measured in a three-electrode system with 1 M NaCl electrolyte solution. (a) CV curves of SDL-A at
a different scan rate from 1 mV to 200 mV s
1, (b) CV curves of SDL-C and SDL-A at a scan rate of 5 mV s
1, (c) specific capacitance of SDL-C and
SDL-A at different scan rates and (d) Nyquist profiles of the electrodes under the influence of an ac voltage of 5 mV, inset are the electrical
equivalent circuit used for fitting the impedance spectra of SDL-C and SDL-A.

the electrodes and consequently improves electrochemical electronic elements in an equivalent circuit are shown in Table
performance.41 S1.† The tted Rct values of SDL-C and SDL-A were 61.9 and 4.1
U respectively. The smaller Rct value of SDL-A shows how faster
charge transfer can be achieved at the electrode and electrolyte
3.2 Electrochemical performance interface due to a highly connected porous structure. At the low-
Quasi-rectangular CV curves were observed for SDL-C and SDL-A frequency region, a straight line with the steepest slope was
at different scan rates (Fig. 4a and S3†) which implied a non- observed for SDL-A which implied a low Warburg impedance
ideal capacitive behavior. We also observed a larger CV loop (Zw) and hence, the highest diffusion coefficients. SDL-A also
for SDL-A which meant it possessed a larger capacitance than showed the smallest x-intercept which implied the lowest
SDL-C (Fig. 4b). A summary of specic capacitances is presented internal resistance (Re). Thus, we expect SDL-A to possess the
in Fig. 4c and the highest capacitance obtained by SDL-A was best electrosorption performance.
270 F g
1 as opposed to SDL-C at 125 F g
1. This high capaci-
tance could be attributed to factors such as high surface area,
hierarchical pore size distribution and surface wettability of 3.3 Desalination performance
SDL-A. For both samples, a decrease in specic capacitance A positive electrical potential is applied to the MCDI across the
corresponding to an increase in scan rate was observed due to two electrodes during the charging step to adsorb ions from
insufficient time for the electrolyte to diffuse into the inner a solution of known NaCl concentration and a negative poten-
pores of the electrode material. tial was applied to discharge the electrodes and expel the ions in
Results of our EIS experiments are presented in Fig. 4d. The a waste stream. Fig. 5a shows the desalination performance of
Nyquist plots of SDL-A and SDL-C showed a small semicircle at SDL-A at an initial concentration of 2500 mg L
1 at various
the high-frequency region which corresponded to the charge electrical potentials from 0.8 V to 1.2 V in NaCl solution. In all
transfer resistance (Rct) at the electrode and electrolyte interface applied potentials, a sharp decrease in solution conductivity
while a vertical line observed at the low-frequency region could was rst observed during charging and as adsorption
be attributed to ion diffusion resistance. The parameters of the approaches equilibrium, the change in conductivity decreases

14888 | RSC Adv., 2019, 9, 14884–14892 This journal is © The Royal Society of Chemistry 2019
 View Article Online

Paper RSC Advances

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

Fig. 5 (a) Electrosorption behaviours of the SDL-A electrode at different voltages in NaCl solution with an initial concentration of 2500 mg L
1
with a flow rate of 50 mL min
1, (b) its corresponding current response, (c) salt removal capacity of SDL-A electrode at different voltages ranging
from 0.8 V to 1.4 V in NaCl solution with an initial concentration of 2500 mg L
1 and, (d) deionization and regeneration curves of SDL-A
electrodes in a 2500 mg L
1 NaCl solution at an applied voltage of 1.2 V.

until a plateau is reached. Fig. 5b shows a positively correlated The MCDI performance was further evaluated different initial
current transient with increasing potential while Fig. 5c shows NaCl concentrations ranging from 500 to 2500 mg L
1 at an
the SAC of SDL-A electrode at different applied potentials. From applied voltage of 1.2 V as shown in Fig. 6b. SAC gradually
0.8 V to 1.2 V, the SAC increased from 23 to 38 mg g
1, indi- increases with an increase in the initial concentration with SDL-
cating that a higher cell potential can enhance electrosorption A electrode exhibiting higher SAC as compared to SDL-C. For
capacity. The observed phenomenon can be attributed to the instance, at a concentration of 500 mg L
1, the SAC of SDL-A is
presence of strong electrostatic forces and the formation of 38 mg g
1 and 22 mg g
1 for SDL-C electrodes. With an increase
a thicker EDL at an increased voltage. Fig. 5d shows the cyclic in saline concentration, electrosorption capacity increases due
stability of SDL-A electrode in 2500 mg mL
1 NaCl solution over to the enhancement of ion diffusion inside the pores and
20 charge–discharge cycles without any apparent loss in a reduction of EDL overlapping effects.44 A Ragone plot was also
removal capacity. This result further conrmed that the SDL-A used to understand the deionization performance of the elec-
electrode can be fully regenerated and reused without any trodes in our MCDI system.45 Fig. 6c shows the Ragone plot of
declination, making it a promising choice of electrode material SAR vs. SAC of the SDL-A electrode under different applied
for MCDI applications. potentials. The performance of SDL-A electrode at an applied
Fig. 6a shows the electrosorption capacity and charge effi- potential of 1.2 V shied towards the upper right region, indi-
ciency related to SDL-C and SDL-A. At each applied voltage, the cating that SAR and SAC increase at higher applied potentials.
removal capacity and corresponding charge efficiency of SDL-A The observed increase in SAC and SAR at high potentials was
were higher than SDL-C electrodes. The charge efficiency due to an increase in coulombic interaction between the elec-
increased gradually from 83 to 87% with an increase in applied trode and charged ions, which increases the thickness of EDL
voltage from 0.8 to 1.4 V for SDL-A and 70 to 80% for SDL-C and thus enhancing the desalination performance at high
electrodes. This observation is consistent with reports in liter- voltage.46 Seen in Fig. 6d is the Ragone plot for SDL-C and SDL-A
ature.42,43 The highest SAC for SDL-A was at 38 mg g
1 at electrodes in 2500 mg L
1 NaCl aqueous solution at an applied
a charge efficiency of 80% when the applied potential was 1.2 V. potential of 1.2 V. The performance of SDL-A is marked at the

This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 14884–14892 | 14889
 View Article Online

RSC Advances Paper

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

Fig. 6 Comparison of SAC and charge efficiency of SDL-C and SDL-A electrodes (a) at different applied voltages 0.8–1.2 V, (b) at different initial
concentrations of NaCl solutions at an applied voltage of 1.2 V, Ragone plots of SAR vs. SAC for (c) SDL-A electrode at different applied voltages in
2500 mg L
1 NaCl solution and (d) for SDL-C and SDL-A electrodes in 2500 mg L
1 NaCl solution at an applied voltage of 1.2 V with a flow rate of
50 mL min
1.

upper right corner which indicates an increase in SAC and SAR adsorption kinetics equations as shown below were used for
as compared to the SDL-C electrode. This observation could be our analyses.
attributed to an increase in surface area and development of Pseudo-rst-order
a hierarchical micro/mesoporous structure due to the activation
of carbon. qt ¼ qe(1
exp
k1t) (6)
We had also studied the adsorption kinetics of the adsorp-
tion process.47,48 Pseudo-rst-order and pseudo-second-order

Table 1 Coefficient of pseudo-first and second-order kinetic models

Kinetic models and parameters 0.8 V 1.0 V 1.2 V

qe (exp) (mg g
1) 25.5 28 38

Pseudo-rst-order
qe (cal) (mg g
1) 23.7 26.1 35.15
k1 (min
1) 0.1180 0.1327 0.1120
R2 0.9336 0.9415 0.9427

Pseudo-second-order
qe (cal) (mg g
1) 27.2 29.6 40
k2 (g mg
1 min
1) 0.0053 0.0056 0.0033 Fig. 7 Effect of applied voltage on MCDI energy consumption using
R2 0.9847 0.9904 0.9876 SDL-C and SDL-A electrodes at an initial concentration of
2500 mg L
1 NaCl aqueous solution with a flow rate of 40 mL min
1.

14890 | RSC Adv., 2019, 9, 14884–14892 This journal is © The Royal Society of Chemistry 2019
 View Article Online

Paper RSC Advances

Pseudo-second-order biocarbon as electrodes. The advantage of biocarbon electrode

is the low cost and eco friendly process. The low cost of biowaste
k2 qe t
2
qt ¼ (7) materials refers to easily available in abundance, and are easily
1 þ k2 qe t
recyclable. Further they can be efficiently converted to green
where qe (mg g
1) and qt (mg g
1) are the amounts of ions carbon electrodes by physical or chemical activation. Thus, SDL-
adsorbed at equilibrium and at a certain time t (min) respec- A electrode possessed a larger SAC (38 mg g
1) at 1.2 V in a NaCl
tively. k1 (mg g
1 min
1) and k2 (g mg
1 min
1) are the concentration of 2500 mg L
1 with lower energy consumption
adsorption rate constants of pseudo-rst-order and pseudo- (132 kJ mol
1) as compared to SDL-C electrodes (22 mg g
1,
second-order equations respectively. Fitting between experi- 142 kJ mol
1). Part of this low energy consumption was attrib-
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

mental data and the equations are as shown in Fig. S4a.† The utable to the inclusion of ion exchange membranes.
values of the coefficients (qe, k1, k2) were calculated from the
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

slopes and intercept values of the plots in Fig. S4a† and sum-
marised in Table 1. Pseudo-second order kinetics models t the 4 Conclusions
experimental data the best with regression coefficients for In summary, Luffa derived carbons were successfully prepared
second order kinetics >0.99. using Luffa sponges. The prepared carbon materials exhibited
Adsorption isotherms were also used to predict the electro- a micro/mesoporous hierarchical structure with a high specic
sorption behaviour and predict the maximum SAC of SDL-A surface area of 2062 m2 g
1 and excellent electrochemical
electrodes in NaCl solution.48 The Langmuir isotherm was properties. They were developed as highly efficient electrodes
used to t the experimental data and equations are as shown as for MCDI to remove salt from brackish water. We managed to
follows: achieve a high SAC of 38 mg g
1 in a 2500 mg L
1 of NaCl
Ce 1 Ce solutions at 1.2 V for SDL-A electrodes, which was 72% higher
¼ þ (8)
Qe qm KL qm than that of SDL-C electrode (22 mg g
1). A combination of
meso- and micropores in SDL-A electrode decreases the resis-
where Qe (mg g
1) is the amount of NaCl adsorbed per gram of
tance of NaCl ions when diffusing through the porous carbon
adsorbent, Ce (mg L
1) is the equilibrium concentration, qm (mg
and micropores provide a large surface area for ion adsorption.
g
1) is the maximum adsorption capacity corresponding to
Our results show efficient KOH activated electrodes can be
complete monolayer adsorption, and KL (L mg
1) refers to the
developed for MCDI desalination using low cost biowaste
Langmuir constant. Fig. S4b† shows the electrosorption
products.
isotherm of SDL-A electrode at a cell potential of 1.2 V and
different initial concentrations ranging from 500–2500 mg L
1
 
NaCl solution. The plot of specic adsorption
Ce
against the
Conflicts of interest
Qe
equilibrium concentration shows a good correlation with the The authors declare no competing nancial interest.
Langmuir model. The Langmuir parameters and the regression
coefficients were determined from the slope and the intercepts Acknowledgements
are summarized in ESI Table S2.† Accordingly, qm (45 mg g
1) of
SDL-A calculated from the Langmuir isotherm equation was This research work is supported by the National Research
higher than that of SDL-A electrode (38 mg g
1), implying an Foundation of Singapore, Prime Minister's Office under its
enhanced electrosorption behaviour. Environment & Water Research Programme with Grant No.
The energy consumed by SDL-C and SDL-A electrode to 1301-IRIS-17 and administered by the Environment & Water
remove an amount of salt from saline water at an applied Industry Programme Office (EWI) of the PUB, Singapore's
potential is shown in Fig. 7. The specic energy consumption National Agency.
(kJ mol
1 of salt removed) was calculated using eqn (9) below:
Ð
  Ecell Idt References
Econsumption for salt removal kJ mol
1 ¼ (9)
V ðCo
Ce Þ
1 M. A. Anderson, A. L. Cudero and J. Palma, Electrochim. Acta,
where Ecell (V) is the potential applied to the MCDI system and 2010, 55, 3845–3856.
all other parameters were as dened previously. Energy 2 A. Subramani and J. G. Jacangelo, Water Res., 2015, 75, 164–
required to remove an ion (kT per ion) was calculated based on 187.
a literature reported elsewhere.11 As observed in Fig. 7, specic 3 S. Lattemann, M. D. Kennedy, J. C. Schippers and G. Amy, in
energy consumption increased with an increase in the applied Sustainability Science and Engineering, ed. I. C. Escobar and A.
potential for both SDL-A and SDL-C electrodes, which is in I. Schäfer, Elsevier, 2010, vol. 2, pp. 7–39.
accordance with previous reported work. The energy 4 L. F. Greenlee, D. F. Lawler, B. D. Freeman, B. Marrot and
consumption obtained for Luffa derived carbons was around P. Moulin, Water Res., 2009, 43, 2317–2348.
18–28 kT, which is comparable with values reported for 5 H. T. El-Dessouky, H. M. Ettouney and Y. Al-Roumi, Chem.
commercial MCDI.11 Even though the energy consumption of Eng. J., 1999, 73, 173–190.
commercial MCDI is comparable with our MCDI system using 6 Y. Oren, Desalination, 2008, 228, 10–29.

This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 14884–14892 | 14891
 View Article Online

RSC Advances Paper

7 Y. Huajie, Z. Shenlong, W. Jiawei, T. Hongjie, C. Lin, 28 S. Zhao, T. Yan, Z. Wang, J. Zhang, L. Shi and D. Zhang, RSC
H. Liangcan, Z. Huijun, G. Yan and T. Zhiyong, Adv. Adv., 2017, 7, 4297–4305.
Mater., 2013, 25, 6270–6276. 29 G.-X. Li, P.-X. Hou, S.-Y. Zhao, C. Liu and H.-M. Cheng,
8 S. Porada, L. Borchardt, M. Oschatz, M. Bryjak, J. S. Atchison, Carbon, 2016, 101, 1–8.
K. J. Keesman, S. Kaskel, P. M. Biesheuvel and V. Presser, 30 C. Feng, Y.-A. Chen, C.-P. Yu and C.-H. Hou, Chemosphere,
Energy Environ. Sci., 2013, 6, 3700–3712. 2018, 208, 285–293.
9 J. Choi, P. Dorji, H. K. Shon and S. Hong, Desalination, 2019, 31 A. N. Wennerberg and T. M. O'Grady, US Pat., US4082694A,
449, 118–130. 1978.
10 M. D. Andelman and G. S. Walker, US Pat., US6709560B2, 32 M. Fatnassi and M. Es-Souni, RSC Adv., 2015, 5, 21550–
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

2004. 21557.
11 R. Zhao, P. M. Biesheuvel and A. van der Wal, Energy Environ. 33 T. Wu, G. Wang, Q. Dong, F. Zhan, X. Zhang, S. Li, H. Qiao
Open Access Article. Published on 14 May 2019. Downloaded on 9/18/2025 8:44:24 AM.

Sci., 2012, 5, 9520. and J. Qiu, Environ. Sci. Technol., 2017, 51, 9244–9251.
12 B. Jia and W. Zhang, Nanoscale Res. Lett., 2016, 11, 64. 34 J. Shen, Y. Min Xie, X. Huang, S. Zhou and D. Ruan, J. Mech.
13 H.-H. Jung, S.-W. Hwang, S.-H. Hyun, K.-H. Lee and Behav. Biomed. Mater., 2012, 15, 141–152.
G.-T. Kim, Desalination, 2007, 216, 377–385. 35 K. Sun, S. Yu, Z. Hu, Z. Li, G. Lei, Q. Xiao and Y. Ding,
14 C. J. Gabelich, T. D. Tran and I. H. M. Suffet, Environ. Sci. Electrochim. Acta, 2017, 231, 417–428.
Technol., 2002, 36, 3010–3019. 36 X. Gu, C. Lai, F. Liu, W. Yang, Y. Hou and S. Zhang, J. Mater.
15 W. Shi, H. Li, X. Cao, Z. Y. Leong, J. Zhang, T. Chen, H. Zhang Chem. A, 2015, 3, 9502–9509.
and H. Y. Yang, Sci. Rep., 2016, 6, 18966. 37 W. Qian, F. Sun, Y. Xu, L. Qiu, C. Liu, S. Wang and F. Yan,
16 Y. Zhi-Yu, J. Lin-Jian, L. Guo-Qian, X. Qing-Qing, Z. Yu-Xia, Energy Environ. Sci., 2014, 7, 379–386.
J. Lin, Z. Xiao-Xue, Y. Yi-Ming and S. Ke-Ning, Adv. Funct. 38 F. Niu, L.-M. Tao, Y.-C. Deng, Q.-H. Wang and W.-G. Song,
Mater., 2014, 24, 3917–3925. New J. Chem., 2014, 38, 2269–2272.
17 Z. Y. Leong, G. Lu and H. Y. Yang, Desalination, 2019, 451, 39 H. Zhu, X. Wang, X. Liu and X. Yang, Adv. Mater., 2012, 24,
172–181. 6524–6529.
18 L. Wang, M. Wang, Z.-H. Huang, T. Cui, X. Gui, F. Kang, 40 J. Yang, F. Chen, C. Li, T. Bai, B. Long and X. Zhou, J. Mater.
K. Wang and D. Wu, J. Mater. Chem., 2011, 21, 18295–18299. Chem. A, 2016, 4, 14324–14333.
19 K. Shi, M. Ren and I. Zhitomirsky, ACS Sustainable Chem. 41 X. Yang, M. Li, N. Guo, M. Yan, R. Yang and F. Wang, RSC
Eng., 2014, 2, 1289–1298. Adv., 2016, 6, 4365–4376.
20 A. G. El-Deen, N. A. M. Barakat, K. A. Khalil and H. Y. Kim, J. 42 T. Kim, J. E. Dykstra, S. Porada, A. van der Wal, J. Yoon and
Mater. Chem. A, 2013, 1, 11001–11010. P. M. Biesheuvel, J. Colloid Interface Sci., 2015, 446, 317–326.
21 M. Ding, K. K. R. Bannuru, Y. Wang, L. Guo, A. Baji and 43 R. Zhao, P. M. Biesheuvel, H. Miedema, H. Bruning and
H. Y. Yang, Adv. Mater. Technol., 2018, 3, 1800135. A. van der Wal, J. Phys. Chem. Lett., 2010, 1, 205–210.
22 L. Zou, L. Li, H. Song and G. Morris, Water Res., 2008, 42, 44 G.-X. Li, P.-X. Hou, Z. Shiyong, C. Liu and H.-M. Cheng,
2340–2348. Carbon, 2016, 101, 1–8.
23 L. Li, L. Zou, H. Song and G. Morris, Carbon, 2009, 47, 775– 45 T. Kim and J. Yoon, RSC Adv., 2015, 5, 1456–1461.
781. 46 H. Li, L. Pan, C. Nie, Y. Liu and Z. Sun, J. Mater. Chem., 2012,
24 M.-W. Ryoo and G. Seo, Water Res., 2003, 37, 1527–1534. 22, 15556–15561.
25 J.-H. Choi, Sep. Purif. Technol., 2010, 70, 362–366. 47 X. Xu, L. Pan, Y. Liu, T. Lu, Z. Sun and D. H. C. Chua, Sci.
26 S. Porada, R. Zhao, A. van der Wal, V. Presser and Rep., 2015, 5, 8458.
P. M. Biesheuvel, Prog. Mater. Sci., 2013, 58, 1388–1442. 48 Y. Liu, J. Ma, T. Lu and L. Pan, Sci. Rep., 2016, 6, 32784.
27 D. Xu, Y. Tong, T. Yan, L. Shi and D. Zhang, ACS Sustainable
Chem. Eng., 2017, 5, 5810–5819.

14892 | RSC Adv., 2019, 9, 14884–14892 This journal is © The Royal Society of Chemistry 2019