Subscriber access provided by Northern Illinois University
Article
Fluoride Removal from Brackish Groundwaters by Constant Current Capacitive Deionization (CDI) Wangwang Tang, Peter Kovalsky, Baichuan Cao, Di He, and T. David Waite Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b03307 • Publication Date (Web): 08 Sep 2016 Downloaded from http://pubs.acs.org on September 12, 2016
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 33
Environmental Science & Technology
1
Fluoride Removal from Brackish Groundwaters by Constant
2
Current Capacitive Deionization (CDI)
3 4
Wangwang Tang†, Peter Kovalsky†, Baichuan Cao ‡, Di He†, T. David Waite†*
5 6
†
7
NSW 2052, Australia
8
‡
9
Beijing Jiaotong University, Beijing 100044, P. R. China
School of Civil and Environmental Engineering, University of New South Wales, Sydney,
Department of Municipal and Environmental Engineering, School of Civil Engineering,
10
Email addresses:
[email protected] (Wangwang Tang);
[email protected] 11
(Peter Kovalsky);
[email protected] (Baichuan Cao);
[email protected] (Di He);
12
[email protected] (T. David Waite)
13 14 15 16
Environmental Science & Technology
17
(Revised and submitted, September 2016)
18 19 20 21 22
_____________________________________________
23
*Corresponding Author: Professor Trevor David Waite, School of Civil and Environmental
24
Engineering, University of New South Wales, Sydney, NSW 2052, Australia; Phone: +61-2-
25
9385-5060, E-mail:
[email protected] 1 ACS Paragon Plus Environment
Environmental Science & Technology
26
ABSTRACT
27
Charging capacitive deionization (CDI) at constant voltage (CV) produces an effluent stream
28
in which ion concentrations vary with time. Compared to CV, charging CDI at constant
29
current (CC) has several advantages, particularly a stable and adjustable effluent ion
30
concentration. In this work, the feasibility of removing fluoride from brackish groundwaters
31
by single-pass constant-current (SPCC) CDI in both zero-volt and reverse-current desorption
32
modes was investigated and a model developed to describe the selective electrosorption of
33
fluoride and chloride. It was found that chloride is preferentially removed from the bulk
34
solution during charging. Both experimental and theoretical results are presented showing
35
effects of operating parameters, including adsorption/desorption current, pump flow rate and
36
fluoride/chloride feed concentrations, on the effluent fluoride concentration, average fluoride
37
adsorption rate and water recovery. Effects of design parameters are also discussed using the
38
validated model. Finally, we describe a possible CDI assembly in which, under appropriate
39
conditions, fluoride water quality targets can be met. The model developed here adequately
40
describes the experimental results obtained and shows how change in the selected system
41
design and operating conditions may impact treated water quality.
42 43
KEYWORDS
44
Capacitive Deionization, Constant Current, Fluoride Removal, Low-salinity Groundwater,
45
Ion Selectivity
2 ACS Paragon Plus Environment
Page 2 of 33
Page 3 of 33
Environmental Science & Technology
46
INTRODUCTION
47
Capacitive deionization (CDI) is an emerging and fast-growing electrochemical
48
technology which usually employs porous carbon electrodes to remove dissolved and charged
49
species from aqueous solutions during a charging step, followed by ion release to regenerate
50
the electrodes during a discharging step.1,
51
techniques such as reverse osmosis, electrodialysis and distillation, CDI offers the advantages
52
of low-pressure operation, low operating and maintenance costs, high energy efficiency and
53
environmental friendliness, primarily for water with a low to moderate salt content.1-4 Recent
54
research interests in the field of CDI include synthesis of new electrode materials (e.g.,
55
controllable surface functionalities),4-6 design of novel CDI architectures (e.g., flow-electrode
56
CDI)7 and investigation of effects of Faradaic reactions (e.g., hydrogen peroxide generation).8
57
In addition to the removal of major ions from waters, removal of other charged
58
species including cations such as copper and zinc, oxyanions such as arsenic and a range of
59
anions including nitrate and fluoride may be required.2, 9-11 The presence of fluoride (F−) in
60
groundwaters, as a consequence of both anthropogenic and natural processes, is causing
61
increasing concern worldwide12,
62
consumption of the fluoride-contaminated groundwater with excessive concentrations of F−
63
(>1.5 mg L−1) will result in permanent bone and joint deformations, and dental or skeletal
64
fluorosis, although an F− level of 0.5−1.5 mg L−1 has beneficial effects on human teeth and
65
bones, especially for young children.14, 15
13
2
As an alternative to other water treatment
and is the focus of the work described here. Human
66
Removal of single electrolytes (such as NaCl and KCl) using CDI has been widely
67
studied and models that reliably describe ion transport within such electrochemical devices
68
have been developed. Hemmatifar et al.16 formulated and solved the first two-dimensional
69
model for capturing ion adsorption/desorption dynamics in a CDI cell under constant
70
charging voltage. Jande and Kim17-19 developed models for both CDI batch-mode operation
3 ACS Paragon Plus Environment
Environmental Science & Technology
71
and single-pass operation under constant charging voltage or current and examined the effects
72
of various input parameters on ion adsorption and desorption. Porada et al.1 successfully
73
proposed a simplified dynamic CDI transport model for batch-mode operation based on the
74
modified Donnan model and was able to reasonably describe the variation in NaCl
75
concentration over time in the recycle vessel. Kim et al.20 ably described observed dynamic
76
ion transfer and charge storage by including the improved modified Donnan model in the
77
CDI porous electrode transport theory. Due to the similarities of NaF and NaCl, these models
78
can be easily extended to the description of F− removal. However, very few studies have been
79
conducted, either theoretically or experimentally, to investigate F− electrosorption in the
80
presence of other electrolytes such as NaCl as is typical in environmental scenarios. In fact,
81
previous studies have indicated that ion selective removal is possible in CDI even for ions
82
with the same valence and very similar hydrated radius,21-23 not to mention those with
83
different charge and hydrated radius.24, 25
84
In our previous work, we investigated the removal of F− in the presence of NaCl by
85
batch-mode CDI21 and single-pass CDI22, in both cases using constant charging voltage, and
86
developed appropriate models to describe the dynamic outlet concentration of both F− and
87
Cl−. Nevertheless, constant charging voltage may not be the most practical operational mode
88
since the effluent ion concentrations vary greatly in time. To obtain purified water with
89
relatively constant ion concentrations below a specified limit (say, an F− level of 0.5−1.5 mg
90
L−1), it is more advantageous to operate in constant current (CC) mode rather than constant
91
voltage (CV) mode in which a constant charging voltage is applied across the electrodes.
92
In this study, we investigate both the adsorption and desorption behavior of F− and Cl−
93
by single-pass constant-current (SPCC) CDI in both zero-volt desorption (ZVD) and reverse-
94
current desorption (RCD) modes. Also, a much more realistic CDI design was used in this
95
work rather than the single pair and a model was developed which enables prediction of the
4 ACS Paragon Plus Environment
Page 4 of 33
Page 5 of 33
Environmental Science & Technology
96
ion selective transport in a system much closer to reality than that in our previous work.22
97
Effects of various parameters including operating and design parameters on three important
98
performance indicators, namely, the effluent ion concentration, the average fluoride
99
adsorption rate (AFAR) and water recovery (WR) are discussed with theoretical calculations
100
compared to experimental data. Finally, possible combinations of CDI modules able to
101
achieve a specific F− target (0.5−1.5 mg L−1) are presented and limitations of CDI systems
102
operated in the SPCC mode analyzed.
5 ACS Paragon Plus Environment
Environmental Science & Technology
103
MATERIALS AND METHODS
104
CDI Module. The CDI module used in this study (AQUA EWP, USA) and the
105
schematic diagram of the inner structure of the CDI module are displayed in Figures S1 and
106
S2, respectively. The CDI module contains a stack of N = 100 cells in parallel. One cell
107
consists of two graphite sheets as current collectors which are alternatingly positively and
108
negatively biased, and two porous carbon electrodes composed of powdered activated carbon
109
and a polymer binder (polytetrafluoroethylene). The electrode is 10 cm × 10 cm in area and
110
100 µm in thickness with a BET surface area of 1068 m2 g−1. Each carbon electrode pair is
111
separated by a 200 µm thick non-conductive nylon cloth to prevent electrical short circuit and
112
to act as a spacer channel. The water is pumped into the module through an opening located
113
in one of the four corners and flows out from the opening in the middle.
114
Experimental Methods. The schematic diagram of the experimental setup used in
115
SPCC CDI tests is shown in Figure S3. The system consists of a feed vessel, a diaphragm
116
pump (SEAFLO, China), a flowmeter (SCINTEX, Australia), a CDI module, a DC power
117
supply (WEP, Yihua Electronic Equipment Co., Ltd, China), a digital electrical conductivity
118
(EC) meter (F-54, HORIBA, Japan) and an effluent vessel. The cell voltage of the CDI
119
module, measured by digital multimeter (Jaycar Electronics, Australia), should be 0 V prior
120
to each test. The module was fully flushed using the tested feed solution until the effluent
121
conductivity equaled the influent conductivity. During adsorption, a constant electrical
122
current was applied to the module until the cell voltage reached the final charging voltage of
123
1.6 V. At this moment we switched from adsorption step to desorption step. During
124
desorption, we either short-circuited the module (ZVD mode) or applied a constant reverse
125
electrical current until the cell voltage dropped back to 0 V (RCD mode). Preliminary tests
126
indicated that the initial cycle did not differ appreciably from the steady periodic behavior of
127
the system (Figure S4), thus, only the first cycle was examined in the studies described here.
6 ACS Paragon Plus Environment
Page 6 of 33
Page 7 of 33
Environmental Science & Technology
128
Analytical grade sodium chloride (NaCl) and sodium fluoride (NaF) were used for the
129
preparation of feeding solutions. The methods used to determine the effluent ion
130
concentrations as well as the experimental and theoretical specific ion j electrosorption per
131
unit mass of electrodes have been described previously.22
7 ACS Paragon Plus Environment
Environmental Science & Technology
132
Page 8 of 33
MODEL DERIVATION
133
To quantitatively describe the effluent concentration of F− and Cl− within a full cycle,
134
as a function of various geometrical and experimental parameters for SPCC CDI in both ZVD
135
and RCD modes, a plug flow model was developed based on our previous work.21,
136
output of a single charge/discharge cycle was determined by numerically solving the
137
equations describing key processes operating in the SPCC CDI system as outlined below. As
138
shown in Figure 1, the spacer channel of a CDI cell is divided into M (M = 100) sequential
139
sub-cells to describe the ongoing ion transport in the flow direction. In each sub-cell, the
140
solution is considered to be ideally stirred, thus leading to a uniform concentration of the
141
specific ions (F−, Cl− and Na+).20, 26 The porous CDI carbon electrodes are considered to
142
consist of macropores where ions migrate and micropores where electric double layers
143
(EDLs) are largely formed and ions are primarily adsorbed.1 In the macropores, the ion
144
concentration gradient is neglected and the ion concentration assumed to be equal to that in
145
the spacer channel,27 an approximation which may not be particularly accurate when using a
146
full porous electrode transport model.16,
147
model, valid in the limit of strongly overlapped EDLs, is used to represent the EDLs’
148
structure which is characterized by a relatively constant value of diffuse layer potential and
149
ion concentration (non-varying with pore position).1, 31 Moreover, it is assumed that ions are
150
not adsorbed onto the carbon surface within electrodes until a non-zero charging voltage is
151
applied.
152 153 154
28-30
22
The
In the micropores, a modified Donnan (mD)
Based on these assumptions, in the i-th sub-cell, the concentration of ion j in the micropores (,, ) (mM) is given by ,, = , · exp (− ∙ ∆, )
(1)
155
where , is the concentration of ion j (mM) in the macropores and spacer channel, is the
156
ion charge number (+1 for the cation and −1 for the anion) and ∆, is the dimensionless 8 ACS Paragon Plus Environment
Page 9 of 33
Environmental Science & Technology
157
Donnan potential (positive for the anode and negative for the cathode). Meanwhile, in the
158
macropores and spacer channel, local ion electro-neutrality is maintained, i.e.,
159
∑ · , = 0
160
The micropore volumetric ion charge density, , (mM), is expressed as
161 162 163
, = ∑ · ,, = −2 · , · sinh∆,
(2)
(3)
The micropore charge density , relates to the dimensionless Stern potential, ∆!", (positive for the anode and negative for the cathode), according to , ∙ # = −∆!", ∙ $!" ∙ %&
164
(4)
165
where F is the Faraday’s constant (C mol−1), $!" is the volumetric Stern capacity of the cell (F
166
m−3), and VT is the thermal voltage (V). , in the anode is also related to the ion current
167
density ' (mol m−2 s−1) by
168
*
() ∙ *+ , ∙ , = −'
(5)
169
where () is the thickness of single electrode (m), , is the electrode microporosity, ' is
170
defined to be positive during adsorption and to be negative during desorption. Based on Eq.
171
(5), assuming zero charge is maintained at the electrode surface at t=0, we can account for
172
charge accumulation at each moment in time and each sub-cell, ,,+ , according to ,,+ =
173
,,+-*+ − ',+-*+ ∙ .//(() ∙ , ). Meanwhile, the ion current density ' is given by
174
' = '12, + '4, − ', = 512 · 12, + 54 · 4, + 5 · , · (-7 )66 ∙ ∆"8,
175
where ', is the flux of ion j (mol m−2 s−1), as given by an approximation of the Nernst-Planck
176
equation, ' = −5 (. /.9 + ./.9) , under the assumptions of gradient-less
177
concentration profile and a linearized potential profile. Dj is the effective diffusion coefficient
178
of ion j (m2 s−1), ()66 describes the total effective ion transport resistance of the spacer and
179
both electrodes (m).27 ∆"8, is the dimensionless voltage drop driving ion transport and is
180
related to the charging voltage, %:;8
181
(7)
182
where %:;8