Impact of nanofluidic electrolyte on the energy storage capacity in vanadium redox flow battery

Accepted Manuscript Impact of nanofluidic electrolyte on the energy storage capacity in vanadium redox flow battery

Jungmyung Kim, Heesung Park PII:

S0360-5442(18)31286-6

DOI:

10.1016/j.energy.2018.06.221

Reference:

EGY 13263

To appear in:

Energy

Received Date:

20 March 2018

Accepted Date:

29 June 2018

Please cite this article as: Jungmyung Kim, Heesung Park, Impact of nanofluidic electrolyte on the energy storage capacity in vanadium redox flow battery, Energy (2018), doi: 10.1016/j.energy. 2018.06.221

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

1

Impact of nanofluidic electrolyte on the energy storage capacity in vanadium redox flow battery

2 3 4

Jungmyung Kim, Heesung Park*

5

Department of Mechanical Engineering, Changwon National University

6 7

Abstract

8

The limitation of energy storage capacity in vanadium redox flow battery impedes further

9

commercialization of the battery. The concept proposed in this study is to overcome the limit

10

by using nanofluidic electrolytes. Multi-walled carbon nanotubes (MWCNTs) are chosen to

11

disperse in electrolytes due to their high surfaces to volume ratio. Nanofluid electrolytes with

12

three electrolyte weight percent MWCNT (0.05, 0.1, 0.2 wt%) were tested and compared with

13

the pristine electrolyte. Half-cell test with cyclic voltammetry has shown that electrochemical

14

reaction performance is proportional to the content of MWCNT in nanofluidic electrolytes. The

15

redox reaction of nanofluidic electrolytes are enhanced by the increased electrochemical

16

activity and reversibility in addition to the lower polarization effect. Meanwhile, single-cell

17

test reveals that the optimum weight percent of nanofluidic electrolytes is 0.1% of MWCNT

18

because the electrolyte containing 0.2% of MWCNT induces the unwanted precipitation at the

19

electrodes during the electrochemical reaction. After completion of 62 charge/discharge

20

cyclings, nanofluidic electrolyte with 0.1% MWCNT retains specific discharge capacity of

21

31.7 Ah L-1 while pristine electrolyte does 26.0 Ah L-1. This corresponds to 22% enhancement

22

of energy storage by using the nanofluidic electrolytes. We conclude that nanofluidic

23

electrolytes can considerably improve the energy storage capacity with optimized content of

24

MWCNT.

25 26 27

Keywords: Active area, Electrochemical performance, Energy storage capacity, Nanofluidic electrolyte, Vanadium redox flow battery

28 29 30 31 32

Corresponding author Mechanical Engineering Department of Changwon National University, 20 Changwondaehak-ro, Changwon City, 51140 South Korea Telephone: +82-55-213-3609 1

ACCEPTED MANUSCRIPT

33

Email address: [email protected]

2

ACCEPTED MANUSCRIPT

34

1. Introduction

35

The current energy supply policy is facing a problem of inconsistency of power

36

demand/supply and limited conditions of large-scale power utility [1,2]. The energy storage

37

system is a promising technology to tackle the problem by compromising the power demand

38

and supply [3]. Vanadium redox flow battery (VRFB) is a strong candidate for the energy

39

storage system (ESS) application due to the benefits of extensibility, independent capacitance,

40

high energy efficiency, long lifetime, and free from cross-over contamination [4-6]. Various

41

redox couples [7], cell components [8,9], and optimum operating conditions [10,11] have been

42

under investigation to increase electrochemical performance and optimum efficiency.

43

Although the energy storage of VRFB extends to 4–40 MW for the distributed and smart grid

44

applications [12], there still requires a breakthrough technology to overcome low energy

45

storage density of VRFB. Zhang et al. have shown that the discharge capacity of VRFB was

46

enhanced by effectively mixing the electrolyte [13]. Nonetheless, lithium-ion battery offers up

47

to 500 Wh L-1 whereas VRFB does 33 Wh L-1 due to the limited solubility of vanadium species

48

(1-2 M) [14,15]. Therefore, a progressive technology is required to overcome the limitation of

49

energy storage density in VRFB.

50

On the other hand, novel approaches to energy efficiency and capacity enhancement have

51

been investigated, most of which have demonstrated increased electrochemical properties

52

using nano-sized materials. Especially, nanoparticles have drawn attention to increase

53

electrochemically active sites of carbon felt electrode in VRFBs. Bismuth nanoparticles [16]

54

and niodium oxide nanorods [17] were employed to replace noble metal with offering high-

55

performance electrodes for VRFBs. Li et al. [18] demonstrated that single-walled carbon

56

nanotubes served an electrode catalyst for VRFBs. Wu et al. proposed cost effective and a high-

57

performance electrode by growing N-doped carbon nano-spheres on graphite felt fibres [19].

58

They demonstrated the superior performances of energy efficiency and capacity retention by

59

conducting single cell test. Li et al. investigated carbon electrode with hollow nanofibers which

60

provided more active sites, higher pore volume, and short diffusion pathways for ions and

61

electrons resulting in enhanced energy storage [20]. Blasi et al. achieved high electrochemical

62

performance in VRFB by using the electrode based on carbon nanofiber with Mn3O4

63

nanoparticles [21]. Meanwhile, nanofluids have been also used to improve thermal properties,

64

lubrication and combustion characteristics [22-24]. There have been a few attempts [15,25] to

65

develop electroactive nanofluids for VRFB; however, significant changes are inevitable to

66

apply nano-sized material technology to the already commercialized VRFB energy storage 3

ACCEPTED MANUSCRIPT

67

devices. The motivation of our study is to economically and effectively increase

68

electrochemical performance and energy storage capacity by dispersing nano-sized materials

69

in electrolytes used in existing installations. In this regard, nanofluidic electrolyte is

70

manufactured by using multi-walled carbon nanotubes (MWCNT). The liquid produced in the

71

laboratory is a colloidal suspension of MWCNT solid nanoparticles in aqueous electrolyte. We

72

have chosen MWCNTs as nanoparticles due to their high porosity and surface-to-volume ratio,

73

which offers increased electroactive sites resulting in the enhanced electrochemical kinetics of

74

VRFB. The electrochemical performance has been compared between pristine (zero MWCNT)

75

and nanofluidic electrolytes with different MWCNT concentrations. It is shown that the

76

MWCNT nanofluidic electrolytes significantly increase reversibility and active area at the

77

reaction zone. After conducting 62 cycling test by using single cell, nanofluidic electrolyte with

78

0.1 weight percent (wt%) MWCNT is found to retain specific discharge capacity of 31.7 Ah

79

L-1 which is 22% higher than pristine electrolyte.

80 81 82

2. Experimental 2.1 Characterization of materials

83

The nanostructures of MWCNTs were analyzed by field emission transmission electron

84

microscopy (TEM, JEM 2100F, JEOL) and X-ray diffractometer (D8 Discover, BRUKER)

85

operated at 160-200 kV and 40 kV, respectively. Raman spectroscope (JP/NRS-3300, JASCO)

86

equipped with a 100 mW solid-state primary laser at 532 and 785 nm was used to identify the

87

compositions of the materials. The surface morphologies of carbon felt electrodes were

88

inspected by using a field emission scanning electron microscope (SEM, Merlin compact,

89

ZEISS) operating at 5 kV. Optical microscope (JP-E600, Nikon) was also used to visualize the

90

dispersion of MWCNTs in nanofluidic electrolytes.

91 92

2.2 Preparation of nanofluidic electrolytes

93

The pristine electrolyte used in the experiments was prepared by electrochemical

94

decomposition of 1.6 M VOSO4 and 4 M H2SO4 solution [26]. The positive and negative

95

electrolytes were divided into electrolytes of different valence states (V3+/VO2+) after

96

conducting 40 mA cm-2 constant current on 10 charge/discharge cycles. With these electrolytes,

97

different wt% MWCNT nanofluidic electrolyte were produced to secure experimental

98

consistency. Graphene has a hexagonal carbon ring bond of carbon atoms in a two-dimensional

99

plane, and when it is connected in the form of a hollow cylinder, it becomes a carbon nanotube. 4

ACCEPTED MANUSCRIPT

100

There are two types of carbon nanotubes depending on their shape. Having a cylindrical single-

101

layer graphene is Single-walled Carbon nanotube (SWCNT). MWCNT has about 50-cylinder

102

graphene layers [22]. The MWCNTs of outer diameter in the range of 30–50 nm (40-45 walls)

103

with length ranging from 10 to 20 μm were supplied by Carbon Nanotubes Plus, USA, with

104

specifications as listed in Table 1. Fig. 1a and b show the TEM and SEM images of a single

105

MWCNT and MWCNT powder, respectively. The multiple graphitic edge planes of MWCNTs

106

can be observed in the image. As shown in Fig. 1c and d, three different weights of MWCNT

107

powders were homogeneously dispersed by the 490 W sonicator in the pristine electrolytes for

108

comprising 0.05, 0.1 and 0.2 wt% nanofluidic electrolytes, respectively. It should be denoted

109

that the temperature of sonicator was kept at 23 ± 1 °C to prevent unwanted precipitation of

110

the electrolyte as illustrated in Fig. 1e. The dispersion of nanofluidic electrolytes was visually

111

inspected by optical microscopy as displayed in Fig. 1f. It can be seen that the MWCNTs were

112

homogeneously dispersed for all nanofluidic electrolytes.

113

114 115

Fig. 1. Schematic diagram of manufacturing process of nanofluidic electrolytes; (a) TEM (b)

116

SEM images; (c) predefined amounts of MWCNTs; (d) positive and negative electrolytes; (e) 5

ACCEPTED MANUSCRIPT

117

temperature controlled sonicator; (f) dispersed MWCNTs images by taking sample nanofluidic

118

electrolytes after completion of dispersing process.

119 120

Table 1 Details information of multi-walled carbon nanotubes. Item Manufacturer Product number Average particle size Surface area (SSA) Purity Tap density Aspect ratio Electrolytic conductivity

121

Specification Carbon Nanotubes Plus, USA GCM335 30-50 nm diameter, 10-20 μm length >100 m2 /g >98 wt% 0.22 g/cm3 333-400 >100 s/cm

122

2.3 Half-cell test

123 124

The electrochemical properties of the pristine and nanofluidic electrolytes were measured

125

by using cyclic voltammetry (CV, NuVant System, Powerstat-05) test with three electrodes.

126

Electric potential variations were recorded by using a saturated silver-silver chloride double-

127

junction cell as a reference electrode, while the counter and working electrodes were a spiral

128

platinum (Ag/AgCl) and glassy carbon (7.065 mm2), respectively. The volume of electrolyte

129

used in the test was 15 mL for all cases. The applied scan rate was varied from 5 to 200 mV s-

130

1

131

the current density was measured for 50 cycles with the scan rate of 20 mV s-1.

to investigate the energy storage capacity of the nanofluidic electrolytes. For CV cycle test,

132 133

2.4 Single cell test

134

The VRFB single cell with 25 cm2 of active area was composed of polytetrafluoroethylene

135

gaskets, bipolar plates (BLUE 100-BP, Standard Energy), 4.6 mm thickness carbon felt

136

electrodes (SIGRACELL, GFD 4.6 EA, SGL) in thickness of 3 mm after compression, a Nafion

137

membrane (NR-117) and an oxygen-free copper current collecting plates as illustrated in Fig.

138

2 [26]. In order to improve the electrical conductivity, corrosion resistance and brittleness of

139

copper, oxygen-free copper prepared so that the oxygen content is 0.008% or less through

140

deoxidizing agent was used. The cell was connected with a power supply (Sorensen HPD series,

141

AMETEK) and an electronic load (PLZ164W, KIKUSUI) for conducting charge/discharge 6

ACCEPTED MANUSCRIPT

142

cycle test. The electrolyte flow rate was defined to 50 mL min-1 with the volume of electrolyte

143

of 50 mL, while the cell potential was limited between 1.7 V and 0.8 V for charge and discharge,

144

respectively. A specially designed pneumatic fixture applied the constant compression force of

145

0.5 MPa between two endplates. With this, it was confirmed that the sealing of the single cell

146

including fitting parts was secured by preventing unwanted leak of electrolytes. During

147

experiment, purge gas (N2) was supplied to each bottle to eliminate air intrusion. The

148

experiments of charge/discharge cycles were conducted by applying constant electrical current

149

of 120 mA cm-2. After the completion of experiments, the used carbon electrodes were taken

150

and analyzed by SEM.

151

(a)

(b)

PTFE housing

Bipolar plate Membrane

Outlet

Current collector

Inlet

Active area Inlet Outlet

152 153

Fig. 2. (a) Schematic diagram of 25 cm2 single cell. (b) Schematic illustration of MWCNT

154

nanofluidic electrolytes in the active area.

155 156 157

3. Results and discussion 3.1. Physico-chemical analysis

158

The X-ray diffraction patterns of MWCNT and carbon felt electrode are depicted in Fig. 3a.

159

It can be seen that the peak intensities of 2θ = 25.6 ° and 2θ = 23.9 ° reflect the graphite

160

composition of (002) and (100), respectively. Fig. 3b compares the Raman spectra of MWCNT

161

and carbon felt electrode. The D bands (first peak) of the two materials were commonly

162

presented at 1340 cm-1. Since the D band reflected the content of graphitic edge which became

163

the active sites (defects) of the graphite materials [15], it could be expected that the presented

164

D band in MWCNT increased the active sites during the electrochemical reaction in VRFB.

165

Furthermore, the G band (second peak) commonly was found at 1580 cm-1. The intensity ratio

166

of the D band to the G band is an indicator of the amount of defect present in the carbon 7

ACCEPTED MANUSCRIPT

167

materials [27]. It can be seen that the carbon felt electrode had more active sites than the

168

MWCNT.

169

The rheological behavior of the nanofluidic electrolytes at different wt% MWCNT was

170

studied by using glass-capillary viscometer. The measured viscosity was 4.455, 4.785, 5.081

171

and 7.481 mPa s for pristine, 0.05, 0.1 and 0.2 wt% electrolyte, respectively. The increment of

172

viscosity by nanofluidic electrolyte was 7.4 and 14.1 % for 0.05 and 0.1 wt% MWCNT,

173

respectively while it was 68.1 % for 0.2 wt% MWCNT. The viscosities in the range of 4.4 –

174

5.08 mPa s were very close to that measured by Li et al. [28]. Nonetheless, significant increase

175

of viscosity for 0.2 wt% MWCNT might be inappropriate in the practical application.

176 G band

D band

177 178

Fig. 3. (a) X-ray diffraction patterns and (b) Raman spectra of MWCNTs and carbon felt.

179 180

3.2. Electrochemical behavior of nanofluidic electrolyte

181

To investigate the effect of nanofluidic electrolyte on the electrochemical reaction, CV tests

182

were performed for positive (VO2+/VO2+) and negative (V2+/V3+) electrolytes with 0 (pristine),

183

0.05, 0.1, and 0.2 wt% MWCNT, respectively. Fig. 4a and b show the oxidation-reduction CV

184

curves when the scan rate was 5 mV s-1 for the positive and negative electrolytes. It should be

185

denoted that the potential window was defined to 0.2-1.6 V (positive electrolyte) and -1.2-0.4

186

V (negative electrolyte). For both of the positive and negative electrolyte, the measured

187

oxidation-reduction peak current densities were in the order of MWCNT content. The increased

188

peak current density was caused by the improved electron transfer kinetics of VO2+/VO2+ and

189

V2+/V3+ couples [29]. Therefore, it can be seen that the electrochemical activities of the positive 8

ACCEPTED MANUSCRIPT

190

and negative nanofluidic electrolyte were enhanced by increasing MWCNT content. This

191

increase in activity is due to the graphitic region of the MWCNT, which can increase the active

192

area and the active interface provided by conventional electrodes. The increased active sites of

193

nanofluidic electrolyte also improved the oxidation potential from 1.066 V (0 wt%) to 1.028 V

194

(0.05, 0.1, 0.2 wt%). The nanofluidic electrolyte with 0.2 wt% MWCNT exhibited the good

195

electrochemical performance considering large peak current density and small oxidation-

196

reduction potential difference. Furthermore, the peak current density variations with respect to

197

different nanofluidic electrolytes and scan rates are presented in Fig. 4c and d. The peak current

198

density was linearly proportional to the square root of the scan rate by using Randles-Sevcik

199

equation, indicating that the electrochemical reaction at each electrode was controlled by

200

diffusive mass transfer at double-layer [30]. It is clearly observed that the slope of the curve

201

increased with the MWCNT content of the nanofluidic electrolyte due to accelerated mass

202

transfer rate. In this, the nanoparticles in the nanofluidic electrolyte enhanced the

203

electrochemical performance through a double layer mechanism.

204

The peak current densities (Ipox and Ipre, for oxidation and reduction, respectively), ratios of

205

peak current density (Ipox/Ipre) and oxidation-reduction peak potential differences (ΔEp) are

206

summarized in Table 2. Although peak current density increased with MWCNT content in

207

electrolyte, the minimum ratio of peak current density and peak potential difference were 1.38

208

and 0.14 V for 0.1 wt% positive nanofluidic electrolyte. Nonetheless, minimum peak potential

209

difference was presented when 0.05 wt% MWCNT negative nanofluidic electrolyte was used.

210

The impact of nanofluidic electrolyte on the electrochemical reaction could be seen by the

211

experiment; however, the optimal wt% of MWCNT was still unclear. In this regard, CV cycle

212

test was also carried out to further investigate the effect of MWCNT content in electrolyte on

213

the long-term stability.

9

ACCEPTED MANUSCRIPT

30

30

(b)

20

-2

Current density (mA cm )

-2

Current density (mA cm )

(a) 10 0 -10 -20

Pristine 0.05 wt% 0.1 wt% 0.2 wt%

VO2+ VO2 + Reduction

-30 0.2

0.4

0.6

0.8

1.0

1.2

1.4

Voltage (V vs. Ag/AgCl)

160 -2

120

Peak current density (mA cm )

-2

Peak current density (mA cm )

214

0 -10 Pristine 0.05 wt% 0.1 wt% 0.2 wt%

-20

-1.0

-0.8

-0.6

-0.4

6

8

-0.2

0.0

0.2

0.4

10

12

14

16

Voltage (V vs. Ag/AgCl)

80

(c)

140 100 80 60

Pristine 0.05 wt% 0.1 wt% 0.2 wt%

40 20 0 -20 -40 -60 -80 -100

10

-30 -1.2

1.6

Oxidation V2+ V3+

20

0

2

4

6

8

1/2

(Scan rate)

10

-1 1/2

12

14

16

(d)

60 40 20 0 -20 -40

Pristine 0.05 wt% 0.1 wt% 0.2 wt%

-60 -80 -100

0

2

4

1/2

(Scan rate)

(V s )

-1 1/2

(V s )

215

Fig. 4. Cycle voltammograms of the tested electrolytes: (a) positive electrolyte, (b) negative

216

electrolyte at a scan rate of 5 mV s-1. Plot of the peak current density according to Randles-

217

Sevcik equation for (c) positive and (d) negative electrolytes, respectively.

218 219

220

Table 2 Summary of cyclic voltammetry results for the pristine and nanofluidic electrolytes. Sample Ip ox (mA cm-2 ) Positive electrolyte Pristine 19.6 0.05 wt% 27.8 0.1 wt% 31.0 0.2 wt% 32.1 Negative electrolyte Pristine 3.08 0.05 wt% 6.09 0.1 wt% 6.69 0.2 wt% 9.19

Ip re (mA cm-2 )

Ip ox /Ip re

ΔEp (V)

-13.7 -19.9 -22.4 -22.5

1.43 1.40 1.38 1.43

0.22 0.14 0.14 0.14

-9.73 -16.7 -15.9 -24.0

0.32 0.36 0.42 0.38

1.04 0.89 0.89 0.96

221 10

ACCEPTED MANUSCRIPT

222

The electrochemical stability of the nanofluidic electrolytes were investigated by conducting

223

50 charge/discharge cycling test. Fig. 5a shows that the peak current densities of positive

224

pristine electrolyte were considerably reduced for oxidation-reduction reaction after 50

225

charge/discharge cycles, whereas nanofluidic electrolytes with 0.1 and 0.2 wt% MWCNT

226

showed unchanged peak current densities during the cycles. Meanwhile, negative pristine and

227

0.2 wt% nanofluidic electrolytes indicated clear reductions of peak current densities after 50

228

cycles as depicted in Fig. 5b. In this regard, the peak current densities corresponding to the

229

cycle number are displayed in Fig. 5c and d for oxidation reaction. It can be found that the

230

reduction ratios of peak current density were 19.8% (pristine), 6.84% (0.05 wt%), 1.11% (0.1

231

wt%), and 1.61% (0.2 wt%) for positive electrolytes, while the ratios were 50.4% (pristine),

232

12.9% (0.05 wt%), 6.73% (0.1 wt%), and 37.3% (0.2 wt%) for negative electrolytes. The

233

irreversible electrochemical reaction reduced electrochemical activity and peak current density

234

during the cycles. Meanwhile, the MWCNT dispersed in nanofluidic electrolyte retained the

235

activity through the double layer mechanism at the graphitic region of MWCNTs. It is evident

236

that the nanofluidic electrolytes offered more stable electrochemical reaction after 50

237

charge/discharge cycles than pristine electrolyte. The MWCNTs dispersed in electrolyte

238

contributed to increase active sites (see Fig. 3b) resulting in high stable electrochemical

239

reaction. It is also denoted that the positive and negative nanofluidic electrolyte with 0.1 wt%

240

MWCNT showed almost unchanged peak current densities during the cycles indicating

241

minimum irreversible loss.

11

ACCEPTED MANUSCRIPT

60

30

30 20 10

20 -2

Current density (mA cm )

40

-2

Current density (mA cm )

50

(b)

(a)

st

Pristine 1 Cycle th Pristine 50 Cycle st 0.05 wt% 1 Cycle th 0.05 wt% 50 Cycle st 0.1 wt% 1 Cycle th 0.1 wt% 50 Cycle st 0.2 wt% 1 Cycle th 0.2 wt% 50 Cycle

0 -10 -20 -30 0.4

0.8

1.2

Voltage (V vs. Ag/AgCl)

st

Pristine 1 Cycle th Pristine 50 Cycle st 0.05 wt% 1 Cycle th 0.05 wt% 50 Cycle st 0.1 wt% 1 Cycle th 0.1 wt% 50 Cycle st 0.2 wt% 1 Cycle th 0.2 wt% 50 Cycle

-10 -20 -30

-1.2

-0.9

-0.6

-0.3

0.0

0.3

0.6

Voltage (V vs. Ag/AgCl) -2

(c) 55 50 45 40

30 Pristine 0.05 wt% 0.1 wt% 0.2 wt%

25

(d)

20 15 10 5

2+

3+

Pristine 0.05 wt% 0.1 wt% 0.2 wt%

35

0

-40 -1.5

1.6

V V peak current density (mA cm )

2+

242

0.0

60

+

-2

VO VO2 peak current density (mA cm )

-40 -0.4

10

30

0

10

20

30

40

50

Cycle number

0

0

10

20

30

40

50

Cycle number

243

Fig. 5. CV curves of (a) positive electrolytes, and (b) negative electrolytes at the scan rate of

244

20 mV s-1; peak current density variations of (c) positive and (d) negative during oxidation

245

reaction against cycle number.

246 247

3.4. VRFB single cell performance

248

The performance evaluation of VRFB single cell is significant to characterize the effect of

249

nanofluidic electrolyte on the porous electrode. Four different positive and negative

250

electrolytes which contained pristine, 0.05, 0.1 and 0.2 wt% MWCNT were prepared. To

251

prevent contamination between the sample electrolytes, the components of single cell were

252

replaced for each experiment. VRFB single cell performance was evaluated by conducting 62

253

charge/discharge cycles test with a constant current of 120 mA cm-2. The energy storage

254

capacity was evaluated by multiplying the applied current and time at cut-off voltage (0.8 V)

255

for the reservoir volume of electrolyte. Fig. 6a and b display the voltage variations for the

256

nanofluidic electrolytes at first and sixty-second cycles. Comparing Fig. 6a and b, the capacity 12

ACCEPTED MANUSCRIPT

257

retention of the pristine electrolyte was 84.5% after 62 cycles, but the nanofluidic electrolytes

258

showed 90.1, 92.6, and 88.8% capacity changes, respectively. The retention of energy storage

259

capacities was improved when the nanofluidic electrolytes were used. It can be seen that the

260

nanofluidic electrolyte increased the VRFB single cell performance and energy storage

261

capacity during the charge/discharge cycles. This corresponds to the reduced potential

262

difference (decreased polarization effect) as listed in Table 2.

263 1.7

1.7

1.6

1.6 2.0

1.2 1.1

1.0

1.0

1.5

1.3 1.2 1.1

1.0

1.0

(a)

0.9 0.8

Pristine 0.05 wt% 0.1 wt% 0.2 wt%

1.4

B

Cell voltage (V)

1.5

1.3

2.0

1.5

Pristine 0.05 wt% 0.1 wt% 0.2 wt%

1.4

B

Cell voltage (V)

1.5

0

0.5

5

10

15

0.0

20

25

30

0.8

35

0

5

-1

Specific capacity (Ah L )

264

(b)

0.9

0.0

0.5

1.0

0.5

10

15

20

25

0.0

Specific capacity (Ah L )

1.5

2.0

30

35

-1

0.0

0.5

A

A

265

Fig. 6. Voltage variations during charge/discharge cycling test at 120 mA cm-2; (a) 1st cycle,

266

(b) 62nd cycle.

35

42.0

34

40.7

33

39.4

32

38.1

31

36.8

30

35.5 34.2

29 Pristine 0.05 wt% 0.1 wt% 0.2 wt%

27 26 25

0

10

20

32.9 31.6

Energy density (Wh L-1)

Specific discharge capacity (Ah L-1)

267

28

30.3

30

40

50

60

29.0

268

Cycle number

269

Fig. 7. The variations of specific discharge capacity during the single cell cycle test for the

270

pristine and nanofluidic electrolytes. 13

1.0

1.5

ACCEPTED MANUSCRIPT

271

To further investigate electrochemical performance during the cycling test, the variations of

272

the specific discharge capacity and energy density against cycle number are illustrated in Fig.

273

7. There can be seen the significant increase in the specific discharge capacity for nanofluidic

274

electrolytes at the beginning of the cycle test. In addition, the retentions of specific discharge

275

capacity during cycling test were 87, 94, 101 and 90% for pristine, 0.05, 0.1, and 0.2 wt%

276

electrolytes, respectively. This indicates the superior long-term stability of the nanofluidic

277

electrolytes in accordance with CV test depicted in Fig. 5. It is noted that the 0.1 wt%

278

nanofluidic electrolyte retained 22% more specific discharge capacity than pristine electrolyte

279

after 62 cycles. For the pristine electrolyte, the small increase of the capacity (3.3 % of initial

280

value) until 25th cycle was caused by imbalance of vanadium valence states or electrolyte

281

volume change [31]. In this, the capacity at 25th cycle was 30.4, 30.5, 34.1 and 32 AhL-1 for

282

pristine, 0.05, 0.1, and 0.2 wt% electrolyte, respectively. The capacity decay rate from 25th to

283

62nd cycle was 0.4, 0.18, 0.18 and 0.3 % per cycle for the same order of MWCNT concentration

284

electrolyte. It can be found that 0.1 wt% electrolyte produced highest capacity and capacity

285

retention.

286

Meanwhile, the significant fluctuation of capacity during the cycles for 0.2 wt% electrolyte

287

was detected in Fig. 7. In this regard, carbon felt electrodes were sampled after completion of

288

each experiment to investigate the surface morphology. Fig. 8 shows the observed images taken

289

by SEM for the pristine and nanofluidic electrolyte. In particular, considerable coagulations of

290

electrolyte were observed in Fig. 8g and h (0.2 wt% nanofluidic electrolyte), which was enough

291

to impede electrolyte flow as well as decrease active area of the porous electrode. This is the

292

reason why the nanofluidic electrolyte with 0.2 wt% MWCNT presented the significant

293

fluctuation capacity and lower electrochemical performance during the cycle test than the

294

nanofluidic electrolyte with 0.1 wt% MWCNT as shown in Fig. 7. In this, 0.1 wt% nanofluidic

295

electrolyte was optimum to enhance energy storage capacity with stable electrochemical

296

reaction in VRFB. In this experimental analysis, the electrical double layer established at each

297

nanoparticle increased the energy storage capacity. Although MWCNT concentration of 0.1 wt%

298

was used, however, the stored energy at each particle due to electrical double layer was

299

transferred to the porous electrode by diffusion and convection. Consequently, there were two

300

different mechanisms. One is the electron pathway established between electrolyte and

301

electrode interface. The other is direct transportation of each nanoparticles to electrode.

302

Nonetheless, 0.2 wt% MWCNT increased viscosity up to 68.1 % than pristine electrolyte and

303

induced large amount of precipitation at the porous electrode. 14

ACCEPTED MANUSCRIPT

304 (a)

10μm

(e)

305

(b)

10μm

10μm

(g)

(f)

10μm

(c)

10μm

10μm

(d)

10μm

(h)

10μm

306

Fig. 8. SEM images of carbon felts after completion of single cell test; (a, b) pristine electrolyte;

307

(c, d) 0.05 wt% MWCNT electrolyte; (e, f) 0.1 wt% MWCNT electrolyte; (g, h) 0.2 wt%

308

MWCNT electrolyte.

309 310

4. Conclusions

311

In this study, the experimental results were introduced to present the feasibility of nanofluidic

312

electrolyte in VRFB. The MWCNTs were chosen as the nanoparticles to manufacture

313

nanofluidic electrolyte. The physio-chemical analysis of the MWCNTs revealed the increased

314

active sites of the electrochemical reaction. CV test was carried out to investigate the

315

electrochemical reaction kinetics of V2+/V3+ and VO2+/VO2+ vanadium redox couples by using

316

nanofluidic electrolytes. The measured peak current densities and electric potential differences

317

indicated the improved reversibility and polarization effect resulting in enhanced

318

electrochemical performance in the order of MWCNT content. In addition, it is evident that the

319

nanofluidic electrolytes offered more stable electrochemical reaction after 50 charge/discharge

320

cycles than pristine electrolyte. The optimum weight percent of MWCNT in electrolyte was

321

revealed to 0.1 wt% by conducting VRFB single cell test whereas 0.2 wt% nanofluidic

322

electrolyte induced large amount of precipitation which impeded the electrolyte flow and

323

decreased active sites resulting in low electrochemical performance and energy storage

324

capacity. Comparing with pristine electrolyte, 0.1 wt% nanofluidic electrolyte increased energy

325

storage capacity 26.0 to 31.7 Ah L-1 after 62 cycles which corresponded to 22% increase. The

326

nanofluidic electrolytes exhibited a breakthrough that ensures high energy storage capacity and

327

electrochemical efficiency by increasing the reversibility of VRFB, maintaining most of the 15

ACCEPTED MANUSCRIPT

328

existing facilities, and securing energy capacity by a convenient method. In addition, we plan

329

to use our single-cell evaluation of nanoparticles containing various carbon or functional

330

groups.

331 332

Acknowledgments

333

This research was supported by Basic Science Research Program through the National Research

334

Foundation of Korea funded by the Ministry of Education (No. 2015R1D1A3A01019588) and the

335

National Research Foundation of Korea grant funded by the Korea government (No. NRF-

336

2017M1A3A3A02016566).

337 338

16

ACCEPTED MANUSCRIPT

339 340 341

References [1] Ferreira HL, Garde R, Fulli G, Kling W, Lopes JP. Characterisation of electrical energy storage technologies. Energy 2013;53:288–98.

342

[2] Liu J, Zhang JG, Yang Z, Lemmon JP, Imhoff C, Graff GL, et al. Materials science and materials chemistry

343

for large scale electrochemical energy storage: From transportation to electrical grid. Adv Funct Mater

344

2013;23:929–46.

345 346 347 348

[3] Yin C, Gao Y, Guo S, Tang H. A coupled three dimensional model of vanadium redox flow battery for flow field designs. Energy 2014;74:886–95. [4] Wang W, Luo Q, Li B, Wei X, Li L, Yang Z. Recent progress in redox flow battery research and development. Adv Funct Mater 2013;23:970–86.

349

[5] Cunha Á, Brito FP, Martins J, Rodrigues N, Monteiro V, Afonso JL, et al. Assessment of the use of vanadium

350

redox flow batteries for energy storage and fast charging of electric vehicles in gas stations. Energy

351

2016;115:1478–94.

352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375

[6] Skyllas-Kazacos M, Chakrabarti MH, Hajimolana SA, Mjalli FS, Saleem M. Progress in Flow Battery Research and Development. J Electrochem Soc 2011;158:R55. [7] Wang W, Kim S, Chen B, Nie Z, Zhang J, Xia G-G, et al. A new redox flow battery using Fe/V redox couples in chloride supporting electrolyte. Energy Environ Sci 2011;4:4068. [8] Liu Q, Sleightholme AES, Shinkle AA, Li Y, Thompson LT. Non-aqueous vanadium acetylacetonate electrolyte for redox flow batteries. Electrochem Commun 2009;11:2312–5. [9] Wei L, Zhao TS, Zhao G, An L, Zeng L. A high-performance carbon nanoparticle-decorated graphite felt electrode for vanadium redox flow batteries. Appl Energy 2016;176:74–9. [10] Pan J, Huang M, Li X, Wang S, Li W, Ma T, et al. The performance of all vanadium redox flow batteries at below-ambient temperatures. Energy 2016;107:784–90. [11] Badrinarayanan R, Tseng KJ, Soong BH, Wei Z. Modelling and control of vanadium redox flow battery for profile based charging applications. Energy 2017;141:1479–88. [12] Dunn B, Kamath H, Tarascon JM. Electrical energy storage for the grid: A battery of choices. Science (80- ) 2011;334:928–35. [13] Zhang Y, Liu L, Xi J, Wu Z, Qiu X. The benefits and limitations of electrolyte mixing in vanadium flow batteries. Appl Energy 2017;204:373–81. [14] Akinyele DO, Rayudu RK. Review of energy storage technologies for sustainable power networks. Sustain Energy Technol Assessments 2014;8:74–91. [15] Dubal DP, Gomez-Romero P. Electroactive graphene nanofluids for fast energy storage. 2D Mater 2016;3:31004. [16] Li B, Gu M, Nie Z, Shao Y, Luo Q, Wei X, et al. Bismuth nanoparticle decorating graphite felt as a highperformance electrode for an all-vanadium redox flow battery. Nano Lett 2013;13:1330–5. [17] Li B, Gu M, Nie Z, Wei X, Wang C, Sprenkle V, et al. Nanorod niobium oxide as powerful catalysts for an all vanadium redox flow battery. Nano Lett 2014;14:158–65.

376

[18] Li W, Liu J, Yan C. The electrochemical catalytic activity of single-walled carbon nanotubes towards

377

VO2+/VO2+ and V3+/V2+ redox pairs for an all vanadium redox flow battery. Electrochim Acta 17

ACCEPTED MANUSCRIPT

378 379 380 381 382

2012;79:102–8. [19] Wu L, Shen Y, Yu L, Xi J, Qiu X. Boosting vanadium flow battery performance by Nitrogen-doped carbon nanospheres electrocatalyst. Nano Energy 2016;28:19–28. [20] Li L, Peng S, Lee JKY, Ji D, Srinivasan M, Ramakrishna S. Electrospun hollow nanofibers for advanced secondary batteries. Nano Energy 2017;39:111–39.

383

[21] Di Blasi A, Busaccaa C, Di Blasia O, Briguglioa N, Squadritoa G, Antonuccia V. Synthesis of flexible

384

electrodes based on electrospun carbon nanofibers with Mn3O4 nanoparticles for vanadium redox flow

385

battery application. Appl Energy 2017;190:165–71.

386

[22] El-Seesy AI, Abdel-Rahman AK, Bady M, Ookawara S. Performance, combustion, and emission

387

characteristics of a diesel engine fueled by biodiesel-diesel mixtures with multi-walled carbon nanotubes

388

additives. Energy Convers Manag 2017;135:373–93.

389

[23] Al-Shamani AN, Alghoul MA, Elbreki AM, Ammar AA, Abed AM, Sopian K. Mathematical and

390

experimental evaluation of thermal and electrical efficiency of PV/T collector using different water based

391

nano-fluids. Energy 2018;145:770–92.

392

[24] Sadeghinezhad E, Mehrali M, Saidur R, Mehrali M, Tahan Latibari S, Akhiani AR, et al. A comprehensive

393

review on graphene nanofluids: Recent research, development and applications. Energy Convers Manag

394

2016;111:466–87.

395 396 397 398 399 400 401 402 403 404 405 406 407 408

[25] Timofeeva EV, Katsoudas JP, Segre CU, Singh D. Rechargeable Nanofluid Electrodes for High Energy Density Flow Battery. NSTI-Nanotech 2013;2:679–82. [26] Kim J, Park H. Experimental analysis of discharge characteristics in vanadium redox flow battery. Appl Energy 2017;206:451–7. [27] Wang S, Zhao X, Cochell T, Manthiram A. Nitrogen-doped carbon nanotube/graphite felts as advanced electrode materials for vanadium redox flow batteries. J Phys Chem Lett 2012;3:2164–7. [28] Li X, Xiong J, Tang A, Qin Y, Liu J, Yan C. Investigation of the use of electrolyte viscosity for online stateof-charge monitoring design in vanadium redox flow battery. Appl Energy 2018;211;1050-59. [29] Kabtamu DM, Chen J-Y, Chang Y-C, Wang C-H. Water-activated graphite felt as a high-performance electrode for vanadium redox flow batteries. J Power Sources 2017;4:11472–80. [30] Pour N, Kwabi DG, Carney T, Darling RM, Perry ML, Shao-Horn Y. Influence of edge-and basal-plane sites on the vanadium redox kinetics for flow batteries. J Phys Chem C 2015;119:5311–8. [31] Wang K, Liu L, Xi J, Wu Z, Qiu X. Reduction of capacity decay in vanadium flow batteries by an electrolyte-reflow method. J. Power Sources 2017;338;17-25.

18

ACCEPTED MANUSCRIPT

 Energy storage capacity increased by nanofluidic electrolyte.  Electrochemical activity and reversibility increased by MWCNT.  Nanofluidic electrolyte with 0.1 wt% of MWCNT enhances 22 % of specific discharge capacity.