- Email: [email protected]
Innovative anode catalyst designed to reduce the degradation in ozone generation via PEM water electrolysis
Accepted Manuscript Innovative anode catalyst designed to reduce the degradation in ozone generation via PEM water electrolysis Jyun-Wei Yu, Guo-Bin Jung, Chi-Wen Chen, Chia-Chen Yeh, Xuan-Vien Nguyen, Chia-Ching Ma, Chung-Wei Hsieh, Cheng-Lung Lin PII:
S0960-1481(17)30335-X
DOI:
10.1016/j.renene.2017.04.028
Reference:
RENE 8723
To appear in:
Renewable Energy
Received Date: 30 January 2017 Revised Date:
9 April 2017
Accepted Date: 13 April 2017
Please cite this article as: Yu J-W, Jung G-B, Chen C-W, Yeh C-C, Nguyen X-V, Ma C-C, Hsieh C-W, Lin C-L, Innovative anode catalyst designed to reduce the degradation in ozone generation via PEM water electrolysis, Renewable Energy (2017), doi: 10.1016/j.renene.2017.04.028. 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
Innovative anode catalyst designed to reduce the degradation in ozone generation via PEM water electrolysis
3 4
Jyun–Wei Yu*, Guo–Bin Jung, Chi–Wen Chen, Chia–Chen Yeh, Xuan–Vien Nguyen, Chia– Ching Ma, Chung–Wei Hsieh, Cheng–Lung Lin
5 6
Department of Mechanical Engineering & Fuel Cell Center, Yuan Ze University, Taoyuan 320, Taiwan;
7
*Correspondence: [email protected]
RI PT
1 2
Abstract
9
Membrane electrode assemblies (MEAs) using commercial PbO2 powder as the anode catalyst
10
to generate ozone via water electrolysis were traditionally adopted. We found that commercial
11
MEAs evinced the typical degradation phenomenon after a current interruption and restart
12
during operation, where the performance degraded and partially recovered after the resumption
13
of current. In this study, homemade MEAs using PbO2 powder and additives were developed,
14
which ameliorated the degradation phenomenon. SEM and XRD analysis were used to compare
15
the anode structure of the homemade to commercial MEAs after short–and long–term operation
16
post–resumption of current after an interruption.
17
Keywords: water electrolysis, ozone generation, membrane electrode assembly, membrane
18
degradation
19
TE D
M AN U
SC
8
1. Introduction
21
Currently, hydrogen is considered the best energy storage carrier, can be used on the renewable
22
to process energy unstably situation, proton exchange membrane (PEM) electrolysis is good for
23
renewable and intermittent power sources. It provides a sustainable solution for the production of
24
hydrogen and has advantage of high voltage efficiency, low operating temperature, whereas
25
disadvantages are high cost of components, and acidic corrosive environment [1]. Raising
26
applied voltage accompanied with proper anode catalyst can be used to generate H2/O2/O3, in
27
addition to H2/O2.
28
AC C
EP
20
（anode)：
2H2O→O2+4H++4e- (1.23V)
1
(1)
ACCEPTED MANUSCRIPT
3H2O→O3+ O2+6H++6e- (1.51V)
29
（cathode）：
30
2H+ + 2e- → H2
(2) (3)
31
（overall reaction）： 2H2O → 2H2 + O2 (O3)
32
Ozone has been applied to process wastewater disinfection or degradation pollution for decades
33
[2,3], ozone is more environment friendly than chlorine. Major methods of producing ozone are
34
high voltage corona discharge, ultraviolet ray, PEM water electrolysis. The advantage of PEM
35
water electrolysis is higher concentration of ozone than the other technologies [4]. Anode
36
material is the most important component to a PEM water electrolysis, particularly anode itself
37
must be able to resist oxidation under high corrosive environment. Commonly anode materials
38
include lead dioxide [5–7], antimony-doped tin dioxide [8], platinum [9–10] and glassy carbon
39
[4, 11–13]. The substrate of the anode material needs characteristic of anti–corrosion. Materials
40
for substrate include stainless steel [14], titanium and ceramics [15–17]. The biggest challenge of
41
PEM water electrolysis for producing ozone is the degradation of membrane electrode assembly
42
(MEA). The reason of degradation was found that intermediate product ·OH will attack catalyst
43
or membrane, leading to performance and MEA life decay [14,18,19]. In addition, current
44
interruption/restart will affect electrochemical reaction balance and result in performance decline
45
[16–17].
46
In this study, homemade MEAs were fabricated and compared with commercial MEA, which
47
usually adopt Pt/Nafion/PbO2 as cathode/electrolyte/anode system. To ameliorate the
48
degradation phenomenon, homemade MEAs using PbO2 powder and additives as anode
49
materials were developed and investigated.
RI PT
SC
M AN U
TE D
EP
2
AC C
50
(4)
Method and Process
51
2.1
Structure of PEM water electrolysis cell
52
Figure 1 shows the structure of the PEM water electrolysis cell, which includes a MEA, gaskets,
53
flow–field plates, and current collectors. The active areas of anode and cathode of MEA were
54
both 9 cm2. 2
RI PT
ACCEPTED MANUSCRIPT
SC
Figure 1. Structure of the PEM water electrolysis cell.
55
2.2
Experimental
57
2.2.1
PEM water electrolysis system
58
The PEM water electrolysis cell and testing system were shown in Fig. 2. Two water tanks were
59
connected to the anode and cathode, respectively.
60 61
AC C
EP
TE D
M AN U
56
Figure 2. Experimental system configuration for measuring performance of PEM water electrolysis cell
62 63
2.2.2 Ozone concentration meter The concentration of generated ozone was determined by measuring the oxidation–reduction
64
potential of the ozone bubbled into water using an ORP–15 digital ozone concentration meter. 3
ACCEPTED MANUSCRIPT
2.2.3
MEA preparation
66
The PbO2/Nafion/additives solutions were prepared by mixing PbO2 powder, additives and
67
Nafion solution. The commercially available Nafion 117 membranes (DuPont Corp.) were
68
combined with the as–prepared anode and a gas diffusion cathode (5 g m−2 Pt catalyst loaded on
69
a carbon structure) to fabricate the MEA.
70
2.2.4. Experiment process
71
The electrolysis cell used to evaluate the MEA performance was composed of a pair of porous
72
titanium plates and two stainless steel end plates holding the titanium plates in place. The
73
assembled MEA was protected in between two pieces of rubber gasket, prior to fixing it in the
74
center of the test hardware (Fig. 1).
75
2.2.5 X–ray diffraction
76
Crystals are composed of atoms arranged with a certain periodicity, and the lattice planes are
77
differentiated by their Miller indices. The Miller indices h, k, l of a lattice plane are related to the
78
intercept of the lattice plane with the three crystallographic axes, and these indices also designate
79
the direction or thogonal to the lattice plane (hkl). When a crystal powder or thin film
80
sample is irradiated by mono–chromatic X–rays, there will be diffraction spots or rings when the
81
incident angle satisfies Bragg's law:
TE D
M AN U
SC
RI PT
65
nλ = 2dsinθ
82
(5)
where is the vertical distance between two adjacent lattice planes, λ is the wavelength of the X–
84
rays, θ is the incident angle, and n is the number or class of diffraction. The diffraction pattern
85
can be used to analyze the atomic structure, which can help determine composition.
86
2.2.6 SEM analysis
87
In scanning electron microscopy (SEM), an electron beam is focused on a sample surface, and
88
the surface is scanned point by point. When the sample is a bulk material or particles, the
89
electron beam may generate secondary electrons, backscattered electrons, or electron donating,
90
which is the most important secondary electron imaging signal. The electron beam energy of 5–
91
35 keV is emitted by an electron gun, which is then focused after passing through the second
92
condenser lens and the objective lens to form a fine electron beam with a certain energy,
AC C
EP
83
4
ACCEPTED MANUSCRIPT
strength, and width. This beam scans the sample surface in a grid pattern with a certain time and
94
space sequence. The focused electron beam interacts with the sample, resulting in secondary
95
electron emission (and other emissions). The secondary electron emission amount changes with
96
the sample surface topography, and it is converted into an electrical signal by the detector. The
97
signal is amplified and displayed on a CRT or computer screen. The resulting secondary electron
98
image reflects the surface topography of the sample.
99
3.
RI PT
93
Results
3.1. Properties of homemade and commercial MEA
101
The specifications of the homemade and commercial MEAs were shown in Table 1. The anode
102
loading in the commercial MEA was four times the amount in the homemade MEA, therefore the
103
thicknesses were significantly thicker. The catalyst proportions and test conditions were shown
104
in Tables 2 and 3, respectively.
Table 1. Specifications of homemade and commercial MEA Thickness Active area Anode catalyst loading (mg/cm2) (mm) (mm2)
105
MEA
0.85 0.52
106 107 108
127 16
Table 2. Proportion of anode catalyst PbO2 Nafion Acid(*) 9 1 5 9 1 5 9 1 5 9 1 10 9 1 15
PTFE(*) 1 3 5 2 3
EP
AC C
Sample A B C D E
30 x 30 30 x 30
TE D
Commercial Homemade
109 110 111
M AN U
SC
100
Table 3. Current interruption/restart test conditions
Sample
Voltage
Test1 Test2 Test3
V 4.5 4.5 4.5
Active time h 12 12 12
Test4
4.5
3
Current interruption time 1min/30min/1h 1min–6cycle/10min/1h 2.5V–1 min/2V–1min/ 1.8V–1 min 3h 5
Current recovery time h 1 1 1 3
ACCEPTED MANUSCRIPT
112
3.2. Performance of homemade and commercial MEA Under operating voltage of 4.5 V, the performances of the homemade and commercial MEA
114
are compared in Fig. 3. The commercial MEA shows a slightly higher current than the
115
homemade MEA (5:1), the best among homemade MEAs. The concentration of generated ozone
116
of commercial MEA shows a slightly higher current than the homemade MEA (5:1) and slightly
117
lower compared to homemade MEA (15:3) as shown in Fig. 3. One possible reason is the
118
generated ozone failed to fully mix with the water in the gas–liquid mixing tower and the high
119
current produced more heat, so the ozone of commercial MEA was broken down due to high
120
temperature.
122
Figure 3. Performance of the homemade and commercial MEAs: (L)current output, (R)c oncentration of generated ozone
EP
121
TE D
M AN U
SC
RI PT
113
3.3. Current interruption/restart test
124
When power was interrupted during ozone generation via an MEA, the performance dropped
125
significantly after the resumption of power [16]. Power interruption/restart comparisons of the
126
homemade and commercial MEAs were shown in Fig. 4. When the commercial MEA restarted,
127
current slightly decreased after power interruptions of 1 min and 10 min, but after a power
128
interruption of 1 h there was a significant decrease in current. In comparison, the homemade
129
MEA (5:1) showed only slight decreases in current after power interruptions of 1 min, 10 min,
130
and 1 h. The difference between the homemade and commercial MEA was particularly
131
significant after the 10 min power interruption. All the homemade MEAs showed good
132
resistance to current drop after power interruption/restart. This means that additives in this study
AC C
123
6
ACCEPTED MANUSCRIPT
successfully inhibited performance after power interruption/restart. Beaufils et al [17] have noted
134
that any interference will change the electrolyzer performance in an electrochemical reaction.
135
After a power interruption, the reaction needs time to restabilize; hence, the performance will
136
decrease transiently. The effect of a power interruption must be related to the chemical or
137
electrochemical reaction on the surface of the electrode and membrane. Ozone concentrations
138
stayed at 1–2 ppm for both the commercial and homemade MEAs. One reason for this result is
139
that ozone for an ambient temperature of around 25 °C and the commercial MEA produced more
140
current, heating the system, which caused thermal decomposition of the ozone.
Figure 4. Performance comparisons of the homemade and commercial MEAs restarted after power interruptions.
EP
141 142
TE D
M AN U
SC
RI PT
133
3.4. 1min/ 10 mins/1 h current interruption
144
Operation mode with several 1–min power interruptions was adopted to observe the effect of
145
repetitive short–term power interruptions. Fig. 5 showed that the commercial MEA suffered no
146
significant drop in performance with six 1–min power interruptions, alternating with 10 min and
147
1 h periods of operation. When the current was interrupted for 10 min in the 7th cycle,
148
performance significantly decreased. When the power interruption was 1 h in the 8th cycle, the
149
performance was even more attenuated. It was found that the performance attenuation was
150
independent of the total operating time and only depended on the interruption time. One possible
151
reason was that the commercial MEA has similar colloidal substances that can protect the
152
catalyst of PbO2 for short power interruptions. In this study, the different proportions of additives
AC C
143
7
ACCEPTED MANUSCRIPT
in catalyst were adopted to mitigate performance drop after current interruption. There were not
154
apparent performance drop of homemade MEAs after each power interruption/restart. This was
155
because the structure of homemade MEAs were porous allowing easy access of water to the
156
catalyst layer, which supported the electrochemical reaction and sustain the current generation.
157
According to Ohba et al. [16] and Beaufils et al. [17], the performance is unlikely to recover
158
100% after current interruption. Results in this study showed that performance of some
159
homemade MEAs remained constant after current interruption. After testing long breaks in
160
power and restart, power supply and power outages in 3 h was to observe changes of the
161
performance in a cycle. The result was similar as that before the power is interrupted several
162
times with a short period, performance of the homemade MEA was similar after a power outage
163
and then restart. The reason that the performances of homemade MEAs remain constant even
164
after long power break and restart was that additive–acid was escaped from the catalyst, leaving
165
more space beneficial for the electrochemical reaction.
166
AC C
EP
TE D
M AN U
SC
RI PT
153
167 168
Figure 5. The performance comparison of the homemade and commercial MEAs after 1–min/ 10 mins/1 h power interruptions.
169
3.5. XRD analysis
170
The XRD analyses of scraped PbO2 from the anode after operation are shown in Fig. 6. The
171
intensity of lattice planes (110) and (101) grew significantly after eight straight hours of 8
ACCEPTED MANUSCRIPT
electrolysis for the homemade MEA, while power interruption and re–starts decreased the lattice
173
plane intensity. According to literature, lattice planes will develop toward a state for ozone
174
formation during the electrolysis of water. However, lattice stability will tend toward another
175
new direction due to the electrochemical or chemical reactions that change after power
176
interruptions, which leads to a reduction in ozone generation of the lattice planes. When the
177
current is restarted, the lattice will slowly increase toward the direction of the ozone; hence,
178
performance will respond slowly. The intensity of the lattice planes grew substantially in the
179
homemade MEA after operation. The signal intensity of the first peak (110) grew relative to the
180
second peak (101) after electrolysis by about a factor of 2. The intensity from lattice planes (110)
181
and (101) were basically the same before the experiment. This suggests that power interruption
182
affected lattice plane (110), but not lattice plane (101) or (200).
183
commercial MEA declined slowly, as found in Figs. 6(a) and 6(b). A possible explanation is that
184
commercial MEA contains some unknown additives and the catalyst amount was much higher
185
than the homemade MEA. Changes in intensity are not as obvious for the commercial MEA, but
186
both the (110) and (101) lattice plane peaks intensify after operation. After restarting after the
187
power interruption, the intensity ratio was approximately 1:1 for the (110) and (101) peaks.
188
These results suggest that ozone generation via PEM water electrolysis is related to the intensity
189
of lattice plane (110).
SC
RI PT
172
AC C
EP
TE D
M AN U
The performance of the
190 191 192
Figure 6. XRD analysis of the homemade and commercial MEAs before interruption/restart test (L) and after interruption/restart test (R) 9
ACCEPTED MANUSCRIPT
Images of the commercial and homemade anode surfaces before and after electrolysis are shown
194
in Fig. 7 and 8, respectively. Figure 7 shows white streaks around the commercial membrane
195
after water electrolysis. After electrolysis, the surface particle of homemade will be attached to
196
the current collector slightly because the structure of homemade is relatively loose and close
197
contact with current collector.
M AN U
SC
RI PT
193
198
Figure 7. The commercial anode surface morphology before and after electrolysis.
201
AC C
200
EP
TE D
199
Figure 8. The homemade anode surface morphology before and after electrolysis.
202 203 204 205 206
TEM images of the homemade MEAs made with acid and PTFE in ratios of 5:1 and 15:3 are shown in Figs. 9 and 10. The PbO2 looked like stacked particles before electrolysis, and presented fine crushing block and dense gatherings after electrolysis relatively, which might make it more difficult for the electrolyte to penetrate the interior and react; PbO2 will produce (OH) during ozone generation via PEM water electrolysis, and the reaction is as follows [9]:
207
PbO2 + H2O→PbO2（·OH）＋H +e− ＋
10
(5)
ACCEPTED MANUSCRIPT
The PbO2 is easily absorbed (–OH) and reacts to generate H2O2 [10,11,18]. The power is strong
209
that MEA contact with current collector around and will cause the higher voltage locally. The
210
reaction of H2O2 thus more likely cause corrosion and loss of function on PbO2, and then easier
211
to produce H2O2 causing further corrosion of PbO2 and loss of function finally.
RI PT
208
212 213 214
SC
215 216
M AN U
217 218
Figure 9. SEM images of homemade MEA with additives acid and PTFE in a ratio of 5:1.
220
Anode before the interruption/restart (L), and anode after the interruption/restart (R).
221
AC C
EP
TE D
219
222
Figure 10. SEM images of homemade MEA with additives acid and PTFE in a ratio of 15:3.
223
Anode before the interruption/restart (L), and anode after the interruption/restart (R).
224
4.
225
In this study, an anode with proportions of PbO2/Nafion of 9:1 and catalyst loading of 30 mg/cm2
226
was utilized with additives for ozone generation. When the ratio of acid and PTFE were 5:1, it
227
showed highest performance among all homemade MEAs andhad a better ability to mitigate
Conclusions
11
ACCEPTED MANUSCRIPT
performance degradation after power interruption and restart. Although commercial MEA had a
229
slightly higher current than homemade with additives (acid: PTFE = 5:1), the catalyst loading
230
was four times higher and showed performance drop after interruption/restart test. The structural
231
difference between the homemade and commercial MEAs was that the porosity of the
232
homemade MEA was larger. Large porosity lets water diffuse in the catalyst layer, which lead to
233
resist decay than commercial MEA in current interruption tests. Although the commercial MEA
234
showed good recovery after short–term current interruptions, performance could not recover to
235
the original levels after long–term current interruption. The homemade MEA performed better
236
than the commercial MEA for both short–term and long–term current interruptions, perhaps
237
because additive–acid was escaped from the catalyst, leaving more space beneficial for the
238
electrochemical reaction.
239
References
240
[1] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, A comprehensive review on PEM water
241
electrolysis, International Journal of Hydrogen Energy 38 (2013) 4901–4934.
242
[2] V. Bocci, Ozone as Janus: this controversial gas can be either toxic or medically useful,
243
Mediators of Inflammation 13 (2004) 3–11.
244
[3] M.A. Khadre, A.E. Yousef, J.G. Kim, Microbiological Aspects of Ozone Applications in
245
Food: A Review, Journal of Food Science 66 (2001) 1242–1252.
246
[4] P.C. Foller, G.H. Kelsall, Ozone generation via the electrolysis of fluoboric acid using glassy
247
carbon anodes and air depolarized cathodes, Journal of Applied Electrochemistry 23 (1993) 996–
248
1010.
249
[5] D.V. Franco, W.F. Jardim, J.F.C. Boodts, L.M. Da Silva, Electrochemical ozone production
250
as an environmentally friendly technology for water treatment, Clean Soil, Air, Water 36 (2008)
251
34–44.
252
[6] P.C. Foller, C.W. Tobias, The effect of electrolyte anion adsorption on current efficiencies
253
for the evolution of ozone, Journal of Physical Chemistry 85 (1981) 3238–3244.
254
[7] S. Chen, F. Jiang, X. Xie, Y. Zhou, X. Hu, Synthesis and application of lead dioxide
255
nanowires for a PEM ozone generator, Electrochimica Acta 192 (2016) 357–362.
256
[8] Y. H. Wang, S. Cheng, K. Y. Chan, Synthesis of ozone from air via a polymer-electrolyte-
257
membrane cell with a doped tin oxide anode, Green Chemistry 8 (2006) 568–572.
AC C
EP
TE D
M AN U
SC
RI PT
228
12
ACCEPTED MANUSCRIPT
[9] M.I. Awad, S. Sata, K. Kaneda, M. Ikematsu, T. Okajima, T. Ohsaka, Ozone
259
electrogeneration at a high current efficiency using a tantalum oxide–platinum composite
260
electrode, Electrochemistry Communications 8 (2006) 1263–1269.
261
[10] A. A. Babak, R. Amadelli, V. N. Fateev, Effect of perfluoro compounds on kinetics of the
262
oxygen and ozone formation at the platinum anode, Russian Journal of Electrochemistry 34
263
(1998) 149–152.
264
[11] P.C. Foller, M.L. Goodwin, ELECTROCHEMICAL GENERATION OF HIGH
265
CONCENTRATION OZONE FOR SMALL-SCALE APPLICATIONS, Ozone: Science and
266
Engineering 6 (1984) 29–36.
267
[12] Y. Nishiki, N. Kitaori, K. Nakamuro, Performances of small-sized generator of ozone-
268
dissolved water using boron–doped diamond electrodes, Ozone: Science and Engineering 33
269
(2011) 114–120.
270
[13] L. M. Da Silva, D. V. Franco, L. A. De Faria, J. F. C. Boodts, Surface, kinetics and
271
electrocatalytic properties of Ti/(IrO 2 + Ta 2O 5) electrodes, prepared using controlled cooling
272
rate, for ozone production, Electrochimica Acta 49 (2004) 3977–3988.
273
[14] L. M. Da Silva, D. V. Franco, J. C. Forti, W. F. Jardim, J. F. C. Boodts, Characterisation of
274
a laboratory electrochemical ozonation system and its application in advanced oxidation
275
processes, Journal of Applied Electrochemistry 36 (2006) 523–530.
276
[15] S. P. Tong, C. A. Ma, H. Feng, A novel PbO2 electrode preparation and its application in
277
organic degradation, Electrochimica Acta 53 (2008) 3002–3006.
278
[16] K. Onda, T. Ohba, H. Kusunoki, S. Takezawa, D. Sunakawa, T. Araki, improving
279
characteristics of ozone water production with multilayer electrodes and operating conditions in
280
a polymer electrolyte water electrolysis cell, Journal of the Electrochemical Society 152 (2005)
281
D177–D183.
282
[17] Y. Beaufils, C. Comninellis, P. Bowen, Preparation and characterization of Ti/IrO2/Pb
283
electrodes for ozone production in a SPE electrochemical cell, pp. 191–200.
284
[18] G. Theis, S. Stucki, R. Kötz, G. Theis, H. Devantay, R. Koetz, H. Devantay, H.J. Christen,
285
H.J. Christen, IN-SITU PRODUCTION OF OZONE IN WATER USING A MEMBREL
286
ELECTROLYZER, Journal of the Electrochemical Society 132 (1985) 367–371.
AC C
EP
TE D
M AN U
SC
RI PT
258
13
ACCEPTED MANUSCRIPT
[19] I. Sirés, E. Brillas, G. Cerisola, M. Panizza, Comparative depollution of mecoprop aqueous
288
solutions by electrochemical incineration using BDD and PbO2 as high oxidation power anodes,
289
Journal of Electroanalytical Chemistry 613 (2008) 151–159.
AC C
EP
TE D
M AN U
SC
RI PT
287
14
ACCEPTED MANUSCRIPT
Highlights
Anode with proportions of PbO2/Nafion of 9:1 and catalyst loading of 30 mg/cm2 was utilized with additives for ozone generation. Ratio of additives acid and PTFE 5:1 leads to highest current generation and better ability to
RI PT
mitigate current degradation after current interruption/restart test.
Homemade MEA performed better than the commercial MEA because additive-acid was
EP
TE D
M AN U
SC
escaped from the catalyst, leaving more space beneficial for the electrochemical reaction.
AC C
S0960-1481(17)30335-X
DOI:
10.1016/j.renene.2017.04.028
Reference:
RENE 8723
To appear in:
Renewable Energy
Received Date: 30 January 2017 Revised Date:
9 April 2017
Accepted Date: 13 April 2017
Please cite this article as: Yu J-W, Jung G-B, Chen C-W, Yeh C-C, Nguyen X-V, Ma C-C, Hsieh C-W, Lin C-L, Innovative anode catalyst designed to reduce the degradation in ozone generation via PEM water electrolysis, Renewable Energy (2017), doi: 10.1016/j.renene.2017.04.028. 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
Innovative anode catalyst designed to reduce the degradation in ozone generation via PEM water electrolysis
3 4
Jyun–Wei Yu*, Guo–Bin Jung, Chi–Wen Chen, Chia–Chen Yeh, Xuan–Vien Nguyen, Chia– Ching Ma, Chung–Wei Hsieh, Cheng–Lung Lin
5 6
Department of Mechanical Engineering & Fuel Cell Center, Yuan Ze University, Taoyuan 320, Taiwan;
7
*Correspondence: [email protected]
RI PT
1 2
Abstract
9
Membrane electrode assemblies (MEAs) using commercial PbO2 powder as the anode catalyst
10
to generate ozone via water electrolysis were traditionally adopted. We found that commercial
11
MEAs evinced the typical degradation phenomenon after a current interruption and restart
12
during operation, where the performance degraded and partially recovered after the resumption
13
of current. In this study, homemade MEAs using PbO2 powder and additives were developed,
14
which ameliorated the degradation phenomenon. SEM and XRD analysis were used to compare
15
the anode structure of the homemade to commercial MEAs after short–and long–term operation
16
post–resumption of current after an interruption.
17
Keywords: water electrolysis, ozone generation, membrane electrode assembly, membrane
18
degradation
19
TE D
M AN U
SC
8
1. Introduction
21
Currently, hydrogen is considered the best energy storage carrier, can be used on the renewable
22
to process energy unstably situation, proton exchange membrane (PEM) electrolysis is good for
23
renewable and intermittent power sources. It provides a sustainable solution for the production of
24
hydrogen and has advantage of high voltage efficiency, low operating temperature, whereas
25
disadvantages are high cost of components, and acidic corrosive environment [1]. Raising
26
applied voltage accompanied with proper anode catalyst can be used to generate H2/O2/O3, in
27
addition to H2/O2.
28
AC C
EP
20
（anode)：
2H2O→O2+4H++4e- (1.23V)
1
(1)
ACCEPTED MANUSCRIPT
3H2O→O3+ O2+6H++6e- (1.51V)
29
（cathode）：
30
2H+ + 2e- → H2
(2) (3)
31
（overall reaction）： 2H2O → 2H2 + O2 (O3)
32
Ozone has been applied to process wastewater disinfection or degradation pollution for decades
33
[2,3], ozone is more environment friendly than chlorine. Major methods of producing ozone are
34
high voltage corona discharge, ultraviolet ray, PEM water electrolysis. The advantage of PEM
35
water electrolysis is higher concentration of ozone than the other technologies [4]. Anode
36
material is the most important component to a PEM water electrolysis, particularly anode itself
37
must be able to resist oxidation under high corrosive environment. Commonly anode materials
38
include lead dioxide [5–7], antimony-doped tin dioxide [8], platinum [9–10] and glassy carbon
39
[4, 11–13]. The substrate of the anode material needs characteristic of anti–corrosion. Materials
40
for substrate include stainless steel [14], titanium and ceramics [15–17]. The biggest challenge of
41
PEM water electrolysis for producing ozone is the degradation of membrane electrode assembly
42
(MEA). The reason of degradation was found that intermediate product ·OH will attack catalyst
43
or membrane, leading to performance and MEA life decay [14,18,19]. In addition, current
44
interruption/restart will affect electrochemical reaction balance and result in performance decline
45
[16–17].
46
In this study, homemade MEAs were fabricated and compared with commercial MEA, which
47
usually adopt Pt/Nafion/PbO2 as cathode/electrolyte/anode system. To ameliorate the
48
degradation phenomenon, homemade MEAs using PbO2 powder and additives as anode
49
materials were developed and investigated.
RI PT
SC
M AN U
TE D
EP
2
AC C
50
(4)
Method and Process
51
2.1
Structure of PEM water electrolysis cell
52
Figure 1 shows the structure of the PEM water electrolysis cell, which includes a MEA, gaskets,
53
flow–field plates, and current collectors. The active areas of anode and cathode of MEA were
54
both 9 cm2. 2
RI PT
ACCEPTED MANUSCRIPT
SC
Figure 1. Structure of the PEM water electrolysis cell.
55
2.2
Experimental
57
2.2.1
PEM water electrolysis system
58
The PEM water electrolysis cell and testing system were shown in Fig. 2. Two water tanks were
59
connected to the anode and cathode, respectively.
60 61
AC C
EP
TE D
M AN U
56
Figure 2. Experimental system configuration for measuring performance of PEM water electrolysis cell
62 63
2.2.2 Ozone concentration meter The concentration of generated ozone was determined by measuring the oxidation–reduction
64
potential of the ozone bubbled into water using an ORP–15 digital ozone concentration meter. 3
ACCEPTED MANUSCRIPT
2.2.3
MEA preparation
66
The PbO2/Nafion/additives solutions were prepared by mixing PbO2 powder, additives and
67
Nafion solution. The commercially available Nafion 117 membranes (DuPont Corp.) were
68
combined with the as–prepared anode and a gas diffusion cathode (5 g m−2 Pt catalyst loaded on
69
a carbon structure) to fabricate the MEA.
70
2.2.4. Experiment process
71
The electrolysis cell used to evaluate the MEA performance was composed of a pair of porous
72
titanium plates and two stainless steel end plates holding the titanium plates in place. The
73
assembled MEA was protected in between two pieces of rubber gasket, prior to fixing it in the
74
center of the test hardware (Fig. 1).
75
2.2.5 X–ray diffraction
76
Crystals are composed of atoms arranged with a certain periodicity, and the lattice planes are
77
differentiated by their Miller indices. The Miller indices h, k, l of a lattice plane are related to the
78
intercept of the lattice plane with the three crystallographic axes, and these indices also designate
79
the direction
80
sample is irradiated by mono–chromatic X–rays, there will be diffraction spots or rings when the
81
incident angle satisfies Bragg's law:
TE D
M AN U
SC
RI PT
65
nλ = 2dsinθ
82
(5)
where is the vertical distance between two adjacent lattice planes, λ is the wavelength of the X–
84
rays, θ is the incident angle, and n is the number or class of diffraction. The diffraction pattern
85
can be used to analyze the atomic structure, which can help determine composition.
86
2.2.6 SEM analysis
87
In scanning electron microscopy (SEM), an electron beam is focused on a sample surface, and
88
the surface is scanned point by point. When the sample is a bulk material or particles, the
89
electron beam may generate secondary electrons, backscattered electrons, or electron donating,
90
which is the most important secondary electron imaging signal. The electron beam energy of 5–
91
35 keV is emitted by an electron gun, which is then focused after passing through the second
92
condenser lens and the objective lens to form a fine electron beam with a certain energy,
AC C
EP
83
4
ACCEPTED MANUSCRIPT
strength, and width. This beam scans the sample surface in a grid pattern with a certain time and
94
space sequence. The focused electron beam interacts with the sample, resulting in secondary
95
electron emission (and other emissions). The secondary electron emission amount changes with
96
the sample surface topography, and it is converted into an electrical signal by the detector. The
97
signal is amplified and displayed on a CRT or computer screen. The resulting secondary electron
98
image reflects the surface topography of the sample.
99
3.
RI PT
93
Results
3.1. Properties of homemade and commercial MEA
101
The specifications of the homemade and commercial MEAs were shown in Table 1. The anode
102
loading in the commercial MEA was four times the amount in the homemade MEA, therefore the
103
thicknesses were significantly thicker. The catalyst proportions and test conditions were shown
104
in Tables 2 and 3, respectively.
Table 1. Specifications of homemade and commercial MEA Thickness Active area Anode catalyst loading (mg/cm2) (mm) (mm2)
105
MEA
0.85 0.52
106 107 108
127 16
Table 2. Proportion of anode catalyst PbO2 Nafion Acid(*) 9 1 5 9 1 5 9 1 5 9 1 10 9 1 15
PTFE(*) 1 3 5 2 3
EP
AC C
Sample A B C D E
30 x 30 30 x 30
TE D
Commercial Homemade
109 110 111
M AN U
SC
100
Table 3. Current interruption/restart test conditions
Sample
Voltage
Test1 Test2 Test3
V 4.5 4.5 4.5
Active time h 12 12 12
Test4
4.5
3
Current interruption time 1min/30min/1h 1min–6cycle/10min/1h 2.5V–1 min/2V–1min/ 1.8V–1 min 3h 5
Current recovery time h 1 1 1 3
ACCEPTED MANUSCRIPT
112
3.2. Performance of homemade and commercial MEA Under operating voltage of 4.5 V, the performances of the homemade and commercial MEA
114
are compared in Fig. 3. The commercial MEA shows a slightly higher current than the
115
homemade MEA (5:1), the best among homemade MEAs. The concentration of generated ozone
116
of commercial MEA shows a slightly higher current than the homemade MEA (5:1) and slightly
117
lower compared to homemade MEA (15:3) as shown in Fig. 3. One possible reason is the
118
generated ozone failed to fully mix with the water in the gas–liquid mixing tower and the high
119
current produced more heat, so the ozone of commercial MEA was broken down due to high
120
temperature.
122
Figure 3. Performance of the homemade and commercial MEAs: (L)current output, (R)c oncentration of generated ozone
EP
121
TE D
M AN U
SC
RI PT
113
3.3. Current interruption/restart test
124
When power was interrupted during ozone generation via an MEA, the performance dropped
125
significantly after the resumption of power [16]. Power interruption/restart comparisons of the
126
homemade and commercial MEAs were shown in Fig. 4. When the commercial MEA restarted,
127
current slightly decreased after power interruptions of 1 min and 10 min, but after a power
128
interruption of 1 h there was a significant decrease in current. In comparison, the homemade
129
MEA (5:1) showed only slight decreases in current after power interruptions of 1 min, 10 min,
130
and 1 h. The difference between the homemade and commercial MEA was particularly
131
significant after the 10 min power interruption. All the homemade MEAs showed good
132
resistance to current drop after power interruption/restart. This means that additives in this study
AC C
123
6
ACCEPTED MANUSCRIPT
successfully inhibited performance after power interruption/restart. Beaufils et al [17] have noted
134
that any interference will change the electrolyzer performance in an electrochemical reaction.
135
After a power interruption, the reaction needs time to restabilize; hence, the performance will
136
decrease transiently. The effect of a power interruption must be related to the chemical or
137
electrochemical reaction on the surface of the electrode and membrane. Ozone concentrations
138
stayed at 1–2 ppm for both the commercial and homemade MEAs. One reason for this result is
139
that ozone for an ambient temperature of around 25 °C and the commercial MEA produced more
140
current, heating the system, which caused thermal decomposition of the ozone.
Figure 4. Performance comparisons of the homemade and commercial MEAs restarted after power interruptions.
EP
141 142
TE D
M AN U
SC
RI PT
133
3.4. 1min/ 10 mins/1 h current interruption
144
Operation mode with several 1–min power interruptions was adopted to observe the effect of
145
repetitive short–term power interruptions. Fig. 5 showed that the commercial MEA suffered no
146
significant drop in performance with six 1–min power interruptions, alternating with 10 min and
147
1 h periods of operation. When the current was interrupted for 10 min in the 7th cycle,
148
performance significantly decreased. When the power interruption was 1 h in the 8th cycle, the
149
performance was even more attenuated. It was found that the performance attenuation was
150
independent of the total operating time and only depended on the interruption time. One possible
151
reason was that the commercial MEA has similar colloidal substances that can protect the
152
catalyst of PbO2 for short power interruptions. In this study, the different proportions of additives
AC C
143
7
ACCEPTED MANUSCRIPT
in catalyst were adopted to mitigate performance drop after current interruption. There were not
154
apparent performance drop of homemade MEAs after each power interruption/restart. This was
155
because the structure of homemade MEAs were porous allowing easy access of water to the
156
catalyst layer, which supported the electrochemical reaction and sustain the current generation.
157
According to Ohba et al. [16] and Beaufils et al. [17], the performance is unlikely to recover
158
100% after current interruption. Results in this study showed that performance of some
159
homemade MEAs remained constant after current interruption. After testing long breaks in
160
power and restart, power supply and power outages in 3 h was to observe changes of the
161
performance in a cycle. The result was similar as that before the power is interrupted several
162
times with a short period, performance of the homemade MEA was similar after a power outage
163
and then restart. The reason that the performances of homemade MEAs remain constant even
164
after long power break and restart was that additive–acid was escaped from the catalyst, leaving
165
more space beneficial for the electrochemical reaction.
166
AC C
EP
TE D
M AN U
SC
RI PT
153
167 168
Figure 5. The performance comparison of the homemade and commercial MEAs after 1–min/ 10 mins/1 h power interruptions.
169
3.5. XRD analysis
170
The XRD analyses of scraped PbO2 from the anode after operation are shown in Fig. 6. The
171
intensity of lattice planes (110) and (101) grew significantly after eight straight hours of 8
ACCEPTED MANUSCRIPT
electrolysis for the homemade MEA, while power interruption and re–starts decreased the lattice
173
plane intensity. According to literature, lattice planes will develop toward a state for ozone
174
formation during the electrolysis of water. However, lattice stability will tend toward another
175
new direction due to the electrochemical or chemical reactions that change after power
176
interruptions, which leads to a reduction in ozone generation of the lattice planes. When the
177
current is restarted, the lattice will slowly increase toward the direction of the ozone; hence,
178
performance will respond slowly. The intensity of the lattice planes grew substantially in the
179
homemade MEA after operation. The signal intensity of the first peak (110) grew relative to the
180
second peak (101) after electrolysis by about a factor of 2. The intensity from lattice planes (110)
181
and (101) were basically the same before the experiment. This suggests that power interruption
182
affected lattice plane (110), but not lattice plane (101) or (200).
183
commercial MEA declined slowly, as found in Figs. 6(a) and 6(b). A possible explanation is that
184
commercial MEA contains some unknown additives and the catalyst amount was much higher
185
than the homemade MEA. Changes in intensity are not as obvious for the commercial MEA, but
186
both the (110) and (101) lattice plane peaks intensify after operation. After restarting after the
187
power interruption, the intensity ratio was approximately 1:1 for the (110) and (101) peaks.
188
These results suggest that ozone generation via PEM water electrolysis is related to the intensity
189
of lattice plane (110).
SC
RI PT
172
AC C
EP
TE D
M AN U
The performance of the
190 191 192
Figure 6. XRD analysis of the homemade and commercial MEAs before interruption/restart test (L) and after interruption/restart test (R) 9
ACCEPTED MANUSCRIPT
Images of the commercial and homemade anode surfaces before and after electrolysis are shown
194
in Fig. 7 and 8, respectively. Figure 7 shows white streaks around the commercial membrane
195
after water electrolysis. After electrolysis, the surface particle of homemade will be attached to
196
the current collector slightly because the structure of homemade is relatively loose and close
197
contact with current collector.
M AN U
SC
RI PT
193
198
Figure 7. The commercial anode surface morphology before and after electrolysis.
201
AC C
200
EP
TE D
199
Figure 8. The homemade anode surface morphology before and after electrolysis.
202 203 204 205 206
TEM images of the homemade MEAs made with acid and PTFE in ratios of 5:1 and 15:3 are shown in Figs. 9 and 10. The PbO2 looked like stacked particles before electrolysis, and presented fine crushing block and dense gatherings after electrolysis relatively, which might make it more difficult for the electrolyte to penetrate the interior and react; PbO2 will produce (OH) during ozone generation via PEM water electrolysis, and the reaction is as follows [9]:
207
PbO2 + H2O→PbO2（·OH）＋H +e− ＋
10
(5)
ACCEPTED MANUSCRIPT
The PbO2 is easily absorbed (–OH) and reacts to generate H2O2 [10,11,18]. The power is strong
209
that MEA contact with current collector around and will cause the higher voltage locally. The
210
reaction of H2O2 thus more likely cause corrosion and loss of function on PbO2, and then easier
211
to produce H2O2 causing further corrosion of PbO2 and loss of function finally.
RI PT
208
212 213 214
SC
215 216
M AN U
217 218
Figure 9. SEM images of homemade MEA with additives acid and PTFE in a ratio of 5:1.
220
Anode before the interruption/restart (L), and anode after the interruption/restart (R).
221
AC C
EP
TE D
219
222
Figure 10. SEM images of homemade MEA with additives acid and PTFE in a ratio of 15:3.
223
Anode before the interruption/restart (L), and anode after the interruption/restart (R).
224
4.
225
In this study, an anode with proportions of PbO2/Nafion of 9:1 and catalyst loading of 30 mg/cm2
226
was utilized with additives for ozone generation. When the ratio of acid and PTFE were 5:1, it
227
showed highest performance among all homemade MEAs andhad a better ability to mitigate
Conclusions
11
ACCEPTED MANUSCRIPT
performance degradation after power interruption and restart. Although commercial MEA had a
229
slightly higher current than homemade with additives (acid: PTFE = 5:1), the catalyst loading
230
was four times higher and showed performance drop after interruption/restart test. The structural
231
difference between the homemade and commercial MEAs was that the porosity of the
232
homemade MEA was larger. Large porosity lets water diffuse in the catalyst layer, which lead to
233
resist decay than commercial MEA in current interruption tests. Although the commercial MEA
234
showed good recovery after short–term current interruptions, performance could not recover to
235
the original levels after long–term current interruption. The homemade MEA performed better
236
than the commercial MEA for both short–term and long–term current interruptions, perhaps
237
because additive–acid was escaped from the catalyst, leaving more space beneficial for the
238
electrochemical reaction.
239
References
240
[1] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, A comprehensive review on PEM water
241
electrolysis, International Journal of Hydrogen Energy 38 (2013) 4901–4934.
242
[2] V. Bocci, Ozone as Janus: this controversial gas can be either toxic or medically useful,
243
Mediators of Inflammation 13 (2004) 3–11.
244
[3] M.A. Khadre, A.E. Yousef, J.G. Kim, Microbiological Aspects of Ozone Applications in
245
Food: A Review, Journal of Food Science 66 (2001) 1242–1252.
246
[4] P.C. Foller, G.H. Kelsall, Ozone generation via the electrolysis of fluoboric acid using glassy
247
carbon anodes and air depolarized cathodes, Journal of Applied Electrochemistry 23 (1993) 996–
248
1010.
249
[5] D.V. Franco, W.F. Jardim, J.F.C. Boodts, L.M. Da Silva, Electrochemical ozone production
250
as an environmentally friendly technology for water treatment, Clean Soil, Air, Water 36 (2008)
251
34–44.
252
[6] P.C. Foller, C.W. Tobias, The effect of electrolyte anion adsorption on current efficiencies
253
for the evolution of ozone, Journal of Physical Chemistry 85 (1981) 3238–3244.
254
[7] S. Chen, F. Jiang, X. Xie, Y. Zhou, X. Hu, Synthesis and application of lead dioxide
255
nanowires for a PEM ozone generator, Electrochimica Acta 192 (2016) 357–362.
256
[8] Y. H. Wang, S. Cheng, K. Y. Chan, Synthesis of ozone from air via a polymer-electrolyte-
257
membrane cell with a doped tin oxide anode, Green Chemistry 8 (2006) 568–572.
AC C
EP
TE D
M AN U
SC
RI PT
228
12
ACCEPTED MANUSCRIPT
[9] M.I. Awad, S. Sata, K. Kaneda, M. Ikematsu, T. Okajima, T. Ohsaka, Ozone
259
electrogeneration at a high current efficiency using a tantalum oxide–platinum composite
260
electrode, Electrochemistry Communications 8 (2006) 1263–1269.
261
[10] A. A. Babak, R. Amadelli, V. N. Fateev, Effect of perfluoro compounds on kinetics of the
262
oxygen and ozone formation at the platinum anode, Russian Journal of Electrochemistry 34
263
(1998) 149–152.
264
[11] P.C. Foller, M.L. Goodwin, ELECTROCHEMICAL GENERATION OF HIGH
265
CONCENTRATION OZONE FOR SMALL-SCALE APPLICATIONS, Ozone: Science and
266
Engineering 6 (1984) 29–36.
267
[12] Y. Nishiki, N. Kitaori, K. Nakamuro, Performances of small-sized generator of ozone-
268
dissolved water using boron–doped diamond electrodes, Ozone: Science and Engineering 33
269
(2011) 114–120.
270
[13] L. M. Da Silva, D. V. Franco, L. A. De Faria, J. F. C. Boodts, Surface, kinetics and
271
electrocatalytic properties of Ti/(IrO 2 + Ta 2O 5) electrodes, prepared using controlled cooling
272
rate, for ozone production, Electrochimica Acta 49 (2004) 3977–3988.
273
[14] L. M. Da Silva, D. V. Franco, J. C. Forti, W. F. Jardim, J. F. C. Boodts, Characterisation of
274
a laboratory electrochemical ozonation system and its application in advanced oxidation
275
processes, Journal of Applied Electrochemistry 36 (2006) 523–530.
276
[15] S. P. Tong, C. A. Ma, H. Feng, A novel PbO2 electrode preparation and its application in
277
organic degradation, Electrochimica Acta 53 (2008) 3002–3006.
278
[16] K. Onda, T. Ohba, H. Kusunoki, S. Takezawa, D. Sunakawa, T. Araki, improving
279
characteristics of ozone water production with multilayer electrodes and operating conditions in
280
a polymer electrolyte water electrolysis cell, Journal of the Electrochemical Society 152 (2005)
281
D177–D183.
282
[17] Y. Beaufils, C. Comninellis, P. Bowen, Preparation and characterization of Ti/IrO2/Pb
283
electrodes for ozone production in a SPE electrochemical cell, pp. 191–200.
284
[18] G. Theis, S. Stucki, R. Kötz, G. Theis, H. Devantay, R. Koetz, H. Devantay, H.J. Christen,
285
H.J. Christen, IN-SITU PRODUCTION OF OZONE IN WATER USING A MEMBREL
286
ELECTROLYZER, Journal of the Electrochemical Society 132 (1985) 367–371.
AC C
EP
TE D
M AN U
SC
RI PT
258
13
ACCEPTED MANUSCRIPT
[19] I. Sirés, E. Brillas, G. Cerisola, M. Panizza, Comparative depollution of mecoprop aqueous
288
solutions by electrochemical incineration using BDD and PbO2 as high oxidation power anodes,
289
Journal of Electroanalytical Chemistry 613 (2008) 151–159.
AC C
EP
TE D
M AN U
SC
RI PT
287
14
ACCEPTED MANUSCRIPT
Highlights
Anode with proportions of PbO2/Nafion of 9:1 and catalyst loading of 30 mg/cm2 was utilized with additives for ozone generation. Ratio of additives acid and PTFE 5:1 leads to highest current generation and better ability to
RI PT
mitigate current degradation after current interruption/restart test.
Homemade MEA performed better than the commercial MEA because additive-acid was
EP
TE D
M AN U
SC
escaped from the catalyst, leaving more space beneficial for the electrochemical reaction.
AC C