Int. J. Electrochem. Sci., 12 (2017) 6170 – 6179, doi: 10.20964/2017.07.

72

International Journal of
ELECTROCHEMICAL
SCIENCE
www.electrochemsci.org

Development of An Electrochemical Impedance Immunosensor

for Myoglobin Determination
Libo Sun1, Wenwen Li2, Maojing Wang3, Wei Ding4 and Yang Ji3,*
1
Department of Pharmacy, The Affiliated Hospital of Qingdao University, Qingdao, Shandong, P.R.
China
2
Department of Hematology, Qingdao Commercial staff Hospital, Qingdao, Shandong, P.R. China
3
Department of Cardiology, The Affiliated Hospital of Qingdao University, Qingdao, Shandong, P.R.
China
4
Ophthalmology Department, Huangdao District People’s Hospital, P.R. China
*
E-mail: [email protected]

Received: 28 March 2017 / Accepted: 6 May 2017 / Published: 12 June 2017

This work addressed the fabrication of a bioelectrode via the attachment of monolayer-protected Au
nanoparticles (AuNPs) onto a 3-aminopropyltriethoxysilane (APTES) self-assembled monolayer
(SAM) on an indium-tin-oxide (ITO). An anti-myoglobin antibody (Ab-Mb) was then linked to the
AuNPs with a carbodiimide coupling reaction. The terminal bioelectrode was termed as Ab-
Mb/AuNPs/APTES/ITO. Electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV)
were employed for the characterization of the proposed bioelectrode, which displayed an
electrochemical impedance response to Ag-Mb in phosphate buffer solution (PBS) with a linear range
of 10 ng/mL-1 μg/mL and an LOD of 2.7 ng/mL.

Keywords: Myoglobin; Gold nanoparticles; Impedance immunosensor; Electrochemical impedance

spectroscopy; Bioelectrode

1. INTRODUCTION

Cardiac myoglobin (Mb) is among the earliest biomolecules released into blood circulation
after acute myocardial infarction (AMI) and is, therefore, an early marker for AMI. The level of an
early marker within 6 h after symptoms onset and a definitive marker with remarkable sensitivity and
specificity in blood could be employed for AMI diagnostics, as recommended by the National
Academy of Clinical Biochemistry (NACB) [1, 2]. Myoglobin is rapidly released, circulated and
shows an increase in concentration due to its small size (17.8 kDa). It is unusual for the serum
Int. J. Electrochem. Sci., Vol. 12, 2017 6171

concentration of myoglobin to exceed 110 μg/L, and this is an indication for AMI. Although
performing well in sensitivity with an LOD of no less than 1 ng.mL−1, ELISA and other cardiac
markers detection assays are considered improper for rapid decision-making with respect to heart
attack, since these assays require no less than 6 h to conduct, and they must be performed in central
laboratories of hospitals and clinics. In addition, a large number of reagents and specimens are required
in these large-scale assays. Early, on-site diagnostics with desirable accuracy aid doctors in their
diagnoses. Novel techniques need to be developed for fast biomarker determinations. In terms of on-
site diagnostics, electrochemical biosensors/immunosensors are a present area of interest. Featuring
cost-effectiveness, facile miniaturization and desirable specificity, and sensitivity and simplicity,
electrochemical immunosensors have gained widespread recognition as a promising analysis
technique.
Being portable, highly efficient, easily usable, possible to be miniaturized, rapid in response
time, and direct in transduction (biomolecular recognition events directly leading to electronic signals),
electrochemical immunosensors have triggered a revolution in contemporary chemical analysis as a
significant sensing system. Electrochemical immunosensors have gained extensive use in biological-
warfare agents, biological toxins, biomarker and protein detection in the fields of pharmaceutical
chemistry, environment, food and clinical diagnosis applications since they are significantly cost-
effective, selective, sensitive and accurate [3-7]. On the basis of glassy carbon electrode (GCE)
modified by methylene blue-multiwalled carbon nanotubes (MWCNT) nanohybrid, an Mb-targeted
electrochemical sensor was proposed by Pakapongpan and co-workers [8], with direct electrochemical
reduction of Mb as the basis of determination. The linear range of Mb determination was as extensive
as 0.1 μM-3.0 μM (∼1.78 μg mL−1-53.40 μg mL−1) with respect to the proposed biosensor. It is,
however, still beyond the physiological range of cMb in human blood. Hence, the specimen should be
diluted when applied to sub-μg mL−1 Mb determination. Since the sensing formation of antibody-
antigen and biotin–avidin complexes and DNA-oligonucleotide interactions and other biological
interfaces have witnessed the nondestructive and sensitive performance of electrochemical impedance
spectroscopy (EIS) in characterizing electrical traits, much attention has been paid to this technique [9,
10].
Having gained widespread application in the field of electrochemical sensors, Au nanoparticles
could take the roles of either markers in electrochemical determination or a desirable conducting
medium for electron exchange rate facilitation [11]. Au nanomaterials of varying forms have been
presented in various publications to immobilize redox proteins, where direct electrochemistry has been
achieved. The direct electrochemistry of hemoglobin (Hb) was studied on an Au nanowire array by
Yang and co-workers [12]. To study the electrochemical performance of Hb, polyvinyl alcohol and 1-
octyl-3-methylimidazolium hexafluorophosphate were integrated with an Au nanoparticle by Zhang
and co-workers [13]. With Hb immobilized onto a single-dimensional Au nanoparticle, a new
amperometric biosensor was fabricated by Hong and co-workers [14]. Au nanoparticles have also been
applied in the field of electrochemical DNA biosensors. The activity concerning immobilized
biomolecules, electrocatalytic activity-induced electron exchange rate promotion, as well as the
enhancement in DNA hybridization and immobilization ability could be realized by the existence of
Au nanoparticles on the surface of an electrode [15, 16].
Int. J. Electrochem. Sci., Vol. 12, 2017 6172

To fabricate a bioelectrode to determine Mb as a cardiac biomarker, a platform with

functionalized Au nanoparticles was employed in this study, with full consideration of the strengths of
metal Au nanoparticle traits as well as an EIS strategy. The covalent anchoring of the nanoparticles
occurs on the self-assembled monolayer (SAM) of 3-aminopropyltriethoxy silane (APTES) over an
indium-tin-oxide (ITO) glass plate. Using carbodiimide coupling, Ab-cMb (cardiac protein antibody)
was covalently attached to APTES/ITO glass plates decorated with carboxyl-functionalized Au
nanoparticles. Diverse microscopic strategies were employed for systematic characterization of the
proposed bioelectrode. EIS with the redox probe [Fe (CN)6]3−/4− was employed for the investigation of
the capacity of this bioelectrode to quantitatively estimating Ag-Mb in phosphate buffered saline
(PBS) at pH 7.4.

2. EXPERIMENTS

2.1. Chemicals

Sigma-Aldrich Corp was the source for monoclonal mouse anti-human cardiac myoglobin (Ab-
Mb), 3-mercaptopropionic acid (MPA), sodium borohydride (NaBH4), chloroauric acid (HAuCl4), N-
hydroxysuccinimide 98% (NHS), 3-mercaptopropyltrimethoxysilane (MPMSi), N-(3-dimethyl
aminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC), 3-minopropyltriethoxysilane
(APTES), mouse immunoglobulin-G (Ag-IgG) and human heart tissue-derived myoglobin (Ag-Mb).
Myoglobin standards and the human myoglobin ELISA kit were purchased from Abcam (Shanghai,
China). The remaining reagents were of analytical grade and were employed with no additional
purification.

2.2. Synthesis of monolayer-protected gold nanoparticles

Monolayer-protected Au nanoparticles were prepared and purified through the Weller

reaction [17], with MPMSi in a thiol/AuCl4− mole ratio of 1:1 in 2-propanol. This was followed by the
addition of NaBH4 as a reducing agent cooled to 273 K using an ice water bath for 1 h. The terminal
monolayer-protected Au nanoparticle solution was filtered through PTFE membranes (0.2 μm) for
deposition removal, and then the excess 2-propanol was evaporated in a rotary evaporator. This was
followed by the separation and collection of the crude product of Au nanoparticles via centrifugation.
The Au nanoparticles were then washed with ethanol to remove the unreacted thiols and byproducts.

2.3. Preparation of immunosensor

The ITO-coated glass was ultrasonically cleaned successively in Extran, acetone, ethanol,
isopropyl alcohol and DI water, with vacuum drying following each cleaning process. To raise the
number of hydroxyl groups on the surface of the ITO glass, the cleaned plates were subsequently
exposed to oxygen plasma in a plasma chamber for 5 min. This was followed by immersion of the
glass plates into a 2% ethanolic APTES solution for 90 min at room temperature to obtain APTES-
Int. J. Electrochem. Sci., Vol. 12, 2017 6173

SAM. After rinsing with ethanol, non-bonded APTES was removed from the substrate surface together
with N2 gas flow drying. Aqueous AuNPs (5 mL, 0.1 mg/mL) was added with NHS (30 mM) and EDC
(150 mM) to activate carboxyl groups. The ITO glass plates were immersed into the mixture of
functionalized and activated AuNPs for 60 min, washed with doubly distilled water, and dried under
N2 gas flow to give AuNPs/APTES/ITO. The electrode was treated overnight with PBS (pH 7.4)
containing Ab-Mb (100 μg/mL) at 4 °C. This was followed by PBS washing and N2 flow drying.
Subsequently, the incubation of the Ab-Mb-immobilized electrode was performed in a BSA solution
(1%) for 0.5 h to block the nonspecific binding sites on the surface of the electrode, followed by PBS
washing to remove any physically adsorbed antibodies. The final immunosensor, denoted as Ab-
Mb/AuNPs/APTES/ITO, was dried with N2.

2.4. Characterization and measurement

An SM-240 CCD spectrophotometer (CVI Spectral Instruments, Putnam, CT, USA) was used
to characterize Au and blank nanoparticles in solution via UV–vis absorption spectra. A traditional
triple-electrode system was applied to the electrochemical measurements, where the roles of counter,
reference and working electrode were taken by platinum wire, Ag/AgCl and the prepared electrode.
Cyclic voltammetry was conducted in 10.0 mL of 0.1 M KCl containing 1.0 mM each of
Fe(CN)63− and Fe(CN)64− at a scan rate of 50 mV/s. Electrochemical impedance spectroscopy (EIS)
measurements with a frequency range of 1-100 kHz at an AC voltage of 0.05 V were performed in
PBS, pH 7.4, containing 0.1 M KCl and 2 mM [Fe(CN)6]3−/[Fe(CN)6]4−.

3. RESULTS AND DISCUSSION

Figure 1 shows the UV absorption spectra of Au nanoparticles protected by the monolayer.

Both the ~525 nm absorption band and yellow-brownish red color variation in the mixture of MPMSi
+ AuCl4− solution provide convincing evidence for successful Au nanoparticles synthesis. Since the
mercapto group of MPMSi has significant affinity to Au, a rather narrowed size is attained for the tiny
Au nanoparticles [18]. Based on the SPR location, the size of the Au NPs can be estimated to be 14
nm. Thus, Au nanoparticles are ensured in this way, with the following aminofunctionalization of them
facilitated. Moreover, the elevated ratio of surface-to-volume and nonrandom nanostructure has been
displayed for the Au nanoparticles according to the aforementioned results. Hence, protein
immobilization could be achieved via the two traits of the nanoparticles for subsequent electrochemical
sensor preparation.
Int. J. Electrochem. Sci., Vol. 12, 2017 6174

Figure 1. UV/vis spectra of the colloidal solution of Au nanoparticles in 2-propanol and the unreacted
2-propanol solution with MPMSi and AuCl4− (S/Au ratio of 1:1).

Figure 2. FTIR spectra of APTES/ITO, AuNPs/APTES/ITO and Ab-Mb/AuNPs/APTES/ITO.

Under attenuated total reflection mode, FTIR measurement is employed to characterize Ab-
Mb/AuNPs/APTES/ITO, AuNPs/APTES/ITO and APTES/ITO. Their respective spectra are shown in
Int. J. Electrochem. Sci., Vol. 12, 2017 6175

Figure 2. A Si–O–Si characteristic band is observed at 1055 cm−1 in for APTES. C=O stretching
vibrations of the carboxylic groups of AuNPs account for the band at 1748 cm−1. The –OH bending
and stretching vibrations of carboxylic acid groups are respectively reflected by the 939 and 2962
cm−1 peaks. The N-Au vibration mode has been noticed at 1044 cm−1, suggesting the affliction of the
nitrogen and gold bonding during the synthesis process. The N–H bending and stretching vibrations
are seen in the peaks observable at 1612 and 3387 cm−1, showing successful Ab-Mb immobilization
[19].
We examined the properties of the decorated electrode surface facilely and effectively via
cyclic voltammetry measurement. Electron transfer resistance and other electron exchange rate
constants bear theoretical relation with the variations in peak current and separation of peak potentials
on varying electrode surfaces, as reflected in cyclic voltammograms (CVs). With the redox probe
consisting of a [Fe(CN)6]3 −/4 –mixture in PBS (containing 0.1 M KCl), pH of 7.4, cyclic voltammetry
was employed for the surveillance of every stage of antibody immobilization as well as ITO surface
modification. CV profiles preceding and succeeding each step of electrode surface modification are
shown in Figure 3. Ab-Mb/AuNPs/APTES/ITO is formed due to the subsequent Ab-Mb
immobilization. The original ITO exhibits a peak-to-peak separation between the oxidation and
reduction potentials (∆Ep) of 125 mV via a quasi-reversible CV. This is followed by a decline to
96 mV (∆Ep) after APTES-SAM-induced modification. These changes are attributed to an increased
interfacial concentration of the anionic probe ([Fe(CN)6]3 − /4 − ) due to its strong affinity toward the
polycationic (NH3+) layer of the amino groups of the APTES [20]. Nevertheless, with respect to
AuNPs-decorated APTES/ITO, there is a rise in peak potential and, conversely, a decline in peak
current. Both after blocking with BSA and with Ab-Mb immobilization on the AuNPs/APTES/ITO
surface, there is an additional decline in CV peak current, possibly ascribed to the generation of a
protein-antibody layer on the electrode. Herein, the route for redox probe to electrode surface is
remarkably impeded by this layer that obstructs mass-transfer and electron exchange.

Figure 3. CV of the original ITO; APTES/ITO; AuNPs/APTES/ITO; Ab-Mb/ZnS(MPA)/APTES/ITO

glass as well as Ab-Mb/AuNPs/APTES/ITO in PBS, pH 7.4, containing 2 mM
[Fe(CN)6]3 −/4 − plus 0.1 M KCl.
Int. J. Electrochem. Sci., Vol. 12, 2017 6176

Because the electrochemical courses at the solution/electrode interface concerning the

decorated electrodes can be quantitatively and qualitatively characterized by EIS measurements, it was
selected as the major characterization approach to evaluate immunosensor behavior. We detected its
response to AC signal (small amplitude) adoption at varying frequencies via EIS. The impedance
performance of varying decorated electrodes is indicated in Nyquist plots with Z′ (real part of the
impedance) vs. −Z′′ (imaginary part of the impedance). In contrast, the linear portion coinciding with
the diffusion limited electron shift occurred at comparatively lower frequencies. It can be proposed that
the spectra were similar to those of Randle's equivalent circuit in theory. Diverse decorated electrodes
are characterized via Nyquist plots with equivalent circuits (inset Figure 4). The Ret value with respect
to the original ITO is observed at 84.12 Ω cm2, while that with respect to APTES/ITO greatly declines
to 22.59 Ω cm2. Since the anionic probe [Fe(CN)6]3−/4− has significantly affinity for the polycationic
layer, as the amino groups of APTES are protonated (NH3+) in aqueous solution, there is an increase in
the concentration of the probe, thus reducing the Ret value. Since the negatively-charged AuNPs
carboxyl species at the interface of the solution and electrode are charge repulsive, there is a significant
increase in the Ret value up to 110.4 Ω cm2 after AuNPs are attached to the silane layer. The probe-
electrode surface electron exchange is unrestricted, as suggested by the finding that an
obvious Ret decline down to 17.0 Ω cm2 corresponds to the decline in semicircle diameter with respect
to AuNPs/APTES/ITO. Moreover, the distinct Ret increase to 199.3 Ω cm2 is attributed to MB antibody
immobilization onto AuNPs over the surface of the electrode as well as the obstruction of the BSA-
induced nonspecific binding sites. The probe-electrode surface mass and charge exchange is impeded
due to the insulation property of the protein molecule.

Figure 4. Nyquist plots of original ITO; APTES/ITO; AuNPs/APTES/ITO as well as Ab-

Mb/AuNPs/APTES/ITO in PBS (pH 7.4) containing 2 mM [Fe(CN)6]3−/4− plus 0.1 M KCl.
Int. J. Electrochem. Sci., Vol. 12, 2017 6177

Non-Faradaic impedance biosensors perform impedance measurement in the absence of any

redox probe. Bacteria detection is based on the impedance change upon the attachment of bacterial
cells on an interdigitated microelectrode in the absence of any redox probe in the sample solution. An
antigen-antibody complex is formed by the specific immunoreaction between Ab-Mb and Ag-Mb,
where the latter functions as the complementary target of the former. The interfacial electron exchange
at the interface of the solution and bioelectrode is interrupted by an immunocomplex-induced kinetic
barrier. There is an increase in electron exchange resistance, together with a decrease in capacitance,
with gradually impeded Faradaic reaction of the redox couple. The control group uses an Ag-Mb-free
sample solution, with an Ret value recorded as the control sample response. After aliquots of target
protein antigen with varying concentrations are consecutively added, the corresponding Nyquist plots
obtained are shown in Figure 5A. As the added Ag-Mb increases in concentration, Nyquist circles
exhibit a significant increase in diameter.

Figure 5. (A) Faradaic impedance spectra of the Ab-Mb/AuNPs/APTES/ITO preceding and

succeeding incubation with varying concentrations of Ag-Mb in PBS, pH 7.4, containing 2 mM
[Fe(CN)6]3−/4− plus 0.1 M KCl. (B) Concentration dependent calibration profile with respect to
the bioelectrode.

Table 1. Performance comparison of the Ab-Mb/AuNPs/APTES/ITO and other Mb electrochemical

sensor.

Method Linear range Detection limit Reference

Bio-functionalized Pt NPs 0.01 μg/mL-1 μg/mL 4 ng/mL [23]
TiO2 nanotubes ― 50 nM [24]
Au/graphene/carbon ionic liquid electrode 1.98-238.2 μM 0.66 μM [25]
Fe3O4@SiO2microsphere/carbon ionic 0.2-11 mM 0.18 mM [26]
liquid electrode

Ab-Mb/AuNPs/APTES/ITO 10 ng/mL-1 μg/mL 2.7 ng/mL This work

This is ascribed to the interaction between antibody and antigen. The variation in specific
electron charge exchange resistance versus the logarithmic Ag-Mb concentration (10 ng/mL-1 μg/mL)
is plotted for the bioelectrode in Figure 5B, with an LOD of 2.7 ng/mL based on a signal-to-noise ratio
Int. J. Electrochem. Sci., Vol. 12, 2017 6178

of 3:1. EIS has the potential to be employed for clinical determination due to its comparatively short
experimental course time (roughly 15 min) for the entire assay, in comparison with ELISA, which
requires several days. Hence, among other recently issued polymer-based Mb sensors as well as
semiconductor/metal nanoparticle/carbon nanomaterials, the proposed bioelectrode is the most
desirable platform [8, 21, 22]. The sensing performance of the new sensor was compared with recently
reported sensors, as shown in Table 1.
Mb was determined in spiked, diluted (1:10) synthetic serum samples and EIS measurements
were performed. The linear detection range is from 50 ng/mL to 700 ng/ mL. In all measurements, the
repeatability is always less than 6%. Each target Ag-Mb with varying concentration was added for ten
diverse independently fabricated bioelectrodes. This work thus successfully assessed bioelectrode
reproducibility. Desirable preparation reproducibility as well as precision are displayed by the inter-
assay variation coefficient range of 2–12% at an individual Ag-Mb concentration. The impedance
measurements were repeatedly performed for target Ag-Mb with the same concentration on the
bioelectrode in similar conditions to examine stability. The embodiment of bioelectrode
biocompatibility in either open air or solution is indicated by the absence of observable impedimetric
response variations, even after 5 impedance measurements were conducted. The immunoreactions in
the presence of nonspecific mouse IgG under similar conditions were performed to examine the
specificity of the bioelectrode for Ag-Mb, without obvious Ret variations monitored as aliquots of IgG
was added, possibly due to the attenuated nonspecific interaction or the absence of the interaction
between antibody and antigen, thus verifying the specificity of the proposed bioelectrode.

Table 2. The contents and recoveries of Mb detected in synthetic serum samples (n=3).

Sample Added (ng/mL) Found ELISA result Recovery (%) RSD (%)
(ng/mL) (nM/mL)
1 70 68.7 69.1 98.14 3.6
2 150 154.6 147.2 98.13 2.5
3 300 309.9 295.7 103.3 1.9

4. CONCLUSIONS

Here, we have given an integrated study of an AuNPs-based bioelectrode employed in

bioanalytical immunosensor application. In general, the monolayer-protected AuNPs were covalently
attached over APTES-SAM onto the ITO surface. The electrode is not only desirably biocompatible
but also supplies ample binding sites where Ab-Mb is immobilized on the surface of the electrode for
immunoreaction. The detection range for Ag-Mb with respect to the prepared sensor is 10 µg/mL in
PBS.

References

1. A. Wu, F. Apple, W. Gibler, R. Jesse, M. Warshaw and R. Valdes, Clinical Chemistry, 45 (1999)
Int. J. Electrochem. Sci., Vol. 12, 2017 6179

1104.
2. F. Apple, R. Jesse, L. Newby, A. Wu and R. Christenson, Circulation, 115 (2007) e352.
3. L. Bonel, J. Vidal, P. Duato and J. Castillo, Biosensors and Bioelectronics, 26 (2011) 3254.
4. B. Jeong, R. Akter, O. Han, C. Rhee and M. Rahman, Anal. Chem., 85 (2013) 1784.
5. J. Li, L. Xiao, G. Zeng, G. Huang, G. Shen and R. Yu, Journal of Agricultural and Food Chemistry,
53 (2005) 1348.
6. S. Arya, T. Pui, C. Wong, S. Kumar and A. Rahman, Langmuir, 29 (2013) 6770.
7. Q. Wei, Y. Zhao, B. Du, D. Wu, H. Li and M. Yang, Food Chemistry, 134 (2012) 1601.
8. S. Pakapongpan, R. Palangsuntikul and W. Surareungchai, Electrochimica Acta, 56 (2011) 6831.
9. A. Venkatanarayanan, T. Keyes and R. Forster, Anal. Chem., 85 (2013) 2216.
10. S. Shrikrishnan, K. Sankaran and V. Lakshminarayanan, J Phys Chem C, 116 (2012) 16030.
11. C. Lei, S. Hu, G. Shen and R. Yu, Talanta, 59 (2003) 981.
12. M. Yang, F. Qu, Y. Li, Y. He, G. Shen and R. Yu, Biosensors and Bioelectronics, 23 (2007) 414.
13. Y. Zhang, R. Yan, F. Zhao and B. Zeng, Colloids and Surfaces B: Biointerfaces, 71 (2009) 288.
14. J. Hong and Z. Dai, Sensors and Actuators B: Chemical, 140 (2009) 222.
15. H. Gao, X. Qi, Y. Chen and W. Sun, Anal. Chim. Acta., 704 (2011) 133.
16. J. Pingarrón, P. Yáñez-Sedeño and A. González-Cortés, Electrochimica Acta, 53 (2008) 5848.
17. A. Schroedter and H. Weller, Angewandte Chemie International Edition, 41 (2002) 3218.
18. T. Teranishi, I. Kiyokawa and M. Miyake, Adv. Mater., 10 (1998) 596.
19. A. Barth, Biochimica et Biophysica Acta (BBA)-Bioenergetics, 1767 (2007) 1073.
20. J. Zhao, C.R. Bradbury, S. Huclova, I. Potapova, M. Carrara and D. Fermín, J Phys Chem B, 109
(2005) 22985.
21. E. Matveeva, Z. Gryczynski and J. Lakowicz, Journal of immunological methods, 302 (2005) 26.
22. E. Suprun, A. Shilovskaya, A. Lisitsa, T. Bulko, V.V. Shumyantseva and A.I. Archakov,
Electroanalysis, 23 (2011) 1051.
23. S. Mishra, A. Srivastava and D. Kumar, RSC Advances, 4 (2014) 21267.
24. S. Mandal, K. Narayan and A. Bhattacharyya, Journal of Materials Chemistry B, 1 (2013) 3051.
25. G. Li, T. Li, Y. Deng, Y. Cheng, F. Shi, W. Sun and Z. Sun, Journal of Solid State Electrochemistry,
17 (2013) 2333.
26. X. Wang, Z. You, H. Sha, Z. Sun and W. Sun, Journal of Solid State Electrochemistry, 18 (2014)
207.

© 2017 The Authors. Published by ESG (www.electrochemsci.org). This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/4.0/).