Oxygen evolution on alpha-lead dioxide electrodes in methanesulfonic acid

Electrochimica Acta 63 (2012) 28–36

Contents lists available at SciVerse ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Oxygen evolution on alpha-lead dioxide electrodes in methanesulfonic acid Alexandre Oury a,∗ , Angel Kirchev a , Yann Bultel b a Laboratoire de Stockage de l’Electricité, Institut National de l’Energie Solaire, Commissariat à l’Energie Atomique et Energies Alternatives (CEA), 50 Avenue du Lac Léman – LYNX 3, BP332, 73377 Le Bourget du Lac, France b Laboratoire d’Electrochimie et de Physico-chimie des Matériaux et des Interfaces (LEPMI) UMR 5291 CNRS/Grenoble–INP/UJF/UdS, 1130, rue de la Piscine, BP 75 – 38402 Saint Martin d’Hères, France

a r t i c l e

i n f o

Article history: Received 6 July 2011 Received in revised form 7 December 2011 Accepted 8 December 2011 Available online 17 December 2011 Keywords: Oxygen evolution reaction Lead dioxide Methanesulfonic acid Soluble lead acid flow battery

a b s t r a c t This work examines the oxygen evolution reaction (OER) taking place on ␣-PbO2 electrode in methanesulfonic acid (MSA) medium and in sulphuric acid as a comparison, by means of cyclic voltammetry (CVA) and electrochemical impedance spectroscopy (EIS), for soluble lead acid flow battery applications. The influence of MSA concentration on OER is studied. EIS measurements highlighted the impact of the hydrated lead dioxide layer upon decreasing MSA or sulphuric acid concentration. The evolution of the Tafel curves plotted from EIS measurements and quasi-stationary currents while varying acid concentration was interpreted in the light of this hydrated layer which could enhance the electrocatalytic activity when it is thin, and on the contrary act as an electronic barrier when growing for low acid concentration. Both EIS and CVA revealed that OER on lead dioxide is less favoured in MSA than in sulphuric acid. It is finally concluded that a high-concentrated MSA electrolyte is better for lead acid flow battery application in terms of oxygen evolution. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction A new type of redox flow battery, based on lead(II) dissolved in the electrolyte, was recently proposed by Pletcher et al. in a series of papers [1–10]. This soluble lead flow battery (SLFB) works with a single electrolyte made up of methanesulfonic acid (MSA, formula: CH3 SO3 H) which contains dissolved lead(II) ions in the form of Pb(CH3 SO3 )2 . The reactor consists of two electrodes between which flows the electrolyte. Lead(II) is oxidised into lead dioxide at the positive electrode and reduced into lead at the negative during charge. During discharge, PbO2 and Pb turn back to Pb(II) in solution in MSA. The overall battery reaction is: 2Pb2+ + 2H2 O

charge



PbO2 + Pb + 4H+

discharge

The lead(II) concentration in the electrolyte decreases when the battery is charged. In the same time, the protons concentration increases twice as fast. The system differs from traditional lead acid batteries in that the electrolyte contains the active soluble lead(II) species, so that there is no solid phase reaction. The main advantage of this technology compared to other redox flow systems is that no membrane or separator is required since the electrolyte is the same for both

∗ Corresponding author. Tel.: +33 479601440. E-mail address: [email protected] (A. Oury). 0013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2011.12.028

electrodes. In addition, methanesulfonic acid is a less hazardous and more environmentally friendly electrolyte than other acids [11]. Oxygen evolution reaction (OER) taking place at the positive electrode during charge leads to a decrease in the faradic efficiency of PbO2 deposition, and hence leads to losses in energy efficiency and chemical imbalance at the two electrodes (excess of lead at the negative electrode) [6]. Moreover, it has been established [12] that uniform and well-adherent PbO2 layers are deposited from MSA electrolyte in low current density and low temperature conditions, whereas powdery or pitted PbO2 of poor adhesion are obtained using a high current density and/or high temperature, i.e. conditions favouring O2 evolution. It is highly possible that O2 evolution play a significant role in the loss of PbO2 in the form of small particles in the electrolyte during SLFB cycling encountered by Pletcher’s team [10]. Therefore studying oxygen evolution reaction on lead dioxide material in methanesulfonic acid medium appeared to be worthwhile. While there is quite a lot of the literature concerning oxygen evolution on PbO2 material in several electrolytes like H2 SO4 [13–24], HClO4 [15,22], H3 PO4 [15], or CF3 SO3 H [21], no OER investigation on PbO2 was carried out in methanesulfonic acid, probably due to its recent application in lead acid batteries technology. The general mechanism leading to O2 formation can be written as: H2 O → OH• ads + H+ + e−

(1)

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

OH• → Oads + H+ + e−

(2)

followed by the chemical recombination of oxygen: 2Oads → O2

(3)

Pavlov et al. [17,18] proposed that this mechanism take place on active centers PbO(OH)2 located in a hydrated lead dioxide layer at the PbO2 /electrolyte interface according to: PbO(OH)2 + H2 O → PbO ∗ (OH)2 . . . OH• + H+ + e− +

PbO(OH)2 . . . OH• → PbO ∗ (OH)2 + Oads + H + e

−

(4) (5)

2Oads → O2 The kinetics of OER studied by steady-state polarization (current density vs. electrode potential), and represented by the slopes of the E. vs. log j plots (Tafel coefficient b), depends to a great extent on PbO2 structure (i.e. on the deposition conditions) [19,24] and allotropic form (␣ or ␤) [14,16]. The most commonly encountered b value is 120 mV [15,16,19,22], which corresponds to a single electron transfer reaction with a ˛ transfer coefficient of 0.5 (b = ln 10 × RT/˛zF) at 25 ◦ C. But values as high as 277 mV [24], or 344 mV [23] can be found. The relative kinetics of the two electrochemical reactions is still not clear. Da Silva et al. [25] proposed that the primary water discharge (1) be the rate determining step (r.d.s.), Ho et al. [16] invoke the oxidation of adsorbed OH• as r.d.s. The present paper is dedicated to the study of O2 evolution on lead dioxide in methanesulfonic acid medium. Cyclic voltammetry was first used to calculate the kinetic parameters (Tafel coefficient and activation energy). Then electrochemical impedance spectroscopy was employed for further investigations on the reaction mechanisms. The lead dioxide allotropic form ␣ was chosen because of its recognition by Pletcher et al. [7] to be the most convenient form for SLFB application due to its compact and welladherent structure. 2. Experimental 2.1. PbO2 deposition conditions The disc electrodes on which lead dioxide was deposited were made up of a glassy carbon rod (diameter ∼0.3 cm, cross-section ∼0.071 cm2 ) embedded in Plexiglas tubes filled with epoxy resin. Before each deposition, the glassy carbon surface was pre-treated by anodic etching, subjecting the electrode to a constant anodic current (30 mA cm−2 ) in 1 M NaOH during 10 min, in order to improve PbO2 adhesion due to glassy carbon roughening [26]. ␣-PbO2 (250 ␮m thick) was electrodeposited onto glassy carbon electrode at room temperature from a 1 M Pb(CH3 SO3 )2 + 0.2 M CH3 SO3 H + 5 mM C16 H33 (CH3 )3 N+ solution prepared by adding PbO (99%, Sigma–Aldrich) and hexadecyltrimethylammonium ptoluene sulfonate (C16 H33 (CH3 )3 N·C7 H6 SO3 H, 99%, Sigma–Aldrich) to a methanesulfonic acid solution (from MSA 70%, Alfa Aesar). C16 H33 (CH3 )3 N+ cation was used by Pletcher et al. [5] to improve lead deposition quality in their redox flow battery (5 mM in electrolyte). The electrodeposition was performed with a constant anodic current of 50 mA cm−2 for 1 h, according to a protocol derived from Sirés et al. [12] for electrodeposition of uniform and well-adherent ␣-PbO2 layer, using a Pt counter-electrode. 2.2. Cyclic voltammetry and electrochemical impedance spectroscopy Oxygen evolution on ␣-PbO2 was studied using cyclic voltammetry (CVA) and electrochemical impedance spectroscopy (EIS)

29

in methanesulfonic acid at different concentrations (from 0.25 to 5 M), and in sulphuric acid as a comparison. All measurements were performed using a classical 3-electrode cell (working PbO2 electrode, Pt counter electrode, reference electrode) with a homemade Ag/Ag2 SO4 reference electrode (E = 0.654 V/SHE). The solutions were all magnetically stirred to remove the bubbles from the electrode surface as much as possible, and to be closed to the electrolyte flowing conditions of the battery. Such convection is not expected to have an influence on the kinetics of the reaction because oxygen evolution reaction comes from water which is the solvent. In these conditions, the rate of the process is not expected to be diffusion controlled (no depletion in water molecules can happen near the electrode surface for there are in strong excess) but only kinetically controlled by the reaction itself. Thus the convection will not have any strong effect both on the polarization curves and on the EIS diagrams, which reflect the kinetics of the reaction. CVA measurements was carried out recording four potential cycles under quasi-stationary conditions (potential scan rate 0.5 mV s−1 ), from 1.1 V to 1.3 V, after a 30 min constant polarization step at 1.1 V, for different temperatures (25–65 ◦ C, with 10 ◦ Cincrements). Tafel curves (E vs. log j) were plotted for the anodic sweep of the last cycle, while Arrhenius curves (log j vs. 1/T) for potentials ranging from 1.1 to 1.3 V were obtained calculating the average current value on the anodic sweep over the four cycles for each potential and temperature. EIS consisted in applying a sinusoidal potential (amplitude: 5 mV, frequency range: 5 mHz–65 kHz) added to different d.c. potentials ranging from 1.1 to 1.4 V (25 mV-increments) to the PbO2 electrode. Before each impedance measurement at a given d.c. potential, the electrode was polarized at this potential during 15 min to make sure that the current remained stable and that bubbles formation and detachment had no incidence on the impedance of the electrode. All EIS measurements were performed at 25 ◦ C. 2.3. Instrumental Carbon etching, PbO2 deposition, and cyclic voltammetry were all performed using a Solartron 1470 apparatus monitored by CorrWare. XRD analysis and SEM images were carried out respectively on a Brücker D8 Advance diffractometer and a FEI Nova NanoSEM 630 scanning electron microscope. EIS measurements were performed with a Solartron SI 1250 monitored by Zplot. EIS data treatments (Nyquist diagrams fitting with equivalent circuits) was done with ZView 2 software, and extracted capacities calculated with Matlab. Cell’s temperature was regulated with a Bioblock Scientific thermostat. 3. Results and discussion 3.1. Voltammograms, Tafel and Arrhenius plots A X-ray diffractogram and SEM images of a PbO2 coating deposited in the conditions described in Section 2.1. are shown in Fig. 1. The identification of the diffraction peaks was done using the PDF-2 database from ICDD. Most of the peaks can be identified to the ␣ form of PbO2 , while very few peaks corresponding to the other allotropic form of lead dioxide (␤-PbO2 ) can be found. The fact that no significant peak exclusively associated to ␤-PbO2 can be observed strongly indicates that the large majority of the lead dioxide coating is in the ␣ (orthorhombic) form, as expected. The mean crystallite size, estimated using the Debye-Scherrer equation with the full width at half maximum (FWHM) of the (0 2 0) or (0 2 1) diffraction peak, was found to be around 15 nm, which is in good accordance with Sirés et al. [12] who reported a mean crystallite size of 13 nm for lead dioxide deposited in the same

30

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

(Fig. 3c) and as a function of MSA concentration at 25 ◦ C (Fig. 3d). The Tafel slopes values for each temperature and MSA concentration were averaged over at least two experiments carried out in the same conditions, considering each time the four cycles recorded. In order to give an idea of the uncertainties, the standard deviation on all the b values obtained was calculated. It can be seen from Fig. 3c that Tafel slope ‘b’ is rather unchanged upon increasing the temperature, as reported by Pavlov et al. [19] in sulphuric acid medium. It ranges from 180 to 200 mV over the temperatures studied (25–65 ◦ C) for MSA 4 M. Similar Tafel slopes have been found at 25 ◦ C for all the MSA concentrations studied (Fig. 3d). The high dispersion obtained on ‘b’ (see error bars) makes it difficult to draw a conclusion about the influence of MSA concentration on Tafel slopes in the oxygen evolution process. Tafel slopes in sulphuric acid medium were found to be similar, with a value of 192 ± 20 mV at 25 ◦ C. It can be noticed that the ‘b’ values obtained in this work for both MSA and H2 SO4 electrolytes are rather high compared to the theoretical 120 mV for a rate determining step involving a single electron and a transfer coefficient (˛) equal to 0.5. Abnormal Tafel slope can be attributed to an apparent transfert coefficient (˛app ) which differs from 0.5 due to surface roughness, as mentioned by Rufino et al. [24] who found b value much higher than 120 mV when changing PbO2 deposition conditions leading to a surface roughness increase. Arrhenius plots (see Fig. 3b for the case [MSA] = 4 M) were used to estimate the activation energy of the OER. Indeed, as mentioned by Pavlov et al. [19], the relation between the rate of O2 evolution (expressed by the current density j) and the temperature can be written: Fig. 1. X-ray diffractogram (a) and SEM images (b) of a PbO2 coating prepared at 50 mA cm−2 in a 1 M Pb(CH3 SO3 )2 + 0.2 M CH3 SO3 H + 5 mM C16 H33 (CH3 )3 N+ solution. All investigations on the oxygen evolution in this work were done on coatings prepared in these conditions.

log j =

MSA 3 M MSA 4 M

j / mA.cm-2

3 2.5

(b)

MSA 1 M

2

MSA 5 M

1.5 1 0.5

H2SO4 0.5 M

4 3 2.5 2

MSA 0.5 M

1.5 1

0 -0.5

4.5 3.5

j / mA.cm-2

3.5

(6)

where Ea is the activation energy in J mol−1 , R is the universal gas constant (=8.314 J K−1 mol−1 ), T is the temperature in K. Ea is then deduced from the slope of the Arrhenius curves log j vs. 1/T. It appeared that Arrhenius plots varied substantially from an experiment to another for the same electrolyte composition. Therefore the Ea values, given in Table 1 as a function of MSA concentration for different potentials, have been averaged over several experiments, using the standard deviations observed to estimate the uncertainties. As it can be seen from Table 1, the activation energies display, like Tafel slopes, a high dispersion. Therefore it is not worth comparing Ea value upon changing electrolyte composition. However, it can be noted that our range of values (from 35 to 68 kJ mol−1 ) are in pretty good agreement with those reported by Pavlov et al. [19] and Ho et al. [16] for OER taking place on PbO2 in H2 SO4 medium. In addition, our Arrhenius plots of OER in H2 SO4 0.5 M gave Ea values close to 55 kJ mol−1 , like in MSA 0.5 M.

conditions. Cyclic voltammograms in quasi steady-state conditions (scan rate: 0.5 mV s−1 ) in the potential region where oxygen evolution takes place on ␣-PbO2 electrodes were recorded in MSA for different concentrations and in 0.5 M sulphuric acid. Fig. 2 shows some of the curves obtained on the 4th cycle at 25 ◦ C. It appears that MSA concentration has a weak effect on the anodic current value of the oxygen evolution reaction. The current seems to reach a maximum for [MSA] = 3 M and to drop for a further increase in MSA concentration. For the same acid molarity (0.5 M), oxygen evolution is favoured in sulphuric acid medium compared to MSA (Fig. 2b), which makes this latter a quite interesting electrolyte for batteries application compared to classical lead-sulphuric acid technology. Fig. 3 shows the temperature dependence of the Tafel plots (Fig. 3a), the corresponding Arrhenius plots log j vs. 1/T as a function of the potential (Fig. 3b) and the Tafel coefficients b (slope of the E vs. log j curves) as a function of the temperature for [MSA] = 4 M

(a)

−Ea + const. ln 10RT

0.5 1.1

1.15

1.2

1.25

E vs. (Ag/Ag2SO4) / V

1.3

1.35

0

1.1

1.15

1.2

1.25

1.3

1.35

E vs. (Ag/Ag2SO4) / V

Fig. 2. Cyclic voltammograms obtained for oxygen evolution on ␣-PbO2 in different electrolytes at 25 ◦ C: (a) MSA 1, 3, 4, and 5 M; (b) MSA 0.5 M and H2 SO4 0.5 M. Scan rate: 0.5 mV s−1 , potentials vs. Ag/Ag2 SO4 electrode.

E vs. (Ag/Ag2SO4) / V

(a)

1.35

(b)

-1.5

log (j / A.cm-2)

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

-2

31

25°C 35°C 45°C 55°C 65°C

1.3 1.25 1.2 1.15 1.1 -4

-3.5

-3 -2.5 log (j / A.cm-2)

(c)

-2

1.3 V

-2.5

1.25 V 1.2 V 1.15 V

-3 -3.5 0.0029

-1.5

0.0031

0.0033

0.0035

(1/T) / K-1

(d)

300

b Tafel / mV

b Tafel / mV

230 210 190 170

270 240 210 180 150

150 20

30

40

50

60

0

70

1

2

T / °C

3 4 [MSA] / mol.L-1

5

6

Fig. 3. (a) Tafel plots of the oxygen evolution reaction on ␣-PbO2 obtained in MSA 4 M for temperatures ranging from 25 to 65 ◦ C; (b) corresponding Arrhenius plots; (c) Tafel coefficients b as a function of temperature for [MSA] = 4 M and (d) Tafel coefficients b as function of MSA concentration at 25 ◦ C (the error bars are related to the standard deviation observed on the b values).

Table 1 Activation energies of OER calculated from Arrhenius plots for each MSA concentration and potential. Activation energy Ea (kJ mol−1 )

E (V)

MSA 0.5 M 1.15 1.2 1.25 1.3

53.1 53.8 53.7 53.4

± ± ± ±

MSA 1 M

11.2 10.5 13.7 16.5

57.4 58.6 58.2 56.2

± ± ± ±

MSA 2 M

12.9 12.7 11.3 9.4

65.5 68.5 68.4 67.8

± ± ± ±

3.2. Electrochemical impedance spectroscopy Oxygen evolution reaction on ␣-PbO2 electrodes was also investigated at 25 ◦ C by EIS in the same conditions (same 3-electrodes cell, MSA and H2 SO4 electrolytes) for different d.c. potentials ranging from 1.1 to 1.4 V vs. Ag/Ag2 SO4 reference electrode. The obtained Nyquist diagrams as a function of applied d.c. potential are displayed in Fig. 4 for the case of MSA 5 M. When increasing the electrode potential, the impedance spectra evolves from partially

(a)

MSA 3 M

12.4 13.5 13.4 12.2

-80

48.2 49.6 48.8 47.5

± ± ± ±

MSA 4 M

13.7 14.1 12.6 12.6

35.1 43.6 38.9 38.2

(b)

42.3 43.5 44.2 42.3

± ± ± ±

9.6 9.7 10.0 9.7

-40

1.3 V 1.4 V

0.01

Z'' / Ω.cm²

Z'' / Ω.cm²

MSA 5 M

8.0 9.7 8.7 8.8

defined semi-circle to depressed semi-circle with an exponentially decreasing diameter. Fig. 5 presents the evolution of the complex plane plots as a function of MSA and H2 SO4 concentration for a 1.2 V d.c. potential. It can be noted that upon decreasing the electrolyte acidity, a semi-circular shape appears in the Nyquist diagrams towards high frequencies for both MSA and sulphuric acid media (Fig. 5). This semi-circle, which can be clearly seen for MSA molarities below 1 M, is rather stable with d.c. potential and acts as a constant

-60 0.01

-40 1.1 V

0.1

1.2 V

-20

fit

fit

1 1

1

0

20

40

60

80

Z' / Ω.cm²

100

120

140

0

0.01

0.1

-20

0.1

0

± ± ± ±

0

0.1 0.01

20

40

60

Z' / Ω.cm²

Fig. 4. Nyquist diagrams of the oxygen evolution reaction in MSA 5 mol L−1 , plotted for several d.c. potentials: (a) 1.1 and 1.2 V; (b) 1.3 and 1.4 V. The values reported are the frequencies in Hz.

32

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

Table 2 Equivalent circuits models used to fit the Nyquist diagrams obtained for different MSA concentrations. Constant phase elements (CPE) were used instead of capacitors. [MSA] (mol L−1)

Best equivalent circuit

Meaning of parameters

a. Rel

CPEdl Rct

Rel : electrolyte resistance CPEdl , Rct : double layer capacitance and charge transfer resistance CPEads , Rads : pseudocapacitance and resistance of OER intermediate products

CPEads

1, 2, 3, 4, 5

Rads

b. Rel

CPE1

CPEdl

R1

Rct

CPEads

0.5

Rads

Rel : electrolyte resistance CPE1 , R1 : capacitance and resistance of hydrated zone. CPEdl , Rct : double layer capacitance and charge transfer resistance CPEads , Rads : pseudocapacitance and resistance of OER intermediate products

c. Rel 0.25

CPE1

CPE2

R1

R2

Rel : electrolyte resistance CPE1 , R1 : capacitance and resistance of hydrated zone. CPE2 , R2 : capacitance and resistance of OER

Table 3 Parameters extracted from Nyquist diagrams for MSA concentrations ranging from 1 to 5 M. Equivalent circuit ‘a’ of Table 2. E (V) MSA 5 M 1.1 1.15 1.2 1.25 1.3 1.35 1.4 MSA 4 M 1.1 1.15 1.2 1.25 1.3 1.35 1.4 MSA 3 M 1.1 1.15 1.2 1.25 1.3 1.35 MSA 2 M 1.1 1.15 1.2 1.25 1.3 1.35 MSA 1 M 1.1 1.15 1.2 1.25 1.3 1.325

Tdl (mF cm−2 sn−1 )

ndl

Rct ( cm2 )

34.7 35.8 37.9 51.1 27.4 19.9 13.0

0.76 0.75 0.73 0.72 0.75 0.76 0.81

– – – – 11.2 7.8 4.8

36.5 38.8 42.4 43.6 35.1 23.9 16.0

0.78 0.76 0.74 0.73 0.74 0.77 0.8

60.7 62.3 66.4 62.1 47.3 34.8 56.17 55.74 61.83 57.02 50.23 37.21 63.8 46.4 63.2 89.4 82.1 110.8

Tads (mF cm−2 sn−1 )

nads

Rads ( cm2 )

47.4 40.3 15.7 16.1 20.2 24.1 17.4

0.5 0.54 0.5 0.65 0.66 0.55 0.52

937.8 264.6 191.3 104.3 67.1 32.9 16.5

– – – – 4.8 2.4 1.3

68.2 55.2 33.7 29.3 30.3 37.4 33.7

0.48 0.56 0.63 0.7 0.66 0.59 0.55

660.5 163.8 119.1 59.0 31.1 16.1 8.7

0.77 0.76 0.74 0.74 0.77 0.79

– – – – 6.7 4.9

94.8 90.1 65.1 74.3 58.3 62.8

0.51 0.57 0.6 0.68 0.6 0.57

267.1 117.3 66.4 25.6 15.3 7.1

0.79 0.78 0.76 0.76 0.76 0.78

– – – 8.1 7.8 5.9

97.3 90.7 54.9 49.2 51.8 82.2

0.54 0.59 0.64 0.67 0.71 0.61

323.6 140.1 64.6 27.4 10.2 4.1

0.62 0.68 0.67 0.64 0.68 0.66

1.9 1.8 1.3 1.1 0.4 0.5

41.5 55.6 82.1 66.6 132.7 113.2

0.77 0.8 0.79 0.77 0.73 0.74

225.4 134.2 27.8 8.6 2.6 1.8

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

-70

(a)

-15

(b)

-13

-50

Z'' / Ω.cm²

Z'' / Ω.cm²

0.01

E = 1.2 V MSA 5 M

-30

0.1

MSA 3 M

0.01

MSA 1 M MSA 0.5 M

0.01

-10 1 0

20

60

80

Z'' / Ω.cm²

Z'' / Ω.cm²

0

0

100 1k 1

10k 2

3 10 1k

0.01

E = 1.2 V 0.1

1

H2SO4 3 M H2SO4 0.5 M

1

5

10

3

4

25

1

-0.9 -0.6

0

5

20

1

10

1k

10

-0.3

100

15

Z' / Ω.cm²

-1.2

1

10

0.1

-5

-1.5

-2

10

-7

0

1

-1

-9

120

1

10

0.01

-1

100

Z' / Ω.cm²

-3

-11

-3

MSA 0.25 M

40

33

100

10k 100

0

0.5

1

1.5

2

2.5

Z' / Ω.cm² Ω

Z' / Ω Ω.cm²

Fig. 5. Nyquist diagrams over the all frequency range and, below, towards high frequencies for different electrolyte compositions: (a) MSA from 0.25 to 5 M; (b) H2 SO4 0.5 and 3 M. Solid lines represent the fit from equivalent circuit models of Table 2, and values reported refers to the frequency in Hz – d.c. potential: 1.2 V vs. Ag/Ag2 SO4 reference electrode.

(a)

0.12

MSA 5 M

Tdl / F.cm-2.sn-1

parallel RC element in series with the OER process. It could be interpreted considering the hydrated gel zone theory developed by Pavlov et al. [17], in which oxygen is evolved on active centers PbO(OH)2 located in a hydrated zone at the oxide/solution interface. It can be assumed that the growth of these hydrated zones is favoured in a diluted acid electrolyte for which the number of ‘free’ water molecules is higher than in a highly concentrated acid. In

0.1

MSA 4 M

this way of reasoning, the proportion of hydrated material within the PbO2 layer is expected to be higher when the acid is more diluted. If we assume that the conductivity of the hydrated zone is smaller than the crystalline PbO2 , the presence of an arc towards high frequencies in the Nyquist diagrams recorded for low acid concentrations could be related to the growth of this resistive hydrated layer.

(b)

0.9

MSA 3 M

0.08

1

0.8

ndl

MSA 2 M

MSA 5 M MSA 4M MSA 3 M MSA 2 M MSA 1 M MSA 0.5 M

0.06

0.7

0.04 0.6

0.02 0 1.05

1.15

1.25

1.35

0.5 1.05

1.45

1.15

E vs. (Ag/Ag2SO4) / V 0.15

MSA 5 M MSA 4 M MSA 3 M MSA 2 M MSA 1 M MSA 0.5 M

0.12 0.09

(d)

0.06

1.45

1

MSA 5 M MSA 4 M MSA 3 M MSA 2 M MSA 1 M MSA 0.5 M

0.8 0.7 0.6

0.03 0 1.05

1.35

0.9

nads

Tads / F.cm-2.sn-1

(c)

1.25

E vs. (Ag/Ag2SO4) / V

0.5 1.15

1.25

1.35

E vs. (Ag/Ag2SO4) / V

1.45

1.55

0.4 1.05

1.15

1.25

1.35

1.45

1.55

E vs. (Ag/Ag2SO4) / V

Fig. 6. Capacitance parameters T and n extracted from the CPE of the equivalent circuit models of Table 2, for several MSA concentrations: (a) double layer CPE coefficient T, (b) pseudo-capacitive CPE coefficient T of intermediate products, (c) double layer CPE coefficient n and (d) adsorption CPE coefficient n.

34

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

The equivalent circuit models that gave the best fits with experimental EIS data are summarized in Table 2. In the literature, all found studies [16,20–22] in which EIS has been performed for investigations on OER taking place on a lead dioxide electrode report the equivalent circuit ‘a’ shown in Table 2. This equivalent circuit includes charge transfer and adsorption processes in series, all in good accordance with the mechanism described by Eqs. (1)–(5). In these studies the resistance associated to charge transfer (Rct ) was much lower than the resistance associated to absorption/desorption species, and the authors concluded that the OER process is governed by adsorption/desorption of intermediate products rather than charge transfer. In the present work, the use of this equivalent circuit results in pretty good fits to the experimental data for MSA concentrations ranging from 1 to 5 M. For MSA 0.5 M, a parallel RC element is added to the equivalent circuit ‘a’ to take into account the high frequencies effect of the decrease in MSA molarity, related to the hydrated layer growth as discussed before. EIS data for MSA 0.25 M were successfully fitted with this hydrated layer related parallel RC in series with a single RC element representing the OER reaction (equivalent circuit ‘c’ of Table 2). This latter can be explained by a decrease in Rct to levels too low to be sensed by the EIS method. Constant phase elements were used instead of strictly capacitive elements in the simulations to take the depression of the semicircles in the complex plane plots into account. This phenomenon is induced by electrode surface morphology (roughness, nonuniformity. . .) [22]. The impedance of a CPE element is expressed by: ZCPE =

1 T (jω)

(7)

n

T is the capacity parameter expressed in F cm−2 sn−1 , j = where √ −1, ω the pulsation in s−1 , and n a dimensionless coefficient related to the depression of the semi-circle in the impedance spectra. It was established that the most accurate method to calculate the corresponding capacitance C (in F cm−2 ) is the use of the following equation [27]: C = T.(fm )

n−1

(8)

where T and n are the parameters expressed above, and fm is the frequency (Hz) at which the imaginary part of the impedance of the circuit constituted of R and CPE in parallel reaches a minimum. The parameters obtained by fitting the EIS data with the equivalent circuit models of Table 2 for different MSA solutions are listed in Tables 3–5. It must be noted that the experimental complex plane diagrams consist of the superposition of two semicircles (one related to charge transfer process and the other to intermediate products adsorption) associated with time constants close to each other. Thus they are very difficult to separate and the fitted values of Rct are mostly not significant because too low compared to Rads . This last point implies that adsorption/desorption of intermediate products governs the OER kinetics in MSA, as mentioned in the literature for other acidic media [16,20–22]. The values of the capacitance parameters T and n are displayed in Fig. 6 as a function of d.c. potential. No clear tendency can be seen in the variations of these parameters as a function of

both the potential and the MSA concentration. T ranges between 12 and 140 mF cm−2 sn−1 . Franco et al. [22] reported Tads values lying between 60 and 180 mF cm−2 sn−1 for the OER taking place on ␤-PbO2 deposited in nitric acid bath, in several mixtures of sulphuric and perchloric acid electrolytes. Calculating the capacitance Cdl and Cads from Tdl and Tads with Eq. (8) led to very high values (some of them were higher than 1 F cm−2 ) of several orders of magnitude above those commonly reported (0.03–25 mF cm−2 ) in the literature [16,20–22]. It became clear that these unexpectedly high capacitances values have no real physical meaning and that this could be an artefact arising from the very rough lead dioxide surface as it can be seen on the SEM image of Fig. 1 which displays a submicrometric structure. On one hand, such a structure is expected to give rise to a higher specific area of the material and consequently to higher capacitance values. On the other hand, the very low values of the CPE coefficients ‘n’ (some of them are near 0.5) are consistent with a high roughness of the electrode, very far from an ideal flat surface, whereas the commonly reported values for n in the references cited above are much more close to unity. This results in very depressed semi-circles in the EIS spectra, as it can be seen in Figs. 4 and 5. In such surface conditions, one can assume that extracting capacitances is difficult and the validity of Eq. (8) with very low value for n can legitimately be questioned. It was found that taking arbitrarily n closer to unity in Eq. (8) makes the C values significantly diminish, especially for the lowest n (see nads of Table 3) for which the difference can reach an order of magnitude. If the relation between the very high capacitances obtained and the observed morphology of lead dioxide is almost certain, such a difference in the lead dioxide morphology remains to be explained. In their recent review dealing with the electrodeposited lead dioxide coatings, Li et al. [28] mentioned many parameters that can influence lead dioxide morphology, including the substrate, pre-treatment of the substrate surface, pH, anion of the electrolyte, lead(II) concentration, current density, charge density (thickness), mass transport regime in the platting bath. In the present deposition conditions (see Section 2.1), many of these parameters differ from those applied in the literature dealing with EIS measurements of OER on lead dioxide, mainly the anion (methanesulfonate), the substrate (vitreous carbon), the current density or the concentration in Pb2+ . The lead dioxide studied in this paper was electrodeposited in conditions expected in the lead/AMS flow battery, and the observed morphology is very similar to the one observed in the battery of Pletcher [7]. The difficulties associated with the extraction of capacitance that have a physical meaning are to be regarded as a result, associated to the morphology arising from coatings electrodeposited in such platting conditions. As mentioned by Amadelli et al. [20,21], the plot of E vs. log(R−1 ) gives the Tafel slope of the OER. In their study, E vs. log(R−1 ) gave Tafel slopes similar to E vs. log j (stationary plots with j the current density) when Rads was used. Fig. 7 shows the correlations between the two plots for the whole MSA solutions investigated. Correlations were good for R = Rads + Rct , since Rct cannot be neglected at high potentials. While the physical meaning of the capacitance values was shown to be questionable, the accordance between E vs. log(R−1 ) and E vs. log j indicates that the EIS data does have sense

Table 4 Parameters extracted from Nyquist diagrams for MSA 0.5 M. Equivalent circuit ‘b’ of Table 2 (Rct values are not shown because too low to be numerically extracted). E (V)

T1 (mF cm−2 sn−1 )

n1

R1 ( cm2 )

Tct (mF cm−2 sn−1 )

nct

Tads (mF cm−2 sn−1 )

nads

Rads ( cm2 )

1.1 1.15 1.2 1.25 1.3 1.35

9.5 7.9 4.5 4.0 3.1 3.7

0.66 0.66 0.7 0.71 0.74 0.73

0.64 0.71 0.78 0.81 0.89 1.13

92.4 85.2 59.7 44.4 53.8 39.0

0.63 0.67 0.67 0.71 0.61 0.62

87.6 87.2 68.5 56.6 26.0 23.9

0.71 0.73 0.68 0.71 0.99 0.93

201.7 75.8 28.4 13.0 4.2 2.2

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

1.45 1.4 1.35 1.3

(b)

172 mV

MSA 5 M MSA 4 M MSA 3 M 196 mV MSA 2 M

1.25

E vs. (Ag/Ag2SO4) / V

E vs. (Ag/Ag2SO4) / V

(a)

187 mV

1.2 1.15 1.1 -4

-3.5

-3

-2.5

-2

1.45 1.4 1.35

35

MSA 1 M MSA 0.5 M MSA 0.25 M

380 mV

170 mV

1.3 1.25 1.2

94 mV

1.15 1.1 -3.5

-1.5

-3

-2.5

1.4 1.35 1.3

log (jstat / A.cm ) ~183 mV

MSA 5 M MSA 4 M MSA 3 M MSA 2 M

1.25 1.2 1.15 1.1 -4

-3.5

-3

-2.5 -1

-1.5

-1

log (jstat / A.cm )

(d)

1.45

E vs. (Ag/Ag2SO4) / V

E vs. (Ag/Ag2SO4) / V

1.45

-2 -2

-2

(c)

230 mV

-2

1.4 1.35

MSA 1 M MSA 0.5 M MSA 0.25 M

282 mV

453 mV

1.3 1.25

110 mV

1.2 1.15 1.1 -4

-1.5

138 mV

-3.5

-3

-1

log [ (Rct + Rads) / Ω ]

-2.5

-2

-1.5

-1

log [ (Rct + Rads)-1 / Ω-1]

Fig. 7. Tafel plots of the stationary current density of the OER for (a) high- and (b) low-concentrated MSA electrolytes, and corresponding Tafel plots using 1/(Rct + Rads ) for (c) high- and (d) low-concentrated MSA electrolytes (for 0.25 M MSA, Rct = 0).

1.4

(b) 1.45

H2SO4 3 M H2SO4 0.5 M

1.35

E vs. (Ag/Ag2SO4) / V

E vs. (Ag/Ag2SO4) / V

(a) 1.45

175 mV

1.3 1.25 1.2 1.15 1.1 -3.5

-3

-2.5

-2

-1.5

-1

1.4

H2SO4 3 M H2SO4 0.5 M

1.35 193 mV

1.3 1.25 1.2 1.15 1.1 -3.5

-3

log (jstat / A.cm-2)

-2.5

-2

-1.5

log (R-1 / Ω-1)

Fig. 8. Tafel plots of the OER in H2 SO4 electrolyte 0.5 and 3 M, using (a) stationary current density and (b) inverse of resistance from EIS data. Equivalent circuit models used: equivalent circuit ‘c’ of Table 2 for H2 SO4 = 0.5 M, and a single parallel RC element for H2 SO4 = 3 M.

in term of resistance. Tafel slopes for MSA concentrations ranging from 2 to 5 M, i.e. for highly concentrated electrolyte, are similar in both E vs. log j and E vs. log(R−1 ) plots, with values close to 180 mV. For less concentrated electrolytes (0.25–1 M), Tafel slopes are changed in a complex way, with a lower value for MSA 1 M, higher value for MSA 0.25 M and the appearance of a second higher slope at high potentials for MSA 0.5 M. This complex evolution of Tafel slopes upon decreasing MSA concentration could be a result of the growth of the hydrated layer at the PbO2 /electrolyte interface

once a sufficient amount of free water molecules in the electrolyte is available, as previously discussed. A growth of the hydrated layer, containing the active centers PbO(OH)2 , leads to an enhancement of the electrocatalytic activity of the material as mentioned by Amadelli et al. [21]. This phenomenon is observed on the Tafel plots of Fig. 7b and d in the case of MSA 1 M, for which one can note the emergence of this hypothetical hydrated layer in the complex plane in Fig. 5a. When acid molarity is further decreased, Tafel slopes raise as displayed by Fig. 7b and d for MSA 0.25 and 0.5 M, whose complex

Table 5 Parameters extracted from Nyquist diagrams for MSA 0.25 M. Equivalent circuit ‘c’ of Table 2. E (V)

T1 (mF cm−2 sn−1 )

n1

R1 ( cm2 )

T2 (mF cm−2 sn−1 )

n2

R2 ( cm2 )

1.1 1.15 1.2 1.25 1.3 1.35 1.4

2.8 3.7 3.0 2.3 2.5 2.4 1.6

0.71 0.67 0.66 0.68 0.69 0.71 0.83

1.1 1.2 1.5 1.6 1.5 1.7 1.3

173.6 124.5 95.8 78.2 69.0 54.0 42.4

0.51 0.53 0.53 0.54 0.57 0.58 0.58

166.7 40.9 33.7 20.4 12.2 7.9 5.7

36

A. Oury et al. / Electrochimica Acta 63 (2012) 28–36

plane plots display a pronounced semicircle towards high frequencies (Fig. 5a) probably due to a thicker hydrated layer. This latter could act as a barrier for electronic transfer due to its increased electric resistance, counterbalancing electrocatalytic activity. The presence of a higher slope towards more positive potentials for MSA 0.5 M (Fig. 7b and d) could be ascribed to ohmic drop in the hydrated layer which slows down the OER kinetics. The same behaviour is observed in Fig. 8a and b for sulphuric acid medium with a single 175 mV slope for H2 SO4 3 M and a rising slope upon increasing potential in the stationary plots ‘E vs. log j’ for less acidic electrolyte (H2 SO4 0.5 M). It is interesting to note that the value of 1/R (or jstat ) for a given potential in MSA electrolyte decreases when MSA concentration is increased, and is lower than that in sulphuric acid. 4. Conclusions Since evolution of oxygen is an undesirable reaction in the lead/methanesulfonic acid redox flow battery technology proposed by Pletcher et al., it was investigated on ␣-PbO2 electrodes in methanesulfonic acid at different concentrations, by cyclic voltammetry and electrochemical impedance spectroscopy. Cyclic voltammetry revealed that the OER Tafel slope was rather constant with MSA concentration and temperature, with values ranging from 180 to 240 mV. Activation energies calculated from Arrhenius plots ranged from 35 to 70 kJ mol−1 , without significant variations when changing the pH. The Tafel coefficient b of the voltammetric measurements was found to be similar in MSA and H2 SO4 medium, with higher current densities in H2 SO4 . While the difficulties associated with the calculation of the double layer capacitance and the pseudo-capacitance were ascribed to the high roughness of the lead dioxide surface, impedance measurements revealed the possible growth of a hydrated lead dioxide layer when decreasing MSA concentration below 1 M, which first enhanced the OER kinetics and acts as a barrier for electronic transport for further pH decrease or potential increase. Hence it can be concluded that two kinetics behaviours are possible: one for high acid concentrations with constant Tafel slopes and another for low acid concentrations with varying Tafel slopes when the hydrated layer becomes significant. On the other hand, 1/R value as well as stationary current density for a given potential were found to be lower in MSA than in H2 SO4 medium, as it was found by cyclic voltammetry.

It can also be emphasized that a high MSA concentration is preferable for limiting O2 evolution during the lead/MSA redox flow battery charge. Acknowledgements The authors are grateful to ADEME (French environment and energy management agency) for its financial support of the thesis within which this study was carried out. They also thank David Brun-Buisson for the XRD analysis, and Julien Laurent for the SEM images. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

A. Hazza, D. Pletcher, R. Wills, Phys. Chem. Chem. Phys. 6 (2004) 1773. D. Pletcher, R. Wills, Phys. Chem. Chem. Phys. 6 (2004) 1779. D. Pletcher, R. Wills, J. Power Sources 149 (2005) 96. A. Hazza, D. Pletcher, R. Wills, J. Power Sources 149 (2005) 103. D. Pletcher, H. Zhou, G. Kear, J. Low, F.C. Walsh, R. Wills, J. Power Sources 180 (2008) 621. D. Pletcher, H. Zhou, G. Kear, J. Low, F.C. Walsh, R. Wills, J. Power Sources 180 (2008) 630. X. Li, D. Pletcher, F.C. Walsh, Electrochim. Acta 54 (2009) 4688. J. Collins, G. Kear, X. Li, J. Low, D. Pletcher, R. Tangirala, D. Stratton-Campbell, F.C. Walsh, C. Zhang, J. Power Sources 195 (2010) 1731. J. Collins, X. Li, D. Pletcher, R. Tangirala, D. Stratton-Campbell, F.C. Walsh, C. Zhang, J. Power Sources 195 (2010) 2975. R. Wills, J. Collins, D. Stratton-Campbell, J. Low, D. Pletcher, C. Walsh, J. Appl. Electrochem. 40 (2010) 955. M.D. Gernon, M. Wu, T. Buszta, P. Janney, Green Chem. 1 (1999) 127. I. Sirés, C.T.J. Low, C. Ponce-de-León, F.C. Walsh, Electrochim. Acta 55 (2010) 2163. P. Jones, R. Lind, W.F.K. Wynne-Jones, Trans. Faraday Soc. 50 (1954) 972. P. Ruetschi, B.D. Cahan, J. Electrochem. Soc. 104 (1957) 406. E.R. Kötz, S. Stucki, J. Electroanal. Chem. 228 (1987) 407. J.C.K. Ho, G. Tremiliosi Filho, R. Simpraga, B.E. Conway, J. Electroanal. Chem. 366 (1994) 147. D. Pavlov, J. Electrochem. Soc. 139 (1992) 3075. B. Monahov, D. Pavlov, J. Appl. Electrochem. 23 (1993) 1244. D. Pavlov, B. Monahov, J. Electrochem. Soc. 145 (1998) 70. R. Amadelli, L. Armelao, A.B. Velichenko, N.V. Nikolenko, D.V. Girenko, S.V. Kovalyov, F.I. Danilov, Electrochim. Acta 45 (1999) 713. R. Amadelli, A. Maldotti, F.I. Danilov, A.B. Velichenko, Electrochim. Acta 534 (2002) 1. D.V. Franco, L.M. Da Silva, W.F. Jardim, F.C. Boodts, J. Braz. Chem. Soc. 17 (2006) 746. B. Chen, Z. Guo, H. Huang, X. Yag, Y. Cao, Acta Metall. Sin. 22 (2009) 373. E.C.G. Rufino, M.H.P. Santana, L.A. De Faria, L.M. Da Silva, Chem. Papers 64 (2010) 749. L.M. Da Silva, L. De Faria, J.F.C. Boodts, Electrochim. Acta 48 (2003) 699. A.C. Ramamurthy, T. Kuwana, J. Electroanal. Chem. 135 (1982) 243. C.H. Hsu, F. Mansfeld, Corrosion 57 (2001) 747. X. Li, D. Pletcher, F.C. Walsh, Chem. Soc. Rev. 40 (2011) 3879.