Materials Science Forum Vols. 514-516 (2006) pp. 1391-1395 online at http://www.scientific.net © (2006) Trans Tech Publications, Switzerland

Electrocatalytic Behavior of Perovskite-related Cobaltites and Nickelates in Alkaline MediA S.K. Poznyak,1,2,a V.V. Kharton,1,2 J.R. Frade1, A.A. Yaremchenko1, E.V. Tsipis1, I.P. Marozau1 and M.G.S. Ferreira1 1

Department of Ceramics and Glass Engineering, CICECO, University of Aveiro, 3810-193 Aveiro, Portugal

2

Research Institute for Physical Chemical Problems, Belarusian State University, Leningradskaya St. 14, 220050 Minsk, Belarus a

[email protected]

Keywords: perovskite-related oxides, lantanium-strontium cobaltite, nikelate, oxygen evolution, electrocatalytic activity Abstract Dense ceramic anodes of perovskite-type La1-x-ySrxCo1-zAlzO3-δ ( x = 0.45-0.70; y = 00.05; z = 0-0.20) and K2NiF4-type La2Ni1-xMexO4+δ (Me = Co, Cu; x = 0-0.20), synthesized by the glycine-nitrate technique, were assessed for oxygen evolution in alkaline media. The lowest overpotentials are observed for (La0.3Sr0.7)0.97CoO3-δ, which exhibits a significant oxygen deficiency in combination with high conductivity associated with the A-site cation nonstoichiometry compensation mechanism via Co4+ formation. Perovskite-type cobaltite anodes are essentially stable in alkaline solutions, whilst La2NiO4-based electrodes exhibit degradation at the potentials where the oxygen evolution occurs, probably due to the electrochemical oxygen intercalation in the lattice. Introduction Electrochemical evolution and reduction of oxygen in alkaline media are of considerable interest for numerous technological applications, including secondary metal-air batteries, water electrolysis, low-temperature fuel cells, electrosynthesis and metal processing. One promising group of highly conductive materials, which can be used as bi-functional electrodes with a substantial performance for both reactions, comprises perovskite-related compounds based on (La,Sr)CoO3-δ and La2NiO4+δ [1-3]. Literature data on the electrocatalytic activity of perovskite ceramics, such as La1-xSrxCoO3-δ, are however contradictory. This results, in particular, from microstructural factors due to the use of different synthesis methods and processing conditions. Although a large surface area and, thus, high porosity are desirable for practical applications, optimization of the anode compositions requires to re-assess electrochemical properties of high-density ceramic materials, all synthesized by the same method and studied under similar conditions. The compositions with maximum performance can then be used to develop nanostructured electrodes with a great surface area. The present work was focused on the evaluation of La1-x-ySrxCo1-zAlzO3-δ (x = 0.45-0.70; y = 0-0.05; z = 0-0.20) and La2Ni1-xMexO4+δ (Me = Co, Cu; x = 0-0.20) anodes; their structure and electrical properties were reported in previous works [4-7]. All materials were synthesized via glycine-nitrate process (GNP), a self-combustion method using glycine as fuel and nitrates of the metal components as oxidant [8]. The density of all anodes was > 93% of their theoretical density calculated from X-ray diffraction (XRD) data. Experimental For the electrochemical measurements, dense ceramic disks were mechanically polished, followed by ultrasonic washing with acetone and distilled water. To provide an ohmic contact, Pt paste was

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 193.136.173.41-01/10/06,22:46:52)

1392

Advanced Materials Forum III

deposited on the back side of the pellets and annealed at 1000oC; then a copper wire was connected to the Pt layer using a silver glue and the corresponding side was isolated by an Araldite epoxy resin. The electrochemical measurements were carried out using an Autolab potentiostat and a standard three-electrode two-compartment cell. A Pt-foil and an Ag/AgCl/KCl (saturated) electrode were used as the counter and reference electrodes, respectively. Description of experimental techniques and equipment, used for characterization of the ceramic materials, can be found elsewhere ([4-6] and references cited). Results and discussion The activity towards the oxygen evolution reaction was evaluated at 20oC measuring IR-free voltammograms in a 1 M KOH solution using a rather slow scan rate, 1 mV s-1 (Fig. 1). -1

(b)

(a) -2

-2 -3

La0.3Sr0.7CoO3 (La0.3Sr0.7)0.97CoO3

-4

La0.3Sr0.7Co0.8Al0.2O3 La0.45Sr0.5CoO3

-5 -6 0.4

log j ( A cm )

-2

log j ( A cm )

-1

-2

-3 La2NiO4 La2Ni0.9Co0.1O4

-4

La2Ni0.8Cu0.2O4

La0.55Sr0.4CoO3 La0.65Sr0.3CoO3

0.5

0.6

0.7

0.8

Potential / V vs. Ag/AgCl

Pt -5 0.4

0.6

0.8

1.0

1.2

Potential / V vs. Ag/AgCl

Fig. 1. Anodic polarization curves of La1-x-ySrxCo1-zAlzO3-δ (a), and La2Ni1-xMexO4+δ and Pt electrodes (b) in 1 M KOH solution. The curves were recorded after pretreatment of the electrodes at anodic galvanostatic polarization (10 mA cm-2) for 10 min. Table 1. Electrode kinetics parameters for oxygen evolution reaction (1 M KOH, 20oC). Electrode material

Tafel slope [mV dec-1]

La0.3Sr0.7CoO3 (La0.3Sr0.7)0.97CoO3 La0.3Sr0.7Co0.8Al0.2O3 La0.45Sr0.5CoO3 La0.55Sr0.4CoO3 La0.65Sr0.3CoO3 La2NiO4 La2Ni0.9Co0.1O4 La2Ni0.8Cu0.2O4 Pt

60 57 85 87 57 62 77 80 83 42.5 at lower E 199 at higher E

Current density [mA cm-2] at E= 0.52 V at E= 0.72V 0.0256 47 0.0845 89 0.17 38.5 0.13 14.5 0.0065 22 0.013 20

Cdl [µF cm-2] 122 1610 2710 21 46 -

Table 1 compares the Tafel slope and apparent current densities at two potentials selected in the low and high overpotential regions. To estimate the true current density, we attempted to determine the roughness factor (σ). In principle, such estimations could be made from the values of capacitive current, measured cycling the potential in a narrow range (≈ 50 mV) within a double-layer region near the rest potential, i.e. in the region free of pseudo-capacitive and faradaic currents [2]. Fig.2 presents typical cyclic voltammograms (CVs) for one electrode, La0.45Sr0.5CoO3; similar CVs were obtained for other electrodes. The double layer capacity, Cdl (Table 1), was calculated from the

Materials Science Forum Vols. 514-516

1393

-2 0

20

40

60

80

-1

Scan rate / mV s

0.2

-2

80 60

0.0

40 20 10

0.0 -0.2

-4.0

E = 0.6 V

log j

0.1

log j ( A cm )

0.4

-1

0.2 -2

0.6

j / mA cm

Current density / mA cm

-2

slope of j vs. scan rate plot taken at the middle potential of the scanned range (inset in Fig.2). The roughness factor was then assessed as σ = Cdl/60, assuming the value of 60 µF cm-2 to be the capacitance of an ideal smooth oxide surface under anodic polarization [9]. The Cdl and, thus, σ values are strongly dependent on the electrode composition (Table 1). This is, however, inconsistent with the results of scanning electron microscopy (SEM), which showed quite smooth surfaces and essentially similar microstructures for most anodes. Such a behavior suggests that the method to estimate roughness factors cannot be used for the mixed-conducting materials, where the current recorded near the rest potential may be contributed by the faradaic currents associated with electrochemical oxygen intercalation. Respectively, in this work the data were normalized to the geometrical electrode area.

-4.5

-3

-1

0 -

log [OH ] (mol)

-4

Electrolyte: 0.1 M KOH 0.2 M KOH 2 M KOH

-5 0.10

0.12

0.14

0.16

0.18

0.20

Potential / V vs. Ag/AgCl

Fig. 2. Cyclic voltammograms of La0.45Sr0.5CoO3 electrode at different scan rates in 1 M KOH solution (20oC), and the plot of pseudocapacitive current vs. scan rate (inset).

-6 0.4

0.5

0.6

0.7

0.8

Potential / V vs. Ag/AgCl

Fig. 3. Tafel plots and plot of log j against log [OH-] (inset) for oxygen evolution on La0.55Sr0.4CoO3 electrode in KOH solutions with different concentration.

For most cobaltites, the Tafel slope is close to 60 mV dec-1, in agreement with data on similar compounds [2, 10]. In the high-overpotential range the electrocatalytic activity decreases in the order (La0.3Sr0.7)0.97CoO3 > La0.3Sr0.7CoO3 > La0.3Sr0.7Co0.8Al0.2O3 > La0.55Sr0.4CoO3 ≈ La0.65Sr0.3CoO3 > La0.45Sr0.5CoO3. Despite the electrochemical reaction mechanisms which will be analyzed in a separate work, this trend shows that the behavior of cobaltite electrodes in alkaline media is affected by several factors, namely, oxygen nonstoichiometry, Co4+ content determining the level of p-type electronic conductivity and surface concentration of catalytically active centers, and phase purity. The key role of the oxygen vacancy concentration, which increases with increasing strontium content, is obvious considering overpotentials as function of the La/Sr ratio. On the other hand, further Sr additions may lead to the secondary phase segregation and to increasing tendency to degradation due to hydroxide formation at the electrode surface, and are thus undesirable. For the materials with La/Sr ratio close to unity, the behavior seems strongly influenced by the cobalt oxide segregation detected by SEM and XRD (Table 2). Note that no phase separation was earlier observed in La0.65Sr0.3CoO3-δ prepared by the standard ceramic route [7], and a substantially high activity was indeed found for the latter material. For the GNP-synthesized ceramics studied in this work, the lowest overpotentials are observed in the case of A-site deficient (La0.3Sr0.7)0.97CoO3-δ. This phase exhibits a significant oxygen nonstoichiometry in combination with a relatively high conductivity (Table 2) due to the dominant mechanism of the cation vacancy compensation mechanism via Co4+ formation [6]. For instance, the Co4+ concentration per unit formula of (La0.3Sr0.7)0.97CoO3-δ, calculated from thermogravimetric data, is 0.447; in the case of La0.3Sr0.7Co0.8Al0.2O3-δ, this value is 0.443 at room temperature. In the range of relatively low overpotentials, the electrochemical activity of cobaltite-based electrodes correlates with Cdl values,

1394

Advanced Materials Forum III

indicating that oxygen intercalation plays a primary role. Compared to the perovskite-type cobaltites, the overpotentials of nickelate-based electrodes are substantially higher. The reaction order with respect to [OH-] was analyzed studying the effect of alkali concentration (0.1 to 2 M) at 20oC. The corresponding anodic polarization curves recorded for La0.55Sr0.4CoO3 are presented in Fig.3; the reaction order, pOH, was approximately 0.75. Similar values were also observed for other cobaltite electrodes studied in this work. Table 2. Properties of the perovskite-related materials at room temperature. Phase impurities

Composition

Structure

perovskite La0.3Sr0.7CoO3-δ perovskite La0.3Sr0.7Co0.8Al0.2O3-δ perovskite (La0.3Sr0.7)0.97CoO3-δ CoO (~1%) perovskite La0.45Sr0.50CoO3-δ CoO (~1%) perovskite La0.55Sr0.40CoO3-δ CoO (~1%) perovskite La0.65Sr0.30CoO3-δ K2NiF4-type La2NiO4+δ K2NiF4-type La2Ni0.9Co0.1O4+δ K2NiF4-type La2Ni0.8Cu0.2O4+δ * The lattice parameters are given for pseudocubic unit cell

Lattice parameters [nm]

δ

a = 0.3836 a = 0.3847 a = 0.3837 a = 0.3832* a = 0.3844* a = 0.3836* a = 0.3865; c = 1.2691 a = 0.3864; c = 1.2664 a = 0.3858; c = 1.2787

0.26 0.13 0.16 0.15 0.17 0.12

0.80

-1 E = 0.48 V E = 0.55 V

j = 22 mA cm

-3 -4 -5

-3

2.8

3.0

3.2 3

3.4

1/T x 10 / K

-1

o

T = 20 C o T = 40 C o T = 62 C o T = 78 C

-4

-5 0.2

0.4

0.6

Potential / V vs. Ag/AgCl

Fig. 4. Tafel plots and Arrhenius plots (inset) for oxygen evolution reaction on La0.55Sr0.4CoO3 electrode in 1 M KOH solution at different temperatures.

Potential / V

log j

-2

log j ( A cm )

-2

-2

Conductivity [S/cm] 2.1×103 5.4×102 3.6×103 3.3×102 3.8×103 39 3.8 24

-2

0.75

after changing the electrolyte

0.70

0.65

0

5

10

15

20

25

30

35

40

Time / h

Fig. 5. Time evolution of the electrode potential of La0.55Sr0..4CoO3 under galvanostatic polarization (1 M KOH, 20±1oC, j = 22 mA cm-2).

The apparent activation enthalpy (∆Hel)E was calculated from the slope of the Arrhenius plot, ∂ log j/∂T-1, at a constant potential chosen in the Tafel region of the anodic polarization curves (Fig.4). In the case of La0.55Sr0.4CoO3-δ electrode, the value of the ∆Hel was about 72 kJ mol-1. The long-term stability was examined under galvanostatic polarization in 1 M KOH solution; one typical example is presented in Fig.5. The results show fairly stable overpotentials, which even decreased during the first 5-7 hours and then became essentially constant. For cobaltite anodes, SEM revealed no considerable alterations of the surface morphology (Figs.6a and 6b). Only a very thin film with interference colors appears at the surface after prolonged anodic polarization; Co-rich grains, if formed during the synthesis, become more visible. On the contrary, the morphology of nickelate-based anodes exhibits dramatic changes after the tests in alkaline media, including a presence of relatively thick dark-brown films and an increase of the surface

Materials Science Forum Vols. 514-516

1395

roughness (Figs. 6c and 6d). The energy-dispersive spectroscopy (EDS) showed that the Ni/La ratio in these films is lower than initial. This indicates that facile electrochemical intercalation of oxygen into the surface layers of La2Ni1-xMexO4+δ may cause a fast decomposition of nickelate anode, probably via formation of NiOOH and La(OH)3. A similar mechanism, in combination with low concentration of catalytically active centers on the nickelate surface with respect to the perovskitetype cobaltites, seems responsible for the lower activity of La2NiO4-based electrodes (Fig.1b). (a) (c)

(b)

(d)

Fig. 6. SEM micrographs of the surface of La0.55Sr0.4CoO3 (a,b) and La2NiO4 (c,d) electrodes before electrochemical measurements (a,c) and after anode operation in 1 M KOH solution at 20oC (b,d). Acknowledgements Financial support from the ULCOS program and from the FCT, Portugal (projects BD/6827/2001 and BPD/11606/2002) is gratefully acknowledged. References [1] S. Trasatti, and G. Lodi: in Electrodes of Conductive Metallic Oxides, Part B, S. Trasatti, Editor (Elsevier, Amsterdam 1981). [2] J. O’M. Bockris and T. Otagawa: J. Electrochem. Soc. Vol. 131 (1984), p.290 [3] S. Bhavaraju, J.F. DiCarlo, D.P. Scarfe, A.J. Jacobson and D.J. Buttrey: Solid State Ionics Vol. 86-88 (1996), p.825 [4] V.V. Kharton, A.A. Yaremchenko, A.L. Shaula, M.V. Patrakeev, E.N. Naumovich, D.I. Logvinovich, J.R. Frade and F.M.B. Marques: J. Solid State Chem. Vol. 177 (2004), p. 26 [5] A.V. Kovalevsky, V.V. Kharton, V.N. Tikhonovich, E.N. Naumovich, A.A. Tonoyan, O.P. Reut and L.S. Boginsky: Mater. Sci. Eng. B Vol.52 (1998), p.105 [6] V.V. Kharton, E.V. Tsipis, I.P. Marozau, A.A. Yaremchenko, A.P. Viskup, J.R. Frade and E.N. Naumovich: Mater. Sci. Eng. B (2005), submitted [7] V.V. Kharton, P.P. Zhuk, A.A. Tonoyan, T.E. Zhabko and A.A. Vecher: Inorg. Mater. Vol. 27 (1991), p.2240 [8] M.W. Murphy, T.R. Armstrong and P.A. Smith: J. Am. Ceram. Soc. Vol. 80 (1997), p. 165 [9] S. Levine and A.L. Smith: Discuss. Faraday Soc. Vol. 52 (1971), p. 290 [10] R.N. Singh, B. Lal: Int. J. Hydrogen Energy Vol. 27 (2002), p.45