Preferred position for the reference electrode in solid state electrochemistry
Introduction
The proper positioning of the reference electrode (RE) in solid state electrochemistry (SSE) has been in dispute in recent years [1], [2], [3], [4], [5]. The controversy arose when examining the electrode overpotential in solid oxide fuel cells (SOFCs) but this issue is, of course, of a wider interest. The common form of the samples considered for SOFCs is a planar one with a plane dimension of a few centimeters and a thickness of 1 mm or less. The common placement of the RE is alongside with the working electrode, WE, under test, i.e. on the same surface of the solid electrolyte. [6] This arrangement is experimentally convenient and allows to expose the RE and the WE to the same gas composition. Thick solid electrolytes (SEs) allow to place the RE on the side of the sample rather than on the flat surface. The RE is then exposed to a fixed gas, usually room air. However, the different gas phases at WE and RE and the high ohmic resistance of the thick SE, that may exceed by more than an order of magnitude the impedance of modern electrodes, make the determination of the overpotential difficult.
Winkler et al. [5] have shown that the first electrode arrangement is extremely sensitive to the exact positioning of the electrodes when the WE edge, facing the RE, and the counter electrode edge are opposite. Then, any misalignment of the WE edge with respect to the counter electrode edge introduces large changes in the current density near the edge of the electrodes and hence also in the overpotential as measured with respect to a nearby RE. Therefore, Winkler et al. suggested a different design of the cell in the form of a short tube closed at one end. This configuration which works well serves, however, for electrode evaluation only.
We shall show here that one can stick to the planar configuration and yet evaluate the electrode performance if the RE, WE and counter electrode, CE, are properly designed. We notice in passing that the sensitivity of the relative positioning of the working and counter electrode in the first arrangement can be overcome by making one larger than the other. The same can be achieved when the edge of one electrode is opposite the more central part of the opposite electrode as shown in Fig. 1.
Two problems remain, however, with the former arrangements. One is the ohmic contribution to the measured overpotential. To eliminate it, one uses current interruption. This requires fast switching on the microsecond scale to make sure that the electrode overpotential has no time to relax before being measured. The second problem arises from the nonuniformity of the current density in the WE even when the electrode is uniform. The current density near the electrode edge is different from that more towards the center of the electrode. This generates also a nonuniform distribution in the local overpotential, as the overpotential depends on the current density. The single value measured, overpotential, η, is thus an average of that distribution. An RE placed on one side of the WE, at a distance normally small with respect to the diameter of WE, yields a distorted average overpotential in which the weight of the local overpotential from the nearby edge is enhanced.
Section snippets
Suggested electrode arrangement
In designing an RE, one should take into consideration the following:
(a) It is desired to reduce the ohmic overpotential measured between WE and RE. This is achieved in liquid state electrochemistry by placing the RE inside the liquid electrolyte close to the WE. Obviously, this is not practical in SSE and a different solution has to be found to place the reference close to the WE.
(b) The overpotential should reflect the true average for the WE.
(c) The reference electrode should have only a
Numerical simulation
The question of interest is the voltage distribution in the gaps in the WE where stripes of the RE are located. We examine the deviation of the voltage in these gaps from the voltage, Vapp, applied onto the WE (and measured relative to CE). In calculating the potential distribution, we take into consideration that the RE is made of a highly conductive material imposing a uniform potential under each RE stripe. We do not “connect” all RE stripes in parallel in the calculation, i.e. we do not
Results of numerical simulations
Fig. 3, Fig. 4, Fig. 5 exhibit the potential distribution along IF, the surface of the SE just under the WE and RE, for low-impedance WE. The configurations considered are shown in Fig. 2d–f and δ′=δ″=δ‴=δ =t, δ/d=0.01. Fig. 6, Fig. 7 show the potential distribution there for a WE with impedance. The horizontal short lines are equal potential curves just under the stripes having a high conductivity.
When WE is highly conductive as is RE (σel=105 S/cm), then a uniform potential exists under both
Cylindrical samples
The I–V relations in cells with cylindrical symmetry can also be evaluated numerically in a straightforward manner. Such a cell would have a cross section in the form of Fig. 2d–f, being symmetric with respect to reflection through a vertical line at the center of the rectangle. The WE and RE are then concentric circles. (For practical reasons, the circles have to be cut, joined near the cut to form an electrical contact, but we neglect this in our discussion here). Due to the cylindrical
Experimental
We have so far tested gold electrodes. These electrodes exhibit a high overpotential. In this case, the ohmic IR drop in the bulk is relative small and η is hardly dependent on the position of a reference electrode. This was indeed observed. Gold electrodes of a form modified from Fig. 2e were applied on a single crystal of Y2O3-stabilized ZrO2 (YSZ) with a thickness of 0.5 mm and a size of 50×50 mm2. YSZ is a solid electrolyte, conducting oxygen ions at elevated temperatures. The modification
New measurement possible with the given electrode arrangement
The present electrode design, in particular that shown in Fig. 2c with δ′∼δ″, allows for characterization of the SE/electrode interface by measurements in the lateral direction along that interface. A current is driven between WE and RE, being further controlled by a bias between WE and CE. This configuration may yield new information on the electrode properties as well as on the stoichiometry of the SE near the interface and its effect on the electrode overpotential. We have used this for
Summary
Using modern methods of accurate pattern formation in thin layers enables to place the RE at a close distance, δ‴, of a few micrometers, from the WE. In particular, it is possible to place the RE in narrow gaps in the WE. The distance, δ‴, between the RE and WE nearest parts is significantly smaller than the thickness, d, of common solid electrolytes. The technique required for preparing the electrode is still not widely used in SSE but the effort may be justified for investigating the basic
Acknowledgements
This work was supported by a grant form the Ministry of Infra Structure, The State of Israel.
References (11)
- et al.
Solid State Ionics
(1994) - et al.
Solid State Ionics
(1997) - et al.
Solid State Ionics
(1994) - et al.
Solid State Ionics
(2000) - et al.
J. Electrochem. Soc.
(1989)