Air Electrode

Air Electrode

 

This work concerned the positive electrode in the rechargeable zinc/air battery.  From the outset, it was recognised that two separate but interacting developments were essential to constructing a successful battery

(a) the development of a catalyst that would allow both oxygen reduction and evolution at low overpotentials in concentrated alkali and this catalyst should not contain precious metals.

(b) placing this catalyst in an electrode structure that allowed both oxygen reduction and evolution at high current density.  This electrode structure should be free of carbon components as these are subject to corrosion in concentrated alkali at the temperatures envisaged for battery operation (323 -353 K). It was clear that the electrode design would need to be significantly different from that used either in alkaline fuel cells or alkaline water electrolysers.

Catalyst Development

Initial studies were carried out using voltammetry with catalyst/PTFE coated, glassy carbon rotating disc electrodes in 1 - 8 M KOH at both 298 and 333 K. A large number of possible catalysts were screened (mainly spinels and perovskites) and compared with Pt; the minimum difference in potential between those for oxygen reduction and oxygen evolution being sought.  These experiments showed that the spinel NiCo2O4 was a good choice. Doping this spinel with other elements did no improve performance but it was found that the preparation procedure did influence performance. A material prepared by precipitation of Ni/Co with a mixture of hydroxide and carbonate followed by washing and thermal treatment at 648 K was preferred. This and other samples were characterised by   elemental analysis, X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), BET surface area determination and electrochemical methods. These suggested that particle size was important.

The mechanisms of both oxygen reduction and evolution at NiCo2O4 were investigated [1]. Firstly, it is clear that the reactions occur with the catalyst with transition metal ions in different oxidation states.  Figure 3(a) shows  a voltammogram for the spinel in 1 M KOH. A well-defined anodic peak is seen prior to O2 evolution and a rather more drawn out cathodic process on the reverse scan. The charge in the anodic and cathodic peaks are equal but represent only a fraction of the charge required to fully oxidise the spinel layer – the peak corresponds only to surface oxidation.  Also the 1st and 10th cycles are very similar confirming that a complete cycle returns the spinel to its original oxidation state. The open circuit potentials following O2 evolution and reduction relax, figure 3 (b), in different ways. After reduction, the potential takes up its initial value almost immediately but after O2 evolution the potential decays over a period of an hour as the higher oxidation states react slowly with water to return them to the initial state.

Figure 3  (a) 1st and 10th cyclic voltammograms. (b) relaxation of the open circuit potential following periods of O2 evolution and reduction. NiCo2O4 in 1 M KOH at 298 K. Stationery disc.

 

Figure 4 show RDE responses for O2 reduction. Figure 4(a) compares the responses at 4 surfaces. It can be seen that both the reduction potential and the limiting current densities depend strongly on the surface. In terms of potential, NiCo2O4 is not as good as Pt but in the battery this difference is diminished because NiCo2O4 is a better catalyst than Pt for O2 evolution. At C and Co3O4 the reduction is a 2e- reduction to H2O2 while at NiCo2O4 and Pt it is a 4e- reaction. This conclusion was confirmed by RRDE studies when at NiCo2O4 almost no H2O2 was detected at the ring electrode.

Figure 4  (a) Voltammograms for O2 reduction at Pt black, XC72, NiCo2O4 and Co3O4 (b) rotation rate dependence for NiCo2O4. Oxygen saturated 1 M KOH at 298 K.

 

 

Figure 4(b) shows the rotation rate dependence for O2 reduction ay NiCo2O4; it can be seen that  the limiting current depends strongly on the rotation rate but the reduction is not mass transport controlled (IL not proportional to ω1/2).  Overall, we believe that there is a rate limiting chemical reaction in a 4e- reduction (probably O-O bond cleavage on the spinel surface). Certainly, the fact that the spinel supports the full 4e reduction is highly advantageous to battery performance.

 

Electrode Structure

In this work package the GDE produced were 1 cm diameter discs and were tested in glass cells. Scale up to A4 electrodes in realistic flow cells was tackled in the unit cell and stack activities. Again it should be stressed that it was considered essential to develop GDE free from C powders and C paper and this required the testing of new materials and novel fabrication procedures. GDE were fabricated with current collectors consisting of both Ni and steel cloths and meshes as well as GDE supported on OH- conducting membranes. Superior long term stability was, however achieved with GDE where (a) a NiCo2O4/PTFE layer or (b) a spinel coated Ni powder/PTFE layer was pressed into a Ni foam current collector.  An example of the performance of these electrodes is reported in figure 5.

The influence of current density is illustrated by the data in figure 6 for Ni foam GDE prepared with a  NiCo2O4 coated Ni powde/PTFE layer.

 

Figure 5 (a) Comparison of Ni foam GDE prepared from Ni powder and NiCo2O4 powder and (b) cycling of NiCo2O4 powder GDE. 50 mA cm-2. 8 M KOH. 333K.

ass=MsoNormal>In this work package the GDE produced were 1 cm diameter discs and were tested in glass cells. Scale up to A4 electrodes in realistic flow cells was tackled in the unit cell and stack activities. Again it should be stressed that it was considered essential to develop GDE free from C powders and C paper and this required the testing of new materials and novel fabrication procedures. GDE were fabricated with current collectors consisting of both Ni and steel cloths and meshes as well as GDE supported on OH- conducting membranes. Superior long term stability was, however achieved with GDE where (a) a NiCo2O4/PTFE layer or (b) a spinel coated Ni powder/PTFE layer was pressed into a Ni foam current collector.  An example of the performance of these electrodes is reported in figure 5.

 

Figure 5 (a) Comparison of Ni foam GDE prepared from Ni powder and NiCo2O4 powder and (b) cycling of NiCo2O4 powder GDE. 50 mA cm-2. 8 M KOH. 333K.

A4 electrodes in realistic flow cells was tackled in the unit cell and stack activities. Again it should be stressed that it was considered essential to develop GDE free from C powders and C paper and this required the testing of new materials and novel fabrication procedures. GDE were fabricated with current collectors consisting of both Ni and steel cloths and meshes as well as GDE supported on OH- conducting membranes. Superior long term stability was, however achieved with GDE where (a) a NiCo2O4/PTFE layer or (b) a spinel coated Ni powder/PTFE layer was pressed into a Ni foam current collector.  An example of the performance of these electrodes is reported in figure 5.

 

Figure 5 (a) Comparison of Ni foam GDE prepared from Ni powder and NiCo2O4 powder and (b) cycling of NiCo2O4 powder GDE. 50 mA cm-2. 8 M KOH. 333K.

A4 electrodes in realistic flow cells was tackled in the unit cell and stack activities. Again it should be stressed that it was considered essential to develop GDE free from C powders and C paper and this required the testing of new materials and novel fabrication procedures. GDE were fabricated with current collectors consisting of both Ni and steel cloths and meshes as well as GDE supported on OH- conducting membranes. Superior long term stability was, however achieved with GDE where (a) a NiCo2O4/PTFE layer or (b) a spinel coated Ni powder/PTFE layer was pressed into a Ni foam current collector.  An example of the performance of these electrodes is reported in figure 5.

 

The influence of current density is illustrated by the data in figure 6 for Ni foam GDE prepared with a  NiCo2O4 coated Ni powde/PTFE layer.

Figure 5 (a) Comparison of Ni foam GDE prepared from Ni powder and NiCo2O4 powder and (b) cycling of NiCo2O4 powder GDE. 50 mA cm-2. 8 M KOH. 333K.

Figure 6 Potential vs. time responses during current density cycling of a spinel coated Ni/PTFE GDE. Current densities: 20, 50, and 100 mA cm-2 in 8 M NaOH at 333 K. Fresh GDE. Oxygen feed rate: 200 cm3 min-1.

 

 

Consortium Websites

CEST logo

E-on logo Fuma-Tech logo
GreenPower logo DNV KEMA logo
University of Seville logo University of Southampton logo

 

Funded by

FP7 Logo