The majority of published R&D in supercapacitors relies on testing at a small coin or Swagelok cell size. This is often due to a lack of availability of equipment and knowledge required to move to a larger scale. Whilst comparisons between materials and other components can be made at the small cell level, the actual performance in larger devices does not necessarily follow to the same extent. In many cases, a scaling factor can be applied to the values of energy and power density for an active material to give the corresponding values in large devices but this is not always the case. For the extrapolation of performance from small to large cell to be valid, many factors must be considered and matched to real devices such as electrode thickness, relative electrolyte amount, current collector properties and separator type. Whilst allowing an estimate of large device properties in this way from small cell level is useful, this approach does not consider many of the problems encountered when scaling up that also differ between cell geometries (cylindrical, pouch or prismatic) can result in predicted performances never being achieved.

In the current work, the development from coin cell to >800 F A5 sized pouch cell (the same geometric area as the batteries of the Nissan Leaf) is reported using commercially available activated carbon (YP50-F, Kuraray, Japan) as the active material and TEABF₄in acetonitrile electrolyte, with comparisons made throughout between the small and large scale. The scale-up of mixing of electrode ink is demonstrated from the 50 g to 10 Kg scale, with discussion of the effects of various components on the solid content achievable at the viscosities required for coating. Various parameters in electrode production are discussed in relation to their effects on cell performance at the coin and large pouch level including electrode thickness/loading and drying temperature profiles, the later of which being surprisingly shown to be responsible for delamination of electrode from current collectors and complete failure in large cells over time (not observed in coin cells). The effect of calendaring of electrodes at different temperatures is shown to produce a range of electrode densities, with the effect on adhesion, specific capacitance and ESR investigated in cell. It is shown that whilst electrode capacitance is, to a broad degree, unaffected by calendering, much benefit can be gained in terms of ESR and adhesion but that over-calendering has a major detrimental effect.

The production of A5 sized pouch cells from the electrodes produced is illustrated, with early cell testing results used to highlight the problems faced in supercapacitor pouch cells resulting from early capacitance fade, gas production and a lack of compression. These problems are addressed through the application of external compression (as is provided in module assemblies and by the spring in coin cells) and through the float ageing of cells to a stable capacitance and subsequent gas removal, with the differences in time and energy required at coin and pouch cell level to achieve this highlighted.

In conclusion, the effects and contributions of electrode thickness, number of electrodes, electrolyte mass and "parasitic" components such as packaging in pouch cells are considered in terms of both volumetric and gravimetric energy and power densities, contrasting the real device values obtained in scale-up with those of the active materials often reported and those predicted from coin cell results.

Figure 1