Charge storage in a carbon-electrode electric double layer capacitor (EDLC) occurs primarily via physical processes. Nevertheless, some charge is also stored via chemical processes, at least initially, which is due to active groups usually present on the surface of a fresh carbon electrode. Fundamentally, chemical charge storage processes are not as reversible as physical processes and they fade relatively quickly, usually after tens of hours on charge at elevated temperature or after a few tens of charge-discharge cycles. Importantly, charge storage irreversibility often generates gas that swells the EDLC package and eventually causes it to rupture. Physical processes, on the other hand, are highly reversible and typically provide reliable charge storage for thousands of hours and millions of charge/discharge cycles. The approach used to initially charge the capacitor may help to deactivate functional groups on an electrode’s surface and thus reduce its gas generation.

Electrolyte decomposition can also generate gas. In a carbon-electrode EDLC, decomposition is expected to be concentrated on the sharp points in an electrode where the electric field strength is highest. On a new electrode, the application of high voltage for short times (the “burn-in”) may help erode such points and thus reduce the electric field strength, which in this case is the driving force for gas generation. Then burn-in details could help to reduce gas generation and increase capacitor life.

Our study examines capacitor life as a function of the burn-in procedure. Groups of sealed EDLCs were fabricated in the Case Western Reserve University’s Electrochemical Capacitor Prototyping Laboratory (http://energy.case.edu/CECFF). A two-factor experimental design was used to assign the burn-in conditions for each of the nine groups, the factors being “applied voltage” and “voltage application time”. After burn-in, each group was aged at 65^oC, 2.7 V to measure its life. Responses measured during aging included the standard electrical parameters reported for EDLCs and the gas generation rate. EDLC life was analyzed using actual or defined failures, for example, package rupture caused by pressure build-up due to generated gas or 30% fade in capacitance. Weibull life distributions were derived for each group. Results from this study are presented with burn-in optimization details.