Multiple model impedance spectroscopy techniques for testing electrochemical systems

Rapid battery impedance techniques have become mandatory in most battery management programs. Methods available today use a one-frequency technique at relatively high frequencies 83-90 Hz Huet, F. (1998) for this type of analysis. The results so far demonstrate reasonable accuracy at determining degradation between 0-75% battery capacity Markle, G.J. (1993). However, the same data also indicates a reduced correlation when the capacity of the battery is higher than 80% Noworolski, Z., et al. (2002). According to IEEE 450, any battery below 80% must be replaced. This creates a serious problem since the detection of the replacement boundary is directly in the region where high frequency impedance techniques become least reliable. A possible solution to this problem is to lower the frequency and investigate if better correlation may be attained. This assumption is supported by the literature where Tenno et al. provides experimental evidence for greater correlation at low frequency Tenno, A., et al. (2002). However, this raises a practical dilemma since cells must be tested in reasonable timescales. If several low frequency acquisitions are performed using standard techniques, the timescale for a complete test including modeling calculations would require /spl sim/1 minute per cell. This is approximately 6 times longer then currently available techniques and is not practically feasible. This problem may be resolved by inspecting our initial conditions and assumptions. Physically in combined solid and liquid electrochemical systems, a strong increase in nonlinearity is induced when voltages and currents increase Macdonald, J.R. (1987). However, if we choose the excitation voltage to be less than the thermal voltage, it may be demonstrated that the response will be linear.