Impedance spectroscopy of dye-sensitized solar cells: Analysis of measurement and fitting errors

Abstract

Errors in impedance measurements made on dye-sensitized solar cells (DSCs) and in subsequent model-based parameter extraction have been analyzed. The variance in the impedance was determined from replicate measurements using a measurement model approach. Measurement models were also used to identify corrupt data that were inconsistent with the Kramers-Kronig relations, providing justification for their rejection prior to further analysis. A common equivalent circuit model containing a transmission line was fitted to the refined data, which was weighted by the experimentally determined variance. For comparison, fits were also performed with the common modulus weighting strategy. Parameter confidence intervals were constructed from approximate standard errors and also from Monte-Carlo simulations based on the measured variance, which are presumably more accurate. Both weighting strategies yielded practically identical results for the main transmission line parameters, validating use of modulus weighting for routine work on DSCs. However, spectra could not be fitted to within the noise level, indicating that the model may need to be improved. Monte-Carlo confidence intervals were in all cases 10-20 tighter than those calculated from parameter standard errors for variance weighted fits and 3-15 times tighter than those for modulus weighted fits, implying the latter can be viewed as being pessimistic. © 2012 The Electrochemical Society.