An uncertainty analysis of the spectral correction factor

Whenever a photovoltaic device is evaluated with respect to a reference spectrum, there is a spectral mismatch error. The photocurrent can be corrected for this error with a spectral correction factor, as is routinely done for primary and secondary terrestrial reference solar cell calibrations at laboratories throughout the world. The spectral correction factor has also been used to accurately measure the performance of cells and modules of any technology with respect to an arbitrary set of reference conditions. This uncertainty analysis considers the sensitivity of the spectral correction factor, and thus the measurement to which it applies, to choices of integration limits and uncertainties in the source irradiance and spectral response measurements from which it is derived. The uncertainty analysis involves Monte Carlo simulations of wavelength-dependent random errors and addition of systematic (bias) errors to the calculation's factors. It is demonstrated that error can be reduced or increased by reducing the wavelength range of the irradiance or spectral response measurement. The uncertainty in the spectral correction factor for standard lamp calibrations is found to exceed that for outdoor global normal calibrations. Correction factor sensitivities to resolution in spectral response data, and temporal instability of a Xenon arc-lamp solar simulator are discussed. Correction factor uncertainty is found to rise with correction factor magnitude.<<ETX>>