The impact of spectral variation on the thermodynamic limits to photovoltiac energy conversion

Research into the fundamental limitations to photovoltaic power conversion has historically used a single predetermined set of conditions to define device performance limitations. This fails to account for the many variables involved in real-world situations. Previous work describing thermodynamic losses in solar energy conversion has typically used an analytical approach, precluding the use of real-world spectra. This paper describes a model which marries the advantages of the analytical approach with a numerical detailed balance calculation, enabling analysis of maximum attainable power conversion efficiency and associated loss mechanisms in photovoltaics under more representative conditions.

Input spectra in the model are treated as separate beam and diffuse components, both in terms of power and subtended angle. Differences in conversion show that diffuse light is effectively under maximum concentration. This does not result in an efficiency gain since the equivalent energy is instead accounted for in the Carnot loss. The Carnot limit for the diffuse portion of the spectrum is therefore lower than that for direct light.

Simulated hourly “clear sky” spectra across a year were analysed for five geographically disparate locations. Results showed that at higher latitudes narrower band gap devices have a similar maximum efficiency to those with wider band gaps, whilst at lower latitudes wider band gap devices have a slightly higher maximum efficiency. This is compounded by increased irradiance at lower latitudes. Irrespective of band gap, annual energy conversion shows little variation at lower latitudes, with greater conversion in summer being offset by reductions in winter at higher latitudes.