Time-dependent coupled optical and electric modeling of dye-sensitized solar cells

We present a time-dependent coupled optical and electrical through-plane model of dye-sensitized solar cells. The optical model is based on a ray-tracing algorithm and accounts for coherent and incoherent optics [1,2]. The electrical model accounts for the generation, transport and recombination of conduction band electrons. The charge generation rate profile derived from the optical model serves as a source term in the continuity equation for conduction band electrons. The time-dependent part of the model is based on linear perturbation theory and allows to simulate time-dependent perturbations around an arbitrary stationary state of the dye-sensitized cell. We extract model parameters from measurement data of small perturbation photocurrent and photovoltage decays. The time-dependent experiments are performed on test cells by varying the white bias light intensity and the TiO2 film thickness. The light is incident from the front- or from the backside of the cell. The photovoltage decays are measured at open-circuit and the photocurrent decays at short-circuit using a low intensity red (642 nm) perturbation light pulse. The time-dependent model explicitly takes into account an exponential distribution of trap states. We assume that during the time-dependent perturbation, the quasi-static approximation is satisfied [3]. First, the relevant model parameters of the steady-state (e.g. electron injection efficiency and electron diffusion length) are determined by comparing the measured quantum efficiencies for illumination from the front- and backsides to the calculated ones [2,4]. In a second step, we simulate the photocurrent and photovoltage decays for the whole set of experiments and determine the values of the model parameters that reproduce the measured quantities in the best way. Discrepancies between measurement and simulation appear in particular for the illumination from the back side at low intensity. This may indicate the presence of additional recombination pathways, which are currently not taken into account in our model (e.g. recombination from surface states) [5,6]