The influence of the excitation intensity (density of injected carriers p(i)) on the time-resolved photoluminescence (PL) of the n-GaAs-electrolyte interface was studied using computer simulation. Simulations taking a photocurrent into account were carried out using the two-dimensional semiconductor analysis package (TOSCA) for n-GaAs (n(0) = 5 x 10(17) cm(-3)) ranging from low level injection (p(i) < n(0)) to the high injection (p(i) > n(0)) under depletion layer conditions. It was found that the PL decay time depends on the injection level (p(i)/n(0)) in a characteristic manner. For low injection levers (p(i) < n(0)), a lower limit of PL decay time is observed over a wide range of excitation intensities if the upward band bending is high in relation to the density of injected carriers. Depending on the relation between the initial band bending and photocurrent, the PL decay time as a function of the injection level may show a maximum. This behavior of the PL decay time can be explained by using spatial and time-resolved profiles of the carrier densities and the potential calculated by TOSCA during and after the excitation pulse. It is possible to roughly estimate the band bending and the exchange photocurrent density from the injection-level at the maximum of the PL decay time and from the shape of the curve.