We analyzed the influence of parameters of deep levels in the bulk and conditions on the surface on transient charge responses of semi-insulating samples (CdTe and GaAs). We studied the dependence on the applied bias step used for the experimental evaluation of resistivity in contactless measurement setups. We used simulations based on simultaneous solutions of 1D drift diffusion and Poisson's equations as the main investigation tool. We found out that the resistivity can be reliably determined by the transient contactless method in materials with a large density of deep levels in the bulk (e.g., semi-insulating GaAs) when the response curve is described by a single exponential. In contrast, the materials with the low deep-level density, like semiconductor radiation detector materials (e.g., CdTe, CdZnTe, etc.), usually exhibit a complex response to applied bias, depending on the surface conditions. We show that a single exponential fit does not represent the true relaxation time and resistivity, in this case. A two-exponential fit can be used for a rough estimate of bulk material resistivity only in a limit of low-applied bias, when the response curve approaches a single-exponential shape. A decreasing of the bias leads to a substantially improved agreement between the evaluated and true relaxation time, which is also consistent with the approaching of the relaxation curve to the single-exponential shape.
Keywords: CdTe; contactless resistivity; nonexponential response; surface defects.