Impedance spectroscopy of unetched CdTe/CdS solar cells—equivalent circuit analysis

Abstract

A detailed study of electric and structural properties has been carried out on CdTe/CdS solar cells which deliberately were not subjected to etching by a nitric-phosphoric (NP) or bromine-methanol (Br-Me) acids, conventionally employed for the formation of Te-rich layer before back contacting. In the previous work [J. Appl. Phys. 101, 014505 (2007)] we have shown that cells that were not etched provide more extensive information on sample/material properties than the etched ones, as analyzed by admittance spectroscopy. Although seemingly being able to describe the distribution of defect energy levels, the admittance spectroscopy approach has a significant drawback because the underlying theoretical formulation does not take into account the frequency-dependent contribution from the back contact together with its influence on the trap contributions. In this work we use an alternative methodology for analysis of impedance data measured in dark conditions, which applies an equivalent circuit model to the experimental spectra. In particular, a complete model consisting of 10–12 elements is suggested, which describes all the sets of data taken at different temperatures, unambiguously separating the respective roles of p-n junction parameters, defect trap levels, back contact, as well as spatial inhomogeneities within the cell. It is essential that the values of the parameters used to describe ac response from trap levels and that from the back contact are found to be consistent with admittance and I-V measurements. In addition, the temperature dependence of the dark conductance (GJ) and capacitance (CJ) of the main p-n junction, as well as temperature dependence of back contact resistance (RB), were obtained and analyzed. It was found that GJ(T) follows exp(T/T0) behavior which is characteristic of temperature-assisted tunneling, while CJ(T) agrees well with values of the high-frequency capacitance of the cell CHF(T). The T dependence of RB is found to follow activation behavior defined by a Schottky barrier with a height of (0.545±0.015) eV, that being close to the value obtained from dark I-V measurements.