Quantifying V_oc Loss Pathways in Polycrystalline Solar Cells

The power conversion efficiency of thin film photovoltaic (PV) devices has improved considerably over the past few decades, but thermodynamic limits and practical limits determined by device simulations suggest that there is yet significant room for improvement. In addition, the typically reported efficiency of champion devices represents the tail end of a wide distribution; narrowing that distribution is critical to the continued success of the thin film PV. An ongoing challenge is determining exactly where in the devices the greatest losses are occurring. Even in crystalline devices, identifying and quantifying the various loss pathways requires multiple test structures and characterization techniques. The difficulty is further exacerbated by the polycrystalline nature of thin film PV due to the presence of grain boundaries, incommensurate interfaces, a wide array of bulk deep defects, large band tails, various crystallite facets, nonuniformities, and uncertain dopant activation energies.<br/> <br/>The goal of this work is to explore the possibility of developing a universal method for independently quantifying V_oc losses due to front interface, back interface, and bulk recombination mechanisms. If a universal approach is not feasible, then best practices and general guidelines will be developed. From a practical perspective, the scope of work considers determination of three key parameters: i) the effective front surface recombination velocity; ii) effective back surface recombination velocity; and iii) effective bulk lifetime. These parameters, while not exhaustive, are critical for diagnosing the regions of most significant performance losses and providing guidance for further improvement. Our approach is based on considering combinations of characterization techniques, test structures, and numerical simulations to minimize the coupling between the key parameters. Hypothesis testing of such decoupling will be done by employing numerical simulation of characterization techniques such as current-voltage as a function of temperature and intensity (JVTi), quantum efficiency (QE), capacitance-voltage (CV), time resolved photoluminescence (TRPL), and others. The role of band tails in V_oc loss, such as Urbach tails, will also be considered through analysis of sub-band gap features in photoluminescence (PL) data. Numerical simulations are conducted using the finite element method to solve to the semiconductor equations with time and temperature dependence and customized photogeneration scenarios. Cd(Se,Te) PV devices will be used as specific case study.<br/> <br/>One approach considered so far employs a set of test structures, TRPL, and simulations in a sequential procedure to minimize the number of fitting parameters in each step. Three test structures are required: i) well-passivated double heterostructures (DHs) of various thicknesses (e.g. Al₂O₃/Cd(Se,Te)/Al₂O₃); ii) DHs of various thicknesses with one contact emulating the device back contact (e.g. Al₂O₃/Cd(Se,Te)/b.c); and iii) a complete device. Initial results from a thickness series of DHs quantified the front and back surface recombination velocities and the bulk lifetime. A challenge to this approach is maintaining consistent absorber properties while varying the DH thickness and between DHs and devices. Modifications to this approach and additional methods will be explored to spatially quantify recombination and identify regions of V_oc loss.