Holistic yield modeling, top-down loss analysis, and efficiency potential study of thin-film solar modules

The world's increasing energy consumption and associated CO2 emissions necessitate the development and widespread adoption of sustainable energy technologies. Thin-film solar modules offer a promising solution due to their high cell efficiencies, short energy payback times, and potential for low-cost mass production. However, to compete with silicon-based technologies, thin-film technologies need to achieve comparably high efficiencies. Computer-aided modeling approaches provide a powerful tool for analyzing internal physical processes and optimizing module performance. While cell-level simulations are common, a holistic approach considering the entire module under real-world conditions is crucial for achieving high energy yields. This requires integrating multiple physical levels: optoelectronic behavior of semiconductor devices, current conduction, and operating conditions that often deviate from standard testing conditions (STC). This paper describes a novel simulation platform that bridges the gap between cell simulation and system design software by integrating these levels, enabling a comprehensive understanding of losses and optimization opportunities.

Existing research extensively uses computer simulation to analyze thin-film solar devices, employing electrical simulations and drift-diffusion models to study single cells. However, for a complete picture, simulations need to move beyond single-cell laboratory tests and encompass the behavior of entire modules in the field. Studies focusing on module-level losses often provide a breakdown of loss mechanisms but lack the comprehensive, integrated approach presented here. Previous attempts to model modules have often relied on separate simulations for various aspects, lacking the seamless integration offered by this approach which explicitly links material parameters to final energy yield predictions. The Shockley-Queisser model provides a theoretical efficiency limit, but this study advances further by quantifying actual losses within a working module operating under real conditions.

The authors developed a unique simulation platform that integrates three distinct simulation levels: 1. **Optical Simulation:** A modified transfer-matrix method (TMM) accounts for the optical properties of the module, including the effects of reflection, absorption, and interference from rough interfaces. This model uses complex refractive indices for all layers, determined experimentally or from literature. The CIGS absorber layer's absorption coefficient is sourced from literature, showing good agreement with measured external quantum efficiencies. 2. **Electronic Semiconductor Simulation:** A drift-diffusion model within a one-dimensional finite element method simulates the p-n junction behavior, including carrier transport and recombination. Material parameters for the CIGS and other layers (donor and acceptor densities, electron and hole mobilities) were obtained from literature and refined using a Reverse Engineering Fitting (REF) procedure, leveraging measured module I-V characteristics. 3. **Electrical Module Simulation:** A quasi-three-dimensional finite element method, solving Poisson's equation using a Delaunay-triangulated mesh, accounts for electrical transport within the entire module, considering the effects of series resistance, shunts, and local maximum power point (MPP) mismatches. The model incorporates geometrical details like edge and interconnect areas. The REF procedure uses a gradient-free downhill simplex algorithm, optimizing material parameters to match the simulated module I-V curve to measured data. The methodology allows for bidirectional calculations: predicting energy yield from material parameters or inferring material parameters from measured module data. The simulation is applied to a 1.2 × 0.6 m² CIGS module from NICE Solar Energy GmbH. The model accounts for temperature and irradiance fluctuations using measured meteorological data obtained on a clear day (Sept 9, 2020) and a cloudy day (Sept 6, 2020). The simulated results were compared to measurements under various conditions.

The study achieved high accuracy in simulating the module's performance, demonstrating a coefficient of determination (R²) of 0.997 between the simulated and measured I-V characteristics under standard testing conditions (STC). The holistic model accurately predicted the module's power output for both clear and cloudy days, with the integrated daily yield showing remarkable agreement between simulation (698.9 Wh) and measurement (698.7 Wh). The top-down loss analysis allows detailed quantification of each loss mechanism. The largest loss was found to be recombination, highlighting the importance of improving material quality. The study also reveals a higher electron mobility in the CIGS layer (μ = 200 cm²/Vs) than commonly assumed (100 cm²/Vs). A sensitivity analysis identified key areas for potential improvements, such as reducing sheet resistance in the TCO contact layer, reducing optical absorption in the TCO and encapsulant, and reducing edge and interconnect areas. Specifically, it showed that improving these areas could increase the PCE from 14.27% to 17.9% without changing the CIGS absorber material. Temperature coefficients for open-circuit voltage (-2.0 mV/K) and fill factor (-0.06%/K) were determined and validated against literature values. The model showed that the open-circuit voltage (*V*_oc) remains relatively constant throughout the day, while the short-circuit current (*j*_sc) closely follows irradiance fluctuations. The fill factor (FF) shows variations due to irradiance and temperature effects.

The findings demonstrate the ability to accurately simulate thin-film solar modules under real-world conditions, enabling precise prediction of module power output and a comprehensive understanding of loss mechanisms. The high agreement between simulated and measured data validates the model's accuracy and reliability. The top-down loss analysis provides valuable insights for targeted research and development efforts. The sensitivity analysis identifies specific areas for improvement that could substantially enhance efficiency without requiring changes to the active absorber layer. The results emphasize the importance of holistic modeling for improving both efficiency and energy yield of thin-film modules. The accurate prediction of performance under varying conditions enhances the capacity for design optimization and predicting the performance of solar power plants in various locations. The backward calculation capability allows for determination of difficult-to-measure material parameters.

This study successfully demonstrates a three-staged simulation model that precisely predicts the performance of thin-film solar modules under realistic operational conditions. The model's accuracy is validated through comprehensive loss analysis and close agreement between simulated and measured daily energy yield. The integrated approach, linking optical, electronic, and electrical aspects of the module's behavior, provides unique insights for optimization. Future research should focus on improving the simulation's spatial resolution for temperature modeling and exploring more advanced optimization algorithms for refining the REF procedure.

The model assumes a constant equilibrium temperature across the module, which may not accurately reflect the temperature distribution in real-world conditions. A more sophisticated, spatially resolved thermal model could enhance the accuracy, particularly for highly variable temperature profiles. The study focuses on a specific CIGS module; further validation with other module types and technologies is needed to ensure generalizability.