Scholarlink

<p>This study focuses on optimizing the performance of Cs<sub>2</sub>HgPdI<sub>6</sub>-based perovskite solar cells (PSCs) by integrating machine learning (ML) models with the SCAPS-1D simulator. A thorough dataset was generated to analyze the effect of important parameters, counting variations in the electron transport layer (ETL), hole transport layer (HTL), absorber thickness, as well as the effects of defects, doping, and impurities in the Cs<sub>2</sub>HgPdI<sub>6</sub> absorber. An artificial neural network (ANN) model was developed, demonstrating high predictive accuracy, with statistical analyses revealing a strong correlation between predicted and actual values. The Pearson correlation coefficient of 0.950 confirmed the model’s reliability in predicting power conversion efficiency (PCE). SCAPS-1D simulations, guided by ANN-optimized inputs, identified an optimal device architecture comprising a 100&#xa0;nm C<sub>60</sub> (ETL), an 800&#xa0;nm Cs<sub>2</sub>HgPdI<sub>6</sub> absorber, and a 500&#xa0;nm Cu<sub>2</sub>O (HTL), achieving a PCE of 30.61%. These findings highlight the effectiveness of combining ML techniques with conventional simulation tools to accelerate PSC design and optimization. Furthermore, the results provide a strong foundation for experimental validation and represent a significant advancement toward scalable, high-efficiency photovoltaic technologies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hybrid machine learning and SCAPS-1D simulation for the design and performance optimization of Cs2HgPdI6-based perovskite solar cells

  • 期刊论文

  • 发表日期：

    2026年6月30日

查看全文
  • Okba Saidani,
  • Abderrahim Yousfi,
  • Yehya Belhadad,
  • Abdsamed Chenouf,
  • Anis Chouya,
  • Girija Shankar Sahoo

摘要

This study focuses on optimizing the performance of Cs2HgPdI6-based perovskite solar cells (PSCs) by integrating machine learning (ML) models with the SCAPS-1D simulator. A thorough dataset was generated to analyze the effect of important parameters, counting variations in the electron transport layer (ETL), hole transport layer (HTL), absorber thickness, as well as the effects of defects, doping, and impurities in the Cs2HgPdI6 absorber. An artificial neural network (ANN) model was developed, demonstrating high predictive accuracy, with statistical analyses revealing a strong correlation between predicted and actual values. The Pearson correlation coefficient of 0.950 confirmed the model’s reliability in predicting power conversion efficiency (PCE). SCAPS-1D simulations, guided by ANN-optimized inputs, identified an optimal device architecture comprising a 100 nm C60 (ETL), an 800 nm Cs2HgPdI6 absorber, and a 500 nm Cu2O (HTL), achieving a PCE of 30.61%. These findings highlight the effectiveness of combining ML techniques with conventional simulation tools to accelerate PSC design and optimization. Furthermore, the results provide a strong foundation for experimental validation and represent a significant advancement toward scalable, high-efficiency photovoltaic technologies.