Scholarlink

<p>This study numerically investigates the potential application of cesium-based lead-free double perovskite solar cells. We systematically correlated bulk defect density, interface defect states, carrier lifetime, diffusion length, and band alignment effects within a unified optimization framework. For this, we investigated electrical parameters by optimizing the thickness, total defect density (N<sub><i>t</i></sub>) and total interface defects (N<sub><i>tf</i></sub>) of the absorber layer under standard AM 1.5G solar illumination light using SCAPS-1D software. The novel configuration Al/FTO/IGZO/Cs<sub>2</sub>AgBiBr<sub>6</sub>/GQDs/Au achieved power conversion efficiency (PCE) of 22.64%, short-circuit current density (J<sub>sc</sub>) of 24.8376&#xa0;mA/cm<sup>2</sup>, open-circuit voltage (V<sub>oc</sub>) of 1.1177, and fill factor (FF) of 81.57% at an optimal level. Effect of thickness of absorber layer, N<sub><i>t</i></sub> and N<sub><i>tf</i></sub> optimization were analysed. In order to justify our results, the training and testing of the data of electrical output parameters were done to predict the efficient performance of the solar cells by the machine learning (ML) algorithm. We employed multiple machine learning algorithms to predict perovskite solar cell efficiencies using interface defect density and absorber-layer thickness as key inputs. The models, particularly K-Nearest Neighbors (KNN) and ensemble methods, deliver near-perfect accuracy, demonstrating their capability to accelerate device optimization and reduce dependence on conventional iterative fabrication approaches. We have compared the performance of proposed solar cell not only with previous work but also before and after optimization. This research’s practical design and significant findings may lead to an affordable lead-free double perovskite solar cell made of Cs<sub>2</sub>AgBiBr<sub>6</sub>.</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

Machine learning assisted SCAPS-1D simulations for performance optimization of cesium-based double perovskite solar cells

  • 期刊论文

  • 发表日期：

    2026年5月7日

查看全文
  • Abhishek Kumar Srivastava,
  • Md. Ali,
  • Manisha Bajpai,
  • C. K. Pandey,
  • Ramesh Sharma

摘要

This study numerically investigates the potential application of cesium-based lead-free double perovskite solar cells. We systematically correlated bulk defect density, interface defect states, carrier lifetime, diffusion length, and band alignment effects within a unified optimization framework. For this, we investigated electrical parameters by optimizing the thickness, total defect density (Nt) and total interface defects (Ntf) of the absorber layer under standard AM 1.5G solar illumination light using SCAPS-1D software. The novel configuration Al/FTO/IGZO/Cs2AgBiBr6/GQDs/Au achieved power conversion efficiency (PCE) of 22.64%, short-circuit current density (Jsc) of 24.8376 mA/cm2, open-circuit voltage (Voc) of 1.1177, and fill factor (FF) of 81.57% at an optimal level. Effect of thickness of absorber layer, Nt and Ntf optimization were analysed. In order to justify our results, the training and testing of the data of electrical output parameters were done to predict the efficient performance of the solar cells by the machine learning (ML) algorithm. We employed multiple machine learning algorithms to predict perovskite solar cell efficiencies using interface defect density and absorber-layer thickness as key inputs. The models, particularly K-Nearest Neighbors (KNN) and ensemble methods, deliver near-perfect accuracy, demonstrating their capability to accelerate device optimization and reduce dependence on conventional iterative fabrication approaches. We have compared the performance of proposed solar cell not only with previous work but also before and after optimization. This research’s practical design and significant findings may lead to an affordable lead-free double perovskite solar cell made of Cs2AgBiBr6.