Scholarlink

<p>This study presents significant advancements in optimizing double perovskite material structures (d-PMS) using SCAPS-1D simulations and machine learning techniques. The simulation effectively modeled device performance, identifying key design parameters for efficiency enhancement. Nd-doped ZnO as an electron transport layer (ETL) demonstrated superior charge transport, higher carrier concentration, and reduced series resistance (Rs), leading to improved power conversion efficiency (PCE) of 31.87%. The optimized device exhibited enhanced voltage (Voc) of 1.21 V and current density (Jsc) value of 32.93 mA/cm<sup>2</sup>, indicating better charge carrier dynamics. Nd doping passivated oxygen vacancies, minimized deep-level trap states, and reduced recombination at the ETL/perovskite interface. The maximum PCE was achieved with the following optimized parameters: t<sub>active</sub> = 500&#xa0;nm, t<sub>ETL</sub> = 250&#xa0;nm, N<sub>Active</sub> = 1 × 10<sup>13</sup>&#xa0;cm<sup>−3</sup>, N<sub>ETL</sub> = 1 × 10<sup>15</sup>&#xa0;cm<sup>−3</sup>, Rs = 0 Ω&#xa0;cm<sup>2</sup>, Rsh = 1000 Ω&#xa0;cm<sup>2</sup>, WF = 4.9 eV, T = 300 K and Illumination Intensity of 1 sun. Machine learning models, particularly random forest and XGBoost, accurately predicted PCE based on material properties with MSE, R<sup>2</sup> value of 0.43, 0.9891 and 0.31, 0.9921 identifying critical factors effecting performance of device such as series resistance, shunt resistance, work function, and defect density. This study highlights the potential of integrating numerical simulations with machine learning for refining d-PMS design, paving the way for efficient and commercially viable solar energy 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

Predictive modeling of CsFABiCuI6-based PSC with Nd-doped ZnO as ETL using machine learning and numerical simulation

  • 期刊论文

  • 发表日期：

    2025年4月26日

查看全文
  • Ghazi Aman Nowsherwan,
  • Saira Riaz,
  • Shahzad Naseem

摘要

This study presents significant advancements in optimizing double perovskite material structures (d-PMS) using SCAPS-1D simulations and machine learning techniques. The simulation effectively modeled device performance, identifying key design parameters for efficiency enhancement. Nd-doped ZnO as an electron transport layer (ETL) demonstrated superior charge transport, higher carrier concentration, and reduced series resistance (Rs), leading to improved power conversion efficiency (PCE) of 31.87%. The optimized device exhibited enhanced voltage (Voc) of 1.21 V and current density (Jsc) value of 32.93 mA/cm2, indicating better charge carrier dynamics. Nd doping passivated oxygen vacancies, minimized deep-level trap states, and reduced recombination at the ETL/perovskite interface. The maximum PCE was achieved with the following optimized parameters: tactive = 500 nm, tETL = 250 nm, NActive = 1 × 1013 cm−3, NETL = 1 × 1015 cm−3, Rs = 0 Ω cm2, Rsh = 1000 Ω cm2, WF = 4.9 eV, T = 300 K and Illumination Intensity of 1 sun. Machine learning models, particularly random forest and XGBoost, accurately predicted PCE based on material properties with MSE, R2 value of 0.43, 0.9891 and 0.31, 0.9921 identifying critical factors effecting performance of device such as series resistance, shunt resistance, work function, and defect density. This study highlights the potential of integrating numerical simulations with machine learning for refining d-PMS design, paving the way for efficient and commercially viable solar energy technologies.