Machine Learning Models for Accurately Predicting Properties of CsPbCl3 Perovskite Quantum Dots

TL;DR

This work addresses predicting CsPbCl3 PQD size, the $1S$ absorption peak, and PL from hot-injection synthesis parameters using six regression approaches. A literature-derived dataset from 59 papers was assembled, including inputs such as injection temperature, chloride and lead sources/amounts, Cs source/amount, ligand volumes, and derived ratios, with outputs size (nm), $1S$ absorption peak (nm), and PL (nm). SVR and NND emerged as the most accurate models, achieving train RMSE around $0.009$–$0.012$ and test RMSE around $0.34$–$0.47$, while other models ranged from moderate to lower accuracy. RF-based feature-importance indicated Cs and OA as key drivers for $1S$ absorption, and Pb and Cs influencing size and PL, with a Pearson heatmap showing strong correlation between $1S$ absorption peak and PL ($r \approx 0.66$). The results demonstrate the potential for machine-learning-guided design of PQDs, enabling more targeted synthesis with improved efficiency.

Abstract

Perovskite Quantum Dots (PQDs) have a promising future for several applications due to their unique properties. This study investigates the effectiveness of Machine Learning (ML) in predicting the size, absorbance (1S abs) and photoluminescence (PL) properties of $\mathrm{CsPbCl}_3$ PQDs using synthesizing features as the input dataset. the study employed ML models of Support Vector Regression (SVR), Nearest Neighbour Distance (NND), Random Forest (RF), Gradient Boosting Machine (GBM), Decision Tree (DT) and Deep Learning (DL). Although all models performed highly accurate results, SVR and NND demonstrated the best accurate property prediction by achieving excellent performance on the test and training datasets, with high $\mathrm{R}^2$ and low Root Mean Squared Error (RMSE) and low Mean Absolute Error (MAE) metric values. Given that ML is becoming more superior, its ability to understand the QDs field could prove invaluable to shape the future of nanomaterials designing.

Figures (14)

• Figure 1: Box plots for the Size, 1S abs, PL providing the data distribution, median, quartiles, and potential outliers for each variable.
• Figure 2: Parity plots of predicted vs. observed values for the Size, 1S abs and PL outputs of the $\mathrm{CsPbCl}_3$ PQDs using DT regression model
• Figure 3: Importance of the variables for 1S abs using the RF regression model.
• Figure 4: Performance metrics for the algorithm trained and tested in 1S abs output, with splitting based on random conditions or to minimize similarity for the training and test data.
• Figure 5: Pearson correlation shows a strong correlation between the three output targets