Workshop Structure
📅 Day 1: ZnO Characterisation, Defect Engineering & Heterojunction Design
Core Objective: Understand how ZnO structure, morphology, optical properties, defects, and heterojunctions influence photocatalytic performance.
Topics Covered
- ZnO crystal structure, morphology, and structure–property relationships
- XRD-based phase and crystallite-size analysis
- SEM/ImageJ-based particle-size and morphology analysis
- UV–Vis/DRS analysis and optical band-gap estimation
- Oxygen vacancies, defect engineering, and doping strategies
- ZnO-based Type-II, Z-scheme, and S-scheme heterojunctions
- Preparation of structural and material descriptors for computational analysis
🛠️ Hands-on Lab
Analyse ZnO structural and microscopy data, estimate particle/crystallite size and band gap, and prepare material descriptors for further analysis.
Hands-on Workflow:
ZnO Structure → XRD/SEM Analysis → Band Gap → Defect & Heterojunction Interpretation → Material Descriptors
🧰 Tools Covered: Materials Project, VESTA, ImageJ, Python, Google Colab
📅 Day 2: Emerging Pollutant Degradation & Photocatalytic Kinetic Modelling
Core Objective: Analyse ZnO-mediated pollutant degradation, understand reaction mechanisms, and convert experimental results into an ML-ready dataset.
Topics Covered
- Semiconductor photocatalysis and electron–hole mechanisms
- Reactive oxygen species and pollutant-degradation pathways
- Dyes, antibiotics, pharmaceuticals, pesticides, and emerging contaminants
- Effects of pH, catalyst dosage, concentration, irradiation time, and light conditions
- Degradation efficiency and pseudo-first-order kinetic modelling
- Rate-constant estimation and experimental-condition comparison
- Mineralization, catalyst stability, recyclability, and photocorrosion
- Preparation of catalyst, pollutant, and process descriptors for machine learning
🛠️ Hands-on Lab
Process photocatalytic data, calculate degradation efficiency and kinetic parameters, compare operating conditions, and build an ML-ready research dataset.
Hands-on Workflow:
Experimental Data → Cleaning → Degradation Efficiency → Kinetic Modelling → Rate Constant → Descriptor Preparation → ML Dataset
🧰 Tools Covered: Python, pandas, NumPy, matplotlib, PubChem, WebPlotDigitizer, Google Colab
📅 Day 3: Explainable Machine Learning & Photocatalyst Optimization
Core Objective: Build interpretable machine-learning models to predict photocatalytic performance and identify optimized catalyst and experimental conditions.
Topics Covered
- Materials informatics and feature engineering for photocatalysis
- Material, pollutant, and experimental descriptors
- Data preprocessing, train-test splitting, and cross-validation
- Random Forest and XGBoost regression models
- Model evaluation using R², MAE, and RMSE
- SHAP-based feature importance and explainable AI
- Prediction of degradation efficiency under new conditions
- Data-driven parameter screening and photocatalyst optimization
- Introduction to inverse design and DFT–ML-assisted catalyst discovery
🛠️ Hands-on Lab
Train and validate ML models, generate SHAP explanations, identify key performance drivers, and screen conditions for improved photocatalytic efficiency.
Hands-on Workflow:
Photocatalysis Dataset → Feature Engineering → RF/XGBoost → Validation → SHAP → Prediction → Optimization
🧰 Tools Covered: Python, pandas, scikit-learn, XGBoost, SHAP, matplotlib, Google Colab
