Intelligent Design of Photothermal Nanomaterials: From Laser Synthesis Optimization to Solar Energy Harvesting Systems

This thesis explores the sustainable synthesis of photothermal nanoparticles (NPs) using laser ablation in liquid (LAL) combined with data-driven approaches. Although noble metal NPs such as Au and Ag show excellent light-to-heat conversion properties, the high cost and limited scalability restrict their application. Conventional chemical reduction generates waste and lacks purity, while LAL, though green and versatile, suffers from low productivity, limited phase control, and empirical optimisation. These gaps motivate the integration of data-driven predictive modelling with experimental synthesis. The overarching aim is to establish a scalable and environmentally sustainable pathway for NP synthesis. Specific objectives are to enhance LAL productivity, enable rational control of NP phase and morphology, and develop cost-effective alternatives to noble metals for photothermal applications. In particular, an artificial neural network was trained on Au NP synthesis data, enabling inverse prediction of laser parameters and increasing productivity to 0.455 g/h, surpassing recent multi-beam systems in efficiency and cost performance. For Cu-based systems, machine learning (ML) models achieved R² = 0.9 in predicting oxidation states, guiding the selective synthesis of metastable Cu(I) nanocrystals. Morphology studies showed that Au nanocorals outperformed other gold NP shapes, reaching surface temperatures of 215 °C and delivering 0.244 V in photo-thermoelectric devices. Finally, a design-of-experiment (DOE) guided model identified optimal conditions for synthesising Fe–Mn–B nanosystems, achieving solar water evaporation rates of 2.40 ± 0.05 kg·m–²·h–¹, exceeding those of Au NPs while maintaining stability and low cost. In addition to advancing our knowledge about the optimisation of NP productivity, phase, morphology, and range of photothermal earth-abundant materials, this work also demonstrates that coupling LAL with data-driven intelligent methods effectively transform NPs synthesis from empirical to predictive, enabling scalable, green, and cost-effective production. More broadly, it establishes a general strategy for accelerating the discovery of functional nanomaterials for energy and environmental applications.