Scalable synthesis, characterization, and DFT-machine learning modelling of Ti₃C₂ MXene quantum dots

Two-dimensional Ti₃C₂T _x MXenes exhibit exceptional electrical conductivity and tunable surface chemistry; however, reduction to zero-dimensional quantum dots (MQDs) affords quantum confinement and enriched functionalization, modulating the optoelectronic and electrochemical properties. In this research study, we developed a protocol for laboratory-level scale-up synthesis using TMAOH, yielding high-quantum-dot production with enhanced surface functionalization. XRD analysis confirmed quantum confinement, with the crystallite sizes sharply reducing from 8.68 nm (nanosheets) to ∼2 nm (MQDs). FTIR spectroscopy and XPS analysis revealed increased oxygen- and hydroxyl-functional groups replacing fluorine terminations on the MQD surfaces. Raman and photoluminescence spectroscopy reflected the combined influence of quantum confinement and surface termination chemistry, manifesting as electronic bandgap widening (1.77 to 2.32 eV), modified vibrational dynamics, and excitation-dependent PL emission. DLS and BET analyses documented size reduction, enhanced dispersion, and an increase in surface area from 31.8 to 140.6 m² g^-1 in MQDs. Electron microscopy validated preserved crystallinity and nanoscale morphology. Electrochemical studies showcased the superior charge transfer, stability, and catalytic activity of MQDs over MXene nanosheets, which was also supported by AFM surface roughness and Bode plot EIS analyses. A detailed DFT investigation quantified the metallic behaviour of 2D MXene with a near-zero bandgap versus the semiconducting behaviour of MQDs with a discrete 0.25 eV direct bandgap due to quantum confinement and oxygen termination. Machine learning models, trained on multimodal descriptors combining experimental and synthetic data, were used as an exploratory screening tool to identify candidate sensor-efficiency predictors, highlighting crystallite size, surface roughness, oxygen content, and particle size as the dominant influencers. MQDs consistently outperformed MXene nanosheets in sensing performance, attributed mechanistically to their nanoscale morphology and surface chemistry. Our work primarily focuses on the first scalable MQD synthesis correlating to the technoeconomic model, which supports a preliminary DFT-ML framework linking dimensionality to next-generation sensing applications.